Toward a Taylor Rule for Fiscal Policy
Posted: 17 Nov 2015
Date Written: June 28, 2013
Abstract
In DSGE models, fiscal policy is typically described by simple rules in which tax rates respond to the level of output. We show that there is only weak empirical evidence in favor of such specifications in US data. Instead, the cyclical movements of labor and capital income tax rates are better described by a contemporaneous response to hours worked and investment, respectively. We show that conditioning on these variables is also desirable from a normative perspective as it significantly improves welfare relative to output-based rules.
Keywords: Fiscal policy, Bayesian model estimation, Identification, Variable selection
JEL Classification: E62, H30, C51
Suggested Citation: Suggested Citation