An empirical approach to the “Trump Effect” on US financial markets with causal-impact Bayesian analysis

Abstract

In this paper, we have tested the existence of a causal relationship between the arrival of the 45th presidency of United States and the performance of American stock markets by using a relatively novel methodology, namely the causal-impact Bayesian approach. In effect, we have found strong causal relationships which, in addition to satisfying the classical Granger Causality linear test, have been quantified in absolute and relative terms. Our findings should be included in the context of one of the main markets anomalies, the so-called “calendar effects”. More specifically, when distinguishing between the subperiods of pre- and post-intervention, data confirm that the “US presidential cycle” represents a process of high uncertainty and volatility in which the behavior of the prices of financial assets refutes the Efficient-Market Hypothesis.

Keywords

• Causality
• Causal-impact Bayesian analysis
• Efficient market hypothesis
• Market anomalies
• Calendar effect
• Economics