To Trend or to Revert - A Portfolio Perspective on Time Series Momentum
To Trend or to Revert - A Portfolio Perspective on Time Series Momentum

To Trend or to Revert - A Portfolio Perspective on Time Series Momentum

Time series momentum or trend following strategies have grown in popularity among institutional investors, with over $300 billion assets under management in managed futures strategies. A portfolio approach on recent return history of 60 liquid futures across different asset classes shows that evidence for both time series momentum and its opposite, reversal, is weak. (Page 1)

Time series momentum (TSM) or more commonly known as trend following is an investment strategy where the exposure to a reference asset is directly proportional to the sign of its past return. Cross-sectional momentum (CSM) is an investment strategy that ranks a given portfolio of assets on past returns and takes long and short positions on top and bottom ranked assets respectively. (Page 2)

Assets under management of the managed futures industry stood at $318 billion as of the fourth quarter of 20191. Assets under management of the factor industry is estimated to be nearly $2 trillion2. (Page 2)

Optimal allocations to different TSM strategies and their significance are summarised in Table 2. Table 3 has the corresponding information for CTS strategies. Evidence for both momentum and reversal is weak. It is interesting to note that 1-month momentum (T SM1), the most widely present (i.e. receives statistically significant positive allocation) among the 12 TSM strategies considered, is significant only in less than a quarter of the universe (14 out of 59). 12-month momentum, which has been the focus of many previous studies, is insignificant in all but two the assets (CAD in currencies and DAX in equity indices). (Page 4)

There is growing evidence that excess performance of many factors have disappeared in the recent years. Arnott et al. (2019) show that CAPM alpha of most factors have vanished post 2003. Publication effect, crowding and overfitting are put forth as some of the key drivers of this decline. (Page 4)

Does a portfolio of an asset and TSM strategies (with different momentum observation windows) on the asset superior outperform the asset? If it is the case, then we might want to know if there is a pattern in significance of TSM strategies on different assets, for example, if 12 month momentum works as well for S&P 500 index futures as it does for FTSE 100 futures. (Page 5)

Examination of time series momentum from a portfolio perspective using recent data shows that the evidence for both momentum and reversal is sparse. While there is limited evidence of short-term reversal contrary to Moskowitz et al. (2012), there is no pattern across asset classes. Lack of strong evidence for time series momentum is consistent with a growing body of literature on decay in abnormal returns of factors and alternative risk-premia. (Page 11)