UC San Diego

Monetary policy in the U.S. has changed substantially in the past few decades. This thesis seeks to understand the effects of monetary policy through household expectations and housing prices. The first chapter proposes and estimates a dynamic model of household inflation expectations. The information flow constraint of the household leads to costly information monitoring. Households use a Bayesian learning model to form and update inflation expectations. The model identifies and corrects for sizable reporting and sampling errors prevalent in household surveys. The estimates show that better-educated households track inflation more closely and report their expectations more accurately. Household inflation expectations are less responsive to changes in the inflation target after the Great Recession. Model-implied household inflation expectations improve the fit of the expectation-augmented Phillips curve. Inattention from households makes it costlier for the Fed to lower inflation than would be the case if everyone is perfectly informed. The second chapter examines the differential effect of monetary policy shocks on U.S. local housing markets. By exploiting the heterogeneity in housing supply elasticity, I provide estimates of local housing price responses to monetary policy shocks in a large sample of metropolitan statistical areas. Given an expansionary shock that decreases the Federal Funds rate by 100 basis points, housing prices increase by 7.2% in cities with a highly inelastic housing supply (e.g., San Francisco), but by only 1.0% in cities with a very elastic housing supply (e.g., Iowa City) at the two-year horizon. To understand the monetary policy transmission mechanism in the housing market, I develop and estimate a structural model of the housing price with information friction. The third chapter surveys a number of important methods on the identification of monetary policy shocks and compares their estimated impacts on output, inflation and unemployment rate for pre- and post-1984 periods. In particular, identification using monetary SVAR or Romer- Romer method suggests substantial changes in the effects of monetary policy shocks. In contrast, the FAVAR method provides relatively consistent estimation for both periods. Tests for structural breaks point to parameter instability during 1979 - 1984. Such instability persists after accounting for GARCH effects.