Tuesday, February 19, 2008

Looking for momentum? Check outside the US

Momentum vs. mean-reversion has been a perennial theme in investing, not least quantitative investing. My contention has always been that momentum strategies are generally less reliable than mean-reversal strategies. (See here or here.) My reader Mr. J. Rigg told me about a recent article in the Financial Times suggesting that momentum strategies are alive and well, according to the research by Prof. Elroy Dimson et al at the London Business School. The strategy is very simple: buy the stocks with the highest returns in, say, the last 12 months, short the ones with the lowest returns, and hold for, say, 1 month. If you run this strategy for the top 100 UK stocks from 1900 to 2007, the average annualized return before costs is about 10%.

There are, however, a number of caveats worth noting in this study:

First, it is very transaction-costly to implement momentum strategies for small or even mid-cap stocks. If you factor in costs, 10% can easily become 5% -- not an impressive number even for a dollar-neutral strategy. (Though one should note that the infrequent rebalancing renders transaction costs consideration less important.)

Second, the drawdown durations are quite lengthy -- sometimes exceeding 2 years. This is not acceptable performance for many hedge funds. Such lengthy drawdowns have been a common feature of many momentum strategies that I have personally studied and traded.

Third, and most interestingly, in the period 2001-2007, this momentum strategy has stopped working altogether for the US market, while continuing to deliver positive returns in other markets!

What may be the reason for this dichotomy between US and international markets? Momentum strategies generally derive their power from the slow diffusion and analysis of information: if all investors are simultaneously aware of all the relevant financial information about a company and can analyze the significance of the information instantaneously, they will have come to a consensus fair market value instantaneously and no momentum in the price will result. Hence perhaps the disappearance of momentum in the US equity market means what most people know already: that it is the most efficient equity market of all.


Anonymous said...

Love the blog but I think your assertion that the US is that much more efficient than other markets so as to produce this momentum "dichotomy" is tenuous at best. The free flow of capital and ideas is global, not just US based. Indeed, I know many traders that suggest the Australian markets are the hardest to trade. I do agree with your preference for mean-reverting strategies however.

All the best. Mark.

Ernie Chan said...

Thank you for your interesting observation about the Australian market. Certainly my belief that the US market is the most efficient is not based on a rigorous study, but merely on my observation over the years that, whenever a stat arb trading model is successful in the US, it will usually be even more profitable in the Japanese market.

Perhaps other readers here can tell us whether it is true of other international markets as well?