Risky Cycles in Stock Price Momentum Strategy Returns

Price momentum strategies are widely used by Quant money managers. They generate high positive returns on average with little systematic risk measured using standard asset pricing models. During the 1002 month period covering July 1927 - December 2010 the returns on one such widely studied strategy using U.S. stocks generated an average return in excess of 1.18%/month with a Sharpe Ratio of 0.15. In contrast the average excess return on the US stock index portfolio was 0.62%/month with a Sharpe Ratio of 0.11. However, momentum strategies incur periodic but infrequent large losses: There were 13 months with losses exceeding 20%/month. We provide an explanation for this pattern in the data, and show that we can characterize such periodic risky cycles in momentum returns with a two state hidden Markov model, where one state is turbulent and the other is calm. All the 13 months with losses exceeding 20%/month occur during turbulent months, i.e., months when the predicted probability of the hidden state being turbulent exceeds 0.5. When months with predicted probability of being in the turbulent states are avoided, the Sharpe Ratio of momentum strategy returns increases to 0.30 and momentum becomes more of a puzzle. The study will provide new insights into the nature of the risk in widely used momentum strategies. We will use return data from CRSP and financial statements data from Compustat. Our findings will be of interest to money managers, plan sponsors, regulators, and academics. Description of the research and potential findings Relative strength strategies have been and continue to be popular among traders. One such strategy, commonly referred to as momentum that involves ranking stocks based on their past performance and buying the winners and selling the losers, has received wide attention in the academic literature as well. Long-Short momentum portfolios have generated large historical average returns with high Sharpe Ratios. For example, ranking stocks based on their past 11 month returns, and holding them for one month into the future (after skipping a month to allow for market microstructure effects) generates a historical average return of 1.18%/month with a Sharpe Ratio of 0.15 during the period 7/1927-12/2010. In contrast the average return in excess of the risk free rate on the portfolio of all exchange traded stocks was 0.62%/month with a Sharpe Ratio of 0.11 during the same period. While stock price momentum strategies produce high abnormal returns, they also exhibit periodic infrequent large losses. We show that such large losses are more likely to occur during turbulent market conditions that can be to some extent anticipated. Levy (1967) was one of the early academic articles to document the profitability of stock price momentum strategies. Jensen (1967) raised several issues with the methodology employed by Levy (1967. The topic did not receive much attention in the academic literature till Jegadeesh and Titman (1993) came up with a clever method for constructing portfolios based on relative price momentum in stocks that is rigorous and replicable. A number of studies have confirmed the Jegadeesh and Titman (1993) findings using data from markets in a number of countries (see e.g. Rouwenhorst (1998)), and in a number of asset classes (Asness et al. (2008)). As Fama and French (2008) observe, the “abnormal returns” associated with momentum “are pervasive”. The literature on momentum is vast, and can be grouped into three categories: (a) documentation of the momentum phenomenon across countries and asset classes (b) characterization of the statistical properties of momentum returns, and (c) theoretical explanations for the momentum phenomenon. We make a contribution to (b) above. Our work is closely related to Grundy and Martin (2001) who explain why the CAPM beta of momentum returns systematically vary in a particular way with past returns on the stock market; Boguth, Carlson, Fisher and Simutin (2011) who document momentum returns having option like features; and Daniel (2010) who finds that large momentum strategy losses occur during periods when the market recovers sharply following steep losses. We argue that the option like features of momentum strategies should be more pronounced during turbulent market conditions. When we avoid investing in momentum strategies during months which are predicted to be turbulent, the Sharpe ratio of momentum strategy returns increase substantially and momentum returns become more of an anomaly The paper will be organized as follows. We will provide a brief review of related literature in Section 2. We will show in Section 3 that the option like features of momentum returns will be more pronounced during turbulent periods. We will develop the econometric specifications of the hidden Markov model for characterizing momentum returns in Section 4; describe the data in Section 5; and discuss the empirical findings in Section 6. We will conclude in Section 7. The appendix will describe the algorithm we use in our estimation.