Can discretionary increases in government spending stimulate the economy? We answer this question by taking into account both the information flow on fiscal measures and the role played by information frictions. Using a novel set of empirical proxies for fiscal news and agents’ misperceptions, our approach identifies three types of innovations to government spending that modify the agents’ information set at different horizons: before, upon and after the actual change materialises. Borrowing from the psychological literature, we name them expected, unexpected and misexpected fiscal changes. By missing this important distinction, we show that standard identification strategies blend unexpected and misexpected changes in a way that leads to significant underestimation of the effects of fiscal policy. An application to US data reveals that once information rigidities are fully accounted for, expected fiscal changes stimulate economic activity and private investments with a cumulative output multiplier around 1.5.
fiscal foresight fiscal shocks government spending government spending news large Bayesian VARs structural VARs Survey of Professional Forecasters