The Mystery of Momentum

The three big anomalies in stock market behaviour that academics have identified over the past 50 years are these:

  • Small cap stocks tend to outperform large cap stocks over time;
  • Value stocks tend to perform better than growth stocks; and
  • Momentum, or buying the winners of the recent past and selling the losers, is a surprisingly effective strategy.

Of these three the most recent to be documented, and in some ways the most surprising, is the last.

After all, if it was so easy to put together a winning strategy by following a simple “buy winners, sell losers” rule, you would expect so many investors to adopt the strategy that it would cease to work. This is exactly what happened in the 1980s, when so many investors tried to implement the recently documented small company effect that the strategy stopped working for several years afterwards.

Yet, as Dimson, Marsh and Staunton, the London Business School academics, reported yesterday as they unveiled in London the latest edition of their annual Global Investment Returns survey, the momentum effect has been both powerful and persistent, despite all the publicity that it has attained in recent years, which should by rights have killed it off as an observable phenomenon.

Using their comprehensive database of market returns since the year 1900, they find that a simple strategy of buying those of the largest 100 stocks in the UK market which have outperformed over the previous 12 months and then holding them for six months has produced an annualised real return of 14.3 per cent per annum, which is well above the long term return on the market as a whole.

In contrast buying the losers, those large cap stocks which have underperformed the market over the previous year, produced a return over the period 1900-2011 of just 3.7% per annum. Allowing for the wonders of compound interest, the winners strategy would have produced a final sum that was 50,000 times as large as the losers strategy over the 110-year period.

The outperformance of momentum, they conclude, is both more powerful and more persistent than any academic theory can currently explain. It has been validated repeatedly in academic studies, and applies not just to the US and UK markets, but to many other global markets as well. Only the Japanese stock market, in this as in so many other things, appears to be immune to the momentum effect.

On average the outperformance over time amounts to a remarkable 0.75% per month, and has been most pronounced in the resource-dominated markets of South Africa , Norway and New Zealand.  It is notably stronger in the UK, intriguingly, than in many other markets. Challenged on this point, Prof Marsh replied: “As we do not have an explanation for the momentum effect, it is difficult to know why it should work better in the UK than elsewhere”.

Given that it is a work of a moment for a hedge fund or quant trading house to programme a simple strategy of this kind, the reason that it has not been more widely adopted remains a mystery. One problem is that it does require a high level of trading. Buying and selling the requisite amount of stocks every six months, as the most common rule tested in academia requires, generates considerable trading costs, which eat into – though they do not eliminate – the excess returns.

A more subtle problem is that the strategy works very well when it works – but when it fails, the failure can be spectacularly bad. The momentum effect, in other words, is persistent, but not reliable. A good example took place in 2009, when the stocks that performed best on a calendar year basis were mostly those that had done worst during the credit crisis, the reverse of what the momentum effect would have led you to believe.

The result is that any hedge fund, wealth manager or family office which attempts to follow a momentum strategy will tend to do well for a number of years, but then run into an unexpected calamitous year which results in a catastrophic and potentially terminal loss of business. In fact there have been three years since the year 2000 when buying the winners and selling the losers has produced big negative returns, of the order of 25% on an annualised basis.

Moreover, these three occasions have all been years when the markets overall have produced positive returns, so the underperformance of momentum investors – losing 25% in a year when the markets have risen strongly – has been even greater. However good your long term returns, you only look foolish as a professional investor if you lag behind in such good times.

For any professional investor worried about losing clients, the business risk of putting all their eggs in the momentum basket is therefore often too great to make it a practical reality. Hedge fund investors, it is well known, for example, rarely tolerate even one year of large drawdown, so only the bravest hedge fund is likely to stick with a strategy that has such potential downside to it. That could be one reason why momentum continues to work as a strategy.

Although the existence of persistent returns to momentum only came to be accepted as an empirically sound observation in the 1990s, Prof Dimson recalls that it had in fact been written up by a number of pioneering academics as early as the 1960s, at a time when comprehensive and reliable long run series of market returns (let alone the computers to analyse them in a matter of minutes) had still to become widely available.

At the time unknown academics such as Sydney Alexander who put forward the idea that momentum might work found themselves roundly derided by the most important figures of the day in academic finance. It has taken 40 years or more for their initial hypotheses to be validated by universally accepted data. Since momentum remains a workable and important factor, it is not surprising that it plays such a large part in many forms of technical analysis.

I know many successful stock market investors who try to incorporate it into their thinking, albeit not in the mechanical portfolio-wide manner that the LBS academics have been researching. Yet the mystery of why it works so well in the first place remains just that – a mystery, and one of the most intriguing which the stock market has to offer.