Is Government Spending Predetermined? A Test of Identification for Fiscal Policy Shocks

Strategies to identify fiscal policies and their effects often use an idea that fiscal instruments cannot respond to realizations of macroeconomic uncertainties within one quarter. I evaluate the validity of this assumption in a standard estimated DSGE model, where informational subperiods are introduced to ensure fiscal policy choices are made before the current state of economy realizes. At the same time, fiscal instruments are allowed to partially respond to macroeconomic shocks, and these responses are estimated using the Bayesian method. The resulting estimates indicate that within one quarter, government spending is adjusted in response to the neutral technology shock, and the tax rate responds to realizations of the preference shock. Moreover, the model capturing contemporaneous responses of fiscal instruments to shocks provides a better fit to the data than the model where fiscal variables are completely predetermined. These results suggest that treating fiscal instruments as predetermined is misleading. Instead of identifying fiscal shocks, such a strategy identifies a combination of the shocks and other macroeconomic uncertainties. I demonstrate that the positive consumption response to the government spending shock in a Cholesky identified structural VAR model reflects the response to the technology shock, while the consumption response is negative in the estimated model.