News Shocks and Business Cycles: Evidence from Forecast Data∗

This paper proposes the use of data on expectations to identify the role of news shocks in business cycles. This approach exploits the fact that news shocks cause agents to adjust their expectations about the future even when current fundamentals are not affected. Using data on expectations, we estimate a dynamic, stochastic, general equilibrium model that incorporates news shocks for the U.S. between 1955Q1 and 2006Q4. We find that the contribution of news shocks to output is about half of that estimated without data on expectations. The precision of the estimated role of news shocks also greatly improves when data on expectations are used. Although news shocks are important in explaining the 1980 recession and the 1993-94 boom, they do not explain much of other business cycles in our sample. Moreover, the contribution of news shocks to explaining short run fluctuations is negligible. These results arise because data on expectations show that changes in expectations are not large and do not resemble actual movements of output. Therefore, news shocks cannot be the main driver of business cycles. JEL classification: C11, E12, E13, E32