Potential gains from predicting the timing of stock market persistence and mean reversion

Abstract

This paper undertakes to investigate the effectiveness of market timing between prior winners and losers in the global equity markets using Monte Carlo simulation over the period from 1 January 1999 to 31 December 2009. The winner and loser portfolios of 100 stocks are constructed based on the prior 36-month U.S. dollar returns of the Dow Jones (DJ) Sector Titans Composite constituents. The market timer is assumed to have varying accuracy in predicting market persistence and mean reversion, and switch between the winner and loser portfolios on a quarterly basis based on his/her predictions. Sensitivity analysis is conducted to determine whether it is more important to predict the timing of persistence versus mean reversion. The study results reveal that an effective market timing strategy could be devised for market timers with modest ability to predict the timing between global equity market persistence and mean reversion. Greater benefits are derived from improvements in the mean reversion timing accuracy versus persistence timing accuracy, even though only 19 out of 44 quarters are classified as periods of mean reversion in the examination period. The results from sensitivity analysis support the view that it is more important to predict the timing of mean reversion correctly than persistence. This outcome could be attributed to the resilient nature of the loser portfolio in turbulent times. The observation that the majority of the persistent quarters are bullish (65.52%) while the majority of the mean reversion quarters are bearish (60%) provides evidence of investor overreaction in the global equity markets.