Why momentum works

Summary: There’s no free lunch. It is difficult to attract and retain investors to strategies that are uncorrelated with popular benchmarks. This “business risk” helps sustain the momentum premium, which is huge on a risk-adjusted basis. Furthermore, asset class level strategies are simply too large to arbitrage away, as they exploit deep, global-scale human movements.


Static efficiency, dynamic irrationality

Stock market beta, everyone gets: in the (very) long-run you are compensated for the volatility of equity pricing by a respectable 4-5% real return for doing nothing more than buying SPY and sitting on your ass. The ever-expanding Efficient Market Hypothesis, or EMH, says that to get a higher return than average you must accept higher volatility. Value stocks have a higher return, but with somewhat higher risk, since often companies are cheap for a reason. Junk bonds pay more than quality corporates, and Swiss treasuries pay almost nothing.

From another perspective, markets are wildly inefficient, manic-depressive even. Day to day, prices can soar on good vibes and a golden narrative, then stumble when there has been no change in the news. The real value of a stock market grows steadily, but the pricing is all over the place. How can that be efficient?

EMH is best applied to a static snapshot of the market: at a given time, prices will be efficient relative to one another. The exceptions at the margins are rapidly arbitraged away. But the way those prices in aggregate change over time is, if not Brownian, chaotic.

The chaos is organic. Prices trace changes in collective emotions, thoughts and possibilities. The herd is open to certain behaviors at certain times, and others at other times. Cyclical booms create positive feedback loops between stock prices, debt issuance, mergers, real estate, incomes, and taxes. They also bring about their own demise through excess and doomed experiments. Busts bring about other moods and behaviors, which may themselves be overdone and create opportunities for those who are early to perceive positive possibilities.

It is this herding behavior that the momentum factor exploits. Prices that have been going up in recent months are somewhat more likely than not to keep going up, and vice versa. Somewhat more likely than not means a statistical anomaly on the order of 51-60%, depending on the asset and other conditions. These are blackjack odds, which are pretty darned good if you don’t have to do the math in your head or accept certain other risks.

I’ll breeze over demonstrating that this factor is indeed persistent, pervasive, robust, investable and intuitive. The literature is good and plenty. I’ll try to remember to drop some links at the bottom.


If you do it, you’ll win. But try doing it.

Now, EMH will demand an explanation for why a factor or premium like this is sustained. Momentum in particular seems to offend dogmatic sensibilities, since strategies that exploit it not only have a higher return than the market, but far lower volatility. Madoff alert! Big red flag!

Chill. There’s a catch. You want a catch. The lunch is great, but to enjoy it you have to make a bargain that you won’t like if you’re evaluated against the S&P. Yes, you can trounce the index every decade with half the downside, but you’ll be down some months, quarters and years when the S&P is up. If you bosses and clients were rational they wouldn’t care, but they are wildebeests. They are looking to their left, to their right, and at the ass in front of them. As Kahneman says, “what you see is all there is,” and if a herding animal is slipping behind the back his hindbrain starts to panic. He’ll take off, and more often than not he’ll switch into a lane that’s about to slow down, which is why most active investors, institutions included, underperform their passive benchmarks.

If you are running a momentum strategy (or any other alpha strategy) as part of a business, you’ll have to contend with mass stupidity. Being ultimately vindicated isn’t as rewarding as you would think – judging from AUM, it is far better as a business to simply play dumb and make the same mistakes as everyone else. There’s an old saying that nobody ever gets fired for buying IBM.


Thank goodness it’s not easy to be different, because if it were it wouldn’t pay. Client acquisition costs and retention risk are huge, and sustain the momentum premium at a level that must seem obscene to the priests of EMH.


You can’t stop a stampede

There may be even more to it – many of these strategies, especially at the individual stock or contract level, cannot be practically implemented with multi-billion dollar funds. But the premium is largest at the asset class level, meaning that you invest in huge categories like “non-Japanese Asian stocks,” “precious metals,” “REITs,” or “Treasury bonds,” which can absorb many billions without a hiccup. Perhaps it gets down to the shear scale of the herd. We’re talking about the entire human enterprise at this level, and a few fund managers aren’t going to make a dent, especially given that they go along with the herd.

Momentum investors may be contrarian in approach, not caring to examine balance sheets or trade balances, but to a casual observer their actions are not contrarian – they buy when things are going up and sell when they are going down. The difference emerges on only close examination – they buy on a objective signal of a nascent upward trend, and sell on an objective signal of price decline. They hold their positions lightly, and bail out of new and old trades with equanimity. They get whipsawed a lot for small losses, but are on the winning side of an astounding number of major trends.

Prices will never be static for long in a market economy, the only kind of economy there is, and to get somewhere new they have to trend up or down. People will concoct explanations for those movements, make judgments, make connections, borrow, build, hire, spend and be taxed all in the context of market themes: housing!, oil!, emerging markets! or Treasuries!, gold!, the S&P! You can’t arb that away.


Academic references on momentum:


It’s worked for 45 years:


It’s worked for a century:


It’s worked for at least 800 years:



Even professors admit it works:



Play with it yourself:












Leave a Reply

Your email address will not be published. Required fields are marked *