Essays on the Identification of Fiscal Policy Behavior

Abstract

The identification of fiscal authority's stance on debt stabilization has become of crucial practical importance. This dissertation studies two different approaches to identify such fiscal policy behavior. The first chapter shows that the current limited-information approaches to identify such fiscal policy behavior, without taking into account the intertemporal equilibrium condition (IEC) of debt valuation and the monetary policy, are, in general, subject to a simultaneity bias problem and, therefore, the corresponding inferences on such fiscal policy behavior may be misleading. The current approaches used include the Ordinary Least Squares (OLS) regression, the Generalized Method of Moments (GMM) and the bivariate Vector Autoregression (VAR). To correctly identify such fiscal policy behavior, the first chapter proposes a simultaneous equations approach, which incorporates both the IEC and the monetary policy and avoids the simultaneity bias problem a priori. Based on the post-World War II U.S. data, the results show that the U.S. primary surplus' response to the state of government debt is, on average, much weaker than those estimated by the OLS and GMM methods. More importantly, the estimated response is statistically insignificant, which is completely different from the OLS and GMM methods. The results also show that the contemporaneous response of the U.S. total government liabilities to a positive surplus shock is not significantly negative, which is in sharp contrast to the typical results of the bivariate VAR. In the second chapter, it is recognized that the identification of such fiscal policy behavior is a general equilibrium problem, which requires that monetary policy behavior always be considered simultaneously. Therefore, the second chapter takes a holistic view and jointly identifies monetary and fiscal policy behavior for the U.S. by estimating a standard New-Keynesian sticky-price model with Bayesian methods. By applying Bayesian model comparison techniques to the U.S. Pre-Volcker and Post-1982 samples, the second chapter finds out that (1) both samples favor determinacy over indeterminacy; (2) active fiscal policy is not detected in either sample. The findings are consistent with the previous literature for the Post-1982 sample, but not for the Pre-Volcker sample.