The Monetary Transmission Mechanism: Some Answers and Further Questions

This is a conference overview and synthesis article, not an original empirical study: Kuttner and Mosser summarize the papers presented at the FRBNY’s April 2001 conference “Financial Innovation and Monetary Transmission,” and every specific empirical finding attributed below to a named author (Boivin-Giannoni, Lown-Morgan, Estrella, etc.) is a secondary paraphrase of that author’s conference paper, not a result Kuttner-Mosser derived themselves. The authors’ own contribution is the organizing taxonomy of channels, the summary of measurement challenges, and the synthesis of conclusions and open questions — the underlying data, methods, and estimates all belong to the conference papers being summarized (p. 15, wiki template note).

Kuttner and Mosser present an “eclectic” schema of six non-mutually-exclusive channels running from open market operations through reserves to the federal funds rate: the interest-rate channel, the wealth channel, the broad credit (financial-accelerator) channel, the narrow credit (bank-lending) channel, the exchange-rate channel, and the monetarist (relative-asset-price) channel (pp. 16-17). For the interest-rate channel — the standard IS-curve mechanism in both Old and New Keynesian models — they note the caveat, citing Bernanke and Gertler (1995), that the observed macroeconomic response to policy-induced rate changes is “considerably larger than that implied by conventional estimates of the interest elasticities of consumption and investment,” which is precisely the observation that motivates looking for the other five channels (p. 16). They stress explicitly that “these channels are not mutually exclusive: the economy’s overall response to monetary policy will incorporate the impact of a variety of channels” (p. 17).

The authors identify simultaneity, the difficulty of separating multiple concurrent channels, and slow-moving structural change as the three core measurement problems (pp. 17-19). On simultaneity: because the Fed eases when the economy weakens and tightens when it strengthens, the raw correlation between real GDP growth and current/near-term funds-rate changes over 1954-2000 is positive, with the contractionary effect of higher rates only becoming apparent after a lag of two quarters or more; comparing 1954-83 to 1984-2000, the correlation between funds-rate changes and subsequent growth is near zero in the later period, a pattern consistent with either less effective policy or more effective stabilization — the simultaneity problem makes the two interpretations hard to distinguish (pp. 17-18). They summarize three common solutions — VAR shock-purging, calibrated theory-based models, and microeconomic cross-sectional approaches — noting each has its own limitation (VARs make it hard to analyze systematic policy; theory models still face simultaneity; microeconomic approaches introduce micro-level endogeneity and require strong general-equilibrium assumptions to extrapolate to macro effects, since “a disproportionate impact on a particular group… will have no macro effect if other firms… are able to ’take up the slack’”) (pp. 18-19). On multiple concurrent channels: when both output and bank lending fall after a tightening, attributing the loan decline to reduced demand versus a supply-side bank-lending-channel effect requires assuming other equations of the system are unchanged when one channel is econometrically “turned off,” which is a strong assumption (p. 19). On structural change: financial innovations such as securitization and consolidation evolved gradually, and “standard statistical methods for detecting structural change work best for distinct, abrupt events, such as the October 1979 shift in Fed operating procedures,” making concurrent, evolutionary changes especially hard to isolate (p. 19).

Krieger and Bennett-Peristiani’s conference papers documented declining reserve balances and the rise of sweep accounts, which have made reserve requirements non-binding for many banks and weakened the link between the funds rate and desired reserve balances; Demiralp and Jordá’s conference paper found that while pre-1994 target changes were accompanied by systematic open-market-operation patterns, those patterns disappeared after 1994 even as the effective funds rate tracked the target more closely than before, leading them to conclude that “announcement effects” have grown in importance and to challenge the conventional view that open-market operations are central to policy implementation (p. 20). Woodford’s conference paper proposed that a “corridor” system with interest-bearing reserves and a Lombard-style lending facility would address foreseeable problems from evaporating reserve demand, and Goodfriend’s proposed interest-bearing reserves at a satiated level as a way to separately control the overnight rate and the quantity of reserves (p. 20). Kuttner-Mosser flag the Demiralp-Jordá “announcement effects” result, via the wiki’s related-concepts note, as an early forerunner of the later forward-guidance literature, though the paper itself does not use that term.

No — Boivin and Giannoni’s VAR-based conference paper and Kahn, McConnell, and Perez-Quiros’s conference paper reach directly conflicting conclusions, and Kuttner-Mosser explicitly flag this as unresolved rather than picking a side. Boivin-Giannoni find that the variance of monetary policy shocks declined sharply after the early 1980s, but argue this shock-variance decline “cannot account for” the reduced volatility of output; instead, they attribute most of the diminished output response to a change in the systematic component of policy — “leaning against the wind” more aggressively — which by implication casts doubt on nonpolicy explanations for the Great Moderation (pp. 20-21). Kahn-McConnell-Perez-Quiros instead attribute the observed reduction in output/inventory volatility to better inventory management enabled by information technology, not to any change in the monetary policy rule; the authors note these results “stand in contrast to those of Boivin and Giannoni,” and that the contrast itself illustrates “the difficulties in measuring and testing for changes in the transmission mechanism” (p. 21). (The wiki notes this Kahn-McConnell-Perez-Quiros claim has not yet been independently verified against the original conference paper in this library.)

Lown and Morgan’s VAR-based conference paper found that bank lending standards have important predictive power for both loan volume and output, but that monetary policy has little effect on those standards — evidence the authors read as offering less support for the bank-lending channel per se (p. 21). Lown-Morgan’s displacement finding, that lending standards partially displace monetary policy shocks in predicting output, is hypothesized by the authors to reflect historical “moral suasion,” and lending standards retained predictive power into the 1990s, possibly as a proxy for broader credit conditions beyond banks specifically. Separately, Estrella’s conference paper found that securitization growth significantly reduced the sensitivity of real output and housing investment to the real funds rate over the 1980s-1990s, even as the sensitivity of mortgage rates to the funds rate “if anything, increased” — interpreted as securitization operating mainly through non-interest-rate (bank-lending/credit) channels rather than the price-of-credit channel (p. 21). English’s conference paper concluded financial consolidation “has thus far had small effects on the implementation of policy and virtually no effect on the transmission of policy changes through the financial system” (p. 21), and Van den Heuvel’s conference paper proposed a “bank capital channel” in which policy-induced rate changes affect bank capital structure, with poorly capitalized banks less likely to lend, so that the distribution of bank capital ratios matters for the macroeconomic impact of policy in ways that “may interact with monetary policy in subtle and hard-to-predict ways” (pp. 21-22).

Ludvigson, Steindel, and Lettau’s structural-VAR conference paper for the US found the wealth channel is weak — “much weaker than it is in conventional structural macro models” — and, if anything, slightly weaker than in the 1960s-1970s despite equities’ growing share of household portfolios, because asset-price responses to funds-rate shocks are largely transitory while consumption responds mainly to permanent wealth changes; the authors note the apparent link between monetary policy and consumption may instead reflect a common response to shared inflation pressures rather than a true wealth-channel effect (p. 22). Separately, McCarthy and Peach’s conference paper found that regulatory changes (the repeal of Regulation Q and the shift from thrift-based to market-based housing finance) mean mortgage rates now respond more quickly to monetary policy than before 1986, but residential investment now responds more slowly and moves concurrently with overall economic activity rather than leading it, leading to the conclusion that “the housing sector is no longer in the vanguard of monetary transmission” (p. 21). In UK data, Aoki, Proudman, and Vlieghe’s conference paper calibrated a Bernanke-Gertler-Gilchrist-style financial-accelerator model and found house prices play a significant role in UK transmission, with flexible refinancing and unsecured credit having raised consumption’s sensitivity to house prices while reducing housing investment’s sensitivity — netting out to smaller monetary-policy effects on UK housing investment and prices but slightly larger effects on consumption (p. 22).

Kuttner and Mosser draw three conclusions: (1) monetary policy’s effects appear somewhat weaker than in past decades, with financial innovation, improved inventory management, and changes in the conduct of policy itself all cited as possible (non-exclusive) causes; (2) thanks to financial innovation and institutional change in housing finance, the housing sector is no longer on the leading edge of the transmission mechanism, having shifted from leading to concurrent movement with overall activity; and (3) neither financial consolidation nor the shrinking volume of reserves appears to be a major factor affecting monetary transmission, “at least not yet” (pp. 22-23). All three conclusions are explicitly hedged as provisional syntheses of a specific set of conference papers rather than the authors’ own independent empirical findings.

The authors flag three remaining open questions: the economy’s sensitivity to interest rates remains contested because Estrella (attributing reduced sensitivity to securitization) and Boivin-Giannoni (attributing it to better systematic policy) reach different explanations for a similarly reduced response, so the underlying simultaneity problem returns unresolved; the causes of growth in securitized lending and of changes in housing finance “resist easy explanation”; and — most pointedly — no paper at the conference addressed the exchange-rate channel at all, which the authors flag as an important lacuna given that the role of net exports in US macroeconomic fluctuations has grown (p. 23). This last point is presented as an explicit gap in the conference’s coverage rather than a null finding: the authors are noting an absence of research, not a result showing the exchange-rate channel is unimportant.