The Feud That Wasn't — Simons and the EMH Guys

The Efficient Market Hypothesis says you cannot beat the market. Not that it is difficult. Not that it is unlikely. That it is, in a meaningful sense, impossible. Prices reflect all available information. Any apparent edge is either luck, survivorship bias, or compensation for risk you have not properly accounted for. The market is efficient. You are not smarter than the market. Go buy an index fund.

This has been the dominant framework in academic finance for sixty years. It earned Eugene Fama a Nobel Prize. It underpins the entire passive investing industry. It is taught in every MBA program on the planet. And Jim Simons spent thirty-five years proving it wrong with compound annual returns of approximately 66% before fees, in a fund that operated continuously from 1988 until his death, with a consistency that makes the word “luck” sound like a joke told by someone who does not understand probability.

The tension between these two facts; the theory and the track record; is one of the most interesting unresolved problems in finance. Not because the answer is complicated. Because the answer is uncomfortable.

What the Hypothesis Actually Claims

The EMH comes in three flavors, and most people who invoke it are not precise about which one they mean.

The weak form says that past price data cannot predict future prices. Technical analysis is useless. Charts are tea leaves. You cannot look at a stock’s price history and extract a trading edge from it. The semi-strong form says that all publicly available information is already reflected in prices. Fundamental analysis; reading balance sheets, studying earnings reports, evaluating management; cannot produce consistent excess returns because the market has already incorporated that information. The strong form says that even private information is reflected in prices. Insider knowledge does not help because it is already priced in through the aggregate behavior of all market participants.

Almost nobody defends the strong form. Insider trading obviously works, which is why it is illegal. The semi-strong form is where most of the debate lives. And the weak form is the one Simons demolished most directly, because Renaissance Technologies built its fortune largely on extracting signals from price data; exactly the kind of data the weak-form EMH says contains no exploitable information.

The critical thing to understand about the EMH is that it is not really a statement about markets. It is a statement about you. It says that whatever information you think you have, other people have it too, and they have already acted on it, and the price has already moved. Your insight is not an insight. Your edge is not an edge. You are, at best, a tourist in a game played by professionals who have already arbitraged away whatever you think you see.

This is a powerful argument. It is also wrong in a specific and important way that the Medallion Fund makes visible.

The History of an Idea

The EMH did not emerge from nowhere. It grew out of a specific intellectual tradition at the University of Chicago in the 1960s, where a generation of economists was trying to turn finance from a trade into a science. The ambition was noble. Finance had been dominated by folklore and gut instinct for centuries. The Chicago school wanted rigor. They wanted testable hypotheses. They wanted mathematics. And the most powerful mathematical result they found was that, in a competitive market with rational participants, prices should already reflect available information. The logic was clean; if a piece of information suggested a stock was undervalued, someone would buy it, and the buying would push the price up until the information was fully reflected. The adjustment should be nearly instantaneous, because money is a powerful motivator and smart people move fast.

Fama formalized this in 1970, and the theory took on the character of revealed truth within academic finance. It was elegant. It was defensible. And it had an enormous practical implication: if markets are efficient, then trying to beat them is a fool’s errand. The entire active management industry; every stock picker, every technical analyst, every macro strategist with a newsletter; was running a scam, or at least a delusion. This was a radical claim, and it made the people who advanced it enormously influential.

The problem is that the EMH’s elegance became a kind of prison. The theory was so clean, so logically satisfying, that it became resistant to empirical challenge. Anomalies were discovered; the value premium, the momentum effect, the small-cap premium; and each one was either explained away as compensation for risk or dismissed as a data-mining artifact. The theory could absorb almost any challenge without breaking, which is either a sign of a robust theory or a sign that it had become unfalsifiable.

And then Simons showed up with three decades of returns that the theory could not absorb at all.

The Track Record That Should Not Exist

Between 1988 and 2018, the Medallion Fund generated average annual returns of roughly 66% before fees and 39% after fees. In the years for which public data exists, the fund had only one losing year, and that loss was small. During the 2008 financial crisis, when the S&P 500 dropped 37%, Medallion returned 82%. Not negative 82. Positive 82.

Put differently: a dollar invested in Medallion in 1988 would have been worth approximately $42,000 by 2018, after the fund took its substantial cut. The S&P 500, over the same period, would have turned that dollar into roughly $20. The gap is not a rounding error. It is several orders of magnitude.

The sample size is not small. Thirty years is not a coin flip. Three hundred and sixty months of trading data is not a lucky streak. The probability of generating Medallion’s track record through chance, given any reasonable model of market returns, is so low that it requires scientific notation to express. The fund did not get lucky. The fund had an edge. The EMH says the edge should not exist.

How the Academics Respond

The academic response to Simons falls into a few categories, none of which are satisfying.

The first is the survivorship bias argument. Thousands of funds try to beat the market. A few will succeed by chance. We hear about the winners and not the losers, so the winners look more impressive than they are. This argument is valid in general but useless against Medallion specifically, because the consistency and magnitude of the returns place it so far outside the distribution of chance outcomes that survivorship bias cannot explain it. You can invoke survivorship bias for a fund that beat the market by two percentage points for ten years. You cannot invoke it for a fund that beat the market by forty percentage points for thirty years. The math does not work.

The second is the risk argument. The fund is not generating alpha; it is generating returns that compensate for risks not captured by standard models. Maybe Medallion is taking on tail risk, or liquidity risk, or some exotic form of exposure that makes the returns fair compensation for danger rather than evidence of skill. The problem with this argument is that Medallion’s worst year was essentially flat. A fund that is being compensated for risk should, by definition, occasionally realize that risk in the form of catastrophic losses. Medallion did not. Its Sharpe ratio; the measure of return per unit of risk; was the highest ever recorded for a fund of its size and duration. The returns were not compensation for hidden danger. They were returns on top of almost no danger at all.

The third response is the most honest. It is silence. A meaningful number of academic finance researchers simply do not discuss Renaissance Technologies. They build models of market efficiency and present evidence for those models and simply do not address the largest counterexample in the history of the field. This is not conspiratorial. It is human. When your life’s work rests on a theoretical foundation, and the most famous data point in your field contradicts that foundation, the temptation to look elsewhere is powerful.

What Simons Actually Proved About Markets

The interesting question is not whether the EMH is wrong. It is wrong in the ways that matter, and Simons proved it. The interesting question is how it is wrong, and what that reveals about the nature of markets.

Markets are not efficient in the way a thermostat is efficient, where deviations from the target are corrected instantly and perfectly. Markets are efficient in the way that an ecosystem is efficient; deviations are corrected, but the correction takes time, and during that time the deviation is exploitable by anyone fast enough and precise enough to capture it.

The EMH assumes that information is incorporated into prices instantaneously. It is not. Information arrives unevenly. Different participants process it at different speeds. Some participants are constrained by institutional rules, or emotional biases, or the simple fact that they are human beings who need to sleep and eat and think. These processing delays create temporary mispricings. The mispricings are small. They are short-lived. They are invisible to anyone trading on fundamental analysis or reading quarterly reports. But they are real, they are systematic, and they can be captured by a sufficiently fast and precise system.

This is what Simons built. Not a system that predicts the future. A system that detects the present fractionally faster than everyone else. The signals Renaissance exploits are not secrets. They are statistical patterns in publicly available data that persist because the speed required to exploit them is beyond the reach of ordinary market participants. The edge is not information. The edge is processing speed and mathematical precision applied to information that everyone technically has access to.

The EMH is correct that you, personally, cannot beat the market by reading the Wall Street Journal and picking stocks. It is correct that the vast majority of active fund managers underperform index funds over long time horizons. It is correct that for most people, in most circumstances, acting as though the market is efficient is the right strategy. What it gets wrong is the universal claim. The claim that no one can beat the market. The claim that excess returns are impossible rather than merely very difficult. Simons did not disprove efficiency. He disproved the universality of efficiency. He proved that the market is efficient enough to defeat almost everyone, but not quite efficient enough to defeat a team of the world’s best mathematicians armed with billions of dollars of computing infrastructure.

The Uncomfortable Implication

The reason the EMH debate matters is not because it changes what ordinary investors should do. It does not. Index funds remain the correct choice for anyone who does not have access to a team of PhDs and a server farm. The reason it matters is what it reveals about the relationship between academic theory and empirical reality.

The EMH is a beautiful theory. It is parsimonious. It is testable. It generates clear predictions. It has enormous explanatory power for a wide range of market phenomena. And it is contradicted by the single most important data point in its domain. The response of the academic establishment has not been to revise the theory. It has been to ignore the data point, or to explain it away with arguments that do not survive scrutiny, or to acknowledge it in footnotes while continuing to teach the theory as though it were unqualified truth.

This is not unique to finance. It is how academic disciplines work. Theories acquire institutional momentum. Careers are built on them. Departments are organized around them. Nobel Prizes are awarded for them. When a counterexample appears, the first instinct is not revision but defense. And when the counterexample is a private fund that publishes no papers and grants no interviews, the defense is easy; you simply do not engage with it.

Simons did not care about winning the academic argument. He was not interested in publishing papers that refuted the EMH. He was interested in making money, and the money was its own argument, and the argument was settled in favor of anyone willing to look at the numbers.

The market is not perfectly efficient. It is approximately efficient, which is a very different thing. The gap between “approximately” and “perfectly” is where Medallion lived for thirty-five years. The gap is real. It is exploitable. And the fact that almost nobody can exploit it does not mean it does not exist.

It means the price of entry is higher than almost anyone can pay.