Bad Luck, Bad Policy, and Learning? A Markov-Switching Approach to Understanding Postwar U.S. Macroeconomic Dynamics

In this paper we analyze changes in the Federal Reserve behavior and objectives since the1960s justified by potentially evolving beliefs—through a real-time learning process—aboutthe structure of the economy and shifts in policymakers' preferences in the late 1970s. In addition, we allow for changes in the volatility of the structural shocks in a Markov-switchingDSGE model. We evaluate the relative contribution of each narrative to the explanation ofthe Great Inflation and the Great Moderation. We argue that the interplay between centralbank learning and a shift in policy makers' preferences explains movements in the monetaryinstrument. In addition, the model captures non-policy related high volatility periods clusteredaround the mid-to-late 1970s, specifically supply side shocks that behaved as destabilizingforces driving macroeconomic fluctuations. We also find that the monetary policy shift towardinflation stabilization (in the late 1970s) changed the composition of output making themarginal efficiency of investment shock the most important contributor to the business cycle.To conclude, we observe that a change in monetary policy objectives, assumptions about policymakers' learning process, and Markov-switching volatility are key to fit the model to U.S.post-war data