Tuesday, July 28, 2020

Overlapping momentum strategies - Gains from diversification of timeframe models

Momentum trading is relatively simple to implement. Look at past performance for a set timeframe, rank the returns from highest to lowest, buy the high decile and sell the low decile. Of course, controlling for other risks is more difficult. Still there is a simple way of improving returns for a momentum portfolio by using more than one timeframe. Call this a simple method of diversification for momentum trading. Testing of this momentum work using two timeframes shows improvement in excess return and  with conditional risk characteristics. This was not the core intent of recent research, but it provides some useful evidence for timeframe diversification. See "Overlapping Momentum Portfolios" by Blanco, De Jesus, and Remesal.)

For an overlapping momentum (OMOM) strategy, the researchers create long (short) portfolios that belong to the top  (bottom) decile of winners using both 6-month and 12-month prior returns. This generates a stronger filter for an equity momentum strategy and reduces the number of stocks by 50%. There are stronger excess returns, alpha, Fama-French 3 and 5 factor alphas over almost holding period periods. The OMOM risk characteristics are less compelling, but the value of OMOM for different sub-periods is also strong.

The rationale for this improvement is based on higher selectivity and differences in the diffusion of information across equities. There is a slow speed of adjustment of prices that likely allows different timeframes to provide useful return contribution. The OMOM does better for stocks that are less closely followed and have slower information diffusion. Additionally, the diversity of timing periods exploits the fact that momentum investors use different timing strategies. Hence, there is different price momentum to be exploited in models. 

The researchers provide theoretical justification for their results and are careful with the construction of their portfolios, but the overall theme is fairly simple. Diversification of momentum timeframes improves a strategy because the investment behavior of momentum traders varies. All traders do not look at momentum the same way. This is no different than the improvement that often occurs with trend followers that use different time frames to improve their signal return performance.

No comments: