Quantitative easing and tapering uncertainty: Evidence from Twitter

In this paper we analyze the extent to which peoples' changing beliefs about the timing of the exit from Quantitative Easing ("tapering") impact asset prices. To quantify beliefs of market participants, we use data from Twitter, the social media application. Our data set covers the entire Twitter volume on Federal Reserve tapering in 2013. Based on the time series of beliefs about an early or late tapering, we estimate a VAR model with appropriate sign restrictions on the impulse responses to identify a belief shock. The results show that shocks to tapering beliefs have profound effects on interest rates, exchange rates and asset prices. We also derive measures of monetary policy uncertainty and disagreement of beliefs, respectively, and estimate their impact. The paper is the first to use social media data for analyzing monetary policy and also adds to the rapidly growing literature on macroeconomic uncertainty shocks.