In part 1 of this article, we looked at momentum from an absolute perspective, let’s explore momentum further, and from a relative perspective.
I constructed a long-short momentum factor that goes long high-momentum stocks but shorts the market. This version of the factor looks steady, because the betas of the long and short portfolios stay more evenly matched through both normal and panic markets.
The risk-based story looks even less defensible when you look at the magnitude and timing of trend-following’s payoffs. Trend-following generates exceptional risk-adjusted returns. Even worse, its best returns come during extreme markets. In stable markets, trend-following performs reasonably well but won’t shoot the lights out, earning about the risk-free rate. Trend-following is like an insurance policy that pays you to own it, an impossible creature.
.png)
Academics seeking to preserve risk-based theories have bent over backward to explain the trendiness of markets, and many of their explanations have an ad hoc feel to them. For equities, it’s time-varying, serially correlated risk premiums or mysterious risk factors. For commodities, it’s the theory of storage, where price momentum is a proxy for tight inventories, which in turn should be compensated for with a “convenience yield.” For currencies, it’s central banks that try to manage exchange rates.
While some of these explanations are undoubtedly true and useful, Hamilton, the much-maligned father of Dow Theory, offers a prescient explanation for trend-following’s success: “Prosperity will drive men to excess, and repentance for the consequences of those excesses will produce a corresponding depression.” The modern interpretation of the behavioral story goes like this: In light of surprising or extreme news, investors anchor new price estimates to old prices and do not fully adjust for its impact. Investors are also loath to realize losses, preferring to keep dogs until they break even, and are too quick to sell winners. Both biases prevent prices from instantly reflecting new information; instead, prices slowly adjust to fair value, creating sustained price movements. Performance-chasers hop on the trend, overextending it, creating a self-fulfilling cycle that draws in more investors. Eventually, Herbert Stein’s law kicks in: “If something cannot go on forever, it will stop.” The trend collapses, leaving the investors who got in at the top holding the bag.
The literature behind this story is impressive and not in much dispute, so I won’t recapitulate it. What are less appreciated are the limits to arbitrage, the sand in the market mechanism that keeps smart money from bringing prices back in line with their fundamental values. Momentum is unusual in that it is perfectly rational for an investor who knows an asset is overpriced to contribute to the bubble or at least not try to fight against it too hard. George Soros, one of the most successful investors of all time, even said, “When I see a bubble forming I rush in to buy, adding fuel to the fire. That is not irrational.” A would-be arbitrager who tries to do the opposite, short-selling a bubble, runs the real danger of going broke, because shorting requires you to get both the path and terminal prices of the asset right. An extreme example is the dot-com bubble, which required almost exact timing to short successfully. A short-seller who was even a few months too early would have been hurt badly; a short-seller who was even earlier would have been driven out of business.
The behavioral story implies that the growing capital dedicated to managed futures will reduce its profitability, possibly to the point where it offers little reward. A “fair” reward for a strategy that does well in bad times is actually a low or even negative expected return. So far, there has been little evidence of this, though statistical tests are usually not powerful enough to detect all but the most dramatic changes. In one of his rare comments on investing strategy, Jim Simons, founder of Renaissance Technologies and dean of quantitative investors, said trend-following had “lost its zip” in recent years. He also noted that “almost any good viable predictive signal will almost certainly erode over five years. You have to keep coming up with new things. The market is working against you.” This was in 2007. Trend-following seems to have defied Simons’ predictions, although he might have been talking about fast trend-following signals, which do seem to have lost their power.
Trend-following looks like the Holy Grail of alternative investments. It’s uncorrelated to stocks and bonds, has an attractive Sharpe ratio, is liquid, and is one of the few strategies that does well in bad times. If you take the historical results of trend-following at face value, there is every reason to dedicate a big chunk of your portfolio to the strategy. However, doing so also requires you to believe that the markets are almost comically inefficient, at least at the macro level, and that other would-be alpha prospectors haven’t tapped the well dry.
