Swedroe: Strategy Cocktails Can Outperform

February 21, 2018

There is a large body of academic research demonstrating that time-series momentum—buying instruments with recent positive returns and selling instruments with negative returns over the same period—produces attractive performance across not only individual stocks, but also equity indexes, fixed income, currencies and commodities. There is also a substantial body of evidence demonstrating that the carry trade—the tendency for higher-yielding assets to provide higher returns than lower-yielding assets—predicts future returns in every asset class.

My latest book, “Your Complete Guide to Factor-Based Investing,” co-authored with Andrew Berkin, both summarizes the research and provides explanations for why we should expect these factors to continue to deliver premiums and add to portfolio efficiency.

The time-series momentum and carry factors appear to be two distinct risk premiums; there is no obvious link between them, such as there is between carry and value (both buy what is cheap and sell what is expensive). Marat Molyboga, Junkai Qian and Chaohua He, authors of the November 2017 study “Time-Series Momentum, Carry and Hedging Premium,” contribute to the literature by investigating whether carry and time-series momentum are related, and then by exploring the potential performance implications of that relationship.


Recent Study

The authors examined the performance of three time-series momentum strategies (long only, short only and combined long-short strategies, including a strategy that targets 40 percent volatility) across 65 futures markets from all major asset classes, including equity indexes, fixed income, currencies and commodities, for the period between January 1975 and December 2016.

They also investigated whether time-series momentum and carry are related to the hedging premium by examining the positions of hedgers in the Commodity Futures Trading Commission’s Commitment of Traders (COT) reports. This permitted the authors to compare hedgers’ positions against the long and short signals of time-series momentum and carry strategies for a subset of 41 futures contracts with position data available in the COT reports over the period from January 1986 through December 2016. The net hedger position is defined as the difference between commercial long and short positions normalized by the total open interest. The net speculator position is calculated as the difference between non-commercial long and short positions scaled by the total open interest.

The following is a summary of Molyboga, Qian and He’s findings:

  • There is significant cross-sectional variability in returns and roll yield (referred to as basis, which is the percentage difference between the spot price and the price of the nearby, or closest-to-maturity, futures contract) across and within asset classes. The annualized return for individual futures markets ranges from -1.8 percent to 7.1 percent for equities, with an asset-wide average value of 3.9 percent; from -19.7 percent to 9.1 percent for commodities, with an asset-wide average value of -1.7 percent; from -0.1 percent to 4.6 percent for fixed income, with an asset-wide average value of 1.9 percent; and from -0.7 percent to 3.2 percent for FX, with an asset-wide average value of 0.7 percent. This result indicates there could be a strong benefit from identifying the factors driving that dispersion.
  • The basis between spot and futures contracts (carry) explains approximately 36 percent of time-series momentum’s performance, indicating that time-series momentum and carry are related.
  • Consistent with prior research, basis has a significant impact on returns. The results were statistically significant at the 1 percent confidence level.
  • Conditioning trading signals on the sign of the basis improves the Sharpe ratio of the time-series momentum strategies by approximately 0.17, to 0.92, and is robust across sub-periods, economic regimes (expansions and contractions), choice of position-sizing in the implementation of time-series momentum strategies, and the lookback period used in the calculation of the basis. The improvement in performance is particularly strong during the early stages of recessions, which tend to exhibit poor stock market performance, providing a hedge for equity exposure when it is needed the most (when marginal utility of wealth is highest). Again, the results were statistically significant at the 1 percent confidence level.
  • Hedgers’ positions load negatively on time-series momentum, suggesting that time-series momentum is likely benefiting from the hedging premium. This finding is robust for lookback periods between one and 12 months for a broad array of markets. In other words, time-series momentum is capturing the hedging premium. There is much weaker evidence that carry captures the hedging premium as well.

The authors concluded: “Time-series momentum and carry are related because both strategies benefit from the time-series and cross-sectional variability in basis, and yet they are distinct because time-series momentum alone is linked to hedging premium.” They add that combining the strategies can “substantially improve” investors’ welfare.



Molyboga, Qian and He have provided new insights that can be used to improve the performance of two well-documented strategies. It will be interesting to see if their findings get implemented into live mutual funds.

Note that the study did not consider transaction costs, so we don’t know their impact on performance. However, the research has shown that even higher-turnover strategies, such as momentum, can survive transaction costs with patient-trading strategies.

Finally, the authors’ findings shouldn’t really come as a surprise because the academic literature is filled with papers demonstrating the benefits of combining factor strategies shown to add explanatory power in the cross-section of returns (they are unique/independent factors) while also providing premiums that are persistent across time and economic regimes, pervasive across the globe and asset classes, robust to various definitions, intuitive (meaning they have risk-based or behavioral-based explanations for why we should expect them to continue), and implementable.

For example, it has been shown that conditioning value strategies on the profitability (or quality) and/or momentum factors improves efficiency. This evidence may also serve as a good reason to consider multi-style equity funds.

Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.

Find your next ETF