Bootstrap rolling-window Granger causality dynamics between momentum and sentiment: implications for investors

Abstract

This paper seeks to examine the unidirectional versus bidirectional Granger causality between investors’ sentiment and momentum strategies. It is based on the full sample Granger causality test and the recent rolling-window bootstrap approach. We also applied a probit model to the extent to which the probability that investors’ sentiment and momentum strategies influence each other. Our results suggest bidirectional Granger causality between investor sentiment and momentum strategy with unstable causality dynamics over time. We find that ADS and VIX positively affect the likelihood that investor sentiment Granger causes momentum strategy and negatively impact the probability that momentum strategy Granger causes investor sentiment. Gold harms the likelihood that investors’ sentiment and momentum strategies affect each other. The research design is unique to combine bootstrap rolling-window Granger causality tests between Sentiment and Momentum to assess investors’ implications in terms of confidence, uncertainty, aggressiveness, or optimism versus Pessimism.