Essays on the analysis of changes in the conduct of monetary policy

This dissertation analyzes the evidence and the effects of changes in the U.S. conduct of monetary policy.;The first chapter revisits the evidence on the stability of U.S. macroeconomic relations. Despite a substantial body of evidence, a clear answer has yet to emerge, as opposite conclusions are still reached, even within very similar empirical settings. I find this evidence to be very sensitive to the testing strategy and the specification of the empirical model. In particular, it varies greatly depending on the tests used and is importantly reduced when the number of parameters entering the empirical model increases. Moreover, some tests, such as functionals of an LM version of the Chow test, suffer from serious size distortions. Based on this analysis I conclude that there is compelling evidence of instability, with monetary policy being a potential source.;The second chapter considers the estimation of time-varying parameter (TVP) models in the linear instrumental variable (IV) framework. Consistency and asymptotic normality are shown for a class of estimators with a multiplicatively separable weighting of the observations, as suggested by the structure of the TVP model. In particular, the class contains estimators that are robust to the presence of serial correlation and heteroskedasticity in the model disturbances. The optimal estimator is derived and applied to a simple forward looking policy reaction function, where a rational expectation assumption requires the use of IV methods. The results indicate substantial movements in the Fed's response to the economy over the 1960:1--1996:4, which do not appear to be completely characterized by a single discrete shift occurring in October 1979.;The third chapter provides a more sophisticated account of the historical conduct of monetary policy and its effect on the economy. Monetary policy is modeled in the context of the Bernanke-Mihov (1998) structural VAR extended to allow explicitly for the Fed's forward looking behavior by including data on its real-time forecasts of inflation and unemployment. Instability is modeled through the TVP model considered in Chapter 2. The results suggest an important shift towards a greater response to inflation. While the bulk of these changes were brought about under Volcker, they fail to explain the high interest rates and the recession of the early 80's. Together with the results obtained for the period around the 1990 recession, the identified changes in monetary policy do not appear to explain the timing of business cycle fluctuations.