Trend-Following: What's Not to Like? | Man Institute | Man Group
Trend-Following: What's Not to Like? | Man Institute | Man Group

Trend-Following: What's Not to Like? | Man Institute | Man Group

Trend-following should be at least as popular as equity investing, right? Well, given equity markets are nearly 300 times the size of trend-following’s assets under management1, either these facts are not well known, or there is some other issue (View Highlight)

Since inception in 1986, the Barclay BTOP50 Index (which comprises of mostly trend-following strategies) has returned 7.0% annualised, only 0.7% shy of world stocks (Figure 1). Trend-following’s risk is significantly lower, however, whether risk is measured in terms of volatility (9.6% versus 14.4%) or maximum drawdown (-16% versus -50%). (View Highlight)

At face value, trend-following is a remarkably simple strategy. Buy something that is going up; and sell something that is going down. (View Highlight)

Try and earn a crust off trend-following in one market, however, and you might end up hungry. There’s only a slight edge, which is why managers run the strategy over tens, if not hundreds, of liquid markets to eke out consistent returns. They use computers for repeatability and because computers don’t get emotional. (View Highlight)

Since its inception, the correlation of the BTOP50 to world stocks and other traditional markets is effectively zero (Figure 2). Intuitively, this is because trend-followers seek to capture trends in all these markets simultaneously, either up or down. Our correlations also show just how little diversification is obtainable within asset classes. (View Highlight)

Long-term low correlation is one thing, but trend-following strategies have another trick up their sleeve; negative correlation to risk assets in times of crisis. (View Highlight)

The performance of a traditional equity portfolio is fairly simple to understand, particularly if it is well-diversified. It hopefully goes up in the long term, but if there is a negative headline, a global pandemic for example, you might anticipate some losses. As we show in Figure 2, diversification across regions, or even a 40% allocation to bonds doesn’t help too much here. (View Highlight)

Trend-following strategies trade many markets across multiple asset classes, not just equities. Indeed, a positive beta to equities, or risk assets in general, could originate from places other than equities; long emerging markets FX or short gold, for example. Complicating things further, this positioning can change as trends in different places emerge or dissipate. A trend-follower’s beta to any asset is dynamic. (View Highlight)

‘Medium-term’ trend-following space spans managers with trend sensitivity between two and six months, in our view. At the shorter end of that spectrum, a manager may be able to shift position quickly, being more responsive in a crisis. (View Highlight)

At six months trend sensitivity, on the other hand, the manager may respond slower to a change in market direction, but should have improved longer-term performance (see 'The Need for Speed in Trend Following’ for more detail). Market choice can also be a factor in understanding performance (see, for example, ‘Gaining Momentum'). (View Highlight)

But trend-following strategies use rules which can be written down, are based on understandable inefficiencies in markets, and computers are utilised for scalability and to remove human emotion. Surely this is a ‘transparent’ box? (View Highlight)

But think about the portfolio. If you hold a trend-following strategy alongside your traditional equity portfolio, in the long term, history suggests that both should be profitable. Further, during crises, trend-following can potentially help cushion losses. Holding the two together gives higher risk-adjusted returns and lower drawdowns than traditional assets alone. (View Highlight)

It remains a surprise to us that trend-following strategies are not more popular than they are. To finish where we started, trend-following performs as well as equities in the long term, is lowly correlated, has better risk-management properties, and generally works well when equities don’t. It works particularly well in conjunction with equity portfolios. What’s not to like about that? (View Highlight)