Momentum effect and investor sentiment : study on the European stock market

Abstract

In the past three decades, the momentum effect has been the object of a tremendous amount of research. Scholars have used various methodologies to try to explain this market anomaly. To this date however, even though studies mostly recognized that investor irrationality was the cause of it, several investor biases were advanced and none of them has been acknowledged as a true explanation. This research focuses on the time variation of momentum rather than its cross-sectional effect. More particularly, it investigates the impact of investor sentiment for the profitability of momentum strategies. This study was conducted on the European stock market and during a period ranging from January 2004 to March 2019. Two questions are asked in this research: Does investor sentiment have a contemporaneous impact on momentum returns? Does investor sentiment have predictive power for the momentum returns? In addition to those, an indirect question of this research is: Does a momentum strategy still yield profits for the period concerned? The answer is no at least for the period post-crisis after 2009. After this period, the momentum returns are not significant. In order to measure investor sentiment, measures are chosen from the literature: one replicating the measure of Antoniou et al. (2013) which computes sentiment levels and one replicating the measure of Da et al. (2015) and Gao et al. (in press) which computes sentiment changes. In addition, two control variables are used for the first measure: market volatility and market state. This choice is motivated by Wang and Xu (2015) who found robust predictive power of market volatility over sentiment. For the second measure, the same set of control variables as Da et al. and Gao et al. are chosen. Concerning the predictive power of investor sentiment, all measures and specifications show that it is not a robust predictor. Concerning the contemporaneous explanative power of momentum, the analyses yield mixed results. Investor sentiment level has a robust explanative power on momentum returns when using the full sample however, this significance disappears when removing the year 2009 or when a Fama-French (Fama & French, 1993) correction is applied to the momentum monthly returns. Overall, the impact of the control variables is also dependent on the sample period and the risk-adjustment of the returns. The influence of investor sentiment change is also not robust when using risk-adjusted returns. The variable with the strongest and robust explanative power on momentum returns is the Business Climate Indicator computed by the European Commission. The influence of investor sentiment on the profitability of momentum strategies is not statistically verified in this study.