Three essays in macroeconomics

This dissertation contains three chapters on macroeconomics and monetary economics with a particular focus on how subjective beliefs and behavioral biases matter for the conduct of monetary policy. In the first chapter, which is joint work with Fabian Seyrich, we develop a new framework for business-cycle and policy analysis. The model features a New Keynesian core and we allow for household heterogeneity and incomplete markets as well as bounded rationality in the form of cognitive discounting. Under cognitive discounting, households' expectations underreact to macroeconomic news, consistent with what we find in the data. The model accounts for recent empirical findings on the transmission mechanisms of monetary policy. In particular, monetary policy affects household consumption to a large extent through indirect effects which tends to amplify the effects of conventional monetary policy on consumption as the incomes of households that exhibit higher marginal propensities to consume are more exposed to aggregate income fluctuations induced by monetary policy. Announcements of future monetary policy changes, in contrast, have relatively weak effects on current economic activity, and the economy remains stable at the effective lower bound. In contrast to existing models, we account for these facts simultaneously without having to rely on a specific monetary or fiscal policy. When abstracting from either household heterogeneity or bounded rationality the model fails to do so. Accounting for these facts simultaneously has important implications for monetary and fiscal policy. We uncover a novel amplification of negative productivity shocks through household heterogeneity, cognitive discounting and the interaction of the two, such that inflation increases substantially more than in existing models. Thus, to stabilize inflation after such a shock, the monetary authority needs to respond more aggressively which fuels inequality and leads to a large increase in government debt. In the second chapter, which is joint work with Klaus Adam and Timo Reinelt, we document several deviations of U.S. households' housing price expectations from rational expectations. In particular, expectations are updated on average too sluggishly; following housing price changes, expectations initially underreact but subsequently overreact; and households are overly optimistic about capital gains when the price-to-rent ratio is high. We show that weak forms of capital gain extrapolation allow to simultaneously replicate the behavior of housing prices and the deviations from rational expectations as an equilibrium outcome. We embed capital gain extrapolation into a sticky price model featuring a lower-bound constraint on nominal interest rates and show that subjective beliefs about housing prices lead to a misallocation of resources. If households become optimistic about future housing prices they invest an inefficiently high amount into the housing sector rather than into non-durable consumption. This misallocation affects the natural rate of interest and we show that lower levels of natural rates increase the volatility of housing prices and thereby the volatility of the natural rate of interest. This channel exacerbates the relevance of the lower bound constraint and causes the optimal inflation target to increase strongly as the natural rate falls. In the third chapter, I propose a new approach to quantify attention to inflation in the data. The approach is based on a model of optimal attention choice subject to information acquisition costs. The result is a law of motion for inflation expectations in which attention governs how strongly agents update their expectations following an inflation surprise. Using micro survey data of professional forecasters and consumers in the U.S., I then use this approach to estimate attention to inflation in the data and show that attention was very low after the 1980s. In the 1970s and 1980s, on the other hand, attention to inflation was substantially higher. Consistent with the underlying model, times of higher inflation volatility and persistence are characterized by higher attention to inflation. This decline in attention has important implications for optimal monetary policy. On the one hand, lower attention helps to stabilize inflation expectations and actual inflation. On the other hand, lower attention renders managing inflation expectations more difficult which can lead to prolonged periods of a binding lower bound and low inflation due to slowly-adjusting inflation expectations. To mitigate this loss of control, the optimal inflation target substantially increases as attention falls. Allowing for permanently higher inflation rates, however, is costly in terms of welfare. Thus, welfare decreases as attention declines whereas it would increase when abstracting from the lower bound constraint