Fiscal Policy Rules and Regime (In)Stability: Evidence from the U.S.

Abstract

We employ Markov-switching regression methods to estimate fiscal policy feedback rules in the U.S. for the period 1960-2002. Our approach allows to capture policy regime changes endogenously. We reach three main conclusions. First, fiscal policy may be characterized, according to Leeper (1991) terminology, as active from the 1960s throughout the 1980s, switching gradually to passive in the early 1990s and switching back to active in early 2001. Second, regime-switching fiscal rules are capable of tracking the time-series behaviour of the U.S. primary deficit better than rules based on a constant parameter specification. Third, regime-switches in monetary and fiscal policy rules do not exhibit any degree of synchronization. Our results are at odds with the view that the post-war U.S. fiscal policy regime may be classified as passive at all times, and seem to pose a challenge for the specification of the correct monetary-fiscal mix within recent optimizing macroeconomic models considered suitable for policy analysis.