Sentiment Shocks as Drivers of Business Cycles

This paper studies the role of sentiment shocks as a source of business cycle fluctuations. Considering a standard New Keynesian model of the business cycle, it introduces agents that update beliefs about the parameters of their forecasting models using newly observed data and exogenous sentiment shocks. The resulting learning model fits U.S. data better than its non-sentiment version and than its rational expectations counterpart. Sentiment is found to be an important driver of economic fluctuations, accounting for up to half of the forecast error variance of aggregate variables at business cycle frequencies. Furthermore, sentiment displays a common pattern for real GDP, investment and consumption growth, where a significant part of the sluggish recovery following a recession can be attributed to the persistent pessimistic views of agents. Sentiment also explains a substantial fraction of the high inflation experienced during the 70’s and early 80’s.