In this paper, I empirically examine how uncertainty about government spending policy affects economic activity by using US time series data. To this end, I build government spending policy uncertainty indexes and estimate a proxy structural vector autoregression (VAR) model. The model shows that an increase in government spending policy uncertainty has negative, sizable, and prolonged effects on economic activity. Firms’ external financing premiums seem to be an important transmission channel of government spending policy uncertainty shocks. The results also imply that the standard recursive VAR model systematically underestimates the adverse effect of government spending policy uncertainty. I also discuss the advantages and disadvantages of the proxy VAR versus the sign restriction VAR.