Absolute and Relative Momentum: A Primer (Part 1)

Despite the overwhelming evidence for relative momentum and its eventual acknowledgement by even diehard efficient-market adherents, absolute momentum (also known as trend-following or time series momentum) remains a touchy subject. Trend-following is market-timing, pure and simple, and the injunction against it is taken seriously by respectable sorts. A 2012 study by Tobias Moskowitz, Yao Hua Ooi, and Lasse H. Pedersen has poked a hole in the dam the same way Jegadeesh and Titman’s study did in 1993. The study was published in the Journal of Financial Economics, the second-most prestigious finance journal. Moskowitz, Ooi, and Pedersen found eerie profitability in a strategy that goes long a futures or forward contract when its trailing 12-month return in excess of cash is positive and goes short when it’s negative. This strategy is economically identical to the classic chartist’s strategy of buying an asset when it’s above its moving average and selling when it’s below. By recasting it and linking it to the existing literature, Moskowitz, Ooi, and Pedersen have wrapped a scientific veneer over an old chartist’s idea.

And indeed it is a very old idea. The great 19th century classical economist David Ricardo is said to have advised friends to “never refuse an option when you can get it,” “cut short your losses,” and “let your profits run on.” Charles Dow, founder and editor of The Wall Street Journal, propounded a system of technical analysis. Upon his death in 1902, his successor, William Peter Hamilton, refined what he called Dow Theory and practiced it until his death in 1929. Dow Theory is a trend-following strategy. In a now-famous study published in 1933, Alfred Cowles showed that Hamilton’s advice would have resulted in lower returns than a simple buy-and-hold portfolio, 12% versus 15.5% annualized. However, Stephen J. Brown, William Goetzmann, and Alok Kumar revisited Dow Theory almost a century later and found that over 27 years Hamilton’s calls generated excess risk-adjusted returns, and his strategy would have worked out of sample. Because Hamilton would go into cash and sometimes even short the market, a portfolio tracking his calls would have exhibited only a third of the beta of the market and an alpha of around 4%. There is no good risk-based theory that addresses the prevalence and profitability of momentum. Sure, some efforts have been made, but none are persuasive. For example, in order to demonstrate that momentum is a risky strategy, defenders of the efficient-market hypothesis point to the momentum factor’s sharp losses in 1932 and 2009. The momentum factor is constructed by taking the average returns of large- and small-cap high prior-return portfolios, minus the average returns of large- and small-cap low prior-return portfolios, refreshed monthly.

That looks like a plausible risk story. Momentum profits are like premiums you get for selling catastrophe insurance. You’re rewarded most of the time but pay for it by suffering the rare and vicious loss. However, these extreme returns are artifacts of the way the momentum factor is constructed. During market panics, the betas of the high- and low-momentum portfolios diverge dramatically, with the high-momentum portfolios comprising the safest, most boring stocks, and the low-momentum portfolios comprising the most-volatile, least-liquid, and riskiest stocks. When the market rebounds, the long-short momentum factor is killed, because the high-momentum safe portfolio greatly lags the low-momentum junk portfolio.

In part 2 of this article, we will explore further into momentum, including looking at it from a relative perspective.