Trend following: equity and bond crisis alpha

From left: Carl Hamill, Sandy Rattray and Otto van Hemert.

Over the past few years we have seen futures trend-following being used as a tool for institutional investors looking for “crisis alpha” — a term first used by Kathryn M. Kaminski and defined as profits gained by exploiting persistent trends occurring across markets during times of crisis. For investors, this means a strategy that has the potential to deliver outperformance during (typically equity) market stress periods.

Indeed, with prices of many traditional asset classes at or near all-time highs, we could see this search continue. Futures trend-following has historically shown potential to provide crisis alpha; in the 2007-'09 financial crisis, the MSCI World index lost 49.3% while the Barclay BTOP50 index, which is an index of the largest managed futures funds, gained 16.3%. Similarly, when the Internet bubble burst at the turn of the century, the MSCI index lost 45.7% and the Barclay BTOP50 gained 36.8%. Both of these instances are clearly visible in Figure 1. What this chart also shows is a near straight-up performance for bonds since 1985. Unsurprisingly then, trend-following strategies might have benefited from the nearly consistent upward trend by holding a (predominantly) long bond position over the past 30 years.

This strong bond performance was driven by significant rates compression. Many economists believe yields cannot fall much further, so it is natural to ask if trend-following strategies have the potential to maintain performance in the absence of a bond market tailwind, and indeed if they can protect against potential stress in bond markets of the form seen in the 1960s, '70s, and early '80s.

In a recent paper, we shed light on these questions by studying simulations of trend-following strategies from 1960 onward. Importantly, this includes the pre-1985 period, which shows a starkly different picture compared to post-1985. Over the 1960-1985 period, bonds experienced negative excess returns on average, while stock markets provided modest positive average excess returns and quite frequent drawdowns. Figure 2 shows the result for different equity (panel A) and bond (panel B) performance quintiles.

Source: Bloomberg, Man database

The results in panel A should be familiar to readers with an interest in trend-following strategies. The simulation presents returns that are strongest when equity markets are at their weakest. It also shows that, historically, trend-following strategies have performed well when equity markets are performing well. These observations give rise to the well-known “equity smile.” What perhaps is less familiar are the results in panel B, which mimic closely those in panel A. This indicates that, historically, trend-following strategies also perform strongest when bond markets are at their weakest or strongest. In other words, there exists a smile in bonds just as there is a smile in equities. If anything, the smile in bonds is even more pronounced than that of equities in our simulations.

The four asset classes — equities, commodities, currencies and bonds — exhibit positive returns across all quintiles, which we feel makes a good case for the potential value of diversification in trend-following strategies. We believe this counteracts the assertion that trend following has only been successful due to the 30-year rally in fixed-income markets.

We also explored in the paper how it might be possible to further enhance the crisis alpha characteristic of trend-following strategies by running versions of the simulations in which, for example, positions in equities are capped at zero. This is designed to help ensure a trend-following strategy is better positioned during periods of equity market decline as it can never be long. We found that position capping improved performance of a model trend-following strategy over the performance of our baseline unrestricted case in quintile one — in other words, during the weakest equity markets. Performance of quintiles three to five and the overall average performance is reduced, however, which can be seen as the price one pays for the enhanced crisis alpha return profile.

In conclusion, our study shows that the well-known “crisis alpha” characteristics of trend-following strategies during equity market weakness have historically had the potential to extend readily to bond markets. Thus we believe our results indicate the potential for futures trend-following to perform positively even in an environment where government bond yields rise.

Sandy Rattray is CIO of Man Group and CEO of Man AHL; and Otto van Hemert is head of commodities and Carl Hamill is quantitative analyst at Man AHL. All are based in London.