Join our community of smart investors
OPINION

Not timing strategies

Not timing strategies
May 19, 2016
Not timing strategies

No, says Cliff Asness at AQR Capital Management. Timing factor investing - that is, investing in not just momentum but other strategies such as value or defensive stocks - is, he says, "very difficult to do well".

To see his point, let's take two factors with a proven track record: momentum and defensives (many other factors are either temporary or spurious).

Momentum has generally trended upwards relative to the FTSE 350 since 2005. There were spurts of great performance in early 2009, the winter of 2013-14 and in late 2015. And there was underperformance in the winter of 2009-10, in 2012 and in the summer of 2014.

Even with hindsight, it's not clear what was unusual about these periods and therefore what indicators might have predicted such variations in relative performance.

Instead, these episodes might vindicate a theory proposed by Andrew Lo at the MIT - that investment strategies naturally "wax and wane".

To see this, think back to mid-2011, when momentum investing had done well for years. Seeing that performance encouraged more people to become momentum investors. This drove up the prices of momentum stocks. But this drove them up too far, with the result that momentum investing did badly in 2012. That poor performance caused fair-weather momentum investors to lose heart and abandon the strategy. Their selling led to momentum stocks becoming underpriced, with the result that in 2013 momentum investing again paid off well.

In this way, there are 'factor cycles' analogous to population cycles in biology, except that they are not so regular or predictable.

Let's take another example - defensive stocks, as measured by my low-risk portfolio, which comprises the 20 lowest-beta stocks, rebalanced every three months.

These have a ratchet-type performance. They beat the market nicely in late 2008, mid-2011 and since January 2015, but with long periods of sideways relative performance in-between.

These bursts of good performance all came when the market generally did badly and defensives did what you'd expect them to do - outperform a falling market. What's odd about defensives is that they hold up well in normal market conditions. They usually only significantly underperform when the All-Share index surges, as it did in 2009. All this implies that timing defensive investing is much the same as timing the general market.

Although many believe that market timing is unwise, there are two ways of doing this. One is to look at dividend yields. Yields predict price changes, at least over longer horizons; a high dividend yield on the All-Share leads to higher returns. You'd expect, therefore, a low yield on the All-Share index to predict defensives outperforming, simply because it points to low returns on the market generally. And this is the case. In weekly data since 2005, there has been a negative correlation between the yield on the All-Share index and the subsequent six-monthly returns on defensive stocks relative to the FTSE 350. However, while statistically significant, this correlation is low, at minus 0.27 - which implies that the dividend yield can explain only 7 per cent of the variation in the subsequent relative performance of defensives. I'm not sure this is a sufficiently strong basis from which to try to time defensive investing.

A second possibility is to look simply at the time of year. We know that equities do worse between May Day and Halloween than they do from Halloween to May Day, so shouldn't defensives outperform over the summer?

They do. Since 2006, defensives have lost an average of 0.7 per cent between May Day and Halloween, compared with an average fall in the FTSE 350 of 1.9 per cent.

And there's the rub - 'lost'. If you're going to use the 'buy on Halloween, sell on May Day' rule, you should switch into cash on May Day, not defensive stocks.

All this suggests to me that defensive investing should be regarded as a buy-and-hold strategy. Defensives lose you less than other stocks in bear markets while keeping pace with the market in all but the strongest of bull markets, delivering outperformance overall on average.

Of course, everybody would like to be able to time factor investing - retail investors because it offers potentially very high returns, and fund managers because it would reintroduce a role for judgment-based investing and the high fees that attracts.

Sadly, however, wishing does not make it so. Factor timing is probably not worth trying. Momentum and defensive investing should be better regarded as long-term buy-and-hold strategies.