Title: Measuring the impact of President Trump's tweets on economic uncertainty: A narrative approach
Authors: Hector Daniel Perico Ortiz - Friedrich-Alexander-Universität Erlangen-Nürnberg (Germany) [presenting]
Abstract: The causal relation between President Trump's tweeting behavior and market uncertainty at a high-frequency level is investigated. We implement an economic narrative approach based on identification of economic narratives from Twitter data using machine learning and estimating the effect of them, using time series regressions, on a high frequency measure of market uncertainty, given by the five-minute change in the VIX index. The results suggest that major economic narratives regarding foreign trade and policy, and monetary policy, have a significant effect on market uncertainty in the period of one hour and three hours after the narrative event. Immigration narrative is also significant at the five hours horizon. Furthermore, behavior events, such as increases in the tweet or retweeted counts above their average, matter for market uncertainty. A similar analysis at the daily frequency level using the EPU index as uncertainty measure provides similar results at longer time horizons.