In an earlier note I showed you something odd about a hundred years of market history. The industries everyone had written off tended to come back and beat the darlings everyone adored, by roughly three points a year over the following decade. The pattern is real and it sits right there in the data. But a pattern without a reason behind it is only a coincidence that has not been caught yet. Why on earth should being unloved turn out to be an advantage?

The answer that finally satisfied me came from a corner I did not expect, a piece of behavioral research published back in 1985. One of its authors was Richard Thaler, who learned his craft directly from Daniel Kahneman and Amos Tversky, the two psychologists whose fingerprints are all over the case for rules on this site. Thaler took what those two had discovered about how ordinary people misjudge probability and pointed it at the market itself. If a single person overreacts to dramatic news, he wondered, might a whole market full of people overreact together?

The study that named it

Working with his doctoral student Werner De Bondt, Thaler ran a test that is almost stubborn in its simplicity. They took decades of New York Stock Exchange returns and sorted the stocks by how they had done over the previous few years, the celebrated “winners” in one basket and the beaten-down “losers” in another. Then they simply waited and watched what each basket did next. Over the three years that followed, the discarded losers went on to beat the beloved winners by something close to twenty-five percentage points. The effect was lopsided, with the losers doing most of the work, and a curious amount of it showed up every January. They published the whole thing under a title that doubles as the finding: “Does the Stock Market Overreact?” Their answer was yes.

Why a crowd overpays for a good story

The mechanism they borrowed from Kahneman and Tversky has an ungainly name, the representativeness heuristic, but the idea is plain enough. People treat a vivid recent story as if it were the permanent truth. A company or an industry that has done wonderfully for five years gets quietly reclassified in everyone’s mind as a wonderful business forever, and the price gets bid up to match that certainty. The one that has stumbled gets written off just as completely, and its price gets marked down to match the gloom. In both cases the crowd has already paid, in advance, for a story it assumes will keep going.

Then ordinary life resumes. The glorious winner turns out to be merely good, which is a disappointment when you have paid for glorious, and the written-off loser turns out to be merely ordinary, which is a pleasant surprise when you have paid for doom. Prices drift back toward the real businesses underneath them. That drift is mean reversion, and its engine, Thaler argued, is nothing more exotic than human beings overreacting and then slowly correcting. It is the same shape I found staring back at me in the French sector data. Winners run too far and then cool, and laggards get left for dead and then recover. The century of numbers showed me that it was happening, and Thaler explained why.

The fair objection

Serious people disagree about this. Eugene Fama, who won his own Nobel for arguing that markets are broadly efficient, would read the very same result a different way. In his telling those beaten-down losers earn more not because anyone blundered, but because they are genuinely riskier, and higher returns are simply the pay for bearing that risk. The odd January clustering, and the plain fact that the effect is slippery and hard to pocket after trading costs, are real objections too. So I do not wave this study around as proof of anything. I treat it as a well-documented, plausible reason for a pattern I can measure independently, which is a very different and more modest claim.

What it means for how we invest

In my hands this stays a piece of understanding rather than a trading rule. I would not buy whatever has lagged or dump whatever has led on the strength of a 1985 paper, and I would not want you to either. What Thaler’s finding supports is the quiet discipline you have heard me describe before, equal weighting and the rebalance back to it. Every time we trim the sector that has run hot and top up the one that has lagged, we are making, on a slow and patient schedule, the exact trade De Bondt and Thaler measured. We are leaning very gently against the crowd’s habit of overpaying for the story of the moment. It is worth remembering that Kahneman collected his Nobel in 2002 and Thaler collected his in 2017, so this sits in the mainstream of what we have learned about how markets and the people in them actually behave.

The bottom line

Mean reversion is the slow settling of prices after a crowd has gotten ahead of itself in one direction or the other. It works on its own schedule, and never on one you can plan around. Equal weighting is simply my way of refusing to join the crowd at its most confident, and the rebalance is how that refusal gets acted on, a little at a time, without my having to feel brave about it. Thaler put a name and a number on the human tendency underneath it all. My job is just to build a portfolio that quietly assumes people will keep being people.

John