Momentum Strategies: Returns From Trend-Following, Data-snooping, and Market Efficiency
This paper focuses on the exploration of the returns to trading strategies based on the factors of momentum, size, and value strategies. Given the evaluation in the literature with respect to the occurrence of momentum as an anomaly, this paper seeks to further explore such returns to momentum while evaluating it on assessments against data-snooping under the mean performance and Sharpe ratio criterion. Additionally, in order to minimize the effects of data-snooping---a problem that occurs when a set of data is used multiple times with different parameters while making ex-post statistical inference---on trading strategies, this paper employs a multiple hypothesis testing framework developed by White (2000), known as the reality check method. Furthermore, to test the specifications and significance in terms of the different combinations to the studied momentum strategy, I utilize the framework proposed by Giacomini and White (2006), known as the conditional predictive ability test. Using monthly U.S equity data from January 1926 to December 2014, this paper finds that the results to momentum from trend following significantly outperforms the returns to the U.S S&P 500 index while accounting for influences of data-snooping in terms of both mean and Sharpe ratio assessments. As a result, this paper indicates evidence against the weak form of the efficient market hypothesis. Moreover, findings also suggest that returns to momentum portfolios are able to outperform that of institutional returns under the mean performance criterion, but not under the Sharpe ratio criterion.