Simultaneous identification of fiscal and monetary policy shocks

Abstract

This research contributes to the literature on the effects of fiscal and monetary policy by exploiting non-Gaussianity of the time series for the identification of a Bayesian structural vector autoregression model. Using quarterly US data from 1954:IV to 2006:IV and from 1985:I to 2020:III, we formally assess the plausibility of theoretically predicted signs to label fiscal policy, monetary policy, and business cycle shocks. The impulse responses of consumption to the fiscal policy shock depend to some extent on the sample period, but the implied fiscal multiplier is always less than unity. On investment, there is a lagging crowding-out effect with a high probability, but it is not strongly evident in the latter sample. As for the responses after a contractionary monetary policy shock, we find a weakening output after some lags consistent with the leading monetary policy literature. The business cycle shock turns out to matter for government spending only in the long run, while it is already important for the federal funds rate in the short run.