Government spending news and surprise shocks: It's the timing and persistence

Abstract

We identify government spending news and surprise shocks based on the Survey of Professional Forecasters (SPF) data. After clarifying the nature of the two news measures used in Ramey (2011b), we show that a measure of the news shock for the recent sample periods is still needed. Moreover, we find that news and surprise shocks may differ not only in timing but also in the level of persistence of government spending responses. Given these findings, we construct a present discounted value measure that properly captures government spending news shock of high persistence. The effects of the news shock strikingly differ from those of the surprise shock. A news shock has a significant expansionary effect on GDP, hours worked, real wage, consumption, and investment, whereas a surprise shock has a contractionary effect. These results are in contrast to previous findings (Ramey, 2011b) and inconsistent with the neoclassical view emphasizing the negative wealth effect.