Quarterly Journal of Economics | Macro Paper Warehouse

Codification, Technology Absorption, and the Globalization of the Industrial Revolution

Thu, 01 Jan 2026 00:00:00 +0000

Layer 1: Overview

Research question and motivation: Why did the First Industrial Revolution (IR) spread to Meiji Japan—and to essentially no other non-Western country—during the first wave of globalization? The paper tests Mokyr’s hypothesis that “technical literacy,” i.e., the codification of engineering, commercial, and industrial knowledge in the local vernacular, was a necessary condition for absorbing IR technologies. The motivating puzzle: after opening to trade (1858) and the Meiji Restoration (1868), 80% of Japanese exports were still primary products as late as ~1883 and real per capita GDP growth was only 0.6%/yr (1870-1883/85); then in a brief 13-year window (1883-1896) the manufacturing export share tripled and stabilized at around 60% of exports until WWII.

Data and setup: The authors build several novel datasets. (1) A cross-language measure of codification: scraping national/major libraries and WorldCat for technical books (agriculture, applied sciences, commerce, industry, technology) in 33 languages, 1500-1930. (2) “British Patent Relevance” (BPR): the cosine similarity (TF-IDF, unigrams+bigrams) between the digitized synopses of all British patents 1780-1852 (from Woodcroft 1857) and a hand-curated corpus of 460 English-language 19th-century technical manuals matched to SITC industries. BPR measures the world supply of codifiable IR knowledge by industry and is deliberately not based on what Japan translated (to avoid endogeneity). (3) The first harmonized, bilateral, industry-level trade dataset for the 19th century: 37 regions, 93 industries, quinquennial 1880-1910, built from reporting countries Japan, US, Belgium, Italy. Outcomes are annualized industry export growth ({1880,1885} to {1905,1910}) and, in robustness, productivity/comparative-advantage growth following Costinot et al. (2012) and Amiti-Weinstein (2018).

Main findings (with magnitudes): A Japanese industry with a one-standard-deviation higher BPR experienced annual export growth ~12 percentage points faster and annual productivity (comparative-advantage) growth ~1.2 percentage points faster (coefficients 0.121*** and 0.012***). Cross-sectionally, the BPR-growth relationship is positive and significant only for Japan and other codifying countries: for non-Japan regions the BPR coefficient is negative (-0.030***), while English-, French-, and the “top-4 codified” (English/French/German/Italian) regions show positive coefficients (0.042**, 0.032**, 0.078***), smaller than Japan’s. Low-income and Asian regions tend negative (divergence), not always significant. Time-series: regressing Japanese export growth from 1875 to varying end-years, the BPR coefficient is negative/significant in the 1875-1880 placebo window (Japan resembled the periphery), flips around 1890, and is positive and significant at 1% by 1895—coinciding with Japan’s catch-up in codification.

Mechanism and the Meiji “natural experiment”: In 1870, 84% of all technical books were in four languages (English, French, German, Italian); an Arabic-only reader had access to just 71 technical books. Japan started ordinary but codified explosively: technical-book growth jumped from 1.6%/yr (1600-1860) to 8.8%/yr (1870-1900); translated technical books rose from 8 (1500-1860) to 608 by 1900; Japanese technical books in the NDL grew from 706 (1880) to 2,823 (1890). State provision solved a public-goods/coordination problem: the government built English-Japanese dictionaries (ETSJ 1862/1866, FSEJ 1871) creating standardized Japanese jargon from Chinese glyphs, and 74% of identified technical-book translators (1870-1885) were government employees. Implication: low-cost vernacular access to technical knowledge was a necessary (not sufficient) condition for IR diffusion; where regions were linguistically/geographically distant from Western Europe, codification required state provision (a Gerschenkronian role for the state).

In depth

Q1. What is the identification strategy and the main threats to it?

Two-pronged. (1) Cross-sectional: regress region-industry export growth on BPR interacted with region-group dummies, with exporter fixed effects, exploiting that BPR is global (not Japan-specific) and that Japan was uniquely a codifier in the periphery. If codification is the mechanism, only codifying regions should show a positive BPR-growth link. (2) Time-series: exploit the sharp timing of Japanese codification (two well-demarcated periods—pre vs. post technical literacy in the 1880s) by estimating the BPR coefficient on Japanese export growth from 1875 to rolling end-years. The 1875-1880 window serves as a placebo (Japan not yet literate). Main threat is omitted-variable bias: that BPR is correlated with distance to the technology frontier, fundamental comparative advantage, Meiji institutional reforms, or industry steam-intensity. The cross-section addresses the ‘BPR matters everywhere’ and income/geography confounds; the timing addresses slow-moving confounds (literacy, Tokugawa culture, gradual reforms) since reforms like tax/banking/railroads were mostly in place by 1875, 15-37 years before the BPR effect appears.

Q2. How are the cross-section and time-series results distinguished from confounders empirically?

In the cross-section, income terciles (High/Medium/Low) and an Asia dummy are added: no region group replicates Japan’s positive pattern; the poorest and Asian regions show negative (divergence) coefficients. The placebo (1875-1880) yields a negative significant BPR coefficient for Japan itself—identical in sign to non-codifiers—then flips positive/significant by 1895, which conventional ‘opening to trade’ (1858) or ‘Meiji Restoration’ (1868) stories cannot explain because the effect appears 37 and 27 years later, respectively.

Q3. What heterogeneity is documented?

Japan’s BPR coefficient is larger (though not always significantly) than that of European codifiers, consistent with Japan having more to learn from British patents as a late industrializer. Among non-codifiers, low-income and Asian regions show negative BPR-growth relationships (divergence). Within codifiers, English- and French-speaking regions individually have positive but smaller and less precisely estimated coefficients; pooling the top-4 codified languages sharpens significance (0.078***). The time-series point estimates for Japan slowly decline after 1900 (not significantly), consistent with Japan shifting to Second Industrial Revolution technologies and becoming less reliant on older IR ones.

Q4. What robustness checks are run?

(1) Alternative patent corpora: results are nearly identical using British patents 1853-1879 (full text and AI-summarized) and US patents 1836-1860 and 1861-1879 (coefficients 0.121, 0.116, 0.111, 0.115), though later/US patents lower the R-squared, suggesting the 1780-1852 IR patents best explain Japanese export growth. (2) Productivity instead of exports (Costinot et al. 2012 comparative-advantage growth): qualitatively the same, 1.2 pp/yr for a 1-SD BPR increase, with deterioration in non-codifiers. (3) Confounders: controlling for British-colony status (insignificant) and industry steam-power intensity (French 1860s data) does not affect results. (4) Sample selection: dropping non-manufacturing sectors, excluding Asian destination markets, and dropping major export products (textiles, iron/metal) all leave the results intact, indicating broad-based change.

Q5. How does this paper relate to and differ from prior work?

It builds on Mokyr (2011) on ’technical knowledge’/‘access costs’ for European industrialization, extending it outside Europe with a Gerschenkronian twist (state as provider of the codification public good). It contributes to the technology-adoption-lags literature (Comin and Hobijn 2010; ~45-year average lags) by offering a friction explanation. It departs from prior Meiji studies (Sussman-Yafeh 2000; Tang; Morck-Nakamura; Bernhofen-Brown) that found banking, railroads, constitutional/monetary reforms had little measurable growth impact—offering codification as the resolution to ‘what drove the Meiji Miracle,’ consistent with Broadberry et al. (2025) dating Japan’s convergence to ~1890 driven by manufacturing productivity. It also extends the knowledge-codification literature (Dittmar 2011; Brown 2024; Abramitzky-Sin 2014) by linking codified vernacular knowledge directly to industry growth rather than indirect outcomes like city growth.

Q6. What are the policy implications and their scope conditions?

Public provision of technical knowledge in the vernacular can relax a critical bottleneck to industrialization, especially for regions linguistically/geographically distant from the technology frontier where the market undersupplies this public good. Scope conditions: codification is necessary but NOT sufficient. The Meiji model required complementary investments—language/jargon standardization, mass education for absorptive capacity (literacy >90% for army conscripts by 1909; ~40% of elementary class time on science), tacit-knowledge acquisition (2,400 hired foreigners providing 9,506 person-years of training; study-abroad missions), and tax capacity (1873 Land Tax Reform). China’s post-1949 codification under Zhou did not yield sustained growth until Maoist policies (Great Leap, Cultural Revolution) ended—’the exception that proves the rule.’

Q7. What external-validity evidence is offered beyond Japan?

The Meiji codification model was studied and transplanted by Park Chung Hee in South Korea (took power 1961; KIST; researcher counts rose sharply) and Zhou Enlai in China (premier 1949; Russian-language translation drive with USSR as the ‘Britain’). In 1950, Japan had ~70,000 technical books, China ~1,000, Korea <100; China surpassed 30,000 by the early 1960s. Korea’s per capita income clearly rises after Park; China’s codification did not translate into growth until after 1976. These are explicitly presented as suggestive/non-causal, plus appendix discussions of British India and Late Imperial Russia.

Q8. What are notable caveats and measurement choices?

BPR uses British 1780-1852 patent synopses and English manuals deliberately (Britain as IR leader; Japan hired British instructors and used British textbooks; avoids endogeneity from Japanese translation choices). It excludes tacit knowledge and secrecy-protected innovation by design. English codification is likely underestimated (British Library was un-scrapable after a 2023 cyberattack; Library of Congress used instead). German patents/trade data were excluded for coverage/reliability reasons. Linguistic-distance evidence on 1870/1913 GDP is explicitly not interpreted causally. The aggregate growth correlations for Japan, Korea, and China are described as suggestive, not causal.

Key Concepts

Codification (of technical knowledge): The creation of a means of transmitting engineering, commercial, and industrial knowledge—via language creation and written messages (manuals, textbooks, dictionaries)—that does not require direct contact between the knowledge originator and the recipient (Cowan and Foray 1997). In the paper’s sense it is a non-rival public good that the market undersupplies.

Technical literacy / technical knowledge: Following Stevens (1995) and Mokyr, the codified engineering, commercial, and industrial practices a practitioner needs to set up and run modern factory-based manufacturing; the paper measures it as the stock of vernacular technical books (agriculture, applied sciences, commerce, industry, technology), excluding theoretical/hard-science and non-firm subjects like medicine.

British Patent Relevance (BPR): An industry-level measure equal to the cosine similarity (TF-IDF weighted) between the vectorized text of British patent synopses (1780-1852) and the vectorized text of English technical manuals for that industry; it proxies how much codifiable IR knowledge a given industry stood to gain, and is independent of what was actually translated into Japanese.

Access costs: Mokyr’s (2011) term for the cost of obtaining usable technical knowledge; the paper argues vernacular codification (dictionaries, translations) lowered these costs, and that linguistic distance from English/Latin-Greek roots and physical distance from Europe raised them.

Technology absorption / absorptive capacity: The complementary conditions needed to use codified knowledge—prior language/jargon development, literacy and scientific training, and tacit knowledge—all of which the Meiji state invested in (dictionaries, compulsory education, ’live machines’/foreign instructors, study-abroad missions).

Defensive modernization (Gerschenkronian state role): The paper’s reading that an existential external threat aligned the Japanese elite behind aggressive state-led adoption of Western science, casting the state as the critical agent supplying the codification public good in late industrialization—a Gerschenkronian extension of Mokyr applied outside Europe.

Technology Sophistication Across Establishments

Thu, 01 Jan 2026 00:00:00 +0000

Layer 1: Overview

Research question and motivation: How sophisticated are the technologies establishments actually use, and how close are they to the world frontier? Traditional measures (since Ryan-Gross 1943 and Griliches 1957) characterize technology by the presence of one or a few advanced technologies, which (i) cover too few technologies and unrepresentative tasks, (ii) say nothing about how non-adopters produce or how far they are from the frontier, and (iii) ignore the intensity with which a technology is used. The authors argue intensity of use matters for explaining income divergence (Comin-Mestieri 2018), so they build a direct, comprehensive measure of technology sophistication.

Data and design: The authors construct “the grid,” a two-dimensional structure with business functions (BF) on the horizontal axis and technologies ranked by sophistication (simplest to world frontier) on the vertical axis. The grid spans 63 business functions (7 general business functions [GBF] relevant to all sectors plus 56 sector-specific business functions [SSBF] across 12 sectors) and a total of 305 technologies. More than 50 industry experts built and ranked the grid before survey administration. The grid is implemented in the Firm Adoption of Technology (FAT) survey, fielded 2019-2023 to 21,055 randomly selected establishments forming nationally representative samples (for establishments with 5+ workers) in 15 countries spanning all income levels (Korea, Poland, Croatia, Chile, Brazil-Ceara, Georgia, Vietnam, four Indian states, Ghana, Bangladesh, Kenya, Cambodia, Senegal, Ethiopia, Burkina Faso), representing a universe of about 2.1 million establishments. The median establishment has 9 workers (mean 34); 20% of workers hold a college degree, 17% are exporters, 18% are multinational-affiliated. FAT records, per BF, which grid technologies are used and which one is “most widely used.”

Two measures are built at the BF-establishment level on a [1,5] affine scale: MAX (sophistication of the most advanced technology used, reflecting adoption) and MOST (sophistication of the most widely used technology, reflecting both adoption and intensity/diffusion within the firm). Establishment-level measures are simple averages across in-house BFs. Cardinalization is validated three ways: linearity of the sophistication-productivity relationship; correlation above 0.98 with a z-score cardinalization (Bloom-Van Reenen 2007); and median correlation 0.95 with an independent productivity-based (“Q”) cardinalization for 18 BFs.

Main findings with magnitudes: (1) Establishments underutilize their most sophisticated adopted technology. In 63% of BFs where multiple technologies are used, MOST is not the most sophisticated available; the MAX-MOST gap appears in 62% of multi-technology BFs. (2) MAX and MOST are distinct upgrading processes: a one-unit rise in the number of technologies (NUM) raises MAX by 0.84 but MOST by only 0.25; MAX explains just 34% of within-establishment MOST variance. (3) Gaps are persistent, not transitory: only weakly related to age (cross-decile correlation -0.29; individual -0.01) and unrelated to time since adoption. (4) Gap frequency falls with income (country-level 51% in Korea to 83% in Burkina Faso; correlation -0.55 with per-capita income) and rises with input scarcity (low human capital, loan denial) and managerial mistakes (perception bias, family ownership, non-exporting). (5) Within-country dispersion in gaps (0.28) is about three times the between-country dispersion (0.09). (6) Establishment-level MAX and MOST average 2.6 and 2.0; both correlate with income (0.78 for MAX, 0.94 for MOST) and with size, human capital, management, exporter and multinational status. (7) Both productivity and profitability rise with sophistication, more strongly for MOST and for agriculture; the association is not smaller in low-income countries, contradicting the “appropriate technology” hypothesis.

In depth

Q1. What are MAX and MOST, and why are they conceptually distinct?

MAX_{f,j} is the sophistication of the most advanced grid technology establishment j uses in business function f; MOST_{f,j} is the sophistication of the most widely used technology in that function. Both lie in [1,5] with MAX >= MOST by construction, and both measure closeness to the world frontier. They are conceptually different: increases in MAX reflect adoption of a new (to the function) more sophisticated technology, whereas increases in MOST can reflect adoption OR the extension/intensification of an already-adopted technology — closer to Mansfield’s (1963) concept of intra-firm technology diffusion. The paper’s central empirical claim is that these are driven by distinct upgrading processes.

Q2. What is the identification strategy, and what does the paper NOT claim?

This is a descriptive/correlational paper, not a causal one. The authors explicitly state their data do not permit causal inference; the productivity, profitability, and characteristic associations are partial correlations from cross-sectional regressions with country and 2-digit sector fixed effects. The BF-level analyses (MAX-NUM, MOST-NUM, MAX-MOST) use establishment and function fixed effects to absorb establishment- and function-specific levels. The main ‘identification’ work is measurement validity, not causal identification.

Q3. How are MAX and MOST shown to be distinct upgrading processes empirically?

Three pieces of evidence. First, regressing MAX on NUM (number of technologies) with establishment and function FE yields a coefficient of 0.84 (s.e. 0.01) — near one-to-one — while regressing MOST on NUM yields only 0.25 (s.e. 0.01). Second, regressing MOST on MAX (with FE) shows MAX explains only 34% of within-establishment MOST variance, so MAX is not a sufficient statistic for MOST. Third, MAX and MOST have different distributions (MOST more skewed), different lifecycle profiles, different correlates, and different associations with productivity.

Q4. Is the MAX-MOST gap transitory or persistent, and how is this tested?

Persistent. Three exercises: (i) across age deciles the gap correlates only -0.29 with age (-0.01 at the individual level), with no clear lifecycle pattern by income or size except a decline only among large establishments aged 16+; (ii) the distribution of years since adopting a top-tier technology is similar for BFs with and without a gap, so time does not close it; (iii) splitting top-tier adopters into early vs. recent adopters yields similar MOST distributions. Together these confirm gaps persist long after adoption.

Q5. What are the two hypothesized drivers of MAX-MOST gaps, and what evidence supports each?

(1) Input constraints — scarcity of skilled labor or finance pushes firms to rely on simpler technologies operable by less-educated workers or needing less capital. Supported by the negative coefficient on human capital (college share) and the positive coefficient on the loan-denied dummy. (2) Managerial mistakes — poor management or biased self-perception of one’s own sophistication causes suboptimal underuse. Supported by positive correlations with perception bias and family ownership, and a negative correlation with exporter status (competitive pressure narrows the gap); the management z-score association is weak. Across subsamples, input scarcity is more prominent in low-income countries while managerial-mistake proxies are more salient among large establishments (likely from the complexity of managing scale).

Q6. What heterogeneity in technology sophistication is documented?

By income: country averages span 1.53 (MAX) and 1.01 (MOST); within-country dispersion (p80-p20) rises with income, more steeply for MOST (0.95 vs 0.33). By sector: agriculture shows greater cross-establishment dispersion in both MAX and MOST than manufacturing or services. Lifecycle: MAX rises gradually with age in all income/size groups, but MOST flattens beyond ~10 years in low-income countries and among small establishments. Size effects on MOST are stronger in high-income countries; on MAX they are similar across income levels. The performance-sophistication link is strongest in agriculture and weakest in services, and is not weaker in low- than high-income countries.

Q7. How much of the variation is across vs. within sectors, and why does that matter?

Following Syverson (2011), sector dummies explain only 14% (2-digit), 20% (3-digit), and 23% (4-digit ISIC) of cross-establishment variance in sophistication — comparable to their explanatory power for productivity (sales per worker). This implies sophistication variation reflects differences in the technologies used to perform similar tasks, not differences in what tasks/goods establishments produce.

Q8. What robustness and validation checks are run?

Cardinalization: linear approximation of the sophistication-productivity relation; correlation >0.98 with z-score cardinalization; median 0.95 (p25-p75: 0.90-0.98) with a productivity-based Q-cardinalization across 18 BFs; establishment-level baseline-vs-Q correlations of 0.90 (MAX) and 0.91 (MOST). Ranking validity: three-stage expert validation (functionality/integration/automation; novelty and cost; ChatGPT replication) on 14 BFs plus an independent relative-productivity exercise on 18 BFs. Data quality: response rates 15-86% (high for establishment surveys); no significant non-response differences in employment, sophistication, wages, or skill; a Kenya back-check pilot showing 80.6% consistency for technology-use reports; external validation against Korea (KED) and Brazil (RAIS) with cross-establishment correlations above 0.93 for sales/employment and 0.73 for labor productivity; ERP adoption in Korean manufacturing of 32% vs. 40% in Chung-Kim (2021). Establishment-level results are robust to controlling for the in-house fraction of functions.

Q9. How does this paper relate to and differ from prior work?

It generalizes the intra-firm diffusion literature (Mansfield 1963; Battisti-Stoneman 2003), which studied a handful of technologies in a few countries, by showing MAX-MOST gaps are widespread and persistent across 63 functions and 15 countries. It parallels Bloom-Van Reenen (2007) on management practices in method (expert rankings, survey scoring, z-scores) and finds supporting evidence for the Bloom-Sadun-Van Reenen (2012) technology-management complementarity. It differs from the US Advanced Business Survey / Acemoglu et al. (2022), which covered five frontier technologies, by being comprehensive and frontier-relative. It contributes new evidence to the agricultural productivity gap (Caselli 2005; Gollin-Lagakos-Waugh 2014) and to the appropriate-technology debate (Basu-Weil 1998; Acemoglu-Zilibotti 2001).

Q10. What are the policy implications and their scope conditions?

Because the sophistication-performance association is not smaller in low-income than high-income countries, advanced technologies appear ‘appropriate’ across income levels — challenging the appropriate-technology hypothesis that poor countries gain little from sophisticated technology. Policy should target not only adoption (MAX) but also the extension of use/intensity (MOST), since MOST is more strongly tied to productivity and profitability. Scope conditions: associations are correlational, not causal; samples are representative only for establishments with 5+ workers; coverage is the 12 surveyed sectors; and the cross-section cannot trace dynamics (the authors plan a longitudinal extension).

Q11. What do the descriptive technology-use patterns show about adoption behavior?

Establishments use about two technologies per function on average; 62.6% of functions use more than one and 28.3% use at least three. Leapfrogging/skipping is rare: among single-technology functions (37.4% of cases), 52.8% use the least sophisticated grid technology, so only about 18% of functions have fully skipped or abandoned simpler technologies. In 70.4% of multi-technology functions one technology used is the least sophisticated available, and sophistication gaps (non-contiguous use) occur in only 25% of functions (27% GBF, 17% SSBF; most common in payments 48%, business administration 34%, sales 28%). Firms thus typically retain dominated technologies rather than abandon them, which is why MAX proxies the full adoption history well. Only 16% of establishments use an ERP system (the most sophisticated business-administration technology).

Q12. Any notable caveats about the measures themselves?

MAX-MOST gaps are ordinal (cardinalization-free), but establishment-level MAX and MOST are cardinal and could be sensitive to the chosen cardinalization — addressed by the validation exercises. Establishment-level measures use only in-house functions (87% of relevant SSBFs and an overwhelming majority of GBFs are in-house; only 3.9% of GBFs not in-house), and results are robust to controlling for the in-house share. The survey deliberately avoided the words ’technology’ and ‘sophistication’ (using ‘methods’/‘processes’) to limit social-desirability bias.

Key Concepts

The grid: A two-dimensional structure mapping each key business function (horizontal axis, task-based) to the range of technologies that can perform it (vertical axis, ranked by sophistication from simplest to the world frontier). Spans 63 business functions (7 general + 56 sector-specific across 12 sectors) and 305 technologies, built and ranked by 50+ industry experts.

MAX: The sophistication (on a [1,5] affine scale) of the most advanced technology an establishment uses in a given business function. Increases in MAX reflect adoption of a technology new to that function; near one-to-one with the number of technologies used (coefficient 0.84).

MOST: The sophistication (on a [1,5] scale) of the most widely used technology in a business function. Changes in MOST reflect both adoption and the intensification/extension of already-adopted technologies — closer to Mansfield’s (1963) intra-firm diffusion than to adoption per se; only weakly tied to the number of technologies (coefficient 0.25).

MAX-MOST gap: A binary indicator equal to 1 when MAX > MOST in a function with multiple technologies in use — i.e., the most widely used technology is not the most sophisticated one adopted. Present in 62-63% of multi-technology functions, persistent over time, and associated with input scarcity, managerial mistakes, and lower productivity.

FAT survey: The Firm Adoption of Technology survey: a cross-section of 21,055 establishments forming nationally representative samples (5+ workers) in 15 countries (2019-2023), implementing the grid plus modules on financials, employment, management practices, and adoption barriers.

Appropriate technology hypothesis: In this paper’s usage, the claim (Basu-Weil 1998; Acemoglu-Zilibotti 2001) that establishments in poor countries underutilize sophisticated technologies because scarce human and physical capital limits the productivity gains those technologies embody. The paper’s finding that the sophistication-performance association is not smaller in low-income countries runs counter to this hypothesis.

"Compensate the Losers?" Economic Policy and the Origins of U.S. Partisan Realignment

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question. Why have less-educated voters in the United States abandoned the Democratic Party over recent decades? The paper argues that the Democratic Party’s evolution on economic policy — specifically its retreat from “predistribution” — is a central, previously understudied driver of partisan realignment by education.

Conceptual Framework. The authors distinguish between two categories of egalitarian economic policy: (1) predistribution — policies that alter the pre-tax-and-transfer earnings distribution, including job guarantees, minimum wage increases, union support, and protectionist trade policies (following Hacker 2011); and (2) redistribution — taxes and transfers. The paper’s central claim is that these two types of policy have sharply different educational gradients among voters, and that the Democratic Party moved away from predistribution beginning in the 1970s, triggering educational realignment.

Data and Methodology. The authors harmonize over 1,000 surveys (N ≈ 2.2 million observations) spanning 1942–2020, drawn from Gallup, ANES, GSS, CCES, and historical survey archives housed at iPoll/Cornell. Education is translated into a common metric (adjusted years of schooling) using Census data, controlling for sex, race, year, and birth cohort to address the changing selectivity of educational categories over time. Congressional roll-call data come from the Comparative Agendas Project (CAP). Campaign finance data come from FEC filings, Congressional hearing records, and watchdog sources. DLC membership data are compiled from official Democratic Leadership Council records (available for 1985, 1986, 1991, 1993, and 1997 onward) and DLC-aligned Congressional caucus lists. House election returns are taken from King and Palmquist (1997) at the minor-civil-division-group (MCDG) level (~60 units per Congressional district), matched to 1980 Census demographic data.

Main Findings.

Voter preferences (demand side): The educational gradient for predistribution is large and negative: averaged across the four predistribution questions (job guarantee, minimum wage, union support, trade protection), each additional year of education reduces support by 0.044 standard deviations (p < 0.001). A college graduate relative to a high school graduate supports predistribution 0.176 standard deviations less — equivalent to roughly half the average Democrat-Republican gap in predistribution support (which is 0.34 standard deviations). This gradient has been stable since at least the 1940s. By contrast, the educational gradient for redistribution (higher taxes on the rich, views on own taxes, welfare spending) is close to zero (summary β = 0.004, not distinguishable from zero in the full sample). The difference between the two gradients is statistically significant (p < 0.001). These results replicate in white-only samples. Notably, the educational gradient on social issues — measured across nine questions on racial attitudes, gender roles, sexual norms — is positive (more education predicts more liberal positions) but has been largely stable since the 1940s, not increasing, conditional on the long-run sample.

Party supply (supply side): Before 1976, predistribution topics accounted for roughly one-quarter of Democratic House roll-call votes when Democrats controlled the chamber. After 1976 (taking Jimmy Carter’s presidency as the start of the “New Democrat” era), this share falls by approximately nine to ten percentage points, while the redistribution share of votes holds steady. Between 1968 and 1980, the union share of total PAC donations to Democratic Congressional candidates falls from approximately 90 percent to 40 percent, coincident with 1970s campaign finance reforms that placed union and corporate PACs on equal legal footing and allowed corporations to exploit their naturally deeper pockets. Corporate PAC share of Democratic donations correspondingly rises from approximately 10 percent to 45 percent over the same period. In individual contributions to primary elections (data beginning in 1980), Democratic primaries rely on increasingly more-educated census tracts relative to Republican primaries; by 2018 Democratic primaries are financed from census tracts averaging 0.41 more years of education than Republican primaries (against a within-year standard deviation of 1.56 years).

The New Democrat/DLC faction: The authors identify the anti-predistribution faction through official DLC membership records and aligned caucus lists. DLC membership as a share of Democratic House seats grows from near zero in the mid-1970s to approximately half by the early 2000s. Roll-call voting analysis (N = 3,428,405 vote-observations) shows DLC members are more conservative than other Democrats overall, and especially so on predistribution: for a 10-percentage-point increase in the share of Republicans voting for a bill, the probability a DLC member votes in favor increases 36 percent more on predistribution bills than on other bills. DLC members show no differential conservatism on redistribution. They are also significantly more socially conservative — more likely than other Democrats to support the Defense of Marriage Act (by 16 pp), the Partial-Birth Abortion Ban (by 7 pp), and restrictive immigration bills (by 10 pp). DLC candidates receive significantly less from labor PACs and significantly more from corporate PACs, and draw their out-of-district individual donations from census tracts averaging more than 0.1 years more educated than non-DLC Democrats.

Voter reaction and the inflection point: Using the N ≈ 2.2 million partisan identification dataset, the authors estimate a structural break in the education-party identification gradient. From the 1940s through the mid-1970s, each additional year of education reduces the probability of identifying as a Democrat by approximately 3 percentage points. A Chow breakpoint test identifies 1976 as the inflection point. Since 1976, the gradient steadily rises; by 2000 it reaches zero; and today (as of the sample period end ~2020) each additional year of education increases Democratic identification by approximately 3 percentage points — an almost exact reversal. The breakpoint for Republican identification occurs later, in 1992, consistent with the Democratic agenda changing first. A Gallup prosperity question (“which party will better keep the country prosperous?”) shows a parallel pattern: controlling for views on parties’ economic performance explains approximately 44 percent of partisan realignment, interpreted as an upper bound on economic policy’s contribution.

Factional tests — hypothetical elections and actual results: In hypothetical general-election matchups from 1972–1992 Democratic primaries (in which most contests pitted a “New Democrat” against an “Old Democrat”), a voter with a college degree is roughly 3 percentage points more likely to vote Democratic when the candidate is a New Democrat rather than an Old Democrat. In 1980s actual House elections using MCDG-level data, DLC candidates out-perform other Democrats in more educated neighborhoods by a magnitude large enough to erase approximately 90 percent of the general Democratic underperformance in highly educated areas. Combining these estimates, the party’s shift toward the DLC accounts for a lower bound of approximately 20 percent, and an upper bound (from the prosperity question) of approximately 50 percent, of educational realignment.

Scope Conditions. The analysis focuses on the United States, 1942–2015 (with some post-2015 discussion in the conclusion). The faction analysis focuses on the Democratic side; Republican faction changes are discussed but not the primary focus. The paper is explicit that between 20–50 percent of realignment is explained, leaving room for other factors, including social issues. The analysis ends mostly before 2016 to avoid complications from the closure of the DLC in 2011 and shifting post-2010 party dynamics.

In depth

Q1. What is the paper’s central conceptual innovation, and how does it differ from prior realignment research?

The paper separates egalitarian economic policies into “predistribution” (pre-tax-and-transfer market interventions such as minimum wages, job guarantees, union support, and protectionism) and “redistribution” (taxes and transfers) and shows these two types have sharply different educational gradients. Prior work typically aggregated all economic policies into a single index, which the authors argue masks essential heterogeneity. By documenting that the educational gradient is large and negative for predistribution but close to zero for redistribution — a pattern stable since the 1940s — the paper reframes the “voting against economic interest” puzzle: less-educated voters leaving the Democratic Party may be responding rationally to changes in the supply of the type of economic policy they actually prefer.

The average coefficient on adjusted years of schooling across the four predistribution questions is -0.044 (p < 0.001), stable over eight decades. A four-year difference in education (high school vs. college) shifts an individual’s support for predistribution by 0.176 standard deviations in the conservative direction — about half the average Democrat-Republican gap in predistribution support (0.34 standard deviations). For social issues, the summary gradient is positive (+0.028, p < 0.001 for the full sample), but this gradient has been largely stable since the 1940s across nine social issue questions, not increasing over time. This stability undermines the interpretation that rising social liberalism among the educated is a new phenomenon driving realignment, at least through the supply of parties’ social positions.

Using the Comparative Agendas Project classification, predistribution topics (labor regulation, industrial policy, public works, trade) accounted for roughly one-quarter of all House roll-call votes during years Democrats controlled the Speakership before 1977. After 1977, this share falls by approximately 9–10 percentage points (a decline of nearly half from its pre-1977 share), and the decline is statistically significant (p < 0.001). The redistribution share of votes holds essentially constant. Party platform data from Hopkins et al. (2022) show a sharp decline in Democratic use of terms like “minimum wage,” “full employment,” and labor-relations language beginning in the 1970s and 1980s, while Republican platforms use these terms sparingly throughout.

Q4. How did 1970s campaign finance reforms change the financial composition of the Democratic Party?

Before the early 1970s, unions enjoyed substantially more freedom than corporations under separate legal regimes governing PAC donations; mid-1970s reforms placed them on equal legal footing, enabling corporations to exploit their deeper pockets. The union share of total PAC donations to Democrats fell from approximately 90 percent in 1968 to approximately 40 percent by 1980, while the corporate share rose from approximately 10 percent to 45 percent. For Republicans, both series barely changed: unions had never donated substantially to the GOP, and the corporate share rose only modestly (from approximately 70 to 80 percent). The authors note the rapid decline cannot be attributed to falling union density in the economy, since both union and corporate PAC donations grew in absolute terms during this period; the relative shift was the result of the regulatory change.

Q5. Who are the “New Democrats” / DLC, and when did they emerge?

The DLC officially operated from 1985 to 2011, but members who would join it began entering Congress in large numbers in the 1970s (“Watergate Babies” of 1974, “Atari Democrats”). The DLC grew to approximately half of all Democratic House seats by the early 2000s. Members were drawn from suburban, affluent districts; their founder Al From explicitly criticized all four predistribution policies the paper studies (minimum wage, job guarantees, unions, and protectionism). The breakpoint test on DLC share in Congress identifies 1975 as the pivotal year — one year before the 1976 inflection point in partisan identification.

Q6. How do DLC members vote differently from other Democrats, and how is this differential conservatism distributed across policy types?

In roll-call regressions (N = 3,428,405 observations, with roll-call fixed effects), a 10 pp increase in the Republican vote share for a bill increases the probability a DLC member votes in favor by 1.48 pp more than for other Democrats (baseline result for all bills). For predistribution-classified bills, this excess alignment with Republicans is 36 percent larger than for non-predistribution bills. Crucially, DLC members are no more conservative than other Democrats on redistribution-classified votes (the interaction with redistribution is near zero and insignificant). DLC members are also differentially more conservative on social issues, a result that proves useful in separating economic from social-issue explanations of realignment.

Q7. Do DLC members finance differently from other Democrats?

Yes. In primary elections, DLC candidates receive approximately 9.7 pp less of their PAC financing from labor unions and approximately 6.7 pp more from corporate PACs (with state fixed effects) relative to non-DLC Democrats. Out-of-district individual contributions to DLC primary candidates come from census tracts averaging more than 0.1 years more educated than those for non-DLC Democrats, while within-district contributions show no significant difference (0.060 years, insignificant). This pattern suggests educated out-of-district donors, rather than local constituency demands, drive DLC candidates’ anti-predistribution orientation.

Q8. When precisely did educational realignment in Democratic party identification begin, and what does the inflection-point analysis show?

Using N ≈ 2.2 million observations from 1,006 surveys, a Bai-Perron breakpoint test on the year-by-year education gradient in Democratic party identification identifies 1976 as the inflection point (with robustness to alternative specifications yielding breakpoints of 1978–1980 for white-only samples and unadjusted years of schooling). Before 1976, each additional year of education reduces the probability of Democratic identification by approximately 3 percentage points (a stable, significantly negative relationship since the 1940s). After 1976, the gradient steadily rises; it reaches zero around 2000 and today is approximately +3 percentage points per year of education — nearly an exact reversal of the baseline. The corresponding Republican inflection point occurs in 1992, about 16 years later, consistent with the Democratic Party’s agenda changing first.

Q9. How do hypothetical presidential matchup surveys test the DLC mechanism?

The authors identify six Democratic primaries from 1972–1992 where a “New Democrat” and an “Old Democrat” were the top two contenders (e.g., Hart vs. Mondale in 1984, Clinton vs. Brown in 1992). Gallup and other surveys asked all respondents — regardless of party — whom they would vote for if either the New or the Old Democrat faced the eventual Republican nominee. A voter with a college BA is approximately 3 percentage points more likely to vote for the Democrat when the candidate is a New Democrat versus an Old Democrat (the “difference in differences” of hypothetical vote shares). This holds after controlling for state × election fixed effects and in five of the six election cycles studied (the 1976 exception is attributed to Mo Udall’s low name recognition, with 28 percent of respondents unfamiliar with him in a May 1976 poll). The result is attenuated but remains marginally significant when excluding non-white respondents, consistent with New Democrats’ success with white voters due in part to their more conservative civil rights positioning.

Q10. What do actual House election results (MCDG-level data) show about DLC electoral performance by neighborhood education?

Using 1980s House returns at the MCDG level (~60 neighborhoods per Congressional district), the authors regress Democratic vote share on neighborhood years of education interacted with a DLC candidate indicator, with Congressional district fixed effects. More-educated neighborhoods generally depress Democratic vote share (reflecting the still-negative overall educational gradient in the 1980s), but DLC candidates dramatically out-perform other Democrats in educated areas: the interaction coefficient is positive and significant, and its magnitude is large enough to erase approximately 90 percent of the general Democratic underperformance in highly educated neighborhoods. This result is robust to including District × Year fixed effects (so the identification comes from within-election, cross-neighborhood variation) and to adding controls for share white and share under age 35.

Q11. How much of educational realignment can the paper’s mechanism account for, and how is this calculated?

Two bounding estimates are provided. Upper bound (~44–50%): controlling for a respondent’s view on which party is better for economic prosperity (from Gallup since 1950) explains approximately 44 percent of the change in the education-party identification gradient (specifically, the total difference in the unconditional gradient between the 1948–1967 baseline and 2001–2020 is 2.411 pp per year of schooling; after controlling for the prosperity question, the unexplained residual is 1.342 pp, leaving a share explained of 44.3 percent). Lower bound (~20%): the difference in the education gradient between matchups involving New versus Old Democrats in Table 4 (~0.75 pp) divided by the total realignment shift (~4 pp from pre-1976 to post-2008 for presidential voting) implies the faction shift accounts for at least approximately one-fifth of realignment. The authors interpret these as bounds because the prosperity question may partly capture party identification itself (upper bound concern), while the hypothetical matchup estimate misses the broader ideological shift not captured in a single election (lower bound).

Three alternative explanations are addressed. (1) Civil Rights: Regional analysis shows that educated white Southerners left the Democrats in the 1940s–1960s (not the 1970s), consistent with their realignment being driven by Democrats’ liberal turn on civil rights rather than economic policy. After the 1960s, the South follows all other regions in the pace of educational realignment. (2) Republican changes: The Republican party identification inflection point occurs in 1992, about 16 years after the Democratic inflection in 1976. Reagan elections in 1980 and 1984 do not appear to have differentially attracted less-educated voters (the “Reagan Democrats” were not differentially less educated). (3) Social issues: The New Democrats were actually more socially conservative than other Democrats (more likely to vote for DOMA, anti-abortion bills, restrictive immigration legislation), yet they disproportionately attracted educated voters. This internal inconsistency rules out a pure social-issues explanation for why educated voters preferred the DLC faction. (4) Religion: Flexibly controlling for religious affiliation explains essentially none of partisan realignment (Appendix Figure A.24).

Q13. What is the role of out-of-district individual donors in shifting Democratic Party positions?

Out-of-district primary donors are analytically important because they influence candidate supply without being able to vote in the election, isolating the “within-party” financial influence of educated supporters. By 1980, out-of-district primary donors to Democratic candidates already come from census tracts more educated than those for Republican candidates, even as local Democratic voters and within-district donors remain less educated than Republican counterparts. Democratic candidates also receive a substantially higher share of out-of-district contributions than Republican candidates — by almost 10 percentage points (Appendix Table A.7). Out-of-district donors thus represent a channel through which educated, anti-predistribution preferences are transmitted into the Democratic Party’s candidate supply before the electoral realignment is visible in vote totals.

Q14. Are predistribution policies becoming less popular overall, which might independently push Democrats away from them?

The paper tests this alternative in Appendix Table A.9 and finds no evidence that predistribution has become less popular relative to redistribution over time. Predistribution appears on average more popular than redistribution across the sample period. If anything, support for predistribution has held steady or slightly risen relative to redistribution over time, conditional on the paper’s survey harmonization. The stability of the educational gradient (shown in Appendix Table A.10 to be unchanged even using educational rank within cohort rather than raw years of schooling) further suggests the negative education-predistribution relationship is a relative, not absolute, phenomenon — consistent with rising average education and stable preferences by education rank.

Key Concepts

Predistribution: Policies that aim to change the distribution of earnings or income before taxes and transfers are applied. In this paper, this comprises government job guarantees, minimum wage increases, support for unions and collective bargaining, and protectionist trade policies. Distinguished from redistribution in that it operates on pre-tax market income rather than post-tax outcomes. The paper uses this term following Hacker (2011): “a focus on market reforms that encourage a more equal distribution of economic power and rewards even before government collects taxes or pays out benefits.”

Redistribution: Policies that change post-market income through the tax and transfer system, including higher taxes on the rich, views on own tax burden, prioritization of tax cuts, and transfers to the poor (welfare spending). In the paper’s usage, redistribution is analytically distinct from predistribution and has a near-zero educational gradient, in contrast to predistribution’s strongly negative gradient.

Educational Gradient: The coefficient on adjusted years of schooling in a regression of an outcome variable (policy preference or partisan identification) on education, estimated separately by time period. The paper’s core finding is that the educational gradient for predistribution is stably negative (approximately -0.044 per year of schooling over the full sample), while the gradient for redistribution is close to zero, and the gradient for Democratic party identification shifts from approximately -0.03 to +0.03 per year of schooling between the 1940s and 2020.

New Democrats / DLC (Democratic Leadership Council): An explicitly anti-predistribution faction within the Democratic Party, identified through official DLC membership records and affiliated Congressional caucus lists. Founded formally in 1985 (operating through 2011), the DLC arose in part from the “Watergate Babies” cohort of 1974. DLC members were more conservative than other Democrats especially on predistribution and social issues, relying differentially on corporate PACs and educated out-of-district donors. The paper treats DLC membership as a proxy for an anti-predistribution faction that gained bargaining power within the Democratic Party from the 1970s onward.

Adjusted Years of Schooling (AdjYearsEduc): The paper’s harmonized education variable across more than 1,000 surveys spanning eight decades. Because raw educational categories change over time and represent different selectivity (e.g., in 1940 only one-quarter of adults had completed twelfth grade, versus nearly 90 percent today), the authors use Census microdata to predict years of schooling as a function of self-reported educational category, sex, race, year, and birth cohort in ten-year bins. This provides a common unit of measurement across surveys with incompatible category systems.

Inflection Point (1976): The structural break in the trend of the education-Democratic identification gradient, estimated using Bai-Perron (1998) methods on N ≈ 2.2 million observations. The data select 1976 as the year at which the previously stable negative gradient begins its upward trajectory. The corresponding Republican inflection point occurs in 1992. The paper argues that identification of this inflection point — not previously documented in the realignment literature — is made possible only by the large historical dataset assembled.

Minor Civil Division Group (MCDG): The granular geographic unit used in the House election analysis for the 1980s, with approximately sixty MCDGs per Congressional district. Matched to 1980 Census demographic data to assign average years of education. Used to test whether DLC candidates out-perform other Democrats in more-educated neighborhoods, within the same Congressional district and election year, to address the concern that DLC candidates sort into more-educated districts.

(Not) Thinking About the Future: Financial Information and Maternal Labor Supply

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates whether information constraints — rather than fully forward-looking choices — contribute to mothers’ reduced labor supply after childbirth, a key driver of gender inequality. The authors deploy two complementary methods in Switzerland: a representative descriptive survey of Swiss mothers aged 25–50, and a large-scale randomized controlled trial (RCT) among approximately 2,400 female public school teachers with children who work part-time.

The descriptive survey first establishes that long-term financial factors are not top of mind for mothers making labor supply decisions: only about 11% of mothers spontaneously mention pensions or long-term career considerations when asked about their post-childbirth employment choices, compared to roughly half who mention child or own well-being. Beyond salience, the survey documents substantial misperceptions: 62% of women over-estimate pension receipt under part-time work by more than 10%, and a similar share believes wage growth under low part-time hours (40% FTE) is at least as high as under 80% employment. The authors label mothers with overly optimistic beliefs on both dimensions “cost-unaware”; 42% of the sample qualifies. Cost-unawareness is more prevalent among less-educated mothers and correlates with less financial interest and more gender-conservative attitudes.

The RCT tests whether providing objective, individualized information shifts financial planning and labor supply. Teachers in treatment schools (two-thirds of all schools) were individually randomized into a treatment group viewing an informational video about the long-run earnings, pension, and life-event consequences of sustained part-time employment, plus access to a Future Calculator tool, or a placebo video on unrelated financial topics. The two-stage randomization (school-level first, then individual within treated schools) allows identification of both direct treatment effects and spillovers. Outcomes are measured in a Wave 1 post-video survey, a follow-up survey two months later, and linked administrative personnel records from the Department of Education one year post-intervention.

Main findings: treated teachers are 31.26 percentage points (58% over the pure control mean) more likely to correctly rank the relative magnitude of long- versus short-term financial factors. Demand for financial planning tools rises by 0.39 standard deviations (SD) overall and by 0.31 SD among cost-unaware women specifically. In terms of stated labor supply plans, the treatment raises planned employment for the next academic year by 1.69 percentage points (ppt) in the full sample and by 4.95 ppt (9% over the pure control mean) among cost-unaware women. These plan effects persist two months later for cost-unaware women but fade for the full sample.

Critically, stated plans translate into verified behavior: linked administrative data one year post-intervention show that cost-unaware teachers increase their contracted employment level by 3.87 ppt, or 7% over the pure control mean of 53.30% FTE. Cost-aware and overly pessimistic women do not reduce their labor supply upon learning they are better off than feared, an asymmetry consistent with agents responding more to perceived losses than gains. If the 3.87 ppt increase were sustained from age 40 onward, cost-unaware teachers would accumulate an additional 130,000 CHF in lifetime income and 40,000 CHF in pension wealth, shrinking the gender gap in lifetime income and pension receipt among teachers by approximately 18% each.

The paper is scoped to Swiss female public school teachers — a population with linear pay scales, no part-time promotion penalty, and relatively low adjustment barriers — meaning the measured lifetime earnings and pension losses likely represent a lower bound relative to other occupations. Short-term RCT findings replicate among a sample of pregnant women in the general Swiss population, and the paper argues that similar labor supply adjustment magnitudes are feasible for a broader segment of part-time working mothers.

Q: What is the central research question and why does it matter? A: The paper asks whether mothers’ post-childbirth reduction in labor supply is partly driven by information constraints — specifically, whether mothers fail to account for the full long-term financial consequences of working reduced hours. This matters because if the child penalty partly reflects uninformed choices rather than deliberate tradeoffs, standard policy tools (parental leave, childcare subsidies) may underperform precisely because their long-term financial benefits are not internalized.

Q: How prevalent is cost-unawareness among Swiss mothers? A: 62% of mothers in the descriptive survey over-estimate pension receipt under part-time work by more than 10%, a similar share believes wage growth under low part-time (40% FTE) is at least as high as under 80% employment, and 42% are overly optimistic on both dimensions simultaneously. Cost-unawareness follows an education gradient: 77% of low-education women over-estimate pension receipt versus 51% of high-education women.

Q: What share of mothers spontaneously considers long-term financial factors when deciding on their labor supply? A: Only about 11% of mothers mention any long-term financial factor (pensions, financial independence, long-term career considerations) in open-ended responses; the share is similarly low across education groups (6% low, 12% mid, 13% high). About 50% mention child or own well-being; roughly 30% raise short-term financial factors such as current childcare costs.

Q: What are the actual long-term financial stakes of the average female teacher’s part-time employment pattern in Switzerland? A: Compared to full-time employment, the average female teacher’s employment trajectory produces a 35% reduction in potential lifetime earnings (approximately 3.34 million CHF versus 5.12 million CHF). Monthly pension receipt under the part-time scenario is 31% lower overall and 43% lower from the occupational second-pillar scheme specifically — a gap comparable to the average 47.5% gender pension gap observed in the second pillar in Switzerland in 2024.

Q: How was the RCT designed and what populations were included? A: The study recruited 2,359 part-time working mothers employed as public school teachers in a German-speaking Swiss canton. A two-stage randomization assigned two-thirds of schools to treatment schools (within which teachers were individually randomized 50/50 to treatment or spillover control) and one-third to pure control schools. This design allows estimation of direct treatment effects and spillover effects. The intervention was timed to precede December–January, the period when teachers communicate their preferred employment levels for the next school year.

Q: What was the treatment intervention? A: Treated teachers watched an informational video following a representative female teacher considering an employment-level increase, covering the impact of part-time work on lifetime earnings, monthly pension receipt, and financial exposure after adverse events such as divorce; it also benchmarked these magnitudes against childcare costs. Treated teachers additionally received individualized access to the Future Calculator, an online projection tool developed with a Swiss bank, calibrated to teachers’ deterministic salary and pension schedules.

Q: Did treated teachers understand and retain the treatment information? A: Yes. Treated teachers were 31.26 ppt (58% over the pure control mean) more likely immediately after the intervention to correctly rank long- versus short-term financial factors in a vignette. Two months later, the treatment group remained significantly more likely to apply the information correctly (22.63 ppt higher), indicating the knowledge was not short-lived.

Q: How did demand for financial planning tools respond to the treatment? A: The treatment raised a financial information/tools index by 0.39 SD overall. For cost-unaware women specifically, demand for financial tools rose by 0.31 SD; cost-aware and pessimistic women showed no significant change. There was no significant average treatment effect on sign-up for an incentivized financial consultation.

Q: How large were the labor supply plan effects in the survey, and did they persist? A: For the full sample, treated teachers planned a 1.69 ppt higher employment level for the next school year immediately after the treatment, and 3.13 ppt higher in 10 years. For cost-unaware women, the short-run planned increase was 4.95 ppt (9% over the pure control mean of about 55%), and plans for 5 and 10 years into the future rose by approximately 4 ppt (6–7% over the mean). The short-run effects for cost-unaware women persisted to the two-month follow-up, while full-sample short-run effects faded.

Q: What do the linked administrative data show about actual labor supply one year post-intervention? A: Cost-unaware women in the treatment group increased their contracted employment level by 3.87 ppt relative to the pure control group (7% over the pure control mean of 53.30% FTE), closely matching the planned increase stated immediately after the treatment. Cost-aware women and the full sample showed no statistically significant shift in actual hours.

Q: What asymmetry did the authors observe between cost-unaware and cost-aware women? A: Cost-unaware (overly optimistic) women increased their labor supply upon learning the true financial costs; cost-aware and overly pessimistic women did not reduce their labor supply upon learning they were better off than expected. The authors interpret this as consistent with agents responding more to perceived losses (bad news for cost-unaware women) than to gains (good news for pessimistic women), and with cost-aware women already having incorporated the financial logic into their decisions even without precise estimates.

Q: What is the estimated lifetime impact of the observed labor supply adjustment? A: If cost-unaware teachers maintain the 3.87 ppt employment increase from age 40 to retirement, they accumulate an additional 130,000 CHF in lifetime income and 40,000 CHF in pension wealth on average. This would reduce the gender gap in both lifetime income and pension receipt among teachers by approximately 18% each.

Q: What emotional and social mechanisms did the paper document? A: The treatment initially produced significantly negative emotional responses (−0.41 SD on an emotions index overall; −0.68 SD for cost-unaware women), consistent with cognitive dissonance from information conflicting with prior beliefs. Two months later, the treatment group reported feeling more in control and less stressed, and cost-unaware women returned to a neutral emotional baseline. Treated women were also 19.61 ppt more likely to have discussed the topic with anyone, with the largest effect on conversations with partners or family.

Q: Did the treatment affect household-level labor supply — specifically, did partners reduce their hours? A: No. The authors found no evidence that partners of cost-unaware women planned to work less in response to the treatment, and women did not plan to adjust future fertility. This suggests the observed hours increase by treated cost-unaware women was not offset by partner adjustments within the household.

Q: Were there social spillover effects within schools? A: Treated teachers were 11.59 ppt more likely to report having discussed the video with colleagues. Two months later, cost-unaware control teachers in treated schools (the spillover group) showed some evidence of absorbing the general treatment message and adjusting short-term labor supply plans upward, and a noisy increase in actual employment of roughly one-third the magnitude of the direct treatment effect, though these estimates were imprecise.

Q: Why might cost-unaware women be uninformed in the first place? A: In both the descriptive survey and the RCT sample, cost-unaware women lean more gender-conservative in their attitudes and report less interest in financial topics. The authors interpret this as suggesting a lack of information (rather than mere salience or forgetting) drives cost-unawareness, implying that passive information delivery through employers or pension funds could be effective.

Q: What constraints to labor supply adjustment did the authors explore? A: In a hypothetical scenario exercise, the scenario producing the largest desired employment increase for both treatment and control groups was if the partner were more engaged (roughly double the adjustment relative to a scenario of higher pay for additional hours). The treatment group adjusted their desired employment level by an additional 0.62–2.03 ppt relative to pure control across all scenarios except relaxing conservative gender norms.

Q: How generalizable are the findings beyond the teacher sample? A: The short-term RCT findings replicated among a sample of pregnant women in the general Swiss population. The authors also document that potential net gains from increasing labor supply — net of additional childcare costs — are large for the broader population of part-time working Swiss mothers, supporting feasibility of similar-magnitude adjustments outside teaching. The teaching context likely represents a lower bound for lifetime earnings and pension losses in other professions due to the absence of a part-time promotion penalty in teaching.

Q: What are the policy implications? A: The findings suggest that default exposure to individualized financial information about the long-term costs of part-time work — delivered by employers, pension funds, or the state — could improve decision quality and labor supply. More broadly, the results imply that policies designed to increase female labor supply (parental leave reforms, childcare subsidies) may underperform if mothers do not fully internalize the financial benefits of additional hours; ensuring that families solve the correct optimization problem is a precondition for unlocking the full potential of such policies.

Child Penalty: The large and persistent reduction in women’s labor force participation and income following the birth of a first child, identified in the paper as the key driver of remaining gender inequality in the labor market in industrialized countries and a source of profound life-cycle financial consequences including reduced lifetime earnings and pension savings.

Cost-Unaware: The authors’ term for women who hold overly optimistic expectations about the financial consequences of part-time work — specifically, who over-estimate pension receipt under low part-time employment by more than 10% and who believe wage growth under low part-time is at least as high as under higher employment levels. In the descriptive survey 42% of mothers qualify on both dimensions.

Future Calculator: An online individualized projection tool developed by the authors in cooperation with a Swiss bank, calibrated to teachers’ deterministic salary and pension schedules, allowing users to estimate the long-term financial implications of different employment levels. Used both in the descriptive survey vignette and as part of the RCT treatment.

Second Pillar (Occupational Pension Scheme, PP): Switzerland’s occupational pension scheme, the pillar most heavily affected by part-time work because contributions are directly proportional to earnings above a minimum annual earnings threshold. The paper documents an average gender pension gap of 47.5% in this pillar in 2024 and a 43% lower monthly pension receipt for the average female teacher’s part-time trajectory relative to full-time employment.

Two-Stage Randomization: The experimental design used to separate direct treatment effects from spillover effects within schools. One-third of schools are assigned to a pure control group; in the remaining two-thirds, teachers are individually randomized into treatment or spillover control (untreated teachers in treated schools), enabling identification of both causal treatment impacts and social learning channels.

Information Constraint: The paper’s central mechanism — mothers’ failure to spontaneously account for the full long-term financial implications of reduced labor supply when making employment decisions, distinct from deliberate forward-looking tradeoffs. The authors document this both through the absence of long-term financial factors in open-ended decision narratives (only 11% of mothers mention them) and through systematic misperceptions of pension and wage outcomes.

Cognitive Dissonance (as used in the paper): The authors use this term to describe the initial negative emotional response (−0.41 SD overall, −0.68 SD for cost-unaware women) when treated women learn that the true financial costs of part-time work are higher than they expected — information that conflicts with prior beliefs and prior choices, producing unpleasant emotions that subsequently reverse into lower stress levels two months later.

A Cognitive Theory of Reasoning and Choice

Mon, 01 Jan 0001 00:00:00 +0000

Bordalo, Gennaioli, Lanzani, and Shleifer develop a cognitive theory of choice in which a decision maker’s attention to the features of options is determined by her categorization of the current problem against a memory database of problems she solved in the past. The core claim is that before solving a problem, the decision maker asks “what kind of problem is this?” and resolves it by selecting the category — indexed by a prototype attention-plus-context vector and a time-discounted frequency — whose similarity to the current problem is maximized. This problem recognition step then pins down which features (price, quality, probabilities) receive attention, which in turn shapes valuation and choice.

The model formalizes two-step choice. In step one (recognition), the decision maker jointly chooses an attention vector alpha_P and a category c* to maximize a separable similarity function S[(alpha_P, kappa_P), (alpha_c, kappa_c)] weighted by category frequency F_c, plus a Type I extreme-value shock that yields a logit probability over categories. In step two, she maximizes perceived value over the menu using the endogenously determined weights. Perceived hedonic value of feature i shrinks toward the menu average when alpha_{P,i} < 1; perceived probabilities compress toward uniform when the event-attention weight falls below 1, producing probability overweighting of unlikely events. Full attention recovers expected utility.

The model yields three structural predictions that hold without changing tastes or information. First, within-person multi-modal attention: because categorization is stochastic, the same person can cluster on entirely different features (e.g., the base rate vs. the likelihood in an inference problem) across otherwise identical choice occasions. Second, systematic context-driven instability: when an irrelevant context feature kappa_{P,i} drifts away from a category’s diagnostic kappa_{c,i}, the probability of that category falls discontinuously, causing a discrete switch in the attention profile and hence in valuation. Third, experience-driven heterogeneity: people more frequently exposed to a category (higher F_c) are more likely to use it, producing persistent differences in price elasticities or probability weighting at constant income and tastes.

Applied to riskless consumer choice, the paper introduces two categories — “buying” (full attention to price, partial to quality: alpha_{M_g}=1 > alpha_{Q_g}=alpha) and “consuming” (full attention to quality, partial to price: alpha_{Q_g}=1 > alpha_{M_g}=alpha). A jam problem categorized as buying yields valuation v = alphaq - etap; categorized as consuming, v = q - alphaetap. The valuation jumps discontinuously as context crosses a threshold kappa*, which shifts when relative category frequency F_{buy}/F_{con} changes. This framework accounts for context-dependent price elasticities (Wakefield and Inman 2003), poverty-driven excess price focus (Shah et al. 2018), de-commoditization through advertising, and mental accounting anomalies including opportunity cost neglect and the sunk cost fallacy — both arising because con neglects capital gains (alpha_{con,Delta_M}=0) and buy neglects quality shocks (alpha_{buy,Delta_Q}=0).

Applied to statistical judgment, the paper introduces two categories — “frequency estimation” (attention alpha_1=1 to a single i.i.d. draw from a known DGP) and “agnostic inference” (attention alpha_S=1 to the share of heads as a sufficient statistic). The threshold N* separates recognition: for sequence length N_P < N*(F_{freq}/F_{inf}), the decision maker categorizes as frequency and correctly assesses odds; for N_P >= N*, she switches to inference and overweights balanced sequences, producing the Gambler’s Fallacy. The same competition between categories also accounts for base rate neglect, conjunction fallacy, and correlation neglect, with the bias strengthening as sequences grow longer.

Applied to risky choice, bottom-up salience — sensory prominence and contrast — interacts with categorization. A publicity shock drawing attention to a low-probability contamination risk raises similarity to “consuming,” triggering a category switch that amplifies attention to quality broadly and reduces attention to price, producing large valuation drops disproportionate to the actual probability shift. This mechanism generates the framing effects of prospect theory without a stable S-shaped utility function: gains and losses frames correspond to different contexts activating different categories.

Scope conditions: the theory applies when features and their values are fully known to the decision maker (no uncertainty about attributes), so the distortions take the form of altered sensitivity to known features rather than missing information. The set of categories C is taken as given in the formal analysis, though the authors discuss endogenization as future work.

Q: What is the paper’s central departure from standard rational inattention and noisy-perception models?

A: Standard models (Sims 2003, Woodford 2012, Enke and Graeber 2023) produce unimodal, stably weighted valuations — the decision maker’s weighting of features is a smooth function of payoff-relevant costs or priors. In this paper, the weighting is determined by problem recognition, which is discrete and stochastic, producing within-person multi-modal attention: the same person can cluster on entirely different features across identical problems. The authors cite direct evidence from Bordalo, Conlon, Gennaioli, Kwon, and Shleifer [20] showing bimodal clustering on base rates vs. likelihoods in statistical problems, a pattern inconsistent with stable-weighting models.

Q: How is perceived value distorted when the attention weight on a hedonic feature is below 1?

A: The perceived value of hedonic feature i is u_i(alpha_P) = alpha_{P,i} * u_i + (1 - alpha_{P,i}) * u_bar_i, where u_bar_i is the average value of that feature across options in the menu. An attention weight of zero collapses perceived variation in that feature to zero; full attention recovers the true value. The implication is that under-attention shrinks the decision maker’s effective sensitivity to a known attribute, causing systematic under- or over-valuation relative to a rational benchmark while tastes (marginal utilities) are held fixed.

Q: How is perceived probability distorted?

A: With attention weight alpha_{P,W} on event W, the perceived probability of event e is P(e)^{alpha_{P,W}} / sum_{e’} P(e’)^{alpha_{P,W}}, which compresses the distribution toward uniform as alpha_{P,W} falls toward 0 and recovers the true distribution at alpha_{P,W}=1. In the jam example, under-attention to the small probability of spoilage causes the decision maker to overestimate the risk of contamination. For multi-dimensional event vectors the formula generalizes multiplicatively, allowing “editing out” of entire event dimensions (e.g., urn selection in a balls-and-urns problem) when their attention weight hits zero.

Q: What is the mechanism for context-dependent price elasticity?

A: When context kappa_P is below threshold kappa*(F_{buy}/F_{con}), the decision maker categorizes the problem as “buying” and her valuation is v = alphaq - etap, giving a high price sensitivity (coefficient eta) and attenuated quality sensitivity (coefficient alpha < 1). Above kappa*, she categorizes as “consuming” and valuation is v = q - alphaetap, reversing the emphasis. Because the threshold kappa* is increasing in relative frequency F_{buy}/F_{con}, a decision maker with more buying experience has a higher threshold and thus acts as more price-elastic at any given context level. These elasticity differences arise without any change in the true marginal utility of money eta or quality q.

Q: How does the model generate the sunk cost fallacy and opportunity cost neglect as a unified phenomenon?

A: Both anomalies arise because buying and consuming categories selectively neglect shocks. In the football example, recognizing the problem as “buying” activates alpha_{buy,Delta_Q}=0, so the blizzard quality shock Delta_q<0 is ignored and the decision maker drives to the game as if the shock did not occur — the sunk cost fallacy. In the wine example, recognizing the problem as “consuming” activates alpha_{con,Delta_M}=0, so the capital gain Delta_p is ignored and the decision maker reports a zero or purchase-price cost — opportunity cost neglect. The unifying mechanism is that each category attends only to the features diagnostic of its prototypical experiences: buying attends to price paid and normal quality; consuming attends to realized quality and partly to price, but not to capital gains.

Q: What comparative static does the model predict for sunk cost susceptibility based on experience?

A: People with higher F_{buy} (more buying experiences, e.g. poverty experiences or having recently purchased but not yet consumed the good) exhibit more sunk cost fallacy and less opportunity cost neglect. Conversely, season ticket holders face many consuming experiences relative to one buying event, raising F_{con} and thus reducing susceptibility to the sunk cost fallacy for sports events. Making the blizzard more salient in the description shifts similarity toward “consuming,” also reducing the sunk cost fallacy through a different channel (bottom-up salience rather than experience).

Q: What is the paper’s explanation for the Gambler’s Fallacy, and what distinguishes it from prior accounts?

A: The Gambler’s Fallacy arises when sequence length N_P exceeds threshold N*(F_{freq}/F_{inf}), causing the decision maker to switch from the frequency category (which attends to the 50:50 fairness of the coin) to the inference category (which attends to the share of heads). Under inference, the decision maker treats balanced and unbalanced sequences as representatives of their “share of heads equivalence class,” and the class of balanced sequences is larger, so balanced sequences receive higher estimated probability — the Gambler’s Fallacy. This differs from Rabin and Vayanos (2010), where the bias stems from a belief that the coin is drawn from a pool; here the decision maker knows the coin is fair (kappa_{P,U}=0.5) but the inference representation causes question substitution rather than a wrong model of the DGP.

Q: How does the model make the Gambler’s Fallacy testable beyond length effects?

A: The model predicts the bias is stronger for decision makers who recently solved many inference problems (lower F_{freq}/F_{inf}), and weaker when the 50:50 nature of flips is made bottom-up salient in the choice context (because salience raises similarity to the frequency category, hindering recognition of inference). These cognitive proxies — experience frequencies and bottom-up salience — are orthogonal to the statistical content of the problem and thus allow identification of the mechanism separately from changes in information or incentives.

Q: How does the model produce framing effects in risky choice without a stable S-shaped utility function?

A: Gains and losses frames are modeled as different context vectors kappa_P that differentially increase similarity to a “safe outcome” category or a “risk” category. Recognizing the problem as the safe-outcome category shifts attention toward the certain option; recognizing it as the risk category shifts attention toward variance. The reversal of preferences between gain and loss frames (the Asian Disease problem, Tversky and Kahneman 1981) thus emerges from context-driven re-categorization rather than from a fixed probability weighting function. The novel prediction is that framing effects should be stronger for decision makers with more experience with the category activated by each frame, and weaker when bottom-up salience of the alternative frame’s features is raised.

Q: How does bottom-up salience interact with top-down categorization in the contamination example?

A: A publicity shock alpha_{delta,Q_b}>0 raises baseline attention to the spoiled-jam quality feature, increasing the similarity of the current problem to the “consuming” category (where quality is focal). This triggers a category switch for marginal agents, activating the full consuming attention profile — which attends to quality broadly, not just to contamination specifically, and reduces attention to price. The resulting valuation drop is therefore disproportionate to the actual probability of contamination and exhibits price insensitivity, because re-categorization shifts the entire attention profile rather than just updating a single probability.

Q: How does the model relate to and distinguish itself from case-based decision theory (Gilboa and Schmeidler 1995) and analogical reasoning (Mullainathan 2002, Fryer and Jackson 2008)?

A: In Gilboa-Schmeidler and related models, the decision maker uses past cases to resolve uncertainty about unknown attributes of current options; attention is full and the mechanism is extrapolation of payoffs from similar cases. In Mullainathan (2002) memory-based model, categories again serve to fill in missing information. In this paper, there is no uncertainty about attributes — features and their values are fully known — and the distortion instead takes the form of altered sensitivity to known features through selective attention. This allows the model to produce biases even in simple problems with full data disclosure, and to explain phenomena like base rate neglect and price insensitivity that are not primarily about missing information.

Q: What does the model predict about within-person versus across-person distributions of valuations?

A: Within a person, attention is multi-modal (bimodal in the two-category case) because categorization is stochastic. However, if many categories are possible across the population, the aggregate distribution of valuations can appear approximately unimodal even though each individual’s distribution is not. This distinction is empirically important: a researcher observing average choices may incorrectly infer smooth preference heterogeneity when the underlying mechanism is discrete category switching.

Q: What cognitive proxies does the model propose for empirical identification?

A: The theory links endogenous attention and choice to three observable (or measurable) proxies: (1) past experience frequencies F_c, measurable from administrative histories, surveys about past exposure, or experimental manipulation of training; (2) contextual similarity, measurable from field or experimental variation in irrelevant context features; and (3) bottom-up salience, experimentally controllable via prominence or contrast manipulations. The key identification logic is that these proxies are payoff-irrelevant — they do not change tastes, information, or the objective choice problem — yet predict systematic shifts in choice through their effect on recognition.

Problem Recognition: The first step in the decision maker’s choice process, in which she jointly selects an attention vector alpha_P and a category c* by maximizing weighted similarity between the current problem (characterized by its context vector kappa_P) and the prototype of a past category (alpha_c, kappa_c), multiplied by the category’s time-discounted frequency F_c. Recognition is not about resolving uncertainty over attributes but about selecting which known attributes to attend to.

Category: A partition element of the decision maker’s memory database, indexed by a prototype attention-plus-context vector (alpha_c, kappa_c) and a frequency scalar F_c. The prototype encodes both the context features diagnostic of experiences in that category (binary alpha_{c,i} for i in Phi_K) and the attention to hedonic and event features (alpha_{c,i} for i in Phi_H union Phi_E) used when solving problems in that category. Examples in the paper: “buying” and “consuming” for riskless choice; “frequency estimation” and “agnostic inference” for statistical judgment.

Attention Weight (alpha_{P,i}): A scalar in [0,1] assigned to feature i of the current problem P. For hedonic features, alpha_{P,i}<1 collapses perceived variation toward the menu average; for event features, alpha_{P,i}<1 compresses perceived probabilities toward uniform. Full attention alpha_{P,i}=1 recovers expected utility. Attention weights are the endogenous output of the recognition step, not fixed preference parameters.

Contextual Similarity S: A separable function measuring how close the current problem (alpha_P, kappa_P) is to a category prototype (alpha_c, kappa_c). It decreases in discrepancies in the attention vector (measured by a strictly increasing, convex function d) and in discrepancies in the values of context features diagnostic of the category (d_i(kappa_{P,i}, kappa_{c,i}) * alpha_{c,i}). Endogenous attention to context is set to reduce sensitivity to discrepancies, not to eliminate them.

Mental Accounting (as categorization): In the paper’s account, non-fungibility, sunk cost fallacy, and opportunity cost neglect all arise because buying and consuming categories selectively attend to different monetary and quality features. The sunk cost effect is alpha_{buy,Delta_Q}=0; opportunity cost neglect is alpha_{con,Delta_M}=0. Mental accounts are not separate budget constraints but the by-product of category-specific attention profiles that were calibrated to normal-state experiences and do not generalize to shocks.

Bottom-up Salience: Exogenous attention to a feature driven by sensory prominence (described by alpha_{delta,i} in the problem’s presentation vector) or payoff contrast (the DM attends more to features where her option’s value deviates more from the menu average relative to total menu variance). Bottom-up salience raises baseline attention to a feature before top-down categorization acts, and can trigger a category switch by raising similarity to the category for which that feature is focal.

Gambler’s Fallacy via Question Substitution: In the model, the Gambler’s Fallacy arises when a long sequence length kappa_{P,N} causes recognition of the “agnostic inference” category, which focuses attention on the share of heads alpha_S=1. The decision maker then treats sequences as representatives of a “share of heads equivalence class,” and since the balanced class is larger than the unbalanced class, balanced sequences are assigned higher estimated probability. This is not a belief that the coin is unfair; it is question substitution induced by the inference representation.

A Monetary-Fiscal Theory of Sudden Inflations

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. Why do sudden inflations and currency crises occur, while symmetric sudden deflations never do? The paper asks whether treating nominal government bonds as analogous to ordinary corporate bonds — with an asymmetric payoff structure capped at face value on the upside but exposed to real losses when fiscal surpluses are insufficient — can generate a unified theory of these crises endogenously from a single model.

Intellectual Lineage and Approach. The paper sits at the intersection of two literatures. The first is the Fiscal Theory of the Price Level (FTPL), originating with Leeper (1991), Sims (1994), and Sargent and Wallace (1985), which links the real value of nominal government debt to expected future surpluses. The second is the safe-asset literature, where Holmstrom (2015) and Gorton (2017) explain that assets can circulate as safe stores of value precisely because their backing is costly to investigate and consumers rationally remain uninformed. The paper applies this information-economics logic to nominal government bonds, so that consumers normally hold bonds without investigating the government’s true fiscal capacity, and only pay the cost to investigate when real repayment doubts become sufficiently severe.

Model Structure. The model is a two-period reduced-form general equilibrium. In period 1, a representative consumer buys nominal government bonds at an interest rate set by the monetary authority. In period 2, the government must repay those bonds. The fiscal authority attempts to hit a price-level target P* by raising tax revenue, but faces a hard ceiling τ_max on the surplus it can collect — arising from Laffer limits on taxation, political constraints on austerity, or the need to fund financial-sector bailouts. The consumer has prior beliefs that τ_max is low (L) with probability π and high (H) with probability 1−π, and can pay a fixed utility cost γ to learn τ_max before deciding how many bonds to purchase.

Bond Payoff Structure and Asymmetry. The key mechanism is the asymmetric, bond-like real payoff of nominal government debt. If τ_max ≥ B1/P*, the government raises enough surplus to repay bonds fully in real terms at the price-level target; the real payoff is flat at face value (the “in-the-money” region). If τ_max < B1/P*, the government sets taxes to the ceiling τ_max and the price level rises above P* to balance the budget constraint, reducing the real payoff proportionally (the “default” region). Critically, because the nominal payoff is capped at face value, there is no upside region: governments will not run surpluses large enough to deliver a windfall to bondholders, so sudden deflations — analogous to a corporate bond being worth more than face value — cannot occur. This asymmetry is the direct source of the one-sided nature of crises.

Two Illustrative Mechanisms for Sudden Inflations. The paper numerically and analytically characterizes two triggering scenarios:

Lower surplus expectations (fiscal stress narrative, corresponding to Burnside et al. 2001 on the 1997 Asian crisis): As the probability π of a low future surplus (e.g., from a prospective banking-sector bailout) rises, the value of information about τ_max increases. In the numerical example (i = 0.05, γ = 0.13, L = 0.1), the value of information equals the cost γ at π = 0.15. For π above 0.15, consumers pay to investigate, learn τ_max = L, and refuse to purchase bonds beyond what will be repaid in real terms (B1 = τ_max = L = 0.1). The price level in period 1 rises discontinuously as a function of π at this threshold.
Interest rate increases (speculative attack narrative): As the monetary authority raises the interest rate to defend a currency, consumers demand more bonds. Larger bond quantities increase the risk that surpluses will be insufficient, raising the value of fiscal information. In the numerical example (π = 0.5, γ = 0.24, 1+i ∈ [1, 1.2]), the value of information equals γ at 1+i = 1.1 (i.e., i = 10%). For interest rates above this threshold, consumers learn τ_max = L, restrict bond purchases to what will be repaid, and the price level in period 1 jumps discontinuously. Further interest rate increases above the threshold produce only upward drift in the price level, not additional monetary tightening effects — illustrating the limits of monetary policy in fiscally stressed environments.

Theoretical Results. Two formal theorems establish generality. Theorem 1 shows that, given bond demand B1(π) such that L < B1 for all π ∈ (0,1), there exist thresholds k and γ > 0 such that the period-1 price level P1 is discontinuous as a function of π on (0, k]. Theorem 2 establishes an analogous discontinuity in P1 as a function of the interest rate i, given that B1(i) > L for all i in the relevant range.

Scope Conditions. The model is a two-period reduced form that abstracts from dynamics, multiple maturities, and secondary market trading. The informational friction is a fixed binary cost γ, not a richer signal structure. The results depend on the existence of a binding surplus ceiling τ_max; when the government is far from this ceiling (i.e., consumers’ beliefs are far from the “default boundary”), shocks produce only small, smooth price-level changes. Large discontinuous price-level jumps require the economy to be near the kink point of the bond payoff curve.

In depth

Q1. What is the fundamental analogy that drives the paper’s theory, and what economic literature does it build on?

The paper analogizes nominal government bonds to corporate bonds (following Sargent 1982’s advice that “government debt is valued according to the same economic considerations that give private debt value”). Like a corporate bond, the nominal government bond pays its face value if the underlying project (government fiscal capacity) delivers a surplus at least equal to the face value, but pays only a share of the realized surplus if the surplus falls short. This bond-like payoff — flat on the upside, proportional to outcomes on the downside — is the direct source of asymmetric crisis dynamics. The paper combines this with Holmstrom (2015) and Gorton (2017)’s framework in which safe assets function because their backing is costly to investigate, so consumers rationally remain uninformed in normal times.

Q2. What is the key information friction, and how does it generate the switch between “normal times” and crisis?

In normal times, consumers are confident that the government’s future maximum surplus τ_max is sufficient to repay bonds in real terms. The fixed utility cost γ of investigating the true surplus exceeds the benefit, so consumers remain uninformed and bonds trade at a price reflecting only uninformed prior beliefs. A crisis arises when the value of information V(.) rises above γ — either because the probability of a low surplus state rises (fiscal stress) or because the interest rate rises and consumers demand more bonds, bringing them closer to the repayment boundary. Once V > γ, consumers investigate and, upon learning τ_max = L (low surplus), refuse to hold bonds that will not be repaid in real terms, triggering a discrete upward jump in the price level.

Q3. How does the bond payoff structure explain the absence of sudden deflations?

The real payoff of a nominal government bond cannot exceed its face value: the bond is capped at face value on the upside because the government will not voluntarily raise tax surpluses to deliver a windfall to bondholders. In the event that surpluses turn out to be higher than needed (τ_max ≥ B1/P*), the government simply sets taxes to exactly repay the bonds at P* and returns no additional real value to bondholders. This is the flat portion of the payoff curve. Because there is no upside kink — no region where learning that τ_max is unexpectedly large causes the price level to fall sharply — there is no mechanism for sudden deflations symmetric to sudden inflations. The 1933 U.S. episode (Jacobson et al. 2019) is cited: when deﬂation from leaving gold would have required fiscal austerity for full real repayment, Roosevelt chose to exit the gold standard rather than allow deflation.

Q4. How does the first numerical example (lower surplus expectations) work quantitatively?

The baseline parameters are: i = 0.05, γ = 0.13, L = 0.1, H ≈ ∞, P* = 1, e1 = e2 = 1, B0 = 1, τ1 = 0.8, β = 1. The analysis is restricted to π ∈ (0, 0.3]. As π (probability that τ_max = L) rises, the value of information V(.) rises. At π = 0.15, V equals the cost γ = 0.13. For π > 0.15, consumers pay to investigate and, upon learning τ_max = L, purchase only B1 = L = 0.1 in bonds — the amount that will be repaid — causing the period-1 price level P1 to jump discontinuously from approximately 0.95 to approximately 1.13. For π ≤ 0.15, consumers remain uninformed and P1 rises only smoothly from below 1 as π increases (fewer bonds demanded as repayment risk rises, even without investigation).

Q5. How does the second numerical example (interest rate increase) work quantitatively, and what does it imply for monetary policy?

With π = 0.5, γ = 0.24, and 1+i ∈ [1, 1.2], as the monetary authority raises the interest rate, consumers demand more bonds, increasing real repayment risk and the value of information. At 1+i = 1.1 (i.e., i = 10%), V equals γ. For 1+i > 1.1, consumers investigate and learn τ_max = L; they then only purchase bonds up to the repayment limit, causing P1 to jump discontinuously to approximately 1.15. For interest rates above the threshold, further increases yield only a smooth upward slope in P1 (bond purchases are fixed in real amount but nominal revenue falls). This illustrates that the monetary authority’s ability to use higher interest rates to lower the price level is limited by the surplus constraint: once the interest rate is high enough to trigger consumer investigation and a fiscal crisis, raising rates further is inflationary rather than deflationary.

Q6. What are the two regions of the deterministic model and how do they differ in fiscal and price-level dynamics?

In the deterministic version (1-π = 0, so τ_max = L with certainty, and there is no uncertainty), the model produces two distinct regions. In the “insufficient surplus” region where τ_max < B1/P*, the fiscal authority sets taxes to their maximum τ_max, the real payoff of bonds is τ_max/B1 < 1, the period-1 price level P1 = B0/(βτ_max), and real bond revenue Π = βτ_max (constant in τ_max). Selling additional bonds does not raise additional real revenue because any extra bonds lead to a proportional rise in P2 and a fall in Q. In the “sufficient surplus” region where τ_max ≥ B1/P*, the government meets its fiscal target (τ2 = B1/P*), P2 = P* is hit, P1 = βB1/(B0P*), and Π = βB1/P* (increasing in B1). In this region, selling additional bonds does raise real revenue and lowers P1 as the government absorbs more money.

Q7. What are the two interest rate regions in the deterministic model, and what is their implication for monetary policy effectiveness?

Using B1 = B0(1+i) (debt rolled over at the chosen rate), the monetary authority has two interest-rate regions. In the “constrained” region where 1+i > τ_max P*/B0 (the surplus ceiling binds), raising i does not change the period-2 surplus (τ2 = τ_max), does not change real revenue (Π = βτ_max), and does not affect P1 — but raises P2 above the target P*. In the “unconstrained” region where 1+i ≤ τ_max P*/B0, raising i increases bond demand, increases real surplus backing, raises real revenue, and lowers P1 while P2 = P* is maintained. The boundary between these regions determines the limit of monetary policy: the monetary authority can reduce P1 by raising i only up to the point where the surplus ceiling would be hit.

Q8. How does the paper relate to and extend prior FTPL literature?

The paper is grounded in the FTPL of Leeper (1991), Sims (1994), and Cochrane (2005, 2020), in which the price level is determined by the requirement that real government liabilities equal the present value of future surpluses. The paper’s contribution is to make the information structure endogenous: consumers’ beliefs and their decision to acquire fiscal information determine whether or not the FTPL logic is operative. In normal times (consumers uninformed), the price level does not respond to changes in the maximum surplus — a result that resembles the “Ricardian” or non-FTPL regime. When consumers investigate and learn the surplus is insufficient, the connection between the surplus and the price level is restored, reproducing FTPL-type dynamics. This provides an endogenous, single-model rationale for the regime-switching behavior between FTPL and non-FTPL environments documented empirically in Bianchi and Melosi (2013, 2017) and Davig and Leeper (2006).

Q9. What is the welfare role of consumer ignorance in this framework?

Consumer ignorance of the government’s true surplus plays a dual role. On one hand, ignorance is individually rational in normal times because the cost γ of investigating exceeds the benefit V (.) when beliefs are comfortably away from the default boundary. On the other hand, following Dang et al. (2017), informed knowledge of the safe asset’s backing destroys the symmetric ignorance that supports the asset’s role as a safe store of value, reducing welfare. In this model the concern is repayment risk rather than adverse selection: the consumer fears not being repaid in real terms and chooses to investigate when that risk is sufficiently high, potentially triggering the very crisis they feared.

Q10. What are the scope conditions and limitations of the model?

The model is explicitly a two-period reduced form designed to illustrate the bond-payoff mechanism in the simplest possible setting. It abstracts from: multi-period bond maturities and secondary market trading; rich heterogeneity among consumers; endogenous monetary and fiscal policy responses beyond the simple rules specified; and the general equilibrium interactions between inflation, output, and labor markets. The information cost γ is modeled as a fixed binary cost rather than a continuous or richer signal structure. The results on discontinuous price-level jumps hold when bond demand is sufficiently large relative to L (i.e., L < B1), ensuring genuine repayment risk; when surpluses are very large relative to bond liabilities, no crisis dynamics arise.

Key Concepts

Maximum Surplus (τ_max). The paper’s name for the hard ceiling on the net tax revenue (taxes minus money transfers) the government can collect in the second period. This ceiling can arise from a Laffer limit on taxable income, political-economy constraints on austerity, or from a banking crisis requiring government transfers to bail out the financial sector. It is the paper’s analogue of a project’s liquidation value: the maximum the “project” (the government) can deliver to bondholders.

Bond-Like Payoff of Nominal Government Debt. The paper’s central structural claim: the real payoff to holding a nominal government bond is capped at face value on the upside (the government will not raise surpluses beyond what is needed to repay bonds at the price-level target) but falls proportionally below face value when τ_max is insufficient for full real repayment. This is precisely the payoff structure of a standard corporate bond — flat on the upside, proportional to recovery on the downside — and it is the source of the asymmetry between sudden inflations and the absence of sudden deflations.

Value of Information (V(.)). Defined as the difference in expected utility between a consumer who learns the true τ_max before making bond-purchase decisions and one who remains uninformed and acts only on prior beliefs π, 1−π. The consumer investigates if and only if V(.) > γ. V is zero when beliefs are certain (limπ→0 and limπ→1), can be hump-shaped in π, and is increasing in the interest rate i (through its effect on bond demand). The threshold condition V = γ defines the boundary between “normal times” (no investigation) and crisis (investigation and possible sudden inflation).

Endogenous Information Structure. The paper’s term for the property that whether consumers choose to learn the government’s fiscal capacity is itself determined within the model by the parameters of the economy (the interest rate, prior beliefs, the cost of investigation). This contrasts with models that exogenously specify whether agents are informed or not. The endogenous information structure is the mechanism by which the paper generates the two apparent regimes (FTPL-active vs. FTPL-dormant) from a single unified model.

Default Boundary. The kink point in the bond payoff curve at τ_max = B1/P*: the level of the maximum surplus at which the government exactly repays bonds in real terms at the price-level target. When beliefs or bond quantities place the economy near the default boundary, small changes in π or i can push the economy across it, triggering large price-level responses. When the economy is far from the boundary (τ_max comfortably above B1/P*), small shocks have only small smooth effects.

Sudden Inflation / Currency Crisis (as defined in this paper). A discrete, discontinuous jump in the period-1 price level P1 that occurs when consumers pass the threshold V(.) = γ and investigate the government’s fiscal capacity, finding surpluses to be insufficient. The mechanism is: informed consumers refuse to hold bonds they know will not be repaid in real terms at P*, forcing the price level to jump to clear the government’s budget constraint with fewer bonds outstanding. The paper treats sudden inflations and currency crises as the same mechanism in different institutional contexts.

Repayment Risk Premium. The markup above the risk-free rate that consumers require on government bonds to compensate for the probability that the government’s surplus will be insufficient for full real repayment (i.e., the probability that the economy is in the τ_max < B1/P* region). This premium is present even when consumers are uninformed (i.e., do not know which state of τ_max will occur), and is reflected in the consumer’s first-order condition for bond demand.

A Theory of How Workers Keep up with Inflation

Mon, 01 Jan 0001 00:00:00 +0000

Aggregation and the Estimation of Quality Change

Mon, 01 Jan 0001 00:00:00 +0000

Errico and Lashkari address two intertwined problems in the measurement of aggregate price indices: how to account for quality change and variety entry/exit when the demand system is not CES, and how to identify flexible demand systems from prices and market shares alone when supply and demand shocks are correlated. The paper makes a theoretical contribution and a methodological one, then applies both to the measurement of US import price inflation over 1989–2016.

The theoretical contribution generalizes the unified CES price index of Redding and Weinstein (2020a) and the Feenstra (1994) variety correction to the full class of smooth, invertible demand systems. The key insight is that the contribution of quality change to the aggregate price index depends on heterogeneous cross-product elasticities of substitution, not a single scalar as in the CES case. For practical implementation, the paper specializes to the Homothetic with Aggregator (HA) family of demand systems — which includes Kimball (1995), CRESH (Hanoch, 1971), and HSA (Matsuyama and Ushchev, 2017) — showing that within this family cross-product elasticities collapse to product-level elasticities, dramatically reducing dimensionality. The resulting approximate price index (Proposition 2) weights each product by its love-of-variety index 1/(epsilon_it − 1), departing from the uniform CES weighting.

The methodological contribution is a dynamic panel (DP) identification strategy that exploits the Markov structure of quality shocks. The paper assumes that innovations to product quality are mean-zero conditional on lagged prices. Under flexible pricing, firms maximize current-period profits without regard to future demand shocks, so lagged prices are valid instruments for current prices. This permits identification of rich demand systems without external cost instruments and without the conventional assumption of uncorrelated supply and demand shocks. The conventional Feenstra–Broda–Weinstein (FBW) approach imposes zero correlation between quality shocks and prices; the paper shows that when quality and marginal cost are positively correlated, FBW produces downward-biased elasticity estimates (endogeneity bias).

The empirical application constructs a dataset covering 155 time-consistent 5-digit NAICS industries over 1989–2018, matching US customs import data with domestic production data and treating country-of-origin varieties as the unit of observation. The paper estimates both CES and Kimball demand systems using the DP approach and compares them to FBW estimates.

Key quantitative findings: First, DP-estimated CES elasticities are larger on average than FBW estimates (weighted mean 5.99 vs. 4.62), confirming a downward endogeneity bias in conventional methods. Second, Kimball mean elasticities exceed CES estimates (weighted mean 3.11 for Kimball vs. 5.99 for CES at the industry level, but the Kimball distribution has a mean of 17.0 and median 4.70), reflecting a heterogeneity bias — CES understates the dispersion of elasticities and thereby understates the elasticity relevant for the base (domestic) product whose market share is declining. Third, quality improvements in imported goods reduced the US import price index by approximately 20.2 percentage points cumulatively (0.67 p.p. annually) under Kimball demand, and 15.9 percentage points cumulatively (0.53 p.p. annually) under CES demand, over 1989–2018. The headline figure cited in the abstract is approximately 0.7 p.p. annually. The aggregate import price index (price plus quality components combined) fell by 8.25 p.p. cumulatively under Kimball and 4.01 p.p. under CES, compared to a BEA PCE index increase of 57.8 p.p. over the same period. Sectorally, machinery and electrical equipment account for roughly 60% of total quality gains (~200 p.p. cumulative). By country, China accounts for approximately 35% of cumulative quality gains, with non-OECD countries collectively contributing ~59%, and China’s quality upgrading accelerating after WTO accession.

Validation using US automobile market data (1980–2018) confirms the DP identification assumption: controlling for current product characteristics, future characteristics are uncorrelated with current prices. The DP approach produces elasticity estimates and quality change measures similar to those obtained using real exchange rate cost-shock instruments, and the Kimball demand closely matches mixed logit (BLP) estimates of both price elasticities and price indices. CES estimates exhibit a measurable downward heterogeneity bias in this validation setting, which the paper traces theoretically and empirically to a positive covariance between demand elasticities and price volatility across products.

Scope conditions: results apply to homothetic (income-invariant) demand; nonhomothetic extensions are provided as a generalization (Proposition 4) but not the primary focus. The import price index measures the cost of imports conditional on given domestic consumption; it does not capture full consumption-side welfare effects including substitution away from domestic varieties.

Q1: What is the core theoretical result on price index measurement beyond CES? Proposition 1 shows that for any smooth, invertible demand system satisfying the connected substitute property, the change in the log aggregate price index can be approximated as a weighted sum of log price changes and log expenditure share changes, with the expenditure share changes premultiplied by the inverse of the matrix Psi_t capturing cross-product elasticities of substitution. In the CES special case this reduces to the scalar (1/(sigma−1)) weight of the Redding-Weinstein (2020a) CUPI. The key departure in general demand is that the weight applied to each product’s expenditure share change is heterogeneous and depends on the full matrix of cross-product substitutabilities, not a single constant.

Q2: How does the HA (Homothetic with Aggregator) family simplify the theoretical results? For HA demand — which nests Kimball, CRESH, and HSA — Lemma 1 establishes that cross-product elasticities sigma_ij depend only on product-level elasticities epsilon_i through simple analytic formulas (e.g., epsilon_i * epsilon_j / epsilon-bar for HDIA), reducing the estimation problem from an N×N matrix to a vector of N scalars. Proposition 2 then gives an approximate price index in which each product’s expenditure share change is weighted by its love-of-variety index 1/(epsilon_it − 1), rather than a common CES scalar. This is the operative formula for the Kimball application.

Q3: What is the endogeneity bias in conventional elasticity estimation and how large is it? Conventional FBW methods assume supply and demand shocks are uncorrelated; when quality improvements are positively correlated with product prices (e.g., higher-quality goods command higher prices and also have higher marginal costs), FBW estimates are biased downward. The paper documents this: for CES demand, the DP-estimated weighted mean elasticity is 5.99 versus 4.62 under FBW, and for median estimates the DP value is 4.27 versus 2.58 under FBW, across 155 industries. The bias matters because underestimated elasticities imply underestimated quality changes and a smaller quality correction to the price index.

Q4: What is the heterogeneity bias and how does it differ from the endogeneity bias? Even after correcting for endogeneity, CES demand imposes a single elasticity per industry, ignoring the cross-product distribution. The paper shows that the CES estimate is an average that does not correctly capture the behavior of the base product (the domestic US variety) whose market share is declining. Because the domestic variety tends to have a lower elasticity than the import average, CES understates this product’s love-of-variety index and thereby understates the quality correction attributable to rising import shares. Theoretically and empirically (Appendix E.4), this bias is larger when demand elasticities covary positively with price volatility across products.

Q5: What is the dynamic panel identification assumption and why does it hold under flexible pricing? The paper assumes that quality shock innovations u_it are mean-zero conditional on lagged log prices: E[u_it | log p_it−1] = 0. Under flexible pricing, firms maximize current-period profits using current variables only; current prices are determined by current quality but are not chosen in anticipation of future quality shocks. Therefore lagged prices are uncorrelated with future quality innovations, making them valid instruments for current prices. This assumption is validated empirically in the automobile market: controlling for current product characteristics (horsepower, weight, fuel economy), future characteristics are not correlated with current prices.

Q6: What are the headline findings on quality change in US import prices? Under Kimball demand, quality improvements in imported goods reduced the US import price index by 20.2 percentage points cumulatively over 1989–2018, equivalent to 0.67 p.p. annually (the abstract rounds this to approximately 0.7 p.p. annually). Under CES demand, the quality contribution is 15.9 p.p. cumulatively (0.53 p.p. annually). The aggregate import price index combining price and quality changes fell by 8.25 p.p. under Kimball and 4.01 p.p. under CES over the same period. These figures imply that official import price statistics substantially overstate import price inflation by failing to account for quality improvements.

Q7: Which sectors and countries drive the quality gains? Machinery and electrical equipment account for approximately 60% of total cumulative quality gains, with roughly 200 p.p. cumulative quality improvement in that sector. Computer and peripheral equipment (NAICS 3341) is a notable contributor — the official import-to-producer price ratio shows a nearly five-fold increase between 1989 and 2018, but after quality adjustment this ratio reverses direction. By country of origin, China accounts for approximately 35% of cumulative quality gains; other non-OECD countries collectively contribute approximately 59%; OECD countries contribute approximately 7%. China’s quality upgrading is documented to accelerate following its WTO accession.

Q8: Why does CES understate the quality correction relative to Kimball? The primary mechanism is that the US domestic variety — which serves as the numeraire for quality measurement — has a declining market share over the sample period. In Kimball demand, products with declining market shares are assigned lower elasticities (higher love-of-variety indices), amplifying the quality correction associated with import share gains. CES imposes a uniform elasticity, failing to capture this asymmetry. The paper shows that the key driver of the CES-Kimball gap in the import price index is CES underestimating the love-of-variety index of the base domestic product.

Q9: How is the identification approach validated in the automobile market? Using the Berry-Levinsohn-Pakes dataset extended by Grieco et al. (2024) for 1980–2018, the paper first verifies empirically that future product characteristics (horsepower, weight, fuel efficiency) are uncorrelated with current prices after controlling for current characteristics. It then compares DP estimates for both CES and Kimball demand against estimates obtained using real exchange rate (RER) variation as a cost-shock instrument, finding similar results in both cases. Finally, it compares Kimball and CES estimates against mixed logit (BLP) demand: Kimball closely matches BLP price elasticities and implied quality changes, while CES shows a downward heterogeneity bias.

Q10: What does the automobile market validation imply for the import price index methodology? Since Kimball demand matches the richer mixed logit demand in the auto setting — where product characteristics are observed — the validation provides evidence that Kimball demand serves as a good approximation to rich heterogeneous-elasticity models when characteristics are unavailable. The paper constructs price indices for the US auto industry based on mixed logit, mixed CES, Kimball, and standard CES, and shows that the Kimball index is closer to the mixed logit and mixed CES indices than is the standard CES index.

Q11: How does the paper handle product entry and exit? Proposition 3 generalizes Proposition 1 to accommodate product entry and exit. The expression includes a variety correction analogous to Feenstra (1994) but generalized to non-CES settings via the mean love-of-variety index of entering and exiting products. In the CES special case this reduces exactly to the Feenstra (1994) correction. In the empirical application to US imports, entry and exit of country-of-origin varieties within industries is a relevant margin given the expansion of trading partners over the sample.

Q12: How does the paper relate to Redding and Weinstein (2020a)? Redding and Weinstein (2020a) derive a price index formula under CES demand that accounts for taste shocks, applied to US retail scanner data where quality is constant at the barcode level. The present paper generalizes their CUPI formula beyond CES to general and HA demand systems, and extends their identification strategy to settings where demand changes partly reflect quality changes rather than pure taste shocks. The paper also shows that the CES assumption used in Redding-Weinstein may overstate the contribution of taste shocks to cost-of-living indices, since part of the expenditure share variation attributed to taste shocks under CES would be reassigned under heterogeneous-elasticity demand.

Q13: Does the paper address welfare implications beyond the import price index? The paper explicitly notes that the import price index does not capture the full consumption-side welfare effects of rising imports, since gains from lower import prices may be partly offset by substitution away from domestic varieties. The paper also notes that it abstracts from nonhomotheticity (income effects), pointing to Jaravel and Lashkari (2021) for that extension. The primary welfare-relevant quantity reported is the quality-adjusted change in the cost of the imported goods basket, which is the import price index in the conventional sense.

Love-of-variety index: For a product i, defined as 1/(epsilon_it − 1) where epsilon_it is the product-level demand elasticity in an HA demand system. It measures the welfare value of having access to that variety and serves as the weight applied to expenditure share changes in the generalized price index formula (Proposition 2). In the CES special case all products share the same love-of-variety index 1/(sigma−1).

Homothetic with Aggregator (HA) demand: A family of income-invariant (homothetic) demand systems — including Kimball (1995), CRESH (Hanoch, 1971), and HSA (Matsuyama and Ushchev, 2017) — in which preferences are represented by a utility function with a specific aggregator structure. The key property exploited in the paper is that cross-product elasticities of substitution sigma_ij depend only on product-level elasticities epsilon_i through simple analytic formulas, reducing the dimensionality of the estimation problem from an N×N matrix to N scalars.

Endogeneity bias (in elasticity estimation): Downward bias in estimated elasticities of substitution arising from a positive correlation between product quality shocks and prices. When higher-quality products command higher prices and also have higher marginal costs, conventional methods (FBW) that assume zero correlation between supply and demand shocks will attribute part of the price variation to supply, underestimating how much demand responds to price. The paper documents this bias as the gap between DP and FBW estimates.

Heterogeneity bias (in elasticity estimation): Additional downward bias in CES elasticity estimates relative to the mean of Kimball elasticities, arising from CES imposing a single elasticity per industry when the true elasticities are heterogeneous across products. The bias is stronger for differentiated products and is theoretically traced to a positive covariance between demand elasticities and price volatility across products.

Dynamic panel (DP) identification: The paper’s proposed identification strategy, which exploits the Markov structure of quality shocks. The key moment condition is that quality shock innovations are mean-zero conditional on lagged prices, which holds under flexible pricing. Lagged prices (and higher-order lags and nonlinear transformations) serve as instruments for current prices, permitting identification of demand parameters without external cost instruments.

Quality shock (phi_it): An unobserved product characteristic that shifts demand for product i at time t, defined through the utility function as a scalar multiplying the quantity consumed. Quality is identified from residual demand — the component of demand not explained by price — following the approach of Khandelwal (2010) and Hallak and Schott (2011). The paper models quality shocks as following a stationary AR(1) process with product-specific means.

Unified CES price index (CUPI): The price index formula of Redding and Weinstein (2020a) for CES demand, which decomposes the aggregate price change into a price component (expenditure-share-weighted price changes) and a quality/taste component proportional to (1/(sigma−1)) times expenditure share changes. The present paper’s Proposition 2 generalizes CUPI to HA demand by replacing the scalar 1/(sigma−1) with product-specific love-of-variety indices.

Are Inflationary Shocks Regressive? A Feasible Set Approach

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question. The paper asks whether inflationary shocks are regressive, and demonstrates that the answer depends critically on the source of the shock. A single aggregate inflation statistic conceals radically different distributional consequences depending on whether inflation is driven by an oil supply contraction or by expansionary monetary policy.

Framework. The authors develop a “feasible set approach” grounded in the envelope theorem. They show that the first-order money-metric welfare effect of any macroeconomic shock on a household is summarized by the present discounted value of changes to five components of the household’s budget constraint: (1) consumption prices, (2) wage income, (3) asset dividends, (4) asset prices, and (5) government transfers. Because the envelope theorem implies that endogenous substitution responses are not welfare-relevant to a first order, no assumption about the utility function’s form or the economy’s general equilibrium structure is required. The framework is valid for generic stationary shocks that do not directly shift household preferences.

Empirical Strategy. The welfare formula requires two inputs: (i) impulse response functions (IRFs) for all prices, dividends, wages, and unemployment, estimated using internal-instrument SVAR methods applied to two identified shocks — the Kanzig (2021) oil supply news shock (instrumented by oil futures surprises around OPEC announcements) and the Gertler-Karadi (2015) monetary policy shock (instrumented by fed funds futures surprises in 30-minute windows around FOMC announcements) — and (ii) cross-sectional data on consumption bundles, labor income, and asset portfolios from the CEX, CPS, SCF, and SIPP for three education groups (high school or less, some college, college-educated) across the full lifecycle. The baseline cross-section uses 2019 data. Shocks are normalized to produce comparable aggregate inflation responses: a 10% WTI oil price increase and a 25 basis point decline in the one-year Treasury yield each generate roughly 15–16 basis points of CPI-U inflation on impact, rising to approximately 34–35 basis points after two quarters.

Main Findings. Oil supply contractions are regressive and monetary expansions are progressive, and this divergence is primarily driven by the asset price channel, not the consumption price or labor income channels.

For the 10% oil supply shock: middle-aged households with high school education or less must be paid approximately $870 (around 2% of annual consumption) to be made whole relative to their pre-shock utility; college-educated middle-aged households, by contrast, gain the equivalent of approximately $833 (1.1% of annual consumption). Younger college-educated households (still net equity accumulators) gain around $572.

For the 25 basis point monetary rate cut: low-education households approximately break even (net welfare effect near $23), while middle-aged college-educated households must be paid approximately $4,051 (around 5.5% of annual consumption) to restore their pre-shock utility. Older college-educated households must be paid approximately $851.

Why asset prices dominate. Oil supply contractions reduce equity prices (S&P500 falls approximately 2% one year post-shock) and depress dividends (approximately 82 basis points), while leaving house prices and bond prices largely unaffected. Because middle-aged college-educated households are the primary accumulators of equities, they benefit from the price decline (cheaper future accumulation), making oil shocks progressive through this channel — but regressive overall once the consumption and labor income channels (both mildly regressive) are included. Monetary expansions do the opposite: equity prices rise approximately 3 percentage points on impact, house prices rise approximately 1.5% after three years, and dividends increase. These asset price increases hurt those in the accumulation phase — disproportionately middle-aged college-educated households — creating a progressive distributional pattern.

Consumption and labor income channels. Both shocks generate disproportionate inflation in motor fuel and fuel and utilities, and low-education households spend a larger share of their budget on these goods, making the consumption channel mildly regressive for both shocks. The labor income channel differs sharply: oil shocks raise unemployment (approximately 0.15 log points for low-education households two years post-shock) and reduce weekly earnings by 0.2–0.6 log points, mildly harming low-education workers; monetary expansions reduce unemployment (approximately 0.83 log points for low-education workers one year post-shock) and similarly benefit low-education households through the labor market, pushing toward progressivity.

Scope conditions. Results apply to short-run first-order welfare effects of identified stationary macroeconomic shocks (four-year horizon). The framework does not incorporate uncertainty shocks, preference shocks, or the role of hedging motives in portfolio choice. Results concern policy shocks rather than policy rules.

Robustness. Qualitative conclusions hold across six alternative specifications: incorporating borrowing constraints (with or without empirical death rates), adjusting for unemployment insurance replacement rates (approximately 6% true average replacement rate), allowing for log-linear trends in no-shock choices, and dropping aggregate CPI controls from IRF estimation.

In depth

Q1. What is the “feasible set approach” and how does it differ from prior work on inflation incidence?

A: The feasible set approach measures welfare effects through changes in the household’s entire budget constraint — consumption prices, wage income, asset dividends, asset prices, and government transfers — rather than focusing on any single channel. Prior work either examined the Fisher channel (net nominal positions), or consumption price heterogeneity, or labor income responses in isolation. The key insight is that the envelope theorem implies substitution responses are not welfare-relevant to a first order, so the money-metric welfare change is simply the discounted sum of changes in the five budget constraint components evaluated at pre-shock choices, without requiring knowledge of the utility function’s form or the economy’s general equilibrium structure.

Q2. Why is the asset price channel — rather than consumption prices — the dominant channel in both shocks?

A: Asset holdings are large relative to annual consumption (net worth averages $1.5 million for college-educated and $260,000 for high-school-educated households in 2019), so even modest percentage movements in asset prices generate large dollar welfare effects. By contrast, the budget shares on the goods most responsive to both shocks (motor fuel, fuel and utilities) are relatively modest, so the consumption channel, while mildly regressive, is quantitatively small relative to the portfolio channel. The portfolio channel accounts for roughly 0.5% of consumption gains for middle-aged college-educated households under the oil shock, while the consumption channel produces losses of only about 0.1% for college-educated and 0.25% for low-education households.

Q3. How does the direction of the equity price response differ between oil and monetary shocks, and why does this create opposite distributional effects?

A: An oil supply contraction reduces equity prices (approximately 2% decline one year post-shock) and dividends (approximately 82 basis points decline), while a monetary expansion raises equity prices (approximately 3 percentage points on impact, approximately 4% higher after four quarters) and increases dividends. The welfare effect of asset price changes falls on those who trade the asset, not those who merely hold it at a constant level: middle-aged college-educated households are the primary net accumulators of equity, so falling prices benefit them (they can buy more cheaply) while rising prices hurt them. This is the principal reason oil shocks appear progressive through the portfolio channel — but regressive overall — while monetary expansions are regressive through the portfolio channel and progressive overall.

Q4. What are the precise welfare numbers for oil supply shocks by education group (baseline, ages 22–65)?

A: From Table 3 (baseline row, lifecycle-weighted averages for ages 25–65): households with high school or less experience a welfare loss of approximately $798; those with some college experience a loss of approximately $816; and college-educated households experience a welfare gain of approximately $494. These numbers reflect the sum of the consumption, labor income, portfolio, and transfer channels over a 16-quarter horizon, discounted at the one-year Treasury yield.

Q5. What are the precise welfare numbers for monetary policy shocks by education group (baseline, ages 25–65)?

A: From Table 3 (baseline row): households with high school or less experience a small welfare gain of approximately $23; those with some college experience a welfare loss of approximately $1,278; and college-educated households experience a welfare loss of approximately $3,055. These losses for college-educated households are driven overwhelmingly by rising equity and house prices that raise the cost of planned asset accumulation.

Q6. How does the life cycle interact with the distributional incidence of both shocks?

A: There is substantial heterogeneity within education groups across the life cycle because asset accumulation and decumulation patterns are age-dependent. Under oil shocks, younger college-educated households (who are net equity accumulators) gain approximately $572, middle-aged college-educated households gain approximately $833, while older college-educated households lose approximately $69 (because they hold large equity positions and lose dividend income). Under monetary shocks, middle-aged college-educated households lose the most (approximately $4,051) because they are simultaneously accumulating equities and housing, both of which become more expensive. Older college-educated households lose less (approximately $851) because rising dividends on existing holdings partially offset the asset price cost. Low-education households are approximately flat across the life cycle under monetary shocks.

Q7. How does the consumption channel compare across education groups and across the two shocks?

A: The consumption channel is mildly regressive for both shocks, but of similar absolute magnitude across the two shocks because both generate similar inflation in motor fuel and fuel and utilities — the goods with the largest price response. Low-education households spend a larger share on motor fuel and fuel and utilities; as a result, they lose approximately 0.25% of consumption from the consumption channel under the oil shock, compared with less than 0.1% for college-educated households. For monetary shocks, the consumption channel affects all household types roughly equally in proportional terms.

Q8. How does the labor income channel differ between oil and monetary shocks across education groups?

A: Oil shocks raise unemployment disproportionately for low-education workers (approximately 0.15 log point increase after two years, roughly 0.68 standard deviations, compared with near-zero response for college-educated workers) and reduce weekly earnings by 0.2–0.6 log points across groups. Monetary expansions reverse this: a 25 basis point rate cut reduces log unemployment by approximately 0.83 log points for low-education workers and approximately 1.96 log points for college-educated workers after one year, with limited response in conditional wages. Thus the labor income channel pushes toward regressive incidence for oil shocks and toward progressive incidence for monetary expansions, though in both cases it is quantitatively smaller than the portfolio channel.

Q9. What is the role of housing in the portfolio channel?

A: Housing behaves simultaneously as a durable consumption good and a financial asset. A house price increase raises welfare for households planning to decumulate (sell) housing (primarily older households) through the portfolio channel, but also raises the implicit rental cost for those who use housing — a negative consumption-side effect. Monetary expansions raise house prices by approximately 1.5% after three years. College-educated households accumulate housing at a faster rate and earlier in the life cycle than low-education households, making them more exposed to the cost of rising house prices during the accumulation phase. This amplifies the progressive pattern of monetary shocks through the portfolio channel.

Q10. How does the paper handle the dual role of durable goods (vehicles and housing)?

A: Durable goods are treated as both a consumption good and a financial asset. The utility-relevant consumption price of a durable is proportional to the price times the depreciation rate per unit of use, capturing the “implicit rent” of ownership. On the asset side, the durable enters the portfolio channel like a zero-dividend financial asset. This allows the framework to correctly attribute, for example, that a rise in house prices hurts net accumulators (through the portfolio channel) while also raising the implicit cost of housing services (through the consumption channel), rather than treating house price appreciation as an unambiguous welfare gain for homeowners.

Q11. What happens to the main conclusions when borrowing constraints are introduced?

A: Incorporating net worth constraints (with either constant or empirical death rates) dampens the portfolio channel for young and middle-aged college-educated households, because rising asset prices relax borrowing constraints for these households, partially offsetting the welfare cost of more expensive accumulation. Under constant death rates with borrowing constraints, college-educated households’ oil shock welfare gain falls from +$494 to +$76; under empirical death rates, it becomes a loss of -$394. For monetary shocks, the college-educated loss falls from -$3,055 to -$1,718 (constant death rate) or -$1,036 (empirical death rates). Despite these quantitative changes, the qualitative conclusion — oil shocks are regressive, monetary expansions are progressive — holds across all specifications.

Q12. What is the implication of these findings for the policy interaction between oil shocks and monetary tightening?

A: If the monetary authority responds to oil-price-induced inflation with unexpected interest rate increases, it may exacerbate the distributional consequences of the initial oil shock. An oil supply contraction is already regressive (harming low-education households through consumption prices and labor market effects); a disinflationary monetary tightening would additionally harm low-education households through the labor income channel (higher unemployment, lower wages) while partially benefiting college-educated households through lower asset prices. The paper notes this policy interaction as noteworthy, while cautioning that the results concern identified policy shocks rather than policy rules.

Q13. How are the two shocks calibrated to be comparable?

A: The oil shock is normalized to a 10% increase in WTI crude oil prices (approximately one standard deviation of monthly oil price growth). The monetary shock is normalized to a 25 basis point decline in the one-year Treasury yield — chosen because it generates approximately the same aggregate CPI-U inflation response as the oil shock (approximately 15–16 basis points on impact, rising to approximately 34–35 basis points after two quarters). This normalization allows the paper to attribute the different distributional outcomes to the source of inflation rather than to differences in the aggregate inflation magnitude.

Q14. What role does the transfer channel play, and for whom?

A: The transfer channel is small relative to the other three channels for the vast majority of working-age households, because transfer income is less than $100 per month for most households under age 65. Social Security payments — the bulk of transfer income — are explicitly indexed to the CPI; the paper models them as moving with CPI with a one-year lag. The transfer channel exclusively benefits older households (those receiving Social Security), and its quantitative effect is modest even there. Transfer income is more than 20 times smaller than labor and asset income for prime-age households of all education groups.

Key Concepts

Feasible set approach. The paper’s organizing framework, in which the first-order welfare impact of a macroeconomic shock is measured by how the shock changes the household’s budget constraint (consumption prices, wage income, asset dividends, asset prices, and government transfers) evaluated at the household’s pre-shock choices. Substitution responses are not welfare-relevant to a first order by the envelope theorem.

Money-metric welfare gain. The willingness-to-pay measure used throughout: the welfare change from a shock divided by the household’s marginal utility of consumption at time zero, expressed in time-zero dollars. Interpreted as an equivalent variation — the amount the household must be paid or would give up to be indifferent to receiving the shock. Used because it places households with very different utility functions on a common dollar scale.

Portfolio channel. The component of the welfare formula capturing the effect of asset price and dividend changes on household welfare. Asset price changes are welfare-relevant only for households that trade (accumulate or decumulate) the asset: rising prices benefit sellers and harm buyers; falling prices benefit buyers and harm sellers. This is distinct from the “Fisher channel” in prior literature, which focuses on net nominal positions rather than on which households are in the accumulation versus decumulation phase.

Internal instrument SVAR. The time-series estimation procedure used throughout: the pre-estimated identified shock series (oil supply news or monetary policy surprise) is included as a variable ordered first in a recursive structural VAR for each outcome variable. This separates shock identification (using the published instruments and controls from Kanzig 2021 and Gertler-Karadi 2015) from IRF estimation for each outcome variable, allowing the use of the full available sample for each outcome series.

Oil supply news shock (Kanzig 2021). An identified supply shock to oil markets, constructed from changes in oil price futures in tight windows around OPEC production announcements. Used to capture exogenous cost-push inflation driven by supply constraints rather than demand.

Monetary policy shock (Gertler-Karadi 2015). An identified demand-side shock, constructed from federal funds rate futures surprises in 30-minute windows around FOMC announcements, instrumented into a monetary SVAR. Captures exogenous interest rate cuts that generate aggregate demand expansion and inflation.

Borrowing constraint wedge. An additional term that appears in the welfare formula when households face net worth constraints. Proportional to the Lagrange multiplier on the net worth constraint, it discounts future periods more heavily when constraints bind, and adds a term for the welfare value of relaxed constraints when asset prices rise. Identified from deviations from perfect consumption smoothing using CEX lifecycle consumption data.

Automation and Rent Dissipation

Mon, 01 Jan 0001 00:00:00 +0000

Acemoglu and Restrepo examine the effects of automation in economies where labor market distortions cause some workers to earn rents—wages above their opportunity cost or outside option. The central question is how the interplay between automation and these distortions shapes wages, inequality, and productivity. The paper makes three contributions: a theoretical framework identifying a rent dissipation mechanism, reduced-form empirical evidence using US data from 1980 to 2016, and a general equilibrium quantification of automation’s aggregate effects.

The theoretical framework extends the task model of Acemoglu and Restrepo (2022) to incorporate task-specific wage wedges. In this setup, a firm employing labor of type g in task x pays a wage equal to the base wage multiplied by an exogenous wedge capturing rents from efficiency wages, bargaining, licensing, regulations, or norms. Because these wedges artificially inflate labor costs in high-rent tasks, firms have a stronger incentive to automate precisely those tasks—automation saves more in labor costs where rents are highest. Proposition 3 establishes that endogenous adoption decisions are tilted toward high-rent tasks: the rent distribution in automated tasks first-order stochastically dominates the rent distribution across all tasks. This targeting generates the rent dissipation mechanism. The equilibrium is inefficient on both the intensive margin (too little employment in high-rent tasks) and the extensive margin (excessive automation of high-rent tasks that a social planner would prefer to keep labor-intensive).

The rent dissipation mechanism has three consequences identified theoretically. First, it amplifies average wage losses for exposed groups beyond what displacement alone would produce, pushing displaced workers toward lower-paying jobs. Second, it compresses within-group wage dispersion by concentrating losses at higher percentiles of the within-group distribution, generating a U-shaped pattern of wage changes: workers at low percentiles earn no rents and experience only base-wage adjustments, while workers between the 70th and 95th percentiles face the steepest declines due to loss of high-rent jobs. Third, it is inefficient: because the tasks targeted by automation are not those where wages reflect scarcity or skill but rather distortionary rents, a planner would have preferred more labor allocated to these tasks, and rent dissipation offsets part or all of the cost-saving productivity gains from automation.

The empirical analysis covers 500 detailed demographic groups defined by education (five levels), gender, five age groups, five race/ethnicity groups, and nativity. Task displacement is measured as a weighted sum of industry-level automation exposure using three proxies: adjusted industrial robot penetration, specialized software services, and dedicated machinery in value added. Workers in the middle and lower-middle of the wage distribution lost 15–20% of their tasks to automation between 1980 and 2016, while post-college workers saw few tasks automated.

A 10 percentage point increase in task displacement is associated with a 24% decline in group-level relative wages (β = −2.36, s.e. = 0.13), falling to 19% after controlling for gender, education, sectoral demand, and rent shifters (β = −1.90, s.e. = 0.29). The U-shaped pattern in within-group wage changes is clearly visible: wages decline by 25–30% per 10 percentage point task displacement at the 70th–90th percentiles, compared to only 16% at the 5th–40th percentiles. Decomposing the average wage effect, the base-wage component is β = −1.53 (s.e. = 0.33) and the rent-dissipation component is β = −0.37 (s.e. = 0.11), implying a rent dissipation rate of approximately 37%. Across multiple proxies for rents—inter-industry/occupation wage differentials, wage losses after job displacement, and quit rates—the average estimated rent dissipation rate is approximately 35%. Rent dissipation accounts for one-fifth of the overall relative wage decline experienced by groups exposed to automation.

In the general equilibrium quantification (with elasticity of substitution λ = 0.5, average cost savings π = 30%, and average rent in automated tasks of 35%), automation accounts for 52% of the rise in between-group wage inequality since 1980: 42 percentage points via baseline displacement effects on labor demand, and 10 percentage points via rent dissipation. Cost savings from automation increased TFP by approximately 3% between 1980 and 2016, but inefficient rent dissipation offsets 60–90% of these gains, leaving net TFP gains of only 0.3–1.3% and net aggregate consumption gains of only 0.45–1.95% over the 36-year period.

Q: What is the rent dissipation mechanism, and why does it arise? A: Rent dissipation arises because labor market wedges make high-rent tasks artificially costly to staff with workers, giving firms a stronger incentive to automate precisely those tasks. When automation displaces workers from high-rent jobs, workers lose the premium above their opportunity cost that those jobs paid, amplifying wage losses beyond what displacement alone would cause. The mechanism is endogenous: firms do not randomly automate tasks but disproportionately target tasks where rents are highest, since doing so saves the most in labor costs. Proposition 3 formalizes this as first-order stochastic dominance of the rent distribution in automated tasks over the rent distribution in all tasks.

Q: Why is rent dissipation inefficient? A: In a distorted economy, high-rent tasks already feature too little employment at the equilibrium—firms under-hire in these tasks because the wage wedge makes labor artificially expensive. A social planner would want to allocate more labor to these tasks, not less. When automation further removes labor from high-rent tasks, it moves the economy further from the efficient allocation, dissipating rents that reflect distortions rather than true scarcity. The TFP formula shows that this inefficient targeting offsets part or all of the cost-saving gains from automation, and can even reduce aggregate productivity if the cost savings are small relative to the rent losses.

Q: What is the U-shaped pattern of within-group wage changes, and what does it indicate? A: The U-shaped pattern means that wage declines due to automation are smallest at the bottom percentiles of a group’s within-group wage distribution, largest in the 70th–95th percentile range, and then smaller again at the very top. Workers at low percentiles earn no rents, so they experience only the base-wage adjustment from reduced labor demand. Workers in the middle-upper range of the distribution hold the high-rent jobs that are disproportionately automated, so they lose both the base-wage component and the rent component of their wages. This pattern is directly visible in US data 1980–2016, with declines of 25–30% per 10 percentage point task displacement at the 70th–90th percentiles versus 16% at the 5th–40th percentiles.

Q: How is task displacement measured, and which groups are most exposed? A: Task displacement is measured as a weighted sum of industry-level automation exposure, accounting for each demographic group’s specialization in routine tasks within industries. Three proxies are used: the adjusted penetration of industrial robots, the increase in specialized software services, and the increase in dedicated machinery in value added. Workers in the middle and lower-middle of the wage distribution—broadly corresponding to non-college workers—lost 15–20% of their tasks to automation between 1980 and 2016. Post-college degree workers saw few tasks automated.

Q: How large is the rent dissipation rate, and how robust is this estimate? A: The baseline estimate from the U-shaped within-group wage change decomposition implies a rent dissipation rate (μ_Ag/μ_g − 1) of approximately 37% (β = −0.37, s.e. = 0.11). Using inter-industry and occupation wage differentials as a proxy for rents, the estimate is 39% (β = −0.39, s.e. = 0.11). Using wage losses after job displacement, the estimate is 20% (β = −0.20, s.e. = 0.04). After purging compensating differentials from the wage differential proxy the estimate remains 37%; after purging from the displacement-loss proxy it falls to 19%. Quit-rate evidence is consistent with rent dissipation: automation shifts workers toward higher-quit-rate jobs, which are lower-rent jobs. The average across proxies is approximately 35%.

Q: How much of between-group wage inequality since 1980 does automation explain, and what share is due to rent dissipation specifically? A: Automation accounts for 52% of the rise in between-group wage inequality in the US since 1980. Of this 52 percentage points, 42 percentage points are attributable to the baseline displacement effect working through reduced labor demand for exposed groups. The remaining 10 percentage points are attributable to rent dissipation—automation pushing exposed groups away from high-rent tasks into lower-paying employment. Rent dissipation thus accounts for roughly one-fifth (10/52) of automation’s total contribution to between-group inequality.

Q: How large are the productivity gains from automation, and how much does rent dissipation offset them? A: Cost savings from automation increased TFP by approximately 3% between 1980 and 2016. However, inefficient rent dissipation offsets 60–90% of these gains, because automation disproportionately targets high-rent tasks rather than tasks where the efficiency case is strongest. The net TFP increase attributable to automation is only 0.3–1.3% over the 36-year period, and the corresponding net increase in aggregate consumption is only 0.45–1.95%.

Q: How does automation affect within-group versus between-group inequality, and why is this notable? A: Automation increases between-group inequality by reducing relative wages of exposed groups (largely non-college workers) relative to unexposed groups, accounting for 52% of the rise in between-group inequality since 1980. At the same time, automation reduces within-group wage dispersion for exposed groups by compressing wages at higher percentiles. This contrasts with the standard view that inequality is fractal—rising at all levels of aggregation due to skill-biased demand—and helps explain why within-group inequality has risen steadily for college workers since the 1980s while remaining flat and then declining for non-college workers since the 1990s.

Q: What do the propagation matrix and rent-impact matrix represent in the general equilibrium analysis? A: The propagation matrix encodes how task reallocation due to automation in one demographic group creates competition for marginal tasks across other groups, transmitting the wage effects of automation to groups not directly displaced. The rent-impact matrix encodes how this task reallocation changes the rent composition of employment across groups. Both matrices are estimated from US data on task shares and group-level wage elasticities and are used to translate partial-equilibrium estimates of task displacement and rent dissipation into general equilibrium effects on wages and productivity for all demographic groups simultaneously.

Q: What are the policy implications of inefficient rent dissipation? A: Because rent dissipation is inefficient, the social value of automation is lower than what firms and consumers are willing to pay—firms capture all the labor cost savings but do not internalize the welfare cost of destroying high-rent jobs that the distorted equilibrium already under-supplies. Second-best interventions should address the underlying distortions generating rents rather than trying to slow automation directly. The paper suggests that strengthening labor market institutions supporting worker rents in non-automatable tasks could partially counteract the adverse distributional consequences of automation.

Q: How does this paper relate to Bound and Johnson (1992) and Borjas and Ramey (1995)? A: Bound and Johnson (1992) decompose changes in the US wage structure between 1979 and 1988 into technology, supply, and rent components (modeled as exogenous industry wedges), finding that 10–20% of between-group wage changes reflect rent losses. Borjas and Ramey (1995) estimate that trade increased the college premium by 1.3–2.6 log points between 1976 and 1990, with 15–33% due to loss of rents from trade-exposed jobs. Both are comparable to this paper’s finding that rent dissipation accounts for one-fifth of the wage effect of automation, though Bound and Johnson’s estimates include all factors affecting rents while this paper isolates automation specifically.

Worker rents: Wages above a worker’s opportunity cost or outside option, arising from efficiency wages, bargaining, licensing, regulations, or norms. Modeled as task-specific multiplicative wedges (μ_gx ≥ 1) that force firms to pay more than the base wage for labor in particular tasks. Explicitly excludes compensating differentials and skill premia.

Rent dissipation: The loss of above-opportunity-cost wages experienced by workers displaced from high-rent tasks into lower-paying employment. Occurs because automation endogenously targets high-rent tasks where labor is most expensive, and pushes workers into tasks where rents are lower. Quantified as the ratio of average rents in automated tasks to average rents across all tasks, minus one (approximately 35% in US data 1980–2016).

Task displacement: The share of tasks performed by a demographic group that are automated away, measured as a weighted sum of industry-level automation exposure accounting for the group’s specialization in routine tasks. Distinct from employment loss because it captures reallocation of tasks from labor to capital within the production function.

U-shaped within-group wage change profile: The pattern whereby automation generates the largest wage declines at intermediate-to-upper percentiles (70th–95th) of an exposed group’s within-group wage distribution, with smaller declines at the bottom, because high-percentile workers disproportionately hold high-rent jobs targeted by automation. Predicted theoretically and confirmed empirically in US data 1980–2016.

Propagation matrix: A matrix estimated from US data on task shares and group-level wage elasticities that encodes how automation of tasks performed by one demographic group creates competition for marginal tasks with other groups, transmitting wage effects across the demographic distribution in general equilibrium.

Inefficient automation targeting: The mechanism by which labor market distortions cause firms to automate high-rent tasks that a social planner would prefer to keep labor-intensive, since the distorted equilibrium already features too little employment in those tasks. Results in rent dissipation offsetting 60–90% of automation’s direct TFP gains from cost savings.

Rent-impact matrix: A matrix that encodes how task reallocation due to automation changes the rent composition of employment across demographic groups, used alongside the propagation matrix to compute general equilibrium effects of automation on wages and productivity accounting for distortions.

Bargaining and Inequality in the Labor Market

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question. How prevalent is individual wage bargaining in the labor market, what determines firms’ bargaining strategies, how do bargaining encounters unfold for workers, and does heterogeneity in bargaining behavior translate into wage inequality—including the gender wage gap?

Data and Setting. The paper develops and validates novel linked survey data for Germany. A firm survey was fielded by the ifo Institute to senior HR professionals and managers in two waves (September 2021 and January 2022), yielding 772 complete responses across all major sectors and regions. These responses were linked—with consent obtained from 72% of firms—to German Social Security records (the Integrated Employment Biographies, IEB) covering 416,821 full-time employees at matched firms in 2020, and to Orbis balance sheet data for firm productivity proxies. A separate worker survey was fielded by the IAB to 135,000 full-time German workers, with 9,756 completing it; nearly 10,000 responses were used for analysis, with 7,079 workers employed at surveyed firms. The worker survey elicited detailed bargaining histories for workers who had received an outside offer in the prior six months, bargaining at the start of current employment (for workers with tenure of three years or less), and responses to a hypothetical salary expectation scenario.

Definition of Individual Bargaining. The authors define a firm as having a “bargaining strategy” if it differentiates pay between workers in the same position it perceives to have similar productivity—encompassing both variation in initial offers (which may reflect firms using information on workers’ salary expectations) and back-and-forth negotiation. Elicitation distinguishes four employee groups (recent labor market entrants, experienced non-managers, managers, and bottleneck-occupation workers) and two contexts (new external hires and incumbent workers who receive an outside offer).

Prevalence of Bargaining. Approximately 50% of surveyed firms are willing to differentiate base wages for recent labor market entrants, more than 80% for experienced non-managers and managers, and nearly all for workers in bottleneck occupations they are struggling to fill. For incumbent workers facing outside offers, 57% of firms would increase pay for recent entrants, and more than 80% for experienced incumbents, managers, and bottleneck workers. In total, 80% of workers in the sample are in positions where individual bargaining is possible.

Magnitude of Wage Differentiation. For new external hires, the typical firm expects a gap between the highest and lowest offers of 3% for recent entrants, 5% for experienced non-managers, and 10% for managers (conditional on a gap: 6%, 10%, and 12% respectively). For incumbent workers responding to outside offers, the typical firm will adjust pay by 3% for recent entrants, 6% for experienced non-managers, and 10% for managers (conditional on responding: 6%, 7%, and 14% respectively). Forty-four percent of firms report that variation in initial offers is at least as important as back-and-forth negotiation in determining workers’ final pay.

Predictors of Firm Bargaining Strategies. Contrary to models predicting more productive firms are more likely to bargain (Doniger 2015; Postel-Vinay and Robin 2004; Flinn and Mullins 2021), firms that bargain are not more productive—as proxied by firm age, size, or assets per employee—nor do they pay higher mean wages. A variance decomposition shows that employee-group dummies alone explain 33% of variation in bargaining strategies for new hires, comparable to more than 500 firm dummies. Labor market factors—particularly whether a position is hard to fill—are systematically associated with bargaining willingness. Collective bargaining agreement (CBA) coverage and East German location are negatively correlated with bargaining flexibility.

How Bargaining Unfolds. In 57% of worker-firm interactions, the worker provides salary expectations before the firm makes its initial offer; 29% of firms require this information. About one-third of applicants ask for more after the initial offer, requesting on average a 3% increase; conditional on asking, about half of firms raise the offer, but fewer than one-third match what was requested, with the typical worker improving the offer by 1.5%. The majority of outside offers are rejected: only 9% of workers who received an outside offer in the prior six months chose to move to a new firm. Of the 91% who remained at their incumbent firm, 13% successfully renegotiated their pay. Back-and-forth dynamics—where offers are accepted or rejected only after multiple rounds—are consistent with models of two-sided incomplete information.

Worker Heterogeneity and Wage Inequality. Workers with better self-assessed outside options are 9 percentage points more likely to ask for an increase after the initial offer and 7 percentage points more likely to successfully negotiate a raise, relative to same-occupation coworkers with worse outside options. Women are 6 percentage points less likely to successfully negotiate their pay upward and show lower salary expectation provision rates, including in a hypothetical scenario in which pay range information is equalized. These gender differences in bargaining are not explained by women negotiating more over non-wage amenities; controlling for outside options and risk tolerance shrinks the female coefficient by at most 15%. Among surveyed workers, after controlling for occupation-establishment fixed effects, there is no gender wage gap at firms that do not bargain, but a 4–5 percentage point gender wage gap at firms that do bargain. Across specifications, firms that engage in individual bargaining have a 3 percentage point higher gender wage gap. A simple decomposition suggests that at surveyed firms, 44% of the residual gender pay gap can be attributed to bargaining. For workers at bargaining firms, a 10 percentage point higher pay premium at the prior firm is associated with 0.5 percent higher pay at the current firm, conditional on occupation-establishment fixed effects; this relationship is statistically insignificant for workers at non-bargaining firms.

Scope Conditions. Results apply to full-time private-sector workers in Germany between ages 25 and 50, with the firm sample over-representing medium and large firms (median size 50–249 employees). CBA coverage in the sample (41%) reflects Germany’s institutional context where firms retain the right to pay above CBA floors. Results are robust to re-weighting to match the overall distribution of German firm size and sector.

In depth

Q1. How do the authors define “individual bargaining” and why is this definition broader than standard labor economics usage?

The authors define a firm as having a bargaining strategy if it differentiates pay between workers in the same position it perceives to have similar productivity, covering both tailoring of initial offers and back-and-forth negotiation. Standard labor economics definitions typically condition on wages being set ex post once outside options are revealed, and focus on back-and-forth negotiation alone. The authors’ definition is most analogous to standard definitions of price discrimination. Empirically, the vast majority of firms that differentiate initial offers (93%) are also willing to engage in back-and-forth negotiation.

Q2. How was the firm survey designed to elicit bargaining strategies reliably, and what is the “protocol question”?

The protocol question asked: “How much more could a person maximally receive compared to the fixed compensation you would have offered based on the person’s qualification/fit for the position alone?” with options ranging from “0%/no adjustments possible” to “more than 40%.” Wording was developed through over 100 conversations with HR professionals; “qualifications and fit” was the phrase most closely aligned with HR professionals’ concept of productivity. The survey was fielded by the ifo Institute—an organization with decades of experience surveying this population—with a 51% response rate, 83% completion rate, and median response time of 11 minutes.

Q3. What validation exercises support the reliability of the elicited firm bargaining measures?

Four exercises are reported. First, intra-respondent reliability: the cross-tabulations between the protocol and incidence questions show most mass on or below the diagonal (incidence-implied spread no greater than the protocol-implied flexibility). Second, inter-respondent reliability: among 37 firms with multiple respondents, there is significant overlap in independently provided answers. Third, external validity using publicly available data: for 90% of firms reporting no CBA, no CBA evidence is found; for 99% reporting no pay information in job ads, none is found in online postings; for 82% reporting no salary expectation elicitation, no evidence of it appears in online application forms. Fourth, the elicited firm strategies are highly correlated with the matching workers’ survey responses—e.g., workers at firms stating they elicit salary expectations are significantly more likely to report having provided these expectations.

Q4. Is firm productivity associated with whether a firm engages in individual bargaining?

No. Firms that bargain and those that do not are similar with respect to firm size, firm age, and total assets per employee, and they also do not differ significantly in their AKM wage premium. These findings are inconsistent with theoretical models predicting that more productive firms are more likely to set pay via bargaining (Doniger 2015; Postel-Vinay and Robin 2004; Flinn and Mullins 2021). The result holds for both binary and continuous measures of bargaining, and is not overturned by machine learning prediction attempts.

Q5. What firm characteristics other than productivity predict bargaining strategies?

CBA coverage is negatively correlated with wage flexibility—CBA-covered firms report less flexibility even for managers who are typically exempt from CBAs and for groups not covered by CBAs, suggesting institutional norms or culture matter. Firms headquartered in East Germany are less likely to bargain with workers in all groups. Publicly traded firms (stock-based corporations) are more likely to set wages flexibly. These correlations are consistent with the view that managerial style and firm culture (rather than productivity) shape wage-setting strategies.

Q6. What does the variance decomposition say about the relative importance of firm versus market factors in predicting bargaining strategies?

Employee-group dummies alone explain 33% of the variation in bargaining strategies for new hires. After adjusting for the number of fixed effects used, four employee-group dummies explain as much variation as more than 500 firm dummies. Adding firm characteristics or coarse industry dummies does not significantly improve the adjusted R-squared relative to a model containing only group dummies. This supports models emphasizing market-level factors (worker replaceability, labor market tightness) over firm-level factors.

Q7. How common is it for workers to provide salary expectations before receiving an initial offer, and what do firms do with this information?

In 57% of worker-firm interactions, the worker provides salary expectations before the firm makes its initial offer. Twenty-nine percent of firms require this information; most ask for it. Forty-four percent of firms report that variation in initial offers is at least as important as subsequent back-and-forth negotiations in determining workers’ final pay. HR professionals and prior research indicate firms interpret variation in stated expectations as reflecting outside options rather than productivity.

Q8. What fraction of outside offers are rejected, and what happens when workers stay at the incumbent firm?

Only 9% of workers who received one or more outside offers in the prior six months chose to move to a new firm. Of the 91% who remained at the incumbent firm, 13% successfully renegotiated their pay at the incumbent. A follow-up survey fielded in spring 2024 corroborates this finding, showing approximately 80% of workers who received an outside offer remained at the incumbent firm; even recoding all job-to-job transitions as accepted offers implies no more than 26% of offers lead to a transition.

Q9. What do the back-and-forth dynamics imply for appropriate theoretical models of wage bargaining?

That many offers are accepted or rejected only after multiple rounds of negotiation is difficult to rationalize with models assuming either firms or workers have perfect information, which typically predict immediate acceptance or rejection. The patterns are consistent with models of two-sided incomplete information (Perry 1986; Chatterjee and Samuelson 1983). Sixty-nine percent of HR professionals in the survey report that decision-makers at their firm only have market-level information on wages, not specific information on what competitors pay.

Q10. How do outside options predict worker bargaining behavior and outcomes, controlling for occupation-establishment fixed effects?

Workers who rated it “easy” or “very easy” to obtain a better outside offer are 9 percentage points more likely to ask for an increase after the initial offer and 7 percentage points more likely to successfully negotiate a raise relative to same-occupation-establishment coworkers who rated it “difficult” or “very difficult.” The same pattern persists during the employment spell: workers with better outside options are 9 percentage points more likely to initiate and 8 percentage points more likely to succeed in renegotiation. These workers are not more likely to receive raises without asking.

Q11. How does risk tolerance predict bargaining, and how does it compare to outside options?

Workers with greater risk tolerance (those rating themselves 7 or above on a 10-point scale) are more likely to engage in wage negotiations and more likely to succeed both at the start of and during employment spells. Gaps in successful negotiations are somewhat larger than gaps in attempted negotiations, suggesting risk-tolerant workers also negotiate more effectively. However, outside options explain more of the between-worker variation in bargaining behavior than risk tolerance does.

Q12. What are the gender differences in bargaining behavior, and can they be explained by differences in outside options or risk tolerance?

Women are less likely to engage in back-and-forth negotiations and are 6 percentage points less likely to successfully negotiate pay upward during an employment spell. Women are also less likely to provide salary expectations and provide lower expectations as a fraction of their current salary in the hypothetical scenario, including when the salary range is provided—women are 6 percentage points less likely to provide expectations above the top of the stated range. Controlling for outside options and risk tolerance shrinks the female coefficient by at most 15%. There is no evidence that women substitute toward negotiating for non-wage amenities. The pattern is most consistent with women finding negotiation uncomfortable, not with a belief that it will not pay off or fear of backlash.

Q13. What is the estimated gender wage gap attributable to individual bargaining?

Among surveyed workers, after controlling for occupation-establishment fixed effects, there is no gender wage gap at firms without individual bargaining (coefficient closes to zero), while a 4–5 percentage point gender wage gap persists at firms with individual bargaining. This difference is robust across measures of pay (total daily pay, base pay, pay conditioning on hours worked), alternative fixed effect specifications, and to including non-surveyed workers at surveyed firms. A simple decomposition suggests 44% of the residual gender pay gap at surveyed firms can be attributed to bargaining. Across the interaction specifications, bargaining firms have a 3 percentage point higher gender wage gap and—in one key specification—a 6 percentage point difference between the gender gaps at bargaining and non-bargaining firms.

Q14. How does a worker’s prior firm wage premium affect current wages, and does bargaining status matter?

In a regression of log current wages on the AKM wage premium of the prior firm (conditional on occupation-establishment fixed effects), a 10 percentage point higher pay premium at the prior firm is associated with 0.5 percent higher pay at the new firm for workers at bargaining firms. For workers whose pay is not set via individual bargaining, the relationship between the prior firm’s pay premium and current pay is statistically insignificant. The result is consistent with the idea that during negotiations with a new firm, workers use their prior firm’s pay policy as an outside option.

Q15. How do AKM person effects relate to bargaining behavior?

Higher-person-effect individuals are more likely to have provided salary expectations when applying to their current firm and ask for a larger fraction of their current salary in the hypothetical scenario (conditional on their wage). These differences persist when controlling for occupation-establishment fixed effects and age and experience. Higher-person-effect workers are not more likely to receive raises without asking. These results are inconsistent with AKM person effects reflecting only productivity differences and instead suggest that fixed differences in individual bargaining behavior contribute to the variance in person effects—which Card, Heining, and Kline (2013) estimated explains a large share (40%) of the growth in German wage inequality.

Q16. Are the bargaining patterns found at surveyed firms representative of bargaining more broadly?

Two robustness exercises support broader representativeness. First, similar bargaining dynamics are found when including a random sample of German workers employed at non-surveyed firms. Second, re-weighting the sample to match the overall distribution of firm size and sector in Germany yields similar results. Because medium and large firms are over-represented in the firm sample, and because small firms hire infrequently and are less likely to have formal bargaining strategies, the true prevalence of individual bargaining among all German firms may be somewhat lower.

Key Concepts

Individual Bargaining Strategy (firm-level). A firm has an individual bargaining strategy if it differentiates pay between workers in the same position that it perceives to have similar productivity. This definition encompasses both tailoring of initial offers (based on, e.g., workers’ stated salary expectations) and back-and-forth negotiation. It is analogous to price discrimination rather than to the standard labor economics distinction between wage posting and Nash bargaining.

Protocol Question. The main survey measure of firm bargaining strategies: firms are asked the maximum percentage by which pay could be increased for a new hire above the fixed compensation the firm would have offered based on qualifications and fit alone, with response bins from “0%/no adjustments” to “more than 40%.” A zero response is used to classify a firm as not bargaining.

Incidence Question. A supplementary survey measure eliciting the expected spread (between highest and lowest offers) that the firm would make to ten candidates with identical qualifications and fit but differing stated salary expectations and competing offers. Used to validate the protocol question and to quantify the importance of initial-offer differentiation relative to back-and-forth negotiation.

Bottleneck Occupation. A firm-defined category of workers in positions that are particularly difficult to fill, drawing on an official German Federal Employment Agency designation. In the paper, bargaining willingness is systematically higher for workers in these positions than for other workers at the same firm, providing evidence that labor market tightness drives bargaining strategies.

Outside Offer Renegotiation. Wage renegotiation at the incumbent firm triggered by a worker receiving an outside offer, without a change in job tasks. The paper documents this is empirically more common than actual job-to-job transitions: of workers receiving outside offers, 91% remain at the incumbent firm, and 13% of those who remain successfully renegotiate their pay.

AKM Person Effect. A worker fixed effect estimated from a two-way fixed effects regression of log wages on worker and firm fixed effects (following Abowd, Kramarz, and Margolis 1999). In this paper, AKM person effects are taken from Bellmann et al. (2020), estimated over 2010–2017 German population data. The paper provides evidence that these effects capture, in part, fixed differences in individual bargaining behavior rather than solely differences in productivity.

AKM Firm Effect (Wage Premium). The firm fixed effect from the same two-way fixed effects regression, representing the pay premium a firm pays relative to what would be expected given its workforce composition. The paper uses the prior firm’s AKM effect as a measure of a worker’s outside option quality when testing whether prior-firm pay policy influences current pay under individual bargaining.

Salary Expectations (Gehaltsvorstellungen). The wage figure a worker provides to a prospective employer, typically before the firm’s initial offer. Legally, German firms (like most US states) cannot ask for salary history but can ask for salary expectations. In the paper, 57% of worker-firm interactions begin with the worker providing expectations; firms report using these to tailor initial offers, interpreting variation in stated expectations as reflecting outside options rather than productivity.

Barriers to Global Capital Allocation

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. Why do observed international investment positions and cross-country differences in rates of return to capital fail to conform to a frictionless capital-market benchmark? The paper asks how large the efficiency and distributional costs of barriers to global capital allocation are, and which frictions — capital income taxes, political risk, and geographic/cultural/linguistic distances — matter most.

Model. The authors develop a multi-country dynamic spatial general equilibrium model in which the entire network of bilateral cross-border investment positions is endogenously determined. Production in each country i follows a three-factor Cobb-Douglas function in reproducible capital, labor, and natural resources, with country-varying income shares. Capital is the only mobile factor. A logit asset demand system governs portfolio shares: the share of country j’s savings invested in country i is proportional to the risk-adjusted expected return on capital in i, scaled by the capital stock of i, and inversely proportional to a bilateral portfolio wedge ∆ij. These wedges can be microfounded via either rational inattention (where wedges reflect the precision of prior beliefs about returns) or extreme-value-distributed transaction costs. The model admits multiple microfoundations but yields the same functional form and the same counterfactual welfare calculations regardless of interpretation.

Frictions measured. Three categories of frictions enter the empirical implementation: (a) bilateral capital income tax rates — a new dataset covering 225 countries (50,625 country pairs), constructed from corporate income tax rates and treaty-adjusted withholding tax rates on dividends and interest, further adjusted for effective tax rates accounting for tax-haven routing; (b) political risk, proxied by an ICRG composite index (excluding socioeconomic conditions) following Alfaro, Kalemli-Ozcan, and Volosovych (2008); (c) geo-political distance, comprising geographic distance, cultural distance (based on 496 World Values Survey questions across 116 countries), and linguistic distance (based on a language-family tree covering 6,737 languages and 242 countries). These distance measures are publicly available at geopoliticaldistance.org. The model covers 96 countries (9,216 dyads), representing 92% of world GDP in 2017.

Gravity Estimation. Bilateral investment data (restated for tax havens using the nationality-basis methodology of Coppola et al. 2020 and Damgaard et al. 2019) are regressed on cultural, geographic, and linguistic distance with origin and destination fixed effects. In OLS, a one-standard-deviation increase in cultural distance (0.023 units) is associated with a 24.0% decrease in foreign assets; geographic distance (0.977 units in logs) with a 78.6% decrease; linguistic distance (0.174 units) with a 51.5% decrease. These magnitudes are robust across OLS, PPML, and IV (using religious distance as an instrument for cultural distance). Under IV, the standardized effect of cultural distance on log foreign assets rises to −76.5%.

Tax haven analysis. A Tobit regression of the share of bilateral investment routed through tax havens on the estimated tax saving from routing through havens yields coefficients of 0.413–0.999 for equity and 1.001–1.777 for debt (across specifications with varying fixed effects), confirming that tax incentives are a primary driver of the discrepancy between residency-based and nationality-based bilateral positions.

Model fit (untargeted moments). The calibrated baseline model produces: (i) a correlation of 0.658 between model-implied and empirical rates of return to capital (vs. 0.325 for the frictionless benchmark), with a standard deviation of 0.417 (vs. 0.091 frictionless; data: 0.496); (ii) a correlation of 0.947 between model-implied and empirical capital per employee (vs. 0.918 frictionless); (iii) a correlation of 0.94 between model-implied and empirical home bias; the model reproduces the mean home bias of 3.973 vs. 4.006 in data and standard deviation of 1.065 vs. 1.224, while the frictionless benchmark produces exactly zero home bias for all countries. Portfolio-share MSE: 1.16 (baseline) vs. 1.86 (frictionless).

Counterfactual findings. Removing all measured barriers raises world GDP by 6.8% relative to the observed equilibrium (equivalent to stating that the distorted equilibrium is 6.8% below the frictionless benchmark). Geo-political distance alone accounts for most of this: when only distance frictions are retained, world GDP is 5.2% below the frictionless level. Capital taxes alone reduce world GDP by 2.6% below frictionless; political risk alone by 0.4%. The standard deviation of log capital per employee is 51.5% higher than it would be without barriers; the standard deviation of log output per employee is 22.5% higher. In the frictionless equilibrium, capital flows from rich to poor countries (the correlation between net foreign assets and development doubles in absolute value), accounting for the Lucas (1990) puzzle. In short-term (one-period) counterfactuals holding wealth fixed, the GDP gain from full barrier removal is 3.6%; the inequality effect remains similar (standard deviation of log capital per employee 48.4% higher with barriers).

Scope conditions. The model focuses on steady-state outcomes; dynamic transition effects are analyzed in extensions but are smaller. Quantitative conclusions are conditioned on: (i) the model sample of 96 countries covering 92% of world GDP in 2017; (ii) the conservative OLS coefficient estimates used for baseline calibration (IV estimates are larger and would amplify results); (iii) the assumption that the logit demand system captures frictions regardless of their microfoundation; (iv) omission of goods-trade frictions from the baseline (when included, the world GDP effect falls to 3.7% and the capital inequality effect to 23.3%).

In depth

Q1. What is the core theoretical prediction about cross-country rates of return when investment barriers exist?

A: In the model’s frictionless benchmark (Propositions 1 and 2), all origin countries hold identical portfolios and risk-adjusted expected returns are equalized across destinations. When bilateral frictions are introduced, countries that are more “peripheral” (harder to access for foreign investors due to high geo-political distance or political risk) receive less inward capital and therefore command higher physical rates of return to capital. Countries that are easily accessible (“central”) attract more capital and exhibit lower rates of return. The Dual Efficiency Theorem establishes that capital is efficiently allocated if and only if marginal products of capital are equalized across countries, which requires that taxes are uniform and that portfolio wedges satisfy a specific cancellation condition.

Q2. How are portfolio wedges measured, and what is the identifying strategy?

A: Portfolio wedges ∆ij are decomposed into a geo-political distance component and a political risk component. The geo-political distance component is specified as a log-linear function of geographic distance, cultural distance, and linguistic distance, with coefficients (β_g, β_c, β_l) estimated from a gravity regression of log bilateral investment on these distances, controlling for origin and destination fixed effects. Because political risk varies only by destination country, it cannot be separately identified from destination fixed effects in the bilateral regression; its elasticity is therefore taken from Alfaro, Kalemli-Ozcan, and Volosovych (2008). The key identification advantage of bilateral data is that origin and destination fixed effects absorb all country-level confounders, so the distance coefficients are identified purely from within-origin, within-destination variation across country pairs.

Q3. What do the OLS gravity regressions find, and are the coefficients stable across specifications?

A: In the baseline OLS specification (Table 2, column 1), the estimated coefficients on cultural distance, geographic distance, and linguistic distance are −11.944, −1.579, and −4.162 respectively (all significant at the 1% level). In standardized terms, a one-standard-deviation increase in cultural distance reduces foreign assets by 24.0%, geographic distance by 78.6%, and linguistic distance by 51.5%. Adding a rich set of control variables (colonial ties, legal origin, currency pegs, trade agreements, effective tax rates) leaves these magnitudes broadly similar: standardized effects on foreign assets are −26.4%, −80.1%, and −47.6%, respectively. Results are also robust across OLS and PPML specifications and across years 2013–2017. Effects are quantitatively similar for foreign equity and foreign debt, though linguistic distance has a somewhat smaller effect on debt.

Q4. How does the instrumental variable strategy address reverse causality in cultural distance, and what does it find?

A: The authors instrument cultural distance with religious distance (based on historical trees of religious affiliation), assuming religious history affects international investment only through its contemporary effect on differences in values and beliefs as captured by the World Values Survey. The instrument is a strong predictor of cultural distance (passes weak-instrument tests comfortably). Under IV, the standardized effect of a one-standard-deviation increase in cultural distance on log foreign assets rises from −24.0% (OLS) to −76.5% (IV). The authors use conservative OLS estimates for their baseline calibration, so the IV results imply the headline counterfactual effects are likely understated.

Q5. How does the model predict home bias, and how well does it match the data?

A: Home bias is defined as the log difference between the domestic portfolio share and the country’s share in the world capital stock. In the frictionless model, Proposition 1 implies that all countries hold identical foreign portfolios, so the model produces exactly zero home bias for every country. The baseline model, by incorporating bilateral frictions, generates home bias endogenously without targeting it. The model-implied home bias correlates with the empirically measured home bias at 0.94 across countries and matches both the mean (3.973 model vs. 4.006 data) and standard deviation (1.065 vs. 1.224) closely. The model also predicts, consistent with Lau, Ng, and Zhang (2010), that home bias and rates of return on capital are positively correlated (model-implied ρ = 0.55), and that rates of return on capital correlate negatively with the log of GDP per employee (model-implied ρ = −0.70).

Q6. What is the quantitative decomposition of the world GDP loss by type of barrier?

A: World GDP in the observed (distorted) equilibrium is measured at $112.9 trillion (PPP), which is 6.8% below the frictionless counterfactual. When all barriers are present except geo-political distance, world GDP is 5.2% below frictionless — meaning distance frictions account for the largest share. When all barriers are present except political risk, world GDP is only 0.4% below frictionless. When all barriers are present except taxes, world GDP is 2.6% below frictionless. These are not exactly additive because the distortions interact; the results confirm that geo-political distance (cultural, linguistic, and geographic) constitutes the dominant source of global capital misallocation among the three measured frictions.

Q7. How do barriers affect the cross-country distribution of capital and income?

A: The standard deviation of log capital per employee is 51.5% higher in the distorted equilibrium than in the frictionless counterfactual; the standard deviation of log output per employee is 22.5% higher. When only geo-political distance distortions are maintained, dispersion in log capital per employee is 38.2% higher and in log output per employee 15.9% higher. Maintaining only taxes raises the dispersion in log capital per employee by 12.9% and log output per employee by 6.0%; maintaining only political risk raises them by 7.3% and 3.8%, respectively. In the frictionless equilibrium, the poorest countries gain the most: some of the poorest countries see capital per employee increase by an order of magnitude and income per employee double.

Q8. Does the model account for the Lucas puzzle (capital not flowing from rich to poor countries)?

A: Yes. In the observed distorted equilibrium, net foreign asset positions correlate only weakly with the level of development, consistent with Lucas’s (1990) observation that capital fails to flow from rich to poor countries. In the frictionless counterfactual, the absolute value of the correlation between net foreign asset positions and log GDP per employee doubles, and capital indeed flows from rich to poor countries as neoclassical theory predicts. The distortions from taxes, political risk, and geo-political distance thus account for the absence of a strong correlation between net positions and development in the data.

Q9. How do extensions incorporating goods-trade frictions, capital controls, and currency hedging costs affect the headline findings?

A: Adding goods-trade frictions (country-specific prices for output and capital installation following Monge-Naranjo et al. 2019) reduces the world GDP effect to 3.7% (from 6.8% baseline) and the dispersion of log capital per employee to 23.3% higher (from 51.5%), but the overall pattern of results is preserved. Replacing political risk with capital controls (using Jahan and Wang 2016 de-jure capital account openness) yields a comparable world GDP loss of 6.6% and a geo-political distance effect of 6.2%, very close to the 6.8% and 5.2% in the baseline. Adding currency hedging costs leaves world GDP loss and inequality effects essentially unchanged relative to baseline. None of these extensions materially alters the headline conclusions.

Q10. How do the authors validate the model against nationality-based versus residency-based bilateral investment data?

A: The model is calibrated to nationality-based positions (restated for tax havens). The MSE for fitting nationality-based external portfolio shares is 1.16, while the MSE for residency-based positions is 1.22. The model was not explicitly designed to distinguish between the two, yet it naturally produces better predictions for nationality-based positions because its frictions incorporate the incentives for indirect investment routing through tax havens. This cross-validation supports the methodological approach of using nationality-restated data and confirms the internal consistency of the model’s treatment of tax-haven routing.

Q11. What are the implications for global tax policy coordination?

A: In the presence of information frictions, simple harmonization of capital tax rates across countries does not improve capital allocation efficiency and could worsen it. The Dual Efficiency Theorem implies that efficient capital allocation in a world with information frictions requires that taxes, risk premia, and information frictions satisfy a joint cancellation condition. From a normative perspective, a global social planner maximizing world GDP should impose lower capital tax rates in countries that are “peripheral” in the network of informational distances, in order to offset the disadvantage created by information frictions for those countries.

Q12. How is the elasticity parameter η calibrated, and how sensitive are the results?

A: The elasticity of substitution among countries’ assets, η, is calibrated at 18.5 based on Koijen and Yogo (2020)’s demand-price elasticities for long-term debt (3.1, converted to a gross-return elasticity of approximately 30), short-term debt (25.2, converted to approximately 24.3), and equity (1.3, converted to approximately 14.8), with weights reflecting the composition of global portfolios. The baseline gravity coefficients are calibrated from OLS with controls (cultural: −13.129, geographic: −1.645, linguistic: −3.850), chosen as conservative estimates relative to IV or PPML. Sensitivity analysis using PPML or IV estimates of β yields broadly similar steady-state GDP losses (around 6%), confirming robustness.

Key Concepts

Portfolio wedge (∆ij): A bilateral distortionary term in the logit asset demand system that captures all frictions reducing the ability of investors from country j to invest in country i. Decomposed empirically into a geo-political distance component and a political risk component. A wedge of 1 means no friction; larger values reduce the share of investment flowing from j to i. Can be interpreted either as prior-belief imprecision under rational inattention or as systematic transaction costs under the extreme-value microfoundation.

Geo-political distance: A composite of geographic distance (population-weighted geodesic distance), cultural distance (expected disagreement in World Values Survey responses between randomly drawn individuals from two countries, constructed with the “flex” method using up to 496 questions), and linguistic distance (normalized tree distance in the Ethnologue language family graph, covering 6,737 languages). Distinct from simple physical distance: it captures the informational and transactional barriers that arise from societal dissimilarity.

Dual Efficiency Theorem: A theoretical result (Theorem in Section 2.8) establishing that capital efficient allocation, equalization of marginal products of capital across countries, and uniform taxes combined with a specific cancellation condition on portfolio wedges are mutually equivalent statements in steady-state equilibrium. This is not a restatement of the First Welfare Theorem; it is a statement about GDP (not welfare) and does not require risk premia to be equalized.

Effective bilateral tax rate (τij): The composite bilateral tax rate on capital after accounting for tax-haven routing. Firms in the destination country optimally choose the share of capital issued through tax havens (solving a quadratic cost optimization), trading off the lower tax rate available through havens against an increasing quadratic routing cost. The effective rate is therefore lower than the statutory (de jure) rate when the tax-haven rate is lower than the statutory rate, with the gap depending on the estimated βth coefficient from the Tobit regressions.

Logit asset demand system: A portfolio allocation rule in which the share of country j’s savings invested in destination country i is proportional to the risk-adjusted expected return raised to the power η (the elasticity of substitution) times the destination capital stock, divided by the portfolio wedge and summed over all destinations. Microfounded either by rational inattention (Matejka and McKay 2015; Pellegrino 2023) or by extreme-value-distributed transaction costs. Produces portfolio gravity analogous to trade gravity when combined with the market clearing conditions.

Home bias: Defined as the log difference between a country’s domestic portfolio share (πii, the share of domestic savings invested at home) and that country’s share of world capital stock (ki/K). In the frictionless benchmark, home bias is exactly zero for all countries by Proposition 1. The baseline model generates home bias endogenously as a consequence of portfolio wedges and reproduces both the level and cross-sectional distribution of empirically observed home bias without targeting these moments directly.

Core-periphery structure: An emergent property of international capital markets under investment barriers: countries that are easily accessible to international investors (low geo-political distance, low political risk, favorable tax treatment) are “central” and attract capital inflows, driving their rates of return to capital lower; “peripheral” countries that are less accessible have smaller capital stocks and higher rates of return, compensating investors for overcoming barriers. This structure generates persistent capital misallocation and cross-country income inequality.

Nationality-based vs. residency-based bilateral investment positions: Residency-based data (e.g., raw IMF CPIS) attributes investment to the immediate counterparty country, including tax-haven shell companies. Nationality-based data (Coppola et al. 2020; Damgaard et al. 2019; Beck et al. 2024) reattributes investment to the country of the ultimate investor and ultimate issuer, bypassing offshore centers. The model fits nationality-based positions better (MSE 1.16 vs. 1.22 for residency-based) because it incorporates frictions that generate incentives for indirect routing, which is what nationality restatement is designed to undo.

Beliefs About the Economy are Excessively Sensitive to Household-Level Shocks: Evidence from Linked Survey and Administrative Data

Mon, 01 Jan 0001 00:00:00 +0000

Bottom-Up Markup Fluctuations

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

The paper asks how firm-level, sector-level, and aggregate markups comove with output at different levels of aggregation, and whether a single structural model can reconcile seemingly contradictory empirical findings about markup cyclicality that arise when researchers use different aggregation schemes.

Model

The authors build a granular macroeconomic model featuring oligopolistic competition with a nested constant-elasticity-of-substitution (CES) demand structure following Atkeson and Burstein (2008). The economy contains N sectors, each with a discrete number of firms competing under Cournot oligopoly with flexible prices. Firm-level markups are endogenously increasing in within-sector market shares: under Cournot, the sectoral markup is a simple function of the sector’s Herfindahl-Hirschman index (HHI), and the aggregate markup is a function of the expenditure-share-weighted average of sectoral HHIs. Firm-level productivity follows a discretized random growth (Gibrat’s law) process as in Carvalho and Grassi (2019), generating fat-tailed firm-size distributions and granular aggregate fluctuations. The baseline calibration features only idiosyncratic firm-level productivity shocks and abstracts from aggregate shocks, because—in the model—aggregate shocks that move all firms proportionately do not affect relative market shares and hence do not affect markups.

Data

The empirical analysis uses French administrative firm-level data from the FICUS-FARE datasets covering the universe of French firms from 1994 to 2019, yielding approximately 9.38 million firm-year observations across 26 years, 22 two-digit sectors, and 275 five-digit NAF sectors. Firm-level markups are estimated following De Loecker and Warzynski (2012) using a translog production function estimated by GMM (following De Ridder et al. 2024) on a subsample of approximately 220,733 firm-year observations where physical output quantity is available from the Enquete Annuelle de Production survey (2009-2019). Using quantity rather than revenue as the output measure avoids the measurement biases documented in Bond et al. (2021).

Main Findings and Quantitative Magnitudes

Markup-market-share relationship (firm level): Regressions of the change in the inverse firm markup on the change in firm market share yield a negative and significant coefficient of approximately -0.268 to -0.293 (depending on fixed-effect specification), consistent with the model prediction that markups rise with market share. Sector-level analogues yield a slope of the change in inverse sector markup on the change in sector HHI of approximately -0.37, which is simultaneously a calibration target (implying sigma = 1.8 given epsilon = 5) and an empirical moment the model closely matches (model counterpart: -0.36).
Within-between decomposition of sector markup changes: In the model under Cournot competition, changes in firm-level markups (the “within” term) account for exactly 50% of changes in sector-level markups, with between-firm reallocation accounting for the other 50%. In the French data, for the median sector, the within term accounts for 59% of changes in sector markups (interquartile range across sectors: 34%-81%).
Firm-level markup cyclicality with sector output (heterogeneous by size): The average firm’s markup is countercyclical with respect to own-sector output (beta_1 approximately -0.073 in levels specification), but this relationship reverses for large firms: firms with market shares roughly above 10% (top 0.1% of the market-share distribution) have procyclical markups (interaction coefficient beta_2 approximately 0.574 in levels). The model qualitatively and roughly quantitatively reproduces this heterogeneity.
Sector-level markup cyclicality with sector output (procyclical): Following Nekarda and Ramey (2013), sector markup changes comove positively and significantly with sector output changes: estimated coefficient of 0.160 (standard error 0.040) in first-differences. The calibrated model yields a median coefficient of 0.139 (std dev 0.057 across 5,000 simulated 25-year samples), close to the data. Consistently, sector concentration (HHI) is also procyclical with sector output (estimated coefficient 0.332, std error 0.067 in first-differences).
Sector-level markup cyclicality with aggregate output (acyclical to weakly countercyclical): Following Bils et al. (2018), the comovement between sector markups and aggregate output is fragile in sign and significance: the French data yields a point estimate of -0.239 (std error 0.116) in first-differences, marginally significant (t-stat 2.06) and with sign sensitive to detrending method. The model without aggregate shocks predicts positive comovement (median coefficient 0.165) that is not statistically different from zero across samples. Adding aggregate productivity shocks (calibrated to match French aggregate output volatility) brings the model-implied coefficient close to zero (median 0.008), with 20-30% of 25-year simulated samples displaying countercyclical sectoral markups relative to GDP—consistent with the ambiguity in the data.
Aggregate output volatility: The baseline calibration with only granular firm-level shocks generates a standard deviation of detrended aggregate output of 0.83%, equal to 26% of the 3.16% observed in the French data. (The comparable granular ratio from Carvalho and Grassi 2019 for a perfectly competitive US model is 30%.) Variable markups dampen granular aggregate volatility: the standard deviation of aggregate output under variable markups is 0.87 times that under heterogeneous-but-constant markups (95% CI: 0.82-0.97), because incomplete pass-through reduces the effective weight of large firms in the price index.
Aggregate markup volatility: In the data, the relative standard deviation of aggregate markup to aggregate output is 0.40-0.50 (depending on detrending). The model generates a relative volatility of 0.36 (median across samples). The correlation between aggregate markup and output in the data is at most 0.06; the model without aggregate shocks implies a counterfactually large median correlation of 0.91, which falls to 0.27 when aggregate TFP shocks are superimposed (with 16% of 25-year samples displaying countercyclical aggregate markups).

Scope Conditions

Results pertain to French private-sector firms (including formerly government-owned firms, most of which privatized during the sample period) across manufacturing and some non-manufacturing sectors at the national-market level. The analysis abstracts from import competition (market shares are computed relative to all French firms in the sector), local geographic markets (relevant for non-tradeable goods where national-level shares understate local concentration), and multi-product firm structure. Findings are for a flexible-price model driven by idiosyncratic productivity shocks; the paper explicitly discusses how nominal rigidities would further strengthen procyclicality at the sector level.

In depth

Q1. What is the central mechanism by which granular firm-level shocks generate markup cyclicality?

A: Because markups are endogenously increasing in within-sector market shares under oligopolistic competition, a firm that receives a positive productivity shock gains market share and therefore raises its markup, while its competitors lose market share and lower their markups. The net effect on the sectoral markup depends on the shocked firm’s initial size: a positive shock to a sufficiently large firm (above a threshold market share) raises the sectoral markup, while a positive shock to a small firm lowers it. Since sectoral expansions in a granular economy are disproportionately driven by large firms, sector output and sector markup tend to comove positively in the medium run.

Q2. Why does the sign of markup cyclicality differ depending on the level of aggregation?

A: Sector-level markups react only to within-sector idiosyncratic shocks, so sectors that happen to be driven by large-firm booms display positive comovement between sector markup and sector output. However, a given sector’s markup is uncorrelated with aggregate output movements coming from other sectors. In small samples (such as 25-year windows), whether a sector’s markup comoves positively or negatively with aggregate output depends on whether the sector happens to lead or lag the aggregate cycle. Over sufficiently long samples, the model implies positive comovement of sector markups with aggregate output, but in finite samples the relationship is indeterminate. This asymmetry across aggregation levels explains why researchers using different reduced-form specifications in the same dataset can reach opposing conclusions about procyclicality versus countercyclicality.

Q3. What is the within-between decomposition of sectoral markup changes and what does it imply quantitatively?

A: Changes in the inverse sectoral markup can be decomposed into (i) a within term—changes in firm-level markups holding market shares fixed—and (ii) a between term—changes in market shares holding firm-level markups fixed. Under Cournot competition, the within and between terms are analytically equal in every period, so each accounts for exactly 50% of the change in sectoral markups; this 50-50 split holds globally (not only to first order). In the French data, for the median sector, within-firm markup changes account for 59% of sector markup changes (interquartile range across sectors: 34%-81%), close to but slightly above the model’s 50% prediction.

Q4. How do variable markups affect granular aggregate output volatility relative to a model with constant markups?

A: Variable markups (endogenous pass-through that is decreasing in firm size) reduce granular aggregate output volatility relative to a model where markups are heterogeneous but fixed. The intuition is that larger firms have lower pass-through rates, so their productivity shocks translate into smaller price changes and therefore smaller output responses than they would under constant markups—effectively reducing the weight of large firms in the aggregate price index in a way similar to a decline in market concentration. Quantitatively, using first-order approximations around equilibrium distributions from the calibrated model, the standard deviation of aggregate output under variable markups is 0.87 times that under heterogeneous-but-constant markups (95% confidence interval: 0.82-0.97). The overall standard deviation under variable and heterogeneous markups is only 1.02 times that under homogeneous and constant markups (95% CI: 0.99-1.14), meaning markup heterogeneity and variability together have limited net effects on aggregate output volatility.

Q5. What does the model predict for firm-level markup cyclicality, and how heterogeneous is this across firm size?

A: Proposition 4 states that, in the asymptotic limit, firm-level markups comove positively with own-sector output for firms with market shares above a threshold, and negatively for firms below it. This occurs because large firms have a disproportionate impact on sector-level price and output (when the product of market share and pass-through rate is increasing in size), so large-firm shocks simultaneously drive sector expansions and raise large-firm markups while compressing small-firm markups. In the French data, the average firm’s markup is countercyclical with respect to sector output (beta_1 approximately -0.073 in log-levels with firm and year fixed effects), but firms with market shares above roughly 10% (top 0.1% of the distribution, since the average market share is only 0.07%) display procyclical markups (interaction coefficient beta_2 approximately 0.574). The model reproduces this qualitative pattern and the order of magnitude of these estimates.

Q6. How does the paper calibrate the key demand elasticities, and what are the resulting pass-through implications?

A: The within-sector substitution elasticity is set to epsilon = 5, a standard value. The cross-sector substitution elasticity sigma is calibrated to match the slope of the inverse sector markup on sector HHI in first-differences. The empirical slope is -0.37; under the model, the slope equals -(epsilon/sigma - 1)/(epsilon - 1), and given epsilon = 5, sigma = 1.8 delivers a model counterpart of -0.36. These parameter values imply own-cost pass-through rates that are decreasing in firm size; for large firms (with market share >= 57%, approximately the top 0.004% of the distribution), the implied pass-through rate is 0.63, within the confidence intervals reported in Amiti, Itskhoki, and Konings (2019) for large Belgian firms.

Q7. Why do aggregate productivity shocks not affect markups in the model, and what are the implications for aggregate markup cyclicality?

A: In the model, firm-level markups are functions of within-sector market shares, not the level of productivity. An aggregate shock that shifts all firms’ productivity proportionately leaves relative market shares unchanged and therefore leaves all markups unchanged. This means aggregate shocks increase aggregate output volatility but leave markup volatility unchanged, reducing the correlation between aggregate markup and aggregate output. When aggregate TFP shocks are added to match French aggregate output volatility, the model-implied median correlation between aggregate markup and output falls from 0.91 (without aggregate shocks) to 0.27 (with aggregate shocks), while 16% of 25-year simulated samples display countercyclical aggregate markups—more consistent with the weak and fragile empirical relationship.

Q8. How does the paper address the potential measurement-error bias in the negative correlation between markups and marginal costs?

A: Since marginal cost is computed as price divided by estimated markup, regressing market shares or markups on marginal costs risks spurious correlation via measurement error in the markup (which appears in both sides). The authors address this concern by constructing an instrumental variable for marginal cost based on firm-specific energy intensity interacted with energy price changes, following Ganapati, Shapiro, and Walker (2020). Table A10 confirms that instrumenting for marginal cost yields negative effects on both markup and market share with larger point estimates than the OLS specifications in Table 4, validating the baseline findings.

Q9. Is the 50-50 within-between decomposition of sectoral markup changes robust to the choice of competition mode?

A: No. The exact 50-50 split of within and between terms in sectoral markup changes is a specific property of Cournot competition and holds globally (not just as a first-order approximation). Under Bertrand competition, the within and between terms are generally not equal to each other. The paper derives analytic results under both competition modes and focuses on Cournot for quantitative work because it generates more markup variation and better matches the estimated pass-through rates and markup-size relationship.

Q10. What do model simulations imply for the magnitude and cyclicality of aggregate markups versus the data, and what is the role of variable versus constant markups?

A: In the data (detrended), the standard deviation of aggregate markup is 1.27% with a relative volatility (to output) of 0.40 and a correlation with output of 0.03. The baseline model with only granular shocks yields a median markup standard deviation of 0.30%, relative volatility of 0.36, and correlation with output of 0.91. The model with aggregate shocks added yields median markup standard deviation of 0.30%, relative volatility of 0.09, and correlation of 0.27. Counterfactually fixing markups at their initial heterogeneous levels while keeping the same market shares and shock variance yields aggregate markup standard deviation approximately 0.93 times the variable-markup value (standard deviation of markups under variable markups is 1.08 times that under constant markups, with a 95% CI of 1.00-1.18), and a correlation with output of 0.92 versus 0.87 under variable markups. Overall, the magnitude and cyclicality of aggregate markups are not substantially different between variable and constant-markup specifications.

Q11. How does the paper reconcile its findings with prior literature on markup cyclicality (Bils et al. 2018 vs. Nekarda and Ramey 2013)?

A: Nekarda and Ramey (2013) find procyclical sector markups with respect to sector output in US data—a result replicated in French data (beta approximately 0.160). Bils, Klenow, and Malin (2018) find countercyclical sector markups with respect to aggregate output in US data. Both results can be generated simultaneously in the model: sector markups are positively correlated with own-sector output because granular booms in a sector are driven by large-firm expansions that raise sector markups; however, a given sector’s markup is weakly and ambiguously correlated with aggregate output because aggregate fluctuations reflect shocks across many sectors, only some of which are in the same sector. The model can therefore simultaneously predict procyclicality with respect to sector output and an acyclical-to-weakly-countercyclical relationship with aggregate output—explaining why both empirical findings can be correct.

Q12. What are the data limitations and how do they affect the interpretation of results?

A: Three limitations are noted. First, market shares are computed relative to total revenue of all French firms in the sector without accounting for imports, so foreign competition is ignored and domestic concentration may be overestimated. Second, revenues are reported at the national level, so for non-tradeable goods (whose relevant market is local) the paper underestimates true local market concentration, attenuating the markup-concentration relationship in those sectors. Third, the model abstracts from entry and exit (the number of firms per sector is held fixed at sector-year averages), though Appendix D demonstrates robustness of main empirical results to restricting the sample to continuing firms.

Key Concepts

Granular macroeconomic model: A model in which the economy consists of a finite (large but discrete) number of firms, so that idiosyncratic firm-level shocks to large firms do not average out and instead generate aggregate fluctuations. In the paper’s usage, granularity means that sectoral and aggregate business-cycle fluctuations are driven primarily by shocks to the largest firms, which also have the highest markups and market shares.

Nested CES demand structure (Atkeson-Burstein): A two-level constant-elasticity-of-substitution aggregation where the final good aggregates N sectors with cross-sector elasticity sigma, and each sector aggregates the output of its Nk firms with within-sector elasticity epsilon > sigma. This structure generates firm-level markups that are endogenously increasing in within-sector market shares (under both Cournot and Bertrand competition) and yields closed-form expressions for sector-level markups as a function of sector HHI and aggregate markups as a function of the expenditure-weighted average of sector HHIs.

Markup elasticity with respect to market share (Gamma_ki): Under Cournot competition, the semi-elasticity of firm i’s log markup with respect to its log market share, equal to (epsilon/sigma - 1)s_ki / (epsilon/(epsilon-1) - (epsilon/sigma - 1)s_ki). This is strictly positive for epsilon > sigma and increasing in market share, implying that larger firms have markups that are more responsive to changes in their competitive position.

Pass-through rate (alpha_ki): The fraction of an idiosyncratic cost shock that is passed into the firm’s price relative to the sectoral price index, given by 1/(1 + (epsilon-1)Gamma_ki). Pass-through is decreasing in market share (larger firms have lower pass-through), which dampens their price response to own shocks and mutes the impact of large-firm shocks on aggregate price volatility—acting like a reduction in market concentration.

Within-between decomposition of sector markup changes: The change in inverse sector markup decomposed into (i) a within term measuring changes in firm-level markups holding market shares fixed, and (ii) a between term measuring reallocation of market shares across firms with heterogeneous markups. Under Cournot competition, these two terms are exactly equal (each 50%) for any firm-level shocks—a result that holds globally (not merely as a first-order approximation)—because the forces that increase the within term (higher markup sensitivity) also raise heterogeneity between firms (increasing the between term).

Sectoral markup (mu_kt): Defined as the ratio of sectoral revenues to total wage payments in the sector, equal to the harmonic mean of firm-level markups weighted by market shares. Under Cournot competition, this is a simple increasing function of the sector’s HHI: mu_kt = (epsilon/(epsilon-1))[1 - (epsilon/sigma - 1)/(epsilon-1) x HHI_kt]^(-1). This mapping between concentration and the markup price-cost wedge gives the central empirical prediction tested at the sector level.

Markup cyclicality (at different aggregation levels): The comovement between markups and output, which the paper distinguishes sharply across three levels: (i) firm markup vs. own-sector output—countercyclical for small firms, procyclical for large firms; (ii) sector markup vs. own-sector output—procyclical (positive covariance) under conditions proven in Proposition 3; (iii) sector markup vs. aggregate output—theoretically positive over long samples but ambiguous and close to zero in short samples, because aggregate output also reflects shocks to other sectors whose markups are uncorrelated with the focal sector’s markups. The paper’s central insight is that the same underlying model generates all three empirical patterns simultaneously.

Business, Liquidity, and Information Cycles

Mon, 01 Jan 0001 00:00:00 +0000

The paper studies how the two roles of stock markets — revealing information about firms’ fundamentals (which guides capital allocation) and providing liquidity — interact, arguing that when stocks are used more intensively for liquidity, their prices reveal less information about fundamentals. The authors build a Grossman-Stiglitz-style trading model with two types of rational traders (‘day’ traders who value liquidity and ’night’ traders who value fundamentals) that generates endogenous noise in prices, derive an analytical measure of price informativeness (PI), and structurally estimate PI from firm-level panel data for 16 countries over 1984-2022, finding that PI declines in periods of insufficient funding liquidity (such as the Great Recession and the COVID-19 pandemic) and that these fluctuations are explained mostly by changes in trading activity rather than information quality. Integrating the trading module into a real business cycle model with heterogeneous firms calibrated to the United States, they simulate recessions: a stand-alone recession is ‘cleansing’ — prices become more informative and allocation improves, mitigating output losses by 4.4% — whereas a recession coinciding with banking distress is ‘sullying’ — agents rely more on stocks for liquidity, prices become less informative, and worsened misallocation magnifies output losses by 22%. A counterfactual with exogenous (rather than endogenous) information implies output would fall about 43% more than in the benchmark, which the authors read as evidence that endogenous information acquisition lets stock markets ’lean against the wind’ in recessions. All magnitudes are model-based and specific to the U.S. calibration.

Summary of a forthcoming paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. What interaction between stock-market roles does the paper study?

The paper studies how the liquidity role of stock markets affects their information role: if stocks are used more intensively for liquidity, prices reveal less information about firms’ fundamentals. While the information and liquidity roles of stock markets are each well studied, their interaction is less understood; the authors ask whether using stocks for liquidity enhances or weakens their information role, how distress in other liquidity sources (such as banks) affects price informativeness, and how this contributes to the depth of recessions.

Q2. How does the trading model generate the information-liquidity tradeoff?

The authors extend Grossman and Stiglitz (1980) by replacing noise traders with two types of rational traders — ‘day’ traders interested in liquidity and ’night’ traders interested in fundamentals — so that each type’s trades act as endogenous noise for the other. In equilibrium a linear pricing function exists in which price informativeness depends on the relative weights of fundamental versus liquidity information in prices, and those weights are determined by how many day and night traders operate, their information choices, and how aggressively they trade. When funding markets malfunction, the economy relies more on stocks for liquidity, there are more day traders, and price informativeness declines.

Q3. What is Price Informativeness (PI), and how is it estimated?

Price Informativeness (PI) is defined analytically as a function of the dispersion of firm productivity, the dispersion of stock-price fluctuations, and their respective price loadings; in a high-PI market, a firm’s high relative stock price is a strong signal of positive information about its fundamentals. The authors estimate PI structurally using firm-level panel data from 16 countries spanning 1984 to 2022. The linear relationship among stock prices, earnings, and stock liquidity holds independently of general-equilibrium considerations, which is what makes the structural estimation tractable.

Q4. What are the empirical cyclical properties of PI?

PI exhibits cyclicality and, more importantly, declines in periods of insufficient funding liquidity, such as the Great Recession and the COVID-19 pandemic. Decomposing PI into its four components, the authors show its fluctuations are mostly explained by changes in trading activity rather than by changes in information quality or the amount of information acquired.

Q5. How is the trading module embedded in a general-equilibrium model and disciplined?

The trading module is integrated into a real business cycle model with heterogeneous firms in which stock prices guide capital allocation, calibrated to the United States with two possibly correlated aggregate shocks — one to aggregate productivity and one to funding liquidity — to capture recessions with and without banking distress. The calibrated model replicates the cyclical properties of the empirical PI measure without targeting them. The authors also discipline how much new information prices convey using price-investment correlations across firms and over time, concluding that new stock-price information is roughly as important as what decision makers already know.

Q6. What are the quantitative real effects in recessions?

In a stand-alone recession, increased uncertainty induces all traders to acquire more information, raising price informativeness and improving allocation, which mitigates output losses by 4.4% (‘cleansing’); when a recession coincides with funding-market distress, heightened liquidity-driven trading makes prices less informative and worsens allocation, magnifying output losses by 22% (‘sullying’). The authors interpret the 22% figure as a sizable real effect of banking problems operating through a novel channel: the weakening of the information and allocative role of stock markets.

Q7. What do the information-structure counterfactuals show?

If information were exogenous rather than endogenously acquired, liquidity distress would reduce PI by more and output would decline about 43% more than in the benchmark — implying endogenous information acquisition lets stock markets ’lean against the wind’ during recessions. The authors further find that halving the cost of information about fundamentals would make output declines about 5% smaller, whereas halving the cost of information about a stock’s liquidity would make declines about 2% larger, leading them to conclude that the welfare effect of transparency is nuanced — easier access to one type of information can make it harder to infer another.

Q8. What are the main limitations and scope conditions?

The authors flag two limitations: the framework assumes no feedback from the real economy back to financial markets (prices affect investment, but investment does not affect prices), and the counterfactuals focus on how the information environment affects price informativeness, abstracting from other channels through which information affects production. Adding two-way feedback would sacrifice the tractability of linear pricing but could introduce additional magnification forces. All quantitative magnitudes are specific to the U.S. calibration.

Key concepts

price informativeness (PI) : the extent to which stock prices reveal to an outside observer the information that informed traders hold about firms’ fundamentals; defined in the paper as an analytical function of productivity dispersion, price-fluctuation dispersion, and their price loadings, and estimated structurally.

day traders vs. night traders : the paper’s two types of rational traders — day traders trade to satisfy liquidity needs, night traders trade on information about fundamentals — whose trades act as endogenous noise for one another, replacing the exogenous noise traders of Grossman-Stiglitz.

funding liquidity vs. market liquidity : funding liquidity is liquidity provided by intermediaries through credit; market liquidity is the ability to trade stocks to meet liquidity needs; when funding liquidity is scarce, agents substitute toward market liquidity, raising liquidity-driven trading.

cleansing vs. sullying recessions : in the paper’s usage, a cleansing recession improves allocation (here via more informative prices), while a sullying recession worsens it; a recession is cleansing without banking distress and sullying when it coincides with funding-market distress.

Changing Opportunity: Sociological Mechanisms Underlying Growing Class Gaps

Mon, 01 Jan 0001 00:00:00 +0000

This paper documents sharp divergent trends in intergenerational economic mobility by race and class in the United States across the 1978 to 1992 birth cohorts, and investigates the causal mechanisms driving those changes. The core empirical facts are two: between 1978 and 1992 birth cohorts, the earnings gap between white children from high-income versus low-income families grew by approximately 28–30% (the “white class gap”), while the earnings gap between white and Black children from low-income families shrank by approximately 27–30% (the “white-Black race gap”). These twin trends — growing class gaps and shrinking race gaps — appear consistently across earnings, employment rates, educational attainment, SAT/ACT scores, incarceration, marriage, and mortality, and they hold in nearly every region of the country.

The data are drawn from de-identified federal income tax returns linked to decennial census records and the Numident database, covering 57 million children born between 1978 and 1992, with information on parental and child incomes, employment, marital status, mortality, and residential location, supplemented by ACS educational attainment and linked SAT/ACT records covering 24.8 million students. Children’s outcomes are measured primarily as household income percentile ranks at age 27.

In dollar terms, the white class gap (mean income difference between children raised at the 25th vs. 75th parental income percentile) grew from $17,720 to $20,950 in real 2023 dollars, while the white-Black race gap for low-income families fell from $20,810 to $14,910. The intergenerational rank-rank slope for white children increased from 0.23 to 0.29. The racial gap in intergenerational persistence of poverty — the probability of a child born to the bottom income quintile remaining there — shrank from 14.7 percentage points to 4.1 percentage points (a 72% reduction), driven roughly equally by improvement in Black children’s chances of escaping poverty and deterioration in low-income white children’s chances. The white class gap in early-adulthood mortality more than doubled, while the white-Black race gap in mortality fell by 77%.

The paper systematically rules out three alternative explanations. Observable family characteristics (parental education, wealth, occupation, and marital status) explain only 7% of the growing white class gap and none of the shrinking white-Black race gap. Neighborhood-level common shocks, tested by including childhood county or Census tract-by-cohort fixed effects, similarly explain only 7% of the class gap and none of the race gap. The divergent trends persist even among children raised in the same Census tract, pointing to forces that operate differentially across race and class groups within the same neighborhood.

The paper’s central finding is that changes in children’s outcomes across cohorts are strongly and positively correlated (r = 0.91 across subgroups) with changes in parental employment rates within the child’s social community, defined as families sharing the same race, class, and childhood county. Low-income white communities experienced sharp relative declines in parental employment rates; low-income Black communities experienced relative improvements. These community-level parental employment changes account for nearly all of the divergent trends.

To establish causation, the paper exploits variation in the age at which children move to counties with changing parental employment rates. Children who moved at younger ages (before age 8) to counties where parental employment was increasing experienced larger improvements in earnings than those who moved at older ages (after age 13), consistent with a causal exposure effect with greater impact for longer durations of exposure. Sibling comparisons — comparing outcomes of younger versus older siblings who moved together — confirm that the age gradient reflects causal exposure rather than family-level selection.

The social interaction mechanism is supported by two sources of variation: children’s outcomes are more strongly related to parental employment rates of their own birth cohort than adjacent cohorts (cohort specificity unlikely to be explained by resources), and outcomes are primarily driven by the employment rates of same-race, same-class community members, with cross-racial influence appearing only in counties where cross-racial interaction is greater (counties with small Black population shares or higher interracial marriage rates). The unified explanation the paper proposes is that children’s outcomes mimic those of the adults in their social communities, following Borjas (1992).

Q: What are the precise magnitudes of the growing white class gap and shrinking white-Black race gap in income percentile ranks? A: The white class gap — the difference in mean household income ranks between white children raised at the 25th versus 75th parental income percentiles — increased from 11.1 to 14.1 percentile ranks between the 1978 and 1992 birth cohorts, a 28% increase. The white-Black race gap for children from low-income families fell from 14.9 to 10.9 percentile ranks, a 27% decrease. The intergenerational rank-rank slope for white children increased from 0.23 to 0.29 (a 28% rise in persistence).

Q: How did the trends in poverty persistence versus upward mobility differ? A: The convergence in white-Black outcomes was driven almost entirely by changes in poverty persistence rather than upward mobility. The racial gap in the probability of remaining in the bottom income quintile shrank from 14.7 percentage points to 4.1 percentage points (a 72% reduction), with roughly half from Black children being less likely to remain at the bottom and half from white children being more likely to remain. By contrast, the white-Black gap in the probability of rising from the bottom quintile to the top quintile fell by only 1.9 percentage points (17%).

Q: How widespread geographically were the divergent trends? A: Outcomes declined for low-income white families in nearly every county, but the largest declines occurred in historically high-mobility areas such as the Great Plains and the coasts. For low-income Black families, outcomes improved in most areas, with the largest gains in historically low-mobility regions including the Southeast and the industrial Midwest. The correlation between county-level changes for low-income white versus low-income Black children is a positive 0.58, meaning the areas where Black families improved most tended to be areas where white families declined least, not most.

Q: Do the trends persist when using non-rank, inflation-adjusted dollar outcomes? A: Yes. The white class gap in mean household income grew from $17,720 to $20,950 in real 2023 dollars, and the white-Black race gap for low-income families narrowed from $20,810 to $14,910. The paper also reports similar patterns for individual earnings (as opposed to household income), ruling out changes in household composition as a driver.

Q: What do the pre-labor-market outcomes show? A: The divergent trends emerge before children enter the labor market. The white class gap in educational attainment grew by 20%, driven by growing gaps in four-year college completion. The white-Black race gap in educational attainment disappeared by the 1992 cohort, driven by narrowing gaps in high school graduation. The white class gap in the share of students taking the SAT/ACT increased by 12.1 percentage points between the 1980 and 1991 birth cohorts, while the white-Black race gap in SAT/ACT-taking decreased by 20.3 percentage points. The white class gap in mean SAT/ACT scores grew by 62% between the 1980 and 1997 birth cohorts among test-takers.

Q: How large is the mortality dimension of these trends? A: The white class gap in early-adulthood mortality (ages 24–27) more than doubled between the 1978 and 1992 birth cohorts, while the white-Black race gap in early-adulthood mortality decreased by 77%. These non-monetary outcomes are invariant to inflation and income measurement choices, confirming the robustness of the broader trends.

Q: How much do family-level characteristics explain? A: Controlling jointly for parental education, wealth, occupation, and marital status reduces the estimated growth in the white class gap by only 7% (from 3.37 to 3.13 percentile ranks). The same controls do not explain the shrinking white-Black race gap — the estimated reduction in the race gap actually becomes slightly larger (4.56 rather than 4.16 percentiles) after controlling for family characteristics, indicating that observable family factors work against the observed convergence.

Q: How much do neighborhood-level common shocks explain? A: Including childhood county fixed effects interacted with birth cohort explains only 7% of the growing white class gap and none of the shrinking white-Black race gap. Including Census tract fixed effects yields essentially identical results. The divergent trends persist among children growing up in the same Census tract, ruling out explanations based on differential exposure to neighborhood-level economic shocks.

Q: What is the community-level parental employment correlation, and what does it explain? A: Changes in children’s earnings, SAT/ACT scores, and educational attainment across cohorts are strongly positively correlated with changes in parental employment rates within the child’s community (same race, same class, same county), controlling for the employment status of the child’s own parents. The correlation between changes in children’s outcomes and changes in community parental employment rates across all race and class subgroups is 0.91. This single community-level factor — as proxied by parental employment rates — accounts for nearly all of the divergent trends by race and class.

Q: What is the quasi-experimental design for estimating causal effects, and what does it assume? A: The paper compares outcomes of children who moved to counties with increasing parental employment rates at younger versus older ages, across earlier versus later birth cohorts. The identification assumption is “constant selection by age”: any selection of families into moving to a given county in years when parental employment is higher may differ across cohorts, but those selection differences must not themselves vary systematically with the age at which children move. The paper treats this as a “constant selection by age” assumption standard in the neighborhood effects literature.

Q: What do the causal exposure results show? A: Children who moved before age 8 to communities where parental employment was increasing show systematically higher earnings in later birth cohorts, while children who made the same move after age 13 show little difference in earnings across cohorts. This pattern — larger effects at younger ages — is consistent with a causal exposure effect of growing up in an improving community, with effects proportional to the duration of exposure.

Q: How do sibling comparisons validate the identification assumption? A: When siblings move together to a community with increasing parental employment rates, the younger sibling — who receives more years of exposure to the higher-employment environment — earns significantly more than the older sibling. The earnings difference is proportional to the age gap between siblings. This rules out explanations based on fixed unobserved family characteristics and supports the constant-selection-by-age assumption.

Q: What evidence distinguishes social interaction mechanisms from economic resource mechanisms? A: Two sources of variation are used. First, children’s outcomes are much more strongly related to the parental employment rates of peers in their own birth cohort than peers in adjacent cohorts — a cohort-specificity that is implausible for economic resource channels (school budgets, local tax bases) which would not vary sharply across adjacent cohorts. Second, outcomes of low-income white children are driven primarily by the employment rates of low-income white parents, not by low-income Black or high-income white parents’ employment rates, and vice versa for low-income Black children — consistent with interaction patterns being stratified by race and class.

Q: What role does cross-racial interaction play? A: In counties where Black children constitute a small share of the population (making cross-racial interaction more likely), Black children’s outcomes are also related to low-income white parental employment rates. Similarly, in counties with higher interracial marriage rates (a proxy for cross-racial interaction), Black children’s outcomes are related to white parental employment rates even after controlling for racial composition. This cross-sectional variation supports the interpretation that the influence channel is social interaction rather than parallel economic shocks.

Q: How do the findings for Hispanic, Asian, and AIAN children compare? A: Changes in economic mobility for Hispanic, Asian, and AIAN children between 1978 and 1992 birth cohorts were much more modest than for white and Black children. For children from low-income families, mean household income ranks were essentially unchanged for Asian children and rose by only about 0.5 percentiles for Hispanic and AIAN children. However, the same community-level parental employment rate mechanism explains the (smaller) changes for these groups as well; the correlation between changes in children’s outcomes and changes in community parental employment rates is 0.91 across all subgroups.

Q: What is the paper’s unified theoretical account of all the divergent trends? A: The paper concludes that a parsimonious theory — that children’s outcomes mimic those of the parents in their social communities, following Borjas (1992) — explains the divergent trends by race and class. Because social interaction is stratified by race and class even within neighborhoods, changes in parental outcomes in the parent generation propagate differentially to white versus Black and high-income versus low-income children, producing growing class gaps and shrinking race gaps through the same underlying mechanism.

Q: What does the paper imply about the malleability of economic mobility disparities? A: Because the causal exposure effects of community environments on children’s outcomes can be detected within a 14-year span (1978 to 1992 birth cohorts), the paper implies that differences in economic mobility by race and class may be malleable in policy-relevant timeframes. This is despite the fact that long-standing disparities partly trace back to historical factors such as slavery, Jim Crow laws, redlining, and the Great Migration.

White class gap: The difference in mean household income ranks in adulthood for white children born to families at the 25th versus 75th percentiles of the national parental income distribution; increased from 11.1 to 14.1 percentile ranks (28%) between the 1978 and 1992 birth cohorts.

White-Black race gap: The difference in mean household income ranks in adulthood for white versus Black children born to families at the 25th percentile of the national parental income distribution; decreased from 14.9 to 10.9 percentile ranks (27%) between the 1978 and 1992 birth cohorts.

Social community: In this paper’s usage, other families who share the same race, class category, and childhood county as a given child; the unit within which community-level parental employment rates are measured and found to be predictive of children’s outcomes.

Causal exposure effect: The effect on a child’s adult outcomes of an additional year spent growing up in a community with higher parental employment rates, estimated quasi-experimentally by comparing children who moved to counties with changing parental employment rates at younger versus older ages; larger effects at younger ages imply a causal, duration-sensitive exposure channel.

Constant selection by age: The identification assumption underlying the quasi-experimental design; requires that any systematic differences in the types of families who move to a county when parental employment is high versus low do not themselves vary with the age at which children move to that county.

Intergenerational rank-rank slope: The OLS slope coefficient from regressing child income percentile rank on parental income percentile rank; for white children, increased from 0.23 in the 1978 birth cohort to 0.29 in the 1992 birth cohort, indicating greater persistence of economic status.

Cohort-specificity of community effects: The empirical pattern that children’s outcomes are more strongly related to the parental employment rates of peers in their own birth cohort than those of adjacent cohorts, used in the paper as evidence favoring social interaction over economic resource channels as the mediating mechanism.

Civil War–Induced Displacement and Human Capital

Mon, 01 Jan 0001 00:00:00 +0000

This paper examines the impact of conflict-driven forced displacement on human capital accumulation using the Mozambican civil war (1977–1992) as the empirical setting. During this war, over four million civilians — roughly a third of the population — fled to rural areas, cities, neighboring countries, or UN-managed refugee camps. The study advances on prior work in three dimensions: it uses the full post-war population census (12 million individuals) rather than a small survey; it studies multiple displacement trajectories in a single framework; and it separately identifies place-based exposure effects from a general uprootedness effect.

The primary data source is the 1997 Mozambican census, which records each individual’s place of birth, residence in 1992 (the war’s end), and residence in 1997. Key outcomes are educational attainment and sectoral employment (agricultural versus services). The authors supplement the census with digitized colonial road and school maps, georeferenced conflict events, and landmine contamination data.

The main identification strategy compares approximately 135,000 siblings (from 45,000 families) separated during the war, using the sibling who stayed behind as a within-family counterfactual. This design controls for household-level characteristics including religious and ethnic background, aspirations, and exposure to violence.

The key findings are as follows. First, rural-born IDPs displaced to cities have a 7.3 percentage point higher likelihood of attending primary school and 0.53 more years of schooling compared to their siblings who stayed behind — roughly one-third of the non-displaced mean. Rural-born IDPs displaced to other rural areas also show gains, with a 3 percentage point higher likelihood of attending school and 0.24 additional years, supporting the uprootedness hypothesis even for displacements that did not reach urban centers. Urban-born IDPs forcibly relocated to the countryside — primarily through FRELIMO’s villagization scheme — experienced 9 percentage point lower primary school attendance and approximately 0.5 fewer years of schooling relative to siblings who remained in cities.

External displacement (to camps in Malawi or Zimbabwe) generated no significant schooling gains relative to staying siblings, despite UN-built schools in camps, likely because scarce employment opportunities reduced perceived returns to education.

Second, the paper jointly estimates place-based and uprootedness effects in a single within-family framework. Place effects are statistically significant: displacement to a district one standard deviation more developed than one’s birthplace raises schooling likelihood by approximately 3 percentage points (OLS) to 5 percentage points (2SLS reduced form). Crucially, a residual uprootedness effect of approximately 2–4 percentage points persists even after controlling fully for destination-origin differences in development and conflict intensity. This uprootedness effect is quantitatively comparable to being displaced to a district one standard deviation more developed than one’s birthplace.

Third, a primary survey of 208 Nampula residents conducted in early 2020 — three decades after the war — confirms lasting educational gains. IDPs displaced to Nampula have a 10 percentage point higher likelihood of completing primary school relative to their siblings who stayed in the countryside, and their educational attainment converged to levels of urban-born, never-displaced residents despite large urban-rural education gaps. However, IDPs report significantly lower social capital, civic participation, and community trust than urban-born respondents, and score significantly worse on mental health indicators, including depression, loneliness, and pessimism. These psychosocial costs persist three decades after the war’s end.

The findings apply to a low-income, post-colonial African setting characterized by widespread illiteracy (over 60%) and subsistence agriculture (over 85% of employment) at the war’s close. The results are robust to alternative age restrictions, extended family comparisons, dropping the oldest sibling, same-sex sibling pairs, and narrowing the age gap between sibling pairs to as few as two years.

Q: What is the core identification strategy and why is it preferred over cross-sectional estimates? A: The authors compare siblings within the same household who experienced different displacement trajectories during the war. Because siblings share household-level characteristics — parental preferences for education, ethnic and religious background, wealth, and local conflict exposure — the within-family design controls for confounders that would bias cross-sectional estimates. The within-family estimates are systematically smaller than cross-sectional ones (e.g., 7.3 pps vs. 24–30 pps for rural-to-urban displacement in primary school attendance), confirming that sorting was present even in the unpredictable civil war setting.

Q: What do the results show for rural-born IDPs displaced to urban centers? A: Within the sibling-pair framework, rural-born IDPs displaced to cities and towns have a 7.3 percentage point higher likelihood of attending primary school and 0.53 more years of schooling compared to their siblings who stayed in rural birthplaces, against a non-displaced sibling mean of approximately 20% primary school access and one year of formal schooling. These IDPs also show a 4 percentage point higher likelihood of non-agricultural employment five years after the war’s end.

Q: What do the results show for rural-born IDPs displaced to other rural areas? A: Even displacement to a different rural district — not a city — generates modest but statistically significant gains: a 3 percentage point higher likelihood of attending school and 0.24 additional years of schooling relative to siblings staying in their birthplace rural district. The authors interpret this as evidence for the uprootedness hypothesis, since rural Mozambique at the time was among the most impoverished and insecure environments in the world, meaning destination quality alone cannot explain the gain.

Q: What do the results show for externally displaced refugees? A: Refugees displaced to camps and settlements in Malawi, Zimbabwe, Tanzania, Zambia, and Swaziland show schooling levels statistically similar to their siblings who remained in their rural birthplaces, despite UN-built primary schools in camps. The authors attribute the absence of gains to low perceived returns to education stemming from scarce employment opportunities at displacement destinations. Externally displaced individuals do show a 5 percentage point lower likelihood of agricultural employment relative to staying siblings.

Q: What are the consequences of urban-to-rural forced displacement? A: Urban-born individuals forcibly relocated to the countryside — primarily through FRELIMO’s villagization and food production programs — have approximately 9 percentage point lower likelihood of attending primary school and 0.5 fewer years of schooling compared to siblings who remained in urban areas. These results indicate that FRELIMO’s coercive relocation policies imposed material human capital costs on the displaced.

Q: How are place-based and uprootedness effects separated empirically? A: The authors construct principal component indices for destination-origin differences in regional development (aggregating population density, Portuguese-speaking share, offspring mortality, road density, colonial market density, and school density) and conflict intensity (conflict events per capita and landmine contamination per capita). They then include these continuous exposure measures alongside a binary displacement indicator in within-family regressions. The coefficient on the binary displacement indicator — conditional on destination-origin development and conflict differences — isolates the uprootedness effect for individuals displaced to districts with identical characteristics to their birthplace.

Q: What are the magnitudes of the place-based and uprootedness effects? A: Under OLS, displacement to a district one standard deviation more developed than one’s birthplace raises schooling likelihood by approximately 3 percentage points. The residual uprootedness effect — displacement per se, controlling for destination quality — raises schooling likelihood by approximately 2 percentage points. Under 2SLS (instrumenting destination-origin development differences with the development of districts within 100 km of birthplace), the place-based effect rises to approximately 5 percentage points in the reduced form, and the uprootedness effect remains significant at approximately 4 percentage points. Both the uprootedness and place-based effects are of comparable magnitude.

Q: What instrument is used in the 2SLS specifications and what is its first-stage strength? A: The instrument exploits the fact that Mozambique’s heavily mined and rudimentary transportation network constrained civilian movement — the median displaced sibling ended up roughly 97 kilometers from birthplace. The authors instrument actual destination-origin development and conflict differences with the predicted differences based on the characteristics of districts within 100 km of the birthplace. The first-stage elasticity between actual and proximity-predicted differences in development is 0.86, and for conflict is 0.88, both precisely estimated.

Q: What do the long-run survey results from Nampula show about educational persistence? A: In a 2020 survey of 208 Nampula residents aged over 35, IDPs who fled to Nampula during the war have a 10 percentage point higher likelihood of completing primary school relative to their siblings who stayed in the countryside. Their educational attainment converges to the level of urban-born, never-displaced Nampula residents, despite large historical and contemporary urban-rural education gaps in northern Mozambique. The majority of IDPs (73%) report that extended relatives or friends advised them to attend school upon arriving in the city, and most believed education was necessary for urban employment.

Q: What are the long-run psychosocial costs documented in the Nampula survey? A: Even three decades after the war’s end, IDPs in Nampula report significantly lower social capital, civic participation, and community trust compared to urban-born never-displaced residents. IDPs also score significantly worse on mental health indicators including depression, loneliness, and pessimism. These findings suggest that forced displacement imposes persistent psychosocial costs that are not remediated by economic or educational convergence.

Q: What drives displacement in the data, and does selection threaten identification? A: Linear probability and multinomial logit models show that conflict intensity and geographic proximity (distance to the border for external displacement; distance to cities for urban displacement) are the primary correlates of displacement type, while differences in destination development are uncorrelated with displacement. Nevertheless, the overall explanatory power of these models is low, confirming many idiosyncratic and unpredictable features of the war. The within-family design addresses residual selection on household characteristics, and the 2SLS design addresses selection on destination-specific characteristics.

Q: How do educational gains translate into sectoral employment outcomes? A: Across specifications, gains in schooling move in tandem with a shift out of agriculture into services. Rural-to-urban IDPs have a 4 percentage point higher likelihood of non-agricultural employment five years after the war, while externally displaced show a 5 percentage point lower likelihood of agricultural employment. Urban-born IDPs displaced to the countryside are more likely to work in agriculture after the war. The authors interpret this co-movement as suggesting that conflict-driven human capital accumulation may contribute to structural transformation away from subsistence agriculture.

Q: How robust are the within-family estimates? A: The authors conduct six sensitivity checks: adding family fixed effects to cross-sectional regressions, restricting to individuals aged 12–18 in 1997 to address co-habitation concerns, extending comparisons to cousins and other relatives, dropping the oldest male sibling to minimize favoritism concerns, restricting to same-sex sibling pairs, and narrowing the age gap to two years. Across all permutations, the qualitative ordering is preserved: refugees show no significant schooling gains, rural-to-urban IDPs show gains of 5–6 percentage points in primary attendance and 0.35–0.5 extra years, rural-to-rural IDPs show small positive gains, and urban-to-rural IDPs show losses.

Uprootedness hypothesis: The idea, traced in the paper to Stigler and Becker (1977) and earlier scholars, that forced displacement incentivizes human capital investment precisely because education is a mobile asset that cannot be expropriated — distinct from place-based effects of destination quality.

Place-based (exposure) effects: The impact on human capital outcomes attributable to differences between the development level and conflict intensity of the displacement destination and the individual’s birthplace, measured as destination-origin differences in a principal component index of regional development.

Separated siblings design: An identification strategy that compares siblings from the same household who experienced different displacement trajectories during the war, holding constant all household-level characteristics including parental preferences, ethnicity, religion, wealth, and local conflict exposure.

Internal displacement (IDP): Conflict-driven movement within national borders to either rural areas or urban centers, constituting approximately 60% of global forced displacement and the majority of displacement in the Mozambican civil war context.

Source text origin: A categorization of the working paper text used for summarization — distinguishing full PDF or HTML text from abstract-only text. Abstract-only text is a hard block for summary generation in the pipeline.

Structural transformation: In this paper’s usage, the shift of workers out of subsistence agriculture into services associated with human capital accumulation triggered by conflict-driven displacement, treated as a potential mechanism of post-conflict recovery.

Psychosocial costs of displacement: Long-run deficits in social capital, civic engagement, community trust, and mental health (depression, loneliness, pessimism) reported by IDPs three decades after displacement, persisting despite convergence in educational attainment and employment.

Collusion with Optimal Information Disclosure

Mon, 01 Jan 0001 00:00:00 +0000

This paper asks how a third-party intermediary (an “algorithm”) that observes market demand or costs superior to competing firms should optimally disclose that information to maximize the firms’ collusive profit in a repeated Bertrand competition setting. The motivation is the rise of algorithmic pricing intermediaries such as RealPage in apartment rentals, A2i Systems in retail gasoline, and Rainmaker in hotel rooms, as well as offline cartel facilitators like AC-Treuhand.

The model extends the canonical Rotemberg–Saloner (1986) repeated Bertrand framework with stochastic demand. The key technical assumption is that firm profit is affine in the unknown state s, so expected profit depends only on the expected state. This holds for binary states, linear demand with unknown intercept (D(p,s) = s − p), and linear demand with unknown per-unit cost. The algorithm observes s and commits to a known disclosure policy mapping s to a public signal. The solution concept is pure-strategy subgame-perfect equilibrium, and the paper solves for the disclosure policy and equilibrium that jointly maximize collusive profit.

The main result (Theorem 1) is that the unique optimal disclosure policy is upper censorship: there is a cutoff ŝ such that demand states s < ŝ are disclosed and result in the corresponding monopoly price p^m(s), while demand states s ≥ ŝ are pooled — only the event {s ≥ ŝ} is disclosed — and result in the monopoly price for the mean concealed state, p^m(s*), where s* = E[s | s ≥ ŝ]. The reduction to a static information design problem (Lemma 1) is the key technical step: optimal collusive profit equals V*, the greatest fixed point of V = max_{G ∈ MPC(F)} E_G[min{π^m(s), δV/((1−δ)(n−1))}]. The “capped monopoly profit” min{π^m(s), π^max} is convex-then-concave in s, and classical results from the static information design literature (Kolotilin 2018; Dworczak and Martini 2019) then imply upper censorship is uniquely optimal.

Two features of the optimal equilibrium are notable. First, prices are rigid (constant at p^m(s*)) whenever s ≥ ŝ — the opposite of Rotemberg–Saloner’s “price wars during booms.” The logic is that pooling high demand states with a lower average state is more profitable than cutting prices, because pooling reduces the current-period deviation gain without sacrificing as much on-path profit. Second, for demand states s ∈ (ŝ, s*), the equilibrium price p^m(s*) exceeds the monopoly price p^m(s) — supra-monopoly pricing occurs for a range of intermediate states. Monopoly pricing is attainable at each such state in isolation, but recommending the higher price p^m(s*) is necessary to make the pooling incentive-compatible at states s > s*.

Comparing to full disclosure, Proposition 1 shows that optimal disclosure leads to strictly higher prices at every demand state, and hence unambiguously lower consumer surplus. Proposition 3 shows that improving the algorithm’s accuracy (a mean-preserving spread of F) reduces expected consumer surplus whenever consumer surplus under monopoly pricing is concave in s — a natural condition. This result is more pessimistic than prior work (Sugaya–Wolitzky 2018; Miklos-Thal–Tucker 2019), which found ambiguous effects because those papers assumed full disclosure.

Comparative statics (Proposition 2): fewer firms or a higher discount factor δ increases collusive profit V* and makes prices more flexible (raises ŝ). Collusion is impossible if and only if δ < (n−1)/n, the same threshold as under full disclosure.

Extensions maintain the core results. With Markov (persistent) demand (Section 4 / Theorem 2), upper censorship remains optimal but the cutoff ŝ(s) depends on last-period demand s: under positive serial correlation, ŝ(s) is decreasing in s, so the algorithm discloses less information following high demand. With differentiated products under a symmetric linear demand system (Section 5 / Theorem 3), the optimal policy censors an intermediate interval [ŝ_L, ŝ_H] and discloses both the lowest and highest demand states, because at high states the absence of an upper bound on equilibrium profit makes disclosure with price-cutting optimal.

Q: What is the core research question and why is it policy-relevant? A: The paper asks how an informed intermediary should optimally disclose demand or cost information to competing firms to maximize their collusive profit. It is directly motivated by antitrust cases against RealPage (sued by the US DOJ in August 2024), A2i Systems/Kalibrate, and Rainmaker, all of which gather market data from competing firms and recommend prices. The theory also applies to offline facilitators like AC-Treuhand, prosecuted by the European Commission for disclosing competitively sensitive information.

Q: What is the affinity assumption and why does it matter? A: The paper assumes that firm profit π(p, s) is affine (linearly increasing) in the demand or cost state s for each price p. This implies that expected profit for any distribution over states equals profit evaluated at the expected state: E[π(p,s)] = π(p, E[s]). As a consequence, any disclosure policy is equivalent, from a profit standpoint, to choosing a distribution G of the firms’ posterior mean beliefs over s, and G must be a mean-preserving contraction of the prior F (by Blackwell 1953). The assumption is satisfied for binary states, linear demand with unknown intercept, and linear demand with unknown cost.

Q: What is the key reduction result (Lemma 1) and what does it achieve? A: Lemma 1 reduces the problem of finding an optimal repeated-game equilibrium to a static information design problem. Optimal collusive profit equals V*, the greatest fixed point of V = max_{G ∈ MPC(F)} E_G[min{π^m(s), δV/((1−δ)(n−1))}], and this is attained by a symmetric, stationary, grim-trigger equilibrium. The reduction works because, under Bertrand competition, static deviation gains are proportional to on-path payoffs, creating a one-to-one correspondence that allows the repeated-game constraint to be folded into a single-period objective.

Q: Why is upper censorship the uniquely optimal disclosure policy? A: The static information design problem has a “capped monopoly profit” objective: min{π^m(s), π^max}, where π^max = δV*/((1−δ)(n−1)) is the maximum per-period profit that satisfies incentive constraints. Because π^m(s) is convex (as the maximum of affine functions) and the cap π^max is constant, the overall objective is convex for s below the cap and constant (then concave) above it — i.e., convex-then-concave in s. Classical results for linear information design (Kolotilin 2018; Dworczak and Martini 2019) imply that the unique optimal policy for a convex-then-concave objective is upper censorship.

Q: What is the supra-monopoly pricing result and why does it arise? A: For demand states s ∈ (ŝ, s*), the equilibrium price is p^m(s*) > p^m(s), meaning firms charge above the monopoly price for the current state. This arises because the pooling policy must recommend a single price for all states s ≥ ŝ, and the recommended price is p^m(s*) where s* = E[s | s ≥ ŝ]. At intermediate states s ∈ (ŝ, s*), this price exceeds the local monopoly price. The algorithm accepts lower profit at these states because it is necessary to maintain the pooled recommendation at higher states where monopoly pricing would otherwise require a price cut.

Q: How does optimal disclosure compare to full disclosure in terms of consumer surplus? A: Proposition 1 shows that collusive prices under optimal disclosure are strictly higher at every demand state compared to full disclosure (Rotemberg–Saloner). In Rotemberg–Saloner, high demand states trigger price cuts (“price wars during booms”) to deter deviation; under optimal disclosure, high states are pooled and prices are instead rigid at p^m(s*). Because prices are higher at all states, consumer surplus is unambiguously lower under optimal disclosure.

Q: What does Proposition 3 say about the effect of algorithmic accuracy on consumer surplus? A: Proposition 3 states that if consumer surplus under monopoly pricing, CS(s), is concave in s, then a mean-preserving spread of F (i.e., improved algorithmic accuracy) reduces expected consumer surplus. This result is more pessimistic than prior work by Sugaya–Wolitzky (2018) and Miklos-Thal–Tucker (2019), which found ambiguous effects. The difference is that those papers assumed full disclosure, so better accuracy tightened incentive constraints and sometimes forced price cuts. Under optimal selective disclosure, a more accurate algorithm always raises average prices because the algorithm withholds information that would have forced price cuts.

Q: What are the comparative statics with respect to the number of firms and the discount factor? A: Proposition 2 establishes that a decrease in the number of firms n or an increase in the discount factor δ increases collusive profit V* and makes collusive prices more flexible (raises ŝ). The intuition for fewer firms making prices more flexible is that with fewer firms, incentive constraints bind for a narrower range of demand states, so less pooling is needed. Collusion is impossible if and only if δ < (n−1)/n, the same threshold as under full disclosure.

Q: How does the model generate empirically testable predictions distinct from other collusion models? A: The model predicts: (1) the equilibrium price distribution has support on an interval [p^m(s_bar), p^m(ŝ)] plus a single mass point at the higher price p^m(s*); (2) prices are pro-cyclical overall but rigidly fixed at p^m(s*) for all but the lowest demand states; (3) the gap p^m(s) − p(s) is non-monotone — zero at low states, negative (supra-monopoly) at intermediate states, and positive at high states; (4) prices are more flexible when firms are more patient or fewer. The rigid high price combined with a flexible interval of lower prices is described as a distinctive collusive marker not present in other models.

Q: How does the model relate to the empirical literature testing Green–Porter versus Rotemberg–Saloner? A: Rotemberg–Saloner predicts counter-cyclical prices (price wars during booms), while Green–Porter predicts pro-cyclical prices. Empirical tests (e.g., Porter 1983, Ellison 1994) have typically found pro-cyclical prices, favoring Green–Porter. The present model generates pro-cyclical prices through a different mechanism — perfect monitoring plus selectively disclosed demand information — showing that pro-cyclical prices are consistent with perfect monitoring when the information intermediary optimally pools high demand states. The paper suggests that distinguishing the theories requires estimating the gap between price and monopoly price over the cycle: under Green–Porter, collusion succeeds better in high demand states; under this model, collusion succeeds better in low demand states.

Q: What narrative evidence from the RealPage case corroborates the model’s predictions? A: The US DOJ complaint against RealPage states that “in down markets… [RealPage] instills pricing discipline in landlords, curbing normal fully independent competitive reactions by substituting them with interdependent decision-making,” and that RealPage advertised that its AI helps clients “avoid the race to the bottom in down markets.” This is consistent with the model’s prediction of flexible monopoly prices at low demand states and a rigid, supra-monopolistic price in normal times. The Kumatori Contractors Cooperative case (studied by Kawai, Nakabayashi, and Ortner 2024) corroborates the censorship result: that organization took drastic steps to limit bidders’ information about costs on the largest projects — exactly the states where deviation is most tempting.

Q: How do results change with persistent (Markov) demand? A: Theorem 2 shows that upper censorship remains uniquely optimal with Markov demand, but the cutoff ŝ(s) now depends on last-period demand s. Under positive serial correlation, ŝ(s) is decreasing in s: the algorithm discloses less information after high demand because firms are more optimistic and thus more tempted to deviate. Under negative serial correlation, ŝ(s) is increasing. The optimal collusive price is no longer always equal to the monopoly price for the disclosed mean demand, and the expected price conditional on last-period demand can be countercyclical (similar to Rotemberg–Saloner), even though the current-period price is always monotone in current demand.

Q: How does the optimal disclosure policy change with differentiated products? A: With a symmetric linear demand system (Section 5, Theorem 3), the optimal policy censors an intermediate interval [ŝ_L, ŝ_H] and discloses both the lowest and the highest demand states. At high demand states s > ŝ_H, the algorithm discloses the state and recommends a price below monopoly (to satisfy incentive constraints), because with differentiated goods there is no upper bound on equilibrium profit and profit is convex in s at high states, making disclosure with price-cutting optimal. Mathematically, the capped monopoly profit is piecewise-convex rather than convex-then-concave, so the optimal policy is intermediate-interval censorship rather than upper censorship. The Appendix A version extends to general demand systems and capacity constraints with the same qualitative logic.

Q: What are the main limitations and directions for future work acknowledged by the authors? A: The paper identifies three main limitations. First, if profit is not affine in s (i.e., expected profit depends on more than the mean state), the information design problem becomes non-linear and upper censorship is typically suboptimal, though it remains approximately optimal when the problem is close to linear. Second, the model assumes the algorithm’s objective is to maximize industry profit; if the intermediary is a profit-maximizing seller of software (as in Harrington 2022), the objective may instead be to maximize the profit differential between adopters and non-adopters. Third, the model assumes all firms use the algorithm; allowing partial adoption would require modeling firms’ incentives to subscribe. The paper notes that incorporating these considerations “could be an interesting direction for future research.”

Upper Censorship (disclosure policy): A disclosure policy in which demand states below a cutoff ŝ are revealed to firms (along with the corresponding monopoly price recommendation), while states above ŝ are pooled — only the event {s ≥ ŝ} is disclosed — with a single monopoly price recommendation p^m(s*) for the mean concealed state s* = E[s | s ≥ ŝ]. This is the uniquely optimal disclosure policy in the baseline model.

Capped Monopoly Profit: The per-period profit objective in the reduced static information design problem: min{π^m(s), π^max}, where π^max = δV*/((1−δ)(n−1)) is the maximum industry profit attainable in a single period without violating incentive constraints. This function is convex-then-concave in s, which drives the optimality of upper censorship.

Supra-Monopoly Pricing: Equilibrium prices that exceed the monopoly price for the realized demand state. In the model, this occurs for states s ∈ (ŝ, s*), where the algorithm’s pooled recommendation p^m(s*) is above the local monopoly price p^m(s). It arises because the pooled recommendation must be incentive-compatible at the highest concealed states.

Price Rigidity: The feature of the optimal equilibrium in which the collusive price is constant at p^m(s*) for all demand states s ≥ ŝ. The algorithm achieves this by withholding information about high demand states, preventing the “price wars during booms” predicted by Rotemberg–Saloner (1986) under full disclosure.

Algorithmic Accuracy: In the paper’s terms, the informativeness of the algorithm’s signal about s, formalized as the precision of the distribution F. Improving accuracy corresponds to a mean-preserving spread of F (Blackwell 1953). A more accurate algorithm always increases collusive profit; under the concavity condition on consumer surplus, it also reduces expected consumer surplus.

Mean-Preserving Contraction (MPC(F)): The set of distributions G of firms’ posterior mean beliefs over s that are consistent with Bayesian updating of the prior F. By Blackwell (1953), a disclosure policy is feasible if and only if it induces a distribution G ∈ MPC(F). This is the feasibility constraint in the static information design problem.

Affinity in the state: The assumption that π(p, s) is affine (linearly increasing) in s for each price p. This implies E[π(p,s)] = π(p, E[s]), so expected profit is determined entirely by the expected state, enabling the reduction of the disclosure problem to choosing a distribution of posterior means.

Complete Pass-Through in Levels

Mon, 01 Jan 0001 00:00:00 +0000

Costs of Financing U.S. Federal Debt Under a Gold Standard: 1791-1933

Mon, 01 Jan 0001 00:00:00 +0000

Overview

This paper constructs a new dataset of US federal bond prices and uses it to estimate the full term structure of yields on gold-denominated US federal debt from 1791 to 1933 — the entire gold standard era. The core research question is how the costs of financing US federal debt evolved over this period and what monetary, fiscal, and financial policy changes drove that evolution, with the ultimate aim of understanding how the US built fiscal capacity and transformed its debt from a “junk bond” into a global “safe asset.”

Data and Methodology. The authors compile monthly prices, quantities, and descriptions of all US Treasury securities from 1776 to 1960 (the Hall et al. 2018 dataset). Bonds with less than one year to maturity are excluded from the main estimation due to liquidity premia. The primary estimation uses a Dynamic Nelson-Siegel (DNS) model with stochastic volatility (Diebold and Li 2006; Hautsch and Yang 2012), estimated by Bayesian MCMC. A key methodological innovation is the addition of bond-specific idiosyncratic pricing errors (Assumption 3), which allows the authors to include bonds with heterogeneous contract features — call options, indefinite maturities, conversion features — that characterize 19th-century US debt without either dropping them from the sample or having their idiosyncrasies distort the common yield curve. The data are “big” in the time-series dimension but sparse in the maturity (cross-sectional) dimension, frequently offering fewer than five price observations per month; the DNS framework pools information across time to address this sparsity.

For the greenback period (1862–1878), the authors extend the approach by modeling the greenback yield curve as a function of the gold yield curve and a time-varying VAR model of exchange rate expectations (Assumptions 4–5). Only nine greenback-denominated bonds exist in the sample, most of them short-term; the VAR is estimated jointly using exchange rate data and the relative prices of greenback and gold bonds.

Main Findings.

Long-run decline in yields. The 10-year gold-denominated zero-coupon yield fell from approximately 8% in 1800 to approximately 2% in 1900, consistent with global secular decline trends, but the trajectory stabilized near 2% after 1900 — suggesting US debt began to play a distinctive “safe-asset” role from the turn of the 20th century.
War spikes were much larger than previously understood. The paper’s estimate of the 10-year gold yield reaches a peak of approximately 16% near the end of the Civil War. This is substantially higher than the Homer and Sylla (2004) peak of 6% at the start of the war. The discrepancy arises because Homer and Sylla used bonds trading at par — which did not exist during the Civil War — while this paper uses the full universe of bonds at monthly frequency.
Yield curve slope switched sign. The term spread (10-year minus 2-year gold yield) was typically negative before the Civil War (inverted yield curve) and turned persistently positive afterward. The authors link this switch to a change in long-run inflation predictability: inflation was relatively hard to forecast before the Civil War and easier to forecast after, consistent with a negative inflation-risk premium in the pre-war period.
Default risk premium disappeared around 1905. Comparing hypothetical gold-denominated US consols to UK consols (the 19th-century benchmark safe asset), US yields were persistently above UK yields until approximately 1905, when US yields fell below UK yields. This indicates that US federal debt acquired safe-asset characteristics well before World War I, foreshadowing the shift in global reserve asset status during and after Bretton Woods.
Nominal anchor during the Civil War. Despite a 60% depreciation of the greenback against gold during the Civil War (100 greenback dollars could be purchased for as few as 40 gold dollars in summer 1864), investors expected greenbacks to eventually return to gold parity. Estimated long-run exchange rate expectations remained anchored at one-for-one parity throughout the period. This kept greenback-denominated bond yields flat at approximately 6% — bonds traded around par — explaining the “Civil War yield puzzle” noted by Friedman and Schwartz (1963).
Short-rate disconnect. Short-maturity government bonds (less than one year) traded with a premium of approximately 0.25 to 0.5 percentage points relative to model-implied yields throughout most of the 19th century, reflecting scarcity of money-like assets. This premium effectively disappeared from the 1880s until World War I — coinciding with the National Banking Era — and then reappeared in the 1920s after the Federal Reserve created a secondary market for Certificates of Indebtedness.

In depth

Q1. Why does the paper restrict estimation to bonds with maturity greater than one year?

Short-maturity Treasury notes exhibited particularly large estimated bond-specific pricing errors in preliminary analysis, which the authors attribute to a liquidity premium: short-term government debt was used for transactions and thus commanded a money-like premium that a common discount function cannot accommodate. To keep this liquidity premium from distorting estimates of the longer end of the curve, these bonds are excluded from the main estimation. Short-maturity bonds are then studied separately as an “out-of-sample” exercise (the short-rate disconnect).

Q2. How does the Dynamic Nelson-Siegel model with stochastic volatility solve the cross-sectional sparsity problem?

The DNS model parameterizes the entire yield curve at each date using only three latent factors — level (L), slope (S), and curvature (C) — which follow a driftless random walk. The stochastic volatility component, captured in the covariance matrix Σt, governs how much information is pooled across adjacent time periods. When Σt → 0, the yield curve is assumed constant (full pooling); when Σt → ∞, estimates are date-by-date (no pooling). By allowing Σt to vary, the model pools more heavily in sparse periods and less during wars when yields change rapidly. The companion paper (Payne et al. 2023a) confirms via information criteria that stochastic volatility and correlated shocks improve fit without overfitting.

Q3. What is the bond-specific pricing error and why is it essential for historical data?

Assumption 3 adds to each bond i a Gaussian pricing error with mean zero and bond-specific standard deviation σ(i)_m (scaled by Macaulay duration to approximate yield-space errors). This allows bonds with idiosyncratic contract features — call options, conversion clauses, ambiguous payment currency — to inform the common yield curve without unduly distorting it. Bonds with larger σ(i)_m receive less weight in estimation. In modern datasets, researchers pre-select homogeneous bonds and use time-specific pricing errors; the historical sparsity prevents that approach here.

Q4. How large were Civil War yields compared to prior estimates, and why does the discrepancy arise?

The paper’s posterior median for the 10-year gold zero-coupon yield peaks at approximately 16% near the end of the Civil War. Homer and Sylla (2004) report a peak of 6% at the start of the war. The discrepancy arises because Homer and Sylla used bonds trading close to par, but during the Civil War no federal bonds traded at gold-price par (Lincoln’s re-election was uncertain in summer 1864; 100 greenback dollars could be purchased for 40 gold dollars, implying 6% coupon bonds were priced at 40% of par, implying yields in excess of 15%). This paper uses the full universe of Treasury bonds at monthly frequency and allows all bonds — regardless of trading price — to inform the yield curve.

Q5. When did US debt cease to carry a default risk premium relative to UK debt, and how is this measured?

The authors compare yields-to-maturity on gold-denominated UK consols to those on hypothetical gold-denominated US consols promising the same coupon flows. Because both countries were on a gold standard for most of the period and UK consols were the 19th-century safe asset, the spread is interpreted as a risk premium on US debt. US yields fell below UK yields persistently after approximately 1905, indicating that US debt was priced as a safe asset well before World War I. US yields were temporarily close to UK yields in the 1820s but the spread re-widened after the Jacksonian era, state defaults in the 1840s, and the Civil War. The spread closed only after Civil War disruptions resolved, the National Banking System matured, and gold-greenback parity was restored in 1879.

Q6. What is the “nominal anchor” finding during the greenback era, and what econometric method uncovers it?

During 1862–1878, the federal government issued non-convertible greenback dollars alongside gold bonds. The greenback depreciated substantially (to 40 cents per gold dollar in 1864), yet greenback-paying bonds traded near par, implying greenback yields near 6%. The authors model the greenback yield curve as a product of the gold discount function and a “multiplier” z(j)_t capturing the expected future gold-to-greenback exchange rate at each horizon j (Assumption 4). The exchange rate expectations are estimated via a time-varying VAR(2) model of the gold-to-greenback and gold-to-goods exchange rates (Assumption 5), jointly constrained by the prices of greenback bonds via an interest-rate parity condition. The resulting estimates show that throughout the greenback era — even during large wartime depreciations — investors’ long-run expectations of the exchange rate remained anchored near gold parity, consistent with anticipated eventual resumption.

Q7. How did political events affect exchange rate expectations during and after the Civil War?

The time-varying VAR captures shifts in exchange rate expectations associated with identifiable political events. Grant’s victory in 1869 (which resolved uncertainty about whether debts would be honored in gold) coincided with an increase in the price of greenbacks, a decrease in expected greenback appreciation, and a closing of the gap between greenback and gold 10-year yields. In the early 1870s, following the Panic of 1873 and uncertainty about resumption, investors came to expect that gold-greenback discrepancies would persist almost indefinitely, causing gold and greenback yields to converge. The Resumption Act of January 1875 then shifted 2-year and 10-year expectations back toward parity.

Q8. What is the short-rate disconnect and what does it reveal about the National Banking Era?

The short-rate disconnect is the difference between observed yields-to-maturity for bonds with less than one year to maturity and the yields-to-maturity implied by the model estimated on bonds with more than one year maturity. A positive disconnect means short-maturity bonds yielded less than long-maturity bonds conditional on the model — indicating a liquidity premium on short-term debt. The authors find a persistent premium of 0.25 to 0.5 percentage points through most of the 19th century, reflecting scarcity of money-like assets when state bank notes circulated at variable discounts. The premium disappeared from approximately the 1880s to World War I, coinciding with the mature National Banking Era after greenback-gold parity was restored in January 1879. The authors interpret this as evidence that the National Banking Acts (1862–1866), which allowed National Banks to issue standardized bank notes backed by long-term US government bonds, ultimately succeeded in supplying liquid assets and equalizing the pricing of short- and long-term federal debt — but only after the currency risk from the greenback period had been resolved.

Q9. How does the composite long-term yield series (Officer-Williamson / Homer-Sylla) distort historical narratives?

The composite series combines Homer and Sylla US federal yields (1798–1861), New England Municipal bond yields (1862–1899), and corporate bond yields (1900–1940). The paper shows that this composite series substantially underestimates the increase in US federal borrowing costs during Civil War deficits (peak of 6% vs. this paper’s 16%) and overstates post-Civil War borrowing costs by mixing in riskier private obligations. The authors argue that earlier findings of no strong association between 19th-century interest costs and deficits (Evans 1985, 1987) may reflect the composite series’ failure to accurately capture federal borrowing costs during large deficit episodes.

Q10. How did the yield curve slope change after the Civil War and what explains it?

The term spread (10-year minus 2-year gold yield) was typically negative before the Civil War and positive after the late 1870s. Major wars caused sharp temporary decreases (inversions). The authors connect the sign switch to a change in long-run inflation dynamics documented in a companion paper (Payne et al. 2023b): long-run inflation was hard to predict before the Civil War and easier to predict after, suggesting gold bonds provided a better inflation hedge in the pre-war period (negative inflation-risk premium), which is consistent with asset pricing theory producing a downward-sloping yield curve. After the Civil War, as inflation became more predictable, the inflation-risk premium became positive and the yield curve turned upward-sloping.

Q11. What did the National Banking Acts seek to do and was the puzzle of bank note under-issuance resolved?

The National Banking Acts (1862, 1863, 1865, 1866) authorized federally chartered banks to issue bank notes up to 90% of the par or market value of eligible US Treasury bonds deposited as collateral, subject to a 1% annual tax on notes outstanding (0.5% after 1900), compared to a 10% tax on state bank notes. The intended goals were to increase the supply of short-term liquid assets and to increase bank demand for long-term federal debt, thereby lowering long-term yields and eliminating the short-rate disconnect. A long-standing puzzle (Friedman-Schwartz, Cagan, Champ, Calomiris-Mason) held that yields on eligible Treasuries did not fall enough to equal the note tax rate, implying under-issuance. The paper’s analysis of the short-rate disconnect offers a resolution: if one focuses on the disconnect rather than the yield-tax spread, the National Banking Acts appear to have largely achieved their goals by the 1880s — but only after greenback-gold parity was restored, suggesting that currency devaluation risk had initially restrained bank note issuance, as hypothesized by Cagan (1965).

Key Concepts

Dynamic Nelson-Siegel (DNS) model with stochastic volatility: A parametric yield curve model (Diebold-Li 2006) parameterizing zero-coupon yields at each date as a function of three latent factors — level (L), slope (S), curvature (C) — following a driftless random walk. The paper extends this with time-varying shock volatilities (stochastic volatility) to allow the degree of information pooling across time periods to vary with institutional and wartime disruptions. Used here to handle cross-sectional sparsity in historical bond data.

Bond-specific pricing error: A Gaussian pricing error with bond-specific standard deviation σ(i)_m (scaled by Macaulay duration) added to each bond’s observed price. Allows bonds with heterogeneous and idiosyncratic contract features (call options, conversion clauses) to inform a common discount function without distorting it, by automatically down-weighting “peculiar” bonds through higher estimated σ(i)_m.

Short-rate disconnect (liquidity premium): The systematic difference between observed yields-to-maturity on bonds with less than one year to maturity and yields implied by a pricing kernel fitted on bonds with more than one year to maturity. Interpreted as a money-like convenience yield (liquidity premium) on short-term debt: when money-like assets are scarce, short-term bonds are overpriced (lower yields) relative to the term structure implied by longer maturities. Measured here as an out-of-sample fit residual from the DNS model.

Denomination risk: The risk that the unit of account in which bond payments are promised may change in value relative to gold. During the greenback era (1862–1878), bonds denominated in greenbacks carried denomination risk because greenbacks could depreciate against gold. The paper distinguishes denomination risk from default risk by estimating separate gold and greenback yield curves and modeling exchange rate expectations.

Nominal anchor: The phenomenon in which long-run market expectations of the gold-to-greenback exchange rate remained anchored near gold parity (one-for-one) even during large short-run depreciations during the Civil War. Inferred from the observation that greenback-denominated bonds traded near par (yield ~6%) while the spot greenback depreciated by up to 60% against gold, implying investors anticipated eventual full appreciation.

Default risk premium (US-UK yield spread): The difference between yields on hypothetical gold-denominated US consols and yields on UK consols. Since both were on a gold standard (so inflation expectations are similar), and UK consols were the 19th-century benchmark safe asset, the spread is interpreted as the compensation investors demanded for the risk that the US might default or alter payment terms. Persistently positive until approximately 1905, then became negative.

Convenience yield: An implicit yield that accrues to holders of money-like or safe assets because of their use in transactions or as collateral. In this paper, it emerges as the spread between yields on US federal bonds and other low-risk bonds in the late 19th century, reflecting increased demand for Treasuries as reserves under the National Banking System. Historically identified via the short-rate disconnect disappearing in the National Banking Era.

Digital Distractions with Peer Influence

Mon, 01 Jan 0001 00:00:00 +0000

This paper estimates the causal effects of mobile app usage on college students’ academic performance, physical health, and labor market outcomes, while separately identifying behavioral (endogenous) and contextual (exogenous) peer effects in app usage — the first study to do so within a unified empirical framework. The analysis draws on administrative data for three freshman cohorts (2018–2020) at a mid-tier Chinese university, linked to individual-level mobile phone usage records from a major telecommunications carrier covering 6,430 students over four years (excluding COVID semester). High-frequency GPS data, hourly app usage records for the 2020 cohort, and two waves of university surveys supplement the main dataset.

The identification strategy addresses three challenges: endogeneity of own app usage, endogeneity of peer group formation, and the reflection problem in peer effects. For own usage, two instrumental variables are used: (1) a shift-share instrument interacting the September 2020 launch of the blockbuster game Yuanshen with students’ pre-college app usage intensity; and (2) China’s October 2019 minors’ game restriction policy (prohibiting under-18s from playing online games 10 p.m.–8 a.m. and capping weekday gaming at 90 minutes/day) interacted with the evolving number of underage pre-college friends. For peer effects, the university’s random dormitory assignment within gender-class units provides exogenous peer variation; behavioral peer effects are further isolated using the minors’ restriction policy interacted with roommates’ pre-college underage friend networks, an instrument that affects roommates but not the focal student. Contextual peer effects are recovered by subtracting the estimated behavioral component from reduced-form estimates.

The main findings are as follows. First, app usage is contagious: a one standard deviation (s.d.) increase in roommates’ in-college total app usage raises a student’s own usage by 5.8% (IV). Behavioral peer effects dominate: contextual peer effects are small and statistically insignificant. Second, own app usage severely harms academic performance: a one s.d. increase in total app usage reduces GPA for required courses by 36.2% of a within-cohort-major s.d. (IV), and a one s.d. increase in game app usage alone reduces GPA by 56.6% of a within-cohort-major s.d. The direct disruption effect of roommates’ app usage reduces GPA by a further 20.6% of a within-cohort-major s.d.; combining the indirect channel (behavioral contagion), the total roommate effect reaches 22.7% of a within-cohort-major s.d., more than 60% of the own-usage effect. Third, the effect on physical education scores is roughly four times larger than on required-course GPA: a one s.d. increase in own app usage reduces PE scores by 2.74 points, while roommates’ app usage has no direct effect on PE. Fourth, a one s.d. increase in own in-college app usage reduces initial wages upon graduation by 2.3% (12.1% of within-cohort-major wage s.d.); a one s.d. increase in roommates’ usage reduces wages by 0.9% directly, with a total effect (including the contagion channel) of approximately 1.0% (5.3% of within-cohort-major s.d.). Controlling for cumulative GPA reduces the gaming-to-wage coefficient by roughly one-third, indicating that academic performance is an important but partial mediator.

A back-of-the-envelope policy simulation extending the minors’ gaming cap (3 hours/week) to college students — binding for 34.3% of student-month observations — projects an average wage increase of 0.9% at graduation, approximately half the wage premium from one additional year of work experience in developing countries.

Mechanism evidence from GPS data shows that Yuanshen’s launch caused students to arrive at study halls 18.2 minutes later and leave 23.4 minutes earlier per day. High-frequency sleep data show that a one s.d. increase in nighttime app usage reduces sleep duration by approximately 30 minutes and raises the probability of sleeping late by 34 percentage points. Survey evidence indicates that heavy app users recognize the addictive nature of gaming, pointing to self-control problems rather than lack of awareness.

The scope conditions are: single mid-tier Chinese university; 2018–2020 cohorts; outcomes through initial job placement only; peer group restricted to dormitory roommates; findings rely on IV exclusion restrictions conditional on student and time fixed effects.

Q: What is the core research question? A: The paper asks how individual and peer mobile app usage affect college students’ academic performance, physical health, and early labor market outcomes, and it separately identifies the behavioral (endogenous) versus contextual (exogenous) components of peer influence in app usage. This is claimed as the first study to disentangle these two types of peer effects within a unified empirical framework.

Q: What data does the paper use? A: Administrative records for 7,479 undergraduates across three freshman cohorts (2018–2020) at a medium-sized mid-tier Chinese university are linked to monthly mobile app usage records from a telecommunications provider covering 75% of the provincial population; 6,430 students are matched. The dataset also includes GPS location data at 5-minute intervals, hourly app usage for the 2020 cohort (used to infer sleep), and two waves of voluntary annual surveys with 1,798 respondents (24% response rate). Labor market outcomes — employment status, wages, post-graduate admissions — are available for the 2018 and 2019 cohorts.

Q: How does the paper address the endogeneity of own app usage? A: Two sets of instruments are used. The first interacts the September 2020 launch of Yuanshen (the most popular game in China, with over 13 million Chinese users by 2021, the majority under age 25) with students’ pre-college app usage, forming a shift-share instrument under the assumption that the game launch is orthogonal to unobserved GPA determinants conditional on student fixed effects. The second interacts China’s October 2019 minors’ game restriction policy with the evolving count of a student’s underage pre-college friends; event studies confirm no pre-trends and a sharp, transitory drop in app usage post-policy that dissipates as friends age out of the restricted group.

Q: How does the paper solve the reflection problem and separate behavioral from contextual peer effects? A: Three-step procedure: (1) random dormitory assignment within gender-class units yields reduced-form peer effect estimates using roommates’ pre-college app usage as the exogenous peer shifter; (2) behavioral peer effects are isolated via an IV using the minors’ restriction policy interacted with roommates’ (not the focal student’s) underage pre-college friend networks — an instrument that shifts roommates’ app usage but is orthogonal to the focal student’s outcomes; (3) contextual peer effects are recovered as the residual from subtracting the estimated behavioral effect from the reduced-form estimate.

Q: How large and significant are the behavioral versus contextual peer effects in app usage? A: A one s.d. increase in roommates’ in-college total app usage raises own usage by 5.8% (IV estimate, significant). For game apps alone the behavioral spillover is 10.7%, and for games plus video it is 6.5%. Contextual peer effects (identified from roommates’ pre-college characteristics) are much smaller and statistically insignificant, indicating that peer influence operates primarily through the direct imitation of peers’ actions rather than their background traits.

Q: What is the effect of own app usage on GPA? A: The IV estimate shows a one s.d. increase in total in-college app usage reduces GPA for required courses by 0.716 points, equivalent to 36.2% of a within-cohort-major GPA s.d. (significant at 1%). For game apps alone, a one s.d. increase reduces GPA by 1.119 points, or 56.6% of a within-cohort-major s.d. OLS estimates are biased toward zero, likely because negative health shocks reduce both GPA and app usage simultaneously.

Q: How large is the total peer effect of roommates’ app usage on a student’s GPA? A: Roommates’ app usage directly lowers GPA by 0.408 points (20.6% of within-cohort-major s.d.) through disruption of the dormitory study environment or crowding out of group study. The behavioral contagion channel (5.8% increase in own usage per s.d. of roommates’ usage) adds an additional 0.042 points, bringing the total effect to approximately 0.450 points, or 22.7% of a within-cohort-major s.d. — over 60% of the own-usage effect.

Q: What is the effect on physical education (PE) scores, and why do roommates’ app usage not matter there? A: A one s.d. increase in own total app usage reduces PE scores by 2.74 points (IV), approximately four times the magnitude of the effect on required-course GPA, consistent with health literature on excessive screen time. Roommates’ app usage has no statistically significant direct effect on PE, which the authors attribute to the irrelevance of dormitory noise and study disruptions for outdoor physical activity.

Q: What are the effects of app usage on wages at graduation? A: Doubling total app usage during college reduces initial wages by approximately 2% (IV). A one s.d. increase in own usage reduces wages by 2.3%, or 12.1% of a within-cohort-major wage s.d. A one s.d. increase in roommates’ usage directly reduces wages by 0.9% (4.8% of within-cohort-major s.d.); including the behavioral contagion channel, the total roommate effect is approximately 1.0% (5.3% of within-cohort-major s.d.). Controlling for cumulative GPA reduces the game-usage-to-wage coefficient by about one-third, implying GPA is a partial but not complete mediator.

Q: What does the policy simulation of the gaming cap say? A: Extending the minors’ game restriction (3 hours/week cap) to college students would bind for 34.3% of student-month observations, reducing average monthly gaming from 12.1 hours to 8 hours (a one-third decrease). Incorporating the behavioral peer multiplier for gaming (0.078), average gaming further converges to approximately 7.65 hours in steady state. The implied wage gain at graduation is 0.9%, approximately half the wage premium from one additional year of work experience in developing countries (Lagakos et al., 2019 estimate).

Q: What does the GPS evidence show about time allocation? A: Following Yuanshen’s launch, the average student arrives at the study hall 18.2 minutes later and returns to the dormitory 23.4 minutes earlier per day. The minors’ restriction reverses this: students with the average number of minor friends arrive at study halls 17.4 minutes earlier and return to the dorm 19.8 minutes later. Both game shocks also shift tardiness and absence rates for major-required courses in the expected directions, and the effects intensify over time with Yuanshen’s growing popularity.

Q: What do the sleep data show? A: A one s.d. increase in nighttime app usage (9 p.m.–3 a.m.) is associated with roughly 30 minutes less sleep (7% of the mean), a 34 percentage point higher probability of sleeping late, and a 4.5 percentage point higher probability of waking up late. Daytime app usage (8 a.m.–9 p.m.) is also associated with 7.2 fewer minutes of sleep (1.8% of mean) and a 3.7 percentage point higher probability of late wake-up. These results are descriptive (from the 2020 cohort hourly data) rather than IV-based.

Q: What does the survey evidence show about mechanisms and self-awareness? A: Heavier app users report worse physical health and higher stress, are less likely to have obtained professional certifications by graduation, submit fewer job applications, and express lower satisfaction with job offers. Notably, heavier users are more likely to acknowledge the addictive nature of apps and games, suggesting a self-control problem rather than informational deficiency. They also report better relationships with roommates and greater likelihood of following roommates’ advice on post-graduation choices, a potential direct channel for peer labor market effects.

Q: How representative is the sample, and what are the key scope conditions? A: The university is a mid-tier institution in southern China with students predominantly from the 30th–80th CEE score percentile among provincial college-admitted applicants; it is less female (42% vs. 53% nationally) and more rural (40% vs. 27% nationally). Survey respondents oversample less advantaged backgrounds and are re-weighted. Findings pertain to dormitory roommates as the peer group; all labor market outcomes are initial wages upon graduation; the sample covers 2018–2021 with COVID semester excluded. The peer effects estimates rest on random dormitory assignment, which the authors verify by showing no within-dorm correlation in pre-college characteristics.

Behavioral (endogenous) peer effects: The mechanism by which a peer’s actual behavior — here, contemporaneous app usage — directly influences a focal individual’s own behavior. In this paper, identified via IV using the minors’ game restriction policy interacted with roommates’ underage pre-college friend networks, which shifts roommates’ usage but not the focal student’s characteristics.

Contextual (exogenous) peer effects: The influence of peers’ pre-determined background characteristics (e.g., pre-college app usage, reflecting motivation, study habits, attitudes toward academics) on a focal individual’s outcomes, independent of peers’ actual in-college behavior. Recovered as the residual after subtracting estimated behavioral peer effects from reduced-form estimates; found to be small and insignificant in this setting.

Shift-share instrument (Yuanshen): A quasi-experimental instrument constructed by interacting the mid-sample launch date of the blockbuster game Yuanshen (September 2020) with students’ pre-college app usage intensity, under the assumption that pre-college usage predicts differential susceptibility to the shock while the launch itself is orthogonal to the university’s academic environment.

Minors’ game restriction policy: China’s October 2019 policy prohibiting individuals under 18 from playing online games between 10 p.m. and 8 a.m. and capping weekday gaming at 90 minutes per day (tightened to 3 hours/week in September 2021). Used both as an instrument for own app usage (via underage pre-college friends) and as an instrument for roommates’ usage (via roommates’ underage friends) to isolate behavioral peer effects.

Reflection problem: The identification challenge first articulated by Manski (1993) arising because an individual’s behavior both affects and is affected by peers simultaneously, making it impossible to separately identify the direction of influence from observational data without exogenous variation in peer behavior.

Source text origin: The paper’s own data provenance category distinguishing whether summaries are based on full working paper text (pdf or oa-html) versus abstract only — a distinction the paper itself does not use but that is relevant to the review pipeline running this analysis.

Within-cohort-major GPA standard deviation: The unit used to scale all GPA effect sizes, defined as the standard deviation of GPA within students of the same graduation cohort and declared major. This normalization accounts for systematic differences in grading across fields and years, making effect magnitudes comparable across specifications.

Disaggregated Economic Accounts

Mon, 01 Jan 0001 00:00:00 +0000

This paper develops and implements a system of disaggregated economic accounts that breaks down national accounting positions into bilateral flows between small groups of consumers, producers, the government, and the rest of the world. Standard national accounts document aggregate income and production plus input-output trade between producer industries; they contain no comprehensive data on which consumers buy from which producers or which producers pay income to which consumers. The paper fills this gap by measuring, for Denmark, all 36 positions in the UN System of National Accounts (SNA) — consumer spending, labor compensation, profit income, intermediates trade, government transfers and taxes, and foreign trade — as bilateral cell-to-cell flows, satisfying all national accounting identities at the level of individual cells and at the aggregate level. The data reveal systematic stylized facts about domestic spending shares, gravity of spending, urban bias, and assortative matching between consumer and producer characteristics. Combining the disaggregated accounts with a general equilibrium model with nominal wage rigidities, the paper shows that fiscal transfer multipliers vary substantially across consumer cells — from below 1 to above 2 — depending on the spending intensity of recipient cells on the slack (unemployed) portion of the economy. Applying the framework to a hypothetical U.S. tariff shock on Denmark (calibrated to July 2025 effective tariff levels on China), the paper demonstrates that the cells generating the highest multipliers are not those directly exposed to the shock or even those made slack, but those whose spending intensity on slack cells is high. The disaggregated accounts allow the government to select more effective fiscal policies: choosing transfers targeting high-spending-intensity cells saves approximately 0.4–0.7% of Danish GDP relative to programs targeting low-intensity cells, for the same GDP stimulus.

Measurement framework (Section II): The paper assigns every Danish adult to one of approximately 2,744 consumer cells, defined by the interaction of 98 municipalities (regions) and 28 industries (industry of main employment). Every production establishment is assigned to one of approximately 2,646 producer cells by region and industry. Median consumer cell contains 658 adults; median producer cell contains 47 establishments. The circular flow includes: (i) consumer spending on domestic and foreign producers; (ii) labor compensation paid by producer cells to consumer cells; (iii) profit income (dividends, mixed income, owner-occupied housing surplus) from producers to consumers; (iv) intermediates trade between domestic producers; (v) foreign trade; (vi) government taxes, transfers, and spending. A “bottom-up” approach uses microdata — geocoded transaction records from Danske Bank (largest Danish bank) and administrative government registers — to directly measure bilateral flows; a “top-down” approach distributes aggregate flows using assignment algorithms. Year: 2018. Data available at disaggregatedaccounts.com.

Stylized facts (Section IV):

1. Domestic spending shares (§IV.B): The share of a consumer cell’s spending going to domestic rather than foreign producers ranges from 75% to almost 100% (average 92%). Rural (small-population) cells, older cells, and less college-educated cells have higher domestic spending shares. Population size, average age, and college share jointly explain about half of the cross-cell variation in domestic shares; the patterns hold within industry and within region. The majority of foreign spending goes to travel-related and specialized retail categories (hotels, airlines, food away from home, clothing).

2. Gravity (§IV.C): Consumer spending declines with distance (log-log gradient = −1.33, column 1 of Table II). On average, roughly 50% of spending stays in the home region and an additional 10% goes to regions within 25 km. The distance gradient is steeper for groceries and fuel (local, in-person purchases) and shallower for telecommunications, insurance, and hotels. Rural, older, and less college-educated consumers spend more locally (stronger distance gradient, consistent with higher domestic shares).

3. Urban bias (§IV.E): Consumer spending flows disproportionately toward large cities. The 15 largest regions receive 34% of national consumer spending while accounting for only 27% of consumers. Urban bias is absent for everyday purchases (groceries) and strong for irregular or remote purchases (telecommunications, specialized retail). Rural consumers also visit urban regions in person, so urban bias is present in card payments too.

4. Assortative spending (§IV.D): Consumers tend to spend on producer cells employing workers with similar characteristics. Age of consumers and average age of workers in receiving cells are positively correlated (β = 0.178); college share similarly (β = 0.120); domestic spending share similarly (β = 0.203). The slopes are well below 1 (consumers purchase from many cells), but mild assortative spending reinforces first-order domestic spending patterns through higher-order connections.

5. Triangular flows (§IV.F): A distinctive cross-regional pattern: consumer spending and intermediates trade flow on net from rural to urban regions (urban regions run a net internal trade surplus); rural regions run a net external surplus (rural manufacturers export; e.g., Novo Nordisk in Kalundborg, Vestas in Nakskov); urban regions import relatively more from abroad. This triangular flow arises from urban consumption amenities and urban business service concentration.

Spending intensity (§IV.G): The paper constructs a reduced-form measure capturing, for each consumer cell i, how much its spending contributes to the income of a target group of cells — accounting for all higher-order connections (the infinite sum over indirect spending chains). The domestic spending intensity of cell i is defined recursively as the sum over all domestic producer cells j of (spending share αji × domestic spending intensity of producer cell j). Values range from roughly 0.4 to 0.9. The measure is strictly greater than the direct domestic spending share because the recursive formula incorporates second- and higher-order domestic connections. Domestic spending intensity is higher for rural, older, and less college-educated cells (consistent with the stylized facts). A spending intensity on slack cells can be constructed in the same way by replacing the target group with cells experiencing demand-driven unemployment.

General equilibrium model (Sections V–VI): The model is a static small open economy with many consumer and producer cells. Consumer utility is Cobb-Douglas over goods from all producer cells and foreign goods. Each producer cell’s production function is Cobb-Douglas with decreasing returns to scale (equivalent to a fixed factor). The key friction is downward nominal wage rigidity: Wi ≥ (1−δ)W̄i. When demand for a cell’s labor falls sufficiently (more than fraction δ), the wage rigidity binds and some workers in that cell become slack (unemployed demand-determined). A fiscal transfer to consumer cell i raises its income, which stimulates spending, which flows through the disaggregated network to raise labor demand across cells. The multiplier is higher when recipient spending flows disproportionately to slack cells, generating additional employment. The model is calibrated using the measured disaggregated accounts: spending shares αji, profit shares κij, labor shares λij, intermediates shares ωjj′, and tax rates are all taken directly from the disaggregated data. Baseline elasticity of substitution = 1 (Cobb-Douglas); robustness checks use short-run elasticities (< 1) and long-run elasticities (> 1), with no material change in conclusions.

Analytical result (Proposition 1): In an economy-wide recession (all cells slack), the vector of transfer multipliers is µ = ϕ′ · (I − M)⁻¹ · M · D((1 − τ̄ᵢ)⁻¹), where M is a transformed Leontief-style spending matrix incorporating the disaggregated accounts and τ̄ᵢ are fiscal externalities. The key insight is that the multiplier of cell i’s transfer is closely linked to its spending intensity on all other domestic cells, with all higher-order connections captured by the (I − M)⁻¹ M term. A cell’s multiplier is high when: (i) it spends domestically rather than on imports; (ii) it spends on producers that in turn employ domestic workers in slack cells; and (iii) these higher-order effects amplify through the circular flow.

Economy-wide recession: quantitative multipliers (Table III):

Transfer policy	Multiplier	Cost to raise GDP by 5% (bn DKK)
Uniform (all adults)	1.04	96.08
Top 10% domestic spending intensity	1.21	81.99
2018 child tax credit	1.02	97.85
2022 inflation relief to elderly	1.13	88.11
2023 housing rent inflation support	1.03	96.45
Construction worker support	1.23	81.16
Consulting/IT worker support	0.95	105.22

High-multiplier policies (construction workers, 2022 elderly relief) target rural, older, less college-educated cells with high domestic spending intensity. Low-multiplier policies (consulting/IT workers, 2023 housing relief, 2018 child tax credit) target urban, young, or college-educated cells with lower domestic intensity. The gap between the best and worst policies amounts to savings of roughly 15 bn DKK (≈ 2.4 bn USD), or 0.4–0.7% of Danish GDP, for the same aggregate GDP impact.

U.S. tariff shock application (Section VII): The paper analyzes a hypothetical U.S. tariff increase to 41.4% (the July 2025 effective U.S. tariff on China) on Danish exports, motivated by Greenland tensions. The shock reduces export revenue by 41.4% for each producer cell, with direct exposure varying by region: Billund (Lego headquarters), Kalundborg (pharmaceuticals), and a Copenhagen manufacturing hinterland face the largest direct declines — up to 8% of total regional sales. The shock propagates through the disaggregated network; cells whose income falls by more than 4% become slack. Key findings:

Regional slackness follows direct exposure but is also shaped by proximity to other exposed regions (urban bias propagates the shock to cities) and isolation (Billund has high direct exposure but low slackness relative to exposure because it is geographically isolated from other high-exposure cells)
Transfer multipliers for this heterogeneous recession (Proposition 2) depend on spending intensity on slack cells, not on direct exposure or own slackness
Table IV (R² for multiplier): slack cell indicator alone explains R² = 0.015; direct spending share on slack raises R² to 0.366; spending intensity on slack cells raises R² to 0.769 (column 3); adding both spending share and spending intensity on slack reaches R² = 0.840 (column 4)
Billund, despite high exposure, has low multiplier because its spending (often local to a low-exposure vicinity) does not create labor demand for slack cells elsewhere
Some of the highest-multiplier regions are themselves non-slack but are surrounded by many slack cells, so their spending effectively employs slack workers

Dynamic model (Section VIII): The paper extends to a dynamic OLG (Blanchard-Yaari) model with heterogeneous marginal propensities to consume (MPCs) calibrated from a 2009 Danish fiscal policy. Key result: static and year-4 dynamic multipliers are closely correlated (slope ≈ 0.898). Long-run cumulative multipliers exactly equal static multipliers (formally proved in Appendix V.F): in the long run, all transfers are fully spent. MPCs and domestic spending intensity are complementary determinants of dynamic multipliers — targeting high-MPC cells amplifies short-horizon (year 0–2) multipliers, while targeting high-spending-intensity cells shapes both short- and long-run multipliers. The paper’s main mechanism (spending intensity on slack cells) is robust at all horizons.

Robustness (Section IX): (i) Counterfactual accounts with reversed stylized patterns (e.g., rural cells spending like urban cells) lead to substantially different multipliers — the specific measured patterns drive the results. (ii) Imposing standard simplifying assumptions (consumer spending flows only to local producers; spending flows across regions in proportion to intermediate trade) misses most of the multiplier variation. (iii) The mechanism is similarly important in less open economies. (iv) Low short-run and high long-run substitution elasticities (from the trade literature) produce similar multiplier rankings across cells.

Scope conditions: The implementation is a proof of concept for Denmark, using existing micro data from a single large bank and government registers; full coverage of all banks and complete data on within-firm flows would strengthen measurement. Capital-related transactions (saving, investment, financial assets) are aggregated into a single capital accumulation cell — disaggregating these would require different data. The model is intentionally static (with a dynamic extension), abstracting from price adjustment dynamics beyond the NK wage rigidity. The analysis is a partial equilibrium in the sense that monetary policy response is not modeled; the fixed exchange rate assumption is realistic for Denmark (pegged to the Euro) but may not transfer to economies with flexible rates. The proof of concept suggests that national statistical agencies could benefit substantially from measuring disaggregated flows through refined surveys.

Summary of a forthcoming paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. What is missing from standard national accounts that this paper’s system provides?

Standard national accounts measure aggregate consumer spending, income, and output, plus intermediates trade among producer industries (input-output tables); what they do not measure is which specific consumer groups buy from which specific producer groups, or which specific producer groups pay labor and profit income to which specific consumer groups. This means that propagation of a shock through the circular flow — e.g., a tariff shock that reduces exports by rural manufacturers, which reduces income for rural workers, who then reduce spending on urban services, which reduces urban workers’ income — cannot be traced without simplifying assumptions (like “spending flows only to local producers”) that the disaggregated data shows to be empirically inaccurate. The paper provides a proof of concept demonstrating that measuring these bilateral consumer-to-producer and producer-to-consumer flows, while satisfying all national accounting identities, is feasible with existing micro data and yields policy-relevant variation in fiscal multipliers.

Q2. Why do rural, older, and less college-educated consumer cells have higher fiscal multipliers during an economy-wide recession?

These groups have higher domestic spending intensity — a higher fraction of their spending reaches domestic consumers rather than leaking abroad — because they spend less on international tourism, less on imported goods accessed through online retail or urban services, and more on local goods purchased in person. The gravity patterns (stronger distance gradient) and direct domestic spending shares document this directly: rural consumers allocate ~92–100% of spending to domestic producers versus ~75–80% for urban young college-educated consumers. When all cells are slack, a transfer to a high-domestic-intensity cell circulates more within the country, generating more rounds of domestic income and employment before leaking to imports. The mild assortative spending pattern further reinforces the first-order effect: spending by rural older consumers flows toward producer cells employing workers with similar characteristics, who also spend domestically, so higher-order connections amplify rather than dilute the domestic spending effect.

Q3. Why does targeting directly exposed or slack cells not guarantee a high transfer multiplier after the U.S. tariff shock?

A transfer raises GDP by increasing spending, which creates labor demand for other consumer cells; a transfer to a slack cell only generates a high multiplier if that cell’s spending flows toward other slack cells (directly or through indirect chains) — not if it flows toward non-slack cells or abroad. The tariff shock creates isolated pockets of slackness in rural manufacturing regions (e.g., Billund for Lego) that are geographically far from other slack regions; Billund consumers spend locally (gravity) and their locality is not itself a center of other slack cells. In contrast, regions near Copenhagen with moderate direct exposure may have high multipliers if they are close to many other slack manufacturing cells — their spending generates employment across the slack network. The R² decomposition confirms this: knowing a cell is slack explains only 1.5% of multiplier variation (R² = 0.015), while knowing its spending intensity on slack cells explains 76.9% (R² = 0.769).

Q4. How does the paper ensure that the disaggregated flows satisfy national accounting identities?

The system is designed so that every cell’s total inflows equal total outflows (a cell-level balance sheet constraint), and the sum of all cell-level flows equals the corresponding national aggregate from the SNA — both conditions are imposed by construction, not just approximated. For most positions, a bottom-up approach uses observed bilateral microdata (e.g., card payments from Danske Bank directly measure consumer spending by consumer cell i at producer cell j); for positions without direct microdata, a top-down algorithm distributes an aggregate total across cells using assignment rules grounded in the microdata. This dual approach ensures national comprehensiveness (the sum of disaggregated flows equals aggregate national accounts) and individual consistency (cell-level identities hold), unlike existing regional accounts or social accounting matrices that satisfy only one of these constraints.

Q5. What is the relationship between spending intensity and the standard fiscal multiplier formula?

The cell-level multiplier (dGDP/dTi) in Proposition 1 equals approximately the cell’s spending intensity on domestic cells, corrected for fiscal externalities and price effects of the fixed factor. The formal difference is that the model multiplier involves the matrix (I − M)⁻¹M where M incorporates both spending and production shares (through which price changes for the fixed factor enter), while the reduced-form spending intensity uses only the spending matrix. Despite this difference, the two measures are highly correlated empirically: the regression of cell-level multipliers on domestic spending intensity has a slope of approximately 1.66 for static multipliers. The spending intensity can thus be calculated directly from the disaggregated accounts without solving the full general equilibrium model, making it a practical statistic for policy guidance.

Q6. How does the dynamic model reconcile the fact that rural, older, and less college-educated cells have high spending intensities but typically lower MPCs?

MPCs and spending intensities are complementary but distinct determinants of dynamic multipliers at short horizons: high-MPC cells spend the transfer quickly (year 0–1), generating a large immediate impact, while high-spending-intensity cells ensure that spending, whenever it occurs, circulates domestically and reaches slack labor markets. At long horizons (year 4+) the two effects converge because all cells eventually spend their full transfer (long-run MPC = 1) and the multiplier converges to the static model’s value, which depends only on spending intensity. The practical implication is that policies targeting rural/older/less-educated cells (high intensity, lower MPC) may have lower immediate multipliers than policies targeting high-MPC urban consumers, but converge to higher long-run multipliers. The year-4 cumulative multipliers from the dynamic model closely resemble the static model, suggesting a 3–5 year business cycle horizon is well captured by the static analysis.

Q7. What does the triangular flow pattern imply for understanding regional inequality and fiscal redistribution?

The triangular flow — rural regions receive net income from foreign exports; rural consumers spend net inflows toward urban regions; urban consumers spend net toward abroad — means that rural regions’ incomes depend on export competitiveness while urban regions’ incomes depend on domestic consumption demand; fiscal transfers to rural consumers thus have high domestic multipliers because their spending boosts urban income (via the rural-to-urban spending flow), which then circulates domestically before leaking abroad. This pattern is also consistent with the political economy finding that high-multiplier cells (rural, older, less educated) are more likely to vote for right-wing populists and feel politically disenfranchised — they are the “left behind” groups that economic research associates with exposure to globalization and automation, but whose spending patterns happen to generate large domestic multipliers during recessions.

Key concepts

disaggregated economic accounts : a system that breaks down all national accounting positions — consumer spending, labor and profit income, intermediates trade, government transactions, foreign trade — into bilateral flows between consistently defined region-by-industry consumer cells and producer cells, satisfying national accounting identities both at the cell level and in aggregate; the paper’s proof of concept is implemented for Denmark using 2,744 consumer cells and 2,646 producer cells in 2018.

spending intensity : a cell-level, reduced-form statistic capturing how much a consumer cell’s spending contributes to the income of a target group of cells (e.g., all domestic cells or all slack cells), accounting for all indirect higher-order connections through the circular flow; formally defined as a recursive sum that incorporates the full disaggregated network structure; ranges from 0.4 to 0.9 for domestic spending intensity and is systematically higher for rural, older, and less college-educated cells.

slack cell : in the paper’s NK model, a consumer cell for which demand-driven unemployment occurs because the nominal wage rigidity binds — labor supply exceeds demand when the cell’s income declines by more than a threshold δ due to a negative demand shock; fiscal transfers with high multipliers are those whose spending reaches slack cells (directly or through higher-order network connections).

triangular flows : the cross-regional spending pattern documented for Denmark in which net consumption spending flows from rural regions to urban regions (urban bias), net foreign export revenue flows to rural regions (rural manufacturing), and net foreign import spending flows from urban regions; implies that rural-to-urban spending flows act as an important transmission channel for fiscal stimulus targeted at rural consumers.

bottom-up vs top-down disaggregation : the two methodological approaches for constructing bilateral cell-to-cell flows; the bottom-up approach uses individual-level microdata (e.g., bank transaction records) to directly observe cell-to-cell payment flows; the top-down approach allocates an aggregate national accounting position across cells using assignment algorithms informed by microdata; both approaches are designed so that the resulting disaggregated flows sum to the corresponding SNA aggregate.

Distributional Growth Accounting: Education and the Reduction of Global Poverty, 1980–2019

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Core Argument

This paper constructs the first estimates of the aggregate and distributional effects of worldwide educational expansion since 1980 by developing a “distributional growth accounting” framework that isolates the contribution of schooling to economic growth by income group. The framework integrates the canonical labor supply-and-demand model of education and the wage structure (à la Goldin and Katz 2007) with standard growth accounting tools, applied to a new microdatabase covering household surveys in 150 countries and representative of approximately 95% of the world’s population, alongside new country-specific estimates of private returns to primary, secondary, and tertiary schooling. Under conservative assumptions — relying on standard Mincerian returns, assuming capital income is unaffected by schooling, and abstracting from human capital externalities — education can account for approximately 50% of global economic growth, 70% of income gains among the world’s poorest 20% of individuals, and 40% of extreme poverty reduction since 1980; it also explains over 50% of improvements in the share of labor income accruing to women. A key mechanism is imperfect substitutability between skill groups: as educational expansion raises the supply of skilled workers, their relative wage falls, redistributing income toward low-skilled workers and amplifying education’s equalizing effect at the bottom of the distribution — a channel that canonical cross-country growth accounting misses, causing it to underestimate education’s contribution to poverty reduction by a factor of approximately three. Combining these indirect investment benefits from education with direct government redistribution (from a companion paper) brings the total contribution of public policies to extreme poverty reduction to at least 50%.

In depth

Q1. Q: What is distributional growth accounting and how does it differ from standard growth accounting?

A: Standard growth accounting (as in Barro and Lee 2015) combines cross-country data on average years of schooling with a uniform return to derive a counterfactual average income absent educational progress. Distributional growth accounting instead starts from microdata on the joint distribution of income and education within 150 countries, constructs income-group-specific counterfactuals, and accounts for both direct wage effects on individuals whose education changed and general equilibrium supply effects that alter relative wages across all workers. The standard approach is found to underestimate education’s contribution to the poorest 20%’s income growth by a factor of roughly three (23% vs. 71% in the benchmark specification), because cross-country averages cannot accurately locate the world’s poorest individuals and because two key channels — labor income shares being greater at the bottom, and supply-side wage redistribution — are omitted.

Q2. Q: How is the counterfactual world income distribution constructed?

A: In five steps applied to the 150-country microdata. First, education levels are downgraded within each survey until matching the 1980 distribution of educational attainment (using the Barro–Lee database), prioritizing individuals closest to the target level. Second, the earnings of downgraded workers are reduced using the “true” return to schooling, which lies between the initial return (prevailing before expansion, computed from the CES production function using the 2019 elasticity) and the final return observed in 2019 — for plausible parameterizations, the true return weights initial returns at 50–70%. Third, relative wages are adjusted to reflect supply effects: the increase in skilled-worker supply lowers their relative wage by 1/σ log points per log-point increase in relative supply. Fourth, counterfactual labor income is combined with unchanged capital income to yield counterfactual total income. Fifth, the share of actual income growth attributable to education is computed as the gap between the actual and counterfactual growth rates, expressed as a fraction of actual growth.

Q3. Q: What role does imperfect skill substitution play, and how is σ calibrated?

A: Imperfect substitution between skill groups (elasticity σ in a CES production function) is the mechanism through which educational expansion redistributes income. When skilled-worker supply rises, their relative wage falls and low-skilled workers’ relative wage rises, so the income gains from education are shared more broadly than individual returns alone would suggest. With perfect substitutes (σ → ∞), supply effects vanish and education’s distributional impact is determined entirely by who directly received schooling. The elasticity is calibrated from the recent macroeconomics literature; in sensitivity analysis, the paper bounds the contribution of education to the poorest 20%’s income growth between 60% and 90% across plausible values of σ and private returns.

Q4. Q: Why are the estimates described as conservative?

A: Three reasons, each biasing the estimates downward. First, standard Mincerian returns are used, which are systematically lower than causal estimates from natural experiments — a meta-analysis of 15 papers and the paper’s own quasi-experimental validation (India, Indonesia, United States) confirm this; if anything, the framework underestimates schooling’s benefits in those settings. Second, capital income is assumed unaffected by schooling, abstracting from potential effects on capital accumulation and returns. Third, human capital externalities — for which there is now substantial empirical evidence — are ignored entirely. These conservative choices are deliberate; relaxing them would increase all headline estimates.

Q5. Q: How does skill-biased technical change interact with the education contribution?

A: In the CES model, the return to schooling is increasing in the skill bias of technology (AH/AL): a higher skill bias raises the marginal product of skilled workers relative to unskilled, making schooling more profitable. The benchmark counterfactual holds technology fixed at its 2019 value and reduces education to its 1980 level. An alternative counterfactual would hold technology at its 1980 value and increase education to its 2019 level; the difference between these two exercises identifies the contribution of skill-biased technical change in amplifying the benefits of schooling. Because 1980 microdata on the world income distribution are unavailable, this decomposition can only be performed for the subsample of 33 countries with surveys around 2000; for that sample, skill-biased technical change accounts for 20–30% of the income benefits of schooling, meaning education would still have yielded large gains even absent technological progress.

Q6. Q: What do the quasi-experimental validations in India, Indonesia, and the United States show?

A: Three large-scale schooling policy interventions — a school construction program in India (studied in Khanna 2023), Indonesia’s INPRES program (Duflo 2001 and 2004), and US compulsory schooling laws (Acemoglu and Angrist 2000) — are used to externally validate the framework. Using regional variation in exposure to each program and rich microdata on the income distribution, the paper documents two findings: (1) educational expansion had large causal effects on aggregate regional incomes comparable in magnitude to individual returns estimated in the same contexts; and (2) all three policies disproportionately benefited low-income earners, substantially reducing inequality. The distributional growth accounting framework reproduces both findings with “a remarkable degree of accuracy,” and if anything underestimates the benefits of schooling, providing validation of the methodological foundation.

Q7. Q: How does the paper quantify education’s role in gender inequality reduction?

A: The framework is extended to gender by constructing a counterfactual for how large gender labor income gaps would be absent educational improvement since the early 1990s (the period for which female labor income share data are available). The counterfactual accounts for three gender-specific channels: differential educational expansion between men and women, heterogeneous returns to schooling by gender, and differential effects of schooling on female labor force participation. Comparing the counterfactual to actual trends in female labor income shares, education can explain 50–80% of the observed reductions in gender inequality, depending on specification and world region.

Q8. Q: How do public policies as a whole contribute to extreme poverty reduction?

A: The paper’s estimate of education’s indirect investment benefits (40% of extreme poverty reduction) is combined with a companion paper’s (Gethin 2023) estimates of direct government redistribution — cash and in-kind transfers together accounting for approximately 30% of global poverty reduction since 1980, with in-kind transfers alone accounting for approximately 20%. Because the two contributions overlap (e.g., public education spending is both an indirect investment benefit and an in-kind transfer), the combined lower bound is reported as “at least 50%” of extreme poverty reduction attributable to public policies.

Q9. Q: Why does the distributional approach yield such different results from the standard approach for the poorest 20%?

A: Two main reasons. First, cross-country data cannot accurately measure the incomes of the world’s poorest, because the poorest individuals are not all concentrated in the poorest countries — distributional accounting within countries is necessary to locate them precisely. Second, the standard approach misses two progressive channels: (a) labor income shares are higher at the bottom of the income distribution than average, so gains from schooling translate into larger income increases for the poor; and (b) supply effects redistribute schooling gains from high-skilled to low-skilled workers, a mechanism that is entirely absent from cross-country averages but directly captured in the microdata-based counterfactual.

Key Concepts

Distributional growth accounting: A framework, introduced in this paper, that combines a model of education and the wage structure with household microdata to construct income-group-specific counterfactuals, isolating the contribution of human capital accumulation to growth at each point of the income distribution rather than at the national-average level.

True return to schooling (r):* In the CES framework with imperfect skill substitution, the “true” aggregate return to schooling used in the counterfactual lies strictly between the initial return (prevailing before educational expansion, counterfactually higher because skilled-worker supply was lower) and the final return (observed after expansion, lower due to skill-supply pressure). The true return is the return that equates the model’s predicted output loss to the actual output loss from reducing education; for plausible parameters it weights initial returns at 50–70%.

Supply effects (general equilibrium effects of schooling): When the supply of skilled workers rises, their relative wage falls and the relative wage of unskilled workers rises. These wage adjustments are not captured by individual-level Mincerian returns but are modeled via the CES elasticity of substitution σ. Supply effects are central to education’s progressive distributional impact: they compress the skill premium and raise earnings at the bottom of the distribution.

Imperfect substitution between skill groups: The CES production specification in which skilled (H) and unskilled (L) labor are combined with elasticity σ < ∞. This governs the magnitude of general equilibrium wage effects: a lower σ means a larger wage compression per unit of skilled-supply increase, amplifying the redistributive role of education. The paper calibrates σ from the macroeconomics literature and bounds results over plausible ranges.

Skill-biased technical change (SBTC): Technology that raises the marginal product of skilled workers relative to unskilled (captured by the ratio AH/AL in the CES production function). SBTC amplifies returns to schooling; in the subsample of 33 countries with around-2000 surveys, SBTC accounts for 20–30% of schooling’s income benefits, but education would still have generated substantial income gains absent SBTC.

Conservative assumptions (scope condition): All headline quantitative results (50% of aggregate growth, 70% of poorest-20% income gains, 40% of extreme poverty reduction, >50% of gender inequality reduction) are explicitly conditioned on conservative assumptions: Mincerian rather than causal returns, no effect on capital income, and no human capital externalities. The paper argues these assumptions bias all estimates downward.

Summary based on HAL working paper (halshs-04423765v1, Working Paper 2023/25, November 2023). Period covered in working paper text: 1980–2022. AI-assisted, human review pending.

Diversifying Society's Leaders? Determinants and Causal Effects of Admission

Mon, 01 Jan 0001 00:00:00 +0000

This paper studies why children from high-income families are more likely to attend Ivy-Plus colleges (Ivy League, Stanford, MIT, Duke, Chicago — 12 colleges total) and whether attending these colleges causally improves post-college outcomes. The authors construct a de-identified panel dataset linking federal income tax records, Department of Education college attendance data, College Board and ACT test scores, and application and admissions records from several Ivy-Plus and flagship public colleges covering approximately 2.4 million students across entering classes from 1998–2015.

The central finding on the input side is that students from families in the top 1% of the income distribution (income above $611,000) are 2.3 times more likely to attend an Ivy-Plus college than middle-class students (defined as the 70th–80th percentiles of the national parental income distribution, approximately $91,000–$114,000) with comparable SAT/ACT scores. Two-thirds of this gap is attributable to higher admissions rates at Ivy-Plus colleges for high-income applicants; conditional on SAT/ACT scores, top-1% applicants are 58% more likely to be admitted than middle-class applicants. The remaining third splits between differences in application rates (roughly 20% of the total attendance gap) and matriculation rates (roughly 12%). In contrast, admissions rates at flagship public colleges are essentially uncorrelated with parental income conditional on test scores.

Three admissions practices drive the high-income admissions advantage at Ivy-Plus colleges. First, legacy preferences: legacy applicants from the top 1% are admitted at more than five times the rate of non-legacy applicants with comparable test scores, demographics, and admissions ratings; children of alumni of a given Ivy-Plus college are not more likely to be admitted to other Ivy-Plus colleges, confirming that legacy status is not merely a proxy for unobservable credentials. Legacy preferences account for 52 of the estimated 168 “extra” top-1% students per average Ivy-Plus class (enrollment ~1,650). Second, non-academic ratings: students from the top 1% have markedly stronger non-academic credentials (extracurricular activities, leadership ratings) partly because they disproportionately attend private high schools whose students receive higher non-academic ratings despite no higher academic ratings; this accounts for 35 additional extra top-1% students. Third, athletic recruitment: the share of recruited athletes rises from 5% among admitted students from the bottom 60% to 13% among those from the top 1%, accounting for 27 additional extra top-1% students.

On the output side, the authors estimate causal effects of attending an Ivy-Plus college using a new research design based on waitlisted applicants. The key identification assumption is that idiosyncratic variation in admissions decisions across waitlisted applicants at one Ivy-Plus college is uncorrelated with admissions decisions at other Ivy-Plus colleges — which the authors verify empirically. Under this assumption, comparisons of admitted vs. rejected waitlisted applicants identify causal effects for marginal students. The marginal student who attends an Ivy-Plus college instead of the average flagship public is approximately 50% more likely to reach the top 1% of the earnings distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm. Attending an Ivy-Plus college increases mean earnings by $101,000 at age 33 relative to a counterfactual mean of $143,000 at state flagships. Effects are concentrated in the upper tail of earnings — the impact on reaching the top quartile is small and statistically insignificant, while impacts on reaching the top 1% far exceed what a constant percentage treatment effect would predict. Effects are larger for students with weaker fallback options (i.e., whose home-state colleges channel fewer students to the top 1%).

Critically, the three credentials driving the high-income admissions advantage — legacy status, athletic recruitment, and high non-academic ratings — are uncorrelated with or negatively correlated with post-college success once the college attended is held constant. Academic credentials (SAT/ACT scores, academic ratings) remain highly predictive of outcomes.

Counterfactual simulations show that eliminating all three high-income admissions preferences and replacing those slots with students having the same test score distribution would increase enrollment from the bottom 95% of the parental income distribution by 8.8 percentage points — comparable in magnitude to the effect of race-based affirmative action on Black and Hispanic enrollment shares. Such a policy would have small effects on monetary leadership outcomes (e.g., Fortune 500 CEO share from bottom-95% families rises by only 0.4 pp, because Ivy-Plus graduates are a small fraction of all top earners) but larger effects on non-monetary leadership positions: the share of senators from the bottom 95% would rise by 1.7 pp and the share of Supreme Court justices by 5.4 pp. With need-affirmative policies (giving low-income students preferences comparable to those currently given to legacy applicants), the share of Supreme Court justices from families in the bottom 60% would rise by 17.5 pp. These predictions assume that the causal share of Ivy-Plus attendance in explaining observational differences in leadership outcomes is the same as that estimated for early-career outcomes, and they ignore general equilibrium effects.

Q: How much more likely are top-1% students to attend an Ivy-Plus college than middle-class students with the same test scores? A: Students from families in the top 1% (income above $611,000) are 2.3 times more likely to attend an Ivy-Plus college than students from the 70th–80th percentile of the parental income distribution (approximately $91,000–$114,000) with comparable SAT/ACT scores. This “missing middle” pattern is stable across entering classes from 1998 to 2018 and persists after controlling for race and ethnicity.

Q: How is the overall attendance gap decomposed into application, admissions, and matriculation? A: Differences in admissions rates explain two-thirds of the gap in Ivy-Plus attendance between top-1% and middle-class students conditional on test scores. Of the estimated 168 “extra” top-1% students per average Ivy-Plus class, 87 come from higher admissions rates for non-recruited athletes, 27 from athletic recruitment, and the remaining slack from application rate differences (accounting for roughly 20% of the overall attendance gap) and matriculation differences (roughly 12%).

Q: How large is the admissions advantage for top-1% applicants at Ivy-Plus colleges? A: Conditional on SAT/ACT scores, applicants from the top 1% are 58% more likely to be admitted to Ivy-Plus colleges than middle-class applicants. Students from the top 0.1% are 2.5 times more likely to be admitted than middle-class applicants with comparable test scores. At flagship public colleges, admissions rates are essentially constant across the income distribution conditional on test scores.

Q: What is the magnitude of legacy preferences and how is it established that legacy is not just a proxy for other credentials? A: Legacy applicants from the top 1% are admitted at more than five times the rate of otherwise comparable non-legacy applicants at the college their parents attended. The paper isolates the legacy effect by showing that children of alumni at a given Ivy-Plus college are only slightly more likely to be admitted at other Ivy-Plus colleges — and the predicted counterfactual admissions rate for legacy students at other colleges closely matches their actual admissions rate — confirming that legacy status is not merely a proxy for other unobservable credentials. Legacy applicants constitute 2.5% of the overall applicant pool but over 9% of top-1% applicants.

Q: How do non-academic credentials differ by parental income, and what drives the difference? A: Top-1% applicants have markedly stronger non-academic ratings (measuring extracurricular participation and leadership traits) compared with other applicants, while the share achieving high academic ratings is essentially constant across the income distribution. Students from the top 1% are much more likely to have attended private high schools, whose applicants receive substantially higher non-academic ratings than students from public high schools with the same SAT/ACT scores. Non-academic ratings account for 35 of the estimated 168 extra top-1% students per Ivy-Plus class.

Q: What is the research design for estimating causal effects, and what is the key identification assumption? A: The authors focus on applicants who are waitlisted at a given Ivy-Plus college and compare those ultimately admitted versus rejected from the waitlist. The key identification assumption is that if different colleges’ admissions committees make correlated assessments of underlying student merit but uncorrelated idiosyncratic admissions errors, then residual variation in admissions outcomes for waitlisted applicants at one college is orthogonal to students’ long-run potential. The authors validate this empirically by showing that waitlist admission at one Ivy-Plus college is uncorrelated with admissions decisions and internal ratings at other Ivy-Plus colleges.

Q: What are the causal effects of attending an Ivy-Plus college on post-college outcomes? A: For the marginal student (one who attends an Ivy-Plus college instead of the average flagship public), attending an Ivy-Plus college increases the probability of reaching the top 1% of the earnings distribution at age 33 by approximately 50%, nearly doubles the probability of attending an elite graduate school, and increases the probability of working at a prestigious firm by approximately 2.5 times. Mean earnings at age 33 increase by $101,000 (relative to a counterfactual mean of $143,000 at state flagships). Effects on reaching the top quartile of earnings are small and statistically insignificant, while effects at the very top tail are disproportionately large.

Q: Why do the findings differ from Dale and Krueger (2002) and related studies finding little effect of selective college attendance on earnings? A: The authors replicate the matriculation design of Dale and Krueger (comparing outcomes conditional on the set of colleges to which students were admitted) and obtain estimates statistically indistinguishable from their waitlist design — the research designs are not the source of disagreement. Instead, the differences arise because (1) the authors have direct college fixed effects rather than relying on average test scores as a proxy for college quality, and (2) the authors focus on upper-tail outcomes (top 1% earnings, elite graduate schools, prestigious firms) rather than log mean earnings, where Ivy-Plus colleges have their largest effects.

Q: Are the credentials that drive the high-income admissions advantage — legacy, athlete status, high non-academic ratings — predictive of better post-college outcomes? A: No. Recruited athletes, students with higher non-academic ratings, and legacy students have equivalent or lower chances of reaching the upper tail of the income distribution, attending an elite graduate school, or working at a prestigious firm than comparable Ivy-Plus applicants once the college attended is held constant. By contrast, SAT/ACT scores and academic ratings are highly positively predictive of all three post-college outcome measures.

Q: How much could changing admissions practices diversify Ivy-Plus enrollment and subsequently society’s leadership? A: Eliminating legacy preferences, non-academic rating weights, and the differential recruitment of high-income athletes — and filling those slots with students having the same test score distribution as the current class — would increase enrollment from families in the bottom 95% of the parental income distribution by 8.8 percentage points, a magnitude comparable to race-based affirmative action’s effect on Black and Hispanic enrollment shares. For leadership positions, predicted effects are small for monetary outcomes (Fortune 500 CEOs from the bottom 95% would increase by only 0.4 pp) but larger for positions where Ivy-Plus graduates are a larger share: senators from the bottom 95% would increase by 1.7 pp and Supreme Court justices by 5.4 pp. A stronger need-affirmative policy (giving low-income students preferences equivalent to current legacy preferences) would increase the share of Supreme Court justices from the bottom 60% by 17.5 pp.

Q: How are “elite” and “prestigious” employers defined in this study? A: Elite firms are defined as those that disproportionately employ Ivy-Plus graduates relative to flagship public graduates, pulling firms from the top of that ratio ranking until 25% of Ivy-Plus attendee employment is accounted for. Prestigious employers are defined by the residual of that ratio after controlling for the firm’s predicted top-1% income probability — they are firms that disproportionately employ Ivy-Plus graduates conditional on their salaries, capturing high-status jobs that do not necessarily lead to the highest earnings. The paper validates this algorithmic approach against external rankings (Vault.com for law and consulting firms; Scimagoir for hospitals), finding substantial overlap.

Q: How are treatment effect estimates adjusted for heterogeneity in students’ fallback options? A: Causal effects of Ivy-Plus attendance are much larger for students with weaker fallback options — specifically, students whose home-state flagship colleges channel fewer students to the top 1% of earnings. The authors exploit this heterogeneity to estimate the treatment effect for the marginal student who actually switches from a flagship public to an Ivy-Plus college. This heterogeneity also implies that the average causal effect across all admitted students may differ from the effect for the marginal admitted student.

Q: What share of the overrepresentation of top-1% families at Ivy-Plus colleges is attributable to pre-application factors versus admissions practices? A: Of the 245 “extra” top-1% students in an average Ivy-Plus class relative to an unconditionally income-neutral benchmark, 77 (31%) are attributable to the higher test scores of top-1% students (a pre-application factor). The remaining 168 (69%) reflect higher attendance rates conditional on test scores, of which the large majority is attributable to admissions practices (legacy, non-academic ratings, athletic recruitment) rather than application or matriculation rate differences.

Ivy-Plus colleges: The twelve highly selective private colleges comprising the eight Ivy League institutions plus Stanford, MIT, Duke, and the University of Chicago — the focus group of the study, which together account for more than 10% of Fortune 500 CEOs, a quarter of U.S. senators, and three-fourths of Supreme Court justices appointed in the last half century despite enrolling less than 0.5% of Americans.

Missing middle: The pattern by which attendance rates at Ivy-Plus colleges conditional on SAT/ACT scores are lowest for students from the middle class (70th–80th percentile of the parental income distribution, approximately $91,000–$114,000) — lower than both the top 1% and, slightly, the bottom 40% — producing a non-monotone income gradient in attendance.

Legacy preference: An admissions advantage given to applicants whose parent(s) obtained an undergraduate degree from the college to which the student is applying. In the paper’s data, legacy applicants from the top 1% are admitted at more than five times the rate of non-legacy applicants with comparable test scores, demographics, and admissions ratings; the preference is college-specific (children of alumni are only slightly more likely to be admitted at other Ivy-Plus colleges).

Waitlist research design: The paper’s primary identification strategy for causal effects, which exploits idiosyncratic variation in admissions decisions among waitlisted applicants. The design’s validity rests on the empirical finding that waitlist admissions at one Ivy-Plus college are uncorrelated with admissions decisions and internal ratings at other Ivy-Plus colleges, implying that residual variation conditional on being on the waitlist is orthogonal to students’ long-run potential outcomes.

Prestigious employers: Firms defined by the paper’s algorithm as disproportionately employing Ivy-Plus graduates conditional on those firms’ predicted top-1% income probability — capturing high-status employment that does not necessarily lead to the highest earnings (e.g., prominent law firms, consulting firms, elite hospitals). Validated against external rankings (Vault.com, Scimagoir).

Non-academic ratings: Numerical scores assigned by admissions officers measuring aspects of an application outside academic achievement, such as extracurricular activities and leadership traits. In the paper’s data, non-academic ratings differ substantially by parental income — particularly because top-1% applicants disproportionately attend private high schools whose students receive higher non-academic ratings — while academic ratings do not differ across the income distribution.

Surrogate index: A prediction of later earnings outcomes (specifically, probability of reaching the top 1% at age 33 and mean income rank) constructed from individuals’ graduate school attendance and employer fixed effects at ages 22–25, used to extend the outcome window for cohorts observed only early in their careers. The approach follows the terminology and methodology of Athey et al. (2019).

Do Financial Concerns Make Workers Less Productive?

Mon, 01 Jan 0001 00:00:00 +0000

Do Financial Concerns Make Workers Less Productive?

Research Question

The paper tests whether financial concerns distract workers sufficiently to meaningfully reduce their productivity, and whether receiving cash — by alleviating those concerns — can raise output even when total compensation is held fixed.

Setting and Sample

The experiment involves 408 low-income male agricultural casual laborers in rural Odisha, India, recruited from 47 villages across five worksites in four districts. The study takes place during the lean agricultural season (March–June 2017 and 2018), when formal employment is scarce (workers found paid wage work on only 1.9 days per week on average). During this period, 86% of workers reported being “worried” or “very worried” about their finances, 68–71% carried outstanding loans, and 64–66% said they would have difficulty coming up with Rs. 1,000 (roughly four days of wages) in an emergency. Workers bring these burdens to the job: on a given day, approximately one in two workers reported thinking about financial worries while working.

Experimental Design

Workers were employed for twelve days in a piece-rate manufacturing task — stitching sal tree leaves into disposable plates for restaurants. The payment-timing manipulation is the core of the identification strategy. Control workers received all accrued earnings as a lump sum on the final day (day 12). Treatment workers received their earnings in two installments: an interim payment of earnings to date on day 8 or 9 (randomly staggered across waves), with the balance paid on day 12. Total compensation was held constant across groups; only the timing of receipt differed. On day 5 (the “announcement day”), each worker learned his payment schedule individually. The design thus separates the announcement period (days 5 through the interim payment day, when workers know their schedule but have not yet received cash) from the post-pay period (days after the interim payment until the contract end). This enables the authors to test whether productivity effects arise from information about impending cash, or only once cash is physically in hand.

First Stage: Effects on Financial Strain

Within three days of receiving the interim payment, treated workers increased loan repayments by Rs. 271, a 287% increase relative to the control group mean (p < 0.001), and were 40 percentage points (222%) more likely to repay any loan (p < 0.001). The majority of repayments occurred on the same evening as the cash disbursement — a 746% single-day increase in loan payments. Household expenditures on food, clothing, and essentials rose by 40% (Rs. 150) over three days (p < 0.001). Treatment workers also reported feeling more focused on the work task (11.5 percentage points more likely, p = 0.032) and were less likely to report thinking about financial worries while making plates (13.7 percentage points, p = 0.044).

Main Productivity Results

In the post-pay period, treated workers increased output by 0.109 SD (6.9%) relative to the control group (p = 0.020). No treatment effect emerged during the announcement period (0.014 SD, p = 0.685); the post-pay and announcement-period effects are statistically distinguishable (p = 0.008). Because work hours are fixed and daily attendance is 98.3% with no treatment effect on attendance, these gains reflect improvements in how quickly workers produce plates per hour of work.

Effects are concentrated among workers with below-median baseline wealth (fewer assets, less liquidity): for this subgroup, the interim payment increases output by 0.204 SD (13.0%, p = 0.003). For workers with above-median wealth, the effect is close to zero and statistically insignificant (p = 0.819).

Attentiveness Results

Beyond total output, the authors measure attentiveness through three markers embedded in the finished plates: the number of “double holes” (paired stitching holes indicating a removed mistaken stitch), the number of leaves used, and the number of stitches used. These measures are collected unbeknownst to workers and combined into an “attentiveness index.” After receiving the interim payment, treated workers’ attentiveness index increased by 0.077 SD across all workers (p = 0.092); among poorer workers, attentiveness increased by 0.17 SD (p = 0.041). This improvement occurred simultaneously with higher output speed — workers were producing plates faster while also making fewer mistakes, suggesting improved cognitive engagement rather than mere effort intensification.

Piece-Rate Comparison

In separate supplementary rounds with 150 experienced workers, the authors varied piece rates (Rs. 2, 3, or 4) while holding overall earnings constant. Each one-rupee increase in the piece rate raised output by 0.020 SD (p = 0.042). Critically, piece-rate increases produced no detectable change in the attentiveness index (point estimate negative, statistically insignificant), and the piece-rate effect on output differs significantly from the attentiveness effect (p = 0.001). This indicates that consciou effort and automatic attentiveness can move independently: higher incentives increase pace but do not reduce attentional lapses, whereas financial relief increases both pace and attentiveness.

Alternative Explanations Ruled Out

The authors systematically address gift exchange/fairness, trust, nutrition, and sleep. Fairness and gift-exchange stories are inconsistent with: (i) no detectable announcement-period effect; (ii) no decline in control-worker effort when treatment workers are paid before them; (iii) the pattern of effects being concentrated among poorer workers; and (iv) attentiveness being affected when it is not a sanctioned quality dimension for payment. Nutritional channels are inconsistent with overnight effect onset (nutritional stock changes are too slow biologically), no treatment effect on breakfast consumption patterns, and productivity effects persisting through the end of each workday. Sleep channels are inconsistent with no treatment effect on hours or quality of sleep.

Scope Conditions and Implications

The effect operates through the actual arrival of cash, not its anticipation, consistent with a model in which automatic cognitive inputs — unlike consciously chosen effort — respond to current financial strain rather than expected future income. Effects are concentrated among more financially constrained workers within an already-poor sample. The authors do not identify the specific psychological mechanism (worry, anxiety, affect, or rumination) but interpret results as evidence that financial strain, at least partly through psychological channels, reduces earnings exactly when money is most needed.

In depth

Q1. Why does the experiment focus on payment timing rather than an outright transfer of additional money?

Varying only payment timing — not total pay — holds constant both the piece-rate incentive and total wealth across treatment and control. An outright cash transfer would raise total lifetime income, potentially reducing effort through a neoclassical income effect (more lifetime wealth lowers the marginal utility of current consumption). By holding total compensation fixed and only shifting when it arrives, the design isolates the effect of financial strain per se, separable from any wealth or incentive effect.

Q2. Why is there no treatment effect during the announcement period, and why does this matter?

Between day 5 (when workers learn their payment schedule) and the interim payment date, treated workers know cash is coming but have not yet received it. Output in this window shows no treatment effect (0.014 SD, p = 0.685), and the announcement effect is significantly smaller than the post-pay effect (p = 0.008). This matters because it rules out mechanisms that should operate on information alone — including gift exchange, trust updating, or effort responses to higher discounted expected income — and is consistent with a model in which financial strain falls only when cash is physically received (e.g., moneylenders do not relent until the loan is actually repaid).

Q3. What is the attentiveness index and how was it constructed?

The attentiveness index averages three plate-level markers: (i) number of “double holes” — pairs of stitching holes indicating a mistaken stitch was removed; (ii) number of leaves used; and (iii) number of stitches used. Each component was normalized using the control group’s post-pay mean and standard deviation, then averaged and reverse-coded so that higher values denote better attentiveness (fewer mistakes, fewer leaves, fewer stitches). Workers were unaware these dimensions were being measured. The index thus captures the number of unforced steps a worker took to complete a plate — a behavioral trace of cognitive lapses.

Q4. How do the piece-rate rounds demonstrate that effort and attentiveness are separable?

In supplementary rounds (150 workers, 2019), piece rates were experimentally varied among Rs. 2, 3, and 4 per plate with the base wage adjusted to hold total earnings constant, so financial strain was unchanged. A one-rupee increase in the piece rate raises output by 0.020 SD (p = 0.042), consistent with a standard effort response. The same increase produces no discernible change in the attentiveness index (point estimate: negative but not significant), and the output and attentiveness effects are significantly different from each other (p = 0.001). This shows that workers can speed up via conscious effort without reducing attentional lapses, whereas the cash infusion raises both pace and attentiveness simultaneously — a pattern inconsistent with pure motivation as the mechanism.

Q5. What does the staggered timing within the treatment group (Wave A vs. Wave B) contribute to identification?

Treatment workers were randomized to receive their interim payment on day 8 (Wave A) or day 9 (Wave B). On day 9, Wave B workers have not yet been paid while Wave A workers have. If fairness concerns drove control workers to reduce effort upon seeing colleagues paid first, control workers on day 9 — having observed Wave A payments the evening before — should work less hard relative to Wave B treatment workers (who have also not yet been paid). The authors find no such pattern: the triple interaction (Cash × Payment Day × Wave B) is close to zero and insignificant, ruling out effort reductions from seeing peers paid earlier.

Q6. What are the magnitudes and timing of the spending response to the cash infusion?

Within three days of the interim payment, treatment workers spent Rs. 900 in total — roughly two-thirds of the average interim payment of over Rs. 1,400. On the day of the payment itself, loan repayments rose by Rs. 169 (746% increase), and household expenditures rose by Rs. 70 (68% increase). Over three days, loan repayments increased by Rs. 271 (287%), the probability of repaying any loan rose by 40 percentage points (222%), and total household spending rose by 65% (Rs. 371). These patterns indicate that the two main sources of financial stress cited by workers — outstanding debt and inability to meet household essentials — were directly addressed, suggesting a meaningful reduction in financial strain.

Q7. Why are the productivity effects concentrated among poorer workers, and what are the two interpretations?

Workers with below-median baseline wealth (fewer assets, lower liquidity) show a 0.204 SD (13.0%) productivity gain, while workers above the median wealth threshold show essentially no effect. The authors offer two interpretations. First, poorer workers may start from a higher level of financial strain, giving the intervention more scope to reduce it. Second, since all workers in the sample are objectively poor and report similar baseline financial worries and loan levels, the more likely explanation is that the interim payment is larger relative to the wealth and income buffer of poorer workers, making the same nominal cash infusion more meaningful for them. Both richer and poorer workers in the sample use the interim payment to repay loans and cover household needs.

Q8. How do the authors rule out nutritional channels?

Two tests address nutrition. First, workers were not at subsistence — 94% reported missing no meals the prior week — and increased food spending cannot change the nutritional stock overnight (the medical literature indicates nutritional-stock effects on cognition operate over longer time horizons). Second, and more precisely, all food consumed at the worksite during the workday was provided by the researchers, so differential pre-worksite breakfast consumption is the only plausible same-day biological channel. The authors find no treatment effect on breakfast consumption (whether workers had breakfast, how much, or what they ate). Further, if blood sugar or satiety drove effects, they should attenuate over the workday as all workers are given the same afternoon meal; instead, treatment effects persist and if anything increase through the final hours of the workday.

Q9. What does the self-report evidence on focus and worry show, and why is it treated as suggestive rather than primary?

Two days after the interim payment, workers were asked an open-ended question about what they were thinking about while working. Treatment workers were 11.5 percentage points (15.5%) more likely to report feeling focused on the task (p = 0.032) and 13.7 percentage points (32.7%) less likely to report thinking about financial worries (p = 0.044). A supplementary test showed treated workers were 10 percentage points (31%) more likely to generate explanations for a low-income person’s negative affect that were unrelated to financial concerns (p < 0.05), suggesting a broadening of cognitive scope. These measures are treated as suggestive because they were collected only at a single point and are self-reported; the primary evidence rests on objective production data because it is more objective and collected at fine hourly resolution throughout the post-pay period.

Q10. What does the paper say about optimal payment frequency as a policy implication?

The authors are cautious in drawing a direct policy inference about paying workers more frequently. While the positive productivity effect of early payment points toward more frequent paydays reducing financial strain, this must be weighed against workers’ self-control problems in consumption. In settings where workers face lumpy expenditure needs (e.g., monthly rent), more frequent payments could cause under-saving and worsen strain at the time of lumpy bills. The authors suggest payment frequency or size that matches expenditure needs, or more generally financial products that allow workers to time income receipts to coincide with expenses, as potentially more robust solutions — noting that such products appear largely absent in these markets.

Key Concepts

Financial strain (as used in the paper): A psychological burden arising from pressing present needs for resources — defined in the authors’ model as increasing in both the current marginal utility of consumption (i.e., how valuable an additional rupee would be today) and the level of outstanding debt (including lender harassment pressure). Strain is present-oriented: it responds to current cash-on-hand and debt levels, not to expected future income, which is why anticipating a payment does not fully relieve it.

Automatic input (a): In the authors’ behavioral model, one of two inputs into production. Unlike “effortful” input (e), which the worker consciously controls (speed of hands, consciously directed attention), the automatic input captures cognitive functions that are beyond the worker’s full control — background attentional processes that can be degraded by financial strain even when a worker is motivated and exerting high effort. The key behavioral assumption is that a falls when financial strain is high, independently of chosen effort.

Attentiveness index: A composite measure constructed from three unincentivized physical markers embedded in completed leaf plates: (i) number of double holes (pairs indicating a stitch was removed to correct a mistake); (ii) number of leaves used; (iii) number of stitches used. The index is normalized to the control group’s post-pay distribution and reverse-coded so higher values denote better attentiveness. Workers were unaware these dimensions were measured. The index captures attentional lapses — unforced errors that increase the number of steps and time needed to complete each plate.

Announcement period: The days between when workers are individually informed of their payment schedule (day 5) and when the interim payment is actually disbursed (day 8 or 9). This window serves as a within-experiment control: if effects arose from information about impending cash (e.g., through discounting, gift exchange, or trust), they should appear here. The consistent absence of treatment effects during this period is a key identification result.

Post-pay period: The days from the interim payment until the contract end (day 12). The main productivity and attentiveness treatment effects are estimated in this window, comparing treatment workers (who have received cash) to control workers (who have not yet been paid).

Lean season: The months outside the peak agricultural planting and harvesting periods (roughly six to eight months per year in the study area) during which agricultural workers seek intermittent casual employment in manufacturing, construction, and other sectors. Employment rates are low (1.9 paid days per week on average), income is low and variable, and financial strain is correspondingly high. The experiment is intentionally conducted during this period to maximize baseline levels of financial concern.

Piece-rate elasticity of effort: The responsiveness of output to changes in the marginal return per unit produced (the piece rate), holding financial strain constant. In the supplementary rounds, a one-rupee increase in the piece rate raises output by 0.020 SD. The authors interpret this as the upper bound on how much pure motivational effort can move output in this task, and use it to benchmark the cash infusion effects, which are roughly five times larger per unit of treatment variation and additionally move attentiveness (which piece-rate changes do not).

Dollar Dominance and the Transmission of Monetary Policy

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Summary

An emerging view in international macroeconomics contends that dollar invoicing of exports renders monetary policy ineffective for non-U.S. countries: because export prices are allegedly sticky in dollars, exchange rate depreciations cannot shift expenditure toward domestic goods, muting the classical Mundell-Fleming channel. McLeay and Tenreyro argue that this view rests on empirical assumptions that are not borne out by the data: goods priced in dollars tend to have more flexible prices and higher elasticities of substitution, not the monopoly power and sticky dollar prices assumed in dominant currency pricing (DCP) models. They propose a mixed currency pricing (MCP) framework that incorporates heterogeneous price flexibility and intra-sector international competition, and show that even with dollar pricing, depreciating the currency by loosening monetary policy can still boost exports and activity materially. The limit to any expansion is not demand, but supply capacity: after a depreciation, domestic dollar costs fall, flexible-price exporters lower prices slightly and gain large market share due to high demand elasticities, and the expansion runs until rising marginal costs offset the initial depreciation — producing limited reduced-form dollar pass-through as an equilibrium result rather than evidence of nominal stickiness. Empirical tests using monetary policy shocks in a sample of emerging and developing economies, case studies of Canada and Chile as commodity exporters, and three large devaluation episodes all find significant, material increases in exports and aggregate activity following exchange-rate depreciations, consistent with the MCP model’s predictions.

In depth

Q1. What is the specific empirical claim that DCP models rest on, and how do McLeay and Tenreyro challenge it?

DCP models (e.g., Gopinath et al. 2020) posit that exporters invoicing in dollars have monopoly power and face nominal rigidities that keep their dollar export prices sticky. The observable implication used to motivate this assumption was limited exchange rate pass-through to dollar export prices. McLeay and Tenreyro show that low pass-through is equally consistent with a flexible-price, high-elasticity equilibrium. When demand elasticities are high, firms optimally absorb exchange rate changes through quantities rather than prices; the reduced-form pass-through coefficient is small even without any nominal friction. Low pass-through is therefore not informative about the degree of nominal rigidities, and using it to calibrate sticky-price DCP models and draw normative conclusions about exchange rate policy is unwarranted.

Q2. What are the three empirical facts that motivate the MCP framework’s assumptions?

Fact 1: Homogeneous products (commodities and commodity-like goods traded on organized exchanges or reference-priced, following Rauch 1999) represent a large share of goods exports, exceeding 70% for developing economies, around 60% for emerging economies, and around 35% for advanced economies; Sub-Saharan Africa, Latin America, and the Middle East all have shares above 50%. Fact 2: Homogeneous and more competitively produced goods have more flexible prices, documented across multiple countries — for instance, Nakamura and Steinsson (2008) find a median monthly price-change frequency of 10.8% for finished-good producer prices but 98.9% for crude materials. Fact 3: Dollar (vehicle currency) invoicing is most prevalent precisely in these homogeneous, competitive-good sectors; classical work by McKinnon (1979) and Magee and Rao (1980) emphasized that vehicle-currency invoicing facilitates continuous price comparability in competitive markets, and panel regressions corroborate a positive relationship between the share of exports invoiced in dollars and the homogeneous-goods share of exports.

Q3. What is the mechanism through which depreciation boosts exports in the MCP model, and why does this generate low observed pass-through?

With sticky wages (representing non-tradable input price stickiness more broadly), a monetary policy-induced depreciation lowers the domestic cost of production when expressed in dollars. For competitive exporters facing highly elastic demand, even a small reduction in the dollar price translates into a substantial gain in export quantities. Firms therefore lower their dollar prices slightly, trading some profit margin for a large increase in market share. As exports expand, domestic marginal costs rise (firms move up the upward-sloping marginal cost curve), partially offsetting the depreciation’s effect on dollar costs. In equilibrium, the net dollar price movement is small — producing the observed limited pass-through — but the quantity response is large. In the perfectly competitive limit (relevant for commodity exporters), the dollar price is unchanged by the world market, and the entire adjustment is through an expansion of export volumes until rising domestic marginal costs absorb the depreciation. The implied observation is identical to a sticky-price model for prices, but “the implications for export quantities are diametrically opposed.”

Q4. How does the MCP model nest existing frameworks, and what does it add relative to the DCP and PCP benchmarks?

The MCP (mixed currency pricing) framework nests sticky-price DCP as a special case (by setting demand elasticities low and allowing full price stickiness) and produces behavior close to PCP (producer currency pricing) in the flexible-price, high-elasticity limit — restoring the allocative properties of the exchange rate from Obstfeld and Rogoff (1995). The distinctive addition is intra-sector international competition: domestic exporters face competition from international competitors producing highly substitutable varieties of the same good, so substitution elasticities can be high at the variety level even when macro-level elasticities between goods remain low. This follows a bottom-up approach to elasticities as in Feenstra et al. (2018). The model also allows heterogeneous nominal rigidities across producers, with exporters of dollar-invoiced homogeneous goods having flexible prices while non-tradable input prices (wages) remain sticky — the source of monetary non-neutrality and the mechanism for real exchange rate effects.

Q5. What is the role of supply capacity, and why is it “the limit” rather than demand?

In the sticky-price DCP model, the constraint on the export response is on the demand side: dollar prices do not move, so demand is unchanged, and there is no export response at all. In the MCP model, demand responds immediately to the cost reduction — the constraint that eventually stops the expansion is supply capacity, captured by the slope of the marginal cost curve and macroeconomic constraints on non-tradable inputs. With a flat marginal cost curve (plentiful supply capacity), exports expand materially; with a steep curve or hard capacity constraints, the increase in marginal cost fully offsets the depreciation before much quantity adjustment occurs. This supply-side framing reorients the policy question: the limiting factor for monetary policy’s external effectiveness is not whether dollar prices can move, but whether the domestic economy has the productive capacity to expand tradable output. This also connects the paper to the Salter-Swan two-good framework and to Schmitt-Grohé and Uribe (2021).

Q6. What do the macroeconomic empirical tests find, and how do they distinguish the MCP from sticky-price DCP?

The paper uses three empirical exercises. First, using a sample of developing and emerging economies, monetary policy expansions that generate exchange rate depreciations cause significant increases in both exports and aggregate economic activity — consistent with the MCP model’s material export response and inconsistent with the DCP prediction of no export channel. Second, focusing on Canada and Chile as commodity exporters where the MCP assumptions (competitive markets, flexible export prices) are especially applicable, the aggregate results are corroborated and sectoral evidence provides additional support. Third, three case studies of large devaluations in the sample document that they are followed by material increases in exports relative to trend. In all exercises, the direction and magnitude of export and output responses are consistent with a functioning expenditure-switching channel, even where exports are priced in dollars.

Q7. How does the paper reinterpret the pass-through evidence that motivated sticky-price DCP models, and what does this imply for normative conclusions?

Standard reduced-form pass-through regressions relate the change in dollar export prices to changes in the exchange rate. These regressions typically omit or fail to fully capture movements in marginal cost. In the MCP model, flexible-price firms fully pass through changes in marginal cost; the observed limited pass-through to export prices is an equilibrium result of the offsetting rise in marginal costs as export volumes expand, not evidence of a nominal friction. Because the standard regressions omit marginal cost dynamics, they risk attributing the equilibrium quantity-driven equilibrium to a pricing friction. This has direct normative implications: the case made by the IMF (2019, 2020) that dollar invoicing worsens the cost-benefit calculation for flexible exchange rates — and may bolster the case for capital controls — rests on interpreting low pass-through as evidence of stickiness. If low pass-through instead reflects high demand elasticities and supply-side adjustment, the normative argument for constraining exchange rate flexibility is weakened.

Q8. How does the paper relate to the purchasing power parity puzzle and the Mussa puzzle?

The MCP framework offers explanations for two classic international macro puzzles without assuming nominal rigidities in export prices. On the PPP puzzle (the volatility and persistence of the real exchange rate, Rogoff 1996): in the MCP model, exporters’ optimal reset prices move very little after exchange rate changes — not because of stickiness, but because demand is elastic and marginal costs rise quickly. This predicts limited movement in relative export prices, consistent with empirical evidence in Blanco and Cravino (2020) and Itskhoki and Mukhin (2025). On the Mussa puzzle (the large jump in nominal and real exchange rate volatility after the Bretton Woods collapse): the model’s mechanism via sticky wages is consistent with evidence that depreciations produce slow adjustment of non-tradable prices (Burstein, Eichenbaum, and Rebelo 2005), generating real exchange rate movements despite limited response in traded-good dollar prices.

Key Concepts

Dominant currency pricing (DCP): A framework in which non-U.S. exporters set and maintain prices in U.S. dollars, with sticky dollar prices. As formulated by Gopinath et al. (2020), DCP predicts that exchange rate depreciations by non-U.S. countries do not reduce dollar export prices and therefore do not stimulate export demand — muting the expenditure-switching channel of monetary policy.

Mixed currency pricing (MCP): The framework introduced in this paper. It allows heterogeneous price flexibility and market structure across export sectors, nesting both sticky-price DCP and flexible-price PCP as special cases. Dollar-priced exports face elastic demand from international competition, have flexible prices, and respond to depreciations through quantities rather than prices. Non-traded inputs (wages) remain sticky, providing the source of monetary non-neutrality.

Expenditure-switching channel: The mechanism by which exchange rate depreciations redirect spending toward domestically produced goods, boosting exports and aggregate demand. In PCP models, this works through a fall in relative export prices. In the MCP model, it works through an expansion in export quantities even when dollar prices change little.

Exchange rate pass-through (to export prices): The elasticity of dollar export prices with respect to the nominal exchange rate. In sticky-price DCP models, low pass-through reflects a nominal friction (prices cannot adjust). In the MCP model, low pass-through reflects high demand elasticities and offsetting marginal cost increases: it is an equilibrium outcome, not a friction, and therefore does not imply that export volumes are unresponsive.

Intra-sector international competition: The market structure feature central to the MCP framework. Domestic exporters of a given good compete with foreign suppliers of highly substitutable varieties, making their demand elastic at the variety level even if aggregate elasticities across different goods categories are low. This follows Armington (1969) as implemented by Feenstra et al. (2018).

Supply capacity constraint: In the MCP model, the binding constraint on how much a depreciation can boost exports. With high demand elasticities, demand for domestic exports expands freely; the limit is set by how quickly rising domestic marginal costs absorb the improvement in export profitability. The supply constraint replaces the demand constraint that operates (mechanically, via zero price response) in sticky-price DCP models.

Homogeneous goods (Rauch 1999 classification): Goods traded on organized commodity exchanges or reference-priced in trade publications, as opposed to differentiated goods. McLeay and Tenreyro use this classification to establish that dollar-invoiced exports are disproportionately homogeneous, competitive, and flexible-priced — contrary to the DCP assumption of monopoly power and price stickiness.

Summary based on published open-access version. AI-assisted, human review pending.

Enlightenment Ideals and Belief in Progress in the Run-up to the Industrial Revolution

Mon, 01 Jan 0001 00:00:00 +0000

This paper tests Joel Mokyr’s claim that Britain’s industrialization was preceded and enabled by a cultural shift — specifically, that Enlightenment ideals produced a “progress-oriented” view of science that diffused to artisans and craftsmen. The central research question is whether and when the language of science became more progress-oriented in the build-up to the Industrial Revolution, and whether this shift was concentrated in volumes directly linked to industrial production.

The authors assemble 173,031 unique volumes printed in England and written in English between 1500 and 1900, drawn from the Hathitrust Digital Library. Because copyright law prohibits downloading full text, they use HDL’s Extracted-Features “bag of words” dataset. After removing duplicates and Latin-language volumes from an initial set of 420,081, they apply Latent Dirichlet Allocation (LDA) with cross-validated perplexity minimization to identify an optimal T=60 topics. Topic-pair co-occurrence analysis identifies three categories — science, religion, and political economy — each anchored by three defining topics. Volume-level category weights are derived by multiplying each topic’s weight by its category coefficient. The resulting classification yields 50,090 science volumes, 102,565 political economy volumes, and 14,124 religion volumes.

Progressive sentiment is measured using a seven-word dictionary (progress, improvement, stride, betterment, advance, rise, amelioration) assembled from thesaurus synonyms for “progress,” manually vetted by all four authors, and restricted to words attested in the Oxford English Dictionary before 1643 (Newton’s birth year). Sentiment for each volume equals the count of progress-dictionary words divided by total word count. An analogous optimism-sentiment placebo dictionary is constructed separately.

Industrial relevance is scored using the digitized indexes of all five volumes of Appleby’s Illustrated Handbook of Machinery (1877–1903); the top industrial root words are crane (weight 51), electr (42), weight (37), rope (27), and cost (27). Each volume receives an industry score equal to the weighted occurrence of industrial root words normalized by volume length.

Three main findings emerge. First, the language of science and religion showed little overlap beginning in the 17th century — that is, the secularization of science predates the onset of industrialization. Science volumes shifted from approximately 40 percent religious content around 1700 to only about 10 percent by 1850, with scientific content rising correspondingly from roughly 40 percent to over 60 percent. This trend was stable from 1650 through 1900.

Second, while scientific volumes became more progress-oriented during the Enlightenment, this progressive shift was concentrated in volumes at the nexus of science and political economy. Volumes of “pure” science were largely neutral with respect to progress sentiment, and those at the science-religion nexus had on average negative progress sentiment. The marginal effect of scientific content on progress sentiment was greatest for volumes mixing science and political economy, and most of the increase in predicted sentiment at that nexus occurred during the 18th century, remaining stable thereafter. A placebo test using optimism sentiment finds the opposite pattern: volumes at the science-political economy nexus were among the least optimistic, while the most optimistic language appeared at the religion-political economy nexus. This rules out the interpretation that the measured shift reflects a general increase in positive affect rather than specifically progress-oriented language.

Third, volumes employing industrial terminology that also sat at the science-political economy nexus were distinctively progressive beginning in the mid-18th century. At the 90th percentile of industry score, predicted progress sentiment at the science-political economy nexus was positive throughout the sample; at zero industry score, it was negative until the mid-18th century. Volumes at the religion-political economy nexus showed modestly positive and time-stable progress sentiment regardless of industry score.

The paper concludes that it was the pragmatic, applied volumes — those bridging science and political economy, written for artisans and a broader literate public rather than for the human-capital elite alone — that embodied the cultural values Mokyr identifies as central to Britain’s industrialization.

Q: What gap in the existing literature does this paper address?

A: Prior work on the cultural deep roots of economic growth rarely tracks how culture changes over time, relying instead on cross-sectional variation or qualitative case studies. Quantitative evidence that the language of science itself became more progress-oriented — and that this change reached beyond elite thinkers to artisans and craftsmen — had not been marshaled before. The paper provides inaugural quantitative support by analyzing 173,031 volumes spanning four centuries.

Q: Why does the paper restrict the progress-sentiment dictionary to words attested before 1643?

A: Words that entered English only after 1643 (Newton’s birth year) could not have appeared in volumes from the early Enlightenment, so including them would bias sentiment scores toward the later part of the sample. The restriction ensures the dictionary is applicable and unbiased across the full 1500–1900 period. The final retained words are: progress, improvement, stride, betterment, advance, rise, amelioration.

Q: How does LDA classify volumes, and how is T=60 selected?

A: LDA treats each volume as a bag of words and derives a Dirichlet distribution such that observed documents are generated by repeated topic sampling. The number of topics T is selected by minimizing perplexity on held-out data via 4-fold cross-validation, rotating training and test sets across folds; this procedure yields T=60 as optimal. Each volume is then represented as a mixture over those 60 topics.

Q: What are the three categories and their anchor topics?

A: Political Economy is anchored by topics on law/public opinion, governance/parliament, and trade/price/labour. Religion is anchored by topics on church/Christian doctrine, God/faith/sin, and virtue/fame/religion. Science is anchored by topics on engineering/steam/electricity, chemistry/acid/heat, and geometry/equations/trigonometry. These three sets of topics were selected for high corpus-wide importance and mutual independence.

Q: What does the finding on science-religion separation imply for timing?

A: The separation of scientific and religious language was already visible by 1600 and firmly established by the mid-17th century, well before the Industrial Revolution conventionally dated to the mid-18th century. This supports Mokyr’s argument that the secularization of science was an Enlightenment-era precursor to industrialization rather than a product of it. The trend remained stable from 1650 through 1900.

Q: How does the progressive sentiment differ between pure science and the science-political economy nexus?

A: Volumes of pure science were largely neutral with respect to progress-oriented language and in some periods showed slightly negative predicted progress sentiment. The science-religion nexus showed consistently negative progress sentiment. By contrast, volumes at the science-political economy nexus showed the highest level of progressive sentiment beginning in the mid-18th century, and most of this growth in predicted sentiment occurred during the 18th century, after which it remained stable.

Q: What does the placebo optimism test show?

A: The optimism sentiment scores are nearly the mirror opposite of the progress scores: the most optimistic language appears at the religion-political economy nexus, while volumes at the science-political economy nexus are among the least optimistic. This dissociation rules out the interpretation that the measured progress-sentiment rise reflects a general shift toward positive language rather than a specific cultural embrace of science as a tool for improving human welfare.

Q: How is the industrial score constructed and what are the most heavily weighted terms?

A: The authors digitized the detailed indexes of all five volumes of Appleby’s Illustrated Handbook of Machinery (1877–1903), restricted to words attested before 1643, and weighted each industrial root word by its index frequency. Each corpus volume’s industry score equals the sum of (word count × index weight) across all industrial words, normalized by volume length, yielding a score between 0 and 1. The top-weighted terms are crane (51), electr (42), weight (37), rope (27), and cost (27).

Q: What is the key result linking industrial scores to progressive sentiment?

A: At the science-political economy nexus, volumes with industry scores at the 90th percentile had persistently positive predicted progress sentiment throughout the sample, while volumes at that nexus with zero industry score had negative predicted sentiment until the mid-18th century. The shift to positive sentiment for high-industry volumes at this nexus occurred in the mid-18th century — roughly coinciding with the onset of Britain’s industrialization — and those volumes remained the most progress-oriented in the corpus thereafter.

Q: What is the paper’s interpretation of the science-political economy nexus finding in relation to Mokyr?

A: The authors interpret volumes at the science-political economy nexus as pragmatic, applied works aimed at a broader literate audience including artisans and craftsmen, not exclusively the human-capital elite. These are precisely the volumes Mokyr’s “Industrial Enlightenment” thesis predicts would carry progress-oriented cultural values into the mechanical and artisanal pursuits that drove industrialization. The finding that pure-science volumes were not especially progressive, while applied volumes bridging science and political economy were, is consistent with Mokyr’s argument that it was the diffusion of Enlightenment ideals to skilled practitioners — not just to elite scientists — that mattered.

Q: What qualitative examples support the quantitative findings?

A: Martin Clare’s The Motion of Fluids (1735) explicitly addresses “the Unlearned” and states in its preface that the work is meant to be “of singular Use and Benefit to Mankind” — a direct expression of the progress-oriented language the algorithm detects. George Stephenson’s 1831 railway report argues that rail infrastructure would allow Ireland to “reciprocate with England and with other nations, the products of industry,” exemplifying how progress-oriented language pervaded industrial writing by the early 19th century. These examples confirm that the high progress-sentiment scores for industrial volumes at the science-political economy nexus reflect genuine rhetorical content, not measurement artifacts.

Q: What are the paper’s limitations regarding early sample periods?

A: The corpus is thin in earlier eras, particularly around 1550, so results from the earliest decades must be interpreted with caution. The HDL data derive from digitized scans with OCR output of very old books, introducing errors such as the “long-S” misread (e.g., “juftice” for “justice”) that require manual correction. Additionally, the bag-of-words model discards word order, which may obscure some semantic distinctions.

Q: What future research directions do the authors identify?

A: The authors propose applying the same textual analysis techniques to test whether English-language volumes began reflecting greater freedom of expression in the run-up to Britain’s economic takeoff, connecting to the literature on European political fragmentation and the marketplace of ideas. They also suggest applying the approach to corpora in other languages — Dutch (following McCloskey’s argument about bourgeois values) and Spanish (to examine whether the Counter-Reformation and Spain’s economic lag are reflected in cultural attitudes toward progress and science).

LDA (Latent Dirichlet Allocation): An unsupervised generative statistical model that treats each document as a bag of words and extracts latent topics as multinomial distributions over vocabulary; used here to reduce 173,031 volumes to mixtures of 60 topics without imposing prior scholarly interpretations.

Progressive Sentiment Score: The fraction of words in a volume belonging to a seven-word dictionary of progress synonyms (progress, improvement, stride, betterment, advance, rise, amelioration), normalized by total word count; measures the cultural orientation toward the betterment of humankind as embedded in text.

Industrial Score: A volume-level measure equal to the weighted count of industrial root words — derived from the indexes of Appleby’s Illustrated Handbook of Machinery (1877–1903) — normalized by volume length; captures the degree to which a volume’s vocabulary overlaps with industrial production terminology.

Science-Political Economy Nexus: The region of the topic simplex where volumes carry substantial weight in both the science and political economy categories but low weight in religion; the paper finds this is where progress-oriented language was most concentrated from the mid-18th century onward, interpreted as applied science aimed at artisans and a broader literate public.

Industrial Enlightenment: Joel Mokyr’s (2009) concept describing the diffusion of Enlightenment ideals about the practical utility of science into the mechanical and artisanal pursuits that drove Britain’s industrialization; the paper provides quantitative support for this thesis by showing that industrial volumes at the science-political economy nexus were distinctively progress-oriented.

Culture of Growth: Mokyr’s (2016) broader argument that a pan-European network of elite intellectuals fostered a progress-oriented view of science — the idea that scientific understanding could improve the human condition — and that this cultural norm, in combination with Britain’s stock of skilled craftsmen, made industrialization possible.

Bag of Words: A representation of text that records only word frequencies within a document, discarding word order; used here both because HDL copyright restrictions prevent full-text download and because it is the input format required by LDA.

Failing Banks

Mon, 01 Jan 0001 00:00:00 +0000

Correia, Luck, and Verner ask a foundational question in banking: why do banks fail? Specifically, they seek to adjudicate between two theoretical views — the solvency view (failures caused by deteriorating asset quality and insolvency) and the bank runs view (failures caused by depositor coordination failure that can bring down otherwise solvent banks) — using the longest micro-level panel of U.S. commercial bank balance sheets assembled to date.

The authors construct a panel covering approximately 37,000 distinct banks across two samples: a historical sample of all national banks from 1863 to 1941 (sourced from OCC Annual Reports, digitized via OCR) and a modern sample of all commercial banks from 1959 to 2024 (from FFIEC Call Reports merged with the FDIC failure list). More than 5,000 banks fail across the full sample, with 2,887 failures before 1935 and 2,233 after 1959. The sample spans institutional regimes before and after the Federal Reserve (founded 1913) and the FDIC (founded 1933/1934).

Three sets of findings emerge. First, failing banks are characterized by deteriorating fundamentals well before failure: rising non-performing loans and declining solvency (equity-to-assets falls by 8 percentage points in the five years before failure in the modern sample), increasing reliance on expensive noncore funding (rising by 18% of assets in the decade before modern-era failures), and a boom-bust pattern in real assets (expanding by 34% from ten years to three years before failure before contracting). These patterns are consistent across the pre-FDIC and modern eras.

Second, bank failures are highly predictable from publicly available accounting data. Using simple regression models with insolvency risk, noncore funding reliance, and asset growth as predictors, the area under the ROC curve (AUC) for predicting failure within one year reaches 86% in the historical sample and 90–95% in the modern sample. Pseudo-out-of-sample performance is nearly as strong as in-sample performance. A bank in the top 5th percentile of both insolvency risk and noncore funding vulnerability faces a three-year failure probability of 27% in both the historical and modern samples, compared to unconditional rates of 2.5% (historical) and 1% (modern) — a 10- to 25-fold increase.

Third, while large deposit outflows consistent with bank runs were common in pre-FDIC failures — deposits declined on average by 14% immediately before failure in 1880–1934, and by 21% in the period before the banking holiday — failures with runs are as predictable as failures without runs, and they occur in banks with similarly weak fundamentals. Recovery rates on failed banks’ assets averaged only 52% of book value in pre-FDIC failures. Using a framework comparing recovery rates to leverage, the majority of pre-FDIC failed banks appear to have been fundamentally insolvent. Even under the extreme assumption of zero value destruction from failure, runs on banks that were not fundamentally insolvent account for fewer than 8% of pre-FDIC failures; under an assumption of 20% value destruction from failure, this share rises to 22%.

OCC bank examiners classified fewer than 2% of pre-FDIC failures as caused by runs or liquidity issues; most were attributed to losses, fraud, or external shocks. The aggregate failure rate is also largely predictable: regressing the actual bank failure rate on predicted aggregate failure risk yields an R-squared of 40%.

Scope conditions: the historical sample covers only national banks (market share ranging from ~80% in the 1870s to ~45% in the 1930s); the modern sample excludes de novo banks (younger than three years); deposit outflow data for the historical period begin in 1880; and FDIC failure transaction data for the modern period begin in 1993.

Q: What are the two main theoretical views the paper evaluates, and how does the paper distinguish between them? A: The solvency view holds that bank failures are caused by deteriorating asset quality and insolvency, with the runnable nature of liabilities playing no essential causal role. The bank runs view holds that the runnable nature of demandable deposits is central, with depositor coordination failure capable of bringing down otherwise solvent banks (Diamond and Dybvig, 1983) or weak-but-solvent banks (Goldstein and Pauzner, 2005). The paper distinguishes between them using three empirical tests: predictability of failures from fundamentals, deposit outflows before failure, and asset recovery rates in failure.

Q: How predictable are bank failures, and what does predictability imply for the bank runs view? A: In the historical pre-FDIC sample (1863–1934), the in-sample AUC for predicting failure within one year is 86%; in the modern sample (1959–2024) it is 90–95%. Pseudo-out-of-sample AUC is nearly as strong as in-sample AUC. High predictability is consistent with the solvency view and fundamental-based panic run models, but is inconsistent with non-fundamental self-fulfilling runs (Diamond and Dybvig, 1983), which should strike randomly. Predictability also cuts against the assumption of rational, forward-looking depositors in fundamental-run models, since attentive depositors would act on observable signals and accelerate failure, reducing predictability.

Q: What is the boom-bust pattern in failing banks’ assets? A: In the decade before failure, failing banks’ real total assets expand by 34% from ten years to three years before failure, then contract over the final two years. The boom-and-bust pattern is present in both the historical and modern samples but is more pronounced in the modern period. The boom is driven primarily by loan growth (particularly real estate lending and C&I lending in the modern sample) rather than by growth in liquid assets, consistent with the view that rapid credit expansion produces future credit losses.

Q: How does noncore funding behave in failing banks, and why does it matter? A: In failing banks in the modern sample, noncore funding (time deposits plus wholesale funding) rises by 18% of assets over the decade before failure, while demand deposits decline as a share of assets. In the historical sample, noncore (wholesale) funding also rises gradually. Noncore funding is a signal of failure for multiple reasons: it is more expensive than core deposits, eroding profitability; it can finance risky asset growth; it reflects realized losses being funded at the margin; and it increases funding fragility, making banks more vulnerable to shocks.

Q: How strong is the joint signal from insolvency and noncore funding? A: A bank in the top 5th percentile of both insolvency risk and noncore funding vulnerability faces a three-year failure probability of 27% in the historical sample and 27% in the modern sample. The unconditional three-year failure probability is 2.5% in the historical sample and 1% in the modern sample. This amounts to a 10- to 20-fold increase in failure probability, illustrating that the combination of solvency and funding weakness is a powerful joint predictor.

Q: Were deposit outflows common before the FDIC, and did they decline after its introduction? A: In the 1880–1934 historical sample, deposits in failing banks declined on average by 14% between the last call report and failure, with 25% of pre-FDIC failures preceded by outflows exceeding 20%; during the period before the banking holiday the average deposit decline was 21%. In contrast, in the modern sample (1993–2024), average pre-failure deposit outflows were only 2.5%, and outflows exceeding 20% occurred in only 3% of failures, consistent with deposit insurance insulating most depositors.

Q: Are failures with large deposit outflows (runs) less connected to weak fundamentals than other failures? A: No. The paper finds that failures with large deposit outflows are as predictable as failures without large deposit outflows. The relationship between insolvency risk or noncore funding and three-year failure probability is similar for failures with and without large deposit outflows. This implies that runs did not disproportionately strike banks with otherwise strong fundamentals.

Q: What do asset recovery rates reveal about the insolvency status of pre-FDIC failed banks? A: Recovery rates on pre-FDIC failed banks averaged 52% of book value of assets. Under the extreme assumption that receivership destroys zero bank value, runs on non-fundamentally-insolvent (weak but solvent) banks account for fewer than 8% of pre-FDIC failures. Under the equally extreme assumption that failure destroys 20% of bank value, this share rises to 22%. The majority of pre-FDIC failed banks therefore appear to have been fundamentally insolvent.

Q: What did contemporary OCC bank examiners attribute as the causes of bank failures? A: OCC bank examiners classified most pre-FDIC failures as caused by losses, fraud, or external economic shocks. Runs and liquidity issues together account for fewer than 2% of OCC-classified failures, notwithstanding the common occurrence of large deposit outflows before many of these failures. This examiner evidence supports the solvency view.

Q: Can bank-level fundamentals predict systemic banking crises and aggregate failure waves? A: Yes. The authors aggregate out-of-sample predicted failure probabilities to construct a predicted aggregate bank failure rate. The R-squared from regressing the actual aggregate bank failure rate on this predicted rate is 40%, indicating that spikes in bank failures during systemic crises are substantially accounted for by the prior deterioration of bank-level fundamentals.

Q: Why is predictability higher in the modern sample than in the historical sample? A: The authors identify several reasons. Accounting data quality is higher in the modern sample. Historical national banks operated as unit branches with less geographic diversification, making idiosyncratic shocks more important and harder to predict. Modern-era failures are preceded by larger lending booms that produce more predictable downstream losses. Additionally, in the modern context bank failures are largely supervisory decisions, and frictions in the supervisory process may delay closure and thereby increase predictability.

Q: What role do the authors assign to depositor inattention? A: The high predictability of failures combined with the finding that many failing banks had high predicted failure probabilities before actually failing suggests that depositors were often slow to react to observable signals of bank weakness. The authors note this points to behavioral frictions such as neglect of downside risk (Gennaioli et al., 2012) and sleepy or inattentive depositors (Hanson et al., 2015; Jiang et al., 2023), rather than the rational, forward-looking depositor assumption embedded in standard bank run models.

Q: What is the paper’s overall interpretive conclusion about the relative importance of solvency versus runs? A: The primary cause of bank failures is almost always and everywhere a deterioration of bank solvency. Runs were more common in the historical pre-FDIC data as a mechanism triggering failure, but they typically closed banks that were already fundamentally insolvent. Non-fundamental, self-fulfilling runs on otherwise healthy banks appear to be an uncommon cause of bank failures. Under the solvency view, even when runs occur, they are the trigger and final mechanism rather than the root cause.

Insolvency risk: A bank’s proximity to default, proxied in the historical sample by surplus profits relative to equity (capturing profitability and capitalization) and in the modern sample by net income to assets. High insolvency risk reflects declining profitability and eroding capital buffers.

Noncore funding: Expensive, risk-sensitive funding sources outside core demand deposits, including time deposits, wholesale funding (bills payable, rediscounts), and non-deposit wholesale borrowings. Banks relying heavily on noncore funding face higher funding costs, reduced profitability, and greater fragility to funding shocks.

Fundamental run: A run triggered when bank fundamentals are so weak (theta at or below the lower threshold in the Goldstein-Pauzner framework) that all depositors have an incentive to withdraw regardless of others’ actions — the bank is effectively insolvent and failure is inevitable.

Panic-based run: A run triggered when bank fundamentals are moderately weak (below the threshold equilibrium in Goldstein-Pauzner) but the bank would have been able to pay all creditors absent the run; the run itself destroys value and causes failure.

Non-fundamental (self-fulfilling) run: A run on an otherwise solvent bank driven purely by depositor coordination failure, as in Diamond and Dybvig (1983); failure arises from one of two equilibria and is not predicted by fundamentals.

Recovery rate: Funds ultimately collected by the receiver throughout receivership proceedings divided by the book value of assets at suspension; used as a proxy for the degree of fundamental insolvency at failure. Pre-FDIC recovery rates averaged 52% of book value.

Area Under the ROC Curve (AUC): A measure of binary classification performance used to quantify the predictability of bank failures; an uninformative predictor has AUC of 0.5, while AUC of 1.0 indicates perfect classification. In this paper, AUC ranges from 86% (historical, one-year horizon) to 95% (modern).

Boom-bust pattern: The systematic tendency of failing banks to experience rapid loan-driven asset growth in the years preceding failure followed by asset contraction in the final two years before failure — present in both the historical and modern samples, more pronounced in the latter, with real assets expanding by 34% from ten to three years before failure.

Global Working Hours

Mon, 01 Jan 0001 00:00:00 +0000

Drawing on about 5,000 labor force and household surveys from 160 countries that cover 97% of the world’s population, this paper builds a new global database of hours worked and shows that hours worked per adult decline only slightly with GDP per capita and are weakly correlated with economic development overall: the unconditional elasticity of hours with respect to GDP is about -0.04 across countries and -0.01 within countries over time, GDP explains roughly 5% of cross-country and under 1% of within-country historical variation in hours, and the implied reduction is 0-20% over the entire development spectrum. The strong age and gender gradients the authors document are, in their cross-country regressions, driven less by development itself than by institutions: hours worked by the young (aged 15-19) and the elderly (aged 60+) fall with development almost entirely because of rising school attendance and public pension coverage, while prime-age (20-59) hours stay roughly flat but undergo what the authors call a “great gender reshuffling,” in which falling male hours per worker are quantitatively offset by rising female labor force participation. Across countries and over time, labor taxes are strongly negatively correlated with prime-age hours worked; controlling for government transfers only partly reduces this link, which the authors read as ruling out income and substitution effects on labor supply as the only driver, while controlling for working-hours regulations and the size of the formal sector reduces the link much more sharply, suggesting to them that regulation—not just the incentive effects of taxes—plays a large role in shortening intensive-margin hours in richer countries. The authors conclude that collective choices and social norms, often encoded in public policy (schooling, pensions, cultural norms about women’s work, and hours regulation), powerfully shape working hours over and above pure economic development. These are correlational cross-country and time-series patterns rather than identified causal effects, and hours are measured as weekly hours in all GDP-producing jobs (including unpaid agricultural work but excluding unpaid home services).

Summary of a forthcoming paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. What new data does the paper assemble, and how does it improve on prior global hours databases?

The authors mobilize roughly 5,000 nationally representative household and labor force surveys to build a database of hours worked covering 160 countries and 97% of the world population in cross section, plus time series spanning over 20 years in 86 countries. They combine six groups of sources, principally the ILO’s Microdata Repository (about 1,800 surveys in 150 countries since 1990) and the World Bank’s I2D2 database, which include survey data not publicly disclosed by the countries that created them. This extends the most comprehensive prior effort, Bick, Fuchs-Schündeln, and Lagakos (2018), whose core database covered 49 countries (23% of world population) and whose extended database covered 80 countries (41%); large countries such as China and India (35% of world population) that were absent from that study are now included. The authors state they are publishing and plan to regularly update the underlying database at the country×year×age×gender level so that researchers can reproduce their results.

Q2. How seriously does the seasonality concern affect the estimates?

The authors investigate seasonality directly and conclude that monthly seasonality in hours worked is limited in developing countries—actually larger in richer countries because of summer holidays—which gives them confidence that surveys not fielded over the full year still provide reliable annual hours estimates. This matters because Bick, Fuchs-Schündeln, and Lagakos (2018) had restricted their core sample partly out of concern that surveys run in specific months (e.g., around seasonal agricultural work) could bias hours estimates. Resolving this concern is what lets the authors retain the far larger country coverage.

Q3. How much do hours worked actually vary with economic development?

Hours worked per adult slightly decline with GDP but are only weakly correlated with development overall, with an unconditional elasticity of about -0.04 in the cross section and -0.01 in panel data—implying a reduction in hours of 0-20% over the entire development spectrum. GDP explains around 5% of cross-country variation in hours worked and less than 1% of historical within-country variation. Decomposing the margins, employment rates are essentially uncorrelated with development, while hours per worker are bell-shaped: they rise at low levels of development because of structural change (hours in manufacturing and services are very high in middle-income countries, while agricultural hours are moderate and flat with GDP), then flatten. Globally, 59% of the adult population (aged 15+) is employed, working an average of 42 hours per week, which implies about 25 weekly hours per adult; hours are strongly bell-shaped with age, and women supply 35% of GDP-producing hours versus 65% for men, a gap driven mostly by the extensive employment-rate margin.

Q4. Why do hours worked by the young and the elderly fall with development?

In simple cross-country regressions, the decline in hours worked by the young (15-19) and the elderly (60+) as countries develop is entirely driven by rising school attendance for the young and rising public pension coverage for the elderly, in line with a broad body of prior work. In the time series the two margins diverge: the fall in youth work is particularly pronounced, whereas elderly work is stable rather than falling. The authors read this as consistent with developing countries expanding schooling faster, but rolling out elderly pensions more slowly, than frontier economies did historically.

Q5. What happens to prime-age hours, and what is the “great gender reshuffling”?

Prime-age (20-59) hours worked are flat, if not slightly increasing, with GDP per adult, but this stability masks a large compositional shift the authors term a “great gender reshuffling”: female hours rise with development while male hours decline, and the fall in male hours (driven by reduced hours per worker) is quantitatively offset by increases in female employment rates. The authors interpret this as development tending to equalize hours across genders—shortening the long hours of working men while allowing more women into GDP-generating employment. They emphasize considerable heterogeneity across countries and over time in this pattern.

Q6. What role do religion and political history play in female hours worked?

The authors report that Muslim/Hindu religion depresses female hours worked enormously, while former communist status increases them. Grouping countries into former-communist, Muslim/Hindu-majority, and other categories, they show female hours rise with development on average but with large level differences across these groups, which they treat as evidence that cultural and institutional factors—not development alone—shape the gender allocation of work. These are descriptive cross-country associations, not causal estimates.

Labor taxes are strongly negatively related to prime-age hours worked, both in international comparisons and within-country time series; once tax variables are controlled for, GDP per capita is only weakly positively correlated with hours, with an elasticity of around 0.1. The authors probe what drives the tax-hours link. Controlling for social spending (cash or quasi-cash transfers) attenuates it, consistent with income effects from transfers playing some role—but the attenuation is only partial, which the authors read as ruling out income and substitution effects on labor supply as the sole driver. Controlling instead for the share of formal workers and working-hours regulations reduces the link much more sharply. They therefore suggest labor taxes depress hours not mainly through income and substitution effects but rather because high labor taxes correlate with the development of a formal sector with regulated working hours.

Q8. Can a standard labor supply model rationalize these findings?

The authors note that a standard labor supply model with a low uncompensated but large compensated labor supply elasticity can rationalize the joint pattern of weak hours-GDP but strong hours-tax correlations. The logic they invoke from the macroeconomics literature is that economic growth raises the wage rate (an uncompensated labor supply effect, which is weak here) while labor taxes fund transfers (a compensated labor supply effect, which is stronger). The partial attenuation of the tax effect when social spending is controlled is consistent with this account, but the sharper attenuation from regulation and formal-sector controls leads the authors to give regulation a large role alongside—rather than instead of—these labor supply channels.

Q9. What is the paper’s overall interpretation?

The authors conclude that collective choices and public policies—schooling and pension systems, cultural norms regarding women, and regulations on hours worked—have first-order effects on the level and allocation of working hours by age and gender, over and above economic development. They argue that while growth may help develop such institutions, many are only partially determined by it, which is why large cross-country variations in hours worked persist at all levels of development. The paper is framed as documenting and interpreting robust correlations across countries and over time, not as identifying causal policy effects.

Q10. What are the main scope conditions and caveats?

Throughout, hours worked follow international conventions: weekly hours in all jobs that contribute to GDP, including unpaid agricultural work but excluding unpaid home services such as cleaning, cooking, and care. Coverage is 97% of world population, with the missing 3% concentrated in parts of the Middle East and North Africa. The central results on taxes, transfers, regulations, religion, and communist history are correlational—drawn from cross-country regressions and within-country time series—and the authors repeatedly use calibrated language (“correlated,” “suggests,” “consistent with”) rather than claiming identified causal effects.

Key concepts

Hours worked (GDP-producing) : Weekly hours in all jobs that contribute to GDP, following international conventions—this includes unpaid agricultural work (which produces goods counted in GDP) but excludes unpaid home services such as cleaning, cooking, and caring for children or the elderly. Great gender reshuffling : The paper’s term for the pattern in which, as countries develop, declining male hours per worker are quantitatively offset by rising female labor force participation, leaving prime-age (20-59) hours worked roughly stable while its gender composition shifts markedly. Unconditional elasticity of hours with respect to GDP : The raw cross-country (about -0.04) or panel (about -0.01) elasticity of hours worked to GDP per adult before conditioning on taxes, transfers, or institutions; its small size is the paper’s headline evidence that development per se explains little hours variation. Uncompensated vs. compensated labor supply elasticity : In the standard labor supply model the authors invoke, growth raises wages (an uncompensated effect, weak in their data) while labor taxes fund transfers (a compensated effect, stronger in their data); a low uncompensated and large compensated elasticity reconciles weak hours-GDP with strong hours-tax correlations. Formal sector / working-hours regulations : Regulated wage employment in which statutory limits on hours bind; the authors emphasize that the expansion of this regulated formal sector with development, rather than the incentive effects of taxes alone, is the channel that most sharply accounts for shorter intensive-margin hours in richer countries.

Key concepts

Growth Experiences and Trust in Government

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates whether individuals who have experienced stronger GDP growth over their lifetimes are more likely to trust their national government. The authors — Besley, Dann, and Dray — assemble a newly harmonized global dataset comprising approximately 3.3 million respondents across 166 countries since 1990, drawn from 11 major opinion surveys (Afrobarometer, Americasbarometer, Arabarometer, Asiabarometer, European Social Survey, Gallup World Poll, Integrated Values Survey, Latinobarometer, Life in Transition Survey, South Asia Barometer, and World Justice Project). They supplement this with longer-run U.S. evidence from the American National Election Studies (ANES) going back to 1958, covering respondents born as early as the 1880s, and longitudinal Swiss evidence from the Swiss Household Panel (SHP) which allows individual fixed-effects estimation.

The core methodological contribution is the exploitation of country-cohort variation in lifetime GDP growth experiences. Following Malmendier and Nagel (2011), the authors construct a weighted average of past growth realizations across an individual’s lifetime, with weights decaying linearly over time (lambda = 1), so that more recent growth receives greater weight. The baseline specification includes country fixed effects, cohort-by-subcontinent fixed effects, survey-by-survey-year fixed effects, controls for log GDP per capita at year of birth, and individual characteristics (sex, marital status, education, religious denomination). More demanding specifications add country-by-survey-year and country-by-age fixed effects. For Switzerland, individual fixed effects are included, fully absorbing time-invariant personal characteristics.

The main finding is that a one standard deviation increase in lifetime GDP growth experience — corresponding to approximately 2 percentage points of additional growth — is associated with a 2.1 percentage point increase in the probability of trusting the national government, significant at the 1 percent level. This corresponds to roughly 0.042 standard deviations of the trust outcome and approximately 5 percent of the global mean trust in government. The effect is quantitatively meaningful: it approximates between one-quarter and one-half of the difference in average trust between older and younger cohorts in India and Italy, respectively. For the U.S. ANES sample, a one standard deviation increase in growth experience (about 0.2 percentage points) increases trust in the federal government by 2.4 percentage points, explaining more than two-thirds of the average trust gap between Baby Boomers (born 1946–1964) and Millennials (born 1981–1996).

Several scope conditions and heterogeneity findings sharpen the interpretation. First, the growth-trust link is specific to government institutions: there is no statistically significant effect of growth experience on interpersonal trust or trust in religious organizations, indicating the channel runs through perceptions of state performance rather than generalized social capital. Second, a recency heuristic operates: the linearly decaying weighting function (lambda = 1) outperforms both an unweighted lifetime average (lambda = 0) and a formative-years weighting. Growth experienced during formative years (ages 18–25) or before birth has no detectable effect on trust in government; the pre-birth result serves as a placebo test. Third, the positive growth-trust relationship is stronger in democracies than in autocracies, which the authors interpret as democracies producing citizens more responsive to government performance signals. Fourth, a “trust paradox” emerges: unconditionally, average trust in government is lower in democracies than in autocracies, and longer democratic experience is associated with lower trust, which the authors attribute to democratic institutions generating greater citizen skepticism about government performance. Fifth, core results are robust to controlling for other lifetime politico-economic experiences including inflation, banking and currency crises, epidemics, political unrest, executive turnover, stock market returns, and income inequality. The Swiss evidence further shows that private income growth experience does not drive the result — only aggregate macroeconomic growth does.

Q: What is the paper’s core quantitative finding on the growth-trust relationship? A: Using the global harmonized dataset of 3.3 million respondents across 166 countries, a one standard deviation increase in lifetime GDP growth experience (corresponding to approximately 2 percentage points of additional growth) is associated with a 2.1 percentage point increase in the probability of trusting the national government, significant at the 1 percent level. Using only the Gallup World Poll subsample (roughly half the observations), the estimated effect is somewhat larger at 3.6 percentage points per standard deviation increase. These estimates remain statistically significant under more demanding specifications with country-by-survey-year and country-by-age fixed effects, though the magnitudes decrease as these interacted fixed effects absorb variation in recent growth experiences.

Q: How do the authors measure individual lifetime growth experience? A: The growth experience variable is a weighted average of all past annual GDP per capita growth rates since an individual’s birth, with weights that decay linearly over time (lambda = 1 in the Malmendier-Nagel framework). Under this parameterization, the measure simplifies to how much recent economic performance (in the year prior to the survey) exceeds the long-run mean over the respondent’s lifetime, scaled by the respondent’s midpoint of life. This implies younger individuals are more sensitive to recent growth outcomes because their shorter life histories give recent events relatively greater weight. The authors validate this lambda = 1 choice via a grid search over alternative weighting structures using minimum residual sum of squares as the criterion.

Q: How is reverse causality addressed? A: The empirical strategy identifies the relationship using past, cumulative growth experiences measured prior to the survey, so current trust in government cannot cause past growth. Survey-year fixed effects absorb all aggregate time trends simultaneously affecting trust and growth. The authors also conduct a placebo test showing that GDP growth occurring before an individual’s birth has a precisely estimated null effect on their trust in government, which would not be the case if unobserved societal trends were jointly driving both growth histories and political perceptions.

Q: Does growth experience affect interpersonal trust or trust in non-state institutions? A: No. The estimated coefficient on lifetime growth experience is statistically insignificant at conventional levels when interpersonal trust replaces trust in government as the dependent variable, with narrow confidence intervals indicating a precisely estimated null. Similarly, growth experience has no systematic effect on trust in religious organizations such as churches or mosques. The authors interpret these null results as evidence against the alternative explanation that broad modernizing social changes are jointly driving both growth experiences and political trust.

Q: What do the U.S. ANES results add? A: The ANES data, which extends back to 1958 and captures cohorts born as early as the 1880s, provide a within-country test controlling for state fixed effects, generation dummies, and rich individual characteristics including partisan affiliation and partisan strength. A one standard deviation increase in U.S. growth experience (approximately 0.2 percentage points) raises trust in the federal government by 2.4 percentage points, significant at the 1 percent level. This estimate is quantitatively large enough to explain more than two-thirds of the average trust gap between Baby Boomers and Millennials. Results are robust to adding state-by-survey-year fixed effects and birth-state-by-generation fixed effects, and hold for a broader “trust in government index” covering beliefs about waste, corruption, and responsiveness of the federal government.

Q: What do the Swiss Household Panel results contribute? A: The SHP allows individual fixed-effects estimation, exploiting within-person changes in growth experience and trust over time from 1999 onward, which absorbs all time-invariant individual characteristics that could confound the global and U.S. cross-cohort results. The growth experience coefficient remains positive and significant, with a one standard deviation increase yielding a 1.9 percentage point increase in trust in the Swiss federal government (significant at the 1 percent level). The Swiss data also uniquely allow the authors to test whether personal income growth experience drives the result; they find no significant effect of private income growth experience on trust in government, only aggregate macroeconomic growth matters.

Q: Does the recency heuristic hold — does growth in formative years matter? A: No. The authors find no detectable effect of growth experienced specifically during formative years (ages 18–25) on trust in government. Additionally, in a grid-search exercise assessing model fit across different lambda values, the linearly decaying weighting scheme (lambda = 1, giving more weight to recent growth) outperforms both equal-weighted lifetime averages (lambda = 0) and weighting schemes that emphasize earlier life experiences (lambda less than 0). The pre-birth placebo result (null effect) and the absence of a formative-years effect together indicate that the operative mechanism is about evaluating current government performance based on recent macroeconomic experience, not the imprinting of long-lasting political dispositions during youth.

Q: What is the “trust paradox” and how is it documented? A: The trust paradox refers to the empirical finding that average trust in government is lower in democracies than in autocracies at the cross-country level, and that longer experience with democratic institutions within countries is associated with lower levels of trust in government in the micro data. This is counterintuitive given the standard view that good institutions should foster confidence in government. The authors suggest the paradox likely reflects democracies cultivating greater citizen skepticism and more critical judgment of government performance, rather than indicating that democratic governance actually performs worse. Importantly, the positive effect of growth experience on trust remains present in democracies, and the growth-trust relationship is actually stronger in democratic regimes, consistent with citizens in democracies being more responsive to government performance signals.

Q: How is the growth-trust finding related to corruption perceptions and living standards? A: Using the Gallup World Poll, the authors find that stronger lifetime growth experience is associated with lower perceived corruption in government, greater satisfaction with personal living standards, and higher likelihood of feeling one lives comfortably on one’s present income. These results are consistent with citizens attributing economic success to government competence and integrity, and with growth translating into perceptions of improved personal circumstances through both direct income effects and indirect public goods provision.

Q: Are the results robust to controlling for other lifetime politico-economic experiences? A: Yes. When the authors include lifetime experience measures for political unrest, executive turnover, epidemic exposure, banking crises, currency crises, and inflation (both levels and volatility) simultaneously in equation (3), the growth experience coefficient remains consistently positive, stable, and significant across all specifications. Among the other experience variables, only lifetime unrest and epidemic exposure are independently negative and statistically significant at conventional levels. F-tests reject the null hypothesis that the crisis and growth experience coefficients are equal in magnitude. The U.S. results are also robust to adding lifetime experiences with S&P 500 returns, unemployment, and top-income-share inequality measures.

Q: What are the policy implications of the findings? A: The authors note that sustained economic growth may itself be a mechanism for building political trust, with positive downstream effects for policy compliance — a connection they document has been relevant during the COVID-19 pandemic (where higher-trust societies showed lower mobility during lockdowns and higher vaccine acceptance). The growth-trust channel could have implications for increasing compliance across a range of policy domains including climate action and tax morale. Governments that deliver sustained economic growth can expect citizens to update their trust upward, particularly in democracies where citizens are more performance-responsive, while governments that preside over stagnation or contraction face predictable erosion of political legitimacy across cohorts.

Growth experience: A weighted average of all past annual GDP per capita growth realizations since an individual’s birth, with weights that decay linearly over time following Malmendier and Nagel (2011), so that more recent growth receives greater weight. Under the paper’s preferred parameterization (lambda = 1), the measure equals how much last year’s GDP per capita exceeds the respondent’s lifetime mean, scaled by the respondent’s midpoint of life.

Trust in government: A binary dummy variable equal to one if a survey respondent expresses “a great deal” or “quite a lot” of trust or confidence in the national government, constructed from harmonized responses across 11 major opinion surveys. The paper treats this as reflecting respondents’ perceptions of government performance rather than a deep interpersonal trust relationship.

Trust paradox: The empirical regularity documented in the paper whereby average trust in government is unconditionally lower in democracies than in autocracies at the cross-country level, and whereby longer democratic experience within countries is associated with lower individual trust in government. The authors attribute this to democratic institutions generating more critical citizen judgment of government performance.

Recency heuristic: The finding that more recent growth experiences carry greater weight in forming trust in government, as captured by the linear decay weighting scheme (lambda = 1) outperforming equal-weighted or early-life-weighted alternatives. Growth before birth and growth during formative years (ages 18–25) have no detectable effect, while recent macroeconomic performance is the operative signal.

Cohort-level variation: The within-country differences in lifetime growth experiences across birth cohorts that form the paper’s primary identification strategy. Because different cohorts in the same country have lived through different sequences of growth episodes, differences in trust across cohorts within a country can be attributed to differential growth exposure rather than time-invariant country characteristics.

Formative years effect: The hypothesis, tested and rejected in the paper, that economic experiences during ages 18–25 have a lasting imprint on political attitudes analogous to formative-years effects found in other political behavior literatures. The paper finds no statistically significant association between growth experienced during these years and trust in government.

Source text origin: In the pipeline context relevant to this paper’s acquisition, this refers to whether a summary was generated from full working paper text (“pdf” or “oa-html”) versus abstract only (which is hard-blocked). The working paper was obtained from LSE Research Online (eprint 129614), classified as published version under CC BY 4.0.

How Do You Identify a Good Manager?

Mon, 01 Jan 0001 00:00:00 +0000

This paper develops a novel experimental method to identify the causal contribution of managers to team performance, and uses it to evaluate which characteristics predict managerial effectiveness and how manager selection mechanisms affect organizational outcomes.

The core identification challenge is that managers are not randomly assigned to teams in the field, and field managers are a highly non-random sample, making it difficult to infer which traits genuinely predict managerial performance. The authors address this by repeatedly randomly assigning managers to multiple teams in a controlled laboratory experiment, then estimating each manager’s average causal contribution to group output after conditioning on group members’ individual productive skills. The intuition is that a good manager is someone who consistently causes their team to produce more than the sum of their parts.

The experiment was conducted at the University of Essex lab with 555 participants (46% female, mean age 25, ethnically diverse) forming 728 groups of three across four rounds. Each group consisted of one manager and two workers who performed a Collaborative Production Task requiring coordination across three problem-solving modules (numerical, spatial, and analytical reasoning). The team score was the minimum module score — a weakest-link structure making coordination essential. Prior to group testing, all participants completed individual assessments of task-specific skill, fluid intelligence (CFIT), emotional perceptiveness (Reading the Mind in the Eyes Test, RMET), economic decision-making skill (the Assignment Game, which measures resource allocation under comparative advantage), Big 5 personality, and demographic characteristics. Manager selection was randomly varied at the session level: in 20 sessions, the participant with the strongest preference for leadership became manager (self-promotion); in 19 sessions, managers were assigned by lottery.

The main quantitative findings are as follows. First, there are large, stable, and statistically significant manager effects: a manager one standard deviation above average improves team performance by approximately 0.23 standard deviations (p = 0.04). This estimate is roughly 90% the size of the combined productive skill coefficient for the two workers (approximately 0.26 sd), indicating that a good manager is roughly twice as valuable as a good individual worker. Manager contributions predict out-of-sample group performance in a leave-one-out procedure (p < 0.01).

Second, among randomly assigned managers, only two predictors significantly explain managerial performance: fluid intelligence (CFIT) and economic decision-making skill (Assignment Game scores), both significant at below the 1% level. Gender, age, and ethnicity do not predict managerial performance.

Third, self-promoted managers perform substantially worse than lottery-assigned managers, by approximately 0.10 standard deviations — roughly equivalent to being assigned a manager with fluid intelligence one full standard deviation below average. The mechanism is overconfidence: people who strongly prefer management roles are significantly more overconfident (d = 0.41 sd, p < 0.01) and exhibit a strong negative correlation between self-reported social skills and actual emotional perceptiveness on the RMET (r = -0.37, p < 0.001). Among self-promoted managers, self-reported extraversion and political skill are negatively correlated with managerial performance (rho = -0.24 and -0.26, p < 0.05); no such negative relationship appears among lottery managers.

Fourth, selecting managers on economic decision-making skill rather than self-promotion improves average manager quality by 0.6 standard deviations — equivalent to replacing an average worker in every group with a worker at the 99th percentile of individual productivity.

The three mechanisms through which good managers improve performance are: (1) monitoring — good managers (1 sd above average) cut monitoring errors from 16% to 8%; (2) optimal task allocation according to comparative advantage — groups with optimally assigned workers score 0.52 sd higher (p < 0.01); (3) worker motivation in late-stage effort — teams led by a 1-sd-above-average manager solve 0.6 more problems in the final two minutes versus only 0.3 more in the first two minutes.

The experiment was conducted in a university lab in the UK, and the sample skews toward graduate students with limited work experience. Generalizability to field settings is supported by prior evidence that peer productivity spillover experiments yield similar magnitudes in lab versus field settings, and that the estimated manager effects are similar to Lazear et al. (2015) estimates from a large employer dataset.

Q: What is the core methodological innovation of this paper? A: The paper requires repeated random assignment of managers to multiple teams, combined with controls for individual productive skill measured prior to group work. This allows identification of each manager’s average causal contribution to group output, rather than confounding management quality with team composition or individual worker ability. The key estimand is the standard deviation of individual manager effects (sigma_alpha), interpreted as the impact of having a manager one standard deviation above average.

Q: How large is the estimated manager effect, and how does it compare to worker effects? A: A manager one standard deviation above average improves team performance by approximately 0.23 standard deviations (p = 0.04 by randomization inference). This is roughly 90% the size of the combined productive skill effect of both workers together (approximately 0.26 sd), implying a good manager is nearly twice as valuable as a good individual worker. Without conditioning on production skills, the manager effect rises to 0.29 sd.

Q: What characteristics predict managerial performance among randomly assigned managers? A: Only two measures predict managerial performance in the lottery arm: fluid intelligence (CFIT) and economic decision-making skill (scores on the Assignment Game), both significant at below the 1% level. These predictors are robust to controls for demographics, education, work experience, emotional perceptiveness, and personality traits. Gender, age, and ethnicity do not predict managerial performance.

Q: What is the “Assignment Game” and why is it a strong predictor? A: The Assignment Game (Caplin et al., 2024) places participants in a simulated managerial role where they must assign fictional workers to tasks. Performing well requires understanding comparative advantage intuitively, managing an attentionally demanding numerical environment, and avoiding biases such as anchoring. The paper argues its strong predictive power reflects that good managers excel at allocating workers according to comparative advantage — which the experiment directly identifies as a key mechanism.

Q: How do self-promoted managers perform relative to lottery-assigned managers? A: Self-promoted managers perform approximately 0.10 standard deviations below lottery managers, and this gap is robust across model specifications. The performance deficit is roughly equivalent to being assigned a manager whose fluid intelligence is one full standard deviation below average. This finding implies that common organizational practice of selecting managers partly via self-nomination actively reduces team productivity.

Q: Why do self-promoted managers underperform? A: The paper attributes underperformance primarily to overconfidence. People strongly preferring management roles are significantly more overconfident than those without strong preferences (d = 0.41 sd, p < 0.01). Self-promoted managers specifically overestimate their social skills: among them, self-reported people skills are strongly negatively correlated with actual emotional perceptiveness on the RMET (r = -0.37, p < 0.001), and self-reported extraversion and political skill are negatively correlated with managerial performance (rho = -0.24 and -0.26, p < 0.05). None of these negative relationships appear among lottery managers.

Q: Who wants to be a manager, and does it differ by gender? A: The three variables most strongly correlated with wanting to be in charge are extraversion, risk appetite, and being male. The relationship between high extraversion and preference for management is driven largely by men. Women are much less likely to nominate themselves for leadership roles despite being equally or more effective on average — a finding consistent with broader experimental evidence on gender and leadership self-selection.

Q: How large are the potential gains from skill-based manager selection? A: Compared to self-promotion, selecting managers based on economic decision-making skill yields managers who are 0.6 standard deviations better in terms of estimated manager effects. In terms of group performance, this is equivalent to replacing an average worker in every group with a worker at the 99th percentile of individual productivity. Selecting on both economic decision-making and fluid intelligence outperforms random assignment, selection on social skills, or selection on worker task performance (the Peter Principle).

Q: What are the three mechanisms through which good managers improve team performance? A: First, monitoring: good managers (1 sd above average) reduce monitoring errors — defined as having a worker on a module substantially above the minimum score at task end — from 16% to 8% (bivariate correlation with manager performance = -0.40, p < 0.001). Second, optimal task allocation: the probability of finding the optimal comparative-advantage-based assignment is positively associated with manager performance (rho = 0.19, p < 0.01), and groups with always-optimal starting assignments score 0.52 sd higher than those with never-optimal assignments (p < 0.01). Third, worker motivation: team performance in the final two-minute period is about 50% more influential for overall outcomes than the first two minutes (p = 0.038), and 1-sd-above-average managers generate 0.6 more problems solved in the final period versus 0.3 in the first, consistent with differential motivational effects emerging over time.

Q: What is the Peter Principle, and how does this paper relate to it? A: The Peter Principle refers to the practice of promoting employees based on their performance as line workers rather than their suitability for management — promoting individuals to their level of incompetence. Benson et al. (2019) document this selection pattern empirically. This paper shows that selecting managers on worker task skill is inferior to selecting on economic decision-making skill or fluid intelligence, confirming that task skill is not the right criterion for manager selection even if it predicts individual worker output.

Q: How does the paper validate that manager effects are real and not noise? A: The paper uses randomization inference with 5,000 simulated allocations to compute p-values, obtaining p = 0.04 for the main manager effect. Robustness checks include controlling for pre-existing social relationships, manager risk appetite, variance of individual scores, and granular skill measures — all yielding estimates near 0.22 sd. A leave-one-out out-of-sample prediction test confirms manager contributions significantly predict held-out group performance (p < 0.01), while the analogous worker out-of-sample estimate is less than half the magnitude and not statistically significant.

Q: What are the scope conditions on the experimental results? A: The experiment is conducted in a university lab in the UK with graduate students averaging 25 years of age and two years of work experience, limiting direct generalizability to experienced workers or senior management. The task lasts approximately 15 minutes, which may not capture longer-run managerial dynamics. Compensation equalized average earnings between managers and workers, which differs from most real-world settings. The authors note their effect-size estimates closely match Lazear et al. (2015) from a large employer, and that Herbst and Mas (2015) find lab peer-productivity experiments generalize to the field.

Manager Effect (sigma_alpha): The standard deviation of individual managers’ average causal contributions to group performance, estimated via repeated random assignment and conditioning on individual productive skill. Represents the impact of having a manager one standard deviation above average, estimated at approximately 0.23 standard deviations of group output.

Collaborative Production Task: A novel lab group task in which a manager and two workers solve problems across three modules (numerical, spatial, analytical reasoning), with team score defined as the minimum module score (weakest-link structure). Managers are responsible for worker assignment, monitoring, and motivation; workers face no financial performance incentives.

Economic Decision-Making Skill: Defined by Caplin et al. (2024) as the ability to make good resource allocation decisions, assessed via the Assignment Game in which participants must optimally assign workers to tasks under comparative advantage. The single strongest predictor of managerial performance in the lottery arm.

Monitoring Failure: Defined in the paper as having any group member working on a module at task end whose score is substantially greater (e.g., 10 points higher) than the minimum module score — meaning the worker’s effort is not contributing to the group score. Occurs in 16% of groups overall; managers one sd above average reduce this to 8%.

Self-Promotion (as selection mechanism): A treatment condition in which the participant with the strongest stated preference for being manager (on a 1-10 scale) is assigned the managerial role. Contrasted with lottery assignment; self-promoted managers perform approximately 0.10 sd worse than lottery managers.

Overconfidence (in managerial context): The gap between self-assessed skill (particularly social/interpersonal skill) and objectively measured skill (e.g., RMET score). Self-promoters are significantly more overconfident (d = 0.41 sd), and overconfidence is strongly negatively correlated with actual emotional perceptiveness (r = -0.33, p < 0.001).

Comparative Advantage Allocation: The practice of assigning each worker to the module in which they have the highest relative (not absolute) performance advantage. Captured via whether a manager selects the optimal one-to-one assignment given pre-measured individual module scores; groups with always-optimal allocation score 0.52 sd higher.

Ideas Have Consequences: The Impact of Law and Economics on American Justice

Mon, 01 Jan 0001 00:00:00 +0000

This paper quantifies the effect of the Manne Economics Institute for Federal Judges — an intensive two-week economics training program run by the Law and Economics Center from 1976 to 1998 — on the decision-making of U.S. federal judges. The research question is whether exposure to a coherent set of economic ideas can directly shift the policy decisions of sitting policymakers, as distinct from effects operating through partisan affiliation or formal legal rules.

The program trained nearly half of all federal judges over its two decades of operation. By 1990, forty percent of federal judges had attended; by the late 1990s, roughly half of circuit court cases had a Manne-trained judge on the panel. Instructors included Milton Friedman, Armen Alchian, Harold Demsetz, Martin Feldstein, Paul Samuelson, and Orley Ashenfelter, covering supply-and-demand theory, the Coase Theorem, externalities, property rights, and criminal deterrence following Becker (1968). The program was funded by pro-business foundations and had a recognized conservative-leaning orientation, though it invited both Republican- and Democrat-appointed judges and was popular across party lines.

The identification strategy is a differences-in-differences design exploiting staggered attendance timing. Because the program was oversubscribed and admitted judges on a first-come-first-served basis — with applicants bumped to later cohorts when capacity was reached — the timing of attendance within the ever-attending population has a quasi-random component. The preferred control group consists exclusively of other ever-attending judges who had not yet attended, rather than never-attenders, because never-attenders differ systematically on observables and show a pre-existing positive trend in economics language use, likely from ambient diffusion through clerks, law schools, and organizations such as the Federalist Society. Judge fixed effects and circuit-by-year (or courthouse-by-year) fixed effects absorb time-invariant judge characteristics and court-level time trends. Elastic-net-selected covariates predicting attendance timing, fully interacted with year fixed effects, are added as robustness controls. Standard errors are clustered by judge.

The data cover approximately 200,000 published circuit court opinions (1970–2005) from Bloomberg Law, a 5% random sample of circuit cases hand-coded for ideological direction from the Songer-Auburn database, machine-coded regulatory agency outcomes, a newly collected antitrust case dataset, and approximately 1.03 million district court criminal sentencing records (1992–2003) from TRAC.

The main findings are as follows. First, after attending the Manne program, judges increase their use of economics language in written opinions by approximately one-third of a standard deviation, measured via word-embedding similarity to an economics lexicon; this effect is statistically significant in the short-run event-study window but does not persist over the full career. Second, Manne attendance raises conservative voting in economics-related cases (labor and regulation) by approximately one-quarter of a standard deviation — corresponding to judges deciding in the conservative direction about 20 percent more often relative to the mean — with no significant effect on non-economics cases; the interaction effect is robust across specifications including never-attenders. Third, post-Manne judges vote more frequently against federal labor and environmental regulatory agencies, a result that is statistically significant and economically meaningful with no detectable pre-trends. Fourth, post-Manne judges impose longer and more frequent prison sentences, with no increase in sentencing harshness for drug crimes — consistent with Manne instructors having explicitly advocated drug legalization — and with the harshness gap between Manne and non-Manne judges widening after the 2005 Booker decision expanded judicial sentencing discretion. Fifth, there is some evidence of increased voting against antitrust enforcement, though this result is more sensitive to specification. Persuasion rates computed following DellaVigna and Gentzkow (2010) are slightly larger than those estimated for partisan media interventions such as Fox News and are closest to the effect of a 10-week Washington Post subscription on Democratic governor vote share. Neither the legalist model (judges follow statutes mechanically) nor the attitudinal model (judges follow party affiliation) can explain these within-judge, within-party shifts.

Q: What is the central identification challenge and how do the authors address it? A: The key threat is that judges who chose to attend the Manne program — or who attended at a particular time — may differ systematically from non-attenders in ways correlated with their decision trajectories. The authors address this in two steps. First, they restrict the control group to other ever-attending judges who had not yet attended, exploiting the first-come-first-served oversubscription rule that created quasi-random variation in timing among applicants. Second, they use judge fixed effects plus circuit-by-year fixed effects, and add elastic-net-selected biographical covariates (e.g., birth cohort indicators) interacted with year fixed effects as a robustness check. Republican affiliation — the most salient ideological predictor of attendance — is not a statistically significant predictor of attendance timing, supporting the exclusion restriction.

Q: Why are never-attenders excluded from the preferred control group? A: Never-attenders differ from attenders on observables including political party and show a positively trending use of economics language in their opinions even before any treatment, suggesting ambient diffusion of economics ideas through law clerks, law school curricula, and organizations such as the Federalist Society. Including never-attenders in the control group produces a near-zero coefficient on the language outcome, which the authors interpret as reflecting spillovers rather than a true null effect; the coefficient on conservative voting in the interaction specification, however, remains positive and significant even when never-attenders are included.

Q: What is the magnitude of the effect on economics language use? A: The within-judge effect of Manne attendance on the word-embedding similarity between judicial opinions and an economics lexicon is approximately one-third of a standard deviation, statistically significant in the short-run event-study window (covering six years before and after attendance). The effect shrinks and becomes non-significant when the full career of Manne judges is examined (rather than just the event-study window), consistent with broad diffusion of economics language across the judiciary over time rather than a persistent individual-level treatment effect.

Q: How large is the effect on conservative voting, and is it concentrated in particular case types? A: Post-Manne attendance raises conservative voting in economics-related cases (labor and regulation) by approximately one-quarter of a standard deviation, corresponding to judges deciding in the conservative direction about 20 percent more often relative to the mean liberal-conservative decision rate. There is no statistically significant effect on non-economics cases. The interaction coefficient — the differential effect on economics versus non-economics cases — is positive and significant across all specifications including the full sample with never-attenders, making this the most robust directional result in the paper.

Q: What is the effect on regulatory agency voting? A: Post-Manne judges vote more frequently against federal labor agencies (National Labor Relations Board, OSHA, Department of Labor, Federal Labor Relations Authority, Office of Worker’s Compensation Programs) and the Environmental Protection Agency. The event study shows a positive and significant increase that persists across the event-study window with no detectable pre-trends. This result is robust to both the baseline specification and the elastic-net-controls specification.

Q: What is the effect on criminal sentencing, and what heterogeneity is found? A: Post-Manne judges impose both more frequent prison sentences and longer sentences, consistent with Becker’s deterrence framework taught in the program’s criminal law curriculum. The sentencing effects are absent for drug crimes, consistent with Manne instructors — including Milton Friedman — having explicitly advocated against the drug war and for drug legalization. The gap in sentencing harshness between Manne and non-Manne judges widens after the 2005 United States v. Booker decision, which made the Federal Sentencing Guidelines advisory rather than mandatory; this is consistent with the program having shaped latent judicial preferences that are expressed more fully when formal constraints are relaxed.

Q: How do the persuasion rates compare to benchmark media studies? A: The persuasion rates computed following DellaVigna and Gentzkow (2010) are slightly larger than those estimated for partisan media interventions such as Fox News (DellaVigna and Kaplan, 2007) and are closest to the persuasion rates implied by a 10-week subscription to the Washington Post on Democratic governor vote share (Gerber et al. 2009). The comparison contextualizes the Manne program as a moderately high-intensity ideational intervention relative to documented cases of political persuasion.

Q: What do the results imply for theories of judicial behavior? A: The findings are inconsistent with both the legalist/formalist model — under which judges apply statutes and precedent without regard to extra-legal factors, predicting zero effect — and the attitudinal model — under which judges simply follow partisan preferences, also predicting zero effect since the program attended judges of both parties. The within-judge, within-party shifts point to a third channel: judicial worldviews and economic ideas, independent of formal law and partisan affiliation, shape high-stakes precedent-setting decisions.

Q: Can the authors distinguish between a pedagogical (informational) and an ideological persuasion mechanism? A: They cannot definitively distinguish between the two. Both mechanisms predict increased economics language, more conservative rulings in economics cases, deregulatory voting, and harsher non-drug sentences. The drug-crime heterogeneity is somewhat more consistent with a nuanced pedagogical channel, since Manne instructors explicitly discussed drug legalization, but this pattern is also consistent with complex ideological effects. Evidence on decision quality (citation rates, judicial promotion) is mixed and not robust, providing no clean test of the informational mechanism.

Q: What does the antitrust evidence show? A: Post-Manne judges tend to vote against antitrust claimants (i.e., in favor of less antitrust enforcement), but this result is more sensitive to specification than the regulatory agency and sentencing results and is not always statistically significant across specifications. The authors treat it as suggestive rather than conclusive.

Q: How does this paper relate to the literature on economics education and normative beliefs? A: Prior work finds that economics students are less redistributive (Selten and Ockenfels 1998), view surge prices more favorably (Frey and Meier 2005), favor profit maximization (Rubinstein 2006), and that economics professors are less ideologically liberal than other social scientists (Jelveh et al. 2018). The present paper extends this literature by studying established professionals (judges) making high-stakes real-world decisions, and by documenting a direct policy impact rather than a change in survey responses or experimental choices.

Q: What is the scope of the dataset and the program coverage? A: The circuit court dataset covers approximately 200,000 published opinions from 1970 through 2005. The district court sentencing dataset covers approximately 1.03 million cases from 1992 through 2003 (event study sample). The Manne program ran from 1976 to 1998, with roughly twenty judges per cohort; by 1990 forty percent of federal judges had attended, and by the late 1990s roughly half of circuit court cases had a Manne-trained panelist. Biographical information comes from the Federal Judicial Center; program attendance lists come from Butler (1999) supplemented by FOIA-obtained annual reports.

Manne Economics Institute for Federal Judges: An intensive two-week economics training program for sitting U.S. federal judges, run by the Law and Economics Center from 1976 to 1998, covering supply-and-demand theory, the Coase Theorem, externalities, property rights, deterrence theory, and related topics; funded by pro-business foundations; admitted judges on a first-come-first-served basis and trained nearly half of all federal judges over its operation.

Word-embedding economics language measure: A continuous measure of how closely a judicial opinion’s vocabulary aligns with a lexicon of law-and-economics phrases, constructed using word2vec embeddings (Mikolov et al. 2013) trained on the corpus of judicial opinions; measures the semantic proximity of opinion text to the Ellickson (2000) economics lexicon in embedding space, capturing implicit and contextual use of economics reasoning rather than raw phrase counts.

Deterrence theory (Becker model): The framework, drawn from Becker (1968), taught in the Manne program’s criminal law curriculum, which holds that optimal crime deterrence requires setting the expected penalty — the economic cost of punishment times the probability of detection — high enough to outweigh the expected benefits of crime; treated in the paper as the theoretical basis for predicting harsher sentencing among post-Manne judges, and contrasted with retribution- or rehabilitation-based sentencing rationales that dominated before its diffusion.

Conservative judicial decision (economics cases): In the paper’s usage, a ruling against the liberal/pro-plaintiff position in a case involving labor or regulation, as hand-coded by the Songer-Auburn database; includes ruling against a labor agency, rejecting a regulatory claimant, or voting against antitrust enforcement; the paper finds Manne attendance shifts judges in this direction in economics cases but not in non-economics cases.

First-come-first-served oversubscription: The admission rule of the Manne program during its oversubscribed heyday (from the second cohort in 1977 through the late 1980s), under which applicants who did not secure a spot were bumped to the next year’s cohort; the authors argue this rule generates quasi-random variation in the timing of attendance among ever-attending judges, conditional on applying, providing the identifying variation for the differences-in-differences design.

Persuasion rate: A summary statistic, following DellaVigna and Gentzkow (2010), measuring the fraction of the “persuadable” population that is convinced by a treatment; used in the paper to benchmark the Manne program’s effect size against documented media persuasion interventions such as Fox News and Washington Post subscriptions.

Insuring Peace: Index-Based Livestock Insurance, Droughts, and Conflict

Mon, 01 Jan 0001 00:00:00 +0000

This paper provides quasi-experimental evidence that Index-Based Livestock Insurance (IBLI) — a remote-sensing-triggered, automated payout scheme for pastoralists — substantially reduces drought-induced conflict in Kenya over the 2001–2020 period.

The research question is whether a market-based financial instrument can mitigate the causal chain running from drought shocks to violent conflict between nomadic pastoralists and sedentary farmers and other land users. The authors motivate the study by documenting that droughts force pastoralists out of their traditional grazing grounds and into mixed-land-use areas (farms, ranches, urban settlements, nature reserves), where miscoordination with other land users escalates into violence. A case study of the Samburu-Laikipia-Isiolo-Meru region in central Kenya — drawing on georeferenced survey data from Lengoiboni et al. (2010) and ACLED conflict events — validates this spatial mechanism: during droughts, roughly 60–90% of non-pastoral land users report encounters with pastoralists, and conflicts accumulate precisely where drought migration routes cross into non-pastoral land.

The empirical design combines two sources of variation: (1) plausibly exogenous changes in rainfall deficits at the 0.1 × 0.1-degree grid-cell level (roughly 10 × 10 km), derived from NASA GPM satellite data; and (2) the staggered, five-wave rollout of IBLI across 146 insurance districts in Kenya from 2010 onward, which the authors argue was driven primarily by technical challenges rather than pre-existing conflict or drought patterns. The unit of observation is 94,300 cell-periods. Because conflicts due to pastoralist drought migration occur in the neighborhood of affected areas rather than within them, both drought and IBLI coverage are measured as inverse-distance-weighted averages over surrounding cells. The estimating equation is a linear probability model with cell and period fixed effects, interacting neighborhood rainfall deficit with neighborhood IBLI coverage; the coefficient on this interaction term (delta3) is the parameter of interest.

The main finding is that a one-standard-deviation increase in neighborhood IBLI coverage reduces the semi-elasticity of neighborhood rainfall deficit on conflict probability by approximately 23%. In absolute terms, a one-percentage-point increase in the rainfall deficit raises the probability of conflict by 6.92 percentage points at average IBLI coverage; with one additional standard deviation of neighborhood IBLI, that same deficit raises conflict probability by only 5.34 percentage points — a reduction of 1.58 percentage points against a baseline conflict probability of roughly 2.5%.

Scope conditions: the effect is estimated for Kenya specifically, over a pastoralist-heavy population of approximately 8.8 million out of 53 million Kenyans, during 2001–2020. The conflict-mitigating effect is approximately four times larger in mixed-land-use areas (nine times when rollout-cluster-times-period fixed effects are included), consistent with the theoretical expectation that IBLI matters most where pastoralists are most likely to encounter other land users during drought migration.

Two mechanisms are identified. First, IBLI reduces migratory pressure: when pastoral homelands have IBLI coverage, the distance between the ethnic homeland centroid and conflict events involving that group decreases, indicating reduced drought migration. Second, IBLI smooths incomes — corroborated with Afrobarometer geo-coded data — raising the opportunity cost of fighting. An instrumental-variable specification finds that actual IBLI payouts in the neighborhood reduce conflict probability by approximately 150% relative to the baseline risk.

A cost-effectiveness analysis finds that even using conservative World Health Organization or World Bank estimates of the value of statistical life, IBLI delivers fatality savings of between 10 and 22 cents per dollar spent on government subsidies for the program, making it a cost-effective complement to political and institutional conflict-mitigation approaches.

Q: What is the core causal mechanism linking droughts to conflict that IBLI interrupts?

A: Droughts deplete forage in pastoralists’ traditional grazing grounds, forcing them to migrate into mixed-land-use areas — farms, ranches, urban settlements, and nature reserves — where encounters with other land users are more likely to escalate into violence. Without insurance, pastoralists hold excess livestock as precautionary savings, amplifying the extent of necessary migration during dry periods. IBLI payouts allow pastoralists to purchase forage locally, reducing migration distance and intensity, and also smooth income, raising the opportunity cost of engaging in violence.

Q: How does IBLI work technically, and why does it overcome problems of traditional livestock insurance?

A: IBLI uses satellite remote sensing to calculate whether a district-specific drought threshold has been crossed; if so, automated payments are triggered immediately without requiring direct loss assessment or field inspections. This design eliminates moral hazard and adverse selection problems inherent in traditional indemnity insurance, reduces monitoring costs, and enables fast delivery via mobile payment platforms such as MPESA even to remote households. The Kenyan government rebranded the program as the Kenyan Livestock Insurance Program (KLIP) in 2015 and fully subsidizes coverage for up to five tropical livestock units per household.

Q: What is the magnitude of the main conflict-mitigation result?

A: A one-standard-deviation increase in neighborhood IBLI coverage reduces the semi-elasticity of the neighborhood rainfall deficit on conflict probability by approximately 23% (delta3/delta1 = -0.0158/0.0692). In absolute terms, this translates to a reduction from a 6.92 percentage-point increase in conflict probability per one-percentage-point rainfall deficit to a 5.34 percentage-point increase — a decline of 1.58 percentage points against a mean conflict probability of roughly 2.5%.

Q: Why do the authors use a neighborhood rather than cell-level treatment measure?

A: Drought-induced pastoralist conflicts occur primarily not in the pastoral home areas themselves but in neighboring regions where drought migration routes cross into non-pastoral land. The case study documents this pattern directly: ACLED conflict events accumulate where migration routes from Namelok, Lodungokwe, and Ngaremara communities intersect urban or agricultural areas, not within the pastoral zones. The neighborhood approach, using inverse-distance-weighted averages, captures both the probability of migration from surrounding cells and the declining probability of migration with distance.

Q: What is the main identification concern and how do the authors address it?

A: The main concern is that the timing of the IBLI rollout is endogenously determined — areas with a higher latent drought-conflict elasticity might receive coverage earlier or later, biasing the interaction coefficient. The authors show that the pre-treatment drought-conflict elasticity has no systematic correlation with either IBLI eligibility or the timing of coverage receipt. Placebo tests interacting the neighborhood rainfall deficit with pre-treatment eligibility or eventual coverage indicators yield positive, statistically insignificant coefficients, suggesting any bias would run in the direction of underestimating the mitigation effect. A permutation test randomly reassigning IBLI coverage across the six rollout clusters finds the actual point estimate is in the bottom 2.2% of the simulated distribution, indicating it is unlikely to arise from cluster-level confounders.

Q: How do the authors rule out that other programs — cash transfers or development aid — explain the result?

A: The authors control for cell-level and neighborhood-level coverage of Kenya’s Hunger Safety Net Programme (HSNP), which provides unconditional cash transfers to vulnerable households and covers most IBLI-eligible areas, as well as for World Bank agricultural aid projects. Across these specifications, the estimated conflict mitigation ranges from -19.16% to -42.24%, with the baseline estimate of -22.79% remaining robust, indicating neither HSNP nor development aid is a plausible alternative explanation.

Q: What is the alternative identification strategy using within-rollout-cluster variation?

A: The authors exploit pre-determined (1984 government land-use map) variation in mixed-land-use status across cells within the same IBLI rollout cluster-period, including rollout-cluster-times-period fixed effects that absorb any omitted variable related to the potentially endogenous rollout steps. The conflict-mitigating effect of IBLI is approximately four times larger in mixed-land-use cells, and approximately nine times larger in the most restrictive specification with rollout-cluster-times-period fixed effects, consistent with the prediction that IBLI matters most where pastoralists encounter other land users.

Q: How do the authors establish the migratory pressure mechanism?

A: Following Eberle et al. (2023), the authors match conflict actors to ethnic homelands using Murdock (1967) boundaries and test whether IBLI coverage in a homeland reduces the distance between the homeland centroid and conflict events involving that group. They find that it does, indicating that IBLI coverage reduces the spatial range of pastoralist drought migration and thus the probability of conflict-generating encounters with other land users.

Q: How do the authors establish the income-smoothing mechanism?

A: Using geo-coded Afrobarometer survey data, the authors show that IBLI coverage is associated with higher reported incomes among pastoralist households, consistent with Jensen et al. (2017). Higher incomes raise the opportunity cost of fighting (following Grossman, 1991), contributing to the overall conflict-mitigating effect alongside reduced migratory pressure.

Q: What does the instrumental variable specification find?

A: The authors instrument inverse-distance-weighted IBLI payouts in the neighborhood with the interaction of neighborhood rainfall deficit and neighborhood IBLI coverage. The first stage confirms that rainfall deficits trigger payouts conditional on coverage. The second stage finds that the occurrence of payouts in the neighborhood reduces the probability of conflict by approximately 150% relative to the baseline risk, corroborating the reduced-form results.

Q: How do the authors assess cost-effectiveness?

A: The authors predict plausible drought-induced conflict fatalities in Kenya over the pre-treatment period and calculate yearly lives saved from the main estimates, then compare the monetary value of saved lives to government subsidy expenditures on IBLI. Using conservative VSL estimates from the WHO and World Bank, IBLI delivers between 10 and 22 cents of pure fatality savings per dollar of public subsidy expenditure.

Q: How robust are the results to alternative drought and conflict measures?

A: Results are qualitatively similar using an Aridity Index or Dry Matter Productivity (DMP) as drought proxies instead of rainfall deficit. The estimated interaction effect maintains a t-statistic above two for spatial decay functions ranging from distance^-0.5 to distance^-1.5 and for Conley standard error cutoffs from 200 km up to 400 km. Results also hold when restricting to conflict events not involving the government, or to battles, riots, and violence against civilians only, and when excluding the pre-IBLI period (2000–2009) entirely.

Q: What are the policy implications regarding scalability?

A: Pastoralism covers 43% of the African landmass across 36 countries, supporting approximately 268 million people (FAO, 2018). The World Bank and private equity were planning to invest close to 900 million dollars in East African pastoralist programs over 2023–2027. The authors argue that IBLI’s cost structure — high fixed costs of technology and setup but low marginal costs of expansion — gives it a scalability advantage over cash transfer programs or public works schemes that require sustained state capacity. Market-based IBLI complements rather than substitutes for political and institutional reforms.

Index-Based Livestock Insurance (IBLI): A financial instrument that uses satellite remote sensing to automatically trigger preemptive cash payouts to pastoralists when a pre-determined district-specific drought threshold is crossed, bypassing direct loss assessment and thereby eliminating moral hazard and adverse selection problems inherent in traditional indemnity insurance.

Drought-conflict semi-elasticity: The percentage-point change in the probability of conflict associated with a one-percentage-point increase in the rainfall deficit; the paper’s main outcome quantity, estimated at 6.92 percentage points at mean IBLI coverage, reduced by 23% for a one-standard-deviation increase in neighborhood IBLI coverage.

Neighborhood approach: An empirical strategy that measures both drought severity and IBLI coverage as inverse-distance-weighted averages over all surrounding grid cells, reflecting the authors’ finding that pastoralist drought-migration generates conflicts not in the pastoral home area but in neighboring mixed-land-use zones where migration routes intersect other land users.

Migratory pressure: The mechanism by which drought forces pastoralists — who hold excess livestock as precautionary savings in the absence of insurance — to migrate farther from traditional grazing grounds into mixed-land-use areas, increasing the probability of encounters and violent miscoordination with farmers, urban dwellers, and protected-area managers.

Mixed land use: Areas, designated using a 1984 Kenyan government land-use map, where pastoral grazing zones are proximate to farms, ranches, urban settlements, or nature reserves; the paper identifies these as the locations with the highest expected treatment intensity, where IBLI coverage reduces drought-induced conflict approximately four to nine times more than elsewhere.

Tropical Livestock Unit (TLU): The standard unit of account for IBLI contracts in Kenya; one TLU corresponds to one head of cattle or ten goats or sheep; the Kenyan government fully subsidizes IBLI for up to five TLUs per household.

Rollout-cluster-times-period fixed effects: A restrictive set of fixed effects included in the alternative identification strategy that absorbs all omitted variables varying at the level of the six IBLI spatial rollout clusters over time, allowing the authors to identify the conflict-mitigating effect purely from within-cluster variation in mixed-land-use exposure.

International Reserve Management Under Rollover Crises

Mon, 01 Jan 0001 00:00:00 +0000

The paper extends the Cole-Kehoe (2000) sovereign rollover crisis model to include international reserves and derives the joint optimal management of sovereign debt and reserves in a small open economy subject to potential creditor coordination failure. The central results are: (i) reserves are only valuable as a rollover-crisis defense when debt has sufficiently long maturity; (ii) the optimal exit path from the crisis zone requires holding zero reserves while gradually reducing debt, then jumping simultaneously to the optimal safe pair (a*, b*) by issuing new debt while accumulating reserves; (iii) this seemingly paradoxical debt-financed reserve accumulation lowers bond spreads because it moves the economy fully into the safe zone.

Environment: The government issues long-maturity bonds with Macaulay duration 1/δ (δ=1 is one-period debt; δ→0 is a consol). In each period, creditors decide whether to roll over. If the economy is in the crisis zone C (defined below), a sunspot ζ ∈ {0,1} with P(ζ=1) = λ determines whether a coordination failure occurs: if ζ=1 and the government is in C, creditors refuse to roll over, and the government must use reserves to service debt; if reserves are insufficient, the government defaults. The government also holds reserves a ≥ 0 earning the risk-free rate r.

Three-zone structure (Definition 1, Figure 1): the debt-reserve space (b,a) is partitioned into:

Safe zone S: b < b−(a) — government can meet its debt obligations even if the rollover crisis sunspot realizes (ζ=1); reserves are sufficient to cover the redemption shortfall
Crisis zone C: b−(a) ≤ b ≤ b+(a) — a rollover crisis is possible but not inevitable; if ζ=1, the government defaults unless reserves cover the gap; if ζ=0, the government refinances normally
Default zone D: b > b+(a) — the government defaults regardless of the sunspot because its debt burden exceeds any feasible repayment

Proposition 2 — Reserves expand the safe zone: Both boundaries b−(a) and b+(a) are increasing in reserves a. The slope of b−(a) with respect to a is steeper than the slope of b+(a), so as reserves rise: the safe zone expands, the crisis zone narrows, and the default zone shrinks. Reserves improve debt sustainability by shifting both zone boundaries to higher debt levels, but the benefit falls with debt because high-debt governments are closer to the default zone where reserves cannot compensate.

Proposition 3 — Positive reserves require long debt maturity: Optimal reserves a* > 0 requires that debt maturity is long enough (condition (18): δ < δ̄ for some threshold δ̄ < 1). The intuition is mechanical: if there is a rollover crisis with one-period debt (δ=1), the government must immediately repay the full face value b of all outstanding bonds; moderate reserve stocks a « b cannot cover this, making reserves useless. With long-maturity debt (δ<1), a rollover crisis only forces repayment of the near-term cash flow (δb plus coupon), which a much smaller reserve buffer a can cover. Hence reserves only provide value — and are only demanded — when debt has sufficient duration.

Proposition 4 — No reserves with one-period debt: When δ=1 (pure short-term debt), the optimal reserve level is zero: a* = 0. This follows directly from Proposition 3: one-period debt lies above the maturity threshold, so the safe zone cannot be expanded by any feasible reserve level.

Proposition 5 and Corollary 1 — Optimal exit strategy: The optimal exit path from the crisis zone is non-monotone in reserves:

While in the crisis zone, hold zero reserves (a=0) and reduce debt b through primary surpluses
Continue reducing debt until the government can reach the optimal safe pair (a*, b*) in a single period
In that final period, simultaneously issue new debt (increase b) AND accumulate reserves (increase a to a*), jumping directly from the safe zone to (a*, b*)

The counterintuitive simultaneous debt issuance in step 3 lowers bond spreads immediately because the reserve accumulation moves the economy firmly into the safe zone, eliminating rollover risk for creditors who then demand a lower yield premium. The optimal path delays all reserve accumulation until this transition step — building reserves gradually while in the crisis zone is suboptimal because partial reserves still leave the economy vulnerable to sunspot crises while incurring the return cost of holding low-yield liquid assets.

Proposition 6 — One-period exit condition: If the government’s current net foreign asset position NFA = a − q·b exceeds the NFA at (a*, b*), the government can exit the crisis zone in a single period.

Calibration (Italy 2012 sovereign debt crisis as the target economy):

Endowment: y = 1 (normalized); relative risk aversion: σ = 2; risk-free rate: r = 3% annually; discount factor: β = (1+r)^{−1}
Debt maturity: 1/δ = 7 years (corresponding to Italy’s average debt maturity in 2012)
Default cost: consumption floor c = 0.70 (government can guarantee 70% of normal consumption even in default, with the residual representing trade balance adjustment and output losses)
Rollover crisis probability: λ = 0.5% per quarter (calibrated to historical sovereign crisis frequency in the data)
Crisis zone midpoint parameter ϕ calibrated to set the midpoint of the crisis zone at 90% of GDP debt (consistent with Italy’s 2012 position at the crisis zone boundary)
Optimal safe pair: a* = 0.05 (5% of GDP in reserves); b* = 0.93 (93% of GDP in debt)
With reserves a = a*: bond price at b = b* is higher than without reserves; the b+(a) boundary shifts outward, confirming reserves improve debt sustainability
Without reserves (a=0): for the same debt level b = b*, bond price is lower and rollover risk is higher — the counterfactual quantifies the reserves premium

Sensitivity analysis:

Shorter debt maturity (1/δ = 4 years): optimal reserves rise substantially, to approximately 30% of GDP, because shorter maturity means the government must cover a larger fraction of face value in a rollover crisis
Higher risk aversion (σ > 2): optimal reserves increase (the welfare cost of default is higher, raising demand for precautionary reserves)
Higher default cost (lower consumption floor c): optimal reserves decrease (default is so costly to avoid that the government maintains a small debt stock in the safe zone even without reserves)

Policy implication: The standard IMF prescription to immediately accumulate reserves after a sovereign crisis is suboptimal for highly indebted governments. The paper prescribes the opposite sequence: first reduce debt through fiscal adjustment until the government can jump to (a*, b*) in a single step, then execute the jump by simultaneously issuing debt and accumulating reserves. Importantly, this jump increases both debt and reserves relative to the pre-jump position but is welfare-improving because it eliminates rollover risk — the yield reduction from entering the safe zone more than offsets the higher debt service.

Scope conditions: The model abstracts from: reserves serving exchange rate management or import coverage purposes (only rollover crisis defense modeled); a domestic banking sector; capital controls; negotiated renegotiation after default (default is assumed final). The rollover crisis mechanism is purely self-fulfilling (no fundamental triggers); the calibration is specific to Italy’s 2012 maturity structure, output level, and crisis zone midpoint.

Summary of a forthcoming paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. What are the three zones, and how do reserves shift their boundaries?

The safe zone S is the set of (b,a) pairs where the government can repay even under a rollover crisis sunspot (ζ=1), because reserves cover the financing shortfall; the crisis zone C is where self-fulfilling rollover crises are possible but not inevitable (government survives if ζ=0); the default zone D is where the government defaults regardless of the sunspot because debt exceeds any payable amount. Reserves shift both boundaries of the crisis zone to higher debt levels (Proposition 2), with the S/C boundary b−(a) rising more steeply than the C/D boundary b+(a), so the safe zone expands and the crisis zone narrows as reserves increase. This shift is the core channel through which reserves improve debt sustainability: at any given debt level b, a higher a makes it more likely that b < b−(a) (i.e., the economy is in the safe zone).

Q2. Why do reserves only matter for long-maturity debt?

With one-period debt, a rollover crisis forces immediate repayment of the full face value b — a total that any realistic reserve stock a « b cannot cover, so reserves provide zero marginal benefit against rollover risk. With long-maturity debt (duration 1/δ), a rollover crisis only requires repayment of the current-period obligation (δb + coupon), which scales with δ; as δ → 0 (near-perpetuity), this obligation becomes arbitrarily small and any positive reserve stock can cover it. Proposition 3 formalizes this by showing that a* > 0 requires δ < δ̄ (a maximum maturity threshold), and Proposition 4 confirms that δ=1 (one-period debt) implies a*=0 regardless of other parameters.

Q3. Why should a government in the crisis zone hold zero reserves?

Holding reserves while in the crisis zone is costly because reserves earn the risk-free rate r, which is lower than the sovereign’s borrowing rate (which includes a rollover risk premium); the cost of holding reserves is therefore the spread between the sovereign’s borrowing cost and the risk-free rate. The benefit of reserves while in the crisis zone is partial: positive reserves reduce the probability of default in a rollover crisis but do not eliminate rollover risk entirely (the economy remains in C for moderate a). The return on accumulating reserves jumps discontinuously when crossing from C into S — only in the safe zone do reserves entirely eliminate rollover risk. Hence the optimal strategy concentrates all reserve accumulation at the transition step when the economy crosses into the safe zone.

Q4. Why does the optimal exit involve simultaneously issuing debt and accumulating reserves?

The jump to (a, b) requires the government to reach a higher reserve level a* and a higher-than-current debt level b* simultaneously; b* > current b because (a*, b*) is inside the safe zone at a debt level the government can afford, not at the minimum possible debt level.** The debt issuance at the moment of transition is financed at the safe-zone bond price (lower spread) rather than the crisis-zone price, making the gross financing cost of the extra debt affordable. More importantly, the simultaneous reserve accumulation moves the economy into the safe zone, raising the bond price immediately: creditors see that a = a* makes b = b* safe, and they lower the yield premium accordingly. This feedback means the jump is self-financing in terms of expected debt service — the yield reduction partially covers the cost of holding reserves.

Q5. Why is the IMF prescription of immediate reserve accumulation suboptimal?

The standard prescription is to begin accumulating reserves as soon as a crisis episode passes, which keeps the government in the crisis zone longer (because reserve accumulation diverts fiscal resources from debt reduction) while paying the spread cost on all reserves held at crisis-zone yields. The paper’s prescription is to instead prioritize debt reduction until the government can make the one-step exit (Proposition 6: NFA(current) > NFA(a*, b*)), then execute the jump. This path reaches the safe zone with total lower expected cost because: (i) time spent in the crisis zone is minimized; (ii) the carry cost of reserves (spread between borrowing rate and safe asset return) is paid only for the brief period of the transition, not throughout the exit path.

Q6. How do reserves affect bond prices and spreads?

Reserves reduce sovereign spreads through two channels: (i) a direct precautionary channel — for a government already in the safe zone, reserves make the safety guarantee more credible and support the high bond price; (ii) a zone-transition channel — crossing from the crisis zone to the safe zone by accumulating reserves to a eliminates the rollover risk premium that was embedded in crisis-zone yields.* In the calibration, at Italy’s 2012 debt level (≈127% of GDP), zero reserves implies the government is in the crisis zone or default zone — bonds trade at distressed prices. At the calibrated safe pair (a*=5%, b*=93%), bonds price at the risk-free rate plus a default risk premium that excludes rollover-crisis risk. The counterfactual (same b*, a=0) yields a lower bond price, quantifying the reserves’ contribution to debt sustainability.

Q7. What does the Italy 2012 calibration imply for actual Eurozone crisis management?

Italy’s 2012 debt-to-GDP ratio of approximately 127% places it well above the optimal target b=93%, suggesting Italy was not in the safe zone even had it held substantial reserves; the primary prescription for Italy at that moment — debt reduction, not reserve accumulation — follows directly from the model’s exit strategy (Propositions 5-6).* The model also implies that European bailout mechanisms (ESM, OMT) shifted the effective boundary of the safe zone by providing contingent external reserves, consistent with the empirical observation that ECB President Draghi’s “whatever it takes” announcement in July 2012 moved Italy’s bond yields toward safe-zone pricing without any actual reserve or debt movement.

Key concepts

rollover crisis : a self-fulfilling coordination failure in which creditors refuse to roll over maturing sovereign debt not because solvency fundamentals require default but because they expect other creditors to refuse; modeled by a sunspot ζ=1 with probability λ that triggers a crisis when the economy is in the crisis zone C.

safe zone : the set of (b,a) pairs where the government can service its debt even under the worst-case sunspot (ζ=1); defined by b < b−(a); entering the safe zone eliminates rollover risk entirely and immediately lowers bond yields to the risk-free rate plus a pure credit-risk premium.

crisis zone : the set of (b,a) pairs where rollover crises are possible but not certain; b−(a) ≤ b ≤ b+(a); the government survives if ζ=0 but defaults if ζ=1; bonds are priced to include a rollover risk premium while in this zone.

optimal exit strategy : Proposition 5 and Corollary 1 — the welfare-maximizing path out of the crisis zone; involves holding zero reserves while reducing debt, followed by a simultaneous jump to (a*, b*) that increases both reserves and debt, moving the economy immediately to the safe zone and eliminating rollover risk in a single step.

long-maturity debt advantage : the property (Proposition 3) that reserves only provide rollover-crisis protection when debt has sufficiently long maturity (δ < δ̄); with short-maturity debt, a rollover crisis forces repayment of the full face value, which no realistic reserve stock can cover; with long-maturity debt, only the near-term cash flow must be covered.

debt-financed reserve accumulation : the seemingly paradoxical simultaneous issuance of new long-maturity bonds and accumulation of reserves at the moment of exit (a=0→a*, b<b*→b*); welfare-improving because the jump moves the economy into the safe zone, lowering bond yields immediately and making the higher debt affordable.

Leveraging Virtual Contact and Social Networks to Foster Interethnic Harmony

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates whether virtual contact — exposure to an outgroup through a documentary film — can promote interethnic harmony, and whether targeting network-central individuals amplifies effects on untreated community members. The study addresses a context of deep, historically rooted discrimination: the Santal ethnic minority in northwestern Bangladesh have faced colonial-era land dispossession, ongoing violence, labor market discrimination, and structural exclusion by the Bengali ethnic majority. The Santals are the second-largest ethnic-minority group in Bangladesh; in the study villages, their share ranges from 13% to 83% of the population.

The authors conducted a cluster-randomized field experiment across 121 multiethnic villages in the Rajshahi and Naogaon districts of Bangladesh, involving over 3,300 households. Villages were randomly assigned to three arms: a random treatment arm (RR, 40 villages, N=562 Bengalis) in which approximately 14 randomly selected ethnic-majority households per village watched a 45-minute documentary film (“Ami Santal” / “I Am Santal”) portraying Santal culture, economic hardships, and aspirations; a central treatment arm (41 villages) in which approximately 7 randomly selected Bengalis (RC) and 7 network-central Bengalis identified via a diffusion-centrality nomination exercise (CC) watched the same film; and a control arm (40 villages) in which households watched a placebo documentary on flower farming. The documentary, costing approximately $13 per participant, was screened individually at participants’ homes on tablets. Data were collected at baseline (September–October 2022), first end line approximately 3 months post-screening (February–March 2023), and a casual-work field experiment second end line approximately 4.5–5 months post-screening (April–May 2023). Outcomes were measured via lab-in-the-field experiments (dictator game, solidarity game), an experimentally validated interethnic trust survey item (Falk et al. 2018), self-reported behaviors, administrative police complaint data, and facial emotion detection during screening.

The main findings are as follows. First, treated Bengalis in the central arm (RC) gave 14.7% more in the dictator game (p < .01) and exhibited 21.7% greater trust toward Santals (p < .01) compared to controls; RR participants showed a 7.1% increase in solidarity game giving (p < .10) and 11.8% greater trust (p < .01). Effects on reducing negative stereotypes and discriminatory opinions were not statistically significant, suggesting that affective components of prejudice are more responsive to the intervention than cognitive components. About 82% of treated Bengalis reported acquiring new information about Santals, primarily regarding occupational struggles, educational aspirations, and economic potential. Facial expression analysis using emotion-detection software found sadness to be significantly more prevalent among viewers (p < .05), particularly among network-central participants, consistent with an empathetic response.

Second, untreated Bengalis in the central arm — who never watched the documentary — showed 20.9% higher altruism (p < .10), 27.3% higher solidarity (p < .05), and 8.1% higher trust (p < .05) toward Santals relative to controls. No significant effects on untreated Bengalis were found in the random arm. Untreated Santals in both arms exhibited greater trust toward Bengalis (11% increase in random arm, p < .05; 21.7% increase in central arm, p < .01) and higher subjective well-being (p < .01 in both arms). Village-level administrative data show a significant reduction in Bengali police complaints against Santals post-intervention (p < .05), but only in the central arm.

Third, in the casual-work field experiment, multiethnic pairs jointly produced paper bags under piece-rate compensation. Overall productivity increased approximately 5% (p < .05) in the central arm only. Both Bengali and Santal workers increased productivity specifically in the finisher role — the most critical role for determining earnings — in the central arm. The authors interpret Bengali productivity gains as reflecting increased prosociality toward Santal co-workers, and Santal productivity gains as reflecting conformism or peer pressure in response to Bengali effort. The scope of all effects is limited to multiethnic villages in northwestern Bangladesh, a context of historically severe and ongoing majority-minority inequality; the intervention deliberately did not challenge the socioeconomic hierarchy of the villages.

Q: What was the documentary film’s content and design rationale? A: The 45-minute film “Ami Santal” featured three narrative layers: Santal culture (rituals, cuisine, the Baha festival), economic hardships (housing, water access, low incomes, labor market struggles, educational barriers), and aspirational stories of Santals who achieved success. All stories were narrated by non-actor local Santals, filmed outside the study region, and deliberately avoided attributing blame to Bengalis. The film was designed under the supervision of anthropologists at the University of Rajshahi to maintain ethnographic authenticity and a non-moralistic, observational tone (moral judgment language was much lower than in comparison Bangladeshi documentaries and general films, per LIWC-22 analysis).

Q: How were network-central individuals identified and why might targeting them matter? A: In central-arm villages, enumerators surveyed approximately 18–20 randomly selected passers-by at village markets and asked them to nominate the 15 people most effective at disseminating information. The seven most consistently and highly ranked individuals per village were selected as network-central (CC). These individuals were expected to have high diffusion centrality — meaning information they receive spreads widely — so targeting them with the documentary could shift attitudes and behavior among untreated community members through persuasion, visibility, credibility, or diffusion (the paper cannot separately identify which mechanism operates).

Q: What were the primary behavioral effects on treated Bengalis (the ethnic majority who watched the film)? A: Randomly selected participants in the central arm (RC) gave 14.7% more in the dictator game (p < .01) and 8% more in the solidarity game (not statistically significant), and exhibited 21.7% greater trust toward Santals (p < .01), all relative to controls. In the random arm (RR), participants showed a 6.4% increase in dictator game giving (not statistically significant), a 7.1% increase in solidarity game giving (p < .10), and 11.8% greater trust toward Santals (p < .01). Effects on self-reported behaviors — interethnic friendships, social interactions, amount charged to minorities for water — were not statistically significant.

Q: Did the intervention change Bengali stereotypes or discriminatory opinions toward Santals? A: No. Despite treated Bengalis acquiring substantial new information (approximately 82% reported learning new things, primarily about Santal occupational struggles and educational aspirations), the authors find no significant effects on the stereotypes index or the discriminatory-opinions index among treated Bengalis. They propose two explanations: cognitive components of prejudice (stereotypes) are harder to change through indirect contact than affective components (emotions, prosocial behavior), consistent with Tropp and Pettigrew (2005) and Turner, Crisp, and Lambert (2007); and a single documentary may be insufficient to counter deeply ingrained generational biases due to resistance to change.

Q: What emotional responses did the documentary elicit, and how was this measured? A: Field assistants took candid photographs of participants’ faces at a random point during the screening; these were analyzed using Emotimeter software (machine learning-based emotion detection) that assigns scores across seven emotion categories summing to 100%. Sadness was significantly more prevalent among documentary viewers compared to placebo viewers (p < .05), particularly among network-central participants (CC). The authors interpret this as consistent with an empathetic response to the film’s content about Santal hardships, and connect it to increased prosocial behavior via emotion-regulation mechanisms (alleviating sadness through prosocial action).

Q: What were the spillover effects on untreated Bengalis in the central arm? A: Untreated Bengalis in central-arm villages — who never watched the documentary — showed 20.9% higher altruism (p < .10), 27.3% higher solidarity (p < .05), and 8.1% higher trust toward Santals (p < .05) relative to controls. By contrast, untreated Bengalis in random-arm villages showed no statistically significant effects on any of these outcomes. The authors attribute the central-arm spillovers to the presence of network-central individuals being treated in those villages, though whether these patterns reflect persuasion, visibility, credibility, or information diffusion cannot be separately identified.

Q: How did the intervention affect the Santal ethnic minority (who never watched the documentary)? A: Untreated Santals in both arms exhibited greater trust toward Bengalis: an 11% increase in the random arm (p < .05) and a 21.7% increase in the central arm (p < .01) compared to controls. Santals in both arms also reported higher subjective well-being (p < .01). A weakly significant increase in food security was observed among Santals in the central arm (p < .10), possibly reflecting increased material support from Bengalis. No statistically significant effects were found on Santal altruism or solidarity.

Q: What did the village-level administrative complaint data show? A: Using data collected from two police stations covering all 121 villages, the authors find a significant reduction in Bengali complaints against Santals post-intervention in the central arm (p < .05). No significant reduction was found in Santals’ complaints against Bengalis (p > .10) in any arm. Data from village counselors’ offices (shalish arbitration complaints) showed no significant change in any arm. The distinction matters because police complaints involve more serious, violent matters, while village-counselor complaints involve routine arbitration.

Q: How was the casual-work field experiment designed, and what did it find? A: Approximately 4.5 months after the documentary screenings, 720 participants (360 Bengalis, 360 Santals) drawn equally from the three study arms were paired into multiethnic dyads to jointly produce paper bags for a local supplier under piece-rate compensation, with earnings split equally. One worker was randomly assigned the preparer role and the other the finisher role; roles were switched halfway through the three-hour session. The paper finds an approximately 5% overall productivity increase (p < .05) in the central arm only, concentrated in the finisher role (the role most critical for final output). Bengalis and Santals both increased productivity specifically as finishers in the central arm.

Q: What mechanisms explain the productivity effects in the casual-work experiment? A: For Bengali finishers, the productivity gain is interpreted as prosocial behavior: treated Bengalis who showed greater altruism toward Santals worked harder to increase the earnings of their Santal co-workers. For Santal finishers, the productivity gain is interpreted as conformism or peer pressure: Santals increased effort more when they worked as finisher after swapping roles (i.e., after observing Bengalis’ higher effort as finisher first), suggesting responsiveness to the higher productivity of Bengalis rather than an independent prosocial motivation. The authors present a simple theoretical model to formalize these interpretations, citing Rotemberg (1994) on prosocial effort and Kandel and Lazear (1992) and Mas and Moretti (2009) on peer pressure mechanisms.

Q: Why was virtual rather than direct contact used in this intervention? A: The authors argue that encouraging direct contact between Bengalis and Santals in this setting carries specific risks: the unequal status of the groups may generate anxiety during interactions, potentially limiting engagement or provoking backlash. By contrast, the documentary provides an indirect, low-cost ($13 per participant) form of contact that presents Santal lives without disrupting the socioeconomic hierarchy of the villages and without attributing blame to Bengalis. The film’s entertaining veneer and emotional storytelling make it more scalable and logistically feasible in contexts where direct contact is socially difficult or impractical.

Q: What are the primary limitations acknowledged by the authors? A: The authors acknowledge that the study’s sampling protocol relied on a door-to-door skip procedure without systematic records of approached households, raising the possibility of convenience or snowball-type recruitment and potential deviations from random sampling — this is reflected in some imbalances in baseline characteristics across arms. CC-control comparisons are explicitly descriptive (not causal) because network-central individuals were selected on centrality. Differential attrition was found among untreated Santals (both treatment arms had significantly lower attrition than control, p < .05), which could bias estimates for that subgroup. The authors cannot separately identify the mechanisms (persuasion, visibility, credibility, diffusion) underlying spillover effects in central villages.

Q: What are the policy implications of this study? A: The findings suggest that media-based virtual contact interventions are a low-cost, scalable tool for improving interethnic prosociality even in contexts of deep-rooted discrimination where direct contact may be socially impractical. Targeting network-central individuals — identified via a simple nomination exercise requiring no pre-existing network data — amplifies village-wide effects, including among untreated community members and the minority group itself. The productivity gains in multiethnic work teams imply that improved interethnic relations can have tangible economic consequences beyond attitudinal change. However, the null effects on stereotypes and discriminatory opinions suggest that single documentary interventions may not be sufficient to alter deep-seated cognitive biases, and more intensive or repeated interventions may be needed to achieve durable attitude change.

Virtual contact: Indirect exposure to an ethnic outgroup through a documentary film, as distinct from direct intergroup contact; posited to influence majority-group attitudes and behavior by increasing empathy and identification with the outgroup without requiring face-to-face interaction.

Diffusion centrality: A network measure of how effectively an individual can spread information through a community, operationalized via a nomination exercise in which community members identify those best positioned to disseminate information; used to select the seven highest-ranked individuals per village for targeted treatment.

Prosociality (altruism and solidarity): Measured using incentivized lab-in-the-field games — the dictator game (unilateral allocation of an endowment to a passive outgroup recipient) and the solidarity game (precommitted transfers to an outgroup member who may incur a random loss) — capturing willingness to benefit non-coethnic others at personal cost.

Affective versus cognitive components of prejudice: A distinction between emotional aspects of prejudice (feelings, empathy) — which the authors find to be more responsive to the documentary intervention — and cognitive aspects (negative stereotypes, discriminatory opinions) — which show no significant change despite new information acquisition.

Spillover effects (untreated individuals): Changes in behavior or attitudes among community members who did not directly receive the intervention (did not watch the documentary), attributed to the influence of treated individuals in their village, particularly network-central individuals in the central arm.

Piece-rate casual-work field experiment: A second end line in which multiethnic pairs of Bengali and Santal workers jointly produced paper bags for a local supplier, with individual earnings determined by joint piece-rate output; designed to measure whether improved interethnic attitudes translated into higher workplace productivity in ethnically mixed teams.

Source text origin: The provenance classification of the text used to generate a paper summary (full PDF, open-access HTML, or abstract only); the paper’s pipeline rules impose a hard block on abstract-only summarization.

Lives Versus Livelihoods: The Impact of the Great Recession on Mortality and Welfare

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. Does the Great Recession reduce or increase mortality, and what are the welfare implications of incorporating recession-induced mortality changes into standard macroeconomic welfare frameworks?

Setting and Identification. The authors exploit spatial variation in the severity of the 2007–2009 Great Recession across 741 U.S. Commuting Zones (CZs), following the empirical design of Yagan (2019). The primary shock variable is the percentage-point change in the CZ unemployment rate between 2007 and 2009. The key identifying assumption is that no concurrent shocks to mortality coincide with the timing and geographic pattern of the Great Recession shock. Pre-trend evidence supports this: CZs subsequently harder hit experienced a slight relative increase in mortality before 2007, which is the opposite sign from the main effect, supporting the validity of the design.

Data. Mortality data come from CDC restricted-use death certificate microdata (2003–2016) covering the universe of U.S. deaths, combined with SEER population denominators. A 20 percent random sample of Medicare enrollees aged 65–99 provides an individual-level panel that directly addresses concerns about endogenous migration. The main outcome is the log age-adjusted CZ mortality rate; economic indicators come from BLS, BEA, and FHFA; air pollution data from the EPA AQS monitor network (PM2.5); morbidity from the BRFSS; nursing home characteristics from federal certification inspections.

Main Mortality Finding. A one-percentage-point increase in the local unemployment rate between 2007 and 2009 is associated with a 0.50 percent decline (SE = 0.15) in the annual age-adjusted mortality rate in 2007–2009, and a 0.58 percent decline (SE = 0.34) in 2010–2016; the two periods are statistically indistinguishable (p = 0.78). Because the national average unemployment rate rose by 4.6 percentage points, the Great Recession on average reduced the annual age-adjusted mortality rate by approximately 2.3 percent, with effects persisting for at least 10 years. The authors note this is equivalent to approximately two years of secular mortality improvement at the pre-recession trend pace of 1.1 percent per year. For a 55-year-old, the estimates imply that 1 in 25 gained an extra year of life from a shock of this magnitude.

Heterogeneity by Cause of Death. Mortality declines appear across most major causes. Cardiovascular disease (34 percent of 2006 deaths) declines by 0.65 percent per percentage-point unemployment increase (SE = 0.21) and accounts for approximately 48 percent of the total estimated mortality reduction. Motor vehicle mortality falls by 1.7 percent (SE = 0.56) and liver disease by 1.1 percent (SE = 0.43). Suicides show a statistically significant 1.7 percent decline (SE = 0.5) in the 2010–2016 period. The notable exception is cancer (the second-largest cause of death), for which the estimated effect is a precise null of 0.02 percent (SE = 0.11). The null cancer result is interpreted as a specification check: if mortality declines were spurious (e.g., driven by population mismeasurement), cancer mortality should also decline.

Heterogeneity by Demographics. Recession-induced mortality declines are similar in percentage terms across gender and race/ethnicity, and statistically equi-proportional across age groups (p-value for equality across 25–64 versus 65+: 0.76). Because mortality is heavily concentrated in the elderly, those aged 65 and over account for approximately 74.3 percent of averted deaths, roughly proportional to their 72.5 percent share of 2006 mortality. The most striking heterogeneity is by education: the entire mortality decline is concentrated among the approximately 52 percent of the population with a high school degree or less. The estimated 2007-2016 effect is −1.3 percent per percentage-point unemployment increase (SE = 0.56) for those with high school or less, compared to +0.34 percent (SE = 0.68) for those with more than high school (statistically distinguishable at p < 0.01).

Mechanisms. The authors distinguish internal effects (own reduced employment or consumption improving health) from external effects (externalities from reduced aggregate economic activity, holding own employment/consumption fixed). Evidence strongly favors external effects as the primary driver. Three-quarters of averted deaths accrue to the elderly, who experienced no direct income effects from the labor market shock. Moreover, the timing pattern—an immediate mortality drop that does not grow over time—is inconsistent with health-behavior channels (e.g., smoking cessation, improved diet) that would build up gradually. Direct tests find no statistically significant impact on self-reported health behaviors (smoking, drinking, exercise) and no impact on healthcare use among Medicare enrollees.

Among external channels, neither reduced spread of infectious disease nor improved nursing home staffing receives empirical support. Reduced air pollution (PM2.5) is identified as a quantitatively important channel. A one-percentage-point increase in CZ unemployment is associated with a 0.16 µg/m³ decline in PM2.5 (SE = 0.04), a 1.3 percent decline relative to the 2006 national average of 12 µg/m³. A mediation analysis (controlling for the PM2.5 shock) attenuates the estimated mortality effect by 37 percent, from −0.52 percent to −0.33 percent per percentage-point unemployment increase. Back-of-the-envelope calculations combining the PM2.5 decline with external estimates of PM2.5-mortality elasticities suggest pollution can explain 17 to 35 percent of total recession-induced mortality declines.

Lag Structure. Exploiting variation in the speed of post-recession labor market recovery (measured by 2010–2016 EPOP ratio changes) conditional on the initial shock, the authors find that mortality reductions persist in areas that have fully recovered economically by 2016, suggesting lagged mortality effects of the initial economic downturn beyond what contemporaneous economic conditions alone explain.

Welfare Analysis. The authors extend the Krebs (2007) consumption-based welfare cost-of-recessions model to incorporate endogenous mortality. For a 45-year-old with γ = 2 and a value of a statistical life-year (VSLY) of $250k (five times annual consumption), accounting for endogenous mortality reduces the willingness to pay to avoid all future recessions from 2.00 percent of average annual consumption to 0.91 percent—a reduction of approximately 55 percent. Starting around age 55, recessions become welfare-improving on net. For the Great Recession specifically, at age 55 endogenous mortality reduces the welfare cost by approximately 25 percent (from 2.39 to 1.80 percent of average annual consumption). Because mortality declines are concentrated among those with high school or less, accounting for endogenous mortality also substantially mitigates—and at older ages reverses—the finding that the Great Recession was more costly for the less educated.

Scope Conditions and Caveats. (i) The design captures only differential local effects, not nationwide impacts (e.g., stock market collapse, nationwide malaise). (ii) Mortality impacts may not generalize to milder recessions, though the relationship appears approximately linear in shock size. (iii) The analysis excludes morbidity, though limited evidence suggests morbidity is also pro-cyclical and roughly equi-proportional across ages. (iv) The welfare analysis begins at age 35 and does not account for longer-run mortality costs of recession entry for younger cohorts.

In depth

Q1. What is the baseline empirical specification, and why does the design exploit cross-sectional variation rather than time-series panel regressions?

The estimating equation regresses the log age-adjusted CZ mortality rate on an interaction of the CZ-level Great Recession shock (2007–2009 unemployment change) with year indicators, plus CZ and year fixed effects, weighted by 2006 CZ population. The authors prefer this to the standard two-way fixed effects panel approach (area and year FE with contemporaneous unemployment rate) for three reasons: (1) it directly identifies the full dynamic lag structure of the shock rather than imposing contemporaneity; (2) exploiting a single spatially differentiated shock reduces risk of confounding from other concurrent area-level shocks; (3) the panel can be linked to individual-level Medicare data, allowing explicit control for endogenous migration, which the existing literature cannot do.

Q2. How does the paper address the concern that mortality rate declines might simply reflect unmeasured population outflows from hard-hit areas rather than genuine reductions in deaths?

The authors offer two main responses. First, cancer mortality shows a precise null effect despite being the second-leading cause of death; if unmeasured population losses were driving the results, cancer deaths should decline proportionally. Second, using the Medicare individual-level panel, they fix each enrollee’s location at their 2003 CZ and find a statistically significant mortality decline of 0.35 percent per percentage-point unemployment increase in the reduced-form (2007–2009 period). A control function approach that instruments current-year location with 2003 location yields an estimate of −0.37 percent (SE = 0.17), similar to the baseline −0.50 percent from the aggregate specification, confirming that migration bias is not the primary driver.

Q3. How long do the mortality reductions from the Great Recession persist, and does the paper identify whether these are contemporaneous or lagged effects?

The 2007–2009 period estimate is −0.50 percent per percentage-point unemployment increase and the 2010–2016 period estimate is −0.58 percent, and these are statistically indistinguishable (p = 0.78). To identify whether persistence reflects ongoing economic effects or true lagged mortality effects, the authors compare CZs with above- vs. below-median 2010–2016 EPOP recovery (conditional on initial shock decile). Both groups show similar 2010–2016 mortality declines despite the above-median recovery CZs having returned to pre-recession employment levels by 2016. This finding is consistent with lagged mortality effects of the initial economic downturn that persist independently of current economic conditions.

Q4. Are mortality reductions concentrated among individuals already near death (“harvesting”), or do they represent meaningful longevity gains?

The authors use a Medicare auxiliary model to predict counterfactual remaining life expectancy for each enrollee based on age, demographics, and chronic conditions. The marginal life saved has only about 6 percent lower counterfactual remaining life expectancy than a typical decedent of the same age, and this difference is statistically insignificant. Because effects persist over 10 years (not just days or weeks), short-run mortality displacement (harvesting) is not the operative concern. The 6 percent difference is also small enough that the authors do not adjust their welfare analysis for it.

Q5. What is the educational gradient in mortality impacts, and is it explained by age composition or other confounders?

Mortality declines are entirely concentrated among those with a high school degree or less: the 2007–2016 estimate is −1.3 percent per percentage-point unemployment increase (SE = 0.56) for this group versus +0.34 percent (SE = 0.68) for those with more than high school, distinguishable at p < 0.01. This gradient holds within age groups (confirmed in Appendix analysis), and further disaggregation shows no mortality declines for those with some college or college-or-more separately. In Medicare data, the elderly mortality effect is concentrated among the approximately 12 percent enrolled in Medicaid (a proxy for low income), reinforcing the socioeconomic concentration.

Q6. What evidence rules out improved health behaviors (increased exercise, reduced smoking, reduced alcohol) as the main mechanism?

Two types of evidence argue against this channel. First, three-quarters of averted deaths are among the elderly, who experienced no direct income or employment effects from the local labor market shock and would not plausibly change their health behaviors in response to someone else losing employment. Second, the mortality decline is immediate in 2007 and flat through 2016 rather than growing over time; smoking cessation, for example, takes 10–15 years to accumulate mortality effects. Direct tests of behavioral outcomes from BRFSS find no statistically significant impact on smoking, drinking, exercise, or flu vaccination rates, individually or pooled. The pooled average treatment effect on six morbidity measures is statistically significant and negative (suggesting morbidity improvements), but behavioral covariates show no movement.

Q7. What is the evidence for and against improved nursing home care as a mechanism?

Prior literature (Stevens et al. 2015; Konetzka et al. 2018; Antwi and Bowblis 2018) documents that recessions increase nursing home staffing and reduce nursing home deaths in earlier decades. However, the authors find no evidence for this channel in the Great Recession context. Estimated mortality impacts are virtually identical (approximately 0.5 percent per percentage-point unemployment increase) for the 7 percent of the elderly in nursing home care and the 93 percent not in nursing home care. Direct measures of nursing home staffing (direct-care staff hours per resident-day, highly skilled nurses ratio) show no statistically significant change in harder-hit areas: the point estimate for direct-care hours is −0.11 percent (SE = 0.22) in 2007–2009. Nursing home occupancy rates and resident characteristics also show no significant changes.

Q8. How is the quantitative importance of the air pollution channel estimated, and what are the two complementary approaches used?

Approach 1 (back-of-the-envelope): The authors combine their estimate that a one-percentage-point unemployment increase reduces PM2.5 by 0.16 µg/m³ with external estimates from Deryugina et al. (2019) of PM2.5’s effect on elderly daily mortality, rescaled to annual exposure. This calculation implies pollution explains 17–35 percent of total recession-induced mortality declines, depending on which Deryugina et al. mortality estimates are used. Approach 2 (mediation analysis): Adding the county-level PM2.5 shock as an additional control in the mortality regression attenuates the Great Recession mortality coefficient from −0.52 percent to −0.33 percent per percentage-point unemployment increase—a 37 percent attenuation. Both approaches are suggestive rather than definitive, as the mediation analysis requires the strong assumption that the recession shock and PM2.5 shock are conditionally independent of other unmeasured mediators.

Q9. What are the specific calibration parameters in the welfare model and how does the paper set the mortality decline parameter?

The authors extend Krebs (2007)’s income process calibration (pH = 0.03, pL = 0.05, dH = 0.09, dL = 0.21, g = 0.02, σ = 0.01, πH = 0.5) and use 2007 SSA life tables for age-specific mortality rates in normal times. The recession mortality parameter is set to dm = −0.015 for all ages, derived from a 3.1 percentage-point unemployment increase in a typical recession multiplied by the estimated 0.5 percent mortality decline per percentage-point. VSLY values are parameterized at two, five, or eight times annual consumption ($100k, $250k, or $400k at $50k annual consumption). Risk aversion γ takes values 1.5, 2, and 2.5. For the Great Recession-specific exercise, dmA = −0.023 (4.6 × 0.5 percent), dmHS = −0.037, and dmC = 0.0006.

Q10. How does accounting for endogenous mortality change the distributional welfare analysis of the Great Recession by education group?

Under exogenous mortality, the welfare cost of the Great Recession at age 35 is 2.89 percent of average annual consumption for those with high school or less versus 1.23 percent for those with more than high school—the less educated bear roughly twice the burden. Under endogenous mortality, the mortality declines are concentrated entirely among the less educated (dmHS = −0.037 vs. dmC ≈ 0), so accounting for mortality disproportionately offsets welfare losses for that group. By around age 65, the welfare costs of the Great Recession converge across education groups, and after age 65, the less educated bear lower welfare costs than the more educated, reversing the exogenous-mortality ranking. This result depends on the same education differential in mortality impacts that drives the main empirical finding.

Q11. What robustness checks demonstrate that the baseline mortality estimates are not driven by geographic or functional-form choices?

The baseline CZ-level estimate of −0.50 percent (SE = 0.15) is replicated almost exactly at the state level (−0.62, SE = 0.25) and county level (−0.49, SE = 0.10). A Poisson regression yields −0.45 percent (SE = 0.14). Dropping the top/bottom decile of CZs by shock size yields −0.46 percent (SE = 0.16). Adding Census-division-by-year fixed effects attenuates the estimate slightly to −0.38 percent (SE = 0.14) but retains statistical significance. Dropping CZs with high fracking activity and dropping the ten most populous CZs both produce estimates similar to baseline. Quartile regressions show monotone mortality reductions across quartiles of the unemployment shock, consistent with approximate linearity.

Q12. What does the expert survey reveal about prior beliefs, and how does the paper’s finding compare?

In a spring 2023 survey of over 300 experts, 50 percent predicted the Great Recession would increase mortality and only 27 percent predicted a decrease. Of those predicting a decrease, 93 percent gave a magnitude larger (in absolute value) than the paper’s negative point estimate of 0.50 percent per percentage-point unemployment increase, and 82 percent gave a prediction larger than the upper bound of the 95 percent confidence interval. This illustrates that the paper’s finding—mortality is meaningfully pro-cyclical during the Great Recession—was highly surprising to the empirical and policy economics community.

Key Concepts

Pro-cyclical mortality: The phenomenon whereby mortality rates fall during economic downturns and rise during expansions. The paper documents this for the Great Recession using a spatial identification strategy, in contrast to the time-series correlation that had weakened in the two decades before the Great Recession. The term “pro-cyclical” means mortality moves in the same direction as the business cycle (up in booms, down in recessions), implying recessions are associated with fewer deaths.

Internal vs. external effects (of recessions on mortality): The paper distinguishes internal effects—whereby an individual’s own reduced employment or consumption affects her own mortality—from external effects, which are changes in mortality from reduced aggregate economic activity that hold constant one’s own employment and consumption. This distinction has direct welfare implications: external effects (e.g., less pollution from lower industrial output) are genuine welfare improvements for people who did not lose income, while internal effects of behavioral change are mitigated by the envelope theorem if behavior is privately optimal.

Commuting Zone (CZ) shock: The paper’s primary treatment variable, defined as the percentage-point change in the CZ unemployment rate between 2007 and 2009. CZs are aggregations of counties (741 total) designed to approximate local labor markets. The median CZ experienced a 4.6-percentage-point increase, with substantial variation ranging from roughly 2.9 points (bottom quartile) to 6.7 points (top quartile).

Value of a Statistical Life-Year (VSLY): The dollar value placed on one additional year of life in expectation, used in the welfare calibration. In the paper’s framework it equals VSLY = bcγ − c/(γ−1), where b is a preference parameter governing the marginal utility of life-years. Results are reported for VSLYs of $100k, $250k, and $400k corresponding to two, five, and eight times average annual consumption of $50k, following Hall and Jones (2007).

Endogenous mortality in welfare analysis: The paper’s central theoretical contribution is augmenting the Krebs (2007) welfare cost-of-recessions framework to allow mortality to vary with the aggregate state of the economy. When mortality is endogenously lower in recessions, the willingness to pay to eliminate recession risk falls—and at high enough VSLY or old enough ages, recessions become welfare-improving because the mortality benefit outweighs the consumption cost.

Mortality displacement (harvesting): The possibility that short-run mortality declines merely reflect the premature death of already-frail individuals being slightly delayed, without meaningful longevity gains. The paper argues this is not the operative concern given 10-year persistence and uses auxiliary Medicare models to show marginal lives saved have only 6 percent shorter counterfactual life expectancy than average decedents of the same age.

PM2.5 mediation analysis: An empirical approach in which the county-level change in fine particulate matter (PM2.5, in µg/m³) between 2006 and 2010 is added as a covariate in the mortality regression. Under the assumption that the recession shock and the PM2.5 shock are conditionally independent of other unmeasured mediators, the attenuation in the recession-mortality coefficient when controlling for PM2.5 identifies the share of the mortality effect operating through the pollution channel. A 37 percent attenuation is found in the 2007–2009 period.

Making the Invisible Hand Visible: Managers and Worker Allocation

Mon, 01 Jan 0001 00:00:00 +0000

This paper asks why managers matter for firm performance, and specifically whether managers improve productivity by matching workers to better-suited jobs inside firms rather than through supervision, motivation, or selection out of the firm. The setting is the internal labor market of a large private consumer goods multinational enterprise (MNE) operating in more than 100 countries, with annual turnover exceeding EUR 50 billion. The data cover the universe of white-collar workers and managers at the firm — 200,000 workers and 30,000 managers observed monthly over 11 years (January 2011 to December 2021) — linked to payroll, performance ratings, organizational chart, digital platform activity, employee surveys, and an independent sales productivity series for field sales workers in 15 countries.

The paper confronts two identification challenges. First, the author constructs a measure of manager quality — “high flyers” — defined as managers who were promoted to the first managerial work level (WL2) by age 30. This threshold yields 26.2% of managers classified as high flyers. The measure is defined entirely ex ante, before the manager ever supervises the worker under study, which addresses reverse causality. It is validated against ex post performance metrics including future salary growth, probability of promotion to WL3, performance ratings, and anonymous subordinate feedback. Second, to identify causal effects of manager quality on workers, the author exploits the firm’s long-standing policy of rotating WL2 managers laterally across teams as part of their career development, a practice implemented for several decades. Using an event-study design centered on the worker’s first manager transition, the author compares workers who transition from a low-flyer to a high-flyer manager (LtoH) against workers who transition from one low-flyer to a different low-flyer (LtoL), netting out the effect of the transition itself. Pre-event parallel trends are confirmed empirically.

The main findings are as follows. Gaining a high-flyer manager causes substantial reallocation of workers within the firm through lateral job transfers: seven years after the manager transition event, cumulative lateral moves are 40% higher for workers who gained a high-flyer manager relative to those who gained another low-flyer. These lateral moves are not confined to a single organizational margin — transfers rise within-team, across teams in the same function, and across functions — and they involve meaningfully larger shifts in task content, as measured by angular separation across O*NET cognitive, routine, and social task intensity dimensions, with cumulative task distance becoming statistically distinguishable from zero approximately seven quarters post-transition. These gains in lateral mobility translate into persistent wage growth: seven years after the manager transition, workers supervised by a high-flyer earn salaries 13% higher than the comparison group, with divergence beginning only after the transition date. Using independent sales bonus data, three years after gaining a high-flyer manager workers’ sales productivity increases by 0.347 standard deviations, ruling out the interpretation that wage gains merely reflect manager favoritism rather than genuine productivity improvement. Establishment-level data further show that sites with a higher share of workers under high-flyer managers display higher output per worker and lower operational costs per unit.

Effects are asymmetric: gaining a good manager has large positive effects, but losing one (comparing HtoL with HtoH transitions) produces no corresponding negative effects, implying that a single exposure to a high-flyer manager generates durable benefits that survive a subsequent downgrade in manager quality. A mediation analysis finds that 64% of the salary gain is explained by lateral job changes, though the author notes this understates the full allocation channel because it excludes vertical transfers and the gains from remaining well-matched in the current role. These findings hold under multiple robustness checks including restricting to new hires, using the Sun and Abraham (2021) interaction-weighted estimator, varying the age threshold for high-flyer classification, using a tenure-based alternative, and placebo tests with randomly assigned manager types.

The scope conditions are specific to white-collar workers at a large, organizationally homogeneous consumer goods multinational. All workers hold college degrees, mean firm tenure is 8.5 years, team sizes average five workers, and the firm has the same organizational structure across all countries, functions, and years.

Q: How does the paper define “high flyer” managers and what share of managers receive this classification? A: High flyers are managers who achieved the first managerial work level (WL2) by age 30, a threshold derived from continuous age estimates constructed from 10-year age bands in the personnel records. This definition yields 26.2% of managers classified as high flyers. The measure is time-invariant and defined ex ante relative to any interaction with the workers whose outcomes are studied.

Q: What validates the high-flyer measure as capturing genuine managerial ability rather than noise? A: The high-flyer classification is significantly positively correlated with multiple ex post performance metrics recorded after the manager’s own promotion: future salary growth, probability of subsequent promotion to WL3 (director level), annual performance ratings, and anonymous upward feedback scores from subordinates on leadership. High flyers are also 14.5 percentage points less likely to be mid-career recruits, suggesting they are internally developed talent rather than external hires.

Q: What is the source of identifying variation and how does the event-study design address endogeneity? A: The firm has operated a decades-long policy of rotating WL2 managers laterally across teams to broaden their experience and to screen candidates for promotion to WL3. These rotations are asserted by firm executives and HR representatives to be orthogonal to worker and team characteristics. The author verifies this empirically by showing that a wide range of team characteristics measured over the two years before a transition — including team performance, inequality, transfer rates, and team diversity — cannot predict the type of incoming manager. The event-study design compares workers who receive a high-flyer replacement (LtoH) against workers who receive another low-flyer replacement (LtoL), netting out any generic effect of a managerial change, and confirms parallel pre-trends.

Q: What is the effect of gaining a high-flyer manager on lateral job mobility? A: Seven years after the manager transition, workers assigned to a high-flyer manager exhibit lateral moves that are 40% higher relative to workers assigned to another low-flyer. These lateral moves occur across all organizational margins: within the same team, across teams within the same function (the largest contributor), and across functions. Beyond frequency, lateral moves under high-flyer managers also involve larger task-content shifts, with cumulative task distance (measured using O*NET cognitive, routine, and social task dimensions via angular separation) becoming statistically distinguishable from zero approximately seven quarters after the transition.

Q: What is the wage effect of gaining a high-flyer manager and when does it materialize? A: Workers who transition from a low-flyer to a high-flyer manager earn a salary 13% higher than workers who transition to another low-flyer, measured seven years after the transition event. The divergence begins only after the transition date, consistent with the pre-event parallel trends assumption, and accumulates gradually rather than appearing as an immediate jump.

Q: Does the wage gain reflect genuine productivity improvement or simply managerial favoritism in pay decisions? A: The author uses an independent sales bonus series — based on monthly targets set by supply chain demand planning teams, not by managers — for 5,604 field sales workers in 15 countries from 2018 to 2021. Three years after gaining a high-flyer manager, workers’ sales productivity increases by 0.347 standard deviations. This confirms that pay gains correspond to actual productivity improvement rather than inflated ratings for unchanged performance.

Q: How much of the wage gain is attributable to the lateral reallocation channel specifically? A: A mediation analysis attributes 64% of the 13% salary gain to lateral job changes. The author cautions that this is a lower bound because the mediation excludes vertical transfers (which mechanically raise salary) and does not capture gains for workers who remain in their current job because it represents a good match rather than requiring reallocation.

Q: Are the effects symmetric — does losing a high-flyer manager reverse the gains? A: No. Comparing workers who transition from a high-flyer to a low-flyer manager (HtoL) against workers who transition from a high-flyer to another high-flyer (HtoH) reveals no corresponding negative effects. The gains from a single prior exposure to a high-flyer manager are persistent and are not undone by a subsequent low-quality manager. The author interprets this as evidence that a good match, once created, endures independently of the manager who created it.

Q: Does gaining a high-flyer manager raise the rate of worker exit from the firm? A: No. There is no statistically detectable effect on either voluntary exits (quits) or involuntary exits (layoffs), with null results that are not masked by heterogeneity across high- and low-performing workers. This rules out the interpretation that high-flyer managers improve measured outcomes of retained workers by selecting out underperformers.

Q: Do workers move into roles connected to their high-flyer manager’s prior network or follow their manager when the manager moves? A: No. There is no evidence that workers move into roles connected to the high-flyer manager’s prior colleagues; if anything, subordinates of high-flyer managers are less likely to make such moves. Workers also do not follow their high-flyer managers when those managers subsequently rotate to a different team. These findings rule out favoritism, social network access, and information-advantage explanations as primary drivers.

Q: How does the paper rule out on-the-job teaching (human capital transmission) as the primary mechanism? A: If high-flyer managers improved worker outcomes primarily by teaching workers to be more productive in their current job, the prediction would be reduced lateral mobility (workers become too productive to leave their current role). The observed pattern — substantially higher rates of lateral reallocation under high-flyer managers — is the opposite of this prediction, making teaching as the dominant channel unlikely.

Q: What does the manager behavior evidence show about how high flyers spend their time? A: Time-use data from a random sample of approximately 600 WL2 managers in 2019 show that high-flyer managers spend 19% more time in one-on-one meetings with subordinates and engage more in communication and multitasking activities relative to low-flyer managers. Their skill profiles also differ: high flyers are more likely to have strengths in strategy and talent management rather than project management, consistent with a more coordination-intensive and people-development-oriented style.

Q: What heterogeneity is there in who benefits from high-flyer managers? A: Effects are larger when managers and workers are in the same physical office (proximity facilitates talent assessment), when the organizational unit has a more diverse set of job roles (more matching opportunities), and for younger workers who are still discovering their comparative advantages. Critically, benefits are not concentrated among high-baseline performers: workers with low initial pay growth experience gains comparable to those of high performers, suggesting high-flyer managers uncover and deploy hidden talent broadly rather than accelerating only already-visible stars.

Q: Does high-flyer management aggregate to establishment-level productivity? A: Yes. Establishments where a higher share of workers are supervised by high-flyer managers show higher output per worker (tons per FTE) and lower operational costs per unit of output (operational costs per ton), measured using establishment-year data across approximately 150 sites globally over 2019-2021. This is consistent with the individual-level allocation mechanism producing aggregate productivity gains.

Q: What are the organizational design implications of the asymmetric effects? A: Because the gains from a single exposure to a high-flyer manager persist even after a subsequent manager downgrade, firms do not need each worker to be continuously supervised by a high-flyer. It is sufficient to rotate high-flyer managers across teams so that each worker receives at least one exposure. This makes the allocation mechanism resource-neutral relative to hiring, firing, or formal training programs.

High flyer (paper’s definition): A manager who achieved the first managerial work level (WL2) at the firm by age 30 — a time-invariant, ex ante classification representing the firm’s revealed-preference assessment of leadership potential, validated against subsequent salary growth, promotion probability, performance ratings, and subordinate feedback. Constitutes 26.2% of managers in the sample.

Internal labor market (paper’s usage): The system within the firm through which workers are allocated to jobs via lateral transfers and vertical promotions, mediated by managers rather than by external price mechanisms; the institutional context within which manager-worker matching produces wage growth and productivity gains.

Lateral transfer (paper’s usage): A horizontal reallocation of a worker to a different job title, team, subfunction, or function at the same work level, as distinct from a vertical promotion. Captured monthly in personnel records; operationalized as moves involving changes in task content measured by O*NET task distances.

Task distance (paper’s usage): The angular separation between origin and destination occupations across three O*NET task dimensions (cognitive, routine, and social intensity), ranging from zero (identical task profiles) to one (completely distinct profiles), used to characterize the substantive scope of lateral moves induced by high-flyer managers.

Manager rotation (paper’s usage): The firm’s longstanding policy of reassigning WL2 managers laterally across teams within a subfunction, designed to broaden managerial experience and screen for promotion to WL3; treated in the empirical strategy as generating plausibly exogenous variation in the manager type each worker encounters.

Allocation mechanism (paper’s usage): The process by which managers discover workers’ specific skills and match them to specialized jobs inside the firm, operating through lateral reallocation rather than through hiring, firing, or on-the-job training; identified in the paper as the primary channel through which high-flyer managers generate persistent wage and productivity gains.

Asymmetric persistence (paper’s usage): The empirical pattern in which the gains from gaining a high-flyer manager are large and durable, while losing a high-flyer manager (transitioning to a low-flyer) produces no corresponding negative effects on the outcomes of previously well-matched workers, implying that good matches, once formed, survive a change in manager quality.

Marginal Returns to Public Universities

Mon, 01 Jan 0001 00:00:00 +0000

This paper asks whether enrolling in an American public university generates positive net returns for marginal students — those who barely qualify for admission — and whether those returns justify public expenditures. The question is policy-relevant because marginal students have weak academic preparation, face high dropout risk, and the net returns to expanding admission margins are theoretically ambiguous.

The author assembles administrative records spanning all 35 public universities in Texas, covering the universe of Texas public high school graduates from 2004–2014 (approximately 2.7 million students). Texas public universities collectively enroll over 10 percent of all American public university students. The data link high school records (test scores, demographics, coursework, attendance, disciplinary infractions) to college application and admission records, postsecondary enrollment and degree completion records, financial aid packages, institutional expenditure data from IPEDS, and quarterly earnings records from the Texas Workforce Commission unemployment insurance system.

The identification strategy exploits hundreds of decentralized SAT/ACT score cutoffs in university admissions — varying across schools and application years — that generate sharp discontinuities in admission probability. A fuzzy regression discontinuity design compares applicants just above versus just below each cutoff. On average, crossing a cutoff raises the probability of admission by 27 percentage points and the probability of enrolling at the target university by 15 percentage points. Density tests and pre-college covariate balance validate the smoothness assumptions. The typical cutoff complier is more disadvantaged than the average college applicant but comparable to the average Texas high school graduate.

Roughly half of cutoff compliers would fall back to another, typically less selective, four-year institution if rejected; 43 percent would fall back to a two-year community college; and only about 6 percent would forgo higher education entirely. The pooled estimates therefore blend intensive-margin effects (more selective versus less selective four-year college) with extensive-margin effects (four-year college versus community college or no college).

Main causal findings for enrollment compliers: the typical marginally admitted student completes approximately one additional year of credits in the four-year sector and becomes 12 percentage points more likely to ever earn a bachelor’s degree from any institution. About half of the additional four-year credits are offset by 15 fewer credits in the two-year sector, and associate degree or certificate completion falls by 7 percentage points. All bachelor’s degree gains are in non-STEM fields; STEM degree completion shows no detectable increase. Compliers become about 3 percentage points more likely to hold a graduate degree by 10 years out.

On earnings, admitted compliers earn less than rejected counterparts in the first five years due to continued enrollment. Year six is the crossover point; by years 8–12, compliers earn a stable 8.6 percent earnings premium in log terms (8.2 percent in dollar ratio terms, representing a LATE of $3,339 against an untreated complier mean of $40,829), with earnings ranks rising approximately 4 percentiles from a base near the 50th percentile.

Marginally admitted students pay no additional net tuition on average: $4,600 in additional gross tuition is nearly fully offset by grant aid, though they take on $5,300 more in student loans. Society incurs approximately $10,000 in additional educational expenditures per complier. Internal rates of return are 26 percent for students, 16 percent for society, and 7 percent for the government budget. At a 3 percent discount rate, the lifetime net present value of enrolling the typical marginal applicant is approximately $80,000 — $70,000 accruing to the student and $10,000 to taxpayers.

Earnings gains are similar across institutions of varying selectivity, but significantly smaller for low-income compliers, who spend more time enrolled, complete fewer degrees, and major in less lucrative fields. A bounding method shows that extensive-margin compliers (those who would otherwise not attend any four-year college) experience larger effects than intensive-margin compliers.

Q: What is the core research question and why is credible evidence scarce? A: The paper asks whether enrolling marginal students in American public universities generates positive net returns — private, social, and fiscal — and what drives heterogeneity in those returns. Credible evidence is scarce because most existing work is correlational and fails to account for selection bias: individuals with more college education may have had pre-existing advantages, confounding college’s causal effect with systematic sorting into it. Even if average returns are positive, the policy-relevant question is whether the marginal student — who has weak preparation and high dropout risk — represents a good investment.

Q: What is the regression discontinuity design, and what does the first stage look like? A: The author infers hundreds of decentralized SAT/ACT score cutoffs across approximately 700 application cells (combinations of university, year, GPA quartile, and test type) by searching for the score value with the largest discontinuity in admission and enrollment within each cell. This procedure delivers a superconsistent estimator of each cell’s true cutoff. Pooled across all cells, crossing a cutoff raises the probability of admission by 27 percentage points and the probability of enrollment at the target university by a precisely estimated 15 percentage points. The density of applicants and a rich set of pre-college characteristics run smoothly through the cutoffs, supporting the exclusion restriction.

Q: Who are the cutoff compliers, and are they representative of any broader population? A: Compliers — applicants who enroll in the target university if and only if they barely cross its cutoff — comprise approximately 15 percent of marginal applicants. In observable characteristics, compliers are roughly representative of the broader population of marginal applicants at the cutoff. They are significantly more disadvantaged than the average public university applicant, but broadly comparable to the average Texas public high school graduate in terms of academic preparation and family income.

Q: What are the next-best alternatives for marginal applicants who are rejected? A: Approximately 47 percent of compliers would fall back to another Texas four-year college (mostly public), 43 percent to a two-year community college, and approximately 9 percent would not enroll in any Texas institution. National Student Clearinghouse data for the 2008–2014 cohorts confirm that only 4 percent of untreated compliers attend a college outside the THECB universe, meaning approximately 6 percent of all compliers truly forgo higher education altogether if rejected. The empirically relevant extensive margin is therefore between the four-year sector and the two-year sector, not between college and no college.

Q: How does cutoff crossing change the institutional characteristics a complier experiences? A: Compliers are propelled into substantially better-resourced environments: the average math test score of college peers rises by half a standard deviation; peers are 12 percentage points less likely to have been low-income; gross tuition rises by $2,400 (a 42 percent increase over the untreated complier mean of $5,700); educational spending per student rises by $3,200 (43 percent over the untreated mean); peers’ 10-year BA completion rate rises by 28 percentage points; and peer mean earnings 8–12 years after college entry are $6,700 higher.

Q: What are the educational attainment effects? A: Cutoff crossing causes compliers to complete approximately 28 additional credits at any four-year institution (roughly one full year of a four-year program) and increases the probability of ever earning a bachelor’s degree by 12 percentage points, raising the completion rate from approximately 40 percent to just above 50 percent. About 15 fewer two-year sector credits are offset against the four-year gains, and associate degree or certificate completion falls by 7 percentage points. All bachelor’s degree gains are in non-STEM fields; there is no detectable increase in STEM degrees. Graduate degree completion rises by approximately 3 percentage points by 10 years out.

Q: What is the earnings trajectory, and when does the premium materialize? A: Admitted compliers earn less than rejected counterparts in the first five years after application because they remain enrolled longer. Year six is the crossover point. By years 8–12, the earnings premium stabilizes at approximately 8.6 percent in log terms and 8.2 percent in dollar ratio terms (a LATE of $3,339 against an untreated complier mean of $40,829). Earnings rank rises by approximately 4 percentiles from a base near the 50th percentile. These results are robust across sandwich earnings, all-quarters-with-earnings, and zero-imputed specifications.

Q: What does the cost-benefit analysis show? A: Marginally admitted students pay no additional net tuition on average: $4,600 in additional gross tuition is nearly fully offset by additional grant aid. They do borrow $5,300 more in student loans, likely financing higher room, board, and consumption costs at four-year colleges. From society’s perspective, compliers generate approximately $10,000 in additional educational expenditures. Cumulative undiscounted earnings benefits surpass costs after 8 years for students, 11 years for society, and 19 years for taxpayers. At a 3 percent discount rate, the lifetime net present value is approximately $80,000 total — $70,000 accruing to the student and $10,000 to taxpayers — with internal rates of return of 26 percent for students, 16 percent for society, and 7 percent for the government budget.

Q: Does selectivity of the admitting institution predict larger earnings returns? A: No. Compliers at more selective institutions experience substantially larger increases in peer quality than those at less selective institutions, but they are also less likely to be on the extensive margin of four-year enrollment and experience smaller BA attainment gains. These factors roughly offset, producing no systematic difference in earnings gains across institutions of varying selectivity. More selective institutions also impose no additional cumulative cost on society, while compliers actually pay slightly less in additional net tuition at more selective schools.

Q: How does the commonly used measure of college value-added (mean peer earnings) compare to actual complier returns? A: Mean peer earnings overpredicts actual value-added for marginal students by a factor of two: compliers attend an institution with $6,700 higher average peer earnings as a result of admission but gain only $3,300 themselves. The measure also overpredicts the earnings return to selectivity by a factor of three: a 100-SAT-point increase in target school selectivity predicts $3,000 higher peer earnings but only a statistically insignificant $900 higher gain in the complier’s own earnings.

Q: How do earnings returns differ by family income? A: Compliers from low-income families experience significantly smaller earnings gains compared to higher-income compliers. The gap is not explained by differential changes in college quality induced by admission. Instead, low-income compliers gain fewer degrees despite spending more time in college and major in less lucrative fields, consistent with related findings in the literature on family income gaps in degree completion and major choice.

Q: How do earnings returns differ by gender and by race? A: Female and male compliers eventually earn similar log earnings and earnings rank gains, but women reach their gains more quickly — likely because men take longer to finish college. White and Asian compliers experience similar earnings gains and BA completion improvements as Black and Hispanic compliers, despite white and Asian students experiencing larger increases in college selectivity and spending per student as a result of admission.

Q: What is the method for separating intensive- and extensive-margin effects? A: The two complier types are not directly distinguishable in the data. The author first uses an endogenous but strong stratification variable — having at least one other Texas public university admission offer — to identify some mean potential outcomes for each type. He then imposes an empirically-informed rank assumption to bound the remaining unknown mean potential outcomes, delivering tightly informative upper and lower bounds on each margin’s effects without requiring full nonparametric identification. The results show that pooled effects are driven by larger returns for extensive-margin compliers who would not have attended any four-year college, with smaller contributions from intensive-margin compliers shifting between four-year institutions.

Q: How do this paper’s earnings estimates compare to prior studies, and what explains the differences? A: This paper’s 8 percent earnings gain is smaller than the 17–26 percent reported in prior studies (Zimmerman 2014: 22%; Kozakowski 2023: 26%; Smith, Goodman, and Hurwitz 2025: 17%; Bleemer 2024: 21%; Hoekstra 2009: 20%). The differences are likely explained by the much larger educational attainment and institutional quality gains induced by those studies’ natural experiments: in Zimmerman (2014), enrollment compliers gain roughly three additional years of four-year education versus one year in this paper; in Bleemer (2024), compliers experience roughly $30,000 more in institutional spending per student versus approximately $3,000 in this paper.

Q: What are the scope conditions for these results? A: The results pertain to marginal applicants to Texas public universities (excluding UT-Austin, which uses holistic admission with no detectable SAT/ACT cutoffs) from the 2004–2014 high school graduation cohorts. The identified effects are local average treatment effects for compliers — applicants who would enroll in the target university if and only if they barely crossed its admission cutoff — and do not represent effects for always-takers or infra-marginal students. Earnings are measured only for Texas-based workers covered by the state unemployment insurance system, which captures an estimated 90 percent of the civilian labor force.

Cutoff complier: An applicant who enrolls in their target university if and only if their SAT/ACT score barely exceeds that university’s admission cutoff. Compliers are the population whose behavior — and thus whose treatment effects — are identified by the fuzzy RD design. They comprise approximately 15 percent of marginal applicants and are more disadvantaged than the average public university applicant but broadly comparable to the average high school graduate.

Extensive versus intensive margin: The extensive margin refers to the contrast between attending any four-year college versus falling back to a two-year community college or no college. The intensive margin refers to the contrast between attending a more selective versus a less selective four-year institution. Approximately half of cutoff compliers are on each margin; the paper treats them as economically distinct parameters requiring separate identification.

Fuzzy regression discontinuity (RD) design: An identification strategy that uses the discontinuous jump in admission probability at a test score cutoff as an instrument for enrollment, recovering the LATE for compliers via the ratio of the reduced-form discontinuity in outcomes to the first-stage discontinuity in enrollment. “Fuzzy” refers to the fact that crossing the cutoff changes admission and enrollment probabilities with a discrete jump rather than with certainty.

Internal rate of return (IRR): The discount rate at which the net present value of an investment equals zero — here, the discount rate equating the discounted stream of earnings benefits to the discounted stream of costs. The paper estimates IRRs separately for students (26 percent), society (16 percent), and the government budget (7 percent), reflecting different cost and benefit definitions from each perspective.

Rank assumption (bounding method): An empirically-informed assumption about the ordering of mean potential outcomes across latent complier types (extensive vs. intensive margin) that, combined with partial identification from a strong endogenous stratification variable, yields tight upper and lower bounds on each margin’s causal effects without requiring full nonparametric identification.

Net tuition: Gross tuition charges minus grant aid. For the typical marginal complier, gross tuition rises by $4,600 but is nearly fully offset by additional grant aid, yielding approximately zero additional net tuition cost — meaning the private financial cost of attending a public university for marginal students is effectively zero on net, though they take on $5,300 more in student loans to finance room, board, and consumption.

Sandwich earnings measure: A procedure applied to quarterly state earnings data that retains only quarters with positive earnings sandwiched between other quarters with positive earnings, discarding high-variance transition quarters between employment spells. Annualized by multiplying the quarterly average by four; used to reduce noise from entry and exit transitions in administrative earnings records.

Measuring and Mitigating Racial Disparities in Tax Audits

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. Do Black taxpayers face higher IRS audit rates than non-Black taxpayers, despite race-blind audit selection? And if so, why — and what would mitigation look like?

Data and Methodology. The authors use comprehensive administrative microdata covering approximately 148 million individual income tax returns and 780,627 operational audits for tax year 2014, supplemented with 71,878 research audits from the IRS National Research Program (NRP) pooled over 2010-2014. Because neither the researchers nor the IRS observe taxpayer race, the authors employ Bayesian Improved First Name Surname Geocoding (BIFSG), which imputes the probability that a taxpayer is Black from first name, surname, and Census Block Group. They develop a novel partial identification strategy: two estimators (a probabilistic estimator and a linear estimator) that, under conditions verified using a matched North Carolina voter-registration dataset containing self-reported race, asymptotically bound the true racial audit disparity from below and above respectively. To address the selective labels problem — underreporting is observable only for audited returns — the authors combine operational audit data with NRP random-sample audits to simulate counterfactual audit selection algorithms.

Main Findings.

Magnitude of the disparity. The probabilistic estimator implies a racial audit disparity of 0.81 percentage points; the linear estimator implies 1.34 percentage points. Against a base audit rate of 0.54% for the overall U.S. population in 2014, these bounds imply that Black taxpayers are audited at between 2.9 and 4.7 times the rate of non-Black taxpayers.

Role of the EITC. The disparity is concentrated among EITC claimants. The estimated disparity within the EITC population is 1.96 to 2.90 percentage points, compared to only 0.10 to 0.18 percentage points among non-EITC claimants. In relative terms, Black EITC claimants are audited at 2.9 to 4.4 times the rate of non-Black EITC claimants. A formal decomposition attributes 70-73% of the overall disparity to higher audit rates among Black EITC claimants, 20-21% to racial differences in EITC claiming rates, and 7-8% to differential audit rates among non-EITC filers. Within EITC claimants, 78.5% of the observed audit disparity is attributable to the Dependent Database (DDb) program.

Source of the disparity — algorithmic objective. Using counterfactual audit selection algorithms estimated on NRP data, the authors find that allocating EITC audits to maximize detected total underreporting (from any source) would produce audit rates of 0.74% for Black EITC claimants versus 1.63% for non-Black EITC claimants — reversing the disparity. In contrast, the status quo, which prioritizes detecting overclaimed refundable credits, yields 3.00% for Black claimants versus 1.04% for non-Black claimants. The primary driver is a difference in the types of noncompliance that are more prevalent by race: dependent-claiming errors are more common among Black EITC claimants (dependent error rate of 26.6% vs. 16.3% for non-Black), while the highest underreporting via business income underreporting is disproportionately concentrated among non-Black EITC claimants. An algorithm focused on refundable credit overclaims implicitly targets dependent errors and therefore selects Black taxpayers at higher rates.

Prediction model bias. Even conditional on the refundable-credit objective, the status quo disparity (1.96 p.p.) exceeds the disparity that would arise under an oracle that uses actual rather than predicted refundable credit overclaims (1.08 p.p.), suggesting that prediction errors are unevenly distributed by race. The refundable credit prediction algorithm generates a disparity of 1.75 p.p., approximately 60% larger than the oracle. The authors find suggestive evidence of missingness in birth certificate data (paternal information is disproportionately missing for children claimed on Black taxpayers’ returns) and differential predictive accuracy in the DDb risk score across race.

Operational consequences. Switching the objective from refundable credit overclaims to total underreporting would shift the composition of audited returns from predominantly dependent-eligibility issues (80% of refundable credit oracle-selected returns contain a dependent error) toward business income (86% of total-underreporting oracle-selected returns have business income underreporting). EITC returns with substantial business income (gross receipts above $25,000) cost on average $369.70 to audit versus $23.09 for other EITC returns. Holding the audit rate fixed, the switch would raise average examination costs by nearly an order of magnitude, while also increasing detected underreporting (mean adjustment of $22,578 per return under the total underreporting oracle versus $9,595 under the refundable credit oracle).

Scope Conditions. Results pertain primarily to tax year 2014. The paper finds similar patterns for tax years 2010, 2012, 2016, and 2018. The analysis covers Black versus non-Black taxpayers; disparities for other racial and ethnic groups are not the focus. The selective labels identification strategy relies on the NRP random-audit sample and the bounding conditions verified in the North Carolina matched data.

In depth

Q1. Why can’t the disparity be attributed simply to Black taxpayers being more likely to claim the EITC, combined with EITC claimants facing higher audit rates generally?

The authors test this directly by estimating racial audit disparities separately within EITC claimants and non-claimants. If differential EITC claiming rates were the full explanation, the within-EITC disparity would be close to zero. Instead, the disparity among EITC claimants (1.96-2.90 p.p.) is larger in absolute terms than the overall disparity (0.81-1.34 p.p.), indicating that Black EITC claimants face substantially higher audit rates than non-Black EITC claimants even holding EITC claimant status fixed. The formal decomposition attributes 70-73% of the overall disparity to differential audit rates within the EITC claimant population, not to differential claiming rates across the population.

Q2. How does the partial identification strategy work, and what are its key identifying assumptions?

The authors derive two estimators of the racial audit disparity that use BIFSG-imputed race probabilities rather than observed race. The probabilistic estimator weights each taxpayer’s contribution by their estimated probability of being Black; it is downward-biased when there is a positive residual covariance between audits and true race after conditioning on imputed race (E[Cov(Y,B|b)] > 0). The linear estimator regresses audit status on imputed race probability; it is upward-biased when there is a positive residual covariance between audits and imputed race after conditioning on true race (E[Cov(Y,b|B)] > 0). When both covariance terms are positive, the probabilistic and linear estimates bound the true disparity from below and above. The authors verify both conditions are positive and statistically significant (p < 0.01) in the matched North Carolina dataset, for the full population and the EITC population specifically.

Q3. Does the racial audit disparity within EITC claimants disappear when comparing taxpayers with similar levels of underreporting?

No. The authors use NRP data to estimate audit rates by race within each underreporting decile among EITC claimants. Within every decile of the underreporting distribution, the estimated audit rate for Black taxpayers exceeds that for non-Black taxpayers. An oracle algorithm that selects returns in descending order of actual underreporting produces an audit rate of 0.74% for Black EITC claimants and 1.63% for non-Black EITC claimants — the opposite of the status quo pattern (3.00% for Black, 1.04% for non-Black). This rules out total-dollar underreporting as the primary driver of the observed disparity.

Q4. Why does focusing audit selection on refundable credit overclaims specifically lead to higher audit rates for Black taxpayers?

Two mechanisms operate simultaneously. First, EITC eligibility is linked to children, so detecting erroneously claimed dependents generates large refundable credit adjustments. The dependent error rate is higher among Black EITC claimants than non-Black EITC claimants (26.6% vs. 16.3% in the probabilistic estimate, or 30.8% vs. 15.4% in the linear estimate). Second, the highest-dollar noncompliance via underreported business income is disproportionately concentrated among non-Black EITC claimants: among EITC claimants in the top 1% of business income underreporting, the probabilistic estimate shows 0.05% are Black compared to 0.21% non-Black. An algorithm aimed at refundable credit overclaims implicitly targets dependent errors and therefore selects Black taxpayers at higher rates; one aimed at total underreporting would prioritize business income underreporting instead and therefore select non-Black taxpayers at higher rates.

Q5. How do the simulated algorithms compare to the actual IRS algorithms?

The authors cannot directly replicate the IRS’s confidential DDb algorithm, but they provide three pieces of evidence that their refundable credit prediction algorithm is a reasonable proxy. First, public governmental documents describe DDb’s stated goal as identifying taxpayers who do not meet refundable credit eligibility requirements. Second, when selecting audits based on predicted refundable credit overclaims using largely the same features available to IRS, the authors generate a disparity (1.75 p.p.) close to the status quo disparity (1.96 p.p.). Third, operational audits of EITC returns are strongly associated with their predicted refundable credit overclaims measure but show a much weaker association with predicted total underreporting.

Q6. What does the status quo disparity exceeding the refundable credit oracle disparity reveal about prediction model design?

The status quo disparity (1.96 p.p.) is approximately 80% larger than the disparity that would arise if the IRS were perfectly informed about actual refundable credit overclaims and selected accordingly (oracle disparity: 1.08 p.p.). The refundable credit prediction algorithm generates a disparity of 1.75 p.p., approximately 60% larger than the oracle. This gap between the oracle and prediction disparity is consistent with prediction errors being distributed unevenly by race. The authors find that birth certificates of children claimed on Black taxpayers’ returns are substantially more likely to be missing paternal identity information, which may reduce the predictive accuracy of the DDb model for this population. They provide suggestive evidence that modifying the predictive features used could reduce the disparity without substantially degrading credit overclaim detection.

Q7. What are the downstream operational consequences of switching the algorithmic objective?

Switching from refundable credit overclaims to total underreporting would shift audited issues from dependent eligibility (80% of refundable credit oracle-selected returns have a dependent error) toward business income (86% of total underreporting oracle-selected returns have business income underreporting). Auditing business income returns is substantially more resource-intensive: $369.70 per return on average for returns with gross receipts above $25,000, versus $23.09 for other EITC returns. Holding the current EITC audit rate fixed, the share of audited returns with substantial business income would rise from 3% to 93%, raising total examination costs by nearly an order of magnitude. However, because total detected underreporting per audited return would also rise substantially (mean of $22,578 vs. $9,595), the increase in detected noncompliance would exceed the increase in audit costs, and the qualitative pattern persists even when accounting for higher per-return costs.

Q8. Is the disparity consistent across years, and is it driven by a particular audit type?

The authors find comparable audit disparities for tax years 2010, 2012, 2016, and 2018, confirming the 2014 results are not year-specific. The disparity is concentrated in correspondence audits: the estimated disparity in correspondence audit rates is 0.804-1.328 p.p. for the full population, while the disparity in field/office audit rates is only 0.010-0.016 p.p. The disparity is present in both pre-refund and post-refund audits, though pre-refund audits show a larger disparity even among correspondence audits alone. Among EITC claimants, the correspondence audit channel is nearly entirely responsible for the group-level disparity.

Q9. What heterogeneity exists within EITC claimants?

The disparity is especially pronounced among unmarried male EITC claimants with dependents: among this subgroup, the audit rate for Black men exceeds the audit rate for non-Black men by more than 4 percentage points, and both are an order of magnitude above the overall U.S. population audit rate. Disparities are smaller among joint filers, unmarried women, and unmarried men without dependents, though the ratio of Black to non-Black audit rates remains substantial across all subgroups. The concentration of the disparity among unmarried men with dependents is consistent with the role of dependent-claiming errors, which are more likely to arise in family structures characterized by nonmarital cohabitation — a pattern more prevalent among Black Americans due to lower marriage rates.

Q10. Can the disparity be attributed to disparate treatment — i.e., race-conscious selection?

The authors rule out disparate treatment for the EITC population. The DDb audit selection process for EITC returns is automated (no manual review), and IRS does not use race or geography as an input into audit selection. The disparity is therefore the product of disparate impact: race-neutral selection criteria interact with racially correlated patterns of tax return characteristics to produce differential audit rates. For higher-income non-EITC taxpayers, where audit selection may involve human classifiers, the authors cannot rule out disparate treatment.

Key Concepts

Audit Disparity (D). Defined in the paper as D = E[Y|B=1] - E[Y|B=0], the difference in audit rates between Black taxpayers (B=1) and non-Black taxpayers (B=0). This is a group-level difference in selection rates, not conditional on any other characteristic, and is the primary estimand throughout.

Probabilistic Disparity Estimator. An estimator that calculates group-specific audit rates by weighting each taxpayer’s contribution by their BIFSG-imputed probability of being Black (or non-Black). It is shown to be downward-biased when E[Cov(Y,B|b)] > 0, i.e., when there is residual positive association between true race and audits after conditioning on imputed race.

Linear Disparity Estimator. An estimator based on regressing audit status (Y) on BIFSG-imputed race probability (b). It is shown to be upward-biased when E[Cov(Y,b|B)] > 0, i.e., when imputed race probability predicts audits even after conditioning on true race. Together, the probabilistic and linear estimators form bounds on the true disparity under conditions verified empirically.

BIFSG (Bayesian Improved First Name Surname Geocoding). A probabilistic race imputation method that uses Bayes rule under a conditional independence assumption (first name, surname, and geography are independent given race) to compute Pr[Black | first name, surname, Census Block Group]. Applied here to all 148 million tax returns; calibrated and validated against matched North Carolina voter registration data with self-reported race.

Selective Labels Problem. The problem that noncompliance (underreporting) is observed only for returns selected for audit, not for the full filing population. In this paper it means the IRS cannot directly observe the underreporting distribution for unaudited returns. The authors address this using NRP random-audit data, which allows estimation of the unaudited underreporting distribution and construction of counterfactual selection algorithms.

Algorithmic Objective. The paper distinguishes between (1) the prediction component of audit selection — which model to use to forecast noncompliance — and (2) the objective component — what type of noncompliance to predict and pursue (overclaimed refundable credits versus total underreporting from any source). The paper finds that the objective, not just prediction error, is an independent driver of the racial audit disparity.

Dependent Database (DDb) Program. The IRS’s primary EITC audit selection program, responsible for approximately 75% of audited EITC returns in 2014. DDb flags returns based on rules, heuristics, and proprietary risk scores, with the stated goal of identifying taxpayers who do not meet refundable credit eligibility requirements. Selection through DDb is fully automated, without human classifier review.

National Research Program (NRP). A stratified random sample audit program through which the IRS conducts near-line-by-line examinations of a small fraction of the filing population each year (approximately 2% of audited returns in 2014). The paper pools 71,878 NRP audits from 2010-2014 to identify the distribution of underreporting in the full EITC filing population and to estimate counterfactual selection algorithms.

Micro MPCs and Macro Counterfactuals: The Case of the 2008 Rebates

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research question. Do the high marginal propensities to consume (MPCs) estimated in the leading household studies of the 2008 U.S. tax rebates—particularly Parker et al. (2013), which found MPCs of 50–90 percent within three months—imply plausible macroeconomic counterfactuals? And if not, what combination of micro-level bias corrections and general equilibrium forces reconciles the micro evidence with aggregate data?

Setting. The 2008 Economic Stimulus Act distributed approximately $100 billion in tax rebates, totaling eleven percent of January 2008 monthly disposable income. Among the 85 percent of households receiving a check, the average amount was $1,000. Rebates were distributed primarily from April through July 2008, with nearly half delivered in May alone. The timing of receipt was determined by the last two digits of Social Security numbers, providing quasi-random variation exploited by the household-level literature.

Methodology. The paper proceeds in two halves. In the first, the authors construct macro counterfactuals by calibrating a standard medium-scale two-good, two-agent New Keynesian (TANK) model with the micro MPCs from the literature and simulating what aggregate consumption would have been absent the rebate. The model contains life-cycle permanent income households and hand-to-mouth households whose dynamic spending propensities are calibrated directly to match the household-level estimates. General equilibrium effects—including Keynesian income multipliers, real interest rate movements, and changes in the relative price of durable goods—are incorporated. Counterfactual consumption paths are constructed by subtracting model-simulated deviations from steady state from actual NIPA consumption data.

In the second half, the authors revisit both the micro estimates and the macro model. On the micro side, they identify three upward biases in standard two-way fixed effects (TWFE) estimates applied to CEX data: (1) omitted variable bias from excluding the lagged rebate indicator; (2) “forbidden comparisons” bias arising from comparing cohorts with heterogeneous treatment effects, following Borusyak et al. (2022) and Sun and Abraham (2020); and (3) a rebate reporting bias in which households are systematically more likely to report receiving the rebate in the month that coincides with large expenditure increases, causing spurious positive correlation between reported receipt and contemporaneous spending. On the macro side, the baseline model is modified to incorporate an upward-sloping supply curve for durable goods (calibrated to a supply elasticity of 5, midway between House and Shapiro (2008) and Goolsbee (1998)), replacing the baseline assumption of frictionless conversion between nondurable and durable intermediates.

Main findings with quantitative magnitudes.

Implausibility of baseline counterfactuals. When calibrated to Parker et al.’s (2013) micro MPC of 0.9, the baseline model implies that real PCE absent the rebate would have collapsed by 6.0 percent from April through July 2008—a decline exceeded historically only by the Covid-19 lockdowns. Even the more modest micro MPC of 0.5 implies a 2.7 percent three-month PCE decline, comparable only to the 1980 Volcker disinflation with credit controls. For motor vehicle expenditures, the counterfactual drops range from 38 percent (micro MPC = 0.3) to 67 percent (micro MPC = 0.9)—larger than any historical experience, including the 30 percent Covid decline. Contemporaneous professional forecasters (Federal Reserve Greenbooks, Survey of Professional Forecasters, Goldman Sachs) predicted at most small consumption declines in summer 2008. Even the authors’ own pessimistic forecast model—incorporating actual oil price paths and a Lehman Brothers bankruptcy dummy—implies that the cumulative difference between actual and forecast consumption attributable to the rebate was at most $20 billion out of $100 billion in rebates, for an implied GE-MPC of at most 0.2.

Bias correction in micro MPC estimates. Applying all three bias corrections to CEX data (the preferred specification with lagged rebate indicator, cohort-level treatment effects, and lagged expenditure controls), the estimated three-month MPC falls from 0.50 to 0.28 in the full sample and from 0.82 to 0.34 in the rebate-recipients-only sample, with both rounding to approximately 0.3. The Borusyak-Jaravel-Spiess (BJS) imputation method yields an MPC of 0.20 in the full sample and 0.37 in the rebate-only sample, consistent with the OLS corrections.

Composition of spending. In the preferred corrected specification, essentially all of the total expenditure MPC of 0.3 is accounted for by motor vehicle spending: the MPC on motor vehicles is 0.30 in the full sample and 0.26 in the rebate-only sample, while the MPC on all other expenditures is −0.02 (full sample) and 0.08 (rebate-only sample).

General equilibrium dampening via inelastic durable supply. In the model with a calibrated durable supply elasticity of 5, rebate-induced demand for motor vehicles raises the relative vehicle price by approximately 1.1 percent in July 2008. This price increase crowds out durable expenditure by optimizing households through intertemporal substitution. At the preferred micro MPC of 0.3, the general equilibrium MPC (GE-MPC) for total PCE is only 0.07, well below the 0.3 micro estimate. At a micro MPC of 0.5, the GE-MPC is 0.22. The combination of the bias-corrected micro MPC and dampening general equilibrium forces implies a general equilibrium consumption multiplier below 0.2 for the 2008 rebates.

Importance of durable goods composition for HANK models. A model that abstracts from durable goods and calibrates the full expenditure micro MPC to nondurable spending predicts a GE-MPC of 0.36 when the micro MPC is 0.30—five times larger than the 0.07 implied by the model with durable goods. This contrast illustrates that the distribution of spending across nondurable and durable goods is a key determinant of the aggregate fiscal multiplier, in addition to heterogeneity in wealth and income emphasized by the existing HANK literature.

In depth

Q1. What is the central empirical puzzle the paper addresses?

A. The leading household studies of the 2008 rebates estimate very high three-month MPCs (50–90 percent). When these estimates are plugged into a standard New Keynesian model to construct counterfactual consumption paths absent the rebate, the model implies that PCE would have collapsed by 2.7–6.0 percent from April through July 2008 and then sharply recovered just as Lehman Brothers failed in September. No contemporaneous forecaster or narrative evidence suggests such extreme, short-lived macroeconomic stress was present. The Lehman collapse itself caused only a 1.1 percent three-month PCE decline—smaller than all three counterfactual declines implied by micro MPCs of 0.3, 0.5, or 0.9.

Q2. What are the features of the TANK model used to construct the counterfactuals?

A. The model is a two-good (nondurable and durable), two-agent (optimizing life-cycle and hand-to-mouth) New Keynesian model calibrated at monthly frequency, building on Ramey (2021) and Galí et al. (2007). Intermediate goods can, in the baseline, be frictionlessly converted into either nondurable or durable goods (implying a fixed relative price of one). Durable goods (interpreted as motor vehicles) enter household utility, with optimizing households facing a Calvo-type adjustment friction motivated by Evans and Ramey (1992) calculation costs. The fraction of hand-to-mouth consumers and their dynamic propensities to spend are calibrated directly to match the micro MPC estimates from the household literature. The model incorporates a Calvo-style price-adjustment structure for nondurables, sticky wages set by unions, capital with adjustment costs and variable utilization, and an inertial monetary policy rule.

Q3. How does the model translate micro MPCs into macro counterfactuals, and why does it amplify rather than dampen the micro estimates in the baseline?

A. The model’s GE-MPC equals the micro MPC’s direct demand effect plus Keynesian income multiplier effects. Because the rebate is highly transitory, there is little movement in the real interest rate (the Phillips curve is flat and monetary policy is inertial), so the dominant general equilibrium force is the income multiplier. This amplifies, rather than dampens, the micro MPCs. As a result, the GE counterfactuals exhibit even sharper V-shapes than the pure micro counterfactuals.

Q4. What narrative and forecast evidence do the authors use to argue the baseline counterfactuals are implausible?

A. Contemporary forecasts from the Federal Reserve Greenbooks, the Survey of Professional Forecasters, and Goldman Sachs all predicted at most small consumption declines in summer 2008—Goldman Sachs forecast only −0.125 percent (not annualized) per quarter in Q2–Q3 2008. The authors also construct their own “pessimistic” time-series forecast that incorporates actual oil price paths (which rose from $98 to $140 per barrel by July 2008) and a Lehman Brothers bankruptcy dummy; even this forecast lies above all three model counterfactuals in summer 2008 and displays no V-shape. Furthermore, the cumulative difference between actual PCE and the pessimistic forecast over April–October 2008 totals only $20 billion—implying a GE-MPC of at most 0.2 even if the entire gap were attributed to the rebate.

Q5. What is the first bias in standard TWFE estimates of the MPC, and how large is its effect?

A. The first bias is omitted variable bias from excluding the lagged rebate indicator. In a first-differenced panel regression, lagged treatment enters the error term. Because current treatment reduces the probability of past treatment, current and lagged treatment are negatively correlated, and omitting the lag inflates the OLS estimate of the contemporaneous effect. Including a lagged rebate indicator reduces the contemporaneous spending response by $40 in the full CEX sample (from $470 to $434) and by approximately $237 in the rebate-only sample (from $764 to $527).

Q6. What is the “forbidden comparisons” bias and how is it corrected?

A. When treatment effects are heterogeneous across cohorts (e.g., the June rebate cohort has a larger MPC than the September cohort), standard homogeneous TWFE estimates use later-treated cohorts as control groups for earlier-treated cohorts even after accounting for average mean-reversion. Because the mean-reversion of the earlier (larger-effect) cohort is larger than that of the later cohort, this comparison is contaminated, inflating the estimate. The authors correct for this by allowing cohort-specific treatment effects, following Sun and Abraham (2020). This reduces the contemporaneous effect by a further $90 in the full sample; in the rebate-only sample the correction raises the estimate slightly (by $70) because later treatment effects are larger in that sample.

Q7. What is the rebate reporting bias and what mechanism underlies it?

A. The rebate reporting bias arises because households in the CEX are systematically more likely to report receiving the rebate in the interview month that coincides with high expenditure. Although the true timing of rebate checks is determined by Social Security number last-digits (and is thus random), the reported timing may reflect recall issues: households more readily remember and report receiving the rebate when it was accompanied by a large purchase. The empirical signature is a statistically significant negative effect of future rebate receipt on current expenditure (−$863 in the full sample, −$575 in the rebate-only sample at the 10% level), indicating that rebate reporters had unusually low spending in the period prior to reporting receipt. Controlling for lagged expenditure and income decile fixed effects corrects for this bias, reducing the three-month MPC in the full sample from 0.37 to 0.28.

Q8. What are the authors’ preferred bias-corrected MPC estimates, and how do they compare across specifications and estimators?

A. After correcting for all three biases (preferred specification, column 4 of Table 3), the implied three-month MPC is 0.28 in the full sample and 0.34 in the rebate-only sample, both approximately 0.3. The Borusyak-Jaravel-Spiess imputation method, which imposes weaker assumptions and overcomes the first two biases by construction, yields an MPC of 0.20 (full sample) and 0.37 (rebate-only sample), with an average consistent with the OLS-corrected estimates. Both methods point to an MPC around 0.3, substantially below the 0.5–0.9 range from the baseline Parker et al. (2013) approach.

Q9. How is almost all of the total expenditure MPC concentrated in motor vehicles?

A. After bias correction, the MPC on motor vehicles is 0.30 in the full sample and 0.26 in the rebate-only sample. The MPC on all other PCE is −0.02 (full sample) and 0.08 (rebate-only sample), neither statistically significant. This concentration in durables is consistent with Adams et al. (2009) and Aaronson et al. (2012), and is corroborated by CEX vehicle-expenditure data showing a car-purchase response concentrated in the three months surrounding receipt of the rebate.

Q10. How does introducing an upward-sloping supply curve for durable goods change the model’s general equilibrium predictions?

A. In the modified model, durable goods producers face a production externality (or fixed factor) that makes the short-run supply of motor vehicles upward-sloping, with supply elasticity calibrated to 5. When rebate recipients increase demand for motor vehicles, the relative price of motor vehicles rises by approximately 1.1 percent in July 2008 (consistent with the observed 1.5 percent spike in the BLS new vehicle price index relative to core CPI around the rebate distribution). This price increase induces optimizing households to intertemporally substitute away from durable goods. Because durable demand is highly price-elastic (long-run elasticity of −1 to −15 depending on the study), even a modest relative price increase generates substantial crowding out of durable expenditure by non-recipients.

Q11. What are the GE-MPC estimates in the modified model with less elastic durable supply, and how do they decompose?

A. At the preferred micro MPC of 0.3, the GE-MPC for total PCE is 0.07—general equilibrium forces dampen the micro effect. At micro MPC of 0.5, GE-MPC is 0.22 (modest dampening). At micro MPC of 0.9, the GE-MPC rises to 1.42 (amplification). Decomposing by good type at micro MPC of 0.3: the GE-MPC on motor vehicles is 0.09 and the GE-MPC on nondurables is −0.03. The dampening is concentrated almost entirely in durable expenditure.

Q12. How sensitive are the GE-MPC results to the calibration of durable demand elasticity?

A. The baseline calibration uses a long-run vehicle demand elasticity of −15, based on household-level evidence from Bachmann et al. (2021). When the authors instead use the lower-bound estimate of −6.4 from Baker et al. (2019), the GE-MPC at micro MPC of 0.3 rises from 0.07 to 0.12. Even at this lower demand elasticity there is substantial crowding out in general equilibrium, so the qualitative conclusion is robust.

Q13. Why does a nondurables-only model with the same overall MPC substantially overstate the fiscal multiplier?

A. When abstracting from durable goods and calibrating a nondurable MPC of 0.30 (to match the overall expenditure MPC), the model predicts a GE-MPC of 0.36—five times larger than the 0.07 from the two-good model. This occurs because nondurable demand is far less price-elastic than durable demand, and the nearly-flat Phillips curve makes nondurable supply very elastic, so there is no relative-price-driven crowding out channel. The comparison illustrates that the distribution of spending across nondurable and durable goods is a quantitatively important determinant of the fiscal multiplier, independent of the level of the MPC.

Q14. What evidence is provided that the control group in the household regressions is itself affected by the rebate in general equilibrium?

A. Figure 9 in the paper plots motor vehicle spending per household by rebate-receipt status using CEX data. When rebate recipients begin reporting receipt in June 2008, motor vehicle expenditure in the rebate group rises while simultaneously falling in the never-rebate group. This pattern is consistent with the model’s prediction that the rebate-induced rise in relative motor vehicle prices crowds out purchases by non-recipient households. This general equilibrium spillover means the difference-in-differences micro MPC estimate remains valid as a micro estimate (the symmetric crowding out does not affect the treated-versus-control difference), but the aggregate GE-MPC is less than the micro MPC.

Q15. How do the authors verify that their preferred corrected specification recovers true MPCs?

A. In Appendix C.6 the authors simulate household-level data from the modified Section 5 model and apply both the original Parker et al. (2013) specification (Equation 1) and their preferred corrected specification (Equation 5). The Parker et al. specification produces upward-biased MPC estimates in the simulated data, consistent with Kaplan and Violante’s (2014) theoretical argument. The preferred corrected specification recovers the true MPCs from the model, validating the correction methodology.

Key Concepts

GE-MPC (General Equilibrium Marginal Propensity to Consume). The paper’s term for the aggregate increase in total consumer spending per dollar of tax rebate, incorporating both the direct micro-level demand effect of the rebate on hand-to-mouth households’ consumption and the induced macroeconomic income effects from Keynesian multipliers and relative price changes. Distinct from the micro MPC, which captures only the household-level spending response before any general equilibrium feedbacks.

Micro MPC. The causal effect of receiving a temporary lump-sum transfer on a household’s own consumer expenditure, expressed as a fraction of the transfer amount, estimated from household panel data via difference-in-differences event studies. In the paper’s usage, this is a partial equilibrium concept that excludes any impact of the policy on prices, wages, or other households’ incomes.

Forbidden comparisons bias. A form of bias in two-way fixed effects event study estimates that arises when treatment effects are heterogeneous across cohorts and later-treated units are used as control groups for earlier-treated units whose outcomes are still reverting after treatment. Named and formalized in Borusyak and Jaravel (2017) and Borusyak et al. (2022); in this paper it manifests because cohorts receiving rebates in June have systematically larger spending responses than those receiving in September, so using September recipients as a “clean” control for June reversal yields contaminated estimates.

Rebate reporting bias. A bias specific to the CEX survey data in which the timing of a household’s self-reported rebate receipt is correlated with unusually high contemporaneous expenditure (and correspondingly low prior-period expenditure), likely due to recall effects. Because the true rebate timing is random but the reported timing is not, this correlation inflates the difference-in-differences estimate of the spending effect.

Two-good, two-agent New Keynesian (TANK) model. A medium-scale New Keynesian model containing two types of households (optimizing life-cycle consumers and hand-to-mouth consumers who exhaust current income) and two goods (nondurables and durable goods interpreted as motor vehicles). The model is used in this paper as a framework to translate micro MPC estimates into aggregate general equilibrium counterfactuals, calibrated at monthly frequency.

Durable supply elasticity. The elasticity of real durable goods production with respect to the relative price of durable goods, calibrated in the paper to 5. In the baseline model, this elasticity is infinite (the relative price is fixed at one because intermediates convert frictionlessly). With a finite supply elasticity of 5, rebate-induced durable demand causes the relative vehicle price to rise, generating crowding out of optimizing households’ durable expenditure.

Calvo durable adjustment friction. An adjustment friction imposed on optimizing households’ durable goods purchases, motivated by Evans and Ramey’s (1992) calculation cost model. Only a fraction 1−θd of households reoptimize their durable stock each period (with probability drawn randomly), producing a Calvo-type reduced form. This friction limits both the extensive and intensive margins of durable adjustment and prevents unrealistically large intertemporal substitution of durable purchases in response to price changes.

Macro counterfactual. In this paper’s usage, the simulated path of aggregate consumption that would have occurred in the absence of the 2008 tax rebate, constructed by subtracting the model-implied impulse response to the rebate from the actual observed NIPA consumption series. Plausibility of the counterfactual is assessed by comparison to contemporaneous forecasts and to historical episodes of large consumption declines.

Monetary Policy and Sovereign Risk in Emerging Economies (NK-Default)

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1: Overview

This paper develops a New Keynesian small open economy model with endogenous sovereign default — the NK-Default framework — and uses it to study the interplay between monetary policy and sovereign risk in emerging markets. The core finding is that sovereign default risk amplifies inflation volatility through an expectations channel: when default risk rises, forward-looking firms increase prices in expectation of high future inflation and depressed consumption during a potential default, so that current inflation rises even before any default occurs. Conversely, tight monetary policy disciplines government overborrowing by raising the cost of domestic monetary distortions, which the government internalizes by reducing its borrowing. Calibrated to eight emerging-market inflation targeters (Brazil, Chile, Colombia, Mexico, Peru, Philippines, Poland, South Africa) over 2004–2019, the model quantitatively matches the positive comovement of spreads with inflation and nominal rates, and the temporary nature of inflation events (approximately 4.5% inflation spike, 2.3% spread increase, resolved within roughly a year). Counterfactual experiments find that default risk accounts for approximately 50% of both inflation business-cycle volatility and the inflation increase during these events, and that a 1% tighter monetary policy would reduce spreads by about 0.3% during inflation events. An interest rate rule augmented to respond to default risk dominates strict inflation targeting in welfare and reduces mean spreads by 2.2 percentage points; strict inflation targeting is not the optimal monetary regime when sovereign risk is present.

Summary of a published paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. What is the structural architecture of the NK-Default model?

The NK-Default framework combines the standard New Keynesian small open economy model of Gali and Monacelli (2005) with the Eaton-Gersovitz (1981) sovereign default structure extended to long-term foreign-currency debt, producing a model in which households, firms, a monetary authority, and a fiscal government all interact. Households consume domestic and foreign goods and supply labor; intermediate goods producers are monopolistically competitive and set prices subject to Rotemberg (1982) quadratic adjustment costs, generating a forward-looking New Keynesian Phillips Curve (NKPC); the monetary authority follows a nominal interest rate rule targeting domestic goods inflation; and the government borrows internationally in long-term foreign-currency perpetuity bonds, choosing each period whether to repay or default, with default leading to temporary exclusion from international financial markets and a transitory productivity reduction. The bond price schedule compensates risk-neutral international lenders for expected losses from default and falls with the government’s indebtedness. A key methodological choice is the use of global solution methods rather than local approximations, because the nonlinear dynamics around default are central to the mechanisms.

Q2. What is the default amplification mechanism, and how does it transmit to inflation?

Default amplification operates through an expectations channel encoded in the forward-looking NKPC: when default risk rises, firms’ expectations of higher future inflation (during the inflation that would accompany a default event) and lower future consumption (because default depresses productivity and restricts borrowing) both increase, causing firms to raise current prices, generating current inflation without any contemporaneous policy change. Formally, the NKPC relates current inflation π to a unit-cost term and to the expectation term E[Y’u’_C(π’-π)π’], which increases with default risk because default states feature high inflation and high marginal utility. The resulting current inflation increase then triggers the monetary authority’s interest rate rule to tighten, which in turn depresses consumption through the Euler equation, amplifying the monetary distortion (wedge). In the simplified quasi-linear preferences setting, higher borrowing B’ increases functions F and M — the expectation terms in the NKPC and Euler equation — and Proposition 1 establishes formally that higher borrowing raises default risk, inflation, the nominal domestic rate, and the monetary wedge under Assumption 1.

Q3. How does monetary policy discipline sovereign borrowing?

Tight monetary policy disciplines government overborrowing because the government internalizes the additional costs that monetary distortions impose on the economy: when the monetary authority raises interest rates, the resulting monetary wedge — the gap between the marginal product of labor and households’ marginal rate of substitution — acts as an additional cost on borrowing from the government’s perspective, discouraging excessive debt accumulation. Proposition 2 establishes this formally: under Assumption 2 (one-time deviation from constrained efficiency), a policy rate i > i_ST (above the strict-inflation-targeting rate) generates a positive monetary wedge that modifies the government’s optimal borrowing condition with an additional term reflecting the cost to the sovereign of the higher wedge its borrowing induces. Contractionary monetary policy thus reduces the incentive to borrow and lowers equilibrium default risk. The paper also derives Proposition 3: a default-risk monetary rule of the form i = ī·Φ^αD can achieve the constrained-efficient default risk and an arbitrarily small monetary wedge simultaneously, by choosing αD appropriately — meaning that targeting default risk can address both the pricing friction and the overborrowing incentive.

Q4. What are the quantitative findings on default amplification and the disciplining mechanism?

Quantitatively, default risk accounts for approximately 50% of inflation business-cycle volatility and approximately 50% of the inflation increase during the temporary inflation events (4.5 p.p. inflation spike, 2.3 p.p. spread increase, nominal rate rise from baseline 5–6% to 8–9%), based on comparison with a reference model without default. For the disciplining mechanism, panel-data regressions using monetary policy shocks recovered from estimated Taylor rules across the eight countries find that a 1% contractionary monetary shock reduces sovereign spreads, consistent with model predictions. During the inflation events, a 1% tighter monetary policy would have reduced spreads by approximately 0.3 percentage points. Comparing alternative monetary policy regimes against strict inflation targeting (which implements flexible-price allocation): the baseline interest rate rule (responding only to inflation) reduces mean spreads by 0.5 percentage points relative to strict inflation targeting; an augmented rule that also responds to default risk reduces mean spreads by 2.2 percentage points. Welfare under the baseline rule exceeds that under strict inflation targeting, and welfare under the default-risk rule exceeds both, with the ranking holding across all robustness extensions.

Q5. How does the model fit the data across targeted and untargeted moments?

The model is calibrated to match key business-cycle statistics of the eight emerging-market inflation targeters and successfully replicates several untargeted moments, including the positive correlations of spreads with inflation (mean 0.5 across countries in data) and nominal rates (mean 0.3), the relative volatility of inflation to output (mean 0.8), and the mean spread level of approximately 2%. The temporary inflation events — constructed as windows around periods of elevated inflation — are matched with a combination of low productivity shocks and expansionary monetary shocks, and the model’s impulse response functions for inflation, output, nominal rates, and spreads during these events align with the empirical paths. The model also fits the positive elasticity of inflation expectations to default risk and the negative elasticity of spreads to monetary policy shocks, both of which are estimated from data and used as untargeted validation moments. Structurally, the model is parameterized to match the mean and volatility of inflation, spreads, and the correlation of spreads with output (mean -0.5 across countries), among other moments.

Q6. How do the model’s results hold up across extensions, especially local currency debt and discretionary monetary policy?

The main results — default amplifies inflation, tight monetary policy disciplines borrowing, and the default-risk rule dominates strict inflation targeting — are robust across all extension economies, including the case of local currency sovereign debt, alternative default costs (no productivity loss, endogenous domestic financial frictions), and loose monetary policy during defaults. In the local currency debt extension, which introduces the classic incentive to erode debt via inflation, the paper shows that monetary discretion delivers substantially worse outcomes: average inflation doubles relative to the commitment case and — crucially — sovereign spreads also double under discretion, because market participants anticipate the inflationary incentive. This result shows that the disciplining benefits of commitment in monetary policy rules extend to the sovereign debt dimension: the country’s ability to commit to a rule lowers spreads by reducing the expected future inflation that lenders must be compensated for. The endogenous financial frictions extension — in which banking sector health depends on nominal rates and spreads — generates similar monetary-fiscal interactions, confirming that the mechanisms are not specific to the productivity-cost assumption.

Q7. What is the paper’s relationship to the literature on nominal rigidities and sovereign default?

The NK-Default framework differs critically from related papers that introduce downward nominal wage rigidity (e.g., Na, Schmitt-Grohe, Uribe, Yue 2018; Bianchi, Ottonello, Presno 2023) in that price-setting frictions arise from optimal forward-looking pricing by monopolistically competitive firms under Rotemberg costs, not from a mechanical wage floor, so that inflation expectations matter for current inflation and output in a standard NKPC. This means that expected future default events — through their effects on expected inflation and expected marginal utility — transmit to current equilibrium in a way that downward-rigid-wage models cannot replicate. The paper also differs from the literature studying the inflation incentive for local-currency debt dilution (e.g., Calvo 1988; Du, Pflueger, Schreger 2020): the baseline model assumes foreign-currency debt and a rule-based monetary authority that has no incentive to inflate away debt, so the mechanisms operate through expectations and discipline rather than through the debt-erosion channel. The paper connects these strands in the local-currency extension.

Q8. What are the welfare and policy implications for central bank mandates in emerging markets?

The paper provides formal support for monetary policy rules that respond to financial or sovereign-risk conditions — beyond standard inflation targeting — in emerging economies: the welfare ranking is default-risk rule > baseline rule > strict inflation targeting, with the gap between the default-risk rule and strict inflation targeting driven by lower mean and volatility of spreads, which reduce the frequency and severity of default amplification events. Strict inflation targeting, which delivers the flexible-price allocation, is not optimal because it leaves the overborrowing incentive of the fiscal government unchecked, generating excessive default risk that feeds back into inflation volatility through the expectations channel. A monetary rule with sufficient responsiveness to inflation or to default risk disciplines fiscal behavior and reduces welfare costs from both pricing frictions and default risk, suggesting that emerging-market central bank mandates that focus exclusively on inflation targeting at the expense of financial stability considerations may be suboptimal relative to rules that jointly address monetary and fiscal distortions.

Key Concepts

NK-Default framework: the paper’s model combining a New Keynesian small open economy (Gali-Monacelli structure with Rotemberg price-setting frictions and a Taylor-type interest rate rule) with the Eaton-Gersovitz endogenous sovereign default structure extended to long-term foreign-currency perpetuity bonds; the joint treatment of monetary policy and sovereign risk for emerging economies.
default amplification: the mechanism by which elevated sovereign default risk increases current inflation and depresses output through the forward-looking NKPC expectations channel: firms raise prices in anticipation of high future inflation and low consumption during a potential default, so current inflation rises even without any contemporaneous fiscal action; established as Proposition 1 in the simplified model and confirmed quantitatively.
monetary discipline: the mechanism by which contractionary monetary policy raises the cost of government borrowing through monetary distortions (the monetary wedge), inducing the fiscal government to reduce its indebtedness and thereby lowering equilibrium default risk; established as Proposition 2 and confirmed empirically using panel-data regressions of spreads on monetary policy shocks.
monetary wedge: the deviation of the marginal product of labor from households’ marginal rate of substitution between labor and consumption, arising from price-setting frictions; serves as the quantitative measure of monetary distortions and is the channel through which monetary policy affects government borrowing incentives.
default-risk monetary rule: an interest rate rule of the form i = ī·Φ^αD that responds directly to the one-period-ahead default probability Φ; shown in Proposition 3 to achieve both the constrained-efficient level of government debt and an arbitrarily small monetary wedge simultaneously, by incorporating an additional cost of borrowing for the fiscal government through the rule’s response.
temporary inflation events: empirical regularities in eight emerging-market inflation targeters in which inflation, spreads, and nominal policy rates temporarily spike together (inflation rises approximately 4.5%, spreads by 2.3%, within roughly one year) before reverting to lower levels; the model replicates these patterns using a combination of low productivity shocks and expansionary monetary shocks.

Optimal Resilience in Multitier Supply Chains

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

Grossman, Helpman, and Sabal ask what market failures arise in vertical supply chains with multiple production tiers, limited (non-anonymous) supply networks, arms-length transactions, and recurrent risks of disruption at every node. They then ask what government policies would be required to implement the socially efficient (first-best) allocation as a decentralized equilibrium, and — in a second-best environment where subsidies to firm-to-firm transactions are politically infeasible — how optimal policies to promote resilience and network formation differ.

Model and Methodology

The paper develops a general-equilibrium model of a closed economy with an arbitrary number S+1 of vertical production tiers (tier 0 through tier S). A finite measure of “lead” firms in tier S produce differentiated consumer goods under monopolistic competition using labor and a CES bundle of intermediate inputs from tier S-1 suppliers. Firms in each intermediate tier combine labor and inputs from the tier above using a Cobb-Douglas production function. Tier 0 firms produce from labor alone.

Every firm faces an independent, non-zero probability of a catastrophic disruption (complete inability to produce). Firms may invest labor up front to moderate this risk — endogenous “resilience” — or may invest to forge relationships with a larger fraction of potential suppliers in the next upstream tier — endogenous “network thickness.” Each formed relationship costs k units of labor.

After disruption shocks are realized, surviving firms negotiate quantities and payments bilaterally. Bargaining is sequential (beginning with lead firms negotiating with tier S-1, then tier S-1 with tier S-2, and so on to tier 0), and within each round is governed by Nash-in-Nash equilibrium (Horn and Wolinsky, 1988): each firm takes as given the outcomes of its negotiations with all other partners. The Nash surplus is split with exogenous bargaining weight β_s for the downstream buyer in the s-to-s−1 negotiation.

The paper solves the planner’s direct-control problem and then characterizes the three sets of policy instruments needed to decentralize the first best: subsidies to input transactions between adjacent tiers, subsidies to investments in resilience (agility), and subsidies to network formation (redundancy). It then solves the second-best problem in which transaction subsidies are constrained to zero.

Main Findings

Transaction subsidies. In the competitive bargaining equilibrium, each pair of firms undervalues input transactions because the upstream firm anticipates paying a marked-up price when it bargains with its own suppliers. This cascading distortion means the private marginal cost of producing a tier-s good exceeds the social marginal cost. The optimal first-best transaction subsidy on sales by tier s firms (τ*s) equals [γ_s + (1−γ_s)μ{s−1}]^{−1}, where γ_s is the labor share in tier s production and μ_{s−1} is the endogenous markup factor from bargaining at the s-to-s−1 interface. This subsidy depends only on production function parameters and bargaining weights at the immediately adjacent tier. No subsidy is needed at tier 0 (the most upstream tier), and no subsidy is applied to final-good sales. Under Assumption 1 — inputs become weakly less substitutable as goods proceed downstream — the optimal purchase subsidies rise monotonically as one moves downstream along the supply chain.

Resilience subsidies (first best). Two offsetting forces govern the optimal subsidy to resilience investments θ*s at intermediate tiers: (i) firms capture only the fraction (1−β{s+1}) of the joint surplus that their resilience creates for downstream customers, creating underinvestment; (ii) optimal transaction subsidies inflate private profitability, creating a countervailing overinvestment incentive. The net optimal first-best subsidy for intermediate-tier firms is θ*s = (1−β{s+1}) / τ*_s. This formula depends only on technological and bargaining parameters of tier s and the tier immediately adjacent; it does not depend on conditions elsewhere in the chain. When production parameters and bargaining weights are uniform across tiers, the first-best resilience subsidy is the same at every interior tier. If goods become strictly less substitutable downstream, the first-best subsidy for resilience declines monotonically as one moves downstream, and may turn into an optimal tax for middle tiers where the transaction subsidy is large enough to over-incentivize resilience investment. The first-best resilience subsidy always applies at both extreme ends of the chain.

Network formation subsidies (first best). Despite firms’ private incentive to manipulate their number of upstream suppliers to improve bargaining position, the net strategic effect of network formation in general equilibrium exactly cancels the off-equilibrium spillovers to non-partners. As a result, the optimal first-best policy toward network formation at every tier is identical to the optimal policy toward resilience investment.

Second-best policies. When transaction subsidies are unavailable, uncorrected markups downstream from tier s depress demand for tier-s output, reducing profitability and incentives to invest in resilience below the first-best level. Second-best optimal subsidies for resilience and network formation therefore reflect production function parameters and bargaining weights throughout the entire downstream supply chain, not just at the immediately adjacent tier. Specifically, when buyer bargaining weights are non-increasing along the chain (β_{s+1} ≤ β_s for all s), the second-best subsidy to resilience falls monotonically as one moves downstream. This is the opposite pattern from what might be inferred from the first-best analysis when transaction subsidies are available: with non-increasing bargaining weights, second-best subsidies are larger for upstream producers than for downstream producers.

Scope Conditions

Results are derived for a closed economy. Welfare is measured by the CES utility of the representative consumer over differentiated final goods. The sequential bargaining structure assumes contracts are written after disruption shocks are realized. Assumption 1 (σ_1 ≥ σ_2 ≥ … ≥ σ_S > ε, where σ_s is the elasticity of substitution between inputs at tier s and ε is the demand elasticity for final goods) is maintained for sharper monotonicity results on the structure of optimal subsidies across tiers.

In depth

Q1. What is the precise structure of the supply chain in the model, and why does the bargaining take place sequentially rather than simultaneously across all tiers?

A: The economy has S+1 tiers. Tier 0 firms use only labor; tier s firms (s = 1,…,S−1) use labor and a CES bundle of tier s−1 inputs with elasticity of substitution σ_s > 1; tier S firms produce final differentiated goods using labor and tier S−1 inputs under Cobb-Douglas technology. Sequential bargaining is imposed because the vast number of simultaneous negotiations across all tiers makes a grand coalition impractical. The timing is that lead firms (tier S) first negotiate input quantities and payments with their tier S−1 suppliers; those suppliers, now contractually obligated to their downstream customers, then negotiate with tier S−2, and so on up the chain until tier 1 firms contract with tier 0 suppliers.

Q2. How is the markup factor defined, and what parameters determine it?

A: The markup factor μ_s is the ratio of the payment per unit made by tier s+1 firms to the production cost of tier s firms. It equals μ_s = (1−β_{s+1}) · [σ_{s+1}/(σ_{s+1}−1)] + β_{s+1}, where β_{s+1} is the exogenous bargaining weight of the downstream (tier s+1) buyer. When the downstream firm has all bargaining power (β_{s+1} = 1), the markup equals unity (competitive outcome). When the upstream firm has all bargaining power (β_{s+1} = 0), the markup equals the standard monopoly markup σ_{s+1}/(σ_{s+1}−1). For intermediate bargaining weights, the markup is a weighted average. The markup enters the optimal transaction subsidy formula by inflating the private marginal cost of producing tier-s inputs above the social marginal cost.

Q3. Why are no subsidies needed for the most upstream (tier 0) transactions or for final-good sales?

A: For tier 0 transactions: when tier 0 and tier 1 firms bargain, the negotiations occur last sequentially and so do not affect any prior agreements. There are no downstream cascading markup effects — tier 0 firms produce from labor alone, so their private marginal cost equals their social marginal cost. The joint surplus maximization by the pair thus aligns with the planner’s objective, yielding τ*_0 = 1 (no intervention needed). For final-good sales: final producers do mark up above marginal cost under monopolistic competition, but all varieties are symmetric, so the markup affects all goods equally and does not distort relative consumption choices. Hence τ*_S = 1.

Q4. What are the two offsetting forces that determine the optimal first-best subsidy to resilience investments at an intermediate tier?

A: First, a firm in tier s captures only the fraction (1−β_{s+1}) of the joint surplus that its survival creates for its downstream customers (the rest is appropriated through bargaining by those customers), leading to underinvestment relative to the social optimum. Second, the optimal transaction subsidy τ*_s < 1 raises the private profitability of firms in tier s above its social value, because public finances bear part of the cost of their input purchases. This inflated private profitability encourages resilience investment beyond what the planner desires. The net optimal policy is θ*s = (1−β{s+1}) / τ*_s, which may be a subsidy (θ*_s < 1) or a tax (θ*_s > 1) depending on which force dominates.

Q5. Why does the first-best subsidy for resilience at an intermediate tier depend only on local parameters (at tier s and its immediate neighbors), even though resilience investments generate spillovers to firms throughout the network?

A: When optimal transaction subsidies are in place at all tiers, a firm’s value becomes independent of the joint surplus in sales that occur between firms in tiers other than its own. That is, the positive spillovers to all firms farther upstream and downstream in a firm’s own network are exactly offset by the negative spillovers to firms in rival networks (including rival firms in the same tier). What remains after this general-equilibrium cancellation is only the benefit to the firm’s immediate downstream customers and the wedge created by the transaction subsidy. This result implies that the formula θ*s = (1−β{s+1}) / τ*_s does not involve conditions at tiers other than s and s−1.

Q6. Why does the optimal policy for network formation (supplier link investment) equal the optimal policy for resilience investment, despite the fact that network formation also strategically improves a firm’s bargaining position?

A: Firms in intermediate tiers do have a private incentive to form additional supplier links specifically to improve their bargaining position vis-à-vis their upstream suppliers (by improving their outside options) and vis-à-vis their downstream customers (by the same mechanism). However, the authors show by comparing the firm’s first-order condition for link formation with the planner’s first-order condition that this strategic motivation exactly balances the offsetting general-equilibrium effects from rival firms doing the same. After this cancellation, the residual wedge between private and social incentives for network formation is identical to that for resilience investment. Hence #_s = θ_s for all tiers.

Q7. How do second-best policies differ from first-best policies in terms of both the magnitude of subsidies and the information required to set them?

A: In the first best, the subsidy for resilience at tier s depends only on the bargaining weight β_{s+1} and the markup factor μ_{s−1} — parameters relevant to tier s and its immediate neighbors. In the second best, when transaction subsidies are unavailable, the optimal resilience subsidy at tier s is θ†s = J^{−1} · [1 − (cumulative distortion of all downstream tiers)] · (1−β{s+1}), where J captures aggregate labor-market effects of all markups throughout the chain. This formula requires knowledge of production function parameters (labor shares γ_j, markups μ_j, elasticities σ_j) for every tier j downstream from s. The second-best subsidy may be larger or smaller than the first-best subsidy; it is more likely to exceed the first-best subsidy for upstream tiers, where the cumulative downstream distortions (uncorrected markups contracting demand) produce a larger shortfall in private profitability and hence a larger underinvestment in resilience.

Q8. Under what condition do second-best subsidies fall monotonically as one moves downstream, and how does this compare to the first-best pattern?

A: The ratio of second-best subsidies at adjacent tiers (θ†_{s−1} / θ†s) equals [(1−β_s) / (1−β{s+1})] · [τ*s]^{−1}, where τ*s is the first-best transaction subsidy. If buyer bargaining weights are non-increasing along the chain — β{s+1} ≤ β_s for all s — then (1−β_s) ≤ (1−β{s+1}) and, combined with τ*s ≤ 1, the second-best subsidy is larger upstream than downstream (θ†{s−1} ≥ θ†_s). This contrasts with the first-best policy: when parameters are uniform across tiers, first-best resilience subsidies are the same at every interior tier, while second-best subsidies are strictly larger upstream than downstream.

Q9. What role does Assumption 1 (elasticities of substitution non-increasing as goods move downstream) play in the results?

A: Assumption 1 (σ_1 ≥ σ_2 ≥ … ≥ σ_S > ε) ensures that the operating profit function ~v_s(η) is concave in a firm’s network size η, which in turn ensures interior solutions to the network formation problem. It also delivers sharper monotonicity results: under this assumption, if other production parameters and bargaining weights are similar across tiers, the optimal purchase subsidies rise monotonically downstream, and the optimal first-best resilience subsidies decline monotonically downstream (potentially turning into taxes at some interior tiers). The assumption reflects the realistic view that inputs become more differentiated and specialized as they approach the final consumer good.

Q10. What are the limitations the authors identify regarding their model, and what extensions do they suggest?

A: Three main limitations are identified. First, the model assumes bargaining occurs after disruption shocks are realized, ruling out contingent contracts. Pre-disruption bargaining with contingent payments could mitigate double-marginalization inefficiencies and help internalize resilience externalities, though complex network-wide contingent contracts would likely be needed for full efficiency even in the second-best environment. Second, the model assumes symmetric firms within each tier, so downstream firms cannot sort on upstream firms’ observable resilience levels; if observable differences existed, downstream firms could seek out more reliable partners, partially internalizing the resilience externality. Third, the model covers only a closed economy with idiosyncratic (uncorrelated) shocks. Extensions to global supply chains, correlated (geographic) shocks, cross-country differences in wages and technologies, and optimal cooperative versus unilateral policy are identified as important directions for future research.

Key Concepts

Resilience (agility): In the paper’s usage, a firm’s endogenous investment in reducing the probability of a catastrophic disruption to its own operations. A firm in tier s hires r_s units of labor up front, which raises its survival probability φ_s(r_s), with φ’_s > 0 and φ’’_s < 0. Resilience is a relationship-specific investment in the sense that its payoff is realized only conditional on the firm surviving and then trading with its downstream customers.

Network thickness (redundancy): The fraction η_s of firms in the next upstream tier with whom a firm in tier s forms a supply relationship prior to the disruption shock. Forming k units of labor per link creates a thicker network that hedges against supplier disruption, increases input variety (and thus CES productivity), and improves bargaining positions vis-à-vis both upstream suppliers and downstream customers. Distinct from resilience: resilience reduces the firm’s own probability of disruption; network thickness provides substitutability across suppliers should some fail.

Markup factor (μ_s): The ratio of the per-unit payment made by tier s+1 firms to the production cost of tier s firms, as determined by Nash bargaining. Specifically, μ_s = (1−β_{s+1}) · [σ_{s+1}/(σ_{s+1}−1)] + β_{s+1}. The markup distorts private marginal costs above social marginal costs, causing underinvestment in transactions between firms and, transitively, in resilience and network formation.

Nash-in-Nash equilibrium: The bargaining solution concept used in the paper (following Horn and Wolinsky, 1988). Each pair of firms negotiates as if all other bilateral negotiations involving either party proceed at their equilibrium outcomes, both on and off the equilibrium path. This is the appropriate equilibrium concept when grand coalitions across all firms and all tiers are impractical.

Sequential bargaining: The specific timing structure in which negotiations proceed from the most downstream tier (lead firms bargaining with tier S−1 suppliers) sequentially upstream until tier 1 firms bargain with tier 0 suppliers. Each tier of firms, at the time they bargain with their own suppliers, are already contractually obligated to deliver specified quantities to their downstream customers. This obligation anchors the downstream firm’s outside option in any given bilateral negotiation.

First-best transaction subsidy (τ_s):* The fraction of the cost of a tier-s input that, under the optimal policy, the downstream (tier s+1) buyer must pay. Equals [γ_s + (1−γ_s) · μ_{s−1}]^{−1} < 1 for all intermediate tiers, i.e., it is always a subsidy. Designed to align private marginal cost in the bilateral negotiation with the social marginal cost by offsetting the distortion introduced by anticipated markups on the upstream firm’s own inputs.

Second-best subsidy: The optimal policy toward resilience and network formation when subsidizing firm-to-firm transactions is infeasible (constrained to τ_s = 1 for all s). Unlike first-best subsidies — which depend only on local tier parameters — second-best subsidies depend on production function parameters and bargaining weights throughout the entire downstream supply chain due to the uncorrected cumulative markup distortions.

Peer Effects and the Gender Gap in Corporate Leadership

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates whether exposure to a larger share of female peers during an MBA program causally affects the gender gap in senior corporate leadership positions. The research question is motivated by the persistent underrepresentation of women in top management: in S&P 1500 companies, women hold only 6% of CEO positions despite comprising 40% of the workforce.

The authors merge administrative data from a top-10 U.S. business school (graduating classes 2000–2018, excluding 2009) with public LinkedIn profile data covering full employment histories, firm-level data from multiple sources including InHerSight crowdsourced female-employee ratings, and a 2023–2024 alumni survey of female graduates. Senior management is defined as Vice President, Director, Senior Vice President, or C-level executive, identified from exact job titles in LinkedIn CVs.

Identification exploits the quasi-random assignment of incoming MBA students to one of eight sections of approximately 60 students each, based on alphabetical order with balance checks on gender, undergraduate institution, and ethnicity. This assignment generates exogenous variation in the share of female section peers (mean 34%, standard deviation 4 percentage points). Randomization tests following Guryan et al. (2009) and Caeyers and Fafchamps (2021) confirm the assignment is as good as random. The estimating equation is a linear-in-means model with class, year, and class-by-year fixed effects interacted with gender, plus individual and section-level controls.

The paper first documents a baseline gender gap: despite 96% of both male and female MBA graduates entering management within 15 years, women are 24% less likely than men to hold senior management positions. This gap emerges immediately after graduation, persists for at least 15 years, and is partly attributable to lower promotion rates from first-level management (43% of women in first-level management transition to senior management within five years, versus 57% of men).

The main causal finding is that a 4 percentage point (1 SD) increase in the share of female MBA section peers increases the probability of a woman holding a senior management position by 8.4% (a 3.3 percentage point increase off a 39.1% baseline), equivalent to a 26% reduction in the management gender gap. There is no corresponding effect for men. The effect emerges as early as two years post-graduation, peaks around year seven, and persists through the 15-year horizon.

The increase is concentrated in female-friendly firms, defined as those with above-median ratings on InHerSight metrics including maternity leave generosity, flexible work schedules, and professional support. Women with more female peers are significantly more likely to transition into female-friendly firms 6 to 10 years after graduation — a period coinciding with prime childbearing years — where they subsequently attain senior management roles. The effect on senior management in female-friendly firms is statistically distinguishable from the null effect in non-female-friendly firms (p-value = 0.03). The results are largest in male-dominated industries (consulting, tech, finance) where women face greater barriers to informal networks.

A survey of 283 female MBA alumnae (10% response rate) reveals three mechanisms: (i) information sharing, especially gender-specific advice about employer policies and culture; (ii) higher ambitions and self-confidence through role modeling and emotional support; and (iii) increased perceived support from male MBA peers as female section representation rises. Corroborating the information-sharing channel, women with more female peers are more likely to work at the same firms as their female section peers, particularly when those firms are female-friendly.

A counterfactual exercise shows that reallocating the existing stock of female students so that all sections have at least 34% women would yield 2 to 5 additional female senior managers per graduating class (a 2.4% to 8.4% increase), holding the total number of female students fixed.

Q: What is the baseline gender gap in senior management among MBA graduates, and how does it evolve over time? A: Female MBA graduates are 24% less likely than male graduates to hold senior management positions in the 15 years after graduation. The gap emerges immediately after the MBA and persists for at least 15 years without closing. At year 15, 74% of men hold a senior management position compared to 59% of women.

Q: How is female peer share defined and what is its distribution across sections? A: Female peer share is the proportion of female students in an individual’s assigned MBA section of approximately 60 students, excluding the individual themselves. The average section female share is 34% with a standard deviation of 4 percentage points. The distribution ranges from 19% at the 1st percentile to 45% at the 99th percentile, with the interquartile range spanning approximately 32% to 36%.

Q: What is the main causal estimate of female peers on women’s senior management probability? A: A 4 percentage point (1 SD) increase in female section peer share increases the probability of a woman holding a senior management position by 8.4% (3.3 percentage points off a 39.1% mean), averaged across the 15 post-MBA years. This translates to a 26% reduction in the management gender gap. There is no statistically significant effect on men.

Q: When does the effect of female peers emerge and how does it evolve dynamically? A: The effect on women emerges as early as two years after MBA graduation and grows over time, peaking around seven years post-graduation. The effect is persistent across the 15-year horizon studied. Estimates become less precise toward the end of the sample period as recent cohorts contribute fewer observations.

Q: How do female-friendly firms mediate the main result? A: The main effect is entirely concentrated in female-friendly firms (those with above-median InHerSight ratings). The coefficient on female peer share is positive and significant for senior management in female-friendly firms, and statistically indistinguishable from zero in non-female-friendly firms. The difference between the two coefficients is significant at p = 0.03.

Q: What is the mechanism linking female peers to female-friendly firm transitions? A: Women with more female peers are significantly more likely to be employed at female-friendly firms 6 to 10 years after graduation, a window corresponding to prime childbearing years. This suggests female peers facilitate sorting into supportive firm environments when family-work tradeoffs become most acute. Once at female-friendly firms, women attain senior management positions at higher rates.

Q: Does the increase in female senior managers reflect easier paths (smaller firms, lower pay, non-P&L roles)? A: No. The effect is significant for both small (under 500 employees) and large (over 5,000 employees) firms, with no significant effect on the firm size of employment itself. There is no consistent pattern of women being promoted in firms with higher or lower average compensation. The increase in female senior managers includes those with Profit and Loss responsibilities, indicating these are substantive management positions.

Q: In which industries is the effect largest, and what does this imply? A: The effect is concentrated in male-dominated industries (consulting, tech, finance), with no significant effect in female-dominated industries (consumer goods, healthcare). The difference between coefficients is significant at the 3% level. Entry rates into male-dominated industries are not significantly affected, suggesting the mechanism is higher promotion rates within these industries rather than differential sorting into them. The authors interpret this as evidence that female MBA networks are most valuable where women face greater barriers to informal workplace networks.

Q: What does the survey evidence reveal about mechanisms? A: Among 283 survey respondents (10% response rate), three mechanisms emerge: information sharing about gender-specific employer attributes and policies; raising ambitions and self-confidence through role modeling; and increased perceived support from male MBA peers as section female share rises. Women with more female peers are also more likely to work at the same firms as their female section peers, especially female-friendly ones, consistent with referral and information-sharing channels.

Q: Does the effect operate through greater attachment to the corporate pipeline (fewer career breaks, higher entry into management)? A: No. Female peers do not significantly affect employment rates, career break incidence, entry into first-level management positions, or self-employment rates. The results thus reflect higher promotion rates from first-level management into senior management, not changes in pipeline attachment.

Q: What do the randomization tests show about identification validity? A: Two randomization tests confirm as-good-as-random assignment. Following Guryan et al. (2009), the section-level leave-out mean female share is not significantly different from zero after controlling for the class-level leave-out mean. Following Caeyers and Fafchamps (2021), after netting out the asymptotic exclusion bias, the female share coefficient is insignificant across all specifications. A simulation test (Bietenbeck 2020) finds no statistically significant difference between the actual and simulated within-class female share distributions.

Q: What placebo tests are conducted and what do they show? A: Two placebo tests are run. First, 1,000 random reassignments of students to sections within the same class show the true estimated effect for women lies outside the distribution of placebo effects, while the null effect for men lies within it. Second, estimating the main equation for up to three years before MBA enrollment finds no consistent pre-treatment effect of female share on future female graduates, supporting the identification strategy.

Q: What is the counterfactual policy exercise and what does it imply? A: Holding the total number of female students fixed, reallocating them so that all sections contain at least 34% women would yield 2 to 5 additional female senior managers per graduating class (a 2.4% to 8.4% increase). This assumes nonlinearity in the relationship and suggests meaningful gains from rebalancing section composition without increasing overall female enrollment.

Q: How do the results compare to the Thomas (2021) finding that more male peers raise female MBA earnings? A: The authors note several differences: Thomas (2021) focuses on starting earnings while this paper studies senior management positions over 15 years; the two studies use different universities and time periods; and this paper employs gender-by-cohort fixed effects to account for time trends in female labor market outcomes. The authors suggest these design and outcome differences explain the divergent findings.

Section peers: Students assigned to the same MBA section of approximately 60 students who take core classes together and form the primary peer network; sections are assigned quasi-randomly based on alphabetical order with balance adjustments, generating exogenous variation in gender composition.

Female-friendly firms: Firms with above-median ratings on InHerSight, a crowdsourced platform where female employees rate employers on metrics including maternity leave generosity, flexible work schedules, mentorship programs, and female representation in management; defined in this paper’s own terms as firms whose cultures and policies help women balance work-family responsibilities and support career advancement.

Senior management: Positions defined as Vice President (VP), Director, Senior Vice President (SVP), or C-level executive, identified using keyword matching on exact job titles from LinkedIn CVs; distinguished from first-level management (managers and supervisors) and representing the upper rungs of the corporate management ladder.

Female share (treatment variable): The proportion of female students among an individual’s section peers, excluding the individual themselves (leave-out mean); averaged 34% with a 4 percentage point standard deviation across sections, after residualizing by graduating class.

Management gender gap: The 24 percentage point (24%) difference in the likelihood of female versus male MBA graduates holding senior management positions within 15 years of graduation; emerges immediately post-MBA and does not close over the observed horizon.

Information sharing mechanism: The channel through which female MBA peers provide gender-specific advice and information about employer policies, culture, and female-friendliness that is otherwise difficult to observe; evidenced by the co-location of women with more female peers at the same female-friendly firms as their section peers.

Exclusion bias: The systematic negative correlation between an individual’s own characteristic and her leave-out peer mean that arises mechanically when individuals cannot be their own peer under assignment without replacement; addressed via the Caeyers and Fafchamps (2021) correction in randomization tests.

Permanent Capital Losses after Banking Crises

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

This paper investigates two interrelated questions about historical banking crises: (1) whether bank losses during banking crises are primarily temporary or permanent in nature, and (2) whether policy interventions — particularly liquidity-based interventions — are effective at restoring bank capitalization after such crises. The paper positions these questions against a theoretical divide: models stressing temporary price dislocations (binding borrowing constraints, depositor fragility, information frictions) versus models in which crises reflect fundamental and permanent deterioration in the value of bank assets.

Data and Methodology

The authors construct three new historical datasets spanning 46 economies from 1870 to 2019. The first is a country-level panel of annual and monthly bank and nonfinancial equity index total returns, building on Baron, Verner, and Xiong (2021). The second is an individual-bank-level dataset covering the ten largest banks per country across 17 economies (from Jordà, Schularick, and Taylor 2017), containing equity returns, balance sheet quantities, net income decomposed into write-downs and trading income, and equity issuance within ±5-year windows around each crisis. The third is a new database of the monthly starting dates of policy interventions — extraordinary central bank liquidity support, blanket liability guarantees, and government recapitalizations — extending the databases of Laeven and Valencia (2020) and Metrick and Schmelzing (2024).

Bank equity crises are identified using a real-time, data-driven indicator requiring: (1) a greater than 30% annual decline in the bank equity index and (2) the failure of a top-20 bank within the country. This definition yields 76 bank equity crises, nearly all of which overlap with prior narrative-based chronologies (Reinhart-Rogoff, JST, Laeven-Valencia), and results are robust to all alternative crisis definitions examined.

Main Findings

Permanent losses. In the year of a bank equity crisis onset, bank equity experiences average abnormal returns of -68 log-points (or -49% in arithmetic terms), while nonfinancial equity falls by -36 log-points (-30%). Over the subsequent five years, bank equity does not earn elevated returns relative to the country’s unconditional average — point estimates are consistently negative, and significantly so in years three and four after crisis onset. Bank equity does not recover to its pre-crisis level. By contrast, nonfinancial equity earns cumulative abnormal returns of roughly 30 log-points (35% arithmetic) over five years, recovering to pre-crisis trend, consistent with a discount-rate-driven decline for nonfinancial firms.

Earnings-driven, not discount-rate-driven. Panel regressions at both the country and individual-bank level show coefficients of roughly 1 to 2 on the relationship between the initial bank equity return in the crisis year and the subsequent five-year change in real dividends and real earnings. The initial equity decline thus predicts a roughly commensurate long-run decline in banks’ dividends and earnings, inconsistent with the temporary-loss view’s prediction of discount-rate-driven declines that should subsequently reverse.

Short-run bounce-backs are modest and transient. At the monthly frequency, bank equity does rebound modestly from its trough — the bounce-back averages only about 30% of the initial decline, even assuming perfect market timing. This gain partially reverses after approximately twelve months, so cumulative five-year returns remain not elevated above the unconditional average.

Write-downs, not fire sales, drive losses. Realized book losses in the first year of crisis onset account for only about 30% of market-value losses — contrary to what fire-sale models predict. By year five, cumulative book losses reach roughly 35% of pre-crisis book equity and approximately 100% of market-value losses. Decomposing net income, write-downs track cumulative book losses closely and fully account for market-value losses by year five. Trading losses (from securities sales and asset dispositions) account for only a small share on average, though for banks in the top quartile of securities-to-assets ratios, immediate accounting losses are larger and more trading-loss-driven — consistent with fire-sale dynamics being important specifically for banks with large tradable securities portfolios.

Nonperforming loans confirm the mechanism. At the country level, larger bank equity declines are associated with higher peak NPL rates in the subsequent five years (adjusted R² of 0.53 excluding two outliers; 0.606 for the 2008-2010 subsample only). No analogous relationship exists for nonfinancial equity returns.

Policy interventions are insufficient. Liquidity-based interventions (extraordinary central bank support and blanket guarantees) implemented after bank equity crises are followed by an approximately 20% short-run rebound in bank equity, which reverses between months 12 and 36. No large or permanent increase in bank value follows. Government recapitalization programs have historically been small (averaging 24% of pre-crisis book equity and 43% of realized losses), narrow (65% classified as narrow, median of five banks recapitalized), and delayed. Banks cannot self-recapitalize through high post-crisis profitability.

Crisis type matters. Panic-only crises (banking panics without large bank equity declines, N=85) exhibit very different dynamics: bank equity recovers to pre-crisis levels within five years, dividends fall only temporarily, liquidity interventions produce large and permanent rebounds, and macroeconomic output losses are smaller. In 75% of bank equity crises, the bank equity decline strictly precedes the banking panic, indicating that fundamental weaknesses — not liquidity shocks escalating into solvency problems — are the primary driver. Only 19 cases (25%), labelled “mismanaged banking panics” (including the U.S. Great Depression), saw the panic precede the equity decline, mostly in the pre-1945 Gold Standard era. Early liquidity intervention is essentially a necessary condition for averting incipient crises, but it is effective only when a steep bank equity decline has not yet occurred.

In depth

Q1. How do the authors define a “bank equity crisis” and why does the definition matter for their empirical strategy?

A bank equity crisis is defined as the first year when (1) the bank equity index declines by more than 30% in annual excess total returns in any year within the past five years, and (2) a top-20 bank (ranked by assets) fails within the country. This purely data-driven, real-time definition avoids the look-ahead bias inherent in narrative-based chronologies. The authors identify 76 such crises. Results are robust to using Reinhart-Rogoff, JST, Laeven-Valencia, and 30%-decline-only definitions, alleviating concerns that the differential bank versus nonfinancial equity dynamics are mechanical artifacts of the crisis identification approach.

Q2. What is the quantitative magnitude of the initial equity shock to banks versus nonfinancial firms at crisis onset?

In the year of a bank equity crisis, the average abnormal cumulative log excess total return is -68 log-points for bank equity and -36 log-points for nonfinancial equity (corresponding to -49% and -30% in arithmetic abnormal returns, respectively). These are relative to the country’s unconditional average returns, estimated using country fixed effects in panel regressions.

Q3. Do bank stocks earn elevated returns after banking crises, as temporary-loss models predict?

No. Over the five years following crisis onset, bank equity point estimates of cumulative abnormal returns are consistently negative, and significantly so at years three and four. Bank equity does not recover to its pre-crisis level at any horizon out to five years (and Figure A.9 extends to ten years with similar conclusions). This pattern holds across advanced and emerging economies, before and after 1945, excluding the Global Financial Crisis, and across a variety of methods for computing abnormal returns. Even for surviving banks — excluding those that failed or exited — the pattern holds.

Q4. How do the earnings and dividend dynamics of banks versus nonfinancial firms differ after crises?

For banks, both real dividends per share and real earnings per share remain well below their long-term average five years after crisis onset, with no recovery visible by year five. For nonfinancial firms, dividends and earnings decline at crisis onset but rebound, though only slowly through year five. Panel regressions at both the country and individual-bank level find coefficients of approximately 1 to 2 on the relationship between the crisis-year bank equity return and the five-year-ahead change in real dividends and real earnings — indicating a roughly commensurate earnings-driven decline, not a transitory discount-rate shock.

Q5. What is the magnitude of the short-run bounce-back in bank equity, and does it represent a profit opportunity?

Even with perfect knowledge of the crisis trough (which is not available in real time), the rebound in bank equity from trough to peak averages only about 30% of the initial decline. This gain partially reverses within approximately twelve months, so that cumulative five-year abnormal returns remain not elevated above the unconditional average. Trading strategies that account for risk and factor returns (market, value, size, momentum, global equity) yield even lower risk-adjusted returns, strengthening the conclusion that bank equity is not cheap at crisis troughs.

Q6. How do write-downs compare to trading losses in explaining the accounting losses of banks during crises?

Realized book losses in the first year of crisis onset account for only about 30% of market-value losses. By year five, cumulative book losses reach approximately 35% of pre-crisis book equity and roughly 100% of market-value losses. Decomposing net income, write-downs (revaluations of assets remaining on the balance sheet — loan loss provisions, impairments, goodwill write-downs) track cumulative book losses closely and fully account for market-value losses by year five. Trading losses (realized gains and losses from securities trading and all asset sales) account for only a small share of total losses on average.

Q7. Under what conditions do fire sales rather than write-downs dominate the accounting losses?

For banks in the top quartile of the ratio of securities to total assets, immediate accounting losses in the first year of crisis onset are substantially larger and driven to a significant extent by trading losses rather than write-downs. The six bank equity crises with the highest securities-to-assets ratios (weighted across banks) all occurred during the 2007-2008 crisis (Belgium, France, Germany, Switzerland, the U.K., and the U.S.), when fire sales of securitized assets were significant. Banks holding mostly loans (bottom quartile of securities-to-assets) show slower-to-materialize book losses driven predominantly by write-downs.

Q8. How do nonperforming loan rates relate to the magnitude of bank equity declines across crises?

At the country level, more negative unlevered bank equity returns at crisis onset are statistically significantly associated with higher peak NPL rates over the subsequent five years. The adjusted R² for the full available sample is 0.233, rising to 0.533 after excluding two outliers (U.S. 1990, Sweden 1991). For the 2008-2010 crisis episodes only, the adjusted R² is 0.606. No analogous association between NPL rates and nonfinancial equity returns is found, suggesting the mechanism is specific to the banking sector’s asset-quality deterioration.

Q9. Do liquidity-based interventions (central bank support or blanket guarantees) restore bank capitalization after bank equity crises?

No. Following the implementation of liquidity-based interventions during bank equity crises, bank equity prices initially continue to decline for about two months, then rise by approximately 20%, but this gain reverses between months 12 and 36. Bank equity values remain persistently low thereafter. This is inconsistent with models in which forceful lender-of-last-resort interventions accomplish the same result as direct recapitalizations. The authors caution that interventions are not randomly assigned — deeper crises may receive stronger interventions — so the analysis cannot identify counterfactual outcomes.

Q10. What are the historical characteristics of government recapitalization programs?

Based on a new database covering all government recapitalization programs across 17 economies since 1870, recapitalizations have historically been small (averaging 24% of pre-crisis book equity and 43% of realized market-value losses), narrow (65% classified as narrow, with a median of five banks recapitalized), and delayed. Total equity issuance (government and private combined) is only a small fraction of realized losses. Government-funded issuance accounts for about one-fourth of total bank equity issuance. The U.S. TARP after 2008 was unusual in being both broad (over 700 banks) and timely (about one month after the Lehman collapse). Japan’s crisis of the 1990s is a prominent example of extreme delay, with the first recapitalization program implemented in March 1999, nearly a decade after the real estate collapse began.

Q11. How do “panic-only crises” differ from bank equity crises in terms of equity dynamics and policy effectiveness?

Panic-only crises (N=85) are banking panics without a 30% bank equity decline. They feature significant initial negative returns followed by elevated bank equity returns that bring valuations back to pre-crisis levels within five years. Dividends fall only temporarily. Liquidity interventions during panic-only crises produce a full rebound in bank equity in the month of intervention, contrasting sharply with the modest and transient response observed in bank equity crises. Panic-only crises are also associated with shallower real GDP declines and smaller bank credit contractions than bank equity crises.

Q12. In what fraction of bank equity crises does the bank equity decline precede the banking panic, and what does this imply about the root cause?

In 57 of the 76 bank equity crises (75%), the bank equity decline strictly precedes the emergence of the banking panic. This timing implies that most bank equity crises are not liquidity shocks that evolved into solvency problems — rather, fundamental weaknesses in the banking system are already present at the early stages of the crisis. Only 19 cases (25%), called “mismanaged banking panics,” saw the panic precede the equity decline; these occurred predominantly in the pre-1945 period, often in countries on the Gold Standard with limited central bank capacity.

Q13. Under what conditions can early liquidity interventions avert an incipient banking crisis?

Of 183 episodes of incipient liquidity shocks in which a prior 30% bank equity decline had not yet occurred, 126 received early liquidity interventions, of which 92 were successfully averted (approximately 50% of the original 183 episodes). The two strongest predictors of a successfully averted crisis — essentially necessary conditions — are: (1) the pre-panic bank equity decline remains below 30%, and (2) liquidity intervention occurs within one month of the panic. War outbreak and single-bank focus of the run are additional factors that substantially increase the probability of aversion. Combining the small-equity-decline and early-intervention conditions predicts averted panics with a true-positive rate of 99% (91/92), though with a 24% false-positive rate.

Q14. Does cross-sectional heterogeneity at the bank level confirm the permanent-loss interpretation?

Yes. Sorting the ten largest banks by country into five bins by market-to-book (M/B) ratio at crisis onset shows monotonic relationships with five-year outcomes. The most distressed banks (M/B below 0.2) experience reduced credit growth of 26 percentage points and reduced income-to-book-equity of 87 percentage points (both cumulative over five years) relative to the healthiest banks (M/B above 0.8). The M/B ratio at crisis onset is persistently low in subsequent years, because market values crash permanently while book values are sticky (slow write-down recognition). These results hold with crisis fixed effects, meaning the patterns reflect within-crisis cross-sectional variation, not merely crisis-level heterogeneity.

Q15. Do crises preceded by credit booms have worse post-crisis outcomes for banks?

Yes. Crises preceded by above-median growth in the credit-to-GDP ratio (from pre-crisis trough to peak) are associated with an additional 60 log-point abnormal decline in bank equity excess total returns occurring around year three after crisis onset, persisting through year five. By contrast, crises not preceded by credit booms earn bank equity returns similar to the country’s unconditional average after the initial decline. This supports the hypothesis that credit-boom-driven crises involve unexpected future deterioration in asset quality, possibly linked to persistently negative housing returns (which do not recover to pre-crisis levels within five years after banking crises).

Key Concepts

Bank equity crisis (paper-specific definition): An episode identified in real time when two criteria are jointly met for the first time: (1) the bank equity index declines by more than 30% in annual excess total returns within any year of the past five years, and (2) a top-20 bank (ranked by total assets within the country) fails. This definition is purely data-driven and does not require any look-ahead information. It produces 76 crises across 46 economies from 1870 to 2019.

Permanent-loss view: The theoretical interpretation that banking crises primarily reflect fundamental, lasting deterioration in the value of bank assets — arising either from fire sales that permanently destroy value or (more commonly in the authors’ evidence) from deterioration in asset quality (rising nonperforming loans, loan impairments). Under this view, bank equity declines are earnings-driven rather than discount-rate-driven and do not reverse even after funding and market liquidity are restored.

Temporary-loss view: The theoretical interpretation that bank losses during crises are primarily due to temporary price dislocations — assets held by financial intermediaries trade at sharp discounts due to binding borrowing constraints or depositor fragility, but recover their fundamental value once central banks provide liquidity support. Under this view, bank equity should earn elevated future returns after crises, and forceful liquidity interventions should be equivalent to direct recapitalizations in restoring bank value.

Write-downs (paper-specific definition): Revaluations of assets that remain on the balance sheet, reflecting expected future reductions in cash flows. They include loan loss provisions, additions to loan loss reserves, write-downs of fixed assets, and goodwill impairments. Distinguished from trading income (realized gains and losses from securities trading and all asset dispositions). Write-downs are subject to accounting discretion and are recognized slowly over multiple years after crisis onset, while equity markets price in expected total losses rapidly at crisis onset.

Trading income (paper-specific definition): Realized gains and losses from securities trading and all asset sales, including sales of real estate, loans, and subsidiary divisions. Unlike write-downs, trading losses must be recognized immediately (they are realized transactions), so large trading losses at crisis onset would be evidence consistent with fire-sale dynamics.

Panic-only crises: Banking panics (sustained bank runs or depositor withdrawals) that do not coincide with a greater-than-30% bank equity decline. Identified as N=85 in the full sample. These episodes are characterized by temporary equity declines, full recovery within five years, large positive responses to liquidity interventions, and smaller macroeconomic output losses — consistent with the temporary-loss view.

Mismanaged banking panics: The minority of bank equity crises (19 cases, 25%) in which the banking panic occurred first or concurrently with the 30% bank equity decline, rather than the equity decline preceding the panic. Concentrated in the pre-1945 period, often in Gold Standard countries with limited central bank flexibility. The U.S. Great Depression is the prominent example.

Averted crisis: An incipient liquidity shock to the banking sector that fully recedes within two months without any bank failures or 30% bank equity declines. Empirically, all averted crises in the sample had not yet experienced a 30% bank equity decline and all received early liquidity interventions (within one month of the incipient panic onset).

Praying for Rain

Mon, 01 Jan 0001 00:00:00 +0000

This paper studies rainmaking as an instrumental religious belief. The central research question is: why do people believe that prayer can bring rain, even though it does not work? The authors develop a model of cultural evolution in which a religious leader prays for rain at an arbitrary time, and people update their beliefs about whether the leader can cause rainfall based on whether rain follows. The key mechanism is the local rainfall hazard function — the probability of rain conditional on how many days have passed since the last rainfall. In environments where the hazard is increasing (rain becomes more likely the longer a drought continues), a leader who prays during a drought will tend to be followed by rain, creating the illusion of efficacy. In environments with a flat or declining hazard, prayer cannot be systematically followed by rain in a persuasive way. The model yields five predictions: rain ritual traditions will select for prayers correlated with rainfall; the level of average rainfall does not determine persuasiveness; constant-hazard environments cannot support persuasive prayer; increasing-hazard environments are more likely to adopt rainmaking; and higher net benefits of rainfall (e.g., settled agriculture) further increase the likelihood of ritual.

The authors test these predictions with two empirical strategies. First, they use daily data from the Catholic church in Murcia, Spain, covering 1600 to 1836. Church records provide the daily timing of pro pluvia rogations (prayers for rain), while municipal council records — kept independently of the church — record notable rainfall events. Murcia’s rainfall hazard is estimated to be increasing after long dry spells: the hazard rate after a long drought is roughly double the hazard rate two months after the last rainfall. The main finding is that a prayer for rain in the last 30 days predicts a 0.144 percentage-point higher daily probability of notable rainfall (standard error 0.057 pp), relative to a baseline mean daily rainfall probability of 0.203 pp — a 71% increase in the predicted probability. Prayer also Granger-causes rainfall conditional on lags of recent rainfall, and the predictive power holds within a given calendar month, ruling out a purely seasonal coincidence.

Second, the authors construct an original dataset covering rainmaking practice for 1,208 ethnic groups drawn from the Ethnographic Atlas (Murdock, 1967), coded from 370 anthropological sources. They match each ethnic group to its nearest weather station and estimate the rainfall hazard function each group faces in its ancestral location. Of the 1,208 groups, 33% face an increasing rainfall hazard, and 39% of all groups practice rain ritual. The main global finding is that ethnic groups facing an increasing rainfall hazard are 14 percentage points more likely to practice rainmaking (standard error 3.7 pp), relative to a base rate of 30% among groups facing a non-increasing hazard — a 47% increase. This result is robust to continent fixed effects, geographic and climatic controls (longitude, latitude, elevation, distance to coast, ruggedness, mean temperature, mean rainfall, coefficient of variation of rainfall, maximum dry spell length, and the Giuliano-Nunn 2021 climatic variability measure), alternative hazard estimation methods, and linguistic family fixed effects. Crucially, lower average rainfall, longer droughts, and greater climatic variability are not associated with more rain ritual conditional on hazard shape — it is specifically the shape of the hazard function, not aridity or variability per se, that drives adoption.

A second global finding concerns demand: groups dependent on agriculture are 11 pp more likely to practice rainmaking; those dependent on intensive agriculture, 21 pp more likely; and those dependent on intensive irrigated agriculture, 32 pp more likely (on a base of 32%). The scope of the findings is the pre-modern or traditional period captured by the Atlas; the Murcia case covers 1600–1836. The authors conclude that some environments create an illusion of efficacy that sustains instrumental religious belief through cultural selection, without requiring that believers be irrational.

Q: What is the paper’s central theoretical claim about why rainmaking beliefs persist? A: The paper argues that in environments where the rainfall hazard is increasing during a drought, a leader who begins praying during a dry spell will tend to be followed by rain, because the probability of rain rises as the drought lengthens. People who cannot observe the counterfactual hazard (what rainfall would have been without prayer) interpret this coincidence as evidence that prayer works. Cultural selection then favors leaders whose prayer timing is more persuasive, causing the belief to persist across generations even though prayer does not actually cause rain.

Q: What is the rainfall hazard function, and why does its shape determine whether prayer can be persuasive? A: The hazard function h(t) gives the instantaneous probability of rain at time t days after the last rainfall. If the hazard is flat, the probability of rain is the same regardless of whether prayer was offered or not, so there is no systematic correlation between prayer and rainfall to exploit. If the hazard is declining, prayer during a drought will be followed by lower-than-average rainfall probability, undermining the leader. Only if the hazard is increasing does prayer during a long dry spell systematically coincide with a higher probability of rain, creating a persuasive correlation.

Q: What do Propositions 2 and 3 of the model establish? A: Proposition 2 establishes that if the hazard rate is constant and a person’s prior belief that prayer works is below 0.5, then no prayer start time can persuade them to support the leader. Proposition 3 establishes the converse: if the hazard rate is increasing and the prior is below 0.5, there exists a meaningful belief for which a person will support the leader for any prayer start time. Together these propositions identify the increasing hazard as the necessary and sufficient structural condition for persuasive prayer.

Q: What is the main quantitative finding from Murcia, and what identification strategy supports it? A: A prayer for rain in the last 30 days predicts a 0.144 percentage-point higher daily probability of notable rainfall (standard error 0.057 pp) relative to a baseline mean of 0.203 pp, a 71% increase. The authors additionally demonstrate that prayer Granger-causes rainfall conditional on lags of recent rainfall, and that the effect holds within a given calendar month, ruling out the explanation that prayer simply tracks the rainy season. The prayer and rainfall records are kept by independent institutions (church and municipal council), reducing the risk of strategic recording.

Q: How does the hazard rate in Murcia behave, and does it satisfy the model’s key condition? A: The hazard of rainfall in Murcia is initially high just after rain, declines to a minimum roughly two months after the last rainfall, and then increases significantly thereafter, reaching or exceeding its initial level after a long drought. The fluctuations are large: the hazard after a long dry spell is roughly double the hazard two months after rainfall. This U-shaped pattern means the hazard is increasing during a prolonged drought, satisfying the model’s key condition for persuasive prayer.

Q: How was the global rainmaking dataset constructed, and what is its coverage? A: The authors used the Ethnographic Atlas (Murdock, 1967) as a template, covering 1,290 ethnic groups, and combed 370 anthropological sources — primarily group-specific ethnographic monographs — to code rainmaking practice for 1,208 groups. A group is coded as practicing rain ritual only if there is clear evidence of a practice specifically intended to bring rain through supernatural means. The authors treat their measure as a lower bound. They find that 39% of the 1,208 groups practice rainmaking, across every settled continent.

Q: What is the main global regression result and how robust is it? A: Ethnic groups facing an increasing rainfall hazard are 14 percentage points more likely to practice rain ritual (standard error 3.7 pp) relative to a base rate of 30%, a 47% proportional increase. This coefficient is positive and statistically significant across all specifications, including those adding continent fixed effects, a full battery of geographic and climatic controls (longitude, latitude, elevation, distance to coast, ruggedness, mean temperature, mean rainfall, coefficient of variation of rainfall, maximum dry spell length, and the Giuliano-Nunn 2021 climatic variability measure), alternative hazard estimation methods, linguistic family fixed effects, and restrictions to groups with high-quality rainfall data.

Q: Does aridity or climatic variability explain rainmaking adoption? A: No. Lower average rainfall, longer droughts, and greater climatic variability (measured using the Giuliano-Nunn 2021 index) are not associated with more rain ritual practice, conditional on the shape of the hazard function. This rules out the naive hypothesis that people pray for rain simply because they do not get enough, or because their rainfall is unreliable. It is specifically the shape of the hazard — whether it is increasing during a drought — that drives adoption, not the level or volatility of rainfall.

Q: How does demand for rainfall, proxied by agricultural subsistence, affect rainmaking adoption? A: Groups dependent on agriculture are 11 percentage points more likely to practice rainmaking relative to other subsistence modes. Groups dependent on intensive agriculture are 21 percentage points more likely, and groups dependent on intensive irrigated agriculture are 32 percentage points more likely, all on a base of 32%. This gradient is consistent with Proposition 5 and 6 of the model: settled, location-specific agricultural investment raises the net benefit of rainfall control, increasing support for rain ritual independently of the persuasion channel.

Q: What does the model’s cultural evolution mechanism (Proposition 4) predict about how prayer timing changes over generations? A: Proposition 4 states that rituals with high support are more likely to persist. In increasing-hazard environments, random variation in prayer timing means some leaders gain more support than others; those with more persuasive timing are more likely to persist. Each generation then adopts a policy at least as persuasive as the prior generation, so support rises over time and prayers gradually converge toward the timing that maximizes persuasiveness. This mechanism does not require deliberate optimization by any individual leader.

Q: How does the paper’s finding relate to the long-standing anthropological debate between the traditional and revisionist schools on rainmaking? A: The traditional school (following Frazer 1890) holds that belief is instrumental — people engage in rainmaking to make rain, and belief responds to empirical evidence. The revisionist school (Wittgenstein, Durkheim) argues that religious belief and rationality are fundamentally separate, and religious practice is performative rather than evidence-responsive. The paper’s finding that rainmaking is more prevalent precisely where it is more persuasive — i.e., where the environment makes prayer appear to work — supports the traditional, instrumental interpretation that belief responds to evidence of efficacy.

Q: What are the scope conditions for the paper’s conclusions? A: The Murcia case study covers the period 1600–1836, ending when the abolition of tithes reduced the church’s funding and influence; it applies to a sophisticated Catholic institutional context. The global analysis covers traditional practices of pre-modern ethnic groups as recorded in the Ethnographic Atlas and anthropological literature; it does not speak to modern religious practice or to religions after substantial modernization. The persuasion mechanism requires that people cannot directly observe what rainfall would have been without prayer, a condition satisfied in pre-scientific contexts.

Rainfall hazard function: In this paper’s usage, the function h(t) = f(t)/(1-F(t)) giving the instantaneous probability of rainfall at time t days since the last rainfall. Its shape — whether flat, declining, or increasing during a drought — determines whether prayer can be persuasive, not the overall level of rainfall.

Increasing hazard: A hazard rate that rises as the length of a dry spell increases, so that rain becomes more likely the longer the drought has continued. The paper defines this specifically as the derivative of the hazard function evaluated at the 99th percentile of spell length. This is the necessary structural condition for prayer to seem efficacious.

Instrumental religious belief: Belief directed at achieving a worldly outcome (here, rainfall), as opposed to purely expressive or social belief. The paper treats belief as instrumental if it responds to perceived evidence of efficacy and is adopted where it appears to work.

Persuasion (in the model): The process by which a leader’s prayer timing causes people to update their belief that prayer works, by generating a correlation between prayer and subsequent rainfall that exceeds what people expect from the background hazard rate. Persuasion is possible only when the hazard is increasing.

Pro pluvia rogations: The Catholic church’s formal prayers for rain, practiced in Murcia since at least the 14th century. In the paper’s data, these prayers follow a pattern of escalation — increasing in number and intensity — during prolonged droughts, consistent with the model’s prediction about prayer timing.

Cultural evolution: The paper’s framework (drawing on Henrich 2015) in which religious leaders act as cultural entrepreneurs; leaders whose prayer timing happens to be more persuasive gain greater support and are more likely to survive across generations, so prayer traditions drift toward more persuasive timing without deliberate design.

Rain ritual (global measure): A binary indicator coded as one for an ethnic group if the anthropological literature contains clear evidence of a practice specifically intended to bring rain through supernatural means, including dances, sacrifices, prayers, and petitioning of rain deities. Treated by the authors as a lower bound on actual prevalence.

Present Bias Amplifies the Household Balance-Sheet Channels of Macroeconomic Policy

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Summary

Maxted, Laibson, and Moll study fiscal and monetary policy in a partial-equilibrium heterogeneous-agent model in which homeowners have present-biased time preferences (Instantaneous Gratification preferences, the continuous-time limit of quasi-hyperbolic discounting) and naive beliefs, alongside a liquid savings account, an illiquid home, and access to credit card and mortgage debt. Because present bias substantially increases households’ marginal propensity to consume — in the calibrated model the quarterly MPC rises from 4% under exponential discounting to 14% under present bias, and the quarterly marginal propensity for expenditure (MPX) rises from 13% to 30% — present bias powerfully increases the effect of fiscal stimulus. Present bias also amplifies the overall effect of expansionary monetary policy, but at the same time slows down the speed of monetary transmission: interest rate cuts incentivize households to conduct cash-out refinances, which become targeted liquidity injections to households near the liquidity constraint who have especially high MPCs, but present bias with naive beliefs also introduces a motive for households to procrastinate on refinancing their mortgage, which substantially slows the speed at which this channel operates. A noteworthy feature of the model is that present bias amplifies the direct effect of monetary policy on household consumption while simultaneously delivering larger MPCs — a combination that is in contrast to standard heterogeneous-agent models, where modeling choices that amplify MPCs typically deliver smaller consumption responses to interest rate changes. The calibrated present-biased economy also replicates several empirical regularities that are difficult to match with exponential discounting: high-cost credit card borrowing by homeowners, empirically plausible cash-out behavior and loan-to-value ratios, and refinancing inertia.

In depth

Q1. Q: What is the core modeling innovation and why is it needed?

A: The paper introduces naive Instantaneous Gratification (IG) preferences — the continuous-time limit of quasi-hyperbolic (beta-delta) discounting — into a two-asset heterogeneous-agent model with a liquid savings account and illiquid home equity accessible via mortgage refinancing. The naivete assumption (households do not foresee their own future present bias) is essential because it generates procrastination: naive households perpetually intend to refinance “soon” but keep delaying. A model with exponential discounting that merely sets parameters to match empirical MPCs would not generate procrastination behavior, and would require implausible interest rate calibrations (very low credit card rates or very high illiquid asset returns) to simultaneously match low liquid wealth accumulation and high credit card borrowing. Present bias with interest rates taken from the data resolves both issues.

Q2. Q: What are the key quantitative MPC results and why do they matter for fiscal policy?

A: In the exponential discounting benchmark, the quarterly MPC is 4% and the quarterly MPX (which includes nondurables and durables) is 13%. Under the present-bias benchmark, the MPC rises to 14% and the MPX rises to 30%. The empirical literature estimates quarterly nondurable spending responses on the order of 15%–25%, and total expenditure responses typically two to three times larger, so the present-biased model is substantially more consistent with the data. Because fiscal stimulus (modeled as an unexpected one-time lump-sum payment, financed by a flow income tax) operates through household spending propensities, the higher MPCs and MPXs under present bias directly and powerfully increase the aggregate consumption response to fiscal policy relative to the exponential benchmark.

Q3. Q: How does present bias amplify the effect of monetary policy?

A: Interest rate cuts incentivize households to conduct cash-out refinances — they borrow against accumulated home equity, converting illiquid home equity into liquid wealth. Because this liquidity is targeted to households who are near their borrowing constraint (and thus have especially high MPCs), the aggregate consumption response to a given rate cut is amplified. Crucially, present bias amplifies this channel beyond the exponential benchmark precisely because higher MPCs mean each dollar of liquidity injected generates more consumption. This stands in contrast to the standard result in the heterogeneous-agent literature (Auclert 2019; Olivi 2017; Kaplan, Moll, and Violante 2018) that MPC-amplifying modeling choices reduce the consumption response to interest rate changes because MPC enters the substitution effect with a negative sign in standard one-asset models. The two-asset structure with home equity and the cash-out refinance channel breaks this trade-off.

Q4. Q: How does present bias slow the speed of monetary transmission?

A: Present bias with naive beliefs introduces a motive for households to procrastinate on refinancing their mortgage. Refinancing is an immediate-cost, delayed-reward task: it requires the borrower to spend weeks gathering documents, filling out paperwork, and negotiating with lenders, with benefits (lower mortgage payments or extracted home equity) accruing afterward. Naive present-biased households discount current effort costs very heavily relative to future benefits, so they delay, all the while (counterfactually) believing they will complete the task in the near future. This procrastination substantially slows down the speed at which the cash-out refinance channel of monetary policy operates: even though a rate cut eventually incentivizes households to refinance and extract equity, the timing of that response is stretched out relative to what exponential discounters would do.

Q5. Q: What is the role of naive beliefs versus sophisticated (partially or fully aware) present bias?

A: Naivete is necessary to generate procrastination from small effort costs. A fully sophisticated present-biased household (one who correctly anticipates its own future self-control problems) would not indefinitely defer a task it correctly anticipates will keep being deferred. The paper extends the analysis to partial and full sophistication in Online Appendix D.5. The key takeaway is that procrastination — and thus the speed-reduction effect on monetary transmission — is driven by at least partial naivete. The MPC-amplification and fiscal-policy amplification results are more robust across sophistication levels.

Q6. Q: What empirical regularities does the present-biased calibration match that the exponential model cannot easily match?

A: The present-biased economy replicates: (1) empirically plausible levels of high-cost credit card debt held simultaneously with home equity (a puzzle under exponential discounting); (2) cash-out behavior and loan-to-value ratios consistent with data; (3) a buildup of liquidity-constrained households consistent with empirical propensities to spend out of credit card limit increases (Gross and Souleles 2002; Agarwal et al. 2018); (4) consumption function discontinuities at the borrowing constraint consistent with Ganong and Noel (2019); (5) MPCs and MPXs that remain elevated for large shocks (Fagereng, Holm, and Natvik 2021); (6) the intertemporal MPC profile consistent with Auclert, Rognlie, and Straub (2018); (7) differential MPCs out of liquid versus illiquid transfers (Ganong and Noel 2020); and (8) refinancing inertia — the proclivity for households to delay refinancing when financially optimal (Keys, Pope, and Pope 2016; Johnson, Meier, and Toubia 2019; Andersen et al. 2020).

Q7. Q: What is the model’s scope — what does it abstract from?

A: The model is set in partial equilibrium, so general equilibrium effects (e.g., endogenous interest rate responses, aggregate demand externalities) are not captured; the authors describe their results as inputs for a fuller general equilibrium analysis. The model focuses on homeowners (two-thirds of U.S. housing units), abstracting from renters. House prices are fixed (consistent with their slow movement over short horizons), with an extension to house price shocks in Online Appendix D.2.1. The model does not allow for home equity lines of credit, second mortgages, or reverse mortgages, because these products are more commonly used when interest rates are rising, and the paper focuses on the stimulative effect of rate cuts. The interest rate in the model is a long rate (e.g., 10-year TIPS), with the implicit assumption that the Federal Reserve implements the necessary short-rate adjustments.

Q8. Q: How does the present-biased model compare to the standard HANK picture on the monetary-MPC trade-off?

A: In standard one-asset heterogeneous-agent models, a household’s MPC is a sufficient statistic that enters the substitution effect of interest rate changes with a negative sign — so modeling choices that raise MPCs reduce monetary policy effectiveness. The present-biased two-asset model breaks this result: because interest rate cuts trigger cash-out refinances that inject liquidity targeted to high-MPC households near the constraint, higher MPCs translate into larger, not smaller, aggregate consumption responses to monetary policy. Present bias therefore simultaneously amplifies fiscal policy (via higher MPCs) and amplifies the overall effect of monetary policy (via the targeted liquidity channel), while introducing the procrastination-driven speed reduction as the offsetting cost.

Key Concepts

Present bias (Instantaneous Gratification preferences): The paper uses “present bias” to refer to quasi-hyperbolic discounting. In the continuous-time limit (Instantaneous Gratification, or IG, preferences, following Harris and Laibson 2013), the current self discounts all future selves by factor β < 1, while exponential discounting of the future (rate ρ) applies from any future vantage point. This creates a discontinuity in the discount function at t = 0 whenever β < 1. Setting β = 1 recovers standard exponential discounting.

Naive beliefs: Households do not foresee their own future present bias. The current self believes all future selves will be exponential discounters (β = 1), even though this belief is incorrect. Naivete is what transforms present bias into procrastination: the household perpetually expects its future self to complete effortful tasks, but each future self faces the same bias.

Cash-out refinance channel: When market interest rates fall, households have an incentive to refinance their fixed-rate mortgage, locking in a lower interest rate. If the household has accumulated home equity (illiquid), it can simultaneously borrow against that equity — a cash-out refinance — converting illiquid home equity into liquid wealth. In the model, this acts as a targeted liquidity injection to households near their borrowing constraint (who have high MPCs), amplifying the aggregate consumption response to rate cuts.

Procrastination motive: Present bias introduces a motive to procrastinate on immediate-cost, delayed-reward tasks such as mortgage refinancing. The effort and paperwork costs of refinancing are borne immediately, while the financial benefits accrue over time. A naive present-biased household heavily discounts the current effort cost relative to future benefits, leading it to defer refinancing repeatedly. This substantially slows the speed at which the cash-out refinance channel of monetary policy operates.

Marginal propensity to consume (MPC) vs. marginal propensity for expenditure (MPX): The paper distinguishes the quarterly MPC (response of nondurable consumption to a one-unit cash transfer) from the quarterly MPX (which also includes durables). Under exponential discounting, MPC = 4% and MPX = 13%; under the present-bias benchmark, MPC = 14% and MPX = 30%. The higher MPXs are more consistent with empirical estimates (quarterly nondurable responses of 15%–25%; total spending responses two to three times larger).

Refinancing inertia: The empirical regularity that households delay mortgage refinancing even when it is financially optimal to do so. The paper provides a theoretical foundation for this behavior through the procrastination motive generated by naive present bias combined with the small effort cost of refinancing.

Summary based on LSE Research Online published version. AI-assisted, human review pending.

Republican Support and Economic Hardship: The Enduring Effects of the Opioid Epidemic

Mon, 01 Jan 0001 00:00:00 +0000

This paper establishes a causal connection between the opioid epidemic and the political realignment toward the Republican Party in the United States from the mid-2000s through 2022. The authors—Carolina Arteaga and Victoria Barone—exploit rich geographic variation in Purdue Pharma’s initial marketing strategy for OxyContin, drawn from unsealed litigation records, to construct a quasi-exogenous measure of community-level exposure to the epidemic.

The identification strategy rests on a documented feature of OxyContin’s 1996 launch: Purdue initially targeted the established cancer pain market—physicians and patients already using MS Contin—as an entry point into the much larger noncancer pain market. Areas with higher cancer mortality in 1996 received disproportionate pharmaceutical marketing, leading to outsized opioid prescription growth that spilled over from cancer patients to the broader population through shared physicians. The authors use 1996 commuting-zone (CZ) cancer mortality rates as a proxy for this initial targeting, interacted with year fixed effects in an event-study specification with CZ and state-year fixed effects. The sample covers 625 CZs across the continental United States from 1982 to 2022.

The empirical chain runs through three stages. First, the instrument strongly predicts opioid supply: by 2012, a one-standard-deviation higher 1996 cancer mortality rate led to an additional 0.97 opioid doses prescribed per capita, 65% above the baseline mean. Second, the resulting epidemic caused measurable mortality and economic hardship. A one-standard-deviation increase in 1996 cancer mortality caused drug-induced deaths in 2017 to be 46% above the pre-epidemic average; by 2012 the same increase caused prescription opioid deaths to be 61% higher. Excess mortality was concentrated among individuals under age 55, with no significant effects for those aged 55 and older. The epidemic also raised disability applications: SSDI applications rose by 12% and SSI applications by 7.6% by 2012, effects that persisted through 2020. SNAP enrollment in exposed CZs was 8% higher by 2022, equivalent to a 0.14 standard deviation increase.

Third, and centrally, the communities that endured these health and economic shocks shifted persistently toward the Republican Party. By the 2022 House elections, a one-standard-deviation increase in 1996 cancer mortality increased the Republican two-party vote share by 4.5 percentage points. Effects of similar magnitude appear in presidential elections (4.6 percentage points) and gubernatorial elections (4.3 percentage points). The vote-share shift is consistent across gender, age, race, and education, with no detectable change in voter turnout. The shift translates into actual seat gains: beginning in 2012, exposed areas consistently elected more Republican House members, moving the chamber’s roll-call voting in a more conservative direction. The effect is not driven by anti-incumbent sentiment—results hold regardless of which party held the seat at the time.

The paper identifies three reinforcing mechanisms. First, the Republican Party repositioned itself during this period as the advocate of “forgotten America” and working-class economic hardship, a message that resonated acutely in opioid-devastated communities. Second, conservative-leaning newspapers covered the epidemic at higher rates, and their coverage tracked local mortality; liberal-leaning outlets showed no such correlation. Fox News covered opioid stories at 1.5 times the rate of CNN and 1.7 times the rate of MSNBC, emphasizing crime, trafficking, and cartels at twice the frequency of liberal outlets. Third, exposed communities expressed stronger preferences for Republican-favored policy responses: higher police presence, greater sense of safety around law enforcement, and lower support for marijuana legalization on state ballot initiatives.

Pre-trend tests show no relationship between 1996 cancer mortality and outcomes before OxyContin’s launch. Out-of-sample exercises using 1976 cancer mortality find no analogous pattern in the pre-epidemic period (1982–1994). Placebo instruments based on unrelated causes of death yield null results. The baseline findings are robust to controlling for the China import shock, NAFTA, the 1994 Republican Revolution, the 2001 and 2008–2009 recessions, declining unionization, robot adoption, Fox News introduction, deaths of despair, and Southern and rural political realignment.

Q: What is the paper’s central research question? A: The paper asks whether the opioid epidemic causally increased Republican vote share in communities most severely affected by the crisis. It documents a causal chain from pharmaceutical marketing through drug mortality and economic hardship to political realignment, contributing the first causal estimate of a major public health crisis’s effect on partisan voting.

Q: What is the identification strategy, and why is 1996 cancer mortality a valid instrument? A: Purdue Pharma explicitly targeted physicians in the cancer pain market at OxyContin’s 1996 launch, then used those established relationships to expand into the noncancer pain market. CZs with higher cancer mortality in 1996 received disproportionate marketing, generating differential opioid prescription growth unrelated to pre-existing political or economic trends. Pre-trend tests confirm no differential patterns before 1996, out-of-sample tests using 1976 cancer mortality find no relationship with pre-epidemic outcomes, and placebos using unrelated causes of death yield null results.

Q: How strong is the first stage linking 1996 cancer mortality to opioid prescriptions? A: The relationship between 1996 cancer mortality and opioid prescriptions is positive and statistically significant from 1998 through 2020. By 2012—the year prescription rates peaked nationally at 81.3 per 100 persons—a one-standard-deviation higher cancer mortality rate led to an additional 0.97 morphine-equivalent doses prescribed per capita, 65% above the baseline mean. CZs in the highest cancer mortality quartile experienced a 1,800% increase in grams of oxycodone per capita between 1997 and 2010, compared to less than half that in the lowest quartile.

Q: What are the effects on drug-induced mortality? A: Drug-induced mortality (a broad measure covering deaths from prescription opioids, heroin, and fentanyl) rose continuously in exposed CZs after 1996. By 2017, a one-standard-deviation increase in 1996 cancer mortality caused drug-induced deaths to be 46% above the pre-epidemic average. By 2012, the same increase caused prescription opioid deaths specifically to be 61% higher relative to the pre-epidemic average. Excess mortality was concentrated among individuals under age 55, with no statistically significant effects for those aged 55 and older.

Q: What are the effects on disability program take-up? A: Applications for Social Security Disability Insurance (SSDI) rose by 12% and Supplemental Security Income (SSI) applications rose by 7.6% by 2012 for a one-standard-deviation increase in 1996 cancer mortality. These effects persisted: SSDI recipients grew by 15% and SSI recipients by 3.2% by 2020 in similarly exposed CZs. The increases in disability were concentrated among individuals under age 55, paralleling the mortality effects.

Q: What are the effects on SNAP enrollment? A: Exposed CZs showed a continuous increase in SNAP enrollment over two decades following the epidemic’s onset. By 2020, a one-standard-deviation increase in 1996 cancer mortality corresponded to an 8% increase in the share of the population receiving SNAP benefits, equivalent to 0.14 standard deviations. By 2022, the corresponding figure remains 8%, indicating persistent economic strain in exposed communities.

Q: What is the magnitude of the political effects in House elections? A: A one-unit increase in the 1996 cancer mortality rate yielded a 7.9-percentage-point increase in the 2022 Republican House vote share relative to 1996. Scaled to one standard deviation (0.58 units), this corresponds to a 4.5-percentage-point increase in the Republican two-party vote share by the 2022 midterms. The vote-share shift became statistically significant beginning in 2006, but only translated into consistent seat-level Republican gains starting in 2012.

Q: When did opioid exposure start winning Republicans additional House seats? A: Although the Republican vote share in exposed areas began increasing around 2006, actual seat flips did not become consistent until 2012. The paper explains this lag by noting that initial vote-share gains were concentrated in communities with low baseline Republican support, where additional votes did not immediately cross the winning threshold. Starting in 2010, median-baseline-Republican CZs also began shifting, enabling additional seat changes.

Q: How large are the presidential and gubernatorial election effects? A: In presidential elections, a one-standard-deviation increase in 1996 cancer mortality raised the Republican vote share by 4.6 percentage points. In gubernatorial elections, the same increase raised the Republican vote share by 4.3 percentage points after approximately six election cycles (corresponding to 2017–2020). These effects are described as comparable in magnitude to the difference in Republican vote share between the top and bottom quartiles of NAFTA vulnerability.

Q: Does the political shift reflect increased polarization toward extremist candidates? A: No. The paper finds no increase in the probability of electing candidates at the extremes of the Nokken-Poole ideological scale in any given election year. The ideological shift in the House composition arises from changes in which party wins seats rather than from the election of more extreme Republicans. Campaign donations to Republican candidates did not increase; rather, donations to Democratic candidates declined (in 2016, a one-standard-deviation increase in cancer mortality widened the Republican-Democrat donation gap by 0.44 standard deviations). The shift is interpreted as a change in voting preferences in previously Democratic-leaning areas, not heightened polarization.

Q: Is the shift driven by anti-incumbent sentiment? A: The authors test this by splitting the sample by the incumbent’s party at the time of each election and by redefining the outcome as the incumbent’s vote share. Neither exercise produces evidence of a systematic anti-incumbent response. The changes in Republican vote share are not statistically distinguishable based on whether the incumbent was a Republican or Democrat. If anything, after 2016 there is a slight increase in the likelihood of incumbents retaining their seats.

Q: Where geographically are the Republican gains largest? A: Using state-level treatment effects estimated from an in-differences model interacting cancer mortality with state-year indicators, the paper finds a strong positive correlation between the magnitude of the epidemic’s effect on economic hardship (measured by SNAP participation) and the magnitude of the Republican vote-share increase. This correlation is strongest with a lag: SNAP effects measured in 2006 are most predictive of vote-share shifts in 2022, indicating that deterioration in community economic fabric preceded and predicted the political realignment.

Q: How did conservative and liberal media differ in covering the opioid epidemic? A: Republican-leaning local newspapers covered the opioid epidemic more extensively than Democratic-leaning papers throughout the epidemic period, and their coverage tracked local opioid mortality rates; Democratic-leaning coverage showed no such correlation with local incidence. Fox News covered opioid stories at 1.5 times the rate of CNN and 1.7 times the rate of MSNBC. In terms of content, Republican-leaning newspapers showed 23% higher frequency of economic hardship keywords, 19% higher frequency of illegal activity and crime keywords, and 22% higher frequency of rehabilitation and treatment keywords relative to Democratic-leaning papers. Fox News emphasized crime, drug trafficking, and cartels at double the frequency of more liberal outlets.

Q: How did voter policy preferences align with Republican versus Democratic platforms? A: Using 2020 CCES data, the authors find that higher 1996 cancer mortality predicts a greater expressed preference for increasing the number of police officers on the street and a greater reported sense of safety around law enforcement—both consistent with the Republican Party’s law enforcement approach. Conversely, exposure to the epidemic predicts lower support for marijuana legalization on state ballot initiatives across 18 states from 2012 to 2023, indicating opposition to a key Democratic harm-reduction policy.

Q: What role did political actors themselves play in driving the realignment? A: Relatively little. The opioid epidemic was largely absent from House floor speeches until 2015 and from campaign advertising until 2020. Neither party took a clear legislative lead on the issue during the first two decades of the crisis. The authors interpret the political realignment as driven primarily by the Republican Party’s broader repositioning as the champion of working-class economic hardship and by differential media framing, rather than by active legislative competition over opioid policy.

Q: What major confounds are ruled out? A: The authors control for exposure to the China import shock, NAFTA, the 1994 Republican Revolution, the 2001 and 2008–2009 recessions, declining unionization, robot adoption, Fox News entry, deaths of despair (which include but are not limited to opioid deaths), and the political realignment of the South, rural areas, evangelicals, and the population over 65. Results remain robust across all these specifications. Placebo instruments using unrelated causes of death yield null results.

Q: Could the vote-share effects be mechanically driven by opioid-related deaths removing Democratic voters from the electorate? A: The authors perform a back-of-the-envelope calculation and estimate that even if all opioid-related deaths would have been Democratic votes, the mechanical effect on the Republican vote share is at most 0.22 percentage points relative to the observed 2020 vote share—far smaller than the estimated 4.5-percentage-point shift by 2022. The result is also inconsistent with a turnout mechanism, as voter turnout shows no meaningful change with epidemic exposure.

Opioid epidemic exposure instrument: The paper measures community-level exposure to the opioid epidemic using cancer mortality rates in 1996, the year OxyContin launched. This instrument is grounded in Purdue Pharma’s documented marketing strategy of targeting the cancer pain market first; areas with more cancer patients received disproportionate pharmaceutical marketing, generating differential opioid prescription growth that extended well beyond cancer patients to the broader noncancer population through shared physicians.

Commuting zone (CZ): The paper’s unit of geographic analysis, defined to capture local economic markets. There are 720 CZs in the US, encompassing all metropolitan and nonmetropolitan areas. The authors use 625 CZs with more than 20,000 residents, which account for more than 99% of all opioid deaths and total population.

Two-party Republican vote share: The ratio of votes for Republican candidates to the total votes for both Republican and Democratic candidates in a given election. The paper tracks this measure for House, presidential, and gubernatorial elections from 1976 or 1982 through 2020 or 2022, depending on data availability.

Drug-induced mortality: The paper’s broadest mortality measure, covering deaths from poisoning and medical conditions caused by legal or illegal drugs, including prescription opioids, heroin, and synthetic opioids such as fentanyl. It is distinguished from the narrower measures of prescription opioid deaths and all opioid deaths.

Issue ownership: The political science concept, used in the paper to describe how the Republican Party repositioned itself during the epidemic period as the voice of working-class economic hardship, “forgotten America,” and “America left behind.” The paper contrasts this with Democratic ownership of income inequality and argues that Republican ownership of the hardship narrative made the party’s message especially salient in heavily opioid-affected communities.

Path dependency in pharmaceutical marketing: Purdue’s strategy of concentrating initial OxyContin promotion in cancer-market areas, then later focusing on top-prescribing physicians (the highest three deciles of the distribution), meant that areas receiving high initial cancer-market promotion continued to receive disproportionate promotion as the company expanded to the noncancer market. This created a persistent targeting advantage for high-cancer CZs throughout the epidemic’s first wave.

Nokken-Poole ideological measure: A roll-call-based measure of elected House members’ ideology along the liberal-conservative dimension. The paper uses this measure to show that the epidemic shifted the composition of the House toward more conservative members, not by electing more extreme candidates in any given election, but by changing which party won seats over time.

The Earnings and Labor Supply of U.S. Physicians

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. What do U.S. physicians earn, how is that earnings variation structured across geography and specialty, and how much does government healthcare payment policy shape those earnings and — through them — physicians’ labor supply and long-run talent allocation?

Data. The paper builds a novel administrative panel by merging the universe of U.S. federal individual income tax returns (2005–2017) with: the National Plan and Provider Enumeration System (NPPES) physician registry; Medicare billing records with procedure-level Relative Value Unit (RVU) rates (2012–2017); restricted-use American Community Survey responses; Social Security Administration demographic records; and medical school ranking and graduation data. The main sample covers 11.6 million physician-year observations for 965,000 unique physicians aged 20–70.

Earnings Facts. In 2017, average physician total individual income was $350,000 (median $265,000); the distribution is right-skewed — the top 1% of age-40–55 physicians averages $4.0 million. Physicians in aggregate earned $297 billion in pre-tax dollars, equaling 8.6% of total U.S. healthcare spending. The age-earnings profile is steep: earnings are approximately $60,000 during residency, rise to roughly $185,000 by the early thirties, and peak near $425,000 at age 50. Business income — systematically underreported in survey data (ACS estimates are approximately $140,000 lower than tax data during peak career years, almost entirely due to non-reporting of business income) — accounts for nearly one-quarter of earnings at age 50. Earnings differ sharply across specialties: primary care physicians average $201,200 (ages 40–55), about half the sample mean, while surgeons earn roughly twice as much.

Geographic Pattern. Contrary to the pattern for lawyers and workers broadly, physician earnings are not highest on the coasts. A movers-based event study (physicians who changed commuting zones once during 2005–2017) finds that roughly 70% of the cross-location income difference is driven by place rather than worker composition. A two-way fixed-effects variance decomposition reveals pronounced negative physician-location sorting: high-earning physicians tend to locate in lower-income commuting zones, while lower-earning physicians locate in higher-income areas — the opposite of the pattern for lawyers. Medicare’s relatively weak adjustment of reimbursement rates for local costs (the empirical elasticity of the Geographic Adjustment Factor to median household income is 0.09, versus 0.33 for a broader local price index) can, by the authors’ estimates, account for approximately one-third of this unusual geographic earnings pattern.

Government Influence — Medicare Price Changes. Using procedure-specific RVU changes as a simulated instrument for each physician’s Medicare price exposure, the authors find that a 10% increase in the Medicare price instrument leads to a 2.4% increase in professional earnings of physicians aged 40–55. The behavioral supply response is substantial: physicians bill 4.4% more RVUs (supply elasticity of 0.4 after netting out the mechanical component), of which 3.9% reflects more unique procedures and the rest a shift toward higher-paid procedures. Nearly all of the procedure-level supply increase (3.4 out of 3.8 percentage points) comes from treating additional patients rather than more frequent treatment of existing patients. Converting to pass-through: physicians retain $62 of each $100 in additional Medicare spending directly, or approximately $25 of each $100 of any insurance spending once Medicare’s documented spillover into private insurance rates is accounted for. For physicians aged 56–70, a 10% increase in earnings driven by reimbursement changes reduces retirement probability by 0.5 percentage points in that year.

Government Influence — ACA Insurance Expansion. Using county-level variation in pre-ACA uninsurance rates (as of 2013) as a source of differential exposure to the ACA’s Medicaid expansions and Marketplace subsidies (in 24 states expanding Medicaid in 2014 or early 2015), the authors estimate that a 10 percentage point higher baseline uninsurance rate led to 3.9% higher physician earnings four years post-expansion. Scaling by the first stage (a 10 p.p. higher uninsurance rate translating to 4.96 p.p. higher insurance coverage post-expansion), the implied elasticity of physician earnings to the insurance rate is 0.41. The ACA expansion also reduced retirement probability — a 10 p.p. higher insurance coverage rate leads to a 1 p.p. decline in retirement probability — consistent with a medium-run retirement-to-income elasticity of approximately −1.1. In aggregate, 6% of the $110 billion in annual ACA insurance expansion spending accrued to physicians personally, slightly below their 8.6% baseline share of healthcare spending.

Talent Allocation. Specialty choice is sticky and entry-restricted. The authors estimate a discrete-choice model of specialty choice using graduates of top-5 medical schools — physicians with effectively unconstrained specialty access — and an aggregate model using USMLE Step 1 score buckets as ability proxies. At the top of the ability distribution, higher specialty earnings strongly attract physicians: increasing primary care physicians’ hourly income from $98 to $168 per hour (the level of medicine subspecialists) would raise the share of top-5 medical school graduates choosing primary care by approximately 20 percentage points (nearly doubling their representation in primary care). Moving down the USMLE score distribution, the earnings coefficient falls monotonically and turns negative for the lowest score groups — consistent with the model’s prediction that entry restrictions cause higher-paying specialties to displace lower-ability applicants as earnings rise, rather than simply attracting more entrants. A more modest counterfactual — raising internal medicine earnings to dermatology levels — raises the average USMLE score in internal medicine by 10 points (from 230.2 to 239.6).

Scope Conditions. The earnings estimates are for the period 2005–2017. Pass-through estimates use a short-run price instrument; long-run pass-through may differ depending on private market spillovers and entry. The ACA analysis is restricted to 24 early-expanding states. The specialty-choice model is estimated on medical graduates entering the residency match; the extensive margin of entering medicine itself is not modeled. Health outcome effects of changing physician ability distributions are not estimated.

In depth

Q1. What is the level and composition of physician earnings in the tax data, and how do they compare to survey-based estimates?

In 2017, average physician total individual income was $350,000 and median was $265,000; the top 1% of age-40–55 physicians earned $4.0 million on average, more than twice the average of the top 5%. Business income constitutes nearly one-quarter of earnings at age 50 and is concentrated among top earners: 80% of physicians in the top 1% have business income exceeding $25,000, versus 35% overall. ACS survey data for the same physicians underestimate earnings by approximately $140,000 (roughly one-third of the administrative mean) during peak career years, driven entirely by non-reporting of business income on the extensive margin.

Physicians in aggregate earned $297 billion pre-tax in 2017, equaling 8.6% of total U.S. healthcare spending (approximately $913 of the average American’s $10,611 annual healthcare expenditure). After applying a 30% income tax rate, after-tax physician earnings equal approximately 6% of total healthcare spending, or roughly 1% of GDP. The authors note this provides an upper bound on the magnitude of savings available from policies aimed at reducing physician incomes as a strategy for lowering overall healthcare spending.

Q3. How does the age-earnings profile of physicians evolve, and what drives growth during peak years?

Physician earnings average approximately $60,000 during residency, rise to roughly $185,000 by the early thirties, and peak near $425,000 at age 50, before declining gradually to approximately $270,000 in the late 60s. Growth during peak earning years (ages 40–55) is driven almost entirely by business income: average wages are approximately flat at $285,000 across this age range, while business income and the probability of filing Schedule C rise steadily.

Q4. How large and unusual is the geographic pattern of physician earnings, and what is the causal role of location?

Physician earnings are highest in lower-income states (not on the coasts), unlike lawyers and the broader workforce. A movers event study finds that approximately 70% of the cross-commuting-zone income difference is attributable to location rather than worker characteristics; within specialty the estimate rises to approximately 85%. A two-way fixed-effects variance decomposition (with limited-mobility-bias corrections following Andrews et al. 2008 and Kline et al. 2020) reveals pronounced negative physician-location sorting, with the corrected covariance between individual and location effects being 0.6–0.8 times the variance of location effects in magnitude but opposite in sign — a pattern that reverses to positive sorting when the same methods are applied to lawyers.

Q5. What instrument is used to identify the causal effect of Medicare price changes on physician earnings, and why is it valid?

The authors construct a physician-year “Medicare price instrument” by fixing each physician’s service mix at its 2012–2017 average and then multiplying those fixed quantities by annually-updated RVU rates, summing over services. Because the fixed quantity weights exclude behavioral responses, and because national RVU changes from CMS periodic reviews affect physicians differentially according to their pre-determined service mix, variation across physicians and over time is plausibly exogenous to individual physicians’ income shocks. Year-by-specialty fixed effects absorb common specialty-level price trends.

Q6. What are the magnitudes of the earnings and labor supply responses to Medicare price changes?

A 10% increase in the Medicare price instrument raises earnings of 40–55 year-old physicians by 2.4% (reduced-form), with a 2SLS elasticity of income to billed RVUs of 0.17. The total-RVU billing coefficient of 1.437 implies a supply elasticity of 0.437 (subtracting 1 for the mechanical component). At the procedure level, a 10% price increase for a specific code leads to 3.8% more billings for that code, of which 3.4 percentage points reflects treating additional patients. For physicians aged 56–70, a 10% earnings increase reduces that year’s retirement probability by 0.5 percentage points.

Q7. How does the ACA insurance expansion affect physician earnings and retirement, and what is the implied pass-through?

Counties with a 10 percentage point higher pre-ACA uninsurance rate saw 3.9% higher physician earnings by 2017 (four years post-expansion). Scaled by the first stage (4.96 p.p. higher coverage), the elasticity of physician earnings to insurance coverage is 0.41. A 10 p.p. higher insurance coverage rate leads to a 1 p.p. lower retirement probability post-expansion (medium-run elasticity of retirement to income of approximately −1.1). In aggregate, 6% of $110 billion in annual ACA expansion spending — roughly $7.1 billion, or about $8,400 per physician — accrued to physicians.

Q8. How does the earnings-specialty choice relationship vary across the physician ability distribution?

In the individual-level discrete-choice model estimated on top-5 medical school graduates (likely unconstrained in specialty choice), the coefficient on hourly earnings is 0.014. In the aggregate score-group model, the implied earnings coefficient is 0.016 for USMLE scores above 260 and declines monotonically to −0.008 for scores at or below 190. This negative coefficient for low scorers is consistent with the theoretical prediction that higher earnings attract high-ability physicians, leaving fewer slots for lower-ability applicants due to binding entry restrictions — not a reversal of preferences.

Q9. What are the quantitative implications for specialty choice if primary care incomes were raised to subspecialty levels?

Raising primary care hourly income from $98 to $168 (the level of medicine subspecialists) would increase the share of top-5 medical school graduates choosing primary care by approximately 20 percentage points (about 48% would enter primary care, versus the current share), nearly doubling their representation. Nearly half of these reallocations would come from procedural specialties. An analogous exercise raising internal medicine earnings to dermatology levels shifts the average USMLE score in internal medicine from 230.2 to 239.6 — a 10-point increase — as higher-scoring applicants displace lower-scoring ones within a fixed slot constraint.

Direct estimates imply physicians retain $62 of each $100 in additional Medicare spending. Accounting for Medicare’s documented spillover into private insurance rates (following Clemens and Gottlieb 2017), the pass-through drops to $25 per $100 of total insurance spending. The authors note this is substantially higher than the modest rent-sharing found for average workers in response to firm-level shocks (Card et al. 2018), but comparable to rent-sharing with high-skilled workers benefiting from patent rents (Kline et al. 2019).

Q11. Can Medicare’s geographic pricing policy explain the unusual geographic earnings pattern for physicians?

The elasticity of Medicare’s Geographic Adjustment Factor (GAF) to commuting zone median household income is 0.09, compared to 0.33 for a broader local price index. Using the authors’ short-run estimate that a 10% increase in Medicare prices raises earnings by 2.4%, a counterfactual simulation shows that if the GAF-to-income elasticity rose to 0.33 (aligning Medicare rates with the general cost-of-living gradient), the geographic physician earnings pattern would more closely resemble that of lawyers. The authors estimate that the gap in Medicare’s local cost adjustment explains approximately one-third of the unusual physician earnings geography, conditional on the short-run pass-through estimate.

Q12. How does the theoretical model of specialty choice and entry restrictions guide the empirical predictions?

The model features a unit mass of physicians with heterogeneous ability (Pareto-distributed) and idiosyncratic specialty preferences (exponentially distributed). Physicians choose whether to specialize in period 1; government sets reimbursement rates in period 2; physicians choose labor supply in period 3. With a fixed number of residency slots, higher specialty earnings raise the ability cutoff for entry (rationing by ability). This generates a key nonmonotonic empirical prediction: higher-ability physicians respond positively to earnings increases (choosing a specialty more frequently), while lower-ability physicians respond negatively (displaced by the shift upward in the ability cutoff). The model also implies that demand shocks are not moderated by contemporaneous entry, so incumbents capture the full rent — motivating the estimated pass-through.

Key Concepts

Medicare Price Instrument (Simulated RVU Instrument). A physician-year measure of Medicare payment exposure constructed by holding each physician’s service mix fixed at its 2012–2017 average and multiplying those fixed quantities by time-varying national RVU rates, then summing across services. This purges the instrument of behavioral responses, creating exogenous cross-physician variation in price exposure arising from the interaction of fixed service mix with national RVU policy changes.

Relative Value Unit (RVU). The unit by which Medicare defines and reimburses each physician service in the Physician Fee Schedule. RVUs are intended to reflect the time, effort, and resources required to provide each service, but are subject to periodic review by CMS’s RVU Update Committee (RUC) and influenced by political factors. Changes in RVUs translate directly into changes in Medicare reimbursement rates for affected services.

Pass-Through (Reimbursement to Earnings). The share of an additional dollar of Medicare (or insurance) spending that accrues to physicians personally as earnings, after accounting for practice costs, intermediaries, and behavioral responses. The paper estimates $62 per $100 of direct Medicare spending or $25 per $100 of total insurance spending (the latter accounting for Medicare’s spillover into private rates).

Negative Physician-Location Sorting. The empirical finding — robust to limited-mobility-bias corrections — that higher-ability (higher-earning) physicians disproportionately locate in lower-income commuting zones, while lower-earning physicians concentrate in higher-income areas. This is the opposite of the pattern for lawyers and for worker-firm matching in the broader labor literature. The paper attributes part of this pattern to Medicare’s incomplete geographic adjustment of reimbursement rates.

Ability Cutoff (am) in Residency Matching. In the paper’s theoretical model, the minimum ability level required to gain entry into a restricted-entry specialty. Because the number of residency slots is fixed, the cutoff rises when a specialty’s relative earnings increase (attracting more high-ability applicants), displacing lower-ability physicians who would otherwise have entered. This makes the earnings-specialty relationship nonmonotonic across the ability distribution.

Business Income (Pass-Through Entity Income). Income from physician-owned practices organized as sole proprietorships, S-corporations, or partnerships, reported on Schedule C or through pass-through entities rather than on Form W-2. In the tax data, business income accounts for nearly one-quarter of physician earnings at career peak and is the main source of earnings for top physicians, but is systematically underreported in survey data (ACS), leading to a roughly one-third underestimate of total earnings during peak years.

Geographic Adjustment Factor (GAF). A Medicare policy parameter that multiplies the national RVU rate to adjust physician reimbursements for local input costs (specifically physicians’ work, practice expenses, and malpractice). The paper documents that the GAF’s elasticity to local median household income is 0.09 — far below the 0.33 elasticity of the general local price index — constituting an effective subsidy to rural and lower-income markets relative to higher-income areas.

The Effects of Gender Integration on Men

Mon, 01 Jan 0001 00:00:00 +0000

Greenberg, Wasserman, and Weber (2024/2026) ask whether men negatively respond—in terms of job performance, behavior, and workplace perceptions—when women first enter an exclusively male occupation. They exploit the staggered 2017-onward integration of women into U.S. Army infantry and armor combat companies following the 2016 rescission of the Ground Combat Exclusion Policy. The setting offers unusually clean causal identification: integration timing within Brigade Combat Teams was neither systematic nor data-driven, the Army’s rigid pay scales meant integration posed no displacement or wage threat to incumbent men, and roughly 391 companies are observed over 2012–2020. The empirical strategy is a staggered difference-in-differences design with company fixed effects, BCT-by-year-of-arrival fixed effects, and month-of-year fixed effects, applied to an individual-level sample of newly arrived male soldiers. Outcomes come from monthly administrative personnel records (retention, misconduct separations, demotions, criminal investigations, drug tests, medical profiles, physical fitness scores) and the Defense Organizational Climate Survey (DEOCS), a congressionally mandated annual survey with response rates above 50% covering organizational effectiveness, equal opportunity, and sexual assault prevention and response. The main finding is that integrating women into previously all-male combat companies does not negatively affect men’s performance or behavioral outcomes. Estimates are precise enough to rule out small detrimental effects: two years post-integration, the authors can rule out a 3% increase in attrition, a 5% increase in demotions, and a 4% increase in criminal investigations relative to their respective means. One behavioral outcome shows a statistically significant improvement: integration reduces separations for misconduct by 1.3 percentage points (16% of the mean). Drug test positivity also declines. The sole potential negative administrative finding is a 1.8-point decline in physical fitness scores (0.7% of the mean, roughly 5% of a standard deviation), but this does not affect pass rates and becomes statistically insignificant when scores are imputed using observable covariates. An aggregate Performance and Behavior Index rules out reductions of 0.8% of a standard deviation; the No Adverse Outcomes measure rules out a 1.2 percentage point increase (3% of the mean). Despite these null-to-positive performance effects, survey data reveal that integration causes a 5% of a standard deviation decline in men’s overall perceptions of workplace quality. This perception decline is concentrated in companies that received a female officer shortly after integration. Among companies integrated only with female enlisted soldiers (no female officer), men’s workplace attitudes actually improve by 14.7% of a standard deviation. Two mechanisms are examined: increased male awareness of pre-existing workplace problems (supported by higher reported observations of bullying, hazing, and unwanted comments, especially among male officers in female-officer-integrated companies), and negative reactions to women in positions of authority (supported by broader declines in organizational effectiveness perceptions not confined to equal-opportunity items). Crucially, the perception decline does not translate into retaliatory behavior or performance deterioration; companies integrated with a female officer show some performance gains, and female enlisted soldiers in those companies report fewer workplace problems. Scope conditions: findings apply to a high-stakes, traditionally male-dominated, hierarchical occupational setting during 2017–2020, a period when U.S. deployment missions were primarily advise-and-assist rather than direct combat. Integration increased female representation by approximately 4.7 percentage points on average.

Q: What was the policy change studied and why does it offer causal leverage? A: In December 2015, Secretary of Defense Ashton Carter announced that all U.S. military occupations, including infantry and armor combat roles, would open to women starting in 2016. Women did not begin arriving at operational companies until 2017 due to training timelines. Within BCTs, the selection of which companies to integrate was neither systematic nor data-driven, and baseline characteristics of integrated and non-integrated companies are similar after conditioning on BCT and company-type fixed effects, supporting a parallel trends assumption.

Q: What are the main administrative performance findings? A: Integration has a positive but statistically insignificant effect on retention, and reduces misconduct separations by 1.3 percentage points (significant at the 5% level), representing a 16% reduction relative to the mean. Demotions, criminal investigations (including sex-related and domestic violence), and medical profiles show no significant negative effects, with precision sufficient to rule out 5% increases in demotions and 4% increases in criminal investigations. Physical fitness scores decline by 1.8 points (0.7% of mean, approximately 5% of a standard deviation), but pass rates are unaffected and the estimate becomes insignificant when scores are imputed with observable covariates.

Q: What does the aggregate performance index show? A: The Performance and Behavior Index—an equally weighted z-score average of retention, misconduct separations, demotions, criminal investigations, medical profiles, promotions to Sergeant, and physical fitness outcomes—shows a positive but insignificant effect of integration, ruling out reductions of 0.8% of a standard deviation. The No Adverse Outcomes measure rules out a 1.2 percentage point increase (3% of the mean incidence of adverse outcomes).

Q: How do men’s workplace perceptions change after integration? A: The overall workplace quality index constructed from all DEOCS Likert-scale items declines by 5% of a standard deviation following integration, spanning perceptions of organizational effectiveness, workplace inclusivity, and sexual assault prevention and response. This average effect masks critical heterogeneity by the rank composition of integrating women.

Q: What is the key heterogeneity in survey responses? A: The decline in men’s perceptions is entirely driven by companies that received a female officer shortly after integration. In companies integrated only with female enlisted soldiers (17% of integrating companies did not receive a female officer within a month), men’s perceptions improve by 14.7% of a standard deviation. Male officers show a larger negative shift than male enlisted soldiers in officer-integrated companies, and this difference is statistically significant.

Q: What mechanisms explain the negative perception response to female officers? A: Two mechanisms are investigated. First, increased awareness: male soldiers—especially male officers—report observing more bullying, hazing, and unwanted comments after a female officer is integrated but not after integration with only female enlisted, and the decline in perceptions of sexual assault prevention and response is significantly larger among male officers than enlisted men, consistent with shared leadership roles amplifying awareness of workplace problems. Second, negative reactions to female authority: declines in perceptions are more pronounced on organizational effectiveness questions than on equal-opportunity items and extend to issues unrelated to women, suggesting broader dissatisfaction with female leadership alongside heightened awareness.

Q: Is the decline in perceptions related to actual differences in female officer qualifications or preferential treatment? A: No. Female and male officers have similar baseline characteristics including educational background and experience. Companies integrated with female officers perform at least as well as non-integrated companies or those integrated only with enlisted women on administrative metrics. There is no evidence that male officers waited longer for leadership assignments relative to female colleagues, ruling out perceived preferential treatment as a driver.

Q: Do men’s negative perceptions of female officers translate into retaliatory behavior toward women? A: No. Administrative misconduct metrics show some improvements in male behavior when a female officer is present. Female enlisted soldiers in female-officer-integrated companies report fewer workplace problems on the climate survey than female enlisted soldiers in companies integrated without a female officer, indicating that the presence of a female officer generates benefits for female enlisted soldiers rather than backlash against them.

Q: Does heterogeneity by integration intensity or women’s rank affect administrative outcomes for men? A: Integration intensity (number of women initially integrated) and rank composition (female officers vs. only female enlisted) do not produce negative administrative outcomes in any subgroup. The aggregate Performance and Behavior Index shows a positive effect when a female officer is included. Effects also do not vary with male soldiers’ rank (enlisted vs. officer) or their tenure in the company.

Q: What happens in units that deploy to combat zones? A: Approximately one in five integrated companies deployed to a combat zone within two years of integration. Integration does not negatively affect retention, behavior, or performance of men in deploying units. Declines in workplace perceptions are larger for deploying units and are most pronounced when integration occurs shortly after return from deployment, consistent with deployment strengthening in-group identity among male soldiers rather than women performing poorly during combat-zone service.

Q: What do the findings imply for theories of identity economics and the pollution theory of discrimination? A: The null-to-positive behavioral and performance responses to women’s entry contradict the predictions of Akerlof and Kranton’s (2000) identity economics model and Goldin’s (2014) pollution theory of discrimination, which predict retaliatory or otherwise unproductive behaviors when women enter a male-dominated occupation. The paper shows that, to the extent identity concerns shape male responses, these are confined to subjective perceptions and do not manifest in diminished performance, retention, or conduct.

Q: What are the policy implications for employers considering gender integration? A: The paper provides evidence against the argument that men will become less productive when women enter previously male-only occupations, a justification sometimes offered for excluding women from such jobs. The finding that performance and behavior are unaffected—and misconduct actually declines—allows policymakers and employers to weigh these results against concerns about operational or productivity costs of integration. The perception gap between men’s attitudes and actual outcomes points to a need for targeted leadership and organizational interventions, particularly around the introduction of female leaders.

Ground Combat Exclusion Policy (GCEP): The U.S. military policy, rescinded in 2013 and fully eliminated by Secretary of Defense Carter in 2016, that precluded women from serving in infantry and armor positions; the policy whose removal is the source of the integration shock studied. | Staggered difference-in-differences: The empirical strategy exploiting the sequential, non-systematic integration of women into combat companies across years 2017–2023, using never-yet-treated companies as a comparison group with company fixed effects and BCT-by-year-of-arrival fixed effects. | Performance and Behavior Index: An equally weighted average of z-scored administrative outcomes (retention, no misconduct separations, no demotions, no criminal investigations, no medical profiles, promotion to Sergeant, physical fitness pass/fail and score), constructed for enlisted soldiers, oriented so higher values indicate better outcomes. | Leaders First policy: An Army requirement that a female officer be assigned to a combat company before or alongside female junior enlisted soldiers to ensure female leadership presence at integration; adherence was not universal, with 17% of integrating companies not following it within one month. | Defense Organizational Climate Survey (DEOCS): A congressionally mandated, annually administered, anonymous survey of military unit members covering organizational effectiveness, equal opportunity, and sexual assault prevention and response; the source of workplace perception outcomes. | Pollution theory of discrimination: Goldin’s (2014) theory that men may seek to exclude women from occupations because women’s presence is perceived to diminish the occupation’s prestige or status, potentially leading to retaliatory or unproductive behaviors among incumbent male workers. | Perception-performance wedge: The paper’s central finding that men’s subjective workplace quality perceptions decline with integration—especially when a female officer is present—even as objective administrative performance and behavior metrics show null to positive effects, a divergence between attitudes and measurable outcomes.

The Effects of Mandatory Profit-Sharing on Workers and Firms

Mon, 01 Jan 0001 00:00:00 +0000

This paper studies the causal effects of mandatory profit-sharing on workers and firms using a quasi-experimental design arising from a 1990 French reform that lowered the eligibility threshold for mandatory profit-sharing from 100 to 50 employees. The institutional setting is the French RSP (Réserve Spéciale de Participation), a profit-sharing scheme in place since 1967 that requires firms above the threshold to distribute a fraction of their excess profits — defined as net income above 5% of book equity — to employees according to a formula scaled by the firm’s labor share. For the median firm, this amounts to roughly 10.5% of pre-tax income transferred to workers.

The authors employ two primary empirical strategies. First, a bunching analysis exploits the pre-reform distribution of firm employment around the 100-employee threshold as a revealed-preference test of whether firms perceive profit-sharing as a net cost. Second, a difference-in-differences design compares treated firms (55–85 employees in 1989–1990, who become newly subject to the regulation after 1991) against two control groups: small firms (35–45 employees, likely never subject) and large firms (120–300 employees, already subject). Data come from the universe of French corporate tax files (FICAS) and a linked employer-employee panel (DADS) covering approximately 4% of private-sector workers, spanning 1985–1997.

The bunching analysis documents a 22.3% excess density in the 95–99 employee bin before the reform, which disappears after 1991. Three tests — comparing wage bills per employee across the threshold, cross-checking with DADS employment records, and examining profitability patterns — collectively support the conclusion that bunching reflects genuine employment reductions rather than under-reporting. The implied employment loss is approximately 1.67% of total employment among affected firms.

The difference-in-differences results yield the following firm-level findings: (a) the total compensation share (wages plus profit-sharing divided by value added) rises by 1.8 percentage points for firms with positive excess profits; (b) 77% of this increase comes at the expense of firm owners — the profit share falls by 1.37 percentage points; (c) the remainder is borne by the government through a reduction in the corporate income tax share; (d) the wage share (base wages only) is unaffected, indicating that owners do not reduce wages to offset the cost of profit-sharing; (e) investment and total factor productivity show no statistically significant change — effects on productivity are bounded below ±1% for several TFP measures; and (f) the capital-labor ratio shows a small, mostly insignificant negative effect, consistent with a model-implied increase in the cost of capital of only 0.43 percentage points.

Worker-level analysis using the linked employer-employee data confirms that average total compensation rises by approximately 3.5% for workers in treated firms, with no decline in base wages. Critically, this average conceals distributional heterogeneity across the skill spectrum. For low- and medium-skill workers (blue-collar workers, clerks, supervisors, skilled technicians), total compensation rises while base wages are unchanged — consistent with wage rigidity binding for these groups. For high-skill workers (managers, engineers, executives), base wages fall by enough to leave total compensation unchanged, consistent with more flexible wages at the upper end of the skill distribution. This pattern implies that mandatory profit-sharing is a progressive policy within firms, redistributing excess profits predominantly to lower-skill workers.

The paper concludes that France’s mandatory profit-sharing scheme, as implemented, functions as a non-distortive redistributive tool: it transfers excess profits from shareholders to lower-skill workers without generating measurable productivity losses or large investment distortions. The fiscal cost is non-trivial: each dollar transferred to workers costs approximately 20 cents in foregone corporate income tax. The scheme also has an inherent inequality in its redistribution since it exclusively benefits workers in profitable firms, and firms’ excess profits are highly persistent.

Q: What is the French RSP and how does the formula work? A: The RSP (Réserve Spéciale de Participation) is a mandatory profit-sharing fund established by executive order in 1967. The formula is RSP = 0.5 × (wage bill / value added) × max(net income − 5% × book equity, 0). The 5% deduction represents lawmakers’ view of fair compensation to shareholders; any excess is split between shareholders and workers, with the split scaled by the firm’s labor share. For the median firm in the sample — ROE of 12%, labor share of 0.52, corporate tax rate of 37% — the formula yields roughly 9.5% of pre-tax income, and in post-1991 data the realized average is 10.5% of pre-tax income for firms with positive excess profits.

Q: Why can’t a standard regression discontinuity be used at the 100-employee threshold? A: Because firms strategically control their position relative to the threshold — the bunching analysis itself demonstrates this. When firms sort non-randomly around the cutoff, the local randomization assumption underlying RD is violated. The authors instead use a difference-in-differences design exploiting the time variation introduced by the 1990 reform.

Q: How large is the pre-reform bunching and what does it imply? A: The distribution of employment shows 22.3% excess density in the 95–99 employee bin relative to the post-reform counterfactual distribution. Interpreting this as real employment reduction (supported by three empirical tests), the implied employment loss is approximately 1.67% of total employment among firms in the 85–120 employee range. Dynamic bunching analysis shows this is persistent rather than temporary — the 100-employee threshold significantly constrained three-year employment growth for firms in the 85–99 range in the pre-reform period.

Q: How do the authors establish that bunching is real rather than under-reporting of employment? A: Three tests are conducted. First, wage bills per employee show no discontinuity around the 100-employee threshold in either period, ruling out systematic under-reporting of headcount while truthfully reporting wages. Second, employment from DADS payroll records — harder to manipulate — shows only a statistically insignificant gap of roughly 0.5 employees relative to tax-file employment just below the threshold, far too small to shift firms across the 100-employee bin. Third, profitability and value added per employee are significantly higher just below the threshold, consistent with more profitable firms having stronger incentives to bunch through genuine employment reductions.

Q: What is the main identification strategy for the firm-level analysis? A: A difference-in-differences design where treated firms have 55–85 employees in both 1989 and 1990 (newly subject to the mandate after 1991), compared to small control firms with 35–45 employees (likely never subject) and large control firms with 120–300 employees (likely always subject). Specifications include firm fixed effects and county-by-year and industry-by-year fixed effects. Parallel pre-trends are confirmed graphically and in event-study regressions. The design is intent-to-treat: by 1997, 26.7% of treated firms had shrunk below 50 employees and did not actually pay profit-sharing. LATE estimates are obtained via 2SLS.

Q: What are the main firm-level findings on compensation and profit shares? A: For treated firms with positive excess profits, the total compensation share rises by 1.8 percentage points. The wage share (base wages only, excluding profit-sharing) is precisely estimated at zero — owners do not reduce wages. The profit share falls by 1.37 percentage points, accounting for 77% of the increase in total compensation. The remaining approximately 23% is borne by the tax authority through a reduction in the corporate income tax share, since profit-sharing reduces the corporate income tax base. These findings are robust to balanced vs. unbalanced samples and to alternative control group definitions.

Q: Does mandatory profit-sharing raise or lower firm productivity? A: Across five different TFP estimators (Olley-Pakes, Olley-Pakes with Ackerberg-Caves-Frazer correction, Wooldridge, Levinsohn-Petrin, and Ackerberg-Caves-Frazer), the effect of mandatory profit-sharing on productivity is a precisely estimated zero. For several measures, effects larger than ±1% in magnitude can be rejected. Softer measures of effort — sick leave rates and the probability of working extra hours — also show no significant change. This null finding contrasts with the literature on voluntary profit-sharing adoption, which typically finds 3–5% productivity gains, likely reflecting selection bias in that literature.

Q: Does mandatory profit-sharing distort investment? A: The effect on investment is small and mostly statistically insignificant. The theoretical model shows why: the profit-sharing formula is based on excess profits (net income minus 5% of book equity), not total profits. When the firm’s actual cost of equity approximately equals the regulatory 5% benchmark, the distortion to the cost of capital is zero. The calibrated distortion to the user cost of capital is only 0.43 percentage points — approximately 1.9% of the standard user cost — implying an investment ratio reduction of about 0.84 percentage points using estimated elasticities from Chodorow-Reich et al. (2024). Empirically, capital-labor ratios show a small, largely insignificant negative effect.

Q: How does profit-sharing incidence differ across the skill distribution? A: The worker-level DADS analysis reveals that the average 3.5% increase in total compensation masks sharp heterogeneity. For low- and medium-skill workers (blue-collar workers, clerks, supervisors, skilled technicians), total compensation rises while base wages are unchanged. For high-skill workers (managers, engineers, executives), base wages decline sufficiently to leave their total compensation unchanged. The authors interpret this pattern as consistent with wage rigidity being more binding for lower-skill workers — due to the federal minimum wage and collective agreements — than for managers whose pay is more flexibly set.

Q: Why does profit-sharing not affect base wages for low-skill workers? A: Two candidate explanations are considered. The risk channel — that profit-sharing is risky and thus less valuable to risk-averse workers, who demand wage compensation — is rejected empirically because profit-sharing only marginally increases the variability of workers’ total earnings. The wage rigidity channel is supported: France’s binding federal minimum wage and widespread collective agreements constrain downward adjustment in base wages for lower-skill workers, so firms cannot pass through profit-sharing costs as lower wages for this group.

Q: What is the fiscal cost of the profit-sharing scheme? A: Each dollar transferred to workers through mandatory profit-sharing costs approximately 20 cents in reduced corporate income tax receipts, since profit-sharing payments are deductible from taxable income. The paper notes this is a partial fiscal evaluation; a full assessment would also require analyzing personal income tax implications, which are left for future work.

Q: How does this scheme compare to a corporate income tax as a redistributive tool? A: Both instruments reduce firm profits and can benefit workers, but differ in three key respects. First, the tax base differs: profit-sharing targets excess profits above 5% of book equity whereas the corporate income tax applies to all corporate earnings, generating different distortions to investment. Second, profit-sharing goes directly to workers in the same firm, whereas corporate tax revenues are redistributed through general government spending — making the incidence more direct and more closely monitored by workers. Third, workers have stronger incentives to monitor firm compliance with profit-sharing (each euro of diverted excess profit reduces workers’ collective income by roughly 10–15 cents) than with corporate taxes.

Q: How does this paper compare to findings on mandatory profit-sharing in Peru? A: Tolentino (2022) studies a mandatory profit-sharing scheme in Peru exploiting a 20-employee eligibility threshold and finds larger distortions — reductions in both investment and productivity. The authors attribute this difference to two features: the Peruvian scheme applies to the entirety of post-tax profits rather than excess profits above an equity deduction, creating a broader and more distortionary base; and there is pre-existing bunching at the Peruvian threshold even before the scheme was introduced, suggesting confounding pre-existing regulations.

Q: What are the scope conditions on the external validity of the findings? A: The findings apply specifically to mandatory profit-sharing under the French RSP formula — which exempts a 5% equity return from the profit-sharing base, limiting distortions — during 1985–1997, for firms in the 55–300 employee range. The null productivity effect may not generalize to voluntary schemes, where selection on anticipated gains likely produces positive correlations. The redistributive finding (benefiting lower-skill workers) is specific to a context with binding minimum wages and collective agreements that constrain wage adjustment for that group. The fiscal cost calculation also excludes personal income tax effects.

Excess profits: Defined in the paper as net income minus 5% of book equity — the amount above what lawmakers considered fair compensation to shareholders. Only excess profits (not total profits) are subject to the mandatory profit-sharing formula.

RSP formula (Réserve Spéciale de Participation): The statutory formula RSP = 0.5 × (wage bill / value added) × max(net income − 5% × book equity, 0), scaled by the firm’s labor share to reflect labor’s contribution to production. Unchanged since 1967.

Total compensation share: The ratio of (wage bill plus profit-sharing) to value added — the paper’s primary measure of workers’ overall claim on firm output, as distinct from the wage share (wage bill alone divided by value added).

Wage incidence parameter (λ): The fraction of profit-sharing that firms pass through to workers as lower base wages. λ = 1 means full incidence (workers’ total compensation unchanged); λ = 0 means no incidence (workers fully benefit). The paper’s empirical findings are consistent with λ ≈ 0 for low-skill workers and λ ≈ 1 for high-skill workers.

Bunching: The empirical phenomenon whereby firms cluster employment just below the 100-employee regulatory threshold to avoid mandatory profit-sharing. The paper uses the pre- vs. post-reform shift in the employment distribution as a revealed-preference test of whether firms perceive the scheme as a net cost.

Intent-to-treat (ITT) design: The empirical design comparing firms that were in the newly eligible size range (55–85 employees) just before the 1990 reform against firms that were either always or never eligible, regardless of whether treated firms actually ended up paying profit-sharing post-reform. LATE estimates are obtained via 2SLS to recover effects on actual compliers.

Distortion to user cost of capital: The additional cost of capital induced by profit-sharing, equal to ϕ × γ(1−λ) / [1 − γ(1−τ)] × (re − ρ), where ρ = 5% is the regulatory equity benchmark. When the firm’s actual cost of equity equals the 5% benchmark, this distortion is zero — a feature that distinguishes the French scheme from a standard corporate income tax.

The Effects of Medical Debt Relief: Evidence from Two Randomized Experiments

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1: Overview

Research Question

This paper asks whether relieving downstream medical debt — debt that has been sold to third-party debt collectors — causes improvements in financial outcomes, mental and physical health, and healthcare utilization for recipients. The question is motivated by a large correlational literature documenting strong associations between medical debt and adverse outcomes, and by the rapid expansion of government and private debt relief programs that, as of mid-2024, had committed or planned over $14.6 billion in relief.

Data and Design

The authors partnered with RIP Medical Debt (a non-profit that purchases and forgives medical debt for government and private donors) to conduct two randomized controlled trials between March 2018 and October 2020. In total the experiments relieved medical debt with a face value of $169 million for 83,401 people.

Hospital debt experiment: RIP purchased a random subset of debt from a large for-profit hospital system at the juncture when the hospital would normally sell accounts to a debt collector (approximately one year after the medical service). The purchase price was 5.5 cents per dollar of face value. The treatment group consisted of 14,377 people who received $19 million in face-value relief (average of $1,321 per person). The 61,496-person control group had their debt pursued by the collector under normal protocol.
Collector debt experiment: RIP purchased a random subset of older debt already under collection on the secondary market for several years, at a price of less than one cent per dollar. The treatment group consisted of 69,024 people who received $150 million in face-value relief (average of $2,167 per person). The 68,014-person control group retained their debt.
Credit reporting sub-experiment: Partway into the collector debt experiment, the debt collector ceased reporting medical debt to the credit bureaus, reflecting an industry-wide trend. The authors isolate 2,761 accounts (6.8% of wave 1) that were reported prior to treatment assignment to estimate the effects of debt relief when accounts would have been counterfactually reported, compared to the subsequent no-reporting environment.

Outcomes are tracked using quarterly depersonalized credit bureau data from TransUnion (spanning at least four quarters before to four quarters after treatment), collections account data on future bill accrual, and a multimodal survey of 2,888 hospital debt experiment respondents measuring mental and physical health, healthcare utilization, and financial wellness. The primary credit-bureau outcome is the number of accounts past due; the primary survey outcome is the share with at least moderate depression (PHQ-8).

Main Findings

Credit market outcomes (main experiments): In both the hospital and collector debt experiments — where there is no counterfactual credit bureau reporting — debt relief has no average effect on financial distress, credit access, or credit utilization. The effect on the number of accounts past due is -0.01 (statistically insignificant; 95% CI excludes effects smaller than -0.04, relative to a control mean of 1.20). Effects on credit card balances (95% CI: -$42 to $47 relative to a mean of $1,481) and auto loan balances (95% CI: -$235 to $148 relative to a mean of $8,020) are similarly precise nulls. These null effects hold for the hospital debt sample (younger debt, 1.3 years old on average) and the collector debt sample (older debt, 7.0 years old on average), and across all preregistered subgroups.
Credit reporting sub-experiment: When control group accounts are counterfactually reported, debt relief immediately raises credit scores by an economically small average of 3.4 points (p-value 0.021), with a larger 13.8-point increase (p-value 0.008) for persons with no other debt in collections. Credit limits grow gradually, reaching $340 (15.3% of the post-reporting control mean of $2,231; p-value 0.010) after the no-reporting period begins, with larger effects for those with no other debt in collections. Once control group reporting ceases, both the credit score and credit limit effects converge to zero for those with other debts in collections. No effects on borrowing or financial distress measures are detected in this sub-experiment.
Collections account outcomes (bill repayment): Debt relief causes a statistically significant 1.1 percentage-point increase in the probability of having another unpaid bill sent to collections (6.6% of the control mean of 16.2%; p-value < 0.05) and a $15 increase in the dollar amount of future medical debt sent to collections (7.2% of the control mean of $208). The increase is almost entirely attributable to pre-relief medical services, indicating reduced repayment of existing bills rather than greater healthcare utilization.
Survey outcomes: There are no detectable average effects on depression (primary outcome), anxiety, stress, subjective well-being, or general health. Debt relief raises the share with at least moderate depression by a statistically insignificant 3.2 percentage points (p-value 0.097; control mean 45.0%); a 95% CI rules out a reduction of more than 0.6 percentage points, well below the 7.0 percentage-point improvement predicted by the median expert respondent. There are similarly null effects on healthcare utilization and financial wellness as measured in the survey.

Scope Conditions

The study focuses specifically on downstream medical debt in collections — debt that has already been through the hospital billing cycle and sold to third-party collectors. Results do not necessarily apply to upstream debt relief (e.g., financial assistance programs applied closer to the time of the medical event), nor to populations with different baseline financial profiles. The credit reporting results are most relevant to the prior regime of widespread reporting; under the current environment in which most medical debt has been removed from credit reports, the credit-access channel is largely foreclosed.

In depth

Q1. Why did the authors focus specifically on downstream medical debt in collections, and how does this define the scope of their study?

The authors focus on downstream medical debt because this is the target of essentially all large-scale government and private relief programs working with RIP Medical Debt, and because it is the category of debt that is most comprehensively observable. Downstream medical debt is defined as bills that have been or are about to be sold by the healthcare provider to a third-party debt collector. This focus excludes upstream unpaid bills still held by the hospital, bills being paid over time, and medical expenses charged to credit cards. The distinction matters because prior literature on hospital financial assistance programs finds substantial benefits from upstream interventions that relieve debt closer to the precipitating medical event; the authors’ null results are explicitly scoped to the downstream, post-collection stage.

Q2. Why did the purchase price of medical debt (5.5 cents per dollar for hospital debt, less than 1 cent per dollar for collector debt) suggest caution about expected financial impacts ex ante?

The authors argue that in a competitive market, the purchase price of medical debt reflects the sum of expected recovery rates and collection costs. A price of 5.5 cents per dollar implies that actual recovery (what collectors expect to collect from patients) is very low. Even if all of the expected recovery is passed through to the patient as a financial benefit, the direct liquidity gain from debt forgiveness is a small fraction of the debt’s face value. For the collector debt experiment, where the purchase price is less than 1 cent per dollar, the expected direct financial benefit to recipients is even smaller. The authors note that survey respondents expected to pay 54% of their outstanding medical debt and thought it fair to pay 37%, suggesting that perceived (rather than actual) payment obligations may be what connects medical debt to financial behavior.

Q3. How was random assignment implemented in the hospital debt experiment, and what design features ensure the validity of the experiment?

Within each of 18 waves between August 2018 and October 2020, RIP received a portfolio of unpaid bills from the hospital system. Persons were grouped at the individual level and stratified by the amount of debt, state of residence, insurance status, and a collections score predicting repayment likelihood. Within strata, persons were randomly assigned to treatment or control, with approximately 20% treated per wave (varying with donor funding). The hospital was unaware of the intervention, eliminating scope for selection of particularly uncollectible accounts. Treatment notification occurred via two letters sent approximately three and six weeks post-purchase. Balance tests confirm successful randomization: all p-values on baseline characteristics are above 0.05, and F-tests fail to reject joint balance.

Q4. What was the credit reporting sub-experiment and how was it identified?

The debt collector in the collector debt experiment historically reported medical debt to the credit bureaus but largely ceased doing so before the first intervention wave (March 2018), reflecting broader industry concerns about CFPB enforcement and data integrity risk. However, a subset of accounts — 2,761 accounts (6.8% of wave 1, with virtually identical match rates across treatment and control) — were still being reported until 2019 Q1 (three quarters after wave 1 and one quarter after wave 2). This created a natural sub-experiment: for this subset, treatment group accounts were removed from credit reports immediately upon debt relief, while control group accounts continued to be reported for three more quarters before also being removed. The authors identify reported accounts by matching dollar amounts in collections account data to credit bureau tradeline data in the four quarters prior to intervention, and use this variation to estimate effects separately for the “reporting” and “no-reporting” periods.

Q5. What are the exact estimated effects on credit scores and credit limits in the credit reporting sub-experiment?

During the three quarters when control group accounts are still reported to credit bureaus, debt relief raises credit scores by an average of 3.4 points (p-value 0.021) for the full reporting subsample. The effect is concentrated among those with no other debt in collections: 13.8 points (p-value 0.008) versus 1.2 points (p-value 0.440) for those with other debt in collections. Credit limits increase gradually, reaching $340 (15.3% of the post-reporting control mean of $2,231; p-value 0.010) by the four quarters after control group reporting ceases. Among persons with no other debt in collections, this credit limit effect grows to $922 (23% of the control mean; p-value 0.070). Once control group reporting stops, both the credit score effect and the credit limit growth converge to zero for persons with other debts in collections. The event study coefficients show the credit limit effect growing approximately linearly over five quarters post-intervention before leveling out.

Q6. How does the paper rule out the possibility that medical debt relief increases healthcare utilization, thereby causing more future medical bills?

The collections account analysis separates future debt accrual into debt associated with pre-relief medical services (which can only result from reduced repayment of existing bills) and post-relief medical services (which could reflect either increased utilization or changed repayment of new bills). Panel B of Table VI shows that virtually all of the increased debt sent to collections — a $15 increase and 1.1 percentage-point increase in the probability of any future collection — is attributable to pre-relief services. Panel C shows statistically insignificant increases in future debt from post-relief services. The authors therefore attribute the effect to reduced payment of existing bills and conclude they “cannot rule in or rule out effects on healthcare utilization” for the post-relief services channel, but the dominant mechanism is behavioral change in repayment of already-incurred debt.

Q7. What are the three mechanisms proposed to explain the reduction in repayment of existing medical bills, and which mechanism is rejected?

The authors offer three candidate mechanisms for the 6.6% relative increase in the probability of future bill collections: (i) an expectations mechanism, in which beneficiaries reduce payments because they anticipate future debt relief from similar charitable programs; (ii) a targeting mechanism, drawing on Dobkin et al. (2018), in which patients tolerate a certain level of indebtedness — relieving some debt creates “room” in their debt budget, so they reduce payment of remaining bills to return to that target level; and (iii) a confusion mechanism, in which recipients mistakenly believe the relief applied to non-forgiven bills (the notification letter explicitly stated “the forgiveness is for this outstanding bill only” but patients may not have internalized this). The income effect or “flypaper” mechanism — the idea that financial relief of existing debt frees up mental-account resources for paying medical bills, thereby increasing repayment — is explicitly rejected by the data, as the effect goes in the direction of less repayment, not more.

Q8. What did the expert survey predict, and how did those predictions compare to the experimental estimates?

An expert survey conducted between April and May 2022 — after the interventions were completed but before results were released — asked academics, non-profit staff, hospital revenue-cycle practitioners, and policymakers to predict the impact of the hospital debt experiment. The median expert predicted a 7.0 percentage-point reduction in depression (8.0 points when weighted by confidence), a 10.2 percentage-point reduction in borrowing (13.7 points when confidence-weighted), and meaningful improvements in healthcare access. In total, 75.6% of respondents predicted medical debt relief is at least a moderately valuable use of charity resources, and 51.1% thought it very or extremely valuable. The authors estimate a statistically insignificant 3.2 percentage-point increase in depression (not a decrease), and a 95% confidence interval that rules out a reduction in depression of more than 0.6 percentage points — far below the 7.0 percentage-point expert prediction.

Q9. What survey methodology was used, and what response rate was achieved?

The survey, administered by NORC at the University of Chicago, targeted a random subset of 14,922 hospital debt experiment participants who entered the study after September 2019 (waves 6-18) and owed at least $500. The protocol spanned 13 weeks and included five postal mailings (including a $2 upfront incentive and a $5 incentive with the paper survey), twice-weekly email reminders, certified mail delivery of the full survey instrument, and telephone interviews by a US-based call center. Respondents received a $50 completion incentive. The protocol achieved a 19.4% response rate, with 68% responding via web, 10% via telephone, and 23% via mail. The survey was titled “Health and Financial Wellness Study” and made no reference to RIP Medical Debt to avoid priming respondents. Respondents were surveyed on average 13 months after treatment assignment (interquartile range 10 to 17 months).

Q10. What heterogeneity in survey outcomes was detected, and how do the authors interpret the anomalous depression finding for high-debt recipients?

Across all four preregistered heterogeneity dimensions (medical debt amount, age of debt, age of person, amount of other debt in collections), null effects on survey outcomes were found in 15 of 16 subgroups. The exception is persons in the fourth quartile of medical debt eligible for relief, for whom debt relief caused a statistically significant 12.4 percentage-point increase in depression (p-value 0.002) relative to a control mean of 45.9%, with similar patterns for anxiety, stress, subjective well-being, and general health. The authors consider this may be a statistical fluke given the null results across all other 15 groups. They also note potential parallels with findings from unconditional cash transfer experiments, where the receipt of transfers raised the salience of financial deprivation without addressing its underlying causes. A charity-stigma mechanism (recipients did not request the assistance) is also considered. The authors caution against giving this result undue weight in the overall assessment.

Q11. How does the paper position downstream debt relief relative to upstream interventions, and what does prior evidence suggest about upstream alternatives?

The authors highlight that their null results do not extend to upstream medical debt relief. Adams et al. (2022), studying a hospital financial assistance program at Kaiser Permanente that bundled debt relief with reductions in cost-sharing close to the time of the medical event, found substantial increases in high-value healthcare utilization. The Oregon Health Insurance Experiment (Baicker et al. 2013) found that Medicaid reduced depression by 9 percentage points among low-income uninsured adults. The authors suggest several reasons why downstream relief may fail: the intervention occurs too late after the precipitating event (approximately 15 months after the medical service in the hospital debt experiment, and about 7 years in the collector debt experiment), patients may have habituated to the stress of debt collections, the relief amount may be too small relative to overall financial distress, and the direct financial benefit is inherently limited by the low market price of collections-stage debt.

Q12. How do the authors address concerns about differential survey response and external validity?

Treated persons were a statistically insignificant 1.3 percentage points more likely to respond to the survey (p-value 0.056). The authors address this in two ways. First, they estimate specifications that (i) add rich observable controls and (ii) use speed of survey response as a proxy for unobserved response propensity; neither exercise changes the estimates meaningfully. Second, to probe external validity, they test for heterogeneous effects by predicted response propensity (from a logistic regression of a response indicator on baseline characteristics) and by speed of response; neither yields evidence of differential effects for non-respondents. They also compare credit bureau treatment effects for the full hospital debt sample, the survey outreach sample, and the survey respondent sample and find similar estimates across all three groups.

Key Concepts

Downstream medical debt: Medical bills that have already been sent to third-party debt collectors by the healthcare provider after the initial billing cycle, as distinguished from upstream unpaid bills still held by the hospital at or near the time of the medical event. The paper studies debt at this late stage specifically because it is the target of most large-scale relief programs.

Credit reporting sub-experiment: An embedded quasi-experiment within the collector debt RCT, exploiting the fact that a subset of accounts (6.8% of wave 1) were still being reported to credit bureaus at the time of intervention while the debt collector had already ceased reporting for the remaining accounts. This allows separate estimation of debt relief effects with and without counterfactual credit bureau reporting, using the period until 2019 Q1 (when the collector stopped reporting entirely) as the “reporting” window.

Downstream bill repayment effect: The paper’s finding that debt relief increases the probability of a subsequent unpaid medical bill being sent to collections. The paper attributes this primarily to reduced repayment of existing pre-relief medical bills rather than to increased healthcare utilization, consistent with an expectations, targeting, or confusion mechanism — and inconsistent with an income or flypaper effect that would increase repayment.

Targeting a level of indebtedness: A behavioral model (drawn from Dobkin et al. [2018]) in which patients implicitly target a certain level of indebtedness. Under this model, relieving some debt creates headroom in the patient’s implicit debt budget, leading to reduced repayment of remaining bills to restore the targeted level of total indebtedness.

Expert survey (pre-results): A structured elicitation of predicted treatment effects conducted between April and May 2022 — after the interventions were completed but before results were released — from academics, non-profit practitioners, hospital revenue-cycle managers, and policymakers. Used as a benchmark to quantify how far the causal estimates fall below prevailing beliefs, and to document that the null results were ex ante surprising to informed observers.

PHQ-8 (Patient Health Questionnaire-8): An eight-item validated clinical screen for depression, used as the paper’s primary preregistered survey outcome. An indicator for “at least moderate depression” on the PHQ-8 is the main mental health measure against which the debt relief treatment effect is estimated.

Multimodal survey: A survey protocol combining five postal mailings, twice-weekly email reminders, certified mail delivery of a paper survey instrument, and US-based call center telephone interviews, designed to maximize response rates in a hard-to-reach low-income population with medical debt in collections.

The Future in Mind: Aspirations and Long-Term Outcomes in Rural Ethiopia

Mon, 01 Jan 0001 00:00:00 +0000

This paper tests whether a light-touch behavioral intervention targeting aspirations can produce persistent economic effects on a poor rural population. The research question is whether changing how poor people perceive their future opportunities — by raising aspirations — alters their investment decisions in ways that persist over a multi-year horizon. The authors conduct a randomized controlled trial in Doba, a remote mountainous district in rural Ethiopia roughly 380 kilometers from Addis Ababa, selected partly because its extreme isolation meant residents had almost no exposure to television or media, making even a single video screening a memorable event.

The sample consists of 1,152 households (2,112 individuals) across 64 villages. Households were randomly assigned to one of three conditions: a treatment group shown four 15-minute documentaries featuring real rural individuals from similar communities who escaped poverty through goal-setting and hard work; a placebo group shown an Ethiopian entertainment comedy with no aspirational content; and a within-village control group who were only surveyed. Both the household head and spouse in treatment and placebo groups were invited to attend. Compliance was very high, with only 2 percent of individuals not complying with their assigned condition. Data were collected at baseline (2010), six months after screening (2011), and five years after baseline (2015–2016). Attrition was notably low: 96 percent of households were re-interviewed at the five-year endline, and 94 percent of individual respondents.

Five years after the screening, treated households show meaningfully larger investment across three domains relative to the control group, with all headline results significant at 5 percent or less and robust to multiple hypothesis testing. First, on agricultural effort and investment: treated household heads and spouses work approximately one extra hour per day on their own farms (roughly 8.6 percent of the control mean per spouse). Treated households are 10 percentage points more likely to have adopted modern crop inputs (improved seeds, inorganic fertilizer) and 10 percentage points more likely to have invested in modern livestock inputs (feed, veterinary supplies). Holdings of productive tools are 20 percent higher than in the control group. Second, on educational investment: treated households spend approximately 36 percent more on children’s schooling than the control group. Among children who were of school-going age at the time of the intervention (aged 11–15 then, 16–20 at endline), the number completing full primary school is nearly double the control rate (0.16 per household versus 0.07 in the control). Third, on living standards: treated households experienced 0.33 to 0.38 fewer months of food insecurity in the previous year. Their holdings of consumer durables (furniture, kitchenware, phones) are 29 percent higher than the control group in value. Estimated house values are 27 percent higher. However, there is no statistically significant effect on measured food or frequent non-food consumption expenditure, a finding the authors interpret as consistent with households continuing to divert resources toward future-oriented investments rather than current consumption.

The intervention’s effects appear to operate primarily through aspirations — defined in this paper as desired goals for the future that motivate investment and effort. Treated households report significantly higher aspirations and expectations for income, assets, and children’s education five years later. By contrast, the paper finds no persistent changes in time preferences, risk preferences, grit, or beliefs about returns to technology. Locus of control shifted six months after the intervention but did not persist to the five-year endline, and the authors argue that if locus of control were the operative mechanism, investment effects would also have dissipated. The placebo group shows no significant effects relative to the control, ruling out screening exposure or social attention as mechanisms.

The paper is explicit about scope conditions. The study area was deliberately chosen for its extreme remoteness and media isolation, and the authors caution that this may have amplified the intervention’s salience and persistence relative to less isolated populations. External validity beyond comparable settings is uncertain. A back-of-the-envelope cost-effectiveness calculation finds that increases in durable asset holdings alone outweigh intervention costs by a factor of approximately two at reasonable scale.

Q: What was the intervention and what made it distinct from other role model studies? A: Treated households were invited to watch four 15-minute documentary films featuring real rural individuals from similar socioeconomic backgrounds who had escaped poverty through goal-setting, perseverance, and hard work. The films were produced in Oromiffa, the local language, and featured two male and two female role models depicting achievable actions such as installing irrigation or starting a small business. Unlike studies that vary exposure to in-person mentors or peers, participants received no ongoing mentorship, financial resources, or support of any kind beyond the single video screening, isolating the aspirations channel from material or informational transfers.

Q: How were aspirations measured and validated? A: Aspirations were measured using locally validated survey instruments (Bernard and Taffesse, 2014) that asked respondents what level of annual income, asset wealth, and oldest child’s education they would like to achieve in their lifetime. Test-retest reliability over two weeks produced within-respondent correlations of 0.77 to 0.98 across domains, which the authors benchmark against Angrist and Krueger (1999) standards for reliable income and education measures. The measures correlated in expected directions with wealth: mean income aspirations in the upper wealth tercile were 1.5 times those in the lower tercile, and asset aspirations in the upper tercile were 1.9 times those in the lower tercile.

Q: What were the five-year effects on agricultural effort and investment? A: Treated household heads and spouses worked approximately half an hour more per day each on their own farms relative to control, implying roughly one extra hour per day across the typical household’s adult members — an 8.6 percent increase over the control mean. Treated households were 10 percentage points more likely to have adopted modern crop inputs and 10 percentage points more likely to have invested in modern livestock inputs. Holdings of productive tools were 20 percent higher in value than in the control group. The overall agricultural investment index increased by 0.21 standard deviations relative to the control and 0.18 standard deviations relative to the placebo.

Q: What were the five-year effects on children’s education? A: Among children aged 16 to 20 at endline (who were 11 to 15, upper primary school age, at the time of the intervention), the number per household completing full primary school nearly doubled: 0.16 in the treatment group versus 0.07 in the control. These children in treated households also spent on average 33 minutes more per day attending school than the control group. Across all children, schooling expenditures in the treatment group were 36 percent higher than in the control and 30 percent higher than in the placebo. The education index increased by 0.25 standard deviations relative to the placebo and 0.21 standard deviations relative to the control.

Q: Why did consumption expenditure not increase despite improvements in assets and food security? A: The authors argue that the consumption result is theoretically ambiguous: if treated households continue to divert resources toward future-oriented investments (savings, productive assets, durable goods, housing), intertemporal substitution effects could offset income effects within the five-year observation window. The measured consumption variables — food and frequent non-food spending — do not capture the service flow value of accumulated durables or housing improvements, both of which increased substantially. The authors interpret this as evidence that households were still in an investment phase rather than having converted accumulated wealth into current consumption by endline.

Q: What evidence supports aspirations as the operative mechanism rather than alternative channels? A: The treatment group had significantly higher aspirations and expectations for income, assets, and children’s education at the five-year endline, while the placebo group did not. Measured time preferences, risk preferences, grit, and beliefs about returns to technology were all statistically unchanged for treated households. Locus of control shifted six months post-intervention but did not persist to five years, and the authors note that if locus of control were the driver, investment effects would also have dissipated alongside it. The null placebo effect rules out screening exposure, social attention, or information salience from outside facilitators as mechanisms.

Q: How were locus of control and fatalistic beliefs assessed in this population? A: The sample scored twice as high as Western samples on the classic Levenson (1981) fatalism scale. On the Feagin (1975) scale of perceived causes of poverty, the sample was more likely to attribute poverty to structural or fatalistic explanations than Western samples, and both measures of fatalistic beliefs were higher among poorer households within the sample. The study region’s worldview — rooted in traditional Waaqeffannaa religion, local variants of Orthodox Christianity (Fekade Egziabher), and Islam (Qadar) — emphasizes deference to authority, predestination, and resistance to change, providing qualitative grounding for the aspirations deficit being targeted.

Q: What were the effects on food insecurity and subjective wellbeing? A: Treated households reported 0.33 fewer months of food insecurity in the previous year relative to the control group (from a base of 2.71 months in the control), and 0.38 fewer months relative to the placebo. Treated participants scored approximately a quarter of a step higher on the Cantril ladder of self-reported wellbeing than the control group. There was no significant difference on the USDA food insecurity questionnaire, which the authors attribute to that scale’s unsuitability for households that consume largely from own production.

Q: What were the effects on durable goods and housing? A: Treated households reported 29 percent higher value of consumer durables (furniture, kitchenware, phones) than the control group and 32 percent higher than the placebo. Estimated house replacement values were 27 percent higher than the control and 21 percent higher than the placebo. Enumerators directly observed that treated households were more likely to have their own toilet facility, though this result was not significant relative to the placebo. There were no effects on the probability of having a non-organic roof, which the authors note is an especially expensive upgrade.

Q: How does the paper rule out spillover effects from treated to control households? A: The authors collected data on a supplementary sample of non-treated villages to serve as a “pure control” and used this to run a suggestive test for spillovers from treated households to untreated households within the same village. They found little evidence of large spillover effects, although they acknowledge limitations in the power of these tests. The physical design of the screenings — held in rooms with shuttered windows, requiring tickets for entry, conducted separately from placebo screenings — also minimized contamination during the intervention itself.

Q: What were the early (six-month) results and what do they suggest about the timing of effects? A: At six months, the shorter follow-up found increases in savings and investment in education, consistent with behavioral change beginning soon after treatment. Aspirations showed positive but noisier effects at immediate post-screening and six-month follow-ups, which the authors interpret as consistent with aspirations increasing gradually as people experiment with alternative futures (Appadurai, 2004) or as demotivating beliefs shift incrementally (Carvalho et al., 2023), rather than changing abruptly. This gradual pattern is consistent with a learn-by-doing dynamic where small initial investments generate returns that further raise aspirations.

Q: How does this study’s attrition and follow-up compare to the literature? A: The five-year attrition rate was very low: 96 percent of baseline households were re-interviewed and 94 percent of individual respondents. The authors cite Bouguen et al. (2019) as a benchmark, noting this is a high tracking rate relative to recent long-run RCT follow-ups in low- and middle-income countries. The low attrition strengthens confidence that endline estimates are not contaminated by selective dropout.

Q: What is the cost-effectiveness of the intervention? A: A back-of-the-envelope calculation indicates that increases in durable asset holdings alone outweigh the costs of the intervention by a factor of approximately two at reasonable implementation scale. The authors present this as a proof-of-concept estimate, not a full social cost-benefit analysis, and caution that cost-effectiveness may differ in settings with higher baseline media exposure or less extreme isolation.

Q: What are the key scope conditions limiting external validity? A: The study district (Doba) was chosen specifically for its extreme remoteness: at baseline, only 11 percent of respondents watched TV at least weekly and no household owned a television. The authors argue this isolation likely made the screening event especially salient and memorable, potentially amplifying effects relative to what would be expected in less isolated contexts. They are explicit that the findings represent a proof of concept for the aspirations mechanism and that effect magnitudes should not be assumed to replicate in settings with higher baseline media exposure or different cultural belief systems.

Aspirations: Defined in this paper as desired goals for the future that motivate investment and effort in order to attain them (following Bandura, 1977; Locke and Latham, 1990). Measured via validated survey instruments asking respondents the level of income, assets, or children’s education they would like to achieve in their lifetime — distinct from expectations (what one expects to achieve) and from the village maximum (what one believes the most successful person in the village could achieve).

Aspirations gap: The difference between an individual’s aspired level of income, assets, or education and their current reported level. Median aspirations gaps in the sample are 55 percent of median wealth aspirations and 58 percent of median income aspirations, indicating that aspirations exceed current levels by meaningful but not unrealistic margins.

Capacity to aspire: Drawn from Appadurai (2004), defined as a navigational capacity — the ability to read and navigate a map of a journey into the future. In contexts of poverty, this capacity is described as more brittle because poorer individuals have narrower social networks, fewer role models, and less material slack for experimentation with alternative futures.

Role model: A real individual from a similar socioeconomic background whose documented experience of escaping poverty through goal-setting and effort provides vicarious experience that allows audience members to imagine what is possible for people like them. Role models are most effective when their success appears attainable and when the steps to achieve it are visible.

Zero-sum beliefs: The belief that gains for one individual come at the expense of others in the community, documented in the study area as part of a broader fatalistic, deterministic belief system. These beliefs can suppress effort and future-oriented investment by making individual advancement appear normatively transgressive or materially impossible.

Source text origin: A classification in the paper’s pipeline framework distinguishing whether a summary is based on a full working paper PDF or HTML text versus abstract-only text. Abstract-only summaries are blocked as they miss scope conditions, quantitative results, and the full argument structure.

Placebo group: Households randomly invited to watch an Ethiopian comedy entertainment program (with no aspirational content) rather than the role model documentaries. Used to separate the effect of the aspirations content from the effects of the screening event itself, exposure to outside facilitators, or social attention accompanying selection for the intervention.

The Long-Run Impacts of Public Industrial Investment on Local Development and Economic Mobility: Evidence from World War II

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. Does government-led construction of large manufacturing plants in previously under-industrialized regions generate long-run improvements in regional economic development and in the lifetime earnings of the incumbent residents who were already living there at the outset? And, if so, through what mechanism — developmental improvements during childhood or expanded adult labor market opportunities?

Setting and Identification. The paper exploits the United States industrial mobilization for World War II, specifically the construction of 90 large, government-financed, newly-built manufacturing plants (each costing $10 million or more in contemporary dollars, approximately $150 million in 2020 dollars) in dispersed locations outside the major prewar manufacturing hubs. Strategic and security considerations — not economic optimization — drove the military to insist these plants be sited away from congested industrial centers. Because private firms were unwilling to finance construction in isolated locations with uncertain postwar value, the government built them directly as government-owned, contractor-operated (GOCO) facilities through the Defense Plant Corporation. Site selection within the set of sufficiently populated regions was governed by idiosyncratic, short-run factors — the immediate availability of suitable parcels, informal connections to procurement officers, and expedience — rather than systematic economic characteristics of the receiving counties. The paper documents no systematic association between publicly-funded wartime plant construction and prewar county-level economic or demographic characteristics conditional on population size, and finds parallel prewar trends and balanced outcome levels across treatment and comparison counties in all decades leading up to WWII. A placebo test using 1910-to-1940 intergenerational mobility in matched Census records confirms no differential prewar upward mobility in treatment counties.

The comparison group consists of 1,400 counties outside the 100 largest prewar manufacturing counties that did not receive large public plants. Treatment assignment for individuals is based on birth county, not adult county of residence, enabling the paper to track outcomes regardless of where individuals ultimately live.

Data. The analysis draws on the 1945 War Production Board data book for plant-level investment; county-level panels from Decennial and Economic Censuses spanning 1900–2000; the SSA NUMIDENT file (birth county and date); IRS Form 1040 individual income tax returns in 1969, 1974, 1979, and 1984 (covering wage earnings and adjusted gross income); the full-count 1940 Census (parent earnings, demographics); the 2000 Census long form (educational attainment); and W-2 earnings histories from the SSA Detailed Earnings Record matched to a CPS-linked subsample, with employer information linked to the Business Register.

Regional Effects. By 1970, counties receiving large public wartime plants had approximately 30 percent higher manufacturing employment, 20 percent larger populations, and 7–8 percent higher median family income than comparison counties. Manufacturing employment as a share of total employment rose and remained elevated through the 1970s before converging toward parity with the comparison group by 1990. Treated counties were permanently larger — with population stabilizing at a new, persistently higher equilibrium roughly 20 percent above comparison counties by end of century — even after the manufacturing employment share converged, consistent with path dependence and multiple equilibria. Average production worker pay in manufacturing rose by approximately 10 percent, closely tracking value-added per worker, while average retail wages rose by only one-third as much and were not statistically significant in most years. In the 40 years after the war, treated counties saw median family earnings increase by 5–10 percent, concentrated in higher average wages and employment shares in manufacturing and semi-skilled blue-collar occupations, with limited effects on non-manufacturing, white-collar occupations, or female individual income.

Individual Earnings Effects. Men born in treatment counties in the 18 years before the war (birth cohorts 1922–1940) earned approximately $1,200–$1,300 more per year (2020 dollars) in average wage earnings reported on 1040 returns in 1969, 1974, 1979, and 1984 — an increase of 2.5–3 percent and roughly a one-percentile rise in the national earnings distribution. Effects were largest for children of parents at the bottom of the 1939 earnings distribution: children of the lowest-income parents saw adult wage earnings rise by approximately $1,800–$2,000 per year (3–4 percent), with effects declining linearly by parent rank and effectively vanishing for children of the highest-earning parents. Black men experienced larger average earnings effects (4–6 percent, or $1,500–$2,500 in 2020 dollars) than White men (2–3 percent, or $1,000–$1,500), with the racial earnings gap estimated to have narrowed by about 2 percent in the treatment group. When examining Form 1040 returns (tax-unit level), effects are comparable for men and women, but W-2 individual earnings data from the SSA-CPS subsample show no positive effect on women’s own earnings — the 1040 effects for women are entirely driven by their husbands’ higher earnings.

Mechanism. The balance of evidence points to access to higher-wage jobs in adulthood as the primary channel, rather than developmental human capital improvements accumulated during childhood. War plants modestly increased male educational attainment — children from the lowest-earning families completed approximately one-quarter of a year more schooling and were 3 percentage points more likely to graduate high school — but education effects are too small to account for the full earnings increase. Critically, there is no gradient in earnings effects by birth cohort: children who were younger at the start of the war and therefore had longer childhood exposure to improved regions did not benefit more, contradicting a childhood exposure-effect mechanism as in Chetty and Hendren (2018b). Adult earnings effects are entirely accounted for by adult location: conditioning on 1979 county of residence eliminates the treatment effect. Stayers in treatment counties show large earnings differences relative to stayers in comparison counties, while movers show none. Men born in treatment counties are also directly documented to have worked in industries with higher wage premiums as adults, with coarse industry classification alone accounting for approximately one-third of the estimated log wage increase.

Policy Scope Conditions. The paper argues these effects are specific to the WWII postwar institutional context — high global demand for U.S. manufactured goods, limited international competition, labor-intensive production techniques, and strong union bargaining power — conditions that no longer hold. Reexamination of “million-dollar plant” openings in the 1980s and 1990s shows manufacturing employment expanded but average manufacturing wages did not increase, suggesting contemporary plant openings do not generate the same high-wage opportunities. The association between manufacturing employment density and upward mobility visible in 1950 has entirely vanished by the end of the twentieth century.

In depth

Q1. What exactly defines the treatment group, and why were these plants built by the government rather than private firms?

A: The treatment group consists of 90 counties outside the 100 largest prewar manufacturing regions that received at least one new, fully publicly-financed manufacturing plant costing $10 million or more (approximately $150 million in 2020 dollars) under the WWII industrial mobilization. Private firms refused to finance construction in dispersed, isolated locations with highly uncertain postwar value; the Air Force historians recorded that “industrialists’ reluctance to invest in dispersed plant facilities was at odds with the government’s hope that private capital could finance new inland construction.” The government built and owned these facilities as GOCO plants, operated by private firms under contract. The 353 plants meeting the cost threshold (including both large and smaller public plants) account for 70 percent of all spending on new plants during the war.

Q2. How do the authors establish that plant siting was quasi-random conditional on population size?

A: Identification rests on three forms of evidence. First, historical documents show procurement decisions were driven by idiosyncratic factors — availability of a suitable parcel, informal connections to procurement officers, short-run expedience — rather than systematic economic characteristics. Members of Congress had little ability to influence siting, and Rhode et al. (2018) find little evidence that federal politics drove the geographic distribution of wartime spending. Second, balance tests (estimating prewar county characteristics as outcomes in Equation 1) show no significant differences between treatment and comparison counties in earnings levels, demographics, manufacturing development, or industrial composition after conditioning on 1940 population, with a joint p-value of 0.30 (0.36 when also conditioning on geography and infrastructure). Third, a placebo test using children in the 1910 Census matched to the 1940 Census finds no differential economic outcomes or upward mobility rates in counties that would eventually receive treatment plants, conditional on basic region size.

Q3. What are the county-level effects on the structure of the labor market in the medium run?

A: By the 1960s–1970s, treated counties had higher predicted union coverage rates and a greater share of men in semi-skilled production occupations, driven primarily by movement away from farm work and supplemented by higher male labor force participation. Average wages in craftsperson and operator occupations rose by 8 percent in treated counties — more than double the increase in wages for high-skill professional and managerial occupations. Treated counties had 8 percent higher median male individual incomes by 1979. Effects on female median individual income were minimal, and there were no effects on female labor force participation rates.

Q4. What is the estimated magnitude of the individual earnings effects, and how do they vary by parent income?

A: Men born in treatment counties averaged $1,200–$1,300 more per year in real wage earnings (2020 dollars) on 1040 tax returns across the four observation years 1969, 1974, 1979, and 1984, a 2.5–3 percent increase equivalent to roughly one percentile in the national earnings distribution. Heterogeneity by parent rank is pronounced and monotone: children of parents at the very bottom of the 1939 earnings distribution gained approximately $2,000 per year (about 4 percent), while children of the highest-earning parents experienced no significant effect. When county weighting is equalized to eliminate the differential representation of rural (lower-income) counties, effects are roughly constant across the bottom six deciles of the parent earnings distribution and then drop steeply at the top, showing that the earnings gradient was not simply an artifact of plant openings in poorer, smaller counties.

Q5. How did effects differ by race?

A: Wartime plant construction increased annual adult earnings of Black men by 4–6 percent ($1,500–$2,500 in 2020 dollars) and of White men by 2–3 percent ($1,000–$1,500 in 2020 dollars). The racial earnings gap in the treatment group is estimated to have narrowed by about 2 percent. However, the pattern of heterogeneity by parent income differs by race: for White men, effects are largest for children of below-median parents and effectively zero for children of above-median parents. For Black men, the largest effects — 7–10 percent ($4,000–$5,000 in 2020 dollars) — accrue to children of parents with earnings above the pooled-race national median, while effects for lower-income Black families range from 3–6.5 percent, suggesting that Black workers from higher-income backgrounds particularly benefited from wartime anti-discrimination policies and the opening of previously restricted manufacturing occupations.

Q6. Why do the 1040 returns show comparable effects for men and women, while W-2 data show no effect on women’s individual earnings?

A: Form 1040 returns are filed at the tax-unit level — for married couples, they report the combined wages of both spouses. Because more than 80 percent of women in the sample are married, an increase in a husband’s earnings raises the joint 1040 figure for both spouses. The SSA-CPS subsample with individual W-2 records shows that the entire effect on men’s Form 1040 wages directly reflects increases in their own W-2 earnings, while women’s own W-2 earnings show no positive treatment effect. This finding is consistent with county-level evidence of no impact on female individual income or female labor force participation, and with Rose (2018) finding that women were almost universally excluded from manufacturing jobs after the war’s conclusion despite high wartime female manufacturing employment.

Q7. What evidence tests the developmental-effects mechanism?

A: Three tests argue against childhood developmental effects as the primary driver. First, educational attainment effects — while statistically significant for children of the lowest-income parents (approximately one-quarter of a year more schooling, 3 percentage points more likely to graduate high school) — are too small to account for the earnings increase: a Mincer-equation calculation shows that the education effects can explain less than one-half of the estimated effect on 1979 wages. Second, there is no gradient in earnings effects by birth cohort — children younger at the war’s start, who had longer post-treatment childhood exposure, did not benefit more, in direct contrast to the Chetty-Hendren childhood-exposure framework. Third, postwar in-migrants into treatment counties were not drawn from better-educated or higher-income families and did not themselves have more education than in-migrants into comparison regions, ruling out peer effects from selective in-migration.

Q8. What evidence directly implicates adult labor market access as the operative mechanism?

A: Four pieces of evidence point to contemporaneous adult labor market access. First, individuals born in treatment counties lived as adults in counties with 3–4 percent higher median male earnings and higher wages in semi-skilled blue-collar occupations but not in highly-skilled professional occupations — a pattern quantitatively consistent with the individual earnings effects. Second, the entire earnings effect is concentrated among those who remain in their birth counties: stayers in treatment counties show earnings differences of similar magnitude to county-level manufacturing wage effects, while movers show no difference compared to movers from comparison counties. Third, conditioning on 1979 county of residence eliminates the earnings effect entirely (1979 location fixed effects specification). Fourth, using W-2 data matched to the Business Register in the SSA-CPS sample, men born in treatment counties are directly shown to work in industries with higher wage premiums, with coarse industry classification alone accounting for approximately one-third of the log wage increase.

Q9. Is the persistence of regional effects driven by continued Cold War military spending at the plants?

A: No. The paper separates ordnance and ammunition plants — which predominantly became GOCO facilities or Air Force Bases after WWII and received disproportionately more Vietnam War-era defense spending — from general manufacturing plants, which overwhelmingly transitioned to privatized civilian production. Both types of plants show similarly persistent effects on manufacturing employment and comparable impacts on the long-run earnings of local children. Moreover, general manufacturing plants — which did not generate increased postwar military spending — had large permanent effects on overall population growth, while ordnance plants had smaller population effects. The persistence therefore does not appear to reflect continued federal expenditure.

Q10. What mechanism explains the permanent population effect even after manufacturing employment shares converge?

A: The authors interpret the permanent population differential — treated counties remain roughly 20 percent larger than comparison counties even at the end of the 20th century, after manufacturing employment shares converge — as evidence of path dependence and multiple equilibria. Once a region reaches a new, larger equilibrium, self-sustaining forces (expanded non-tradable employment, public infrastructure investment) maintain it. Treatment counties are more likely to have been connected to the interstate highway system in subsequent decades and show positive effects on local government capital outlays for utilities. The medium-term persistence is attributed partly to the sunk costs of site establishment (surveying, local approvals, infrastructure connections), which make reinvestment at existing sites more attractive than greenfield construction elsewhere.

Q11. Do smaller plant openings generate comparable effects?

A: No. Counties receiving smaller publicly-financed plants costing between $1 and $10 million show no detectable effects on manufacturing employment, population, median family income, or individual adult earnings comparable to those from the large plants. The authors cannot rule out the presence of small effects, but the null results for smaller plants — combined with evidence that the largest effects are in counties with the highest investment intensity per 1940 resident — are consistent with threshold effects (“big push”) in regional development, though the wide confidence intervals do not allow the authors to conclusively distinguish threshold effects from a linear-in-investment model.

Q12. What do modern “million-dollar plant” openings reveal about the contemporary relevance of these findings?

A: Reexamining plant openings from Greenstone et al. (2010) using an event-study design, the authors find that 1980s–1990s million-dollar plant openings expanded manufacturing employment (consistent with Greenstone et al.) but had no impact on average manufacturing wages — in sharp contrast to the WWII findings. Slattery and Zidar (2020) similarly find no impacts on county-level incomes for plant openings since 2000. The correlation between manufacturing employment density and upward mobility rates visible in 1950 had entirely vanished by the end of the 20th century. The authors attribute the divergent results to the changed institutional environment: contemporary production is highly automated, relies on interchangeable labor from staffing agencies, faces intense international competition, and is conducted under much weaker collective bargaining institutions.

Q13. What is the paper’s assessment of aggregate welfare implications?

A: The paper is explicit that its local estimates do not allow clean conclusions about aggregate effects. Publicly-financed plant construction in peripheral locations may have crowded out private investment that would otherwise have occurred in major manufacturing hubs. If so, the documented regional gains represent geographic reallocation of manufacturing activity rather than a net increase in the aggregate plant stock. Aggregate gains from reallocation would require that the benefits in the selected dispersed locations exceeded what would have occurred in the counterfactual locations — a plausible conjecture given the paper’s evidence that effects are larger in counties with lower prewar manufacturing employment shares and lower initial market access, but one the authors cannot demonstrate decisively.

Key Concepts

Government-Owned, Contractor-Operated (GOCO) Plants: Manufacturing facilities built and owned by a U.S. government agency (typically the Defense Plant Corporation) during WWII but built and operated by private firms under cost-plus contracts. GOCO status meant the government bore full construction risk and that post-war disposition (sale to private buyers at a fraction of construction cost, or continued GOCO operation for ordnance production) was determined by public agencies, not by the constructing firm’s investment calculus.

Place-Based Predistribution: The paper’s term for the mechanism by which wartime plant construction raised the incomes of existing residents — not through ex-post redistribution of income via taxes and transfers, but by expanding the set of high-wage employment opportunities available to incumbent workers in the region, thereby changing the pre-tax, pre-transfer wage structure facing those workers.

Adult Labor Market Access (vs. Childhood Developmental Exposure): A distinction the paper draws in explaining why children born in treated counties had higher adult earnings. The “developmental exposure” mechanism (as in Chetty and Hendren 2018b) implies benefits scale with the amount of time spent in an improved childhood environment. The “adult labor market access” mechanism means children benefit irrespective of years of childhood exposure because they can access improved local labor market conditions when they reach working age as adults — what the paper operationalizes through the finding that earnings effects are entirely accounted for by 1979 county of residence and are concentrated among individuals who remain in their birth counties.

Upward Mobility (Absolute and Relative): Following Chetty et al. (2014), the paper uses both concepts: absolute upward mobility means children from low-income backgrounds have higher lifetime earnings than comparable children in counterfactual regions; relative upward mobility means their outcomes converge toward those of children from affluent backgrounds. The paper documents both: large earnings effects for the lowest parent-income deciles, declining linearly to zero for the top deciles.

Conditional Independence (Plant Siting as Quasi-Random): The paper’s identification assumption — that among counties with observably similar population sizes and basic geographic/infrastructure characteristics, the specific choice of plant siting locations was driven by idiosyncratic, short-run factors uncorrelated with potential postwar outcomes. This is a level-balance assumption (not merely a parallel-trends assumption), required because individual outcomes are only observed in the post-period.

Industry Wage Premium: The paper uses Krueger and Summers (1988) estimates of inter-industry wage differentials (the portion of a sector’s average wage unexplained by worker characteristics) to classify adult employers of treated individuals. Finding that men born in treatment counties work at employers in higher-premium industries — with industry category alone explaining approximately one-third of the log wage increase — provides direct evidence of the adult labor market access mechanism operating through industry sorting.

Path Dependence / Multiple Equilibria in Regional Development: The paper documents that treated counties remain permanently larger in population than comparison counties even after manufacturing employment shares converge and the original plants begin to close. This self-sustaining population differential, inconsistent with a unique spatial equilibrium, is interpreted as evidence that the temporary wartime shock shifted treated regions into a permanently higher equilibrium, sustained by subsequent infrastructure investment and non-tradable sector expansion proportional to the larger population base.

The Macroeconomic Consequences of Exchange Rate Depreciations

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

How does an exchange rate depreciation causally affect macroeconomic outcomes? The paper asks whether depreciations are expansionary or contractionary, and through which mechanism. The core identification challenge is endogeneity: exchange rate changes are driven by shocks that simultaneously affect output, making causal inference from unconditional variation misleading.

Empirical Strategy

The paper studies “regime-induced” exchange rate depreciations by comparing macroeconomic outcomes for countries that peg their currency to the US dollar versus countries whose currencies float against the US dollar, in response to movements in the US dollar’s value. The identifying variation arises from the interaction between a country’s pre-existing exchange rate regime (peg vs. float) and changes in the US dollar’s nominal effective exchange rate (NEER), as measured by the BIS trade-weighted index against 24 relatively advanced economies (which are excluded from the analysis). This variation — which amounts to roughly 8% of total exchange rate variation in the sample — isolates a component of bilateral exchange rate changes that is orthogonal to idiosyncratic domestic shocks. The empirical specification is a local projection (Jorda, 2005) on annual data from 1973 to 2019 with country fixed effects and region-by-time fixed effects (four regions: Europe, Americas, Africa, Asia/Oceania). The main estimating equation regresses cumulative changes in outcome variables on the interaction term Peg × ΔUSD at horizons h = 0 to 9. Standard errors are two-way clustered by time and country. Exchange rate regime classification follows Ilzetzki, Reinhart, and Rogoff (2019); observations classified in the most ambiguous intermediate categories (coarse category 3) are dropped from the baseline.

Main Empirical Findings

Regime-induced depreciations are strongly and persistently expansionary. In response to a 1% depreciation of the US dollar, the trade-weighted nominal effective exchange rate of pegger countries depreciates by 0.74% relative to floater countries on impact, rising to 0.9% before falling back to about 0.6% over years 3–5. The real effective exchange rate depreciates by a similar but slightly less persistent amount. The GDP response builds gradually, peaking after five years at approximately 0.4% per 1% US dollar depreciation. Expressed in terms of local currency depreciation, a 10% regime-induced depreciation results in a 5.5% increase in GDP over five years. Consumption rises by nearly 0.4% of GDP at peak. Investment also rises gradually, peaking after five years.

Two findings are particularly important for identifying the transmission mechanism. First, net exports fall in response to a regime-induced depreciation. Imports rise more than exports for several years following the depreciation, ruling out an export-led boom driven by expenditure switching as the primary driver. Second, the short-term nominal interest rate rises modestly in pegging countries relative to floaters (by less than 0.1 percentage point per 1% depreciation), and the ex-post real interest rate response fluctuates around zero and is statistically insignificant throughout. This rules out looser monetary policy in pegger countries as the driver of the boom. Together, these two findings rule out a large set of standard open-economy models (including those with expenditure switching, monetary easing, and s = 0 financial frictions).

The booms are concentrated in the service sector. Manufacturing, agriculture, and mining/construction responses are close to zero, indicating a domestic demand-led boom rather than an export-led one. The GDP response is entirely driven by countries with above-median capital account openness (as measured by the Chinn-Ito index); countries with below-median capital account openness show a similar exchange rate response but no significant output response. Results are similar across the early (1973–1995) and later (1996–2019) sub-periods.

The Plaza Accord of 1985 provides a concrete illustration: the log real exchange rate of peggers depreciated by 12% (SE 2.7%) relative to floaters in the first year, while log GDP of peggers was 7.4% (SE 3.1%) higher after five years, implying a GDP response to a 10% depreciation of 6.2%, broadly consistent with the baseline estimates.

Robustness checks controlling for Peg × US GDP growth, Peg × US inflation, Peg × US interest rate, Peg × commodity price changes, and Peg × global financial cycle (Miranda-Agrippino and Rey) leave results virtually unchanged.

Theoretical Framework

To explain these facts, the paper develops a four-region model (US, Euro Area, pegs to USD, pegs to euro) with imperfect financial openness. The model features (i) UIP deviations between the euro and US dollar driven by financial shocks (ψ_t), and (ii) sticky household portfolio shares, so that households invest a fixed fraction s of savings in foreign bonds and do not fully arbitrage cross-currency return differentials. When s = 0 (no household access to foreign assets), standard theory predicts that expenditure switching and real income channels dominate, yielding rising net exports — directly contradicting the data (Proposition 2). When s > 0, a “foreign credit channel” operates: following a regime-induced depreciation, expected future appreciation of the pegger currency makes foreign-currency borrowing cheaper, stimulating domestic consumption and investment, causing imports to rise more than exports (Proposition 3), consistent with the data.

The model also accounts for unconditional exchange rate disconnect and the Mussa facts. Two shocks — UIP shocks (which generate a positive exchange rate–output correlation) and domestic discount factor shocks (which generate a negative correlation, since demand contractions lead to currency depreciations via monetary easing) — together produce a low unconditional correlation between exchange rates and output even though the conditional effect of regime-induced depreciation is large. The same logic explains why switching from fixed to floating exchange rates raises exchange rate volatility dramatically without raising macroeconomic volatility commensurately: pegging eliminates UIP shock exposure (reducing output volatility) but removes the ability to use monetary policy to offset discount factor shocks (raising output volatility), and these two effects roughly offset each other in the quantitative model.

In depth

Q1. What is the core identification strategy, and what assumption is required for it to yield causal estimates?

A1: The strategy compares macroeconomic outcomes in countries pegged to the US dollar versus countries floating against the US dollar when the US dollar’s value changes. The identifying assumption is that peggers are not differentially exposed (relative to floaters) to aggregate shocks that are correlated with the US dollar exchange rate. If this holds, the direct effects of shocks driving the US dollar move pegs and floats symmetrically and are absorbed by region-by-time fixed effects, leaving only the regime-induced component. Differential exposure to US dollar-correlated shocks is the main threat to identification, but the paper shows robustness by controlling for interactions of the peg indicator with US GDP growth, US inflation, US interest rate changes, commodity price changes, and the global financial cycle.

Q2. How is “regime-induced” exchange rate variation defined, and how large is it relative to total variation?

A2: Regime-induced variation is the component of a country’s exchange rate change that arises from its pre-existing regime vis-à-vis the US dollar interacted with the change in the US dollar’s nominal effective exchange rate. It is identified via the interaction term Peg_i,t × ΔUSD_t in the local projection. This variation represents roughly 8% of total variation in exchange rates in the sample, so the strategy isolates a small but clean slice of total exchange rate movements.

Q3. How do nominal and real effective exchange rates respond for peggers versus floaters?

A3: In response to a 1% depreciation of the US dollar, the trade-weighted nominal effective exchange rate of peggers depreciates by 0.74% relative to floaters on impact, peaks around 0.9%, and then gradually declines to roughly 0.6% over years 3–5. The real effective exchange rate depreciates by a similar but slightly less persistent amount. The less-than-one-for-one response occurs because the classification includes imperfect pegs and imperfect floats; however, this misclassification attenuates both the first stage (exchange rate response) and the reduced form (output response) proportionally, so the ratio — the IV-style estimate — remains unbiased.

Q4. What is the quantitative magnitude of the output effect, and how is it computed?

A4: In response to a 1% US dollar depreciation, GDP of peggers rises by approximately 0.4% relative to floaters, peaking after five years and building gradually. To express this as a response to a 10% local currency depreciation: the average nominal exchange rate response over the first five years is roughly 0.7%, so the implied GDP response per 10% depreciation is 10 × 0.4 ÷ 0.7 ≈ 5.5%. The Plaza Accord case study yields a similar magnitude: a 12% first-year real exchange rate differential is followed by a 7.4% differential in log GDP after five years, implying 6.2% per 10% depreciation.

Q5. Why does the behavior of net exports rule out the expenditure-switching mechanism as the primary driver?

A5: Standard open-economy models predict that a depreciation improves competitiveness, boosting exports and reducing imports — generating an improvement in net exports as the engine of expansion. The paper finds the opposite: imports rise more than exports for several years following a regime-induced depreciation, so net exports fall. This is inconsistent with an export-led expenditure-switching boom. The finding is also inconsistent with the real income channel (as formalized in Proposition 2): even with s = 0, standard models predict rising net exports, but the data show the reverse.

Q6. Why does the behavior of interest rates rule out monetary policy easing as the driver?

A6: If the US dollar depreciated because of loose US monetary policy, countries with currencies pegged to the US dollar would share US monetary policy more strongly, and one would expect a relative decline in nominal interest rates for peggers. The opposite is found: the nominal interest rate of peggers rises slightly relative to floaters (by less than 0.1 percentage point per 1% depreciation), and the real interest rate response is statistically indistinguishable from zero throughout the nine-year horizon. This rules out the interpretation that the boom is driven by an easing of monetary conditions in the pegger countries.

Q7. What are ex-post UIP deviations, and what do they imply about the shock driving the variation?

A7: Ex-post UIP deviations measure the excess return to holding assets denominated in pegger currencies relative to floater currencies. After the initial depreciation of pegger currencies, those currencies subsequently appreciate and their nominal interest rates are (if anything) higher than floater interest rates. This means the ex-post return to holding pegger-currency assets is higher than for floater-currency assets — a positive UIP deviation that builds over several years after the shock. These deviations imply that the shocks driving the US dollar depreciation are financial in nature (UIP shocks), not changes in expected near-term monetary policy fundamentals.

Q8. What is the foreign credit channel, and how does it work in the model?

A8: The foreign credit channel (the second term in equation (18) of Proposition 1) operates through the cost of foreign-currency borrowing. When the pegger currency depreciates on impact and then is expected to appreciate subsequently, the exchange-rate-adjusted cost of borrowing in foreign currency falls — that is, expected future appreciation of the domestic currency reduces the real cost of foreign credit. To the extent that households have portfolio shares in foreign bonds (s > 0), this stimulates consumption via intertemporal substitution. The channel is operative only when s > 0; with s = 0 (no household access to foreign assets), net exports must rise rather than fall (Proposition 2), contradicting the data.

Q9. How does Proposition 1 establish that real interest rates and real exchange rates are sufficient statistics for the relative responses of all macroeconomic aggregates in this setting?

A9: Under Assumption 1 (pegs to the US dollar and pegs to the euro face symmetric non-monetary fundamental shocks), the relative responses of consumption, output, exports, and imports of USD-peggers versus euro-peggers are functions only of the relative path of the real interest rate and the real effective exchange rate. This is because the underlying shocks to the US economy and the Euro Area economy are common to both groups of peggers and cancel out in the comparison. The monetary regime of a country is fully summarized by the paths of the real interest rate and the real exchange rate. Since the estimated relative real interest rate response is close to zero, the paper infers that the observed output differential must arise from the real exchange rate path — hence the title.

Q10. Why does the output response differ by capital account openness but not by trade openness?

A10: The GDP response to a regime-induced depreciation is entirely driven by countries with above-median capital account openness (Chinn-Ito index). Countries below the median show a similar real exchange rate response but no significant output response. In contrast, splitting by trade openness (exports plus imports as a share of GDP) yields similar output responses in both sub-groups. This pattern is consistent with the model’s foreign credit channel, which operates through international capital flows (the parameter s representing financial openness). Countries with restricted capital accounts cannot borrow cheaply from abroad when their currencies become “cheap,” so the foreign credit channel is shut down. The result is inconsistent with the expenditure-switching channel, which would predict larger effects for more trade-open economies.

Q11. What is the sector composition of the output boom, and what does it imply about the transmission mechanism?

A11: The bulk of the output response is concentrated in the service sector. Manufacturing, agriculture, and the mining/construction/energy sectors show responses close to zero, with only a modest boom in the latter at very long horizons. Services are predominantly non-tradable, so this sectoral pattern is consistent with a domestic demand-led boom (via the foreign credit channel) rather than an export-led boom (via expenditure switching on tradable goods). The foreign credit channel stimulates domestic demand broadly, which disproportionately raises output in the non-tradable sector.

Q12. How does the model reconcile large conditional effects of exchange rates with unconditional exchange rate disconnect?

A12: The paper introduces two shocks: UIP shocks (ψ_t) and domestic discount factor shocks (β_t). UIP shocks cause the exchange rate to depreciate and output to rise (a positive conditional correlation). Discount factor shocks reduce domestic demand; monetary policy responds by lowering interest rates, which depreciates the exchange rate, but if the monetary response is insufficient to fully offset the shock, output falls — generating a negative conditional correlation between the exchange rate and output. The unconditional correlation between the exchange rate and output is a weighted average of these two conditional correlations. If these effects are of similar magnitude and opposite sign, the unconditional correlation can be close to zero even though each structural shock generates a large conditional response. This is directly analogous to how supply and demand shocks can generate a small unconditional price-quantity correlation in a standard market setting.

Q13. How does the model provide a new interpretation of the Mussa fact?

A13: The Mussa fact is that the collapse of Bretton Woods dramatically increased the volatility of real exchange rates in countries that switched to floating, without a corresponding increase in macroeconomic volatility. In the model, pegging has two opposing effects on output volatility: it insulates the economy from UIP shocks (reducing output volatility) but prevents the use of monetary policy to offset discount factor shocks (raising output volatility). In the quantitative model (Appendix D), these effects roughly offset each other, so moving from a peg to a float raises exchange rate volatility substantially while leaving macroeconomic volatility roughly unchanged — consistent with the Mussa fact. This contrasts with the Itskhoki-Mukhin interpretation, which attributes Mussa facts to exchange rates (driven by UIP shocks) having little effect on output; in the present paper, the conditional effects are large but cancel in the unconditional moments.

Q14. What does the paper imply for the tradeoffs associated with adopting a fixed versus flexible exchange rate regime?

A14: Traditional analyses of the monetary trilemma emphasize that pegging to the US dollar forces a country to follow US interest rate policy. The paper argues that a first-order consequence of pegging — one that may outstrip the traditional monetary policy tradeoff in importance — is that the country imports the financial shocks (UIP shocks) that drive the US exchange rate while potentially reducing its exposure to home-grown financial shocks. When the US dollar depreciates due to financial shocks, pegger countries experience a stimulatory foreign credit inflow. Conversely, when the US dollar appreciates due to financial shocks, pegger countries face tighter financial conditions. The importance of this financial shock trade-off, the paper argues, may greatly exceed the importance of the traditional monetary trilemma in environments where financial shocks are a dominant driver of exchange rate fluctuations.

Q15. How does the paper handle the potential concern that the peg classification is imperfect?

A15: The paper notes that misclassification of pegs and floats attenuates both the exchange rate response (first stage) and the output response (reduced form) proportionally. Since the ultimate quantity of interest is the ratio of the output response to the exchange rate response (analogous to an IV estimate), misclassification in both the numerator and denominator does not introduce bias. This is analogous to an instrumental variables regression where the first stage need not have a high R-squared for the IV estimate to be valid. The paper also shows robustness to alternative treatments of the ambiguous intermediate categories (Ilzetzki-Reinhart-Rogoff coarse category 3), including them as pegs or floats, with similar results in both cases.

Key Concepts

Regime-induced depreciation: A change in a country’s bilateral exchange rate that arises specifically because the country has a pre-existing peg (or float) to a reference currency, and that reference currency’s value changes in world markets. The variation is defined as the component of a country’s exchange rate movement driven by the interaction between its exchange rate regime vis-à-vis the US dollar and changes in the US dollar’s nominal effective exchange rate. This is distinguished from all other exchange rate variation, including that driven by domestic idiosyncratic shocks.

Foreign credit channel: The mechanism in the paper’s model through which a regime-induced depreciation stimulates domestic demand. When the domestic currency depreciates on impact and is expected to appreciate subsequently, the exchange-rate-adjusted cost of borrowing in foreign currency falls. Households with portfolio shares in foreign bonds (s > 0) borrow more cheaply from abroad, stimulating consumption via intertemporal substitution. This channel requires imperfect financial openness (s > 0 but not full UIP arbitrage) and predicts that the output boom is domestic-demand-led with falling net exports — consistent with the data.

UIP shock (ψ_t): An exogenous shock to uncovered interest parity between the US dollar and the euro, interpreted as arising from frictions in international financial markets or from exogenous shifts in demand for one currency over another. A positive ψ_t represents an increase in demand for the euro (relative to the US dollar), depreciating the US dollar. These shocks are the paper’s preferred interpretation of the financial shocks driving the US dollar exchange rate, consistent with the observed joint behavior of exchange rates and interest rates.

Imperfect financial openness (parameter s): The share of household savings invested in foreign bonds. At s = 0, households have no access to foreign assets (as in Gabaix-Maggiori and Itskhoki-Mukhin); at full financial integration with UIP holding (ψ_t = 0), there is no foreign credit channel. The paper’s model is intermediate: s > 0 but portfolio weights are sticky, so households do not fully arbitrage cross-currency expected return differentials. The foreign credit channel is operative only when s > 0, and the strength of the output boom is increasing in s/σ (the ratio of financial openness to the coefficient of relative risk aversion).

Sufficient statistics (real interest rate and real exchange rate): Under Proposition 1, conditional on Assumption 1 (symmetric non-monetary fundamental shocks across pegger groups), the relative responses of all macroeconomic aggregates for peggers to the US dollar versus peggers to the euro are functions only of the relative path of the real effective exchange rate and the relative path of the real interest rate. The full set of underlying shocks — monetary, financial, productivity, or discount factor — does not need to be separately identified; only the paths of these two prices matter for relative macroeconomic outcomes.

Exchange rate disconnect: The empirical finding, documented extensively since Meese and Rogoff (1983), that exchange rates have very low unconditional correlations with macroeconomic aggregates such as output and consumption. In the paper’s sample, real exchange rates of floating countries are three to four times more volatile than GDP and consumption, and the unconditional correlation of the real exchange rate with GDP is mildly negative (around −0.05 to −0.07). The paper offers a new explanation: this low unconditional correlation reflects the cancellation of large but opposite-signed conditional correlations from UIP shocks and discount factor shocks, rather than indicating that exchange rates have small effects on the economy.

Mussa fact: The empirical observation (Mussa, 1986) that when countries switched from fixed to floating exchange rates after the collapse of Bretton Woods, real exchange rate volatility increased dramatically — for floaters roughly 50–60% higher standard deviation in the paper’s sample than for peggers — but the volatility of GDP, consumption, and other macroeconomic aggregates did not increase correspondingly. The paper interprets this through its two-shock model as the result of two opposing effects of pegging: insulation from UIP shocks (which reduces macroeconomic volatility) versus inability to use monetary policy to offset discount factor shocks (which raises macroeconomic volatility), with the two effects roughly offsetting in the quantitative model.

The Macroeconomic Impact of Climate Change: Global Versus Local Temperature

Mon, 01 Jan 0001 00:00:00 +0000

The paper shows that the macroeconomic impact of climate change is an order of magnitude larger than what standard country-level panel estimates suggest. The key identification innovation is to measure the effect of global mean temperature shocks using time-series local projections, rather than using cross-country variation in local temperatures as in the conventional panel literature. A shock to global mean temperature tracks extreme weather events (droughts, heat waves, wind, precipitation anomalies) that affect all countries simultaneously; a local temperature anomaly in one country does not.

Empirical approach: The authors estimate local projections of world GDP growth on exogenous global mean temperature shocks. The shock is the innovation to global mean temperature after removing a 2-year autoregressive component and a low-frequency trend, following Hamilton (2018). Two estimation samples: BU (Barro-Ursúa macro history, 43 countries, 1860–2019) and PWT (Penn World Tables, 173 countries, 1960–2019).

Key empirical results (Section 3): A 1°C shock to global mean temperature causes world GDP to fall by 14% after 6 years in the PWT sample (95% CI: 6%–22%); significant at the 5% level in years 2–8; does not mean-revert within the 10-year sample horizon. In the BU sample, the peak GDP decline is 18% after 5 years (95% CI: 6%–30%). Converting the cumulative IRF ratio to a permanent temperature change yields a 22–34% long-run GDP decline per 1°C of permanent global warming (PWT and BU respectively). By contrast, local temperature shocks — estimated from a standard cross-country panel with country and year fixed effects — generate effects of 1–3% per °C, not statistically significant at the 5% level.

Why global > local (Section 4): Four categories of extreme climatic events (heat waves, droughts, wind, precipitation anomalies) jointly account for roughly half of the estimated global temperature effect on GDP. None of these are strongly correlated with local temperature anomalies because extreme weather reflects ocean-atmosphere dynamics (El Niño/ENSO) that elevate global mean temperature rather than any single country’s local temperature. In addition, capital and investment both decline persistently after global temperature shocks (capital response significant at 5% level), and warm/low-income countries are disproportionately affected.

Structural model (Section 5): A parsimonious neoclassical growth model embeds climate change as aggregate TFP changes. Households maximize ∫e^{−ρt}U(C_t)dt; firms use Cobb-Douglas technology Z_t K_t^α L_t^{1−α}. The damage function governing TFP is:

Z_t = Z_0 exp( ∫0^t ζ_s T̂{t−s} ds )

where T̂_t is excess global mean temperature above baseline and ζ_s = A(e^{−Bs} − e^{−Cs}) is the structural damage function. When ζ_s → 0, shocks have level but not growth effects; no statistically significant evidence of growth effects is found in Figure 3 of the paper. The model is calibrated with: risk aversion γ = 1 (log utility), capital share α = 0.33, annual capital depreciation δ = 0.08, and pure time preference ρ = 0.02. Proposition 1 (model inversion) shows that, to first order, ŷ_t = ẑ_t + α ∫K_{t,s} ẑ_s ds, where K_{t,s} is the sequence-space Jacobian of the neoclassical growth model. This delivers identification: observed output impulse responses recover the structural TFP damage function ζ_s without imposing functional form on the capital channel.

Estimation results (Section 5.3, Figure 12): The estimated damage function implies a 4% peak short-run productivity decline 2 years after a 1°C transitory global temperature shock; the effect decays slowly and remains significant for up to 10 years. The capital response (non-targeted moment) closely matches its empirical counterpart, providing an overidentification check. The local temperature damage function, estimated by targeting the local-panel output IRF, peaks at only 0.5% and is more than 8× smaller in cumulative productivity effect; it is not statistically different from zero at the 5% level.

Business-as-usual counterfactual (Section 6.1–6.2): Temperature rises from 2024, reaching 3°C above preindustrial by 2100 (asymptoting to 3.3°C), equivalent to 2°C of additional warming since 2024. Under the global temperature damage function:

World output by 2050: −28% vs. no-warming baseline
World output by 2100: −53% (accumulated TFP losses reach −40%)
Capital by 2100: −51% (investment initially rises as households anticipate lower permanent income, then decumulates rapidly)
Consumption by 2100: −53%
2024 welfare loss (consumption equivalent): 35%; welfare continues declining as temperatures rise, eventually reaching 56%
95% CI for 2100 output loss: 29%–77%
All effects statistically significant at the 5% level

Under the local temperature damage function with the same warming scenario: long-run output declines only 9%, welfare loss is 5%, and neither is statistically significant at the 5% or 10% level — consistent with conventional estimates (Nordhaus 1992, Dell et al. 2012, Burke et al. 2015).

Social Cost of Carbon (Section 6.2, Panel F): The SCC is defined as the consumption-equivalent amount households would pay at time 0 to avoid one additional ton of CO2, using the temperature-response function from Dietz et al. (2021a). Baseline result: $1,207 per ton (2024 international dollars), more than 6× larger than the $185/ton estimate in Rennert et al. (2022). 95% CI: $399–$2,015 per ton. Climate sensitivity range (half/double median): $600–$2,400 per ton. BU sample (larger damage functions): >$1,500 per ton. Using the local temperature damage function yields an SCC of only $149/ton, consistent with conventional estimates.

Sensitivity (Section 6.4): Higher time preference ρ > 0.04 lowers welfare losses below 20% and the SCC below 3× conventional high-end estimates — the only scenario where results converge toward prior estimates. Near-Stern discount rates (ρ → 0): welfare loss >40% and SCC >$2,500/ton. A 6°C-by-2100 scenario yields welfare losses >60%.

Historical growth accounting (Section 6.3): Starting the model in 1960 and imposing the realized 1960–2019 warming path, then holding temperature constant at its 2019 level, reveals:

World GDP per capita would be 25% higher today without warming since 1960
By 2040, output is 32% below potential from past warming — one-quarter of losses from historical warming are yet to materialize (due to delayed damage function and transitional capital dynamics)
Climate change reduced the annual world growth rate by as much as a third of baseline by the 21st century

Policy implication: Most decarbonization interventions cost ~$80/ton on average (Bistline et al. 2023). Under conventional SCC estimates based on local temperature ($149/ton), the US Domestic Climate Cost (DCC) falls below policy cost, making unilateral emissions reduction prohibitively expensive. Under the paper’s global temperature SCC of $1,207/ton, the DCC of the United States exceeds $80/ton even accounting for the fraction of global climate benefits that accrue domestically — unilateral decarbonization becomes cost-effective for large economies such as the US.

Scope conditions: The neoclassical model abstracts from adaptation, mitigation, trade, urbanization, and endogenous emissions. The identification assumption requires that global mean temperature innovations are uncorrelated with other global economic confounders at business-cycle and trend frequencies; the paper checks robustness against alternative detrending, exclusion of WWII and COVID-19 years, El Niño/ENSO controls, and instrumental variables for temperature based on solar/volcanic forcing. The conversion from medium-run to long-run effects relies on the constrained ζ_s = A(e^{−Bs} − e^{−Cs}) functional form ruling out growth effects — consistent with the data but not formally testable beyond the 10-year horizon. Counterfactuals involve 2–3°C temperature changes substantially beyond the sample’s moderate perturbations; the model’s extrapolation may understate damages if nonlinearities exist at extreme temperatures (the authors note their conservative constrained-form approach).

Summary of a forthcoming paper, AI-assisted and human-reviewed. See the linked original for the authoritative claims and full conditions.

In depth

Q1. Why do global temperature shocks produce GDP effects an order of magnitude larger than local temperature panel estimates?

Global mean temperature shocks are strongly correlated with extreme weather events — heat waves, droughts, wind storms, and precipitation anomalies — that simultaneously affect all countries; these four event categories jointly account for roughly half of the global temperature effect on GDP. Local temperature anomalies in a given country (as measured in standard cross-country panels with year fixed effects absorbed) are not correlated with these same events, because El Niño/ENSO and related ocean-atmosphere dynamics elevate global mean temperature without proportionally elevating any one country’s local temperature. Local panel studies also implicitly allow economic activity to shift toward cooler regions within a given year — an option unavailable when global warming affects all locations simultaneously. The resulting bias in local-panel estimates is not “aggregation bias” in the sense of Jensen’s inequality, but rather an identification problem: local panels identify a different object (the effect of temperature relative to other countries in the same year) rather than the aggregate climate impact the paper measures.

Q2. What is the identification strategy and what are the main threats?

The global temperature shock is identified as the innovation to global mean temperature after removing a 2-year AR component and a Hamilton (2018) low-frequency trend, yielding a shock orthogonal to its own recent history and to long-run trends. The main threats are: (i) global business-cycle confounders (worldwide recessions that simultaneously lower activity and emissions), addressed by controlling for quadratic time trends and global aggregate demand proxies; (ii) reverse causality (economic expansion warming the atmosphere), addressed by IV estimates using solar/volcanic forcing as instruments; (iii) low-frequency correlation between climate trends and productivity growth, addressed by flexible detrending and robustness to sample period. All major specification checks generate quantitatively similar results, and the paper passes placebo tests for large global confounders (WWII, COVID-19).

Q3. How does the structural model translate medium-run shock responses into long-run warming effects?

Proposition 1 (model inversion) shows that the output impulse response decomposes into a direct TFP effect ẑ_t and a capital channel ŷ_t = ẑ_t + α ∫K_{t,s} ẑ_s ds, where K_{t,s} is the sequence-space Jacobian of the neoclassical growth model (Auclert et al. 2021); this allows recovery of the structural TFP damage function {ζ_s} from the observed 10-year output IRF by non-linear least squares, without having to observe TFP directly. The counterfactual for a gradually rising temperature path (BAU scenario with 2°C additional warming since 2024) is then solved via the full nonlinear model — not via the log-linearization used in estimation — because the 2–3°C excursion far exceeds the sample’s modest temperature perturbations. The capital response (non-targeted moment) closely tracks its empirical counterpart, providing a strong overidentification check that the model’s capital dynamics are correctly specified.

Q4. Why does capital initially rise in the BAU counterfactual before declining?

Following standard permanent-income logic, when households learn at date 0 that global temperatures will rise and future TFP will fall, they temporarily increase saving and investment to accumulate buffer capital before the productivity decline materializes; this front-loads some capital accumulation in the early transition years (2024–2030s), briefly pushing capital above baseline, before the accumulated TFP losses overwhelm the saving motive and capital begins an extended decline. The net effect is still a 51% capital shortfall by 2100 because persistently lower TFP reduces the marginal product of capital over decades, depressing investment and allowing the capital stock to drift far below its no-warming balanced growth path.

The SCC is defined as the dollar amount C such that households are indifferent between (a) a world where one additional ton of CO2 is emitted at time 0 and (b) a world in steady-state where the household has paid C at time 0 (equation 7: V^{ss}(K^{ss} − C) = V^{SCC}_0(K^{ss})). The temperature response to a 1-ton CO2 pulse is taken from Dietz et al. (2021a) — temperature peaks at 0.002°C after a 1-gigaton pulse and stabilizes. The model generates the productivity path {Z^{SCC}_t} via the structural damage function, solves for equilibrium capital and consumption paths, and computes the value function V^{SCC}_0. The resulting $1,207/ton exceeds prior estimates by 6× because the global-temperature damage function implies 4% peak TFP losses per 1°C transitory shock, compared to the ~0.5% peak implied by local temperature — and the SCC is essentially the capitalized sum of these future productivity losses, so the ratio scales proportionally.

Q6. Why are historical climate losses so large if year-to-year warming is small?

The key is cumulation: annual warming increments are individually small (tenths of a degree), but the damage function {ζ_s} is persistent (effects last 10+ years), so each year’s increment adds a flow of persistent TFP losses that stack on top of prior increments. The paper’s growth accounting shows that climate change reduced the world growth rate by up to one-third of baseline in the 21st century — a number that appears modest in any single year but, compounded over decades, translates into a 25% GDP per capita shortfall by 2019. Additionally, because the estimated damage function has a 2-year lag before peak TFP impact, a substantial share of past warming’s losses are yet to be realized — the paper estimates GDP will be 32% below its potential by 2040 even with no further warming.

Q7. What does the sensitivity analysis reveal about the robustness of the results?

The key sensitivity is the rate of time preference ρ: at ρ = 0.02 (baseline, consistent with secular interest rate decline), welfare loss is 35%; at ρ = 0.04 (above recent market rates), welfare loss is still above 20%; only at implausibly high discount rates does the welfare loss fall below 15%. The SCC is more sensitive to ρ than welfare because the SCC is a capitalized stock valuation while welfare is an annualized flow. BU sample damage functions (larger IRF) raise welfare loss to 42% and 2100 GDP loss to 61%; these represent the high end of the estimates. The climate sensitivity range ($600–$2,400/ton for the SCC) reflects uncertainty in the physics of CO2-to-temperature conversion, not in the estimated economic damage function. Across all these dimensions, the global-temperature estimates remain order-of-magnitude larger than local-temperature estimates.

Q8. What is the policy implication for large economies considering unilateral decarbonization?

The domestic decarbonization test compares the Domestic Climate Cost (DCC) — the fraction of the global SCC that accrues to the decarbonizing country — against the marginal cost of abatement (~$80/ton average, Bistline et al. 2023). Under conventional local-temperature estimates ($149/ton global SCC), the US DCC falls below $80/ton, implying unilateral action destroys domestic value. Under the paper’s $1,207/ton global SCC, the US DCC comfortably exceeds $80/ton even if the US only captures a fraction of world welfare gains — because global temperature extremes (hurricanes, heat waves, droughts) strike the US directly, the DCC/SCC ratio is much higher than under local estimates where the US appears less exposed. This fundamentally changes the cost-benefit calculus for large-economy unilateral climate policy.

Key concepts

global mean temperature shock: a time-series innovation to world average surface temperature, identified by Hamilton (2018) detrending; captures ocean-atmosphere climate variability (El Niño/ENSO) correlated with extreme weather events affecting all countries simultaneously; the paper’s key identification variable, distinct from local temperature variation used in standard cross-country panels.

global vs. local temperature effect: the paper’s central finding that the GDP effect per 1°C global mean temperature shock (14–18%) is an order of magnitude larger than the effect per 1°C local temperature shock (1–3%); the gap is explained by extreme climatic events (heat waves, droughts, wind, precipitation) that co-move with global mean temperature but not with individual countries’ local temperatures.

structural damage function (ζ_s): the kernel relating excess global mean temperature T̂_{t−s} to log TFP at time t, specified as ζ_s = A(e^{−Bs} − e^{−Cs}); estimated from the PWT output impulse response via model inversion (Proposition 1); implies a 4% peak TFP loss 2 years after a 1°C transitory shock, decaying slowly over 10 years; rules out permanent growth effects consistent with the data.

Social Cost of Carbon (SCC): the one-time dollar amount households would pay at time 0 to avoid one additional ton of CO2; equals (in the linear limit) the present discounted value of all flow consumption-equivalent welfare losses from the induced warming; paper estimates $1,207/ton (2024 international dollars), more than 6× prior estimates, because the global-temperature damage function implies much larger per-degree productivity losses than local-temperature estimates.

committed climate losses: future GDP shortfalls already locked in by past warming, arising because the estimated damage function has a delayed peak (year 2) and slow decay (10+ years) — temperature rises in recent years continue reducing productivity for the following decade; the paper estimates these committed losses alone will lower GDP 32% below potential by 2040 even with temperature held constant at 2019 levels.

BAU scenario: the business-as-usual warming path used for the main counterfactual — global mean temperature reaches 3°C above preindustrial by 2100 (asymptoting to 3.3°C), implying 2°C of additional warming since the 2024 baseline; under this scenario the model implies 53% GDP loss, 51% capital loss, 53% consumption loss, and a 35% consumption-equivalent welfare loss by 2100.

The Optimal Taxation of Couples

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question. What is the optimal joint nonlinear earnings tax schedule for married couples? How should one spouse’s marginal tax rate depend on the other’s earnings? When is individual earnings-based (separable) taxation optimal versus family-income-based taxation, and what determines the sign and magnitude of “jointness” — the dependence of one spouse’s marginal tax on the other’s earnings?

Model. The paper studies a canonical unitary household model in which each couple consists of two spouses who jointly maximize utility subject to a joint budget constraint. Spousal productivities are drawn from a joint distribution F with arbitrary dependence structure. The planner maximizes a weighted sum of couples’ utilities, with Pareto weights that are decreasing functions of productivities. Utility takes a quasi-linear form in consumption and labor disutility with constant labor supply elasticity parameter γ (implying earnings elasticity γ/(γ-1)). The tax problem is equivalent to a two-dimensional mechanism design problem in which the planner chooses allocations as functions of reported productivity types, subject to incentive compatibility and budget feasibility. Because spousal productivities are two-dimensional, the problem is a multi-dimensional screening problem whose properties are poorly understood in general.

Methodology. The authors proceed in two directions. First, they establish conditions under which the first-order approach (FOA) — restricting attention to local incentive constraints — is valid in this bi-dimensional setting. They show, for the special case of the benchmark economy (symmetric, independent types, separable Pareto weights), that FOA validity is equivalent to convexity of a certain transformation of the value function, and derive necessary and sufficient conditions that are strictly weaker than their unidimensional analogs — so the FOA is more likely to hold in two dimensions than in one. For the general economy, they invoke an Implicit Function Theorem argument in Hölder space to show that the FOA holds for Pareto weights sufficiently close to utilitarian (i.e., when the planner is not “too redistributive”). Second, assuming FOA validity, they characterize optimal taxes via a second-order nonlinear PDE. Since this PDE cannot be solved analytically in general, they apply the Coarea Formula to derive closed-form expressions for conditional averages of optimal tax distortions over various subsets of the type space, expressed entirely in terms of structural primitives (labor supply elasticities, Pareto weights, and elasticities of the joint distribution of productivities).

Main Findings.

Average distortions and assortativeness. Average optimal distortions on married individuals are ranked by the degree of positive quadrant dependence (PQD) in spousal productivities: more assortative matching implies higher optimal tax rates. Optimal distortions on married individuals are always weakly lower than on single individuals with the same productivity, same elasticities, and same marginal productivity distribution — strictly so unless matching is perfectly positively assortative. The intuition is that when couples pool resources, intra-family redistribution already occurs, and distortionary taxation crowds this out; more random matching produces more within-family redistribution, reducing the marginal social value of public redistribution through taxation.
Optimality of separable (individual earnings-based) taxation. In the benchmark economy with independent types, optimal taxes are exactly separable (individual earnings-based), and optimal distortions on married individuals equal precisely one-half of those on comparable single individuals. With separable Pareto weights and independent types more generally, taxes remain separable. Once types are positively dependent, however, the planner optimally introduces jointness even under separable social weights.
Jointness and tail (in)dependence. Optimal jointness — whether one spouse’s marginal tax rate increases or decreases in the other’s earnings — depends critically on tail dependence of the joint productivity distribution, captured by the copula and survival copula elasticities. For right-tail dependent distributions (so that extremely productive individuals are likely to be matched with extremely productive partners), positive jointness is optimal at the top (raising taxes on high earners whose partners are also high earners) and negative at the bottom. For right-tail independent distributions (such as the Gaussian copula, which is tail-independent for any finite ρ), the distortion-reducing motive dominates: optimal jointness is negative at the top and positive at the bottom, conditional on standard convergence conditions.
Primary vs. secondary earners. The secondary earner (lower-productivity spouse) faces on average higher optimal distortions than the primary earner when the planner values redistribution to couples with a very unproductive spouse (α(w,0) ≥ 1), because the phasing out of transfers targeted to such couples generates high marginal tax rates on secondary earners. Family earnings-based taxation is optimal only when total family productivity and relative spousal productivity are independent, and when social weights are measurable only with respect to total family output.
Restricted taxation. Optimal distortions under any of the three restricted tax regimes (anonymous, separable, family earnings-based) exactly equal the relevant conditional average of unrestricted optimal distortions. This establishes that the welfare difference between the restricted and unrestricted optimum stems solely from the planner’s inability to tag taxes to individual productivity types within the restricted class.

Quantitative Findings (calibrated to 2020 CPS data on U.S. married couples, ages 25-65, worked ≥ 20 weeks). Spousal productivities are positively but not perfectly dependent, with Kendall’s tau = 0.21 and Pearson correlation = 0.25 for productivities (0.21 for earnings). The joint distribution is well approximated by a Gaussian copula (ρ = 0.33) with Pareto-lognormal marginals (a = 2.95, Gini = 0.31). The Gaussian copula is tail-independent, so consistent with analytical results, optimal jointness is positive for low earners and negative for high earners (the latter arising at earnings above approximately $8.5 million in the benchmark specification). The quantitative magnitude of optimal jointness is small — marginal taxes for one spouse change by at most several percentage points as a function of the other spouse’s earnings. Individual earnings-based taxation provides a good approximation to the unrestricted optimum. By contrast, family earnings-based (joint) taxation is a poor approximation in all specifications, with marginal taxes on family income varying substantially with the earnings share of the secondary earner, and this conclusion holds even when Pareto weights explicitly favor family earnings-based taxation (k = 0 case). The implied top marginal tax rate converges toward approximately 55 percent (corresponding to limiting distortion of ≈1.35 = 1/γa with γ = 0.25, a = 2.95) but the convergence is slow, so optimal marginal rates remain substantially below this limit even at earnings of $300,000.

In depth

Q1. What is the mechanism design formulation, and why is FOA validity a key concern in the bi-dimensional setting?

A: The planner’s problem is cast as a direct mechanism in which couples report their two-dimensional productivity type (w1, w2) and receive allocations (consumption, earnings). Incentive compatibility requires that no couple prefers to misreport. In one-dimensional models (Mirrlees 1971), restricting attention to local incentive constraints (the FOA) yields the standard ODE characterization of optimal taxes and is valid for a broad class of primitives. In two dimensions, solutions to multi-dimensional screening problems generically display “bunching” (Rochet-Choné 1998, Armstrong 1996), and the FOA may fail. The key difference exploited in this paper is the absence of participation constraints in the public finance setting, which eliminates the main force driving FOA failure in industrial organization models.

Q2. What are the necessary and sufficient conditions for FOA validity in the benchmark economy with independent types?

A: (Proposition 1) In the benchmark economy (symmetric, independent types, separable Pareto weights), FOA validity is equivalent to the condition that x·(1 + λ̃(x^{-γ})/2) is increasing in x, where λ̃(t) = [∫_t^∞ (1-α̃(w))g(w)dw] / (γtg(t)). The unidimensional analog requires x·(1 + λ̃(x^{-γ})) to be increasing. Since the bi-dimensional condition multiplies λ̃ by 1/2 rather than 1, the set of primitives satisfying it is strictly larger: every (G, α̃, γ) for which the unidimensional FOA holds also satisfies the bi-dimensional condition, but not vice versa. Economically, the FOA holds as long as the planner is not “too redistributive” — i.e., Pareto weights on low types are not so high as to violate these monotonicity conditions.

Q3. What is the Coarea Formula result (equation 27) and why is it the central technical tool?

A: Given that the optimality conditions form a PDE system that cannot generally be solved pointwise, the authors integrate the optimality condition (equation 20) over subsets of the type space defined by level sets of an arbitrary function Q(w1, w2). The Coarea Formula allows them to express the result as: E[Σ_i λ*_i γ_i (∂lnQ/∂lnw_i) | Q=t] = [1 − E[α|Q≥t]] / [−∂ln P(Q≥t)/∂ln t]. By choosing different Q functions (e.g., Q = w_i, Q = max{k_1 w_1, k_2 w_2}, Q = R(w) for total family productivity, Q = I(w) for relative productivity), the formula delivers closed-form expressions for distinct conditional averages of optimal distortions, all expressed in terms of exogenous primitives. This contrasts with variational approaches (Golosov et al. 2014, Spiritus et al. 2022) that express optimal taxes in terms of endogenous moments.

Q4. How do optimal distortions on married individuals compare to those on single individuals, and what is the exact quantitative relationship in the independent-types benchmark?

A: (Proposition 4) In the benchmark economy with independent types, the optimal distortion on spouse i with productivity t equals exactly one-half of the optimal distortion λ^{sng,}(t) in the corresponding unidimensional economy: λi(t, w{-i}) = (1/2)λ^{sng,*}(t), and this is independent of the partner’s productivity w_{-i}. The intuition: the deadweight cost of taxing any individual depends only on her own characteristics (elasticity, productivity, density), not on whom she is married to. However, the redistributive benefit of taxation depends on matching — when matching is random, every high-productivity individual is married on average to an average person, so the incremental social benefit of extracting tax revenue from her is exactly half of what it would be if she were single (since half the benefit goes to a partner who is already average). More generally (Proposition 5 and Corollary 2), average distortions are weakly lower for married individuals than for singles as long as matching is not perfectly positively assortative.

Q5. What is average jointness and how is it measured?

A: Average jointness J_i(t) is defined as the ratio of average distortions on spouse i conditional on the partner having above-t productivity to average distortions conditional on the partner having below-t productivity, minus one. Jointness is positive if the marginal tax rate on spouse i is on average increasing in the partner’s productivity, negative if decreasing, and zero for separable (individual earnings-based) taxes. The paper characterizes jointness through auxiliary functions H_i(t) (conditional distortion relative to unconditional average), whose behavior is determined by the copula elasticities η_i and survival copula elasticities η̄_i — the percentage change in the conditional quantile of the partner’s productivity when one spouse’s productivity quantile increases by 1%.

Q6. What is the role of tail dependence in determining the sign of optimal jointness?

A: (Proposition 7, Lemma 4) For right-tail dependent distributions — where the probability that an extremely productive person is married to an extremely productive partner remains bounded away from zero as productivity → ∞ — the redistributive benefit of positive jointness (targeting taxes to the richest couples) dominates its distortionary cost, so optimal average jointness is positive at the top. For right-tail independent distributions (where this probability converges to zero), the distortionary cost of positive jointness dominates, and optimal jointness is negative at the top. Exactly symmetric logic applies at the bottom using the survival copula and left-tail dependence. The bivariate lognormal/Gaussian copula is right-tail independent for any finite correlation ρ, while a distribution with perfect assortative matching in the tails would be right-tail dependent. The speed of convergence to tail independence, measured by κ = lim_{u→0} ln(u)/ln(C(u,u)) ∈ [1/2, 1), also matters: slower convergence (κ closer to 1) implies smaller optimal jointness under tail independence.

Q7. When is individual earnings-based (separable) taxation optimal, and when is family earnings-based taxation optimal?

A: (Propositions 4, 8, Corollary 1) Individual earnings-based taxation is optimal when Pareto weights are separable and spousal productivities are independent. When types are positively dependent, the planner introduces jointness even with separable social weights, because conditioning taxes on both spouses’ earnings facilitates redistribution across couple types. Family earnings-based taxation is optimal when: (i) social weights are measurable only with respect to total family productivity r (i.e., the planner cares only about total family output, not the identity or relative productivity of individual spouses), and (ii) total family productivity r and relative spousal productivity ι are statistically independent. When r and ι are not independent, even a planner with an intrinsic preference for family earnings-based taxation will find it optimal to depart from it.

Q8. What does Proposition 9 (Corollary 7) establish about the relationship between restricted and unrestricted optimal taxes?

A: (Corollary 7) For each restricted tax regime (anonymous, individual earnings-based, family earnings-based), the optimal distortions under the restricted tax equal the corresponding conditional average of unrestricted optimal distortions. Specifically: optimal individual earnings-based distortions equal E[λ*_i | w_i = t] (the average unrestricted distortion at productivity t); optimal family earnings-based distortions equal E[weighted average of λ*_i | R(w) = r]. This reveals that the unrestricted and restricted planners solve the same tradeoff between redistribution benefits and distortionary costs, but the restricted planner must apply a single tax rate to groups of couples that cannot be distinguished under the restriction. The welfare loss from restriction comes entirely from this forced bunching, not from a different objective or a different first-order condition.

Q9. What do the quantitative results say about the goodness of approximation of separable vs. family earnings-based taxation?

A: In the calibrated benchmark economy (Gaussian copula, ρ = 0.33, Pareto-lognormal marginals, γ = 0.25, m = 0.35), optimal jointness is quantitatively small — the marginal tax rate on one spouse changes by at most several percentage points as a function of the other spouse’s earnings over the plotted range. Individual earnings-based (separable) taxation therefore provides a good approximation to the unrestricted optimum across all specifications considered. By contrast, family earnings-based taxation is a poor approximation: the marginal tax rate on family income varies substantially with the earnings share of the secondary earner (the ratio min{y1,y2}/(y1+y2)), and the deviation from the optimal unrestricted tax is large. This finding is robust across different Pareto weight specifications (m ∈ {0.35, 1.5}, k ∈ {0, 1, 2}) and holds even when k = 0, i.e., when the planner’s social weights inherently prefer family earnings-based taxation.

Q10. How do the calibration results relate to the analytical comparative statics predictions?

A: The calibration validates the analytical predictions quantitatively. The analytical result (Proposition 5) that optimal distortions in the U.S. lie between those under random matching (1/2 of single-individual rates) and perfect assortative matching (same as single-individual rates) is confirmed: optimal tax rates for married individuals in the calibrated economy lie between the independence and perfect-dependence gray-line benchmarks in Figure 6. The analytical prediction (Proposition 7) that the Gaussian copula implies positive jointness at the bottom and negative at the top is confirmed, with the switch to negative jointness occurring above approximately $8.5 million in earnings. The slow convergence of the Gaussian copula to tail independence (κ = (1+ρ)/2 ≈ 0.665) explains the small magnitude of optimal jointness relative to the FGM copula (which has κ = 1/2, faster convergence, and exhibits more pronounced jointness as shown in the appendix). The analytical limiting distortion of E[λ*_i | w_i = t] → 1/(γa) ≈ 1.35 as t → ∞ (corresponding to a top marginal tax rate of approximately 55 percent) is confirmed, though convergence is slow and rates remain substantially below this limit at $300,000 in earnings.

Q11. How does the paper relate to and advance beyond Kleven, Kreiner, and Saez (2007/2009)?

A: Kleven et al. (2009) studied couples taxation but avoided the multi-dimensional screening complexity by restricting the secondary earner to binary labor supply. The working paper by Kleven et al. (2007) considered the continuous setting but noted the difficulty of the FOA and derived several special-case insights. The current paper extends KKS in several systematic ways: it provides the first formal proof that the FOA conditions are strictly weaker in bi-dimensional than unidimensional settings; generalizes the formula for average distortions to arbitrary joint distributions (not just independent types); characterizes optimal jointness under positive dependence (not just independence); establishes the role of tail (in)dependence in determining the sign of jointness; compares optimal taxes for married vs. single individuals; and derives conditions under which family earnings-based or individual earnings-based taxation is optimal. It also shows that the KKS result on jointness sign (determined by the third derivative of the SWF) applies only under independence and can be reversed even with arbitrarily small positive dependence, as demonstrated with the Gaussian copula example.

Key Concepts

First-Order Approach (FOA) in multi-dimensional taxation. The restriction of the mechanism design problem to local incentive constraints only — dropping global (non-local) incentive compatibility conditions and solving a relaxed problem. In the paper’s context, FOA validity is equivalent to convexity of a specific transformation vx* of the optimal utility function in the “linearized” type space X. The paper shows that the condition for FOA validity is strictly weaker (i.e., a strictly larger set of primitives satisfies it) in the bi-dimensional couples setting than in the corresponding unidimensional model, because the absence of participation constraints eliminates the main force driving FOA failure in industrial organization multi-dimensional screening.

Optimal tax distortion λ_i(w).* The monotone transformation of the marginal tax rate defined by λ_i(w) = [∇_i T(y(w))] / [1 − ∇_i T(y(w))], where ∇_i T is the partial derivative of the tax function with respect to spouse i’s earnings. This transformation maps [−∞, ∞] marginal tax rates to (−1, ∞) distortions. The optimal tax schedule is characterized by the function λ* satisfying a system of PDEs; the paper studies conditional averages of λ* rather than λ* pointwise.

Coarea Formula. A mathematical result from geometric measure theory that, in this context, converts an integral of the PDE optimality condition over a two-dimensional domain into an integral over the level sets of an arbitrary function Q(w). Applied to equation (20), it yields: E[Σ_i λ*_i γ_i (∂lnQ/∂lnw_i) | Q=t] = [1 − E[α|Q≥t]] / [−∂ln P(Q≥t)/∂ln t]. By choosing different Q functions, the formula delivers conditional averages of optimal distortions over different subsets of the type space, all in terms of exogenous primitives. This is the paper’s principal analytical tool for characterizing optimal taxes without solving the PDE explicitly.

Jointness (positive/negative). The dependence of the optimal marginal tax rate on one spouse’s earnings on the other spouse’s earnings. Taxes are positively jointed at w if ∂²T/∂y_1∂y_2 > 0 (so raising one spouse’s earnings increases the marginal tax rate on the other); negatively jointed if this cross-partial is negative; disjointed (separable) if it is zero. Average jointness J_i(t) at productivity t is measured as the ratio of conditional average distortions above and below the partner’s productivity threshold, minus one. Optimal jointness is the paper’s primary policy object for understanding how taxes on one spouse should respond to the other’s earnings.

Copula and survival copula elasticities (η_i, η̄_i). Defined as η_i(t) = ∂ln C(u)/∂ln u_i and η̄_i(t) = ∂ln C̄(u)/∂ln ū_i, where C is the copula of the joint productivity distribution, C̄ is the survival copula, and u_i = G_i(t_i), ū_i = 1−G_i(t_i) are the corresponding quantiles. These elasticities measure the percentage change in the conditional quantile of the partner’s productivity when one spouse’s productivity quantile increases by 1%. They quantify the additional distortionary cost introduced by jointness relative to a separable tax schedule: smaller elasticities (stronger dependence) correspond to larger distortionary costs of jointness at the boundaries of probability mass.

Tail (in)dependence. A joint distribution F is right-tail dependent if lim_{t→∞} P(w_{-i}≥t | w_i≥t) > 0, i.e., extremely productive individuals have a positive probability of being matched with equally extreme partners. It is right-tail independent if this limit is zero. The speed of convergence to tail independence is measured by κ = lim_{u→0} ln(u)/ln(C(u,u)) ∈ [1/2, 1). Tail dependence determines the sign of optimal average jointness in the tails: right-tail dependence favors positive jointness at the top; right-tail independence favors negative jointness at the top. The Gaussian copula is right-tail independent for any finite ρ; a perfectly assortative matching distribution is right-tail dependent.

Positive quadrant dependence (PQD) order. A partial ordering on joint distributions with the same marginals: F^b ≥_{PQD} F^a if F^b(w) ≥ F^a(w) for all w, equivalently if Cov(φ_1(w_1), φ_2(w_2)) ≥ 0 for any two increasing functions. The paper uses this order to rank economies by the “assortativeness” of matching, and shows that optimal average distortions are monotone in this order (Proposition 5): more assortative matching implies weakly higher optimal tax distortions on each married individual.

Pareto-lognormal (PLN) distribution. Used in the calibration to model the marginal distribution of spousal productivities. Defined as G(t) = Φ((ln t − μ)/σ) − a·exp(aμ + a²σ²/2)·Φ((ln t − μ)/σ − aσ), parameterized by location μ, scale σ, and tail parameter a. The PLN family has a lognormal body and a Pareto tail with tail parameter a, making it suitable for capturing the empirical finding of a thin left tail (implying optimal marginal taxes approaching zero as earnings → 0) and a thick right tail (implying a positive limiting marginal tax rate of approximately 1/(1 + 1/(γa)) as earnings → ∞).

The Power of Proximity to Coworkers

Mon, 01 Jan 0001 00:00:00 +0000

This paper studies how physical proximity to coworkers affects on-the-job training and productivity, using software engineers at a Fortune 500 online retailer observed from 2019 to 2024. The authors exploit two quasi-experimental shocks to proximity: the office closures of 2020, which eliminated proximity differentials that previously existed across team types, and the firm’s subsequent return-to-office (RTO) mandates in 2022 and 2023, which restored proximity for co-located teams while leaving geographically-distributed teams apart. The core identification strategy is a difference-in-differences design comparing engineers whose teams were co-located in a single headquarters building to those whose teams were split across two buildings a ten-minute walk apart — a distinction that became immaterial once offices closed.

The central finding is that sitting near teammates substantially increases the digital feedback engineers receive on their code. Before the office closures, engineers on co-located teams received 23.9% (1.92 comments per program) more code review feedback than engineers on multi-building teams. Once offices closed, this advantage narrowed by 18.3% (1.47 comments per program, p-value = 0.0026). The lost comments were disproportionately those predicted by a machine-learning classifier to be helpful, actionable, well-reasoned, and impactful, with high-quality comments declining by 21–23% — exceeding the overall volume decline. Face-to-face and digital communication are complements, not substitutes: proximate engineers drew on a wider pool of reviewers and asked 48.4% more follow-up questions, a differential that vanished once offices closed.

Proximity’s effects are highly heterogeneous. Gains in feedback are concentrated among less-tenured, younger, and female engineers — those with the most to learn. Junior engineers on co-located teams lost 2.03 more comments per program upon office closure than junior engineers already on distributed teams (p-value = 0.001); young engineers lost 2.47 more comments (p-value = 0.0001). Female engineers lost 38.9% more comments than their distributed female counterparts (p-value < 0.0001), partly because women stop asking as many people for feedback when they cannot do so in person.

Proximity improves code quality for inexperienced engineers. Around the second RTO (three days per week), engineers on co-located teams became 2.2 percentage points less likely to add files subsequently deleted — a measure of churn — and 1.4 pp less likely to introduce bugs, relative to distributed teams (p-values of 0.041 and 0.022 respectively). These gains were roughly twice as large for less-tenured and younger engineers. The benefits persist: engineers who spent more pre-closure time on co-located teams continued to write higher-quality code during the fully remote period.

However, mentorship is costly for those who provide it. Senior engineers on co-located teams wrote 0.76 fewer programs per month in the main codebase before closures (p-value = 0.0005), a gap that closed when offices did and widened again during the second RTO. The firm faces a fundamental tradeoff: proximity accelerates junior engineers’ human capital development while reducing experienced engineers’ immediate coding output.

These dynamics shape hiring. The firm shifted toward hiring older, more experienced engineers during closures — buying talent it could no longer build in-house — and back toward younger hires once offices reopened. Nationally, young college graduates in remotable occupations (classified per Dingel and Neiman, 2020) experienced a 0.88 pp increase in unemployment between 2017–2019 and 2022–2024, while older graduates saw a marginal decline of 0.11 pp. A triple-difference estimate finds a 0.65 pp greater increase in young workers’ unemployment in remotable versus non-remotable occupations (p-value = 0.029), a pattern that predates generative AI diffusion and is robust to controlling for AI exposure. Back-of-the-envelope, remote work accounts for an estimated 64% of the total unemployment increase among young college graduates over this period.

The paper also documents that proximity is fragile: a ten-minute walk between two buildings reduces feedback as much as being multiple states away, and even a single distant teammate imposes negative externalities on those who remain co-located, reducing their feedback by 1.71 comments per program (p-value = 0.095) via a “one Zoom, all Zoom” norm.

Q: What is the main identification strategy for the office-closure analysis, and what is the key parallel-trends evidence?

A: The authors compare engineers on co-located teams (all members in one headquarters building) to those on multi-building teams (split across two buildings a ten-minute walk apart), before and after the March 2020 office closures. Co-located teams lost more proximity when offices closed, while multi-building teams experienced a smaller shock, enabling a difference-in-differences design. Pre-closure trends in feedback are parallel across the two team types (Figure I), supporting the identifying assumption. Standard errors are clustered by team, the unit of treatment assignment.

Q: How large is the effect of proximity on total code review feedback, and how is it broken down by feedback source?

A: Before closure, co-located engineers received 23.9% (1.92 comments per program) more feedback than multi-building engineers. The DiD estimate indicates that losing proximity reduced feedback by 18.3% (1.47 comments per program, p-value = 0.0026, Column 3 of Table II). This decline stems entirely from reduced feedback from teammates; there is no detectable effect on feedback from engineers on other teams — a placebo check that supports the identification strategy and rules out explanations based on differential project complexity.

Q: How does proximity affect the quality — not just the quantity — of code review comments?

A: Using a gradient-boosted decision tree trained on 5,377 human-labeled comments, the authors predict comment quality across all 174,014 comments. Losing proximity reduced comments predicted to be helpful, well-reasoned, actionable, and likely to change the code by 21–23% — exceeding the 18.3% overall volume decline. The residual comments were lower quality: 2.9 pp fewer were helpful (p-value = 0.039), 1.7 pp fewer explained their reasoning (p-value = 0.094), and 1.9 pp fewer were likely to change the code (p-value = 0.072).

Q: What mechanisms drive the complementarity between face-to-face interaction and digital feedback?

A: Proximity increases feedback on both the extensive and intensive margins. On the extensive margin, co-located engineers draw on a wider pool of reviewers, returning less frequently to the same commenter. On the intensive margin, losing proximity reduces follow-up questions by 48.4% (0.12 questions per program, p-value = 0.0083), accounting for roughly half of the total feedback decline. The other half comes from reduced initial reviewer feedback. References to other communication channels (e.g., Slack) within code reviews also decline when proximity is lost, confirming that face-to-face and digital communication are complements.

Q: How small a physical barrier is sufficient to reduce feedback substantially?

A: A ten-minute walk between two buildings on the same headquarters campus reduces feedback by as much as being multiple states away — both groups receive significantly less feedback than engineers whose entire team sits in the same building (Figure Ib). This finding aligns with research on academics showing that different floors or buildings reduce coauthorship, and extends it to daily teammates sharing projects.

Q: What are the externality effects of a single distant teammate?

A: Through the firm’s implicit “one Zoom, all Zoom” norm, even one teammate in a different location shifts all team meetings to video calls. Engineers in the same building exchange 14.5% less feedback when even one teammate is in another building versus when all teammates are co-located (p-value = 0.037). When a new hire transforms a co-located team into a multi-building one, feedback between the original co-located teammates drops by 1.71 comments per program (p-value = 0.095); adding a new co-located hire produces no such decline.

Q: How does the effect of proximity on feedback differ by engineer tenure, age, and gender?

A: Less-tenured engineers on co-located teams lost 2.03 more comments per program upon closure than less-tenured engineers on distributed teams (p-value = 0.001). Young engineers (under 29) on co-located teams lost 2.47 more comments per program than young distributed engineers (p-value = 0.0001). Female engineers on co-located teams lost 38.9% (3.71) more comments than female engineers on distributed teams (p-value < 0.0001), partly because women draw feedback from 14.7% fewer people when proximity is lost (p-value = 0.0078), compared to a negligible 2.6% decline for men. The extra feedback women receive in person is of higher quality, not rude or condescending.

Q: How is the effect of proximity on code quality identified using the RTO design, and what are the magnitudes?

A: The RTO design compares engineers on co-located (same-city) teams to geographically-distributed teams across three periods: full closure, first RTO (two days per week), and second RTO (three days per week). The authors predict γ_closed ≈ 0 (office assignment irrelevant when closed) and γ_2nd_RTO > γ_1st_RTO (more in-office days means more proximity). Both predictions are confirmed. During the second RTO, co-located engineers were 2.2 pp less likely to add files later deleted (p-value = 0.041) and 1.4 pp less likely to introduce bugs (p-value = 0.022), with effects roughly twice as large for less-tenured and younger engineers.

Q: Does the benefit of co-location on code quality persist after remote work resumes?

A: Yes. After all engineers returned to remote work, those who had been on co-located teams pre-closure were 2.37 pp less likely to write disposable code (p-value = 0.013) and 3.09 pp less likely to introduce bugs (p-value = 0.0012). Code quality improves monotonically with the number of pre-closure months spent on co-located teams (Figure A.5). These gaps persist when including current team fixed effects, meaning within the same post-closure team, the previously co-located engineer writes higher-quality code.

Q: What is the cost of mentorship for senior engineers, and how does it manifest in coding output?

A: Senior engineers on co-located teams wrote 0.76 fewer programs per month in the main codebase when offices were open (p-value = 0.0005). Once offices closed, this gap disappeared, and senior engineers who lost proximity to their teammates saw a relative increase in output of 0.58 programs per month (p-value = 0.0014). During the second RTO, engineers with more than sixteen months of tenure on co-located teams wrote fewer programs, while no significant difference emerged for less-tenured engineers. Overall, the DiD estimate indicates losing proximity to teammates increases immediate output by 0.48 programs per month (p-value = 0.0002).

Q: How does the firm’s hiring age distribution respond to changes in proximity?

A: When offices were closed, the firm shifted toward hiring older engineers: the share of hires under age 29 fell from over half pre-closure to less than a third during the closure. After the RTOs, the firm shifted back toward younger hires. Geographic variation reinforces this: headquarters-campus hires were 7–10 years younger than those hired into distributed roles when offices were open; this gap narrowed substantially during closures when everyone was far from teammates.

Q: Does proximity affect which engineers are poached by other firms?

A: Yes. During the office closures, 1.2% of co-located engineers were poached per month, compared to 0.9% of multi-building engineers of similar tenure, age, and engineering group (p-value = 0.044). By the end of the closure period, nearly a quarter of co-located engineers had been poached versus a sixth of multi-building engineers. There is a dose response: more pre-closure time on co-located teams predicts higher poaching rates. The effect is concentrated among younger and female engineers, consistent with their feedback building more transferable general human capital. Tenure does not moderate the poaching effect, consistent with less-tenured engineers’ feedback being more firm-specific.

Q: What does national unemployment data show about the scarring effects of remote work on young workers?

A: Between 2017–2019 and 2022–2024, young college graduates (under 29) in remotable occupations experienced a 0.88 pp increase in unemployment (p-value < 0.00001), while older graduates in the same occupations saw a marginal decline of 0.11 pp (p-value = 0.053). A triple-difference regression finds a 0.65 pp greater increase in young workers’ unemployment in remotable versus non-remotable occupations (p-value = 0.029). Back-of-the-envelope, scaling this estimate by the 61% share of young graduates in remotable jobs predicts a 0.4 pp increase in young college graduates’ overall unemployment — equal to 64% of the realized 0.63 pp increase.

Q: Is the unemployment increase among young workers in remotable jobs driven by generative AI rather than remote work?

A: The authors argue against AI as the primary driver on two grounds. First, the uptick in young workers’ unemployment in remotable occupations predates the rapid diffusion of generative AI. Second, the differential increase is not concentrated among occupations with the highest AI task exposure. The triple-difference estimate is robust to controlling for occupational AI exposure using the Eisfeldt, Schubert and Zhang (2023) index. The authors acknowledge that AI may become more important as it diffuses further.

Q: How do young workers’ own office attendance decisions reflect the value of proximity?

A: At the partner firm, engineers under 29 were 8.8 pp (37.6%) more likely to come into the office during the RTOs than older engineers when on co-located teams (solid line in Figure VIIa). This difference was roughly halved on geographically-distributed teams (p-value of difference = 0.0085), indicating that the draw is specifically proximity to teammates. Co-located managers raised attendance by 2.6 pp, while co-located teammates raised it by 5.1 pp. Nationally, Stack Overflow survey data show nearly half of engineers under 25 are in the office each day, versus a quarter of older engineers (p-value < 0.00001).

Q: What does the paper imply about why remote work was rare before the pandemic despite workers’ stated preferences for it?

A: The paper offers a resolution: firms may have recognized that the value of the office lies in training for tomorrow and improving the quality — not the quantity — of work today. Remote work boosts immediate output, especially for experienced workers, but it reduces mentorship and long-run skill development. The tradeoff between current and future productivity, and between individual and collective returns to human capital, explains why firms historically resisted remote work even when workers preferred it and short-run output was unaffected.

Q: What are the implications for gender equity in remote work?

A: The findings suggest remote work has ambiguous gender effects. While remote work may help working mothers remain in the workforce, it appears costly for young women’s professional development, which is especially sensitive to physical proximity. Women receive substantially more high-quality feedback when co-located, draw feedback from a wider network in person, and lose disproportionately more feedback when proximity is lost. Young female engineers on co-located teams were also disproportionately poached — suggesting their human capital gains from co-location are more general and transferable.

Code review feedback: The digital comments engineers exchange when reviewing each other’s code before it is merged into the live codebase; the paper’s primary measure of on-the-job training and mentorship investment, distinct from mere volume because the authors also classify comments by helpfulness, reasoning, actionability, and expected impact using supervised machine learning.

Co-located team: A team in which all members are assigned to the same office building; the treatment group in the difference-in-differences designs, distinguished from multi-building teams (split across two headquarters buildings, a ten-minute walk apart) and geographically-distributed teams (members in different cities or permanently remote).

One Zoom, all Zoom norm: The implicit team practice of holding all meetings virtually if any single teammate cannot be physically present; the mechanism by which one distant colleague generates negative externalities for the remaining co-located teammates, reducing their in-person interaction and feedback.

Proximity fragility: The finding that even small physical barriers — a ten-minute walk between buildings — reduce feedback as much as being multiple states away, implying that the relationship between physical distance and mentorship is highly nonlinear near zero.

Churn (disposable code): Files that are added by an engineer but deleted within the subsequent six months, either because the code was poorly structured or because it introduced a feature later abandoned; used as one of two code quality proxies in the RTO analysis (occurring in 15% of programs).

Bugs (immediate reversions): Programs that are immediately and fully reverted after being merged, typically indicating the engineer’s changes precipitated an emergency requiring rollback to an earlier version; used as the more serious of the two code quality proxies (occurring in 3.5% of programs).

Scarring effects: The persistent adverse impact on young workers’ human capital and labor market outcomes from reduced mentorship during the remote work period; manifested both as lower code quality at the individual level and higher unemployment rates nationally among young college graduates in remotable occupations.

Remotable occupation: An occupation classified by Dingel and Neiman (2020) as feasibly performed from home; used to construct the national triple-difference analysis comparing age gaps in unemployment across remotable and non-remotable jobs before and after the pandemic.

The Price of Housing in the United States, 1890–2006

Mon, 01 Jan 0001 00:00:00 +0000

Lyons, Shertzer, Gray, and Agorastos construct the first consistent, annual, quality-adjusted market rent and home sales price series for American cities spanning 1890–2006. The paper addresses a fundamental data gap: no annual city-level series existed for market rents at any point in the 20th century, and no annual city-level sales price series existed prior to 1975. Existing national series—the BLS Rent of Primary Residence (RoPR) for rents and the Shiller index for sales—carry well-documented methodological limitations that the authors argue have produced materially misleading stylized facts about long-run U.S. housing markets.

The Historical Housing Prices (HHP) dataset draws on just under 2.7 million newspaper real estate listings from 30 U.S. cities across 1890–2006. Listings must contain a price, a size measure (rooms or bedrooms), property type (house or apartment), and a location indicator. The authors construct hedonic price indices using a rolling-windows methodology—baseline three-year rolling windows with annual step size—that controls for size, type, and standardized within-city location, allowing coefficients to vary over time rather than imposing a fixed vector across the full century. City-level indices are aggregated to national indices using population weights from census data interpolated between census years. Listed prices serve as proxies for transaction prices; the authors validate these against census distributions and against post-1975 FHFA and Case-Shiller series.

The paper’s findings revise several established stylized facts. First, real market rents did not fall over the 20th century as implied by the RoPR series. Instead, real rental price levels were approximately 20% higher in 2006 than in 1890, fluctuating within a relatively narrow band. The RoPR series, by contrast, implies a near-halving of real rents between 1914 and 2006. Second, the paper documents a substantial interwar housing boom-bust absent from the Shiller index: real sales prices rose approximately 47% between 1920 and 1928, then fell 27% by 1935, with the 1928 peak not recovered in real terms until 1968. Third, contrary to the Shiller index’s depiction of minimal housing price growth from 1950 to 1995, the HHP series shows real sales prices rising 21% between 1953 and 1974—a period for which Shiller relies on a truncated sample of government-backed mortgages that excluded higher-valued homes.

On the return to homeownership, the paper finds average nominal housing returns across 1890–2006 of approximately 11% per year, composed of 3.8% capital gain and 7.2% rental return. Gross market rental yields exceeded 8% annually for much of 1900–1945, fell to 7% by 1960, and to 3% by 2006. Capital gains were largely unimportant before the 1940s and became the dominant return component only from 1970 onward; the post-1980 period with sustained capital gains is characterized as historically anomalous. Returns varied substantially across cities, with some cities outperforming the S&P 500 in the prewar era while most underperformed equities from 1981–2006.

The paper also examines implications for the CPI. The HHP series implies nominal rents grew at approximately 3.5% per year from 1914 to 2006, versus 2.6% per year for the RoPR component. A back-of-the-envelope alternative CPI using HHP rental data yields overall price growth of 3.3% per year rather than the official 3.1%, suggesting the measured increase in U.S. living standards since World War I may be modestly overstated. Finally, cross-city analysis shows that land constraints and, increasingly, regulatory constraints explain divergence in price growth across cities, with the role of zoning becoming more pronounced after 1980.

Q: What is the core data source and how are the indices constructed? A: The HHP dataset comprises just under 2.7 million newspaper real estate listings from 30 U.S. cities, 1890–2006, sampled from real estate sections (typically the last Sunday of each month). Valid listings require price, size, property type, and within-city location. Hedonic indices are estimated using rolling three-year windows with annual steps, controlling for size, type, and standardized location, allowing hedonic coefficients to evolve over time rather than imposing a fixed vector. City indices are aggregated to national indices using population-weighted census data interpolated between census years.

Q: Why are the HHP series based on listing prices rather than transaction prices, and how is this limitation addressed? A: Transaction-price records require local archival effort infeasible across 30 cities over 116 years, and rental transaction data are essentially unavailable historically. The authors argue that hedonic mix-adjustment makes listed prices strong predictors of selling prices during normal market conditions, and that a substantial share of houses transact at their exact listing price. Validation against census distributions and against post-1975 FHFA and Case-Shiller series supports the approach; the authors acknowledge listing prices may diverge from transaction prices at cyclical peaks and troughs.

Q: What does the paper find about the long-run trajectory of real market rents, and how does this revise existing understanding? A: The HHP series shows real rental price levels in 2006 were approximately 20% higher than in 1890 or 1914, fluctuating within a relatively narrow band over the century. The BLS RoPR series implies real rents fell by nearly half between 1914 and 2006. The HHP findings align with the most influential proposed corrections to the RoPR by Gordon & van Goethem (2007) for 1915–1939 and broadly with Crone et al. (2010) in terms of overall growth levels for 1940–1995, though the HHP series shows a sharper rental spike after World War II rent controls were lifted that the BLS methodology captures only with deliberate lag.

Q: What does the paper find about the interwar housing cycle, and why does the Shiller index miss it? A: The HHP series documents that real sales prices rose approximately 47% between 1920 and 1928, then fell 27% by 1935, with the 1928 nominal peak not regained until 1946 and the real peak not until 1968. The Shiller index for 1890–1934 is based on a 1934 survey of owner recollections of past transaction prices and assessed values, which the authors argue reflects homeowners’ lack of awareness of the changing value of their homes over prior decades. The HHP finding is consistent with census data, Nicholas & Scherbina’s study of New York City, and Fishback & Kollmann’s analysis of New Deal reports.

Q: What does the paper find about the 1953–1974 period, and what explains the divergence from the Shiller index? A: The HHP series shows housing sales prices increased 21% in real terms between 1953 and 1974, while the Shiller index (based on the Home Purchase Component of the CPI) implies a moderate decline of around 10%. The Shiller index for this period uses a truncated sample of government-backed mortgages subject to FHA loan limits; when the authors truncate their own data using the same statutory FHA limits ($30,000 in 1973, $45,000 in 1974, $60,000 in 1977), approximately 50% of their 1971–1979 listings are excluded and their truncated series matches the Shiller index more closely. This supports the Greenlees (1982) critique of downward bias in the Home Purchase CPI component.

Q: What are the long-run return components to homeownership at the national level? A: Average nominal housing returns across 1890–2006 were approximately 11% per year: 3.8% capital gain and 7.2% rental return. Before World War II (1890–1945), average nominal rental returns ranged from 7.9% to 8.3% per sub-period while capital gains averaged near zero or negative in real terms. Only in 1981–2006 did capital gains (averaging 5.8%) exceed the rental return (averaging 5.3%). The return to housing has thus been dominated by rental income over the long run, with the post-1980 era of sustained capital gains constituting a historical anomaly.

Q: How do rental yields evolve over the sample period? A: Gross market rental yields exceeded 8% annually for much of 1900–1945, with spikes after both World Wars and a dramatic fall from nearly 11% to below 7% during the early 1920s boom, consistent with a bubble dynamic before the Great Depression. Yields fell to approximately 7% by 1960 and to 3% by 2006. City-level heterogeneity was substantial: rental returns exceeded 15% in some cities in the two decades before the Great Depression, and most cities saw returns above 10% nominally during 1930–1945, while even by 1981–2006 cities like Phoenix and St. Louis averaged above 12%.

Q: What does the paper find about housing and the business cycle? A: Real growth rates in GDP and housing prices moved in the same direction in 72 of 116 years for sales prices and 65 of 116 years for rental prices. The paper identifies three major downturns where falling rents led falling prices which led falling GDP: the Great Depression (rents fell from 1924, prices from 1929, GDP from 1930), the early 1990s recession (rents from 1988, prices from 1990, GDP from 1991), and the end-of-sample period (rents from 2002). Only after World War I (1920–21) and World War II (1945–46) did clear economic contractions occur without equivalent housing price downturns.

Q: What does the paper find about cross-city variation in housing returns, and what does this imply for the volatility puzzle? A: Capital gains and rental returns vary substantially across cities and time periods; some cities saw returns exceeding the S&P 500 before World War II (including New York and Chicago), while most underperformed equities from 1981–2006. The authors argue that the apparently low volatility of housing returns at the national level documented by Jordà et al. (2019) is partly an aggregation artifact: local housing markets with very different trajectories are combined into a national index, dampening measured variance. The mild positive correlation between city-level capital gains and rental returns has an R² of 0.24.

Q: What are the implications for CPI measurement? A: The HHP series implies nominal rents grew at approximately 3.5% per year from 1914 to 2006, compared with 2.6% per year for the BLS RoPR component, with higher growth concentrated in the years after both World Wars and in the 1965–1985 period. A back-of-the-envelope alternative CPI substituting HHP rental data yields overall price growth of 3.3% per year rather than the official 3.1%. If rental price growth before 1985 is understated in the BLS data, then there has been less improvement in the U.S. standard of living since World War I than was previously understood.

Q: What does the paper find about the role of supply constraints in explaining cross-city price divergence? A: Natural land constraints are positively linked to price growth throughout the 20th century, with the relationship sharpest during 1930–1945 (before the postwar suburban expansion) and again after 1980. Regulatory constraints—measured at the turn of the millennium—have become an increasingly important driver of cross-city price differences, consistent with zoning functioning as a tax (Gyourko & Krimmel 2021). The paper also finds evidence suggesting land-use regulations are partly driven by expectations of future price growth, consistent with the homeowner-voter hypothesis (Fischel 2015; Trounstine 2018).

Q: How does the paper validate its series against existing sources? A: The HHP rental series aligns closely with the Rees and Jacobs (1961) series for 1890–1914. For sales, the HHP series matches the Case-Shiller-Weiss and FHFA repeat-sales indices at both national and city level after 1990 despite methodological differences. The paper finds approximately 25% more price growth than the CSW series over 1975–2006 (117% versus 90% in the 30 HHP cities), attributing some of the divergence to OFHEO appraisal-based valuations before 1992 and the HHP coverage of the broader owned housing market beyond single-family homes.

Historical Housing Prices (HHP) Project: A dataset of just under 2.7 million newspaper real estate listings from 30 U.S. cities, 1890–2006, used to construct annual, quality-adjusted hedonic price indices for both rented and owned housing segments at the city and national level.
Rolling-windows hedonic methodology: An index construction approach that runs sequential hedonic regressions over two-, three-, or five-year overlapping windows with annual step size, allowing the coefficients on size, type, and location to evolve over time rather than imposing a fixed vector across the full sample period, reducing bias from unobserved quality changes.
Market rent vs. contract rent: Market rent (the listing price for a rental unit actively advertised) is conceptually distinct from contract rent (the rent paid by tenants currently in situ), which is what the BLS RoPR series measures. Market rents adjust to vacancy and lease resets faster than contract rents, producing substantially more short-run volatility and a materially different long-run trend.
Gross rental yield (rent-to-price ratio): Annual rental income from a property divided by its market sales price, computed as RI_{c,t} / HPI_{c,t}. Gross yields exceeded 8% annually for much of 1900–1945 and fell to 3% by 2006 nationally, making rental income the dominant component of total housing returns for most of the century.
Total return to housing: The sum of the capital gain (percentage change in sales price) and the rental return (rental income divided by sales price), computed at annual, city, and national frequency for 1890–2006. The average nominal total return was approximately 11% per year, with 3.8% from capital gains and 7.2% from rental income.
Rent of Primary Residence (RoPR): The BLS survey-based series measuring changes in contract rents for a rotating panel of rental units, used as the shelter component of the CPI. The HHP series implies this series understates rental price growth by approximately 0.9 percentage points per year (3.5% vs. 2.6% nominal growth), concentrated in post-World War periods and 1965–1985, due to tenant non-response bias and delayed incorporation of new construction.
Supply constraints and cross-city divergence: Natural land constraints (geographic barriers to development) and regulatory constraints (zoning and land-use regulation) that limit housing supply, both positively associated with price growth, with regulatory constraints becoming increasingly important after 1980 and consistent with the hypothesis that land-use regulations are partly driven by homeowner expectations of future price appreciation.

Traditional Institutions in Modern Times: Dowries as Pensions When Sons Migrate

Mon, 01 Jan 0001 00:00:00 +0000

This paper asks whether dowry — a transfer from the bride’s family to the groom’s household upon marriage, prevalent throughout India — enables male migration by providing liquidity that compensates parents for the old-age support they would otherwise lose when sons leave the village. The core friction is that in patrilocal societies, sons traditionally co-reside with parents and share income in old age; migration disrupts this arrangement and introduces income-sharing frictions (limited commitment, information asymmetries, remittance costs). Dowry attenuates this friction by providing a liquid pool of resources at the time of marriage that the son can transfer to parents, lowering the net return to migration needed for a household to find migration optimal.

The authors develop a collective household model in which parents and sons jointly maximize a Pareto-weighted utility function. The model yields six testable predictions: (1) net marriage transfers can flow in either direction; (2) parents are more likely to take from the dowry when sons migrate; (3) conditional on migration, the probability of parental taking increases in the son’s income and in parental bargaining power; (4) aggregate male migration rates are higher in districts with stronger historical dowry traditions; (5) migration responses to a reduction in migration costs are larger in dowry areas, provided migration rates are relatively low; and (6) parents who receive remittances from migrant sons are more likely to have also taken from the dowry.

To test predictions 1–3 and the remittance auxiliary prediction, the authors collected two original datasets: a Destination Survey of 557 prime-age men in Gurugram (near Delhi) conducted in 2018, of whom 62% were migrants; and an Origin Survey of 2,541 households across 34 districts in six North Indian states conducted in 2020, covering 3,069 sons, 20% of whom were migrants. These are the first quantitative data on property rights over dowry in India. Across the Destination and Origin surveys, 45% and 27% of grooms’ parents, respectively, took from the dowry on net. Parents of migrants are 27 percentage points (Destination) and 8 percentage points (Origin) more likely to take than parents of non-migrants. For migrant sons, a doubling of the son’s occupational score raises the likelihood of parental taking by 19 percentage points; no such relationship exists for non-migrants. When sons report that parents held veto power over the marriage — a proxy for parental Pareto weight — parents of migrant sons are 28 percentage points more likely to be net takers. Parents whose migrant son sends financial remittances are 17 percentage points more likely to have taken from the dowry (coefficient 0.168, SE 0.074).

To test predictions 4 and 5, the authors use the Ancestral Characteristics data (Giuliano and Nunn 2018) to construct district-level measures of dowry tradition strength, validated against 1999 REDS and IHDS survey data, where a one-unit increase in the historical dowry measure is associated with 81–109% higher gross or net dowry payments. Using the NSS Round 64 migration module (2007–08), they find that the continuous dowry tradition measure is associated with a 2.7–3.7 percentage point increase in migration probability against a mean of 23.8%. For the highway construction identification strategy, the authors exploit the staggered rollout of the Golden Quadrilateral and North-South/East-West corridor (5,846+ km, $71 billion), using modern staggered-entry difference-in-differences estimators (Borusyak et al. 2021; Callaway and Sant’Anna 2020). Young men (ages 15–30) in dowry districts exhibit a large, significant increase in out-migration following highway construction with no pre-trends, while the effect for non-dowry males is indistinguishable from zero. Older males (ages 31–45) show no such effect in either group, consistent with the mechanism operating at marriage. The highway effects are concentrated in inter-district, employment-driven migration.

Scope conditions: the migration-enabling mechanism operates through marriage-age liquidity and patrilocal support norms; results are specific to male migration in India. The model assumes parents and sons act collectively, matching is based on grooms’ earning potential, and migration frictions cause income-sharing transfers to be infeasible when the son migrates.

Q: What is the central hypothesis of the paper? A: The hypothesis is that dowry, by providing a liquid transfer at the time of marriage, allows sons to compensate parents for the old-age support that would otherwise be lost when sons migrate. Because migration introduces frictions that prevent optimal post-migration income sharing between parents and sons, dowry lowers the minimum net return to migration required for the household to find migration optimal, thereby enabling more migration.

Q: What is the “Seeking” versus “Satisfied” distinction in the model, and why does it matter? A: “Satisfied” parents are those whose own income plus the maximum feasible marriage transfer (bounded by the bride’s endowment dE when dowry is present) is at least as large as their consumption allocation under no migration; migration then Pareto-improves the household for any non-negative return R. “Seeking” parents have insufficient income plus endowment, so migration reduces their consumption unless the son’s return R exceeds a threshold B(d). Because dowry strictly increases the feasible transfer ceiling, B(d=1) ≤ B(d=0), meaning dowry converts some Seeking households into effectively Satisfied ones and lowers the migration threshold for the rest.

Q: What share of grooms’ parents actually take from the dowry, and how does migration status affect this? A: In the Destination Survey (62% migrants), 45% of parents take from the dowry on net; in the Origin Survey (20% migrants), 27% do. Parents of migrants are 27 percentage points more likely to take in the Destination Survey and 8 percentage points more likely in the Origin Survey, consistent with the model prediction that migration increases net taking.

Q: How does the son’s earnings level affect parental taking, and does this pattern hold for non-migrants? A: For migrant sons, a 100% increase in the son’s occupational score increases the likelihood of parents taking by 19 percentage points. For non-migrant sons, the son’s occupational score has no meaningful association with taking. This asymmetry is consistent with prediction 3: when migration occurs and the alpha income-sharing channel is shut down, parents with higher-income migrant sons have a higher relative marginal return to consumption and thus take more of the dowry.

Q: What is the remittance auxiliary prediction, and is it borne out in the data? A: The model predicts that parents who receive remittances from migrant sons should also be more likely to have taken from the dowry, because households first exhaust the costless dowry transfer before making costly or risky remittances — so remittance-receiving parents are precisely those Seeking households where dowry was already taken. The data confirm this: parents whose migrant son sends financial remittances are 17 percentage points more likely to have taken from the dowry (coefficient 0.168, SE 0.074, significant at 5%) compared to parents of migrants who do not remit.

Q: How is the district-level dowry tradition measure constructed and validated? A: The measure merges the Giuliano and Nunn (2018) Ancestral Characteristics data — which uses ethnographic sources to estimate the share of each district’s current population belonging to historically dowry-practicing groups — with district-level demographic data. Validation against the 1999 REDS shows that a one-unit increase in the historical dowry measure is associated with 81% higher gross dowry payments and 109% higher net dowry payments without region fixed effects, with a still-significant 79% for net dowry including region fixed effects. Additional validation in the IHDS confirms the historical measure predicts gold payments at marriage (coefficient 0.152 without state fixed effects, 0.185 with state fixed effects).

Q: What is the association between historical dowry traditions and migration in nationally representative data? A: Using the NSS Round 64 migration module (2007–08) for males aged 15–45, against a mean migration rate of 23.8%, the continuous dowry measure is associated with a 2.66 percentage point increase in migration probability with no controls (significant at 1%), and 3.67 percentage points with full controls including state fixed effects, year-of-birth fixed effects, caste fixed effects, distance controls, and education controls (significant at 5%).

Q: What is the highway construction identification strategy, and what does it show? A: The authors exploit the staggered construction timing of the Golden Quadrilateral and NS-EW highway corridors (beginning 1999, 5,846+ km, $71 billion investment) across Indian districts, assembling new data on district-level construction timing from a complete capital projects database. Using staggered-entry event study estimators robust to heterogeneous treatment effects, they separately estimate highway effects in districts with and without strong dowry traditions. For young men aged 15–30, dowry districts show a large, significant increase in out-migration after highway construction with no pre-trends; non-dowry districts show an effect indistinguishable from zero. Older men (31–45) show no significant effect in either group.

Q: Why is the age heterogeneity (15–30 vs. 31–45) in the highway results important for the mechanism? A: The model predicts that dowry’s migration-enabling role operates at the time of marriage, when the liquid transfer is made. Men aged 31–45 at the time of highway construction would largely have already been married before the roads were built, so they cannot retroactively benefit from the new liquidity channel. Young men (15–30) are near or below marriage age and can time their marriages and migration decisions in response to reduced migration costs. The null result for older men and the strong result for younger men together confirm the marriage-time liquidity channel.

Q: Why is the highway effect concentrated in inter-district rather than intra-district migration? A: The Golden Quadrilateral connects districts to other districts, and the model’s mechanism relies on migration creating income-sharing frictions that are more severe at longer distances. Intra-district moves are shorter, less likely to disrupt co-residence and informal support arrangements, and less likely to require the dowry’s compensatory role. The concentration of effects in inter-district migration is directly consistent with the proposed channel.

Q: How does the paper address concerns about pre-trends and robustness in the highway analysis? A: The event study plots show no pre-trends in migration for either dowry or non-dowry districts prior to highway construction. Robustness checks include additional geographic controls, caste-by-year fixed effects, time-varying cultural controls, the alternative Callaway-Sant’Anna estimator, adjusted age distributions, and varying dowry tradition cutoffs at 1%, 10%, and 25% thresholds. Results are stable across these specifications.

Q: What do the theory and evidence imply about the modern transformation of dowry’s function? A: While dowry historically served as a pre-mortem bequest to the bride adapted to patrilocal society, the modern practice has evolved so that grooms’ parents frequently capture the transfer. The evidence is consistent with this reallocation of property rights serving a new function: providing parents with a pension substitute when sons migrate and traditional co-residential support breaks down. The authors speculate this functional evolution may partly explain why dowry prevalence has grown despite legal bans, as declining patrilocality creates rising demand for this type of intergenerational transfer mechanism.

Q: What are the policy implications of the findings? A: The paper suggests that policies discouraging dowry — which has many well-documented negative consequences including intimate partner violence, female infant mortality, and adverse resource allocation — may be more effective if paired with expansions of formal pension programs or other mechanisms for old-age support. Without such alternatives, eliminating dowry could inadvertently reduce male migration and associated economic development benefits because the migration-enabling liquidity function of dowry would go unfilled.

Q: Does the mechanism apply equally to households with both sons and daughters? A: The theoretical appendix shows that in a household with a son and a daughter, the daughter’s dowry outflow partially offsets the son’s inflow, reducing but not eliminating the migration-enabling effect. However, the net aggregate effect on male migration remains positive because more sons live in households where sons outnumber daughters, so the dowry inflow for the son exceeds the outflow on average across the population.

Dowry (in the paper’s sense): A transfer from the bride’s family accompanying marriage that in the modern Indian context is liquid at the time of the wedding and over which grooms’ parents frequently exercise property rights — distinct from the traditional anthropological conception of dowry as a pre-mortem bequest to the bride.

Net Taker: A groom’s parent who receives a positive net transfer from the son’s dowry (tau > 0 in the model), meaning the flow of dowry resources is from the son/bride’s side to the groom’s parents.

Seeking vs. Satisfied parents: Model categories distinguishing parents whose consumption needs can be met from own income plus the maximum feasible marriage transfer (Satisfied, no migration distortion) from those whose needs cannot (Seeking, requiring a minimum migration return threshold B(d) > 0 for migration to be household-optimal).

Migration friction (alpha = 0 under migration): The modeling assumption that income-sharing transfers between migrant sons and parents are infeasible or prohibitively costly due to limited commitment, information asymmetries, and remittance costs — the friction that dowry’s lump-sum transfer at marriage is designed to circumvent.

Ancestral Characteristics dowry measure: The district-level variable from Giuliano and Nunn (2018) measuring the share of the current population belonging to historically dowry-practicing ethnic groups, used as a proxy for the strength of local dowry traditions.

Patrilocality: The residential norm in which sons remain with or near their parents after marriage and provide old-age support — the norm whose breakdown via migration creates the income-sharing friction that dowry helps resolve.

Pareto weight (theta): The weight assigned to parents’ utility in the collective household problem, capturing parental bargaining power; empirically proxied by whether sons report that parents held veto power over the marriage choice.

Trust and Innovation Within the Firm

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates whether and how a CEO’s inherited generalized trust enhances innovation within firms, offering a micro-foundation for the well-documented macro-level relationship between societal trust and economic growth. The author argues that trust — by inducing tolerance of failure — encourages researchers to undertake high-risk, explorative R&D rather than safe exploitation of known approaches.

The empirical foundation is a matched CEO-firm-patent dataset covering 5,753 CEOs at 3,598 US public firms during 2000–2011, encompassing 700,000 patents and over one million inventors. CEO trust is measured as an inherited trait: each CEO’s ethnic origin is inferred probabilistically from their last name using de-anonymized US censuses from 1910–1940, and ethnic-origin-specific trust levels are drawn from the US General Social Survey (GSS), restricted to respondents in highly prestigious occupations. The resulting trust measure is the weighted average of ethnic-specific trust scores across a CEO’s likely ethnic composition.

The main empirical strategy exploits within-firm variation across CEO transitions, using firm and year fixed effects to compare patenting before and after a CEO change. The identifying assumption — that the timing of CEO transitions and the new CEO’s trust level are not predicted by prior firm patenting trends — is supported by event-study tests showing flat pre-trends. A one-standard-deviation increase in CEO inherited generalized trust (equivalent to the difference between Greek and English averages) is associated with a 6.2–6.3% increase in patent filings, statistically significant at the 1% level. For the average firm, this equals approximately 1.1 additional patents annually, worth roughly $6.8 million. The effect is larger among exogenous transitions (CEO retirement or death): 8.5% in the restricted sample, and an IV estimate of 8.2%. The back-of-envelope calculation suggests this trust-innovation channel could account for approximately 37% (range: 16–58%) of the effect of trust on GDP per capita growth.

The paper’s central mechanism — risk taking — is tested by examining the distribution of patent quality rather than the mean. Under the risk-taking mechanism, trust should increase the variance of R&D project quality, raising high-quality patents without necessarily increasing low-quality ones. Consistent with this, CEO trust raises only above-median quality patents (measured by forward citation decile), with effects increasing monotonically toward the top decile and no statistically significant effect on below-median patents. Average patent quality as measured by citation-weighted counts or patent value rises by 4–6%. Trust also disproportionately raises the share of explorative patents (those with at least 90% of backward citations outside the firm’s existing knowledge stock) by 1 percentage point over a base of 17%.

The transmission channel is examined using BERT-based classification of nearly one million Glassdoor employee reviews. Under more trusting CEOs, firms exhibit stronger top-down trust sentiment (managers trusting workers), particularly among R&D workers and scientists. The effect materializes within the first two years of a CEO term. Director selection provides an additional transmission mechanism: under more trusting CEOs, newly appointed directors are more trusting and departing directors are less trusting.

A within-CEO design using bilateral trust (toward researchers in specific countries) with CEO fixed effects addresses omitted CEO characteristics. A one-standard-deviation increase in CEO bilateral trust toward a country is associated with a 5% increase in patents by inventors in that country’s R&D lab, controlling for firm-by-year, CEO, and inventor-country fixed effects.

The effect is strongest when CEO trust is matched to a high-quality researcher pool; in firms with mostly low-quality researchers, high trust may be counterproductive. Trust is also a substitute for R&D knowledge: the effect disappears when the CEO holds a non-MBA graduate degree or has prior R&D experience.

Q: What is the main research question? A: The paper asks whether a CEO’s generalized trust causes more and higher-quality innovation within the firm, and through what mechanism. It also asks how trust transmits from the CEO to researchers who rarely interact with the CEO directly.

Q: How is CEO trust measured? A: CEO trust is measured as an inherited trait using a two-step procedure. First, each CEO’s last name is probabilistically mapped to one or more ethnic origins using four de-anonymized US censuses (1910–1940). Second, ethnic-origin-specific trust is computed from GSS respondents in highly prestigious occupations. The CEO’s trust measure is the weighted average across ethnic compositions. This measure is shown to be more precise than an individual-level survey measure and approximately 80% as precise as a game-based measure, without introducing attenuation bias.

Q: What is the baseline patent effect and how large is it economically? A: A one-standard-deviation increase in CEO inherited trust is associated with a 6.2–6.3% increase in patent filings (statistically significant at 1%). For the average baseline firm, this is approximately 1.1 additional patents per year, valued at roughly $6.8 million. When patent quality is accounted for, the effect rises to 9.9% using citation-weighted patent count and 11.5% using patent value based on excess stock returns on grant dates.

Q: Is the effect causal? What identification strategy is used? A: The main strategy uses firm and year fixed effects, identifying the effect from within-firm variation around CEO transitions. Pre-trend tests confirm that neither the timing of CEO changes nor the new CEO’s trust level predicts prior firm patenting. Among exogenous transitions (CEO retirements and deaths), the effect is 8.5%, and an IV estimate using the predecessor’s trust as instrument yields 8.2% (significant at 10%), both comparable to the baseline.

Q: What is the macroeconomic significance of the trust-innovation channel? A: Combining the paper’s trust-to-patents estimate (0.042–0.062) with Akcigit et al.’s (2017) patents-to-GDP-growth estimate (0.026–0.066) and the cross-country trust-to-growth coefficient (0.007), the trust-innovation channel could explain approximately 37% of the effect of trust on growth, with a plausible range of 16–58%.

Q: What is the mechanism linking CEO trust to innovation? A: The conceptual mechanism is that a more trusting manager interprets researcher failure as bad luck rather than bad type, making her more likely to tolerate failure and continue employing the researcher. This increases the researcher’s incentive to pursue explorative, high-risk R&D over safe exploitation of known approaches. The mechanism implies a variance-increasing effect on the R&D quality distribution, rather than a mean shift.

Q: How is the risk-taking mechanism tested against alternative mechanisms? A: The paper examines the distribution of patent quality by citation decile. Under mean-shifting alternatives (delegation, cooperation, relational contracting), trust should raise all quality brackets. Under risk-taking, trust raises only high-quality patents. The results show CEO trust has monotonically increasing effects from low to high quality deciles, with no statistically significant effect on below-median patents, consistent only with the variance-increasing (risk-taking) mechanism.

Q: What patent quality measures are used and what do they show? A: Beyond forward citation deciles, the paper uses explorativeness (patents with at least 90% of backward citations outside the firm’s existing knowledge stock), disruptiveness (Funk and Owen-Smith, 2017), patent importance (Kelly et al., 2021), backward citations to scientific literature, and patent scope. Trust increases all these measures with statistically significant positive coefficients. The share of explorative patents rises by 1 percentage point over a base of 17%. Average citation count and patent value increase by 4–6%.

Q: Does CEO trust raise R&D expenditure? A: No. The coefficients from regressing R&D expenditure on CEO trust are neither statistically significant nor large enough to explain the innovation effect. The patent effect is also robust to controlling for R&D inputs, suggesting that trust affects the type of projects chosen (consistent with risk-taking) or their realized outcomes, rather than the scale of R&D.

Q: How does CEO trust transmit to corporate culture? A: Using BERT-based classification of nearly one million Glassdoor reviews covering 266 firms and 397 CEO terms between 2008 and 2017, the paper finds that CEO trust is associated with stronger top-down trust sentiment (managers trusting workers). The normalized effect of a one-standard-deviation increase in CEO trust on overall trust sentiment is 0.257, on top-down trust 0.531, and on bottom-up trust only 0.141 (statistically insignificant). The effect is strongest among reviewers who identify as scientists, researchers, or engineers, and materializes within the first two years of the CEO term.

Q: What evidence exists for transmission via director selection? A: Under more trusting CEOs, newly appointed directors — especially those who remain until the end of the CEO term — are more trusting, and departing directors are less trusting. The average director trust improves during the CEO’s term. Because 54% of director hirings and 46% of turnovers occur within the first two years, this change also materializes quickly, consistent with the dynamic pattern of trust culture change.

Q: What is the within-CEO bilateral trust result and what does it add? A: Using within-CEO variation in bilateral trust toward researchers from different countries (from Eurobarometer surveys), and controlling for CEO, inventor-country, and firm-by-year fixed effects, a one-standard-deviation increase in CEO bilateral trust toward a country is associated with a 5% increase in patents by inventors in that country’s R&D lab. This design allows CEO fixed effects, ruling out unobserved CEO-level confounders such as management style or R&D ability.

Q: When is CEO trust counterproductive? A: CEO trust is beneficial only when matched to a high-quality researcher environment. Using residual patent output (controlling for observable firm and CEO characteristics) as a proxy for researcher quality, the effect of CEO trust on patents, patent output per R&D dollar, and future sales/employment/TFP is significant only among firms in the top two quintiles of researcher quality. In firms with mostly low-quality researchers, high CEO trust may be counterproductive by failing to screen out bad researchers.

Q: How does the trust effect vary by industry and CEO background? A: The effect is ubiquitous across industries but especially pronounced in pharmaceutical and ICT firms. The timing varies: it manifests quickly in ICT (short R&D lag) and more slowly in pharma (long R&D horizon). The effect vanishes when the CEO holds a non-MBA graduate degree or has prior R&D experience, suggesting trust is a substitute for direct knowledge of R&D processes.

Q: Are the results robust? A: Yes. The paper reports 14 categories of robustness checks including alternative patent transformations, alternative trust measures (LASSO, World Value Survey, Global Preference Survey, alternative GSS questions), alternative standard error clustering, Poisson count models, restriction to granted patents, exogenous transition subsamples, modern difference-in-differences estimators (de Chaisemartin et al., 2024; Sun and Abraham, 2021; Callaway and Sant’Anna, 2021; Borusyak et al., 2024), and leave-one-ethnicity-out. The baseline result is stable across all these checks.

Inherited generalized trust: The paper’s measure of a CEO’s trust disposition, defined as the probability-weighted average of ethnic-origin-specific trust levels (from the GSS) based on the CEO’s likely ethnic composition inferred from their last name and historical census records. It captures the culturally transmitted component of trust, distinct from individual-level noise.

Explorative R&D: In the paper’s framework (building on March, 1991), research activities that involve testing untested paths, carrying high risk of failure but high potential for innovation, as opposed to exploitation of well-known approaches with low failure risk. The paper argues CEO trust encourages researchers to shift toward exploration.

Tolerance of failure: A manager’s propensity to attribute a researcher’s failure to bad luck rather than bad type. Under the paper’s mechanism, a more trusting manager gives greater weight to bad luck, making her more likely to retain the researcher after failure, thereby incentivizing risk taking.

Top-down trust: In the paper’s BERT-based classification of Glassdoor reviews, the direction of trust from managers toward workers (as opposed to bottom-up trust from workers toward managers). The paper finds CEO trust primarily raises top-down trust sentiment, especially among R&D workers.

Patent explorativeness: A patent quality measure defined as the share of its backward citations that fall outside the firm’s existing knowledge stock; patents are classified as explorative if at least 90% of backward citations are outside that stock. The paper uses this as a direct measure of explorative R&D output.

Bilateral trust: CEO d’s directed trust toward individuals from country c, computed analogously to inherited generalized trust but using Eurobarometer survey data on country-pair trust attitudes among European-origin populations. Used in the within-CEO design to control for CEO fixed effects.

Variance-increasing mechanism: The paper’s characterization of the risk-taking channel, in which CEO trust raises the variance (not the mean) of the R&D project quality distribution by encouraging researchers to pursue high-risk, high-reward exploration. Empirically identified by the pattern that trust raises only above-median quality patents with monotonically increasing effects toward the top decile.

Vanguard: Black Veterans and Civil Rights After World War I

Mon, 01 Jan 0001 00:00:00 +0000

This paper provides the first causal evidence on how military service shaped Black civil rights activism in the aftermath of World War I. The research question is whether random induction into the segregated National Army caused Black men to join the nascent NAACP and become prominent community leaders during the New Negro era. The authors leverage the WWI draft lottery — in which each registrant’s unique serial number was drawn from a bowl to determine induction order — as an instrument for military service, a source of exogenous variation not previously exploited in the literature.

To support this analysis, Ang and Chinoy construct an unusually rich dataset by digitizing nearly one million Black draft registration cards from the first registration (June 17, 1917), linking them through the 1930 full-count census to 233,517 NAACP member observations across 227 branches from 1912 to 1940, and supplementing with Veterans Administration records, Army Transport Service passenger lists, and biographical dictionaries of prominent African Americans. The instrument — serial number percentile within draft board and race (SNP%) — is validated against all observed pre-draft registrant characteristics and yields a first-stage F-statistic of 1,051 in the preferred specification.

The main finding is that Black men randomly induced to serve in the military were nearly three times more likely to join the NAACP than observably similar registrants from the same draft board (TSLS coefficient 0.0219, se = 0.0049, against a sample mean NAACP participation rate of 0.8%). The authors estimate that the draft induced more than 10,000 Black men to join the NAACP in total. Military service also raised the probability of appearing in biographical dictionaries of historically prominent African Americans by a factor of roughly 1.6 (TSLS coefficient 0.0027, se = 0.0012, sample mean 0.17%). These results are robust to alternative instruments, flexible polynomial specifications of SNP%, state-year fixed effects, and alternative veteran-status measures from VAMI and ATS records. They are also not explained by differential residential mobility: adding controls for interstate and North-South migration leaves the main coefficient essentially unchanged (0.0217-0.0218).

In contrast, TSLS estimates for all socioeconomic outcomes — literacy, home ownership, employment, census-predicted income, actual 1940 income, and educational attainment — are small and insignificant, ruling out human capital acquisition as a mechanism. Club involvement measured in the census is likewise unaffected, indicating that NAACP membership reflects specifically civil rights activism rather than generically greater social participation.

The mechanism the paper identifies is experienced discrimination. Effects on NAACP participation increase monotonically with the racial gap in induction rates across draft boards (significant at p = 0.01). Effects are large and significant for men assigned to camps that restricted Black soldiers’ access to military training (coefficient 0.0351, se = 0.0104) and to officer promotion (coefficient 0.0360, se = 0.0111), and are large for men in both restriction types simultaneously (coefficient 0.0367, se = 0.0114). In contrast, men attending less discriminatory camps show small and insignificant effects. Among the two all-Black combat divisions, NAACP participation is highest for veterans of the 92nd Division — subjected to constant racial abuse under U.S. command — and lower for the 93rd Division, which served under more hospitable French command. Previously unstudied veteran surveys from Virginia and Connecticut corroborate this narrative: respondents from camps with training and promotion restrictions were more than twice as likely to mention racial injustice, and mentions of injustice were more predictive of postwar civic engagement than any other survey theme.

The scope of the paper is Black male registrants in the first WWI draft registration (men aged 21-30 as of June 17, 1917), linked to a sample of approximately 300,000 in the 1930 census. Effects are attenuated for men from counties with greater racial hostility — proxied by Confederate state status, Confederate monument density, and county lynching rates — consistent with the interpretation that activism was more feasible in less repressive environments.

Q: What is the core identification strategy and why was it not feasible to use it before this paper? A: The paper uses each Black registrant’s serial number percentile within his draft board and racial group (SNP%) as an instrument for WWI military service. Unlike the WWII and Vietnam drafts, which used birthday-based lotteries, the WWI lottery assigned induction order by drawing unique serial numbers from a bowl, making serial number rank the source of quasi-random variation. This source had never been exploited in the literature, partly because the serial numbers had to be hand-captured from digitized draft card images.

Q: How strong is the first stage, and was the lottery truly random? A: The first-stage F-statistic is 1,051, and a ten-percentile decrease in SNP% is associated with a 34.5 percentage point increase in the probability of serving. Bivariate serial numbers show some non-random patterns — nine of 13 pre-draft characteristics correlate with raw SN% — likely because some Southern boards inflated numbers for white registrants. Conditioning on board fixed effects and using SNP% within board-race cells eliminates these correlations; Panel B of Appendix Table A1 shows the largest standardized coefficient falls to 0.006.

Q: What is the magnitude of the effect on NAACP membership and how does the causal estimate compare to a naive OLS? A: The TSLS coefficient is 0.0219 (se = 0.0049) against a sample mean of 0.8%, implying roughly a threefold increase in NAACP membership. The OLS estimate of 0.0116 understates the causal effect, consistent with the marginal man induced by the lottery being observationally weaker than infra-marginal volunteers.

Q: Does the effect reflect simply that veterans moved to Northern cities where NAACP branches were more accessible? A: No. Adding indicators for interstate migration and North-South migration leaves the TSLS coefficient essentially unchanged at 0.0218 and 0.0217, respectively. The Great Migration channel is thus not the operative mechanism.

Q: Did military service improve Black veterans’ economic outcomes? A: TSLS estimates for literacy, home ownership, employment, census-predicted income, actual 1940 income, and educational attainment are all small and statistically insignificant. This contrasts sharply with evidence on Black veterans of WWII and Korea (Greenberg et al., 2022) and is consistent with the documented absence of meaningful postwar benefits or training for Black WWI soldiers.

Q: If it was not human capital or migration, what mechanism does the paper establish? A: The primary mechanism is exposure to institutional discrimination during military service. Three distinct empirical patterns converge: (1) effects increase monotonically with draft board racial disparities in induction rates; (2) effects are large and significant for men at camps that denied training and promotion, and near zero for men at less discriminatory camps; (3) veteran survey mentions of racial injustice are more common among men from discriminatory camps and are more predictive of postwar NAACP membership than any other survey theme.

Q: How do the two all-Black combat divisions differ in their postwar NAACP participation, and what does this reveal? A: Veterans of the 92nd Division, who fought under U.S. command amid constant racial abuse, show the highest NAACP participation rates. Veterans of the 93rd Division, who fought under French command and were received with relative hospitality, show lower (though not statistically significantly lower) participation. Since both divisions received similar formal training and neither group shows socioeconomic gains, the differential reflects discrimination exposure rather than skill acquisition.

Q: What is the quantitative scale of the effect for the most discriminatory camps? A: For men assigned to camps with restrictions on both training and promotion, the TSLS coefficient on NAACP membership is 0.0367 (se = 0.0114) — more than 1.5 times the average estimate of 0.0219. Men at camps without restrictions show coefficients that are small and statistically insignificant.

Q: How does county-level racial hostility moderate the effect? A: The effects of military service on NAACP membership are larger — more positive — for men from counties with fewer Confederate monuments, lower lynching rates, and non-Confederate state status. This is interpreted as evidence that activism in response to discriminatory military experiences was more feasible in less racially hostile local environments, rather than as evidence that discrimination exposure was lower.

Q: What is the paper’s aggregate policy implication regarding the scale of the draft’s effect on the civil rights movement? A: The authors estimate that the WWI draft induced more than 10,000 Black men to join the NAACP. Veterans accounted for nearly 15% of all male NAACP members, against roughly 8% of Black male adults in the population, and were significantly more likely to appear in biographical dictionaries of prominent African Americans. The draft thus constituted a sizable and measurable contribution to the organizational vanguard of the early civil rights movement.

Q: How does the paper contribute to the economics of discrimination beyond documenting discriminatory behavior by majority actors? A: Most economics research on discrimination studies the conduct of white decision-makers (e.g., racial bias in hiring, lending, or bail). This paper examines how experiences of discrimination reshape the political behavior and aspirations of the minority group itself. The results show that institutional betrayal — systematic exclusion, degradation, and denial of training — generated deep discontent that translated into aggressive political mobilization, a dynamic the authors trace through subsequent episodes including the WWII Double V campaign and responses to police killings.

Serial number percentile within draft board and race (SNP%): The instrument constructed by the authors. Each WWI registrant received a serial number from 1 to the size of his draft board; those numbers were drawn in random order to determine induction priority. SNP% measures where a registrant fell in that draw relative to others in his board and racial group, and serves as the source of quasi-random variation in veteran status.

New Negro era: The period of invigorated Black political and cultural assertiveness following WWI, characterized by renewed racial pride, economic independence, and progressive politics. The movement spanned the Harlem Renaissance, the Universal Negro Improvement Association, the American Negro Press, and the Brotherhood of Sleeping Car Porters, and represented a rejection of the “conservatism, parochialism, and political accommodationism” of older Black leaders.

Draft board racial gap: The authors’ measure of draft board discrimination, defined as the difference in induction rates between Black and white registrants within a given draft board. The interquartile range spans roughly 0 to 20 percentage points, with a notable fraction of boards exhibiting gaps exceeding 30 percentage points.

Camp discrimination: The denial of military training and officer promotion opportunities to Black soldiers, documented in War Department reports by military intelligence officers tasked with monitoring the treatment of Black soldiers. The paper classifies each camp as restricted or unrestricted on each dimension and uses this classification to estimate heterogeneous treatment effects.

Institutional betrayal: The paper’s characterization of the U.S. government’s treatment of Black WWI soldiers — drafting them at higher rates than whites, denying them training and promotion, and assigning them to menial labor — as generating a profound sense of injustice that motivated postwar political activism rather than loyalty or accommodation.

NAACP membership as civil rights activism proxy: The paper uses dues-paying membership in local NAACP branches as its primary quantitative measure of civil rights participation. Membership involved active financial cost (annual fees of $1 to $10 at a time when median Black family income was below $500), exposure to harassment and violence in the South, and participation in local protest and legal advocacy, distinguishing it from passive civic engagement.

Voluntary Minimum Wages: The Local Labor Market Effects of National Retailer Policies

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

This paper studies the labor market effects of voluntary minimum wages (VMWs) — company-wide, publicly announced wage floors set by large private employers — in the U.S. low-wage retail and service sector from 2014 to 2023. The central questions are: (1) How do VMWs affect wages and employment at the adopting large retailers? (2) Do VMWs generate wage spillovers to other employers in shared local labor markets?

Data and Setting

The authors use anonymized payroll data obtained from a large U.S. credit bureau, covering the wage distributions and employment of over 4,000 firms and approximately 18 million hourly workers (roughly 22–24% of the U.S. hourly workforce) from January 2013 to August 2023. The database is skewed toward retail and service sectors: over a third of covered workers are in retail, and over half in retail and services combined. Critically, the data also include worker flow information — records of individual workers moving between firms — enabling the authors to define shared labor markets via actual employment transitions rather than broad geographic or industry proxies.

The sample of VMW events consists of 20 voluntary minimum wage policies across 5 large retailers (each with over 150,000 employees nationally), restricted to events with no other major wage policy within six months before or after the focal event. Voluntary minimum wage announcements were identified from an inventory maintained by the National Employment Law Project and independently verified through media sources, then matched to anonymized companies using employer size, industry, and observed shifts in the wage distribution.

Identification Strategy

The authors adapt the gap design from the national minimum wage literature. For each company-by-commuting-zone (CZ) cell, the “gap” measures the percent increase in average hourly wages that would be required to bring all workers in the area up to the company’s new voluntary minimum. The gap is averaged over months −6 to −3 before the event (months −3 to −1 serve as a built-in placebo-in-time check). This variation in bite across CZs — arising because the same nominal VMW level implies different wage increases depending on local wage distributions — is combined with a stacked event study across 20 VMW events. Spillover effects are estimated by regressing log average wages at non-policy establishments on the large retailer’s CZ-level gap measure, progressively narrowing the definition of “labor market” from: (i) all non-policy establishments in the same CZ, to (ii) establishments in industries connected to the large retailer by worker flows (15 three-digit NAICS industries), to (iii) specific establishments with documented pre-event worker flows to or from the large retailer (“connected establishments”).

Main Findings

Own effects: For $15 VMW events, moving from a CZ gap of 0 to a gap of 1 is associated with an approximately 88 log point increase in average hourly wages in the six months after adoption. Given that the average establishment-level gap for $15 VMWs is 0.11, the implied average wage increase is approximately 10.45% (the authors’ estimate is 9–10%, consistent with small wage increases even in zero-gap comparison areas). Employment of workers earning under $30 per hour rose by 4.62% after $15 VMW events, 2.01% after major events (affecting ≥30% of workforce), and 1.25% across all 20 events. These employment increases are entirely attributable to reduced separations rather than new hiring: separation rates fell by 0.42, 0.57, and 1.09 percentage points after all, major, and $15 VMW events respectively — equivalent to reductions of 6.57%, 8.73%, and 15.33% relative to pre-period means. Separations specifically to other database companies fell by 0.07–0.19 percentage points (5.63–13.48% relative to base rates). If anything, new hiring fell modestly after VMW adoption. Total monthly base pay and gross compensation both rose after VMWs, indicating increased total take-home pay without compensatory reductions in hours or bonuses. The total employment elasticity with respect to wages ranges from approximately 0.35 to 0.45, while the quit elasticity is 2.20–2.38 (consistent with dynamic monopsony models in which the labor supply elasticity is twice the quit elasticity).

Spillover effects: Across all three definitions of the labor market, the paper estimates precise, economically negligible cross-employer wage spillovers in the six months following VMW events. Cross-employer wage elasticities are statistically indistinguishable from zero across all specifications. Among the most narrowly defined sample — establishments with documented pre-event worker flows to or from the large retailer — the upper bound of the confidence interval rules out spillovers greater than 0.2% of wages. No wage spillovers are detected for new hires at non-policy establishments either. These null results are confirmed over a 12-month post-event horizon for the subsample of events with no other major policy nearby.

Mechanism: The reason for negligible spillovers is that VMWs reduced labor market churn rather than expanding the large retailer’s total employment. Hiring away from large retailers by connected non-policy firms falls after VMW adoption — consistent with fewer separations to recruit from — but overall hiring by non-policy firms does not decline, as these firms substitute toward other hiring sources. This substitutability across new hire sources in a thick market is the proximate explanation for the absence of wage pressure on competitor firms.

Scope Conditions

Results pertain to large national retailers (>150,000 employees) operating in U.S. commuting zones during 2014–2023. The database covers only employers large enough to participate in credit bureau income verification; smaller employers (representing over 75% of U.S. hourly workers by the BLS comparison) are not observed, and the authors caution that spillover effects on smaller firms cannot be assessed. The authors also explicitly note that their null local spillover results do not rule out national-level strategic wage-setting dynamics — the rapid sequential adoption of VMWs across major retailers may reflect national-level competition rather than local market competition.

In depth

Q1. What exactly are “voluntary minimum wages” and how do they differ from statutory minimum wages?

Voluntary minimum wages (VMWs) are company-wide, publicly announced wage floors set unilaterally by private employers, typically well above the applicable statutory (federal, state, or local) minimum. Unlike statutory minimums, which bind all employers in a jurisdiction, VMWs apply only to the announcing company across all of its geographic operations in the U.S. The paper studies VMWs adopted by retailers with over 150,000 workers, which include wage floors at levels such as $9, $10, $12, and $15 per hour. $15 VMWs were adopted at a time when few states or localities had yet reached that threshold, meaning the policy bit into the company wage distribution far more deeply than prevailing statutory floors.

Q2. How were VMW events identified and matched to anonymized firms in the payroll database?

VMW events were identified from a database maintained by the National Employment Law Project and verified through an independent review of business news articles. These publicly reported announcements were then matched to the anonymized companies in the credit bureau payroll database using employer size, industry, and the timing of observed shifts in the firms’ wage distributions. An additional three events were identified directly from data: months where the share of workers earning below a given wage level dropped by at least 15 percentage points (for non-$15 events) or 10 percentage points (for $15 events) while the share at exactly that wage bin jumped by at least 10–20 percentage points. The final sample of 20 events was restricted to those with no other major wage policy in the six months before or after.

Q3. How does the gap design work and why does it improve on the fraction-affected approach?

The gap for a given company, commuting zone, and time period is defined as the total wage increase needed to bring all sub-$30 workers up to the company minimum, divided by total wage costs — formally a labor-share-weighted average shortfall from the new minimum across wage bins. The gap leverages more cross-sectional variation in treatment intensity than the simple fraction of workers below the minimum: for a $15 VMW, an area where all workers earn $10 has a gap of 0.50 while an area where all earn $12 has a gap of 0.25. The gap is averaged over months −6 to −3 before the event. The period months −3 to −1 then serve as a placebo window: genuine VMW effects should appear only after the policy’s adoption month, not during the period immediately after the gap is measured. If instead the regression picks up mean reversion in noisy wage data, spurious effects would appear in months −3 to −1 rather than at event time 0.

Q4. What is the magnitude of the wage effect on the large retailers themselves?

For $15 VMW events, the stacked event study estimates that moving from a gap of 0 to a gap of 1 is associated with an approximately 88 log point increase in average hourly wages beginning exactly in the month of policy adoption. Given the average establishment-level gap of 0.11 for $15 VMWs, this implies the average establishment raised wages by approximately 9–10% (the authors compute 10.45% from the average gap, consistent with a slight dampening because zero-gap CZs experienced marginally higher wages too). Wage increases are confirmed persistent at 12 months in robustness checks. For all 20 VMW events pooled, effects are somewhat smaller commensurate with the lower average bite.

Q5. How did VMWs affect total employment and its components at the large retailers?

After $15 VMW events, log total employment of sub-$30 workers rose by 4.62%; after major VMW events (≥30% bite), 2.01%; after all 20 events, 1.25%. The increases are entirely driven by retention gains. Separation rates fell by 1.09 percentage points after $15 VMWs, 0.57 p.p. after major events, and 0.42 p.p. after all events — translating to reductions of 15.33%, 8.73%, and 6.57% relative to pre-period means. Separations to other database companies specifically fell by 0.07–0.19 percentage points (5.63–13.48% relative to the base mean). New hiring — measured as year-on-year log change in hires to control for seasonality — fell after VMW adoption, consistent with a reduced need to replace departing workers.

Q6. What do the labor supply elasticities implied by the VMW results look like?

The total employment elasticity with respect to wages ranges from approximately 0.35 to 0.45 across the three event groupings. Under standard dynamic monopsony models, the labor supply elasticity facing the firm equals twice the quit elasticity in steady state (Manning, 2003). The quit elasticity — derived by dividing the proportional reduction in separations by the log wage increase — ranges from 2.20 to 2.38, consistent with the earlier monopsony-based case study of Ford’s $5 workday (Raff and Summers, 1987) and implying substantial firm-level wage-setting power.

Q7. Did VMWs increase total take-home pay or were wage gains offset by reductions in hours or bonuses?

The paper examines log average monthly base pay and log average gross compensation (which includes bonuses and overtime) as additional outcomes. Both measures rose after $15 VMW events, indicating that the wage floor increase translated into genuine improvements in total take-home pay without compensatory reductions in hours or other non-wage compensation. The monthly gross pay series is an average over calendar year-to-date months, so increases appear gradually rather than as a sharp jump at the adoption month; nevertheless the upward trend is evident and consistent.

Q8. What are the estimated spillover effects on wages at non-policy employers?

Across all three definitions of the labor market — all non-policy establishments in the same CZ, establishments in the 15 connected industries in the same CZ, and establishments with documented pre-event worker flows — the estimated cross-employer wage effects are precise zeros. The stacked event study in the post-period shows coefficients centered on zero with small confidence intervals. The difference-in-differences cross-employer wage elasticity (instrumenting the large retailer’s wage change with the gap) is also indistinguishable from zero. Among the most exposed connected establishments, the point estimate is slightly positive but economically negligible; the upper confidence interval bound rules out spillovers greater than 0.2%. Results are confirmed over a 12-month horizon for the clean-event subsample.

Q9. Could the null spillover result reflect mean reversion bias rather than a true zero?

The authors address this concern explicitly. For the policy-company gap design, they build in a placebo-in-time check by measuring the gap over months −6 to −3 and checking that no wage effects appear in months −3 to −1. For the non-policy spillover analysis, they also examine an alternative treatment variable — the gap between non-policy establishments’ wages and the large retailer’s new VMW — and find evidence of mean reversion: wages begin rising in the pre-period in the direction of this gap measure. They correct for this by detrending post-period estimates using a linear extrapolation of the pre-period trend. After detrending, spillover effects remain indistinguishable from zero.

Q10. Why are spillover effects so limited if the large retailer is drawing fewer workers away from competitors?

The paper’s mechanism analysis shows that while the probability of a non-policy firm hiring a worker from the large retailer falls after a VMW event (consistent with fewer separations to recruit from the large retailer), the overall rate of hiring by non-policy firms does not decline. Non-policy firms substitute toward other hiring sources — primarily other non-policy companies — rather than hiring fewer workers overall. This substitutability across recruiting sources in a thick labor market mutes the competitive pressure on competitor wages: since non-policy firms can replace the reduced flow from VMW companies with workers from other sources without changing total employment, they face no pressure to raise wages.

The authors test whether larger local market presence amplifies spillovers by splitting the sample at the median employment share of the large retailer in the CZ. They find no evidence of positive wage spillovers even in CZs where the large retailer’s employment share is above the median, confirming that neither local market size nor market concentration is a mechanism for spillover transmission in this setting.

Q12. How do these VMW spillover results compare to prior evidence on employer wage-setting spillovers?

The main prior U.S. evidence (Staiger et al., 2010) studied a federally mandated wage increase at Veterans Affairs hospitals and found a cross-establishment wage elasticity of approximately 0.19 for registered nurses at neighboring hospitals. The authors note two key differences: first, the VA policy increased both wages and employment at treated facilities, whereas VMWs primarily reduced separations without increasing hiring, so the supply of workers to competitor firms was not squeezed. Second, the market for low-wage retail and service workers is likely thicker (more potential hires available) than the market for registered nurses, allowing competitors to substitute hiring sources without bidding up wages.

Q13. What do the null local spillover results imply about national-level wage dynamics?

The authors explicitly caution against reading the null local spillover result as implying VMWs have no broader effect on the low-wage labor market. The rapid and successive adoption of VMWs across major retailers during 2021–2022 could reflect national-level strategic wage-setting competition — firms mimicking each other’s announcements in an arms-race dynamic during tight labor markets — rather than local competitive transmission. The paper does not test for national-level strategic interactions and calls for further research on this dimension.

Key Concepts

Voluntary Minimum Wage (VMW): A company-wide, publicly announced wage floor set unilaterally by a private employer, applying across all of the firm’s geographic operations in the U.S., typically well above applicable statutory minimums. Distinct from legally mandated minimum wages in that they bind only the announcing firm and arise from the firm’s own strategic or reputational motivations.

Gap Measure: Borrowed from the national minimum wage literature (Card, 1992; Draca et al., 2011), this is the percent increase in a firm’s average hourly wage that would be required to bring all workers in a given commuting zone up to the company’s new voluntary minimum. Formally the labor-share-weighted average shortfall from the VMW across sub-$30 wage bins. A gap of 0 means no workers fall below the new minimum; a gap of 1 means all workers would need to be raised to the minimum, doubling the average wage. Used as a continuous treatment variable capturing the local bite of the policy.

Stacked Event Study: An empirical design in which a separate 12-month panel (6 months pre- and post-event) is constructed for each of the 20 VMW events, these datasets are stacked, and the effect of the continuous gap treatment is estimated jointly across all events, with event-specific indicators interacting all regressors to allow each event to have its own intercept.

Placebo-in-Time Check: A robustness test built into the gap design by computing the gap over months −6 to −3 and verifying that wage effects do not appear in months −3 to −1 (the period between gap measurement and VMW adoption). Genuine policy effects should materialize at the adoption month; spurious effects driven by mean reversion in noisy wage data would appear in months −3 to −1 because the gap would mechanically predict wage reversion toward the mean in the period immediately following its measurement.

Connected Establishments / Poaching and Feeder Establishments: Specific firm-by-CZ cells identified as sharing a labor market with the large retailer via actual worker flows. “Poaching establishments” hired at least one worker from the large retailer in the 12 months before the VMW event. “Feeder establishments” had at least one worker subsequently hired by the large retailer in the same pre-period. These are the most narrowly defined and most economically relevant labor market competitors for testing spillover effects.

Quit Elasticity / Labor Supply Elasticity (Firm-Level): The quit elasticity is the percent change in the separation rate divided by the percent change in wages induced by the VMW. Under standard dynamic monopsony models (Manning, 2003), in steady state the recruit elasticity equals the quit elasticity, and the firm-level labor supply elasticity equals twice the quit elasticity. The authors estimate quit elasticities of 2.20–2.38, implying labor supply elasticities of 4.40–4.76 to the firm — consistent with meaningful but not extreme monopsony power.

Cross-Employer Wage Elasticity: The percent change in wages at a non-policy employer’s establishment associated with a 1% change in wages at the large retailer in the same commuting zone, instrumented using the large retailer’s gap interacted with the post-event indicator. Estimated to be a precise zero across all market definitions and event groupings in this paper.

What Jobs Come to Mind? Stereotypes About Fields of Study

Mon, 01 Jan 0001 00:00:00 +0000

Conlon and Patel test whether students stereotype the link between college majors and occupations — that is, whether they exaggerate the likelihood that majors lead to their “representative” careers (those most overrepresented among a major’s graduates relative to other majors, as measured by a likelihood ratio in US census data). The representative career for each major is intuitive: doctors for biology/chemistry, lawyers for political science, counselors for psychology, journalists for communications, artists for art, and so forth.

The authors draw on three bodies of evidence. First, surveys of first-year undecided undergraduates in Ohio State University’s Exploration program (primarily Fall 2020 and Fall 2021 cohorts, ~80% response rate), asking students their beliefs about the share of US graduates in various careers conditional on major, as well as their beliefs about their own likely career. Beliefs are benchmarked against true career shares computed from the 2017–2019 American Community Survey restricted to college graduates aged 30–50. Second, 40+ years (1975–2018) of the CIRP Freshman Survey from UCLA, covering more than nine million nationally representative US college freshmen, which records intended major and intended career. Third, a field experiment embedded in the 2021 OSU survey with an RD design, in which treated students were shown the true share of their top major’s representative career before reporting beliefs, intentions, and — via administrative records — actual course enrollments and major declarations up to three years later.

The main finding is large, systematic overestimation of representative careers. In the OSU survey, students believe 53% of art majors work as artists (true: 17%), 47% of journalism majors work as journalists (true: 4%), 38% of political science majors work as lawyers (true: 16%), and 43% of psychology majors work as counselors (true: 21%). OLS regressions of beliefs on true career frequency and a representative-career indicator yield a stereotyping coefficient θ of 0.32 p.p. (p < 0.01) without career fixed effects and 0.28 p.p. (p < 0.01) with them, meaning students believe representative careers are roughly 28–32 percentage points more common than equally prevalent non-representative careers. These patterns are similar across gender, ethnicity, and first-generation status, replicate in an MTurk sample (θ = 0.30, p < 0.01) and a nationally representative US adult sample (θ = 0.33, p < 0.01).

In the CIRP data, 63% of biology freshmen expect to become doctors (true: 23%), 62% of psychology freshmen expect to be counselors (true: 21%), 65% of art freshmen expect to be artists (true: 17%), and 42% of communications/journalism freshmen expect to be writers or journalists (true: 4%). The average gap between expected and actual representative-career attainment is 36 p.p., and this gap has been roughly stable since at least the 1970s.

An implicit association test (IAT) administered to 434 OSU students shows that implicit associations between representative major–career pairs are 0.30–0.36 standard deviations stronger than for non-representative pairs (p < 0.01), and remain 0.24–0.28 SDs stronger (p < 0.01) after controlling for true career frequency. A one-SD increase in individual IAT scores predicts 2.8–4.1 p.p. greater stereotyped beliefs (p < 0.01). Knowing someone with a non-representative major–career combination predicts beliefs 16 p.p. lower for the representative career (p < 0.01) — more than half the stereotyping effect — and also predicts lower IAT scores, suggesting associations arise from personal experience.

An equilibrium model shows that stereotyping causes students to infer that representative careers have unusually favorable unobservable attributes, and that this inflates enrollment in the representative major among marginal students who are poorly suited to it. Correlational evidence from the NSCG, SIPP, and SHED confirms that majors subject to greater stereotyping are associated with more job dissatisfaction (+6.0% per SD, p < 0.01), greater job-skill mismatch (+3.1%, p < 0.05), more major-career mismatch (+5.4%, p < 0.05), and more regret about field of study (+4.8%, p < 0.05).

The field experiment shows that correcting beliefs reduces stereotyping and shifts major choices. A 10 p.p. reduction in beliefs about the top major’s representative career lowers intentions toward that major by 3.5 p.p. (p < 0.01), reduces enrollment in that major’s courses by 0.22 credits in the next semester (p < 0.05), and reduces the probability of declaring that major within one year by 6.1 p.p. (p = 0.23). The same information boosts intentions toward students’ second-ranked major by 2.1 p.p. (p = 0.17), increases second-major course enrollment by 0.20 credits (p < 0.10), and raises the probability of declaring the second major within a year by 9.9 p.p. (p < 0.01). Treated students also spend on average 0.21 more semesters undecided before declaring a major (p < 0.05). Effects are concentrated in the first year and partially fade over the two-to-three-year follow-up window.

Q: How do the authors define a major’s “representative career”? A: The representative career of major M is the career c that maximizes the likelihood ratio R(c, M) = p_{c|M} / p_{c|not-M}, where p_{c|M} is the true share of major-M graduates working in career c and p_{c|not-M} is the share of graduates from all other majors working in c. This ratio captures how much more common a career is among one major’s graduates relative to all other graduates. For example, the representative career of communications/journalism is “writers and journalists,” whose graduates are between 155% and 1,751% more likely to hold their major’s representative career than graduates of other majors, even though the absolute frequency of such careers is often modest (ranging from 2% to 60% across fields).

Q: What is the core model of stereotyped belief formation? A: The model draws from Bordalo et al. (2016). Let p_{c|M} be the true career share and π_{c|M} the student’s belief. The model specifies π_{c|M} = (1 − θ) p_{c|M} + θ · 1[c = c*(M)], where c*(M) is the representative career and θ ∈ [0,1] measures the extent of stereotyping. When θ = 0 the student holds rational beliefs; when θ = 1 beliefs assign all probability mass to the representative career. This formulation implies that students overweight representative careers because those careers come to mind more easily, grounded in a representativeness heuristic based on likelihood ratios.

Q: What does the regression test for stereotyping find in the OSU survey? A: The authors regress individual beliefs π_{c|M} on the true frequency p_{c|M} and an indicator for c being the representative career of M, clustering standard errors at the individual and career-by-major level. The estimated θ is 0.32 (p < 0.01) without career fixed effects (Column 1 of Table 1) and 0.28 (p < 0.01) with career fixed effects (Column 2). For self-beliefs about students’ top-ranked major, the estimates are 0.36–0.43 p.p. (p < 0.01 both with and without career fixed effects). These estimates imply that students regard a major’s representative career as 28–43 percentage points more common than an equally prevalent non-representative career for the same major.

Q: Do the OSU results replicate in other samples? A: Yes. An MTurk convenience sample of 430 current college students yields a stereotyping coefficient of 0.30 (p < 0.01). A nationally representative sample of US adults yields a coefficient of 0.33 (p < 0.01); this pattern holds separately for college-educated and non-college-educated respondents and for both younger respondents (aged 18–29) and older respondents (aged 30+). The authors also ran a pre-registered 2021 replication survey in a new OSU Exploration cohort and found similar results.

Q: What does the CIRP Freshman Survey data show about the persistence and scale of stereotyping? A: Pooling more than nine million US college freshmen surveyed from 1975 to 2018, the CIRP data show that students systematically intend to enter their major’s representative career far more often than graduates actually do. Among students who have decided on a major, 63% intend to have their major’s representative career while only 27% of college graduates actually attain it — a gap of 36 p.p. (p < 0.01). The specific examples include: 63% of biology freshmen intend to become doctors (true: 23%), 62% of psychology freshmen expect to be counselors (true: 21%), 65% of art freshmen expect to be artists (true: 17%), and 42% of communications/journalism freshmen expect to be writers or journalists (true: 4%). The gap has been stable over the full 40+ year window, with no sign of convergence, and amounts to 40,000–200,000 students per year expecting careers in representative fields that they will not attain.

Q: Can alternative mechanisms such as overconfidence or motivated reasoning explain the results? A: The authors argue no, for two reasons. First, students overestimate the prevalence of representative careers not only for majors they plan to pursue (where overconfidence or motivated reasoning might apply) but also for majors they do not plan to pursue — the pattern holds for the gray (population belief) bars across all ten majors in Figure 1. Second, a Shapley-Sharrocks decomposition reported in Table A.V shows that the stereotyping mechanism accounts for a larger share of variance in beliefs than any other mechanism tested. A pre-registered survey also rules out unawareness of non-representative occupations as a driver: students are aware of the overwhelming majority of the 100 most common non-representative occupations, and such unawareness as exists is uncorrelated with stereotyped beliefs.

Q: What does the IAT reveal about the mechanism behind stereotyping? A: The IAT was run on 434 OSU Exploration students in Fall 2021, measuring implicit associations between five major–career pairs (Humanities-Writers and Journalists, Sciences-Healthcare, STEM-Business, Social Science-Law, Social Science-Counseling/Education). Participants sorted stimuli faster in “matched” blocks (where the representative career shares a response key with its major) than in “unmatched” blocks, yielding DID-IAT effects of 0.30–0.36 SDs (p < 0.01) for all five pairs. After controlling for true career frequency with career and major fixed effects, the effect shrinks only slightly to 0.24–0.28 SDs (p < 0.01), confirming that associations are driven by representativeness beyond base rates. At the individual level, a one-SD increase in DID-IAT scores predicts 4.1 p.p. greater stereotyped beliefs (p < 0.01) without career-by-major fixed effects and 2.8 p.p. (p < 0.01) with them.

Q: What does the role-model heterogeneity analysis show? A: Students were asked which major–career combinations they knew personally. Controlling for career-by-major fixed effects, knowing someone with a non-representative major–career combination (i.e., a non-default path) predicts beliefs about the representative career that are 16 p.p. lower (p < 0.01). This is more than half the size of the baseline stereotyping effect (28–32 p.p.). Knowing such a person also predicts lower IAT scores (p < 0.01), implying that personal exposure can reduce both implicit associations and explicit stereotyped beliefs.

Q: What does the equilibrium model predict about misallocation? A: The model embeds stereotyped beliefs in a two-stage choice framework: students choose a major first, then choose a career after graduation. It shows two main results (Propositions 1 and 2 in Online Appendix A.1). First, students who perceive the representative career as more common than it is will infer — through a rational expectations mechanism — that the unobservable amenities of that career are particularly favorable, so they will be surprised upon graduation. Second, stereotyping raises misallocation because it draws in marginal students whose career preferences make them poorly matched to the major’s representative career, while the inframarginal students who would have chosen the major anyway are better matched. The misallocation effect increases in the extent of stereotyping.

Q: What correlational evidence links stereotyping to post-graduation mismatch outcomes? A: Using major-level stereotyping estimates from the OSU data merged with three nationally representative surveys (NSCG, SIPP, SHED), the authors find: a one-SD increase in major-level stereotyping is associated with 6.0% more job dissatisfaction (p < 0.01, NSCG), 3.1% more reports that the job does not fit the worker’s skills and experience (p < 0.05, NSCG), 5.4% more reports that the job is unrelated to the field of study (p < 0.05, SIPP), and 4.8% more regret about field of study choice (p < 0.05, SHED). The authors note these are correlational and cannot rule out confounders such as underlying complexity of the career mapping.

Q: How does the field experiment work and what is its identifying strategy? A: The experiment was embedded in the second 2021 OSU survey, with students in the treatment group shown the true share of their top major’s representative career before reporting beliefs and intentions; control students answered the same questions without receiving this information. The main regression relates outcomes to (True Share − Prior Belief), set to zero for controls. Because students with less accurate prior beliefs may be more likely to choose the relevant major, OLS is potentially inconsistent; the authors use an RD design where the running variable is the information shock (True Share − Prior Belief), with the threshold at zero. Students just above (who overestimated) receive negative news; students just below (who underestimated) receive positive news. The RD estimates are combined with a first-stage estimate of belief updating to produce IV estimates of the effect of a 10 p.p. change in beliefs. Balance tests on predetermined demographics confirm no discontinuities at the threshold.

Q: What are the first-stage belief-updating results? A: Students update their posterior beliefs in response to the treatment: in response to information that the representative career is 1 p.p. less likely, students update their posterior beliefs down by 0.37 p.p. (p < 0.01). This under-reaction is consistent with Bayesian updating when priors are informative (Mobius et al. 2022). Students also update beliefs about non-representative careers: a 1 p.p. reduction in the representative career’s stated likelihood increases the expected probability of other careers by 0.27 p.p. (p < 0.01).

Q: What are the effects of the information intervention on major intentions? A: A 10 p.p. reduction in beliefs about the top major’s representative career reduces intentions (stated probability of graduating with that major) by 3.5 p.p. (p < 0.01). This effect is similar across subgroups (Columns 2–4 of Table 2). For students’ second-ranked major, a 10 p.p. reduction in stereotyping boosts intentions by 2.1 p.p. (p = 0.17), which is imprecisely estimated but consistent in sign with all other outcomes.

Q: What are the effects on actual course enrollments? A: In the semester immediately following the experiment, learning that the representative career of the first major is 10 p.p. less likely causes students to enroll in 0.22 fewer credits in that major’s field (95% CI: [−0.41, −0.02], p < 0.05), relative to a mean of 0.85 credits. Learning that the representative career of the second major is 10 p.p. less likely causes students to enroll in 0.20 more credits in the second major’s field (95% CI: [0.004, 0.40], p < 0.10), relative to a mean of 0.36 credits.

Q: What are the effects on official major declarations? A: Within one year of the experiment, students who learned the representative career of their top major is 10 p.p. less likely are 6.1 p.p. less likely to have declared that major (95% CI: [−16.0, 3.8], p = 0.23) and 9.9 p.p. more likely to have declared their second major (95% CI: [2.5, 17.4], p < 0.01); the difference between these two effects is 16.0 p.p. (p < 0.01). By two years out, the effects are more attenuated. Treated students also spend on average 0.21 more semesters undecided before declaring a major (95% CI: [0.02, 0.40], p < 0.05). Effects do not appear to be driven by dropout: treated students are if anything slightly more likely to still be taking classes two to three years later.

Representativeness (likelihood ratio): The representativeness R(c, M) of career c for major M is defined as the ratio p_{c|M} / p_{c|not-M} — how much more common career c is among major-M graduates than among graduates of all other majors. This is a relative, not absolute, frequency measure. The representative career (or exemplar) of a major is the career that maximizes this ratio.

Stereotyping (as exaggeration of a kernel of truth): In this paper’s framework, stereotyping means overweighting the representative career when forming beliefs about a major’s career distribution. The belief model is π_{c|M} = (1 − θ) p_{c|M} + θ · 1[c = c*(M)], where θ > 0 implies beliefs exaggerate how common the representative career is relative to equally prevalent non-representative careers. This is distinct from overconfidence, motivated reasoning, or simple noise.

DID-IAT score (difference-in-differences implicit association test): The paper’s adaptation of the standard IAT to measure relative implicit associations between major and career groups. For a focal major–career pair, the DID-IAT score is the difference in the matched-vs-unmatched IAT D-score for the focal major (relative to a comparison major). A positive score indicates the focal major is more strongly associated with the focal career than the comparison major is. This measures implicit memory-based associations rather than deliberate beliefs.

Misallocation (as used in the model): The welfare loss arising because stereotyped beliefs draw marginal students — those on the margin between choosing the representative major and not — who have career preferences close to the average rather than being the students best suited to that major. These marginal students end up choosing careers other than the representative career after graduation at higher rates, producing major-career mismatch. Misallocation is shown (Proposition 2) to increase in the extent of stereotyping θ.

Information shock: In the field experiment, the information shock for a given student and major is the difference between the true share of the major’s representative career and the student’s prior belief about that share. Positive shocks correspond to students who overestimated (and thus receive bad news); negative shocks correspond to students who underestimated (and receive good news). The RD design uses the threshold at shock = 0 to generate quasi-experimental variation.

Source text origin (implicit in the paper’s design): The paper measures beliefs about career distributions benchmarked against American Community Survey data on actual career outcomes of college graduates aged 30–50, restricting to respondents born 1958–1997. This defines the objective ground truth against which stereotyping is measured throughout the paper.

What Works and for Whom? Effectiveness and Efficiency of School Capital Investments Across the U.S.

Mon, 01 Jan 0001 00:00:00 +0000

What Works and for Whom? Effectiveness and Efficiency of School Capital Investments Across the U.S.

Research Question

This paper investigates which types of school facility investments benefit students (as measured by test scores) and are valued by homeowners (as measured by house prices), and for which student populations these investments are most effective. Prior state-level studies had reached conflicting conclusions about the returns to school capital spending, and no nationwide evidence had distinguished impacts across spending categories or student backgrounds.

Data and Methodology

The authors assemble a novel panel dataset covering approximately 14,000 school bond referenda in 29 U.S. states and 10,146 districts enrolling 71% of all U.S. students, for the period 1990–2017. The dataset combines: (1) ballot-level bond election records including vote shares, proposed amounts, and ballot text; (2) district-level test scores from the Stanford Education Data Archive (SEDA) extended backward to 2003 for all states and as early as 1995 for some, normalized to a national scale via NAEP; (3) a Census-tract-level house price index (Contat and Larson, 2022) aggregated to school districts; and (4) NCES district finance and demographic data.

Bond ballot texts are classified into eight spending categories using text-analysis: classroom construction/renovation; HVAC; other infrastructure (plumbing, roofs, furnaces); safety and health (pollutant removal, building safety); STEM equipment and labs; athletic facilities; land purchases; and transportation vehicles.

The identification strategy exploits quasi-random variation from close bond elections, building on the dynamic regression discontinuity (DRD) framework of Cellini et al. (2010). A key methodological contribution is a stacked DRD design that addresses heterogeneous treatment effects correlated with timing: each treatment cohort (districts that narrowly authorize a bond in year c) is matched against “clean controls” — districts that also proposed a bond in the same cohort but narrowly failed to authorize it and did not authorize any bond in the following ten years. Cohorts are stacked, and a dynamic RD model is estimated controlling for cohort fixed effects and a district’s bond proposal history.

Main Findings with Quantitative Magnitudes

Average effects. Bond authorization raises capital spending by approximately $1,650 per pupil cumulatively over five years. Test scores increase gradually, reaching 0.079 standard deviations (sd) higher five to eight years after authorization, and 0.073 sd higher nine to twelve years after. 2SLS estimates, amortizing spending over a 30-year project life at a 9% depreciation rate, imply that a $1,000 increase in the flow value of capital spending raises test scores by 0.048 sd. House prices rise by approximately 9% eight to nine years after authorization. When house price effects are estimated against only locally-financed capital spending (not state aid), the 2SLS estimate is 0.8% per $1,000 — roughly consistent with efficiency — suggesting that the larger reduced-form house price response is driven primarily by state aid that supplements local funds rather than by an inefficiently low ex ante spending level.

Heterogeneity by spending category. Category-specific estimates reveal that only certain project types raise test scores: HVAC (+0.20 sd, largest effect), safety and health (+0.15 sd), other infrastructure/plumbing/roofs (+0.15 sd), STEM equipment (+0.15 sd implied), and classroom space (+0.10 sd), all measured three to six years post-election. By contrast, bonds for athletic facilities, land purchases, and transportation produce no detectable effects on test scores. The pattern for house prices is the inverse: athletic facilities generate a 17% house price increase; classroom space generates 14%; STEM generates 11% — while HVAC and safety/health bonds produce no significant effect on house prices. The correlation between category-level test score and house price estimates is −0.07, indicating these are largely orthogonal outcomes.

Heterogeneity by student socioeconomic status. Effects are concentrated in districts serving socioeconomically disadvantaged students (top tercile of the share of students eligible for free or reduced-price meals, denoted low-SES). In low-SES districts, bond authorization raises test scores by 0.13 sd after seven years and house prices by 15%; in high-SES districts, neither outcome shows a significant effect. 2SLS estimates confirm that a $1,000 increase in cumulative spending raises test scores by 0.08 sd in low-SES districts but produces no detectable change in high-SES districts. The SES gradient persists after conditioning on spending amounts, spending categories, and baseline capital stock, indicating that students in disadvantaged districts have higher marginal returns to capital improvements independent of these channels. High-minority districts (top tercile of Black and Hispanic share) similarly see a 0.12 sd test score gain and 15% house price gain after seven years, versus 0.04 sd and 3% in low-minority districts.

Role of baseline capital stock. Among districts with below-median capital stock, test score effects are 0.20 sd in low-SES districts seven years post-election. Even among above-median-stock districts, low-SES districts see house price effects exceeding 10% while high-SES districts see no effect. Differences by SES persist after conditioning on capital stock.

Policy simulation. Closing the spending gap between high- and low-SES districts (approximately $1,000 over 10 years) without changing the composition of spending would raise low-SES test scores by roughly 0.08 sd, closing about 8% of the roughly 1 sd achievement gap. Targeting that same additional spending toward HVAC and safety/health (the highest-impact categories) would generate test score increases approximately three times as large, potentially closing up to 25% of the observed achievement gap.

Reconciling prior literature. Replicating state-level estimates, the authors show that Ohio’s positive effects are explained by a high share of bonds in low-SES districts funding infrastructure, while Texas’s near-zero effects reflect a high share of bonds in higher-SES districts funding classrooms and athletic facilities.

In depth

Q1. What is the first-stage effect of bond authorization on capital spending, and does it contaminate other spending categories?

A1: Bond authorization raises per-pupil capital spending by approximately $700 per year at two years post-election and $590 at three years, with cumulative spending $1,650 higher over five years in treated districts relative to districts that narrowly failed to authorize a bond. Bond revenues are legally restricted to capital uses, and the paper confirms that non-capital (current) spending and instructional spending are not affected following authorization. This establishes a clean first stage: bond authorization raises only capital outlays.

A2: The original CFR estimator assumes treatment effects are uncorrelated with the timing of treatment — an assumption potentially violated when, for example, bonds financing HVAC (high-impact) versus athletic facilities (amenity-focused) have different propensities to be proposed at different points in time. The stacked DRD design avoids “forbidden comparisons” by comparing each treatment cohort only against clean controls that propose but fail to authorize a bond in the same year and do not authorize any bond in the subsequent ten years. This ensures consistency even when treatment effects are heterogeneous across cohorts and correlated with timing.

Q3. How do the authors validate the quasi-random assignment assumption of the regression discontinuity design?

A3: Three tests are performed. First, a McCrary (2008) density test on the vote margin distribution shows no discontinuity at the cutoff in the pooled or stacked data (p-values of 0.59 and 0.24, respectively), though discontinuities are found in Arkansas, Missouri, and Oklahoma — those three states are excluded. Second, pre-election district covariates (income, education, SES shares, enrollment, revenues, expenditures) are smooth around the cutoff in both datasets. Third, pre-election trends in test scores and house prices are flat and parallel between marginally approved and marginally rejected districts.

Q4. How are the eight spending categories constructed, and how many bonds are successfully classified?

A4: Categories are drawn from the SchoolBondFinder.com classification produced by The Amos Group, then refined by splitting capital improvements into HVAC versus other infrastructure, splitting construction/renovation into classroom versus athletic facility projects, and adding land purchases as a separate category. Keyword-based text analysis of ballot language successfully assigns 75% of the approximately 14,000 bonds to at least one of the eight categories. More than two-thirds of classified bonds receive multiple category designations, with a mean of 2.9 categories per proposed bond and 3.2 per authorized bond.

Q5. Why do HVAC bonds raise test scores but not house prices, while athletic facility bonds raise house prices but not test scores?

A5: The authors interpret this divergence as reflecting what different types of improvements offer to different stakeholders. HVAC improvements reduce excessive heat and air pollution exposure in classrooms, directly improving students’ learning experiences — consistent with Park et al. (2020) on heat and Gilraine and Zheng (2022) on air pollution. These improvements are not visibly salient to homeowners without school-age children and carry no amenity value for the broader community. Athletic facilities, by contrast, are highly visible and provide a community amenity valued in the housing market regardless of their impact on academic instruction. The near-zero correlation (−0.07) between category-level test score and house price estimates confirms that the two outcomes respond to largely distinct features of capital investments.

Q6. What are the three candidate explanations for the larger effects of bond authorization in low-SES districts, and which explanations survive empirical scrutiny?

A6: The three candidates are: (1) larger spending increases after authorization in low-SES districts; (2) a different composition of spending categories (more toward high-impact HVAC and safety); and (3) higher marginal returns per dollar for disadvantaged students, holding spending size and composition fixed. The data confirm all three operate, but the third is the residual: 2SLS estimates show a $1,000 increase raises test scores by 0.08 sd in low-SES districts versus a statistically zero effect in high-SES districts, and within-category estimates show HVAC bonds raise scores by 0.27 sd in low-SES districts but have no detectable effect in high-SES districts. Differences by SES also persist after conditioning on the estimated baseline capital stock, though low capital stock accounts for part of the gap.

Q7. How does the role of state aid alter the interpretation of the house price effect for spending efficiency?

A7: A 9% house price increase after bond authorization, if taken at face value under Brueckner’s (1979) efficiency test, would suggest the ex ante level of school capital spending was inefficiently low. However, state grants that partly match local bond revenues raise actual spending without raising local property taxes proportionally. When the 2SLS house price effect is estimated against only locally financed capital spending (using proposed bond size as the relevant measure), the implied house price increase is just 0.8% per $1,000 — consistent with rough efficiency on average across the full sample. The authors conclude that the large reduced-form house price response is driven primarily by the capitalization of state aid, not by an undersupply of capital investments at the aggregate level.

Q8. Does household sorting account for the observed test score and house price gains following bond authorization?

A8: Bond authorization produces small but detectable compositional changes: the share of high-SES students is approximately 3 percentage points higher seven years after an election (a roughly 4% increase relative to an average share of 0.73), while enrollment and the share of white students are largely unaffected. However, controlling for district-by-year shares of each sociodemographic group only slightly attenuates the test score and house price estimates, indicating that sorting accounts for a small share of the observed gains.

Q9. Are the findings robust to alternative research designs?

A9: The results are robust to five alternative estimation approaches: (1) the original one-step TOT estimator of Cellini et al. (2010); (2) a version of the stacked DRD where clean controls are districts that do not approve any bonds in the full [c−5, c+10] window; (3) a version that matches treated and control districts in each cohort based on bond history; (4) a version not controlling for future bond history; and (5) the extended two-way fixed effects (ETWFE) estimator of Wooldridge (2021). Results are also robust to linear polynomials with different slopes and quadratic polynomials of the vote margin.

Q10. How does the capital stock measure illuminate mechanism, and what are its limitations?

A10: The authors construct a district-level capital stock as the 30-year depreciated sum of capital spending from Census of Governments data (1967–2017) at a 5% depreciation rate. This stock is negatively correlated with the share of low-SES students, confirming that more disadvantaged students attend schools in worse structural condition. Conditioning on this proxy, the SES gradient in bond impacts is reduced but remains. Among districts with below-median capital stock, low-SES districts see test score gains of 0.20 sd after seven years, while among above-median-stock districts the gap narrows to approximately 0.10 vs. 0.05 sd. A key limitation is that detailed school-condition data are unavailable nationally, so the capital stock is a proxy only.

Q11. What is the quantitative policy implication of the targeting exercise?

A11: On average, low-SES districts receive about $97 per pupil per year less in capital spending than high-SES districts, so closing this gap over ten years implies approximately $970 in additional cumulative spending. Without changing spending composition, this would raise test scores by roughly 0.08 sd in low-SES districts, closing about 8% of the approximately 1 sd achievement gap between high- and low-SES districts. Redirecting that same additional spending toward the highest-impact categories (HVAC and safety/health) would generate test score gains roughly three times larger, potentially closing up to 25% of the observed achievement gap.

Q12. How do the cross-state differences documented in prior literature map onto the paper’s heterogeneity findings?

A12: The authors replicate earlier state-level estimates and show that Ohio’s relatively large positive effects — found by Conlin and Thompson (2017) — are explained by a high concentration of bonds in low-SES districts funding infrastructure, while Texas’s near-zero effects — found by Martorell et al. (2016) — reflect a high share of bonds in higher-SES districts funding classrooms and athletic facilities. Wisconsin and Michigan, which showed null effects in earlier studies, similarly have bond compositions and student demographics that predict small impacts under the paper’s heterogeneity framework.

Key Concepts

Stacked Dynamic Regression Discontinuity (Stacked DRD). The paper’s primary estimation strategy, which combines the dynamic RD framework of Cellini et al. (2010) with a stacked-cohort design adapted from the staggered difference-in-differences literature. For each treatment cohort (year in which a bond barely passes), “clean controls” are defined as districts that also proposed a bond in the same year but narrowly failed to authorize it and did not authorize any subsequent bond within ten years. Cohort-specific datasets are stacked and estimated jointly with cohort fixed effects, ensuring that estimates are robust to treatment effect heterogeneity correlated with timing.

Clean Controls. Districts used as the counterfactual for treated districts in a given cohort: those that propose a bond in the same year as the treated cohort, barely fail to authorize it, and remain untreated for ten subsequent years. Their “clean” status is quasi-random because their future non-authorization results from narrow electoral loss rather than any endogenous district choice.

Bond Spending Categories. Eight mutually-non-exclusive classifications of bond spending derived from ballot text using keyword analysis: classroom space; HVAC; other infrastructure (plumbing, roofs, furnaces); safety and health (pollutant removal, compliance upgrades); STEM equipment and labs; athletic facilities; land purchases; and transportation. These categories are defined in the paper not by administrative accounting codes but by the stated intended use of funds in ballot language.

Treatment-on-the-Treated (TOT) Estimator. The CFR estimator that captures the effect of bond authorization against the counterfactual of never authorizing a bond in the foreseeable future, achieved by including leads and lags of a district’s bond proposal history as controls. This addresses the problem that multiple elections over time make simple treated-vs-control comparisons confounded by past and future bond activity.

Capital Stock (District-Level Proxy). A measure of each district’s accumulated school facility capital at a given point in time, constructed as the depreciated 30-year running sum of capital expenditures from the Census of Governments, using a 5% annual depreciation rate. Used as a proxy for facility conditions in the absence of nationally available building-quality data, and confirmed to be negatively correlated with district share of low-SES students.

Brueckner Efficiency Test. An application of the theoretical framework linking public good provision levels to house price responses. If a spending increase raises house prices, the initial spending level was below the efficient level; if it lowers house prices, spending was too high. In this paper, the test is refined to use only locally-financed capital spending as the explanatory variable, to strip out the capitalization of state aid and isolate the efficiency assessment for locally-determined spending.

Socio-Economic Status (SES) Terciles. Districts are ranked by the share of students eligible for free or reduced-price school meals as of 1995. “Low-SES districts” refers to those in the top tercile of this share (most disadvantaged); “high-SES districts” refers to those in the bottom tercile (least disadvantaged). Effects are estimated separately for these subsamples throughout.

When Did Growth Begin? New Estimates of Productivity Growth in England from 1250 to 1870

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Research Question. When did sustained productivity growth begin in England? This paper constructs new estimates of the evolution of productivity in England from 1250 to 1870, with the goal of both dating the onset of growth and using that dating to discriminate between competing theories of why growth began.

Methodological Innovation. The core challenge is that real wages over this period were heavily distorted by Malthusian population dynamics. Plague-induced population collapses (most dramatically the Black Death of 1348, which killed roughly 25% of England’s population) drove enormous swings in real wages that reflect movements along a stable labor demand curve, not changes in productivity. A naive regression of wages on labor supply is therefore inconsistent, because in a Malthusian world productivity growth induces population growth, making labor supply endogenous to productivity.

The authors address this by writing down and structurally estimating a full Malthusian model of the economy. Output is produced with fixed land and variable labor (and, in an extended model, capital) via a Cobb-Douglas production function. The labor demand curve equates the real wage to the marginal product of labor. Population growth is increasing in real per-capita income (the Malthus law of motion), capturing both preventive and positive checks. Productivity follows a random walk with drift, and the paper allows for two structural breaks in the average drift rate mu. Exogenous population shocks, modeled as infrequent, sizable plague draws from a beta distribution plus a Gaussian noise term, provide identification: plague shocks and productivity shocks generate observationally distinct dynamics – plague shocks cause an immediate population drop that gradually reverts, while productivity shocks cause an immediate wage jump followed by a slow population rise to a new steady state. The model is estimated via Bayesian Hamiltonian Monte Carlo (Stan), and structural break dates for mu are chosen by maximizing the Bayes factor (marginal likelihood) over the observed data on real wages, population, and days worked per worker.

Key Data. Real wages are from Clark (2010) unskilled building workers series. Post-1540 population is from Wrigley et al. (1997); pre-1540 population trends are from Clark (2007b) manorial records. Days worked per worker are from Humphries and Weisdorf (2019). All series are used as decadal averages.

Main Findings.

Onset of growth: 1600. Productivity growth was zero before 1600. The Bayes factor strongly favors a first structural break in mu at 1600; break dates before 1590 and after 1640 are clearly rejected.
Two-phase post-1600 growth. Between 1600 and 1810, average productivity growth was 4% per decade (posterior mean; 95% credible interval approximately 2%-6%). After 1810, productivity growth accelerated sharply to 18% per decade (95% CI approximately 12%-23%). The second break date is estimated to 1810 (the only alternative not clearly rejected is 1800).
Magnitude of productivity change. By the authors’ estimates, productivity in England was approximately 540% higher in 1850 than in 1500. This contrasts sharply with Clark’s (2010) dual-approach TFP series, which implies essentially no change over this period. The authors attribute the discrepancy to mismeasurement in Clark’s land rent series.
Productivity growth preceded the Glorious Revolution. Productivity rose by an estimated 48% between 1600 and 1680, well before the Glorious Revolution of 1688 and the English Civil War (1642-1651). This supports the view that economic change contributed to causing the bourgeois institutional reforms of the 17th century, consistent with the Marxist tradition (Hill, 1940, 1961), rather than that institutional change preceded and caused growth.
Weakness of Malthusian population force. The elasticity of population growth with respect to real income (gamma) is estimated at 0.09. Combined with a slope of the labor demand curve (alpha) of 0.53, this implies a half-life of plague-induced population dynamics of approximately 150 years. A doubling of real per-capita income stimulated population growth by only 6 percentage points per decade – indicating Malthusian forces were sufficiently weak to be overwhelmed by post-1800 productivity growth. The model implies that the post-1810 productivity growth rate would have produced a 28-fold long-run increase in steady-state real wages even without the Demographic Transition.
Capital extension. When capital is explicitly incorporated, using rates of return on agricultural land and rent charges to infer the capital stock, results are broadly similar: productivity growth from 1600-1810 is 3% per decade and post-1810 is 14% per decade. Capital’s production function exponent is estimated at 0.18, confirming that capital accumulation explains only a modest share of growth.

Scope Conditions. All estimates are for England specifically. The model assumes competitive factor markets, a Cobb-Douglas (or CES) production function, and a log-linear Malthusian population law of motion. Results are robust to alternative wage series (farm laborers, craftsmen, Allen’s series), alternative population sources (Broadberry et al., 2015), constant-days-worked assumption, and alternative prior distributions.

In depth

Q1. Why can’t standard OLS regression of wages on labor supply recover productivity in this setting?

In a Malthusian world, productivity growth causes population growth, which in turn raises labor supply. This means labor supply and productivity are positively correlated, biasing OLS estimates. The authors demonstrate this concretely: from 1300 to 1450 (plague era), wages and labor supply moved in opposite directions along a stable labor demand curve, while after 1630 the same data points begin shifting off that curve – a pattern that OLS would confound with changes in the slope rather than shifts in the intercept.

Q2. How do the authors distinguish empirically between a plague shock and a productivity shock?

The two shocks generate fundamentally different dynamics. A plague shock causes an immediate, large drop in population and a corresponding spike in wages; over time, high wages induce population growth and both wages and population gradually return to their pre-plague levels. A permanent productivity shock, by contrast, causes an immediate rise in wages with no contemporaneous population change; population then slowly rises and wages partially revert until a new, higher steady-state population is reached. The model exploits these different impulse-response signatures in the joint data on wages and population to identify the two shocks separately.

Q3. What is the Bayes factor evidence for the 1600 break date?

Figure 8 in the paper shows the Bayes factor for models with different first break dates (all holding the second break at 1810). The Bayes factor rises sharply from 1580 to 1600 and falls more gradually from 1600 to 1650. Break dates before 1590 and after 1640 are clearly rejected using the standard rule of thumb that a Bayes factor of 10 constitutes strong evidence. The 1600-1810 pair of break dates yields the highest marginal likelihood of any combination considered.

Q4. How does the paper’s productivity estimate compare to Clark’s (2010) dual-approach TFP series?

Clark’s series implies productivity in England was essentially unchanged between the 15th and mid-19th centuries – a result the paper argues is implausible and inconsistent with Allen’s (2005) agricultural TFP estimates (which show a 162% increase in agricultural TFP between 1500 and 1850). The authors’ baseline estimate implies productivity was approximately 540% higher in 1850 than in 1500. The authors conjecture that a key driver of the difference is mismeasurement in Clark’s land rent series, which appears essentially flat from 1250 to 1600 despite enormous plague-induced swings in the land-labor ratio over this period.

Q5. What does the Malthusian model imply about “Engel’s Pause” – the apparent stagnation of real wages during early industrialization?

Between 1730 and 1800, real wages fell slightly despite what the model estimates to be substantial productivity growth. The conventional explanation is that the gains from early industrialization accrued to capitalists rather than workers. The authors offer an alternative Malthusian explanation: England’s population grew rapidly over this period, and in the Malthusian model this population growth depressed wages relative to productivity. The authors do not reject the distributional explanation but show that Malthusian forces alone are sufficient to explain the wage-productivity divergence.

Q6. How quantitatively important are days worked (the Industrious Revolution) for the productivity estimates?

The authors find that their productivity estimates are largely insensitive to whether the Humphries-Weisdorf (2019) days-worked series or a constant-days assumption is used. The qualitative pattern – zero growth before 1600, modest growth 1600-1810, rapid acceleration post-1810 – and the quantitative magnitudes remain similar. What does change is the estimated slope of the labor demand curve alpha: assuming constant days makes the labor demand curve steeper. This robustness is reassuring given that the Industrious Revolution is a contested empirical phenomenon.

Q7. What does the model imply about the speed of Malthusian population dynamics, and how does this compare to prior estimates?

The estimated elasticity of population growth to real income gamma = 0.09, combined with alpha = 0.53, implies a half-life of population dynamics of approximately 150 years. This is consistent with but lies between prior structural estimates: Lee and Anderson (2002) find a half-life of 107 years, and Crafts and Mills (2009) find 431 years. All estimates agree that Malthusian dynamics in England were slow relative to the conceptual ideal of rapid subsistence convergence.

Q8. Can the model explain the post-1750 population explosion without invoking the Demographic Transition?

Yes. The authors simulate predicted population paths from 1740 to 1860 taking real wages and days worked as given and using their estimated gamma and alpha. Despite the weak Malthusian population force, the model can explain the vast majority of the observed population growth from 6 million in 1740 to nearly 20 million in 1860 (10.4% per decade). The key mechanism is that days worked increased substantially over this period, raising per-capita income well above what real wages alone would suggest.

Q9. How does incorporating capital change the productivity estimates?

In the capital-augmented model, the capital stock is inferred from rates of return on agricultural land and rent charges (Clark 2002, 2010). The capital exponent beta is estimated at 0.18, indicating a modest role for capital in pre-industrial England. Average productivity growth from 1600-1810 falls from 4% to 3% per decade, and post-1810 growth falls from 18% to 14% per decade. The authors conclude that the vast majority of growth from 1600 to 1870 cannot be attributed to capital accumulation. From 1600 to 1860, the estimated capital stock grew by a factor of five (8% per decade).

Q10. What theories of the onset of growth are consistent vs. inconsistent with the authors’ timing evidence?

Inconsistent: The North-Weingast (1989) view that the Glorious Revolution of 1688 was the key institutional trigger, since productivity had already risen 48% between 1600 and 1680. Also inconsistent: gradual-growth theories (Kremer 1993, Galor-Weil 2000) in which there is no discrete acceleration. Consistent: Marxist accounts (Hill 1940, 1961) that economic change drove 17th-century institutional change; Acemoglu-Johnson-Robinson (2005) accounts linking Atlantic trade enrichment to the demand for secure property rights (timing broadly consistent, though growth rates do not visibly accelerate after the Civil War or Glorious Revolution); cultural-change accounts (Mokyr, McCloskey) tracing the onset of growth to the spread of literacy and scientific rationalism around 1600; Allen’s (2009a) directed-technical-change theory linking 17th-century wage growth to the later profitability of labor-saving innovation in the Industrial Revolution.

Q11. What does the model imply about the long-run real wage consequences of post-1810 productivity growth, even counterfactually assuming Malthusian forces persisted?

The steady-state real wage in the Malthusian model is w-bar = mu/(alpha*gamma) minus subsistence-related terms. For mu = 0.018 (the post-1810 estimate), this formula implies a long-run real wage 28 times higher than the steady state under zero productivity growth. In other words, even if the Demographic Transition had not occurred and birth and death rates had remained sensitive to income, post-1810 productivity growth was fast enough relative to the weak Malthusian force to generate substantial sustained rises in living standards.

Key Concepts

Labor demand curve (in the paper’s sense). The equilibrium relationship between real wages and labor supply derived from competitive profit maximization by landowners facing a fixed land endowment: w_t = phi - alpha*l_t + a_t. Productivity is identified as shifts in this curve across time periods. The slope alpha is not simply the land share under a CES production function but equals one minus the labor share divided by the elasticity of substitution between labor and land.

Malthusian population force. The feedback mechanism by which higher real wages induce faster population growth, expanding labor supply and pushing wages back toward a steady state. Its speed is governed jointly by gamma (elasticity of population growth with respect to income) and alpha (slope of the labor demand curve); the half-life of wage/population dynamics after a shock equals log(0.5)/log(1 - alpha*gamma). In the paper’s estimates, this force was sufficiently weak (half-life approximately 150 years) that post-1800 productivity growth overwhelmed it.

Plague shock (xi_1t). An infrequent, large, exogenous negative population shock modeled as a draw from a beta distribution occurring with probability pi. Plagues are the primary source of identifying variation for the pre-1600 period: they generate movements along a stable labor demand curve and allow the slope alpha and the (lack of) productivity trend to be separately identified from labor demand shifts.

Structural break in average productivity growth (mu). The drift parameter in the random-walk model for the permanent component of productivity. The paper allows two breaks in mu, with break dates chosen to maximize the marginal likelihood (Bayes factor). The best-fitting breaks are at 1600 (zero to 4% per decade) and 1810 (4% to 18% per decade).

Permanent vs. transitory productivity component. Productivity is decomposed into a permanent component a-tilde_t (random walk with drift, sigma_epsilon1) and a transitory component epsilon_2t (iid noise, sigma_epsilon2). The paper reports and interprets the permanent component as the meaningful measure of underlying technological change; transitory shocks are treated as measurement error and short-run fluctuations.

Industrious Revolution. The hypothesized long-run increase in days worked per worker in England, associated with de Vries (1994, 2008). The paper uses Humphries-Weisdorf (2019) estimates showing a sharp drop after the Black Death followed by a sustained rise from 1350 onward. A key robustness result is that the paper’s productivity estimates are insensitive to whether this Industrious Revolution is assumed to have occurred.

Bayes factor (model selection). The ratio of marginal likelihoods p(y|M_t)/p(y|M_t’) for two competing models, used here to select structural break dates for mu. A factor of 10 is treated as strong evidence. The bridge sampling method of Gronau, Singmann, and Wagenmakers (2020) is used to compute marginal likelihoods.

Who's Afraid of the Minimum Wage? Measuring the Impacts on Independent Businesses Using Matched U.S. Tax Returns

Mon, 01 Jan 0001 00:00:00 +0000

Layer 1 — Overview

Research Question

This paper asks how independent (pass-through) businesses in the United States accommodate minimum wage increases — specifically whether they reduce employment, compress profits, pass costs through to customers, or exit — and what happens to the low-earning workers and business owners affected by these adjustments.

Data and Methodology

The authors construct a novel linked firm-worker-owner panel dataset from the universe of U.S. tax returns, covering approximately 235,000 pass-through firms (S-corporations, partnerships, and LLCs) per year in highly exposed industries over 2010–2019. “Highly exposed” industries are defined as those where at least 15% of workers earned below the full-time equivalent of the federal minimum wage ($15,080 per year) in 2013. The dataset links annual business income tax returns to the individual income tax returns and W-2 information reports of all workers and owners.

The causal identification strategy exploits the six state minimum wage increases that took effect in 2014 (California, Connecticut, Delaware, Michigan, Minnesota, and New Jersey) relative to 24 states that did not change their wage floors at any point from 2012–2018. The empirical workhorse is a panel difference-in-differences event study (Equation 1), augmented by DFL re-weighting (DiNardo et al., 1996) to improve comparability of treatment and control firms on observables. The analysis covers cumulative effects through 2018, by which point the average minimum wage across treatment states had risen 30.6%.

Main Findings

Employment: The average exposed independent firm does not meaningfully reduce employment. The authors estimate an own-wage elasticity of -0.209 (s.e. = 0.0112). Employment adjustments manifest as moderately lower hiring rather than layoffs of existing workers. Reduced hiring is wholly concentrated among teenagers and very part-time jobs paying less than $3,900 annually (with 67% earning less than $1,000 per year).
Worker earnings: Despite the hiring reduction, low-earning workers employed at exposed independent firms experience average earnings gains of approximately $2,000 per year by 2018, relative to comparable workers in untreated states. Young individuals aged 20–26 without a 2013 job earn roughly $4,000 more per year by 2018; teenagers without a 2013 job gain approximately $1,000 per year. Workers in these groups are no less likely — and in some cases slightly more likely — to be employed five years after the minimum wage increase.
Wage bills: Average wage bills among surviving treated firms rose 7.03% (s.e. = 0.0153) by 2018. Earnings gains are concentrated among workers earning $15,600–$35,000 annually, with no evidence of reduced earnings for higher-paid workers. The 7% average wage bill increase amounts to only 1.4% of 2013 firm revenues, easing pass-through.
Revenue and profits: Revenues of surviving treated firms grew approximately 2.1% more than control firms by 2018. On average, this revenue increase fully offsets the higher wage bill, yielding a small net profit increase of roughly $3,360 (s.e. = $1,123) per owner by 2018, or about 2.7% of mean 2013 owner income.
Firm exit: On average across all highly exposed industries, minimum wages increased the five-year exit probability by 0.9 percentage points (s.e. = 0.0029), relative to a baseline raw exit rate of approximately 29%. Exit effects are driven entirely by restaurants: by 2018, restaurants in treated states were 1.85 percentage points (s.e. = 0.0039) more likely to have exited, while the exit response for non-restaurant exposed firms is a precisely estimated zero.
Heterogeneity by productivity within restaurants: Exit is concentrated entirely in the bottom productivity quartile (coefficient = 0.0254, s.e. = 0.0079), with no significant effect in the upper three quartiles. Profits among surviving small restaurants rise by $5,941 (s.e. = $1,546) by 2018 relative to 2013. Among small restaurants, the profit gains are larger for firms in the higher productivity quartiles (Q3: +$7,915; Q4: +$9,161). Surviving restaurants also increase non-labor input spending by 2.53% (s.e. = 0.0101), consistent with expanded output following competitor exits.
Entrant characteristics: Post-reform restaurant entrants in treatment states have higher wage bills (13.8% higher in logs), higher revenues (4.0% higher), higher value-added (8.4% higher), and higher productivity (net income/revenue ratio 2.24 percentage points higher) than entrants in control states, indicating the minimum wage raises the productivity floor for new entrants.
Owner outcomes after exit: Owners of small restaurants forced out by the minimum wage are significantly less likely to own an independent business five years later, but earn no less on average in wages plus business income. Policy-induced exiters are significantly less likely to report negative incomes, suggesting substitution away from risky or marginally profitable business ownership.

Theoretical Framework

The authors present a Cournot competition model with heterogeneous firm productivity and fixed production costs. A minimum wage cost shock raises marginal costs, narrowing margins for all firms. Firms whose cost increases exceed the market price increase cannot cover fixed costs and exit. Remaining firms gain higher markups and larger market shares as demand is reallocated from exiting firms. Selection on ex-ante productivity (the least productive firms exit) limits the distortion to market quantity and amplifies profit gains among productive survivors. The model predicts profit increases only in markets with firm exit, which matches the data: profits rise among restaurants (where exit occurs) but not among retailers (where exit is a precisely estimated zero).

Scope Conditions

Findings pertain to the short-to-medium run (up to five years post-legislation) of phased-in minimum wage increases averaging 30.6% in six U.S. states. The sample covers pass-through (independent) businesses in highly exposed industries. Longer-run effects may differ if entrants adopt production technologies that rely less on low-wage labor or incumbents reconfigure inputs. Border-county retailers appear to be less able to pass through costs than interior firms, suggesting product market competition is a key moderating factor.

In depth

Q1. Why do the authors focus on pass-through businesses rather than publicly traded corporations?

Pass-throughs (S-corporations, partnerships, and LLCs) comprise 78% of non-sole-proprietorship businesses and 79% of firms with fewer than 20 employees. They represent the majority organizational form for independent businesses in virtually all two-digit NAICS industry groups except utilities and enterprise management. Because minimum wage concerns are disproportionately raised on behalf of small independent businesses, and because most minimum wage workers in restaurants are employed at pass-throughs, studying pass-throughs directly addresses the policy debate. Additionally, pass-through tax returns link business income directly to the individual tax returns of each owner, enabling the authors to separately identify employee versus owner responses.

Q2. How do the authors define “highly exposed” industries and why does this matter for identification?

Highly exposed industries are defined as four-digit NAICS industries where at least 15% of workers earned below the full-time federal minimum wage equivalent ($15,080 per year) in 2013, using tax data to construct a proxy for minimum wage workers. The analysis focuses on these industries because minimum wage workers are extremely concentrated — the vast majority are in Leisure/Hospitality and Retail. Restricting to highly exposed industries allows the authors to estimate average effects within affected markets and conduct heterogeneity analysis across firm characteristics within those markets, including comparing firms with different baseline shares of low-earning workers that nonetheless all face the market-level cost shock.

Q3. How do the employment effects decompose into hiring versus retention?

The average firm subject to a higher wage floor does not lay off existing workers (the retention line is flat in event study estimates). By 2018, firms in treated states hire roughly one fewer worker on average than similar firms in control states, entirely through reduced hiring. This reduced hiring is wholly concentrated among teenagers in very part-time jobs: the missing hires consist entirely of workers who would have earned less than $3,900 annually, with 67% earning less than $1,000 per year. Simultaneously, workers already employed at exposed firms are 2 to 4 percentage points more likely to remain with their 2013 employer by 2016, with prime-age low-earning workers exhibiting the largest retention increases.

Q4. What happens to low-earning workers and young people in individual-level panels?

Low-earners (those earning below $25,000 in each year from 2012–2014) at exposed independent firms experience average earnings gains of approximately $2,000 per year by 2018 relative to similar workers in untreated states, including teenage low-earners. Young individuals aged 20–26 with no job in 2013 experience a relative earnings increase of approximately $4,000 per year by 2018; teenagers without jobs in 2013 gain approximately $1,000 per year. These workers are no less likely — and often slightly more likely — to be employed relative to their counterparts in control states, so the earnings gains are not offset by employment losses at the individual level.

Q5. What is the magnitude of the cost shock for firms and how does it compare to revenues?

By 2018, the average wage bill among surviving firms in treated states was 7.03% (s.e. = 0.0153) higher than comparable firms in control states. This is consistent with a back-of-envelope calculation: low-earning workers account for about 21% of wage bills at these firms, and states raised minimum wages by 30.6% on average (0.21 × 0.306 = 0.064). However, the 7% wage bill increase amounts to only approximately 1.4% of 2013 firm revenues, making cost pass-through relatively modest. Higher minimum wages have no discernible impact on pension contributions but slightly reduce deductions for other benefits including health insurance.

Q6. How do surviving firms finance the increased wage bill, and what happens to profits?

Surviving firms finance the wage increase primarily through higher revenues. By 2018, revenues of firms in treated states grew approximately 2.1% more than revenues of firms in control states. On average, this revenue increase outpaces the higher wage bill, resulting in a net profit increase of approximately $3,360 (s.e. = $1,123) per owner by 2018, representing about 2.7% of mean 2013 owner income. There is no evidence of redistribution from middle- or high-income workers within firms; wage bill increases are concentrated among workers earning $15,600–$35,000 annually, consistent with minimum wage spillovers to workers slightly above the statutory floor.

Q7. Why do restaurants experience exit effects but retailers do not?

The asymmetry stems from the intensity of low-wage labor in production. While low-earning workers account for a similar share of labor costs at restaurants (41.8%) and retailers (38.5%), labor costs overall are more than twice as large at restaurants relative to retailers. Wage bills account for 39% of variable costs and 27% of revenues at restaurants, but only 16% of variable costs and 13% of revenues at retailers. As a result, raising the minimum wage raises variable costs by 5.76% at restaurants. Non-restaurant exposed firms are able to fully pass through their smaller cost shock, yielding flat profits and neither employment nor exit impacts.

Q8. Why is firm exit concentrated in the lowest productivity quartile of restaurants rather than among the most exposed firms?

The Cournot framework predicts exits among firms with the lowest ex-ante productivity (highest marginal costs), the largest cost shock (highest share of low-wage labor per unit of output), or a combination. Empirically, productivity is the primary determinant: restaurants across all productivity quartiles use similar shares of low-earning workers (40–44% of wage bills for Q1 through Q4). Exit rises significantly only among restaurants in the bottom productivity quartile (coefficient = 0.0254, s.e. = 0.0079), with no significant effects in Q2–Q4. Among the lowest-productivity restaurants, those most dependent on low-earning labor face the largest exit rates.

Q9. How do the model’s predictions about profit heterogeneity match the data?

The Cournot model predicts profits should rise only in markets with firm exit (via increased margins and market share reallocation to survivors). This is exactly what the data show. Among restaurants, where exit is concentrated in the bottom productivity quartile, profits among surviving small restaurants rise by $5,941 (s.e. = $1,546) by 2018. Among small restaurants specifically, profit gains increase with productivity: Q3 restaurants gain $7,915 (s.e. = $3,326) and Q4 restaurants gain $9,161 (s.e. = $2,127), while Q1 and Q2 gains are statistically indistinguishable from zero. In non-restaurant exposed industries where the exit effect is a precise zero, profits are also flat — exactly as the model predicts.

Q10. What happens to the characteristics of new restaurant entrants after the minimum wage increase?

Post-reform restaurant entrants in treatment states are systematically more productive than entrants in control states. They have wage bills 13.8% higher (in logs), revenues 4.0% higher, value-added 8.4% higher, and productivity ratios (net income/revenue) 2.24 percentage points higher than new entrants in control markets. This implies the minimum wage raises the minimum viable productivity threshold for entrant restaurants, consistent with Sorkin (2015)’s insight that minimum wages shape the capital and technology choices of entering firms. The restaurant industry thus becomes more productive on average through both the exit of the least productive incumbents and the entry of more productive new firms.

Q11. How do worker transition patterns reflect the reallocation of output to surviving firms?

Workers at large independent businesses (top revenue quartile) are 3.52 percentage points more likely to remain with their 2013 employer in 2018 and 2.36 percentage points less likely to switch to another large firm. The large firms that retain more of their existing workforce also reduce their hiring of very part-time teenagers the most — in the top revenue quartile, firms shed roughly 4.5 employment relationships on average, comprising higher retention of 4.15 existing workers offset by reduced hiring of 8.67 very part-time teenage workers. Workers originally at smaller exposed firms are more likely to be found working at larger firms five years out, consistent with demand reallocation from exiting and shrinking small firms toward larger, more productive survivors.

Q12. What happens to owners of restaurants that exit due to the minimum wage?

Policy-induced exiters of small restaurants are significantly less likely to own an independent business five years later and less likely to receive all earnings from business ownership, relative to owners of restaurants that exited for other reasons in control states. However, their average incomes (wage income plus ordinary business income) are no lower. This income stability is partly explained by the fact that policy-induced exiters are significantly less likely to report negative incomes five years out, suggesting they substitute away from potentially risky or marginally profitable business ownership toward wage employment or other activities. The utility implications are ambiguous: these former owners may have preferred business ownership even if it did not yield higher income.

Q13. What is the role of product market competition in mediating pass-through, as evidenced by border-county analysis?

The border county robustness analysis reveals that product market competition is central to pass-through success. Retailers near state borders, where consumers can cross-state-border shop, face more elastic demand and are less able to finance the wage cost shock with new revenues, exhibiting reduced profits and higher exit rates (though estimates are imprecise). Further from the border, where the cost shock is more commonly felt by all potential substitutes (making market demand elasticity rather than firm demand elasticity the relevant parameter), results are very similar to the full-sample aggregate findings. This confirms that the common nature of the minimum wage cost shock — shared by all competing firms in the market — is a key reason firms can pass through costs to consumers.

Q14. How do the findings address the divide among independent business owners on minimum wage policy?

The heterogeneous outcomes rationalize why surveys consistently find business owners divided. Among restaurants, some owners (those operating the least productive small restaurants) face exit and loss of business ownership, while surviving productive restaurateurs see higher profits of $5,941–$9,161 per year. Among non-restaurant exposed businesses, owners are broadly unaffected in terms of profits and viability. Uncertainty about whether a given firm’s demand is elastic enough to bear cost pass-through — given that owners may be more familiar with the elasticity of firm-level demand from prior unilateral price changes, rather than the relevant market-level demand elasticity applying to a common cost shock — may broaden opposition to include even owners who would ultimately benefit.

Key Concepts

Pass-through businesses (independent businesses): Privately owned firms organized as S-corporations, partnerships, or LLCs, taxed by passing income through to the individual returns of owners rather than at the entity level. In 2015, these comprised 78% of non-sole-proprietorship U.S. businesses and 46% of employment. The paper uses “pass-through” and “independent business” interchangeably as the unit of analysis.

Highly exposed industries: Four-digit NAICS industries where at least 15% of workers earned below the annual full-time equivalent of the federal minimum wage ($15,080) in 2013, as measured in the authors’ administrative tax data. This threshold proxies the concentration of minimum-wage workers across industries and drives the sample selection for firm-level analysis.

Own-wage elasticity of employment: The estimated percentage change in employment at a firm associated with a given percentage change in the firm’s minimum wage. The authors estimate this as -0.209 (s.e. = 0.0112), reflecting the average effect across all exposed independent businesses, conditional on the firm’s industry, size, and local market characteristics.

DFL re-weighting (DiNardo-Fortin-Lemieux): A non-parametric reweighting procedure that adjusts the distribution of control-group firms to match the distribution of treatment-group firms on observables (specifically, two-year lagged value-added within three-digit NAICS industries). Used to improve pre-reform comparability of treatment and control firm samples without parametric functional form assumptions.

Firm productivity (in this paper’s sense): Measured as the ratio of net profits to revenues (net income/revenue) at the firm level in the base year 2013, used to assign firms to productivity quartiles for heterogeneity analysis. This is a firm-level profitability measure constructed from pass-through tax returns, not a total factor productivity estimate requiring production function estimation.

Firm exit: An indicator for a firm that filed a tax return in 2013 but did not file a return in a subsequent year t. The average one-year exit rate for highly exposed independent businesses is 5.2%; the cumulative five-year raw exit rate is approximately 29% across treatment and control states.

Cournot competition with heterogeneous productivity and fixed costs: The paper’s conceptual framework, in which N firms compete in quantities with asymmetric marginal costs (reflecting heterogeneous productivity), a common output price, and a fixed cost of production. Under this framework, a minimum wage cost shock narrows margins unevenly, induces exit among firms that cannot cover fixed costs, and generates both demand reallocation and market share gains for productive survivors — rationalizing simultaneous exit and profit increases in the same industry.

Common cost shock: The property that a minimum wage increase raises production costs for all firms employing low-wage workers in the same market simultaneously. Because all competing firms face higher costs, the relevant pass-through parameter is the elasticity of market demand rather than the (higher) elasticity of individual firm demand, facilitating cost pass-through to consumers and distinguishing minimum wages from unilateral price changes by a single firm.

Why Doesn't the United States Have National Health Insurance?

Mon, 01 Jan 0001 00:00:00 +0000

This paper investigates a critical juncture in the development of national health insurance (NHI) in the United States: the post-World War II period when most peer nations moved to establish comprehensive public coverage while the U.S. did not. The authors examine the causal role of the American Medical Association (AMA), which in 1949 hired Whitaker & Baxter’s Campaigns, Inc. — the country’s first political public relations firm — to direct a nationwide campaign opposing NHI and promoting private (voluntary) health insurance (PHI).

The Campaign had two main components. First, a physician outreach component in which AMA members distributed pamphlets to patients warning against “socialized medicine” and encouraging enrollment in private plans, and acted as liaisons to local civic organizations to solicit resolutions against NHI sent to elected officials (nearly 50 million pieces of material were sent to physicians). Second, a mass newspaper advertising component, in which a standard ad was placed across newspapers nationwide, with an additional $19 million (approximately $240 million in current dollars) in coordinated tie-in advertising from roughly 23,000 corporations and industry associations. The messaging framed NHI as “un-American” and associated private insurance with “freedom” and “the American way,” providing little substantive information about insurance products.

The authors construct novel measures of Campaign exposure by combining (a) per capita pamphlets distributed by AMA physicians and (b) per capita advertising circulation scaled by local newspaper readership, using archival data from the Whitaker & Baxter Archives (Sacramento), the National Archives (Washington D.C.), digitized AMA Medical Directories, the N.W. Ayer & Son’s Newspaper Directory, and newly discovered Blue Shield enrollment data from AMA Council on Medical Service annual reports covering 1946–1954.

The primary estimation strategy exploits spatial variation in Campaign intensity combined with its timing, using event studies with state and year fixed effects and design controls for income per capita and unionization. The identifying assumption — that Campaign intensity was conditionally as-good-as-randomly assigned — is supported by balance tests showing no pre-Campaign correlation between exposure and enrollment or sociodemographic characteristics (with the exception of Black population share), and by the historical record that the Campaign was organized hastily following Truman’s unexpected 1948 electoral victory.

Main findings: A one standard deviation increase in Campaign exposure explains approximately 20% of the post-Campaign increase in PHI enrollment, corresponding to roughly 14 million additional enrollees — an effect comparable in magnitude to increasing average per capita income by approximately $100 (about 7 percent). On public opinion, a one standard deviation increase in Campaign exposure led to a six percentage point decline in popular support for NHI per Gallup survey wave, a reversal occurring against a backdrop of 69% pre-Campaign approval that was trending upward. For context, this six-point magnitude approximates the entire gap in NHI support between union and non-union households, or one-third the racial gap in support. Campaign intensity also predicts civic organizations passing resolutions favoring PHI, Republican legislators adopting speech semantically similar to Campaign propaganda, and — by 1952 — AMA members being five times more likely to donate to the Eisenhower-Nixon ticket than non-AMA physicians, with donation rates increasing in Campaign intensity.

Scope conditions: The analysis covers 48 U.S. states from 1946 to 1954, ending at the 1954 IRS tax code change that expanded commercial insurers’ market share. The enrollment data capture Blue Shield (physician-run) plans specifically; the paper explicitly notes that commercial insurer granular data are unavailable for the main Campaign period. The authors argue that multiple subsequent factors — middle-class acquisition of private coverage reducing demand for a public option, incumbent interests defending the status quo, and the persistent ideological linkage of private insurance with freedom — help explain why NHI was not adopted in subsequent decades, though these persistence mechanisms are outside the paper’s direct empirical scope.

Q: What was the AMA’s Campaign, and what prompted it? A: In response to Harry Truman’s unexpected 1948 presidential victory alongside a Democratic Congress — and with a majority of informed voters favoring NHI — the AMA hired Whitaker & Baxter’s Campaigns, Inc. to run the National Education Campaign (NEC). The Campaign had two components: physician outreach (pamphlet distribution to patients, liaison to civic organizations) and mass newspaper advertising. The AMA paid Whitaker & Baxter approximately $1.2 million per year in current terms, and coordinated an additional $19 million in 1950 dollars (roughly $240 million today) in tie-in advertising from allied corporations and trade groups.

Q: How is Campaign exposure measured, and how is it validated as conditionally exogenous? A: Campaign exposure combines two standardized components: per capita pamphlets distributed by AMA physicians (pamphlet quantity from W&B archives scaled by state AMA membership share) and per capita advertising circulation scaled by local newspaper readership (share of adults with more than five years of schooling). The two components are summed and standardized. Exogeneity is supported by balance tables showing no pre-Campaign correlation between exposure and enrollment or Gallup opinion, by the absence of discontinuous changes in income or unionization at Campaign onset, and by the historical fact that Campaign logistics relied on pre-existing networks assembled hastily in response to Truman’s unanticipated victory.

Q: What is the main effect of the Campaign on private health insurance enrollment? A: A one standard deviation increase in Campaign exposure is associated with a two percentage point increase in the share enrolled in PHI in the preferred specification (Column 4 of Table 1, which includes income, unionization, state fixed effects, and year fixed effects; coefficient 0.020, se 0.007, significant at 1%). This accounts for approximately 20% of the overall post-Campaign increase in PHI enrollment, corresponding to roughly 14 million new enrollees. The pre-Campaign coefficient is not statistically significant (coefficient 0.004, se 0.005), and the F-test p-value for pre-trends is 0.958.

Q: What is the effect of the Campaign on public opinion toward NHI? A: Using Gallup survey data, a one standard deviation increase in Campaign exposure led to an approximately six percentage point decline in individual support for NHI legislation per survey wave, against a pre-Campaign approval level of 69% that was trending upward. The F-test p-value for pre-trends in the Gallup event study is 0.179. This six-point effect is approximately equal to the gap in NHI support between union and non-union households, and approximately one-third the racial gap in support.

Q: What evidence links the Campaign to civic organizations and the legislative process? A: The Campaign’s archives document all civic organizations “on record against compulsory health insurance,” meaning they had passed resolutions in favor of PHI. The authors find a positive relationship between Campaign intensity and civic organizations passing such resolutions at the county level. Resolutions sent to elected officials were traced to the Congressional Record and to physical folders in the National Archives; their semantic similarity to AMA-WB propaganda is confirmed. Republican legislators’ speech in the 81st Congress shows increased similarity to Campaign language in proportion to Campaign intensity in their district or state, while Democrat legislators do not show this pattern. NHI and the AMA experienced spikes in mention frequency in the Congressional Record during this period.

Q: Did the Campaign affect physician political behavior beyond the clinic? A: By 1952, when the Republican platform had fully adopted the AMA’s position, AMA members were approximately five times more likely to donate to the Eisenhower-Nixon ticket than non-AMA physicians, with donation probability increasing in Campaign intensity. The authors digitized the donor list from the National Professional Committee for Eisenhower (NPCE) — a separate lobbying entity created because the AMA legally could not endorse candidates — and linked approximately 80% of physician donors to the AMA Medical Directory.

Q: What alternative explanations for PHI growth does the paper address, and how? A: The standard literature attributes PHI growth to the 1942 Stabilization Act wage freeze (which left benefits unconstrained), collective bargaining rights clarified in the late 1940s, and the 1954 IRS tax exemption for employer-paid premiums. The authors include income per capita and unionization as core design controls and show that their Campaign exposure coefficient is stable across specifications with and without these controls (coefficients of 0.025 and 0.020 in Table 1 Columns 1–2 vs. 3–4, respectively). The analysis stops in 1954 before the tax change, and the authors note that by 1952 roughly 63% of households already had some form of medical expense insurance.

Q: What is the conceptual mechanism through which the Campaign operated? A: The authors adapt Sobbrio (2011)’s indirect lobbying model. Voters hold uniform priors over whether NHI enactment yields net positive or negative social surplus. The private-sector advocate (AMA-WB) sends messages that shift voters’ posterior beliefs toward the negative-surplus state and, simultaneously, encourage PHI enrollment, which reduces voters’ private valuation of a public option. Because citizens were likely unaware of the coordinated tie-in advertising across industries and the financial motivation behind physician messaging, the framing operated through naive belief updating. The public-sector advocate (Truman administration, Committee for the Nation’s Health) was vastly outresourced — the CNH raised only $104,000 in 1949 — and faced legal constraints on executive lobbying.

Q: What advertising tactics specifically characterized the Campaign, and what do they imply about mechanisms? A: Campaign pamphlets and ads provided little or no substantive information about insurance products (coverage, eligibility, cost) and instead tied health insurance to ideological symbols: “freedom,” “the American way,” “the voluntary way,” and warnings about “socialized medicine.” Word clouds from Campaign materials confirm “America” and “freedom” as dominant terms. The authors connect this to behavioral models of advertising (Mullainathan, Schwartzstein and Shleifer 2008) whereby advertisers create or exploit associations to influence product beliefs. The absence of informational content is consistent with effects operating through ideology and identity rather than rational product evaluation.

Q: What explains why the U.S. did not adopt NHI in subsequent decades after the immediate Campaign period? A: The authors offer three mechanisms (discussed outside their main empirical scope): First, as middle-class Americans obtained PHI through employers, demand for a public option diminished — the model formalizes this as reduced private valuation of NHI. Second, incumbents who benefit from the private status quo — Blue Cross Blue Shield, AMA, American Hospital Association, and pharmaceutical companies, which today comprise four of the top ten direct federal lobbyists — actively work to maintain it (Acemoglu, Egorov and Sonin 2021). Third, the Campaign’s ideological framing proved durable: ideologically similar rhetoric opposing “socialized medicine” appeared in campaigns against both Clinton-era and Obama-era reform efforts, and has been linked to increased adverse selection and preventable deaths (Bursztyn et al. 2022; Galvani et al. 2022).

Q: What are the paper’s main contributions to the literature? A: The paper provides the first causal evidence on the AMA’s political role in blocking NHI at the post-WWII juncture, contributing to the economic history of U.S. social insurance development. It contributes to the advertising literature by providing credible estimates of a sustained national campaign combining trusted field agents (physicians) with mass media, and to the lobbying literature by documenting indirect lobbying — persuasion of ordinary citizens — as a distinct and effective tool alongside direct lobbying. It also documents physician behavior outside the clinical setting, showing how rents from supply-side constraints were deployed to shape the market for medical services.

Indirect lobbying: In the paper’s usage, persuasion of ordinary citizens via campaigns — as distinct from direct lobbying of policymakers — used to shift median voter beliefs and behavior to achieve legislative goals. Whitaker & Baxter are credited with creating this field through their work at Campaigns, Inc.

Campaign exposure: The paper’s composite treatment variable, constructed as the sum of two standardized components: per capita pamphlets distributed by AMA physicians (physician outreach) and per capita advertising circulation scaled by local newspaper readership (mass communications), then re-standardized to mean 0, standard deviation 1.

Tie-in advertising: Coordinated newspaper advertisements by third-party corporations and trade associations placed simultaneously with the main AMA-WB Campaign ad, sharing the “Voluntary Way is the American Way” slogan. Approximately 60% of newspapers with a main Campaign ad also had tie-in ads, averaging three per issue; third-party spending totaled approximately $19 million in 1950 dollars (~$240 million current).

Voluntary (private) health insurance: In the paper’s framing, the AMA-promoted alternative to NHI — prepaid medical service plans run by state medical societies (Blue Shield) or nonprofit hospitals (Blue Cross) — deliberately labeled “voluntary” to contrast with “compulsory” NHI, embedding the product within an ideological frame of free choice.

National Education Campaign (NEC): The AMA’s official name for the anti-NHI campaign directed by Whitaker & Baxter starting in 1949, characterized as “educational” to provide legal cover; the name itself illustrates the indirect lobbying strategy of framing political advocacy as public information.

Source text origin / abstract-only block: Not a paper-defined concept; excluded.

Naive voter updating: The paper’s modeling assumption (drawn from Sobbrio 2011) that voters held uniform priors on health insurance policy outcomes and updated beliefs via Bayesian message receipt, without awareness of coordination across industries or the financial motivation of physician messengers — making the ideological framing effective.

Physician field agents: In the Campaign’s design, AMA member physicians served as credible, trusted intermediaries who distributed pamphlets to patients and solicited civic organization resolutions, leveraging their social status to amplify the Campaign’s reach into communities where mass advertising alone would be insufficient.

Quarterly Journal of Economics | Macro Paper Warehouse

Codification, Technology Absorption, and the Globalization of the Industrial Revolution

Layer 1: Overview

In depth

Q1. What is the identification strategy and the main threats to it?

Q2. How are the cross-section and time-series results distinguished from confounders empirically?

Q3. What heterogeneity is documented?

Q4. What robustness checks are run?

Q5. How does this paper relate to and differ from prior work?

Q6. What are the policy implications and their scope conditions?

Q7. What external-validity evidence is offered beyond Japan?

Q8. What are notable caveats and measurement choices?

Key Concepts

Technology Sophistication Across Establishments

Layer 1: Overview

In depth

Q1. What are MAX and MOST, and why are they conceptually distinct?

Q2. What is the identification strategy, and what does the paper NOT claim?

Q3. How are MAX and MOST shown to be distinct upgrading processes empirically?

Q4. Is the MAX-MOST gap transitory or persistent, and how is this tested?

Q5. What are the two hypothesized drivers of MAX-MOST gaps, and what evidence supports each?

Q6. What heterogeneity in technology sophistication is documented?

Q7. How much of the variation is across vs. within sectors, and why does that matter?

Q8. What robustness and validation checks are run?

Q9. How does this paper relate to and differ from prior work?

Q10. What are the policy implications and their scope conditions?

Q11. What do the descriptive technology-use patterns show about adoption behavior?

Q12. Any notable caveats about the measures themselves?

Key Concepts

"Compensate the Losers?" Economic Policy and the Origins of U.S. Partisan Realignment

Layer 1 — Overview

In depth

Q1. What is the paper’s central conceptual innovation, and how does it differ from prior realignment research?

Q2. How large and stable is the educational gradient on predistribution, and how does it compare to social issues?

Q3. What happened to predistribution as a share of the Democratic House agenda after the 1970s?

Q4. How did 1970s campaign finance reforms change the financial composition of the Democratic Party?

Q5. Who are the “New Democrats” / DLC, and when did they emerge?

Q6. How do DLC members vote differently from other Democrats, and how is this differential conservatism distributed across policy types?

Q7. Do DLC members finance differently from other Democrats?

Q8. When precisely did educational realignment in Democratic party identification begin, and what does the inflection-point analysis show?

Q9. How do hypothetical presidential matchup surveys test the DLC mechanism?

Q10. What do actual House election results (MCDG-level data) show about DLC electoral performance by neighborhood education?

Q11. How much of educational realignment can the paper’s mechanism account for, and how is this calculated?

Q12. Can social issues, Civil Rights realignment, or Republican changes better explain the 1970s inflection point?

Q13. What is the role of out-of-district individual donors in shifting Democratic Party positions?

Q14. Are predistribution policies becoming less popular overall, which might independently push Democrats away from them?

Key Concepts

(Not) Thinking About the Future: Financial Information and Maternal Labor Supply

A Cognitive Theory of Reasoning and Choice

A Monetary-Fiscal Theory of Sudden Inflations

Overview

In depth

Q1. What is the fundamental analogy that drives the paper’s theory, and what economic literature does it build on?

Q2. What is the key information friction, and how does it generate the switch between “normal times” and crisis?

Q3. How does the bond payoff structure explain the absence of sudden deflations?

Q4. How does the first numerical example (lower surplus expectations) work quantitatively?

Q5. How does the second numerical example (interest rate increase) work quantitatively, and what does it imply for monetary policy?

Q6. What are the two regions of the deterministic model and how do they differ in fiscal and price-level dynamics?

Q7. What are the two interest rate regions in the deterministic model, and what is their implication for monetary policy effectiveness?

Q8. How does the paper relate to and extend prior FTPL literature?

Q9. What is the welfare role of consumer ignorance in this framework?

Q10. What are the scope conditions and limitations of the model?

Key Concepts

A Theory of How Workers Keep up with Inflation

Aggregation and the Estimation of Quality Change

Are Inflationary Shocks Regressive? A Feasible Set Approach

Layer 1 — Overview

In depth

Q1. What is the “feasible set approach” and how does it differ from prior work on inflation incidence?

Q2. Why is the asset price channel — rather than consumption prices — the dominant channel in both shocks?

Q3. How does the direction of the equity price response differ between oil and monetary shocks, and why does this create opposite distributional effects?

Q4. What are the precise welfare numbers for oil supply shocks by education group (baseline, ages 22–65)?

Q5. What are the precise welfare numbers for monetary policy shocks by education group (baseline, ages 25–65)?

Q6. How does the life cycle interact with the distributional incidence of both shocks?

Q7. How does the consumption channel compare across education groups and across the two shocks?

Q8. How does the labor income channel differ between oil and monetary shocks across education groups?

Q9. What is the role of housing in the portfolio channel?

Q10. How does the paper handle the dual role of durable goods (vehicles and housing)?

Q11. What happens to the main conclusions when borrowing constraints are introduced?

Q12. What is the implication of these findings for the policy interaction between oil shocks and monetary tightening?