We study the transmission of fiscal policy under imperfect information where government spending is composed by permanent and transitory components. Agents learn about the previous processes by only observing overall public spending and a noisy signal. Under this setting and employing maximum likelihood techniques, we construct a novel measure of fiscal policy news and show that the estimated variable agrees with the historical narrative evidence for the U.S. economy. We then use macro and micro datasets to document the effects of this proxy on real wages and consumption. The qualitative responses obtained with aggregate data are significantly the same as those using individual PSID data at the median of the empirical distributions – on impact, real consumption falls and real wages do not move, whereas both increase after one year. A potential explanation for these results relies on expectations about future policy adjustments. When we consider the tails of the distributions, real wages fall (rise) upon impact for rich (poor) households. However, the effects on consumption only differ at longer horizons where poor households increase consumption more persistently than those at the top of the distribution.
 

---

Download from REPEC

Download from SSRN

---