Managing momentum crashes in real time : US stock market

Abstract

Though, profitable, momentum is punctuated with crashes which make the strategy risky and unfavourable for an investor who dislikes long tails. Volatility management is one approach that has been introduced to deal with the momentum crashes. Volatility managed portfolios are portfolios that take less risk when volatility is high and high risk when volatility is low. The volatility managed strategy, though implementable, uses ex-post information which can make the results bias and impractical. We propose two tweaks in the volatility management process to make it practical. Firstly, we use cumulative statistics up to the formation month instead of the whole sample to manage the portfolio. Secondly, we use statistics of prior 10- year period to manage the volatility of the next ten years. We find that using prior information instead of ex-post information as in the two modifications do not positive results as claimed and therefore the results of Barroso and Santa-Clara (2015) are biased. We find that the kurtosis, which is an important component of managing momentum crashes, increases rather. Also, combo portfolio, a portfolio that combines value and momentum, is another method of dealing with momentum crashes. The combo portfolio reduces the standard deviation, kurtosis and skewness of momentum strategy while increasing the Sharpe ratio. We further manage the combo portfolio such that we decrease exposure when volatility is high and increase exposure when volatility is low. We find that the Sharpe ratio of the volatility managed combo portfolio of about 1.13 is higher than unmanaged combo portfolio of 0.80. The volatility managed portfolio yields positive large alpha of 0.78 per month even after controlling for exposure to the Fama-French risk factors and the unmanaged portfolio.