Time Calendar and Fiscal Year Calendar Patterns in Momentum Profits

Abstract

A strong conjecture in literature is that momentum profits can be ascribed - at least in part - to the (inefficient) incorporation of new earnings information in the pricing of stocks. A similar hypothesis is set to explain the post-earnings announcement effect. So far empirical research on the momentum effect is strictly separated from empirical research on post-earnings announcement drift. A link between these two strands of literature is hindered by the different timing nature of earnings announcements - (which are events) while momentum strategies involve ranking and holding periods of 3-12 months. This barrier is lifted by proving that past-return strategies with an L-month ranking period can be interpreted as a convolution of L successive past-return strategies with a one-month ranking period. A focus on strategies with a one-month ranking period is sufficient to study momentum strategies. The focus on one-month strategies opens new perspectives to disentangle multiple sources contributing to momentum profits. The confinement to one-month strategies goes together with the possibility to study time calendar, fiscal year calendar and earnings-reporting calendar patterns in the momentum effect. Momentum profits are at a maximum when selecting winner and loser stocks in months with earnings announcements. About half of the momentum profits are related to post-earnings announcement drift. Apart from strong earnings calendar patterns we find strong time calendar patterns, perhaps due to portfolio rebalancing by fund managers at the end of the year.