Essays in Macroeconomics with Imperfect Expectations

The first chapter of this dissertation, coauthored with Martin Eichenbaum and Riccardo Bianchi-Vimercati, addresses the question: how sensitive is the power of fiscal policy at the ZLB to the assumption of rational expectations? We do so through the lens of a standard NK model in which people are level-k thinkers. Our analysis weakens the case for using government spending to stabilize the economy when the ZLB binds. The less sophisticated people are, the smaller the government-spending multiplier is. Our analysis strengthens the case for using tax policy to stabilize output when the ZLB is binding. The power of tax policy to stabilize the economy during the ZLB period is essentially undiminished when agents do not have rational expectations. Finally, we show that the way in which tax policy is communicated is critical to its effectiveness.In the second chapter, coauthored with Bence Bard´oczy, we study the power of state-dependent unemployment insurance (UI) to stabilize short-run fluctuations, allowing for arbitrary deviations from full information and rational expectations. Expectations are critical because higher UI generosity raises consumption partly by lowering precautionary savings. If UI generosity is indexed to the unemployment rate, households must forecast the unemployment rate to anticipate the policy stance. We estimate unemployment expectations in response to identified aggregate shocks. We quantify the consequences of these imperfect expectations through the lens of a Heterogeneous Agent New Keynesian model. First, we work directly with the estimated forecast errors. Our methodological contribution is to use the non-parametric history of forecast errors and forecast revisions to solve dynamic decisions of optimizing agents. By doing so, we sidestep the need to choose a particular model of belief formation (e.g., cognitive discounting or sticky expectations). The estimated model implies that imperfect anticipation substantially affects the stimulative power of UI extensions. Second, we compare alternative ways of implementing UI policies. To run counterfactuals, we estimate a structural model of belief formation. We show that a combination of noisy information and diagnostic expectations fits the data best among a large set of popular alternatives. A UI extension that is announced directly is more stimulative in the very short run than one that is indexed to the unemployment rate.The third chapter studies how belief disagreement across households affects aggregate demand. I develop a model in which households are heterogeneously exposed to business cycles and show that the impact of disagreement can be summarized by a simple statistic–correlated disagreement–which captures the correlation between beliefs and individual business-cycle exposure. I model disagreement as endogenously heterogeneous attention. In this model, attention increases with the exposure to business cycles. Then, I show that disagreement amplifies general-equilibrium effects and acts as a propagation mechanism amplifying business cycles. I also provide evidence of this positive correlation using survey data on expectations. To quantify the implications of disagreement, I extend the analysis to a Heterogeneous-agent New Keynesian model featuring multiple sources of heterogeneity. I show that belief disagreement can substantially amplify business-cycle fluctuations. Finally, I show that targeting spending to the most cyclical workers can significantly increase the spending multiplier.