Probability and seasonality analysis of momentum

Abstract

This paper analyzes momentum by using probability. Winners and losers tend to stay in winner and loser portfolio over at least thirty-six months. In addition, the winner and loser persistency is highly correlated with the momentum strategy profits not only in cross-sectional data but also in time series data. Moreover, the winner and loser persistency tends to have relatively high quarter peaks. This quarter seasonality has an influence on momentum strategy profits. Quarter momentum strategies always outperform non-quarter momentum strategies in our sample period. This strengh of three month interval momentum strategy is independent on idiosyncratic risk and variance of stock returns, as well as earnings momentum.