Using VARs to Identify Models of Fiscal Policy: A Comment on Perotti

Abstract:     
This note comments on Perotti's (2008) estimates of the impact of a government spending shock on the economy. In the process, it makes two points. First, it notes that with enough freedom to pick the dynamics of policy variables, the neoclassical model can generate any set of observations for the non-policy variables. Second, it proposes a method to identify the policy dynamics in theoretical models by using the estimated impulse responses of the policy variables from VARs, and in this way generate testable predictions of the model for the non-policy variables.

Keywords: Fiscal policy, VAR, identification

JEL Classifications: E62, E20, C50

Working Paper Series