A Tale of Two States: An Application of a Markov Switching Model to Anomaly Returns

Abstract

The time-varying profitability of equity anomalies calls for a useful tool to select the winning investment strategies from the loser investment strategies. We offer a new framework for dynamic asset allocation across the anomalies based on a Markov regime switching model. Using a sample of eleven equity anomalies from the US equity market from the years 1963 to 2016 we demonstrate the predictability of their performance. The anomalies forecasted to be profitable significantly outperform the remaining anomalies by 0.15–0.43% per month. The results are robust to many considerations.