Momentum and value investing are two classes of strategies that, historically, alternate ascendancy in terms of which strategy is dominating the other. They are largely opposite strategies: a momentum investor buys a security because it has gone higher (because prices aren’t really a random walk, something which has gone up in price is more likely to continue to go up in price) while a value investor buys a security because it has gone lower (since the lower the buying price, the better the return on a security).

You can imagine the two strategies in the context of horse racing. The “momentum” strategy would be represented by betting on the “favorite,” the horse with the best odds to win as determined by the prior betting. (Some people think the track sets the odds on the horses, but that’s not the case. The payouts are based on the proportion of the entire betting pool allocated to bets on a particular horse, less the track’s vig. So, a horse with “good odds of winning” is simply the horse that has the most money bet on it to win.) That’s pretty close to exactly what a momentum investor in stocks is doing, right? A “value” investor, in the context of horse racing, is the person who bets on the long shots because they have big payoffs when they hit (and the bettor believes, obviously, that these unloved horses are irrationally disliked because most people like betting on favorites and winning frequent, small amounts instead of winning infrequent, large amounts.)

So at the track, sometimes the favorites win and sometimes the long shots win, and there are people in each camp that will tell you their strategy is the better one in the long run. I don’t know that there have been many studies of whether “value” or “momentum” investing in horse racing is the better strategy, but there have been numerous such studies in finance. Both value and momentum have been shown to improve investing strategies, with better risk-adjusted returns than simply buying and holding a capitalization-weighted basket of securities. They tend to have “seasons,” by which I mean long periods when one or the other of these strategies tends to be dominant. But it is very unlikely that either of these strategies could ever be the winner over the long run.

To see why, think of the horse track. Suppose everyone noticed that the favorites were winning, and so more and more money came in on the favorites. What would happen then is that the payoffs on the favorite would get worse and worse, and the payoff on the long shots would get better and better. Eventually, it would be very hard to make money betting the favorites unless they always won. On the other hand, if lots of money were to come in on the long shots, they wouldn’t be long shots for long. So neither strategy can dominate forever.

The same is true in finance. If everyone is betting on the previous winners, then eventually the “losers” become easy money, and vice-versa. The chart below (which is imperfect for a reason I’ll mention in a moment) illustrates the give-and-take. It shows the Russell 1000 “growth” index (RLG, in white) and the Russell 1000 “value” index (RLV, in orange). The source of the chart is Bloomberg.

You can see clearly how “growth” (which has similarities to momentum) outperformed in the Y2k bubble, depressing the heck out of value investors. But then value beat growth for a while, until the next bubble in 2007. The ensuing bear market crushed both strategies.

One caveat here is that the composition of the “growth” and “value” indices doesn’t change every day, and isn’t based on momentum, so that at the peak in 2007 a lot of stocks in the “value” index were not truly value stocks. But you get the general point.

The second, and more important caveat, applies to the years since 2009. This chart would lead one to believe that both value and growth stocks are doing equally well. And they are, given this definition of growth and value. But what this chart really means is that the distinction of “growth” and “value” are now less important than the single factor “momentum.” Whether you have a growth stock with momentum, or a value stock with momentum, is less important than if you compare performance to something else that does not have momentum.

We can illustrate this concept by calculating portfolios that are built to maximize momentum or value for a given risk constraint, and comparing the performance of the portfolios. I’ve done this for a bunch of different types of portfolios (different commodities, equities only, broad investor stock/bond/cash/commodity portfolios, etc) and they all look something like this chart, which shows the total returns of these two competing portfolios:

What I’m doing here is for the security universe in question, I’m calculating for each security a “momentum” score that is simply the year-on-year percentage change in price, and a simple “value” score that is the inverse of the four-year price change.[1] Then I optimize two portfolios, one which maximizes the value score and one which maximizes the momentum score, and then track that portfolio’s performance for the following month (whereupon the portfolios are reconstructed). If there was no memory to the momentum or value processes, these lines would wander around 100…a high momentum score would not increase next month’s performance, e.g.. But, evidently, it does and it has. Over the last three years, for this security universe, the “momentum” portfolio outperformed the “value” portfolios 78% of the time by a cumulative 50%. And this happens for every universe of securities I test. Even within commodities, which are universally hated, the high-momentum commodities are hated less.

Note that this is at the same time that in the first chart above the “growth” and “value” stocks have been performing about the same. This just means that the dispersion between growth and value has been narrow, which is another way that volatility is low.

As a value investor, this situation has been tortuous, and has led me to change the way we do certain things to keep from being purely value all the time. But as I said before, the situation cannot remain this way forever. Every computer is chasing every other computer, for now. But at some point, one of the computers will decide it’s time to lean the other way, and the first ones that do so will be the winners while the other computers start to chase momentum lower.

That might not be as fun for investors as the recent period has been, unless you’re the one who was getting paid on the nag at 200-1 odds.

[1] In this I am taking a cue from Asness, Moskowitz, and Pedersen, “Value and Momentum Everywhere,” Journal of Finance Vol LXVIII, #3, June 2013.