This paper develops a dynamic general equilibrium model which includes financial intermediation and endogenous financial crises. Consistent with the data, financial crises occur out of prolonged (credit) boom periods and are initiated by a moderate adverse shock. The mechanism which gives rise to boom-bust episodes around financial crises is based on an interaction between the maturity mismatch of the financial sector and an agency problem which results in procyclical lending. I show how to model these features in a tractable way, giving a realistic representation of the financial sector’s balance sheet and its lending behavior. The paper provides empirical evidence on the behavior of the U.S. financial sector’s market leverage which is (i) acyclical, (ii) rose mildly prior to the Great Recession, and (iii) increased sharply during the crisis; the model is consistent with these empirical facts. It also predicts and replicates the Great Recession, when confronted with a historical series of structural shocks. Finally, the framework is extended to include price rigidities, nominal debt contracts, and monetary policy. Within this version, I analyze the impact of monetary policy on financial stability and show that a U-shaped pattern of the policy target rate is most likely to increase financial instability.
The Time-Varying Effect of Monetary Policy on Asset Prices
Federal Reserve Bank of San Francisco Working Paper 2017-09
Latest Version: May 2017. [Paper]
This paper estimates the time-varying responses of stock and house prices to changes in monetary policy in the United States. To this end, I augment a time-varying vector autoregressive model (VAR) with a series of monetary policy surprises obtained from federal funds futures, as a proxy for structural monetary policy shocks. The series of surprises enters the model as an exogenous variable and I show analytically that this approach gives identical relative impulse responses compared with an identification that uses the proxy as an external instrument within a constant parameter VAR. However, the exogenous variable approach allows for a convenient and tractable extension to a time-varying parameter VAR that is estimated with standard Bayesian methods. The results show that stock and house prices have been less responsive to monetary policy shocks during periods of high and rising asset prices. Moreover, I find that attempts by the Federal Reserve to lean against the house price boom before the Great Recession would have come with the risk of large deviations from its output target.