Marco Del Negro, Keshav Dogra, and Laura Pilossoph Monetary policy can have a meaningful impact on inequality, as recent theoretical and empirical studies suggest. In light of this, how should policy be conducted? And how does inequality affect the transmission of monetary policy? These are the topics covered in the second part of the recent symposium on “Heterogeneity in Macroeconomics: Implications for Policy,” hosted by the new Applied Macroeconomics and Econometrics Center (AMEC) of the New York Fed on November 12. Inequality and the Transmission of Monetary and Fiscal Policy The first session in the afternoon (the agenda includes links to all of the presentations) asked what we have learned from the new literature on Heterogeneous Agent New Keynesian (HANK) models about the
New York Fed considers the following as important: Adrien Auclert, Fiscal Policy, Gianluca Violante, Greg Kaplan, HANK, heterogeneous agent new Keynesian models, inequality, Keshav Dogra, Laura Pilossoph, Macroecon, Macroeconomics, Marco Del Negro, Michael Woodford, Monetary Policy, monetary policy transmission, Veronica Guerrieri
This could be interesting, too:
New York Fed writes The Fed’s Latest Tool: A Standing Repo Facility
New York Fed writes How the Fed Adjusts the Fed Funds Rate within Its Target Range
New York Fed writes How the Fed’s Overnight Reverse Repo Facility Works
New York Fed writes How the Federal Reserve’s Monetary Policy Implementation Framework Has Evolved
Monetary policy can have a meaningful impact on inequality, as recent theoretical and empirical studies suggest. In light of this, how should policy be conducted? And how does inequality affect the transmission of monetary policy? These are the topics covered in the second part of the recent symposium on “Heterogeneity in Macroeconomics: Implications for Policy,” hosted by the new Applied Macroeconomics and Econometrics Center (AMEC) of the New York Fed on November 12.
Inequality and the Transmission of Monetary and Fiscal Policy
The first session in the afternoon (the agenda includes links to all of the presentations) asked what we have learned from the new literature on Heterogeneous Agent New Keynesian (HANK) models about the transmission and the efficacy of monetary policy. It also asked what we learned from the COVID-19 recession and the associated policy response on the impact of redistributive policy on aggregate demand and supply.
Greg Kaplan from the University of Chicago discussed lessons from the HANK literature regarding monetary policy and its ability to stabilize economic activity. One lesson is that stabilization policy may be harder than envisaged in models with a representative agent. In representative agent models monetary policy can rely on the intertemporal substitution channel—lowering interest rates induces households to consume. This channel is much weaker in HANK models, partly because these models recognize that a number of households have a hard time shifting consumption over time either because they have little wealth or because their wealth is illiquid. Instead, the main transmission channel of monetary policy in HANK models involves changing household disposable income—and in particular affecting income for households with higher marginal propensities to consume. But this often involves redistributing resources among individuals, where some win and others loose (for example, debtors versus creditors) from policy actions.
This finding points to two further lessons for monetary policy. One is that the redistributive effects of policy can no longer be swept under the rug as is the case in representative agent models. The other is that the connection with fiscal policy is also front and center, since many of the consequences of monetary policy in these models involve public debt and the reaction of the fiscal authorities. A final lesson from HANK models, which ties in with the discussion in the first part of the symposium, is that the benefits from aggregate stabilization largely derive from alleviating the hardship of the most vulnerable households during recessions. From Kaplan’s perspective these four lessons suggest that monetary policy should focus less on fine-tuning aggregate demand and more on stabilizing inflation and enhancing the stability of the financial system.
Veronica Guerrieri from the Chicago Booth School of Business discussed monetary and fiscal policy in the context of the COVID-19 pandemic. She first highlighted the striking distributional impact of the pandemic recession, which hit some sectors (for example, leisure and hospitality) and some workers (for example, low-income workers) harder than others. As she shows in her recent paper with Lorenzoni, Straub, and Werning, the pandemic shock can be seen as a supply shock which propagated through the economy via demand channels. In particular, the income losses of workers with high marginal propensity to consume resulted in aggregate demand shortages for a while. Moreover, shocks in some sectors spread to other sectors because of complementarities and supply chains. This demand shortfall called for policy stimulus, and at the same time the asymmetric nature of the shock called for social insurance. To some extent, targeted fiscal transfers accomplished both goals.
Guerrieri then asked whether fiscal (and monetary) stimulus in the United States has been excessive, and whether the inflation that the U.S. is currently experiencing is the outcome of such stimulus. Again, she stressed that differences across households, and specifically the heterogenous savings behavior by income in response to the transfers, are key for understanding the effect of the stimulus on demand, both on impact and over time. She also emphasized the heterogeneity across sectors, with some sectors that have recovered well beyond their pre-pandemic levels and others that are still struggling. In another recent paper with Lorenzoni, Straub, and Werning, Guerrieri and coauthors argued that some of the inflation the U.S. is currently experiencing may be desirable as it helps adjust relative prices across sectors, thereby cooling some of the demand in the booming sectors and directing it toward the weaker ones. In fact, the stimulus may also help the reallocation of labor across sectors as it leads to more sustained wage increases in the strong sectors, which in turn attract workers, thereby lessening supply shortages.
In the discussion that followed the presentations, New York Fed President John Williams said that while he appreciated the insights from the HANK literature on the trade-offs faced by monetary policy, in the end the objective of monetary policymakers remains achieving the dual mandate of price stability and full employment with the tools at their disposal. Does the HANK literature have lessons in terms of which monetary tools (conventional interest rate policy, forward guidance, QE) are best suited to achieve the mandate? Both Kaplan and Guerrieri emphasized that an implication of heterogeneity is that fiscal policy is in many ways better suited than monetary policy to stabilize the economy, because it can be more precisely targeted toward different households.
Gianluca Violante of Princeton University agreed on this point and wondered why fiscal policy is seemingly conducted in a much less sophisticated and thoughtful way than monetary policy. He hinted that one reason may be political economy considerations—the need to achieve compromise—but another reason could be the moral hazard inherent with transfer-based policies. He argued that the willingness of fiscal authorities to act effectively during COVID may be due to the fact that moral hazard was largely absent in these circumstances. St. Louis Fed President Jim Bullard highlighted that if either fiscal or monetary policy are not doing their job in terms of stabilizing the economy, this may limit the effectiveness of the other.
Heterogeneity and Optimal Monetary Policy
The second afternoon session discussed how heterogeneity should affect the design of optimal policy. Adrien Auclert from Stanford began by noting that the normative HANK literature is still in its infancy, relative to the positive literature. He outlined two key outstanding issues. The first issue is that while the New Keynesian (NK) literature is organized around a single tractable “textbook” model and a set of agreed upon methods for solving and estimating quantitative models, no such consensus exists in the HANK literature. But progress has been made: we have a number of tractable HANK models that can be used to understand positive and normative aspects of HANK economies, and we are closer to having efficient tools for solving and estimating HANK models.
The second issue concerns what objective function policymakers should optimize in our normative models. In the canonical NK model, under certain conditions, maximizing household welfare is equivalent to minimizing an “ad hoc” loss function which penalizes deviations of the output gap and inflation from their targets. This equivalence was an important factor in the success of the NK literature, since policymakers already viewed their objective in terms of a simple ad hoc loss function. But when it comes to the HANK literature, there is generally no simple, microfounded loss function that plays the same role.
One approach is to continue to use an ad hoc loss function to analyze optimal policy in HANK. Although this approach is straightforward, it may miss out on important forces within the model and provides no guidance on which measures of inequality, if any, should be incorporated in the policymaker’s objective function. A more technically challenging approach is to solve for policy that maximizes household welfare within a HANK model. In Auclert’s view, the literature that does this has uncovered some findings that should generalize (for example, optimal policy tries to undo the redistributive effects of aggregate shocks), but it is less clear whether other specific findings from particular papers will generalize (some studies find that policy should put more weight on stabilizing the level of output to stabilize consumption inequality, while others find that price stability remains the dominant concern).
Michael Woodford from Columbia University discussed another dimension of heterogeneity, namely heterogeneity of expectations. He argued that while this dimension has been relatively less studied—presumably because it requires relaxing the conventional assumption of rational expectations—it is worth incorporating into our models, both because it is clearly present in survey data, and because algorithmic models of expectation formation can be more tractable than models with model-consistent expectations. Woodford outlined one approach to modelling heterogeneous beliefs based on his work with Yinxi Xie. This approach assumes that households and firms make decisions based on looking forward over a finite planning horizon, implying that heterogeneity of beliefs arises from differences in their planning horizons. In Woodford’s model, this nests fully rational expectations as a limiting special case (when all agents have infinite horizons). An important implication of belief heterogeneity for policymakers is that the same policy announcement will lead different decisionmakers to expect different paths of policy tools and other macroeconomic variables, since some of them put a lot of weight on far future outcomes while others have much shorter planning horizons. Another implication is that policy should depart from fully stabilizing output gaps and inflation in order to reduce distortions arising due to heterogeneity in beliefs.
In the following discussion, Williams emphasized that, while we know heterogeneity is important, and each heterogeneous-agent model has its own particular implications for optimal policy, it will be important to learn which conclusions are robust across these different models. Woodford agreed with Williams (and with Kaplan’s earlier remarks) that it would be too early to follow the policy conclusions from any one paper, but argued that we can still say something about which conclusions are likely to be robust. We can learn which variables in our models have the properties that expectations about those variables have a particularly large effect on outcomes; we should then want to keep those variables stable, so that they are easy for agents to forecast, regardless of how precisely they do that. Kaplan noted that economists generally do not make their models more realistic simply for the sake of realism, but to match some feature of the data that existing models cannot; he wondered what empirical puzzle belief heterogeneity solves. Woodford suggested that finite planning horizons may offer one solution to the “forward guidance puzzle.”
Bullard emphasized the advantage of building on tractable models (for example, lifecycle models with within-cohort heterogeneity), which avoid some of the technical problems Auclert mentioned, but can still match important features of the data. Auclert agreed that simple models are useful to study particular forces (for example, countercyclical movements in the price level can provide insurance to households with nominal debt), but argued that richer models are useful to trade these off against other important forces (for example, in the presence of nominal wage rigidity, countercyclical inflation may have less favorable effects on the distribution of real resources).
In sum, the upshot from the symposium is that taking heterogeneity into account is crucial for policy, both because its effect differs across households—not only fiscal but also monetary policy can significantly affect inequality—and because heterogeneity changes the way we understand how fiscal and monetary policy are transmitted to the economy.
Marco Del Negro is a vice president in the Federal Reserve Bank of New York’s Research and Statistics Group.
Keshav Dogra is a senior economist in the Bank’s Research and Statistics Group.
Laura Pilossoph is a senior economist in the Bank’s Research and Statistics Group.