On front-running momentum and portfolio optimization

Abstract

Most of the empirical research on momentum in finance has been conducted using monthly data and horizons for the formation and holding period of winner and loser portfolio. This research paper studies momentum using a weekly approach and examines strategies that are more flexible than the crowded month-end approach. In particular, this paper is interested in analyzing the legal front-running of month-end momentum strategies by one to five weeks. Furthermore this study analyzes how momentum profits change by using different start dates within a month ("week-effect") as well as within a year ("month-effect") and finds that the second-last week of the month as well as the cluster of months September, October and November exhibit higher Sharpe ratios, more favorable levels of skewness and better protection against downside risk. In addition, this study demonstrates evidence that momentum investing using the widespread "monthend" view is rarely a strictly dominant strategy.