Monetary-Policy-Inequality | Macro Paper Warehouse

Fiscal and Monetary Policy with Heterogeneous Agents

This paper reviews the Heterogeneous-Agent New Keynesian (HANK) literature that has emerged over the past decade, combining the canonical incomplete-markets model of income and wealth inequality (in the Bewley-Huggett-Aiyagari tradition) with the New Keynesian model of price and wage rigidity used to study monetary and fiscal policy. Rather than surveying disparate models, the authors build a single “canonical HANK model” – with sticky wages, flexible prices, and endogenous consumption-saving choice only (no endogenous labor supply) – calibrated to match realistic average marginal propensities to consume (MPCs) and a realistic wealth distribution, and use it to organize the field’s central results. Studying fiscal policy first, they show that a balanced-budget increase in government spending produces an output multiplier of exactly 1, identical to a representative-agent (RA) model, regardless of household heterogeneity (Proposition 1) – but a deficit-financed tax cut has much larger and more persistent effects in the heterogeneous-agent (HA) model than in either a representative-agent or two-agent (TA) model, because households partially save the tax cut, building up “excess savings” that low-MPC, poor households then spend down over time, an effect that “trickles up” toward wealthier households as it winds down. Turning to monetary policy, they show a subtler result: when steady-state government debt is zero, a monetary policy shock has an identical aggregate effect on output in HA, TA, and RA models (Proposition 2, generalizing a result first obtained by Werning 2015), because higher marginal propensities to consume are offset by lower sensitivity to future interest rates. Heterogeneity does not necessarily change the size of monetary policy’s aggregate effect, but it does change its transmission mechanism: decomposing the consumption response shows that “indirect” effects from labor income, capital gains, and government transfers dominate the “direct” interest-rate effect on saving decisions, a finding the authors attribute to Kaplan, Moll and Violante (2018). The paper then surveys a wide set of extensions – cyclical income risk, government debt maturity, nominal (rather than real) assets, behavioral frictions, the fiscal theory of the price level, illiquid two-account models, endogenous portfolio choice, and additional demand components such as investment and durable goods – and closes by noting that the literature has not yet reached a comparably mature theory of optimal monetary and fiscal policy in HANK models, in part because an unrestricted heterogeneous-agent economy typically lacks a well-defined Ramsey steady state to serve as a benchmark.

Fiscal Progressivity and the Time Consistency of Monetary Policy

The conventional division of labour holds that central banks should not target distribution and that fiscal policy, with its targeted instruments, should handle the redistributive consequences of monetary decisions. This paper argues the opposite direction of influence also matters: in a stylized overlapping-generations economy with agents who differ in labour productivity, progressive labour taxation — which is purely costly on efficiency grounds, since it only raises marginal rates and labour-supply distortions — nonetheless serves as an effective instrument for mitigating the inflation bias of discretionary monetary policy, but only when policymakers or voters are concerned about the distribution of consumption. The mechanism runs through distributional conflict: with a flat tax, agents unanimously support financing entirely through the inflation tax, because money holdings of the old are a predetermined and hence non-distortionary tax base; with progressive taxation, lower-productivity agents instead support positive labour taxes to preserve the consumption value of their money holdings and shift the burden of distortionary taxation onto higher-productivity agents, so the median-productivity agent — shown to be the decisive voter — chooses positive labour taxes and thereby curbs the inflation tax. Anticipating that reduction in inflation, agents choosing progressivity behind a veil of ignorance one period in advance (tax inertia) unanimously prefer a strictly positive level of progressivity, even though their individual preferred levels differ and are non-monotonic in productivity. A numerical extension with incomplete markets and idiosyncratic productivity risk, calibrated to US moments (a market-income Gini of 0.48, after-tax Gini of 0.36, public consumption of 15% and transfers of 7% of output), shows that under discretion without progressivity the reliance on inflationary finance generates a collapse of money demand, with lifetime welfare falling to about 0.73 of the commitment benchmark, welfare dispersion rising to about 2.27 times, and output falling from 0.741 to 0.562; adding the calibrated level of progressivity brings both discretionary and majority-voting outcomes back to roughly the commitment benchmark. The analysis is deliberately stylized, and the author reports that the more persistent and volatile idiosyncratic shocks are — and the lower are lump-sum transfers or the higher policymakers’ inequality aversion — the more effective progressive labour taxes are at limiting the inflation bias.

Inflation and the Redistribution of Nominal Wealth

This paper quantifies how an unanticipated bout of moderate inflation, similar in magnitude to the U.S. experience of the 1970s, would redistribute wealth by revaluing nominal (dollar-denominated) assets and liabilities. Combining sector-level data from the Flow of Funds Accounts with household-level data from the Survey of Consumer Finances, the authors construct market-value nominal positions – including indirect positions arising from ownership of financial intermediaries and firms – and their maturity/duration structure for every major class of U.S. nominal asset and liability, then simulate a hypothetical episode of 5 percentage points of extra inflation per year for 10 years starting from a given benchmark year. Because agents’ actual expectations and portfolio-adjustment speed are unobserved, the paper brackets the results between two polar scenarios: “Full Surprise,” in which nominal positions are devalued uniformly regardless of maturity, and “Indexing ASAP,” in which bond markets immediately price in the full future inflation path so only shorter-duration positions bear large losses. Under both scenarios and across benchmark years, the government and (since the 1980s) domestic households gain at the expense of foreign holders of U.S. nominal assets, and within the household sector old, wealthy, bond-holding households lose to young, middle-class households with fixed-rate mortgage debt; for the benchmark year 1989, the paper’s central estimates put the loss to a coalition of rich and old households at 5.7-15.2 percent of GDP and the gain to under-45 middle-class households at up to 45 percent of mean cohort net worth. The paper also documents that financial innovation – chiefly the securitization of mortgages beginning in the early 1980s – roughly halved the surprise-inflation losses of elderly households between the 1989 and 2001 benchmark years by shifting maturity mismatch away from bank/intermediary shareholders and toward long-term bondholders, while leaving households’ exposure to gradual (anticipated) inflation comparatively unchanged. Throughout, the authors are explicit that these are the redistributional effects of revaluing already-existing nominal positions alone, not a full general-equilibrium assessment of inflation’s real effects, and they flag the study of how this redistribution shock feeds back into aggregate consumption, saving, and labor supply, and how fiscal policy might offset it, as questions for companion work rather than this paper.

Monetary Policy and Heterogeneity: An Analytical Framework

This paper builds THANK, a tractable heterogeneous-agent New Keynesian (HANK) model with two household types – savers and hand-to-mouth agents who move between the two states via a Markov process – that nests the representative-agent (RANK) and two-agent (TANK) models as special cases and admits closed-form solutions for dynamic properties that quantitative HANK models can only compute numerically. Its central object is χ, the elasticity of hand-to-mouth households’ income to aggregate income, which pins down whether income inequality is countercyclical (χ>1) or procyclical (χ<1); the paper shows this single statistic governs whether the model’s aggregate Euler-IS equation exhibits “compounding” or “discounting” relative to the representative-agent benchmark. Countercyclical inequality delivers the aggregate-demand amplification and positive fiscal multipliers that much of the quantitative HANK literature is built to generate, but simultaneously makes the model’s Taylor-rule determinacy condition more stringent than the standard Taylor principle and worsens the forward guidance puzzle – the counterfactual prediction that a monetary policy change further in the future moves consumption today by more than a near-term change; procyclical inequality does the reverse, weakening the Taylor principle’s necessity for determinacy and curing the puzzle. Because amplification and puzzle-curing require opposite cyclicalities of the same χ, the author calls this a “Catch-22” for HANK models. The paper offers two classes of resolution: combining cyclical inequality with a separately modeled cyclical income-risk channel of the opposite sign (empirically, the author reports that U.S. disposable-income inequality was mildly procyclical while income risk was countercyclical in the last two recessions), or switching to policy rules – Wicksellian price-level targeting, or a nominal-debt rule – that restore determinacy regardless of the sign of cyclicality. Finally, solving a Ramsey optimal-policy problem to second order, the paper derives a novel “inequality-stabilization” motive that makes the central bank optimally tolerate more inflation volatility whenever inequality is cyclical, while the cyclicality of idiosyncratic risk itself is shown to be irrelevant to the optimal-policy objective (though not to the interest-rate rule that implements it), because the policy target is the perfect-insurance, no-inequality efficient allocation.

Monetary Policy and Redistribution in Open Economies

This paper builds an open-economy heterogeneous-agent New Keynesian (HANK) model in which households differ not only in income and wealth, as in standard closed-economy HANK models, but in their “real integration” (whether they work in a home tradable sector exposed to foreign demand, or a purely domestic nontradable sector) and “financial integration” (whether they can save and borrow internationally, or only in domestic securities priced off the domestic policy rate). Calibrated to Canada, the model is used to revisit three classic questions from Mundell (1963) and Fleming (1962) – the international spillovers of shocks and policies, the comparison of exchange-rate regimes, and the implications of the international price system – but from a distributional rather than purely aggregate perspective. The paper’s central finding is a systematic trade-off between aggregate stabilization and consumption inequality: fixed exchange rates amplify the aggregate response to external shocks (as in standard representative-agent open-economy models) but reduce the cross-household dispersion of that response, because defending a peg requires cutting domestic rates more aggressively, which disproportionately benefits financially non-integrated and nontradable-sector households. A parallel finding is that lower degrees of real and financial integration dampen an economy’s aggregate exposure to external shocks but concentrate their distributional impact on a narrower set of directly-exposed households, leading the authors to conclude that the “discontents” of globalization may stem from integration being insufficiently generalized, rather than from integration itself.

Monetary Policy and the Redistribution Channel

This paper formalizes and measures a “redistribution channel” through which monetary policy affects aggregate consumption – distinct from, and additional to, the standard income and substitution channels present in representative-agent models. Building on Tobin’s (1982) intuition that “aggregation would not matter if… marginal propensities to spend… were the same for creditors and for debtors,” Auclert identifies three specific sources of redistribution set in motion by a monetary expansion: an earnings heterogeneity channel (unequal gains from higher aggregate income), a Fisher channel (unexpected inflation reallocating wealth between nominal creditors and debtors), and an interest rate exposure channel (real rate changes reallocating wealth according to the duration mismatch between a household’s maturing assets and liabilities, captured by a new measure called “unhedged interest rate exposure,” or URE). The paper’s central theoretical result decomposes the first-order response of aggregate consumption into five terms – the two channels present in representative-agent models plus these three redistributive channels – each governed by a sufficient statistic: the cross-sectional covariance between household marginal propensities to consume (MPCs) and the household’s exposure to the relevant aggregate shock. Using household survey data from Italy and the United States, employing three different established methods for measuring MPCs, the paper finds that all three covariances point in the amplifying direction – households who gain from an accommodative monetary shock tend to have higher MPCs than those who lose – with the interest-rate-exposure channel comparable in magnitude to the conventional substitution channel for empirically plausible values of the elasticity of intertemporal substitution, while the Fisher channel, though correctly signed, turns out to be quantitatively modest.

Optimal Monetary Policy According to HANK

This paper studies optimal monetary policy in an analytically tractable heterogeneous-agent New Keynesian (HANK) economy in which households face uninsurable idiosyncratic labor-disutility shocks and can only self-insure through a riskless bond and hours worked. Using CARA preferences and normally distributed shocks – a device the authors also used in earlier work – the model aggregates linearly, so that the entire cross-sectional distribution of consumption collapses to a single sufficient statistic (Sigma_t) that a utilitarian Ramsey planner weighs alongside the standard output-gap and inflation objectives. The paper shows that monetary policy affects this inequality statistic through up to four distinct channels – income risk, self-insurance, unhedged interest rate exposure (URE), and (with nominal debt) the Fisher channel – and derives closed-form optimal policy rules that nest the representative-agent (RANK) case. When income risk is countercyclical (the empirically relevant case), optimal policy curtails the fall in output during recessions more than RANK would, tolerating higher inflation because doing so also limits the associated rise in consumption inequality. The paper’s most novel result is normative and methodological rather than purely quantitative: because a surprise rate cut can redistribute from savers to debtors given existing wealth dispersion, but an anticipated one cannot, the Ramsey-optimal plan is time-inconsistent in a genuinely new way – a benevolent planner who could re-optimize would always want to engineer one more surprise cut. These results are derived under the baseline assumption of real (inflation-indexed) household debt; Section 6 shows they survive, and are reinforced, when debt is nominal and the Fisher channel is reintroduced.

Optimal Monetary Policy with Heterogeneous Agents: Discretion, Commitment, and Timeless Policy

This paper characterizes optimal monetary policy in a canonical one-asset heterogeneous-agent New Keynesian (HANK) model with wage rigidity – a minimal departure from the representative-agent (RANK) New Keynesian benchmark – and systematically revisits the canonical consensus on optimal monetary policy design under discretion, under commitment, and for short-run stabilization. Under discretion, a utilitarian planner has an incentive to overheat the economy beyond the standard markup-correcting level because lowering interest rates redistributes income toward indebted, high-marginal-utility households; since the public rationally anticipates this, the attempt at stimulus is self-defeating and instead produces inflationary bias in the sense of Barro and Gordon (1983), with the paper’s calibration finding this redistribution channel contributes over four times as much to that bias as the conventional markup distortion. Full commitment restores zero inflation in the long-run stationary equilibrium – because inflation and the nominal rate affect household financial income symmetrically, while only inflation is costly – but the standard Ramsey problem still suffers a “time-0” problem that generates short-run inflationary bias, driven both by the usual forward-looking Phillips curve and, newly in HANK, by each household’s forward-looking value function entering as a planning constraint. To resolve this, the authors extend Marcet and Marimon’s (2019) recursive-multiplier approach to continuous-time heterogeneous-agent economies, defining a “timeless” Ramsey problem augmented with an inflation penalty (now shaped by distributional considerations even in HANK) and a novel distributional penalty that specifically counteracts the planner’s incentive to redistribute toward indebted households; this timeless plan eliminates inflationary bias in both the short and long run and can be implemented either by a discretionary planner confronted with the right penalties or by an appropriately designed inflation target. Finally, characterizing optimal stabilization policy under the timeless Ramsey problem, the paper shows that the classic Divine Coincidence result of RANK models – that inflation and output gaps can always be closed simultaneously absent cost-push shocks – generically fails in HANK even with the correct employment subsidy, because the planner now trades off aggregate stabilization against distributional considerations; a quantitative decomposition traces this departure, in response to demand shocks, specifically to the redistribution wedge. The analysis is conducted in a stylized model with a single financial asset and one particular (interest-rate) redistribution channel, and the authors are explicit that while their qualitative logic should generalize, the exact quantitative conclusions – including the sign of the discretionary inflationary bias – depend on the specific pecuniary channels through which policy redistributes in a given model.

Optimal Policy Rules in HANK

This paper characterizes optimal monetary and fiscal policy rules in a rich heterogeneous-agent New Keynesian (HANK) business-cycle model with nominal rigidities, in which the policymaker has two instruments – the short-term nominal interest rate and uniform lump-sum transfer (stimulus-check) payments – and asks whether, and how, household inequality should change how each instrument is set. For a policymaker with a conventional “dual mandate” that targets aggregate output and inflation, the paper proves the optimal interest-rate targeting rule is exactly the same as in the textbook representative-agent New Keynesian model, because in this economy household heterogeneity affects only the demand side, which is a slack constraint once the supply-side Phillips curve is left unchanged; empirically disciplined HANK and RANK models therefore prescribe essentially the same policy-rate paths. The paper then adds an explicit distributional objective – a planner who wants to insure households against business-cycle-driven swings in their consumption shares – and derives a linear-quadratic optimal rule with an additional term governed by the causal effect of each instrument on consumption inequality. Because the calibrated model, built to match evidence on monetary transmission, implies that interest-rate changes move household consumption by roughly similar percentages up and down the wealth and income distribution, this distributional term turns out to matter little in practice: optimal monetary policy stays close to the dual-mandate benchmark even when the planner cares about inequality, because using rates to fight inequality would require costly departures from aggregate stabilization for limited distributional gain. Stimulus checks, by contrast, have strongly progressive effects in the model – both from elevated marginal propensities to consume among low-income, low-wealth households and from a fixed dollar transfer mattering more as a share of low incomes – so they are shown to be an effective complementary tool for offsetting shocks with a strong distributional tilt, such as a simulated income-redistribution shock resembling the Covid-19 recession. These conclusions are explicitly conditional on the paper’s calibration of policy transmission channels; the authors show that alternative model specifications implying larger distributional effects of monetary policy (as in some other recent HANK papers) would restore a more significant role for distributional considerations in interest-rate policy.

The New Keynesian Transmission Mechanism: A Heterogeneous-Agent Perspective

This paper studies how the simplest possible form of household heterogeneity – splitting the representative agent of the textbook New Keynesian model into a “worker,” who receives only labor income, and a “capitalist,” who receives only firm profits – changes the model’s monetary transmission mechanism. Under the standard assumption that only goods prices are sticky and wages are flexible, the authors show that this 2-agent model behaves very differently from its representative-agent counterpart: the real interest rate, inflation, real wages, and profits all respond similarly to a monetary policy shock, but output and employment do not respond at all in the worker-capitalist model, whereas they fall sharply in the standard model. The reason is that with the balanced-growth (King-Plosser-Rebelo) preferences standard in macroeconomics, income and substitution effects on labor supply exactly cancel; once profit income is removed from a worker’s budget (because a worker earns only wages), this cancellation makes hours completely unresponsive to wage movements, so monetary policy only redistributes consumption between workers and capitalists – it does not move aggregate output. The authors then show that the representative-agent model’s own ability to generate an output response rests on an empirically fragile mechanism: profits move countercyclically with the policy rate, making the representative household poorer and inducing it, via a wealth effect, to work more – a channel undermined both by household balance-sheet data showing few households hold much non-labor income, and by the fact that profits are procyclical, not countercyclical, in the data. When wage stickiness is introduced instead of (or alongside) price stickiness, however, workers are pushed off their static labor-supply curve and simply supply whatever hours are demanded; in this case the worker-capitalist model’s impulse responses become nearly indistinguishable from the representative-agent model’s, and the authors show this equivalence strengthens as the degree of wage rigidity increases. The authors confirm these results are robust to allowing limited financial trade between workers and capitalists via a bond market with adjustment costs.

Transmission of Monetary Policy with Heterogeneity in Household Portfolios

This paper builds a two-asset Heterogeneous-Agent New Keynesian (HANK) model – households hold a liquid asset (government and household bonds) and an illiquid asset (real capital, traded only with a per-period probability and priced by q-theory) subject to uninsurable idiosyncratic income risk – to ask how heterogeneity in household portfolios shapes monetary transmission. Its central finding is that monetary transmission works predominantly through investment rather than consumption, and that portfolio heterogeneity systematically dampens the investment response: aggregate investment falls by one-third less after a monetary tightening in the model with portfolio heterogeneity than in an otherwise identical model with a representative portfolio, while the direct effect of the policy rate itself explains 86% of the investment response but only one-third of the consumption response, with equilibrium changes in income accounting for the remaining two-thirds of the consumption effect. The dampening of investment operates through two channels tied to redistribution: first, a monetary tightening redistributes income and wealth toward wealthy households who have a low marginal value of liquidity, which lowers the endogenous liquidity premium (the illiquid asset’s return over the liquid asset’s) by 16 basis points and produces incomplete pass-through of the policy rate to the return on capital; second, because wealthy households have high marginal propensities to invest (MPI) but low marginal propensities to consume (MPC) – close to 40% MPC but under 5% MPI near the borrowing constraint, versus roughly twice the MPI above median wealth – redistributing resources toward them stabilizes aggregate investment while amplifying the fall in consumption among wealth-poor, high-MPC households. The paper also shows the Fisher channel of unexpected disinflation, which redistributes from nominal borrowers to savers, quantitatively amplifies the demand-driven output response by about 9% when prices are sticky, but reverses sign under flexible prices, generating an investment boom instead, because only the marginal-propensity-to-invest heterogeneity is then operative. Two empirical exercises corroborate the mechanism: regressing Survey of Consumer Finances household portfolio data on identified monetary shocks, the paper documents that households below median wealth reduce portfolio liquidity after a monetary tightening while households above median wealth increase it – a differential pattern the model can only replicate when capital is imperfectly liquid – and local projections using two measures of the liquidity premium (the Gomme et al. return on capital and the return on housing, both over the risk-free rate, identified via Romer-Romer narrative monetary shocks) show both premia fall after a tightening, consistent with the model’s prediction. The analysis abstracts from financial frictions on the firm side to isolate household-portfolio frictions specifically.

Unconventional but Different After All? A Unified Series of Narrative Monetary Policy Shocks

Asking whether unconventional monetary policy works differently from an interest-rate cut has been hard to answer because the two kinds of policy have been measured with different tools, so any difference in the estimated effects could be a difference in the measuring rather than in the policy. This paper builds one shock series for both, running Romer and Romer’s (2004) narrative regression — policy-rate changes on the Federal Reserve’s own Greenbook/Tealbook forecasts — but substituting Wu and Xia’s (2016) shadow rate for the federal funds rate over the 2009–16 zero-lower-bound period, which yields shocks for 1969–2008 and 2009–16 built the same way. Estimating local projections on the pooled sample with a structural break at the zero lower bound, the authors cannot reject equality of the peak responses of the interest rate, industrial production, unemployment and the consumer price index across regimes (p-values from 0.09 to 0.56), which they read as unconventional policy being about as effective as rate cuts for aggregate activity. Joint Wald tests do reject equality for wealth shares and for the stock-to-house price ratio, and the sign flips: an expansionary conventional shock normalised to 25 basis points lowers the top 10% wealth share by about 0.3% at the trough, while the equivalent unconventional shock raises it by up to about 0.36% and lowers the middle 40% share by about 0.7%. The authors attribute the difference to asset prices — unconventional easing raises the stock-to-house price ratio by over 3% where conventional easing lowers it by about 3.13% — and a mechanical revaluation of Survey of Consumer Finances portfolios reproduces the pattern, with the equity-heavy top 10% capturing most of the gains and households in the bottom groups, who hold little equity and often no housing, receiving comparatively little. The distributional evidence rests on short samples — 1989–2008 for conventional and 2009–15 for unconventional shocks, at quarterly frequency — and the authors do not address the Fed-information-effect and Fed-reaction-to-news concerns directly, assessing them only indirectly by comparison with shock series that are robust to them.

When do Endogenous Portfolios Matter for HANK?

Most heterogeneous-agent New Keynesian (HANK) models assume households hold a fixed, exogenously given mix of assets – a simplification that is natural because standard first-order or “MIT shock” solution methods leave portfolio choice genuinely indeterminate, but one that sidesteps the fact that agents who perceive aggregate risk and can invest in several assets have a well-defined optimal portfolio near the steady state. This paper develops a new sequence-space method for solving jointly for these “zeroth-order” endogenous portfolios and for the model’s impulse responses, extending the fake-news-algorithm machinery of Auclert, Bardóczy, Rognlie and Straub (2021) with a second-order perturbation of the household portfolio problem evaluated just before shocks realize. When there are at least as many assets as aggregate shocks, optimal portfolios reduce to a simple risk-sharing test – marginal utility must respond proportionally across households to any aggregate shock – and the correction this implies for the model’s sequence-space Jacobians uses the same objects as the ordinary, exogenous-portfolio computation. Applying the method to a simple HANK model with a stock and a bond, the authors find that endogenous portfolios leave the aggregate effects of balanced-budget government spending shocks and of monetary policy shocks unchanged relative to the standard exogenous-portfolio (100%-stock) benchmark, because in both cases the exogenous portfolio already happens to satisfy (or trivially bypass) the risk-sharing condition. Deficit-financed fiscal transfers are different: because such transfers disproportionately raise the consumption – and lower the marginal utility – of poor, high-marginal-propensity-to-consume (high-MPC) households, optimal hedging induces poor agents to take large short positions in the booming stock market, cutting the baseline calibration’s impact transfer multiplier from 0.2 to 0.08 and its cumulative multiplier from 0.77 to 0.53. This result is sensitive to how much gross portfolio exposure is allowed: realistic short-sale and leverage constraints (stocks between -100% and 200% of net worth) bring the multiplier back close to the exogenous-portfolio benchmark, and adding more shocks than assets (incomplete markets) likewise pulls results back toward the exogenous-portfolio case when the additional shocks are hard to hedge. A parallel exercise with nominal assets shows the same logic working in the opposite direction: when households start out highly exposed to a Fisher (debt-deflation) channel, optimal portfolios shrink that exposure toward empirically plausible levels and substantially dampen the response to monetary shocks. The authors conclude that endogenous portfolios can matter a great deal for HANK results, but only when high-MPC agents are permitted to take large gross positions to hedge aggregate risk – a scope condition the paper is explicit about throughout.