Medallion Fund: The ultimate counterexample? The Journal of Portfolio Management. The paper does more than marvel at extraordinary numbers — it systematically dismantles every rational explanation the finance literature offers for exceptional returns, and arrives at a conclusion that sits uncomfortably with decades of academic consensus.
Cornell begins with the performance record itself. From 1988 to 2018, Renaissance Technologies’ Medallion Fund posted a compound gross return of 63.3% annually. USD 100 invested at the start would have grown to USD 398.7 million by the end — a figure that dwarfs even a hypothetical investor with perfect monthly foresight over the same period, who would have ended with only USD 331,288. The fund never had a negative annual return across 31 years, including during the dot-com crash (56.6% return) and the 2008 financial crisis (74.6% return).
The paper then turns to the more serious question: can any standard financial framework explain this?
The risk premium argument fails. Conventional asset pricing theory holds that higher returns compensate investors for bearing greater risk. Cornell tests this directly. Medallion’s market beta was approximately −1.0 — the fund moved inversely to the market, meaning it provided a hedge rather than added exposure. Three-factor regressions incorporating the Fama-French SMB and HML variables yielded negative loadings on both factors, though neither reached statistical significance. The fund had a high standard deviation of returns at 31.7%, but this was offset by a mean of 66.1%, yielding a Sharpe ratio above 2.0. Volatility here reflected the spread of exceptional gains, not the signature of risk-bearing.
The scale argument is equally telling. One might expect that as assets under management grew, returns would compress — more capital chasing the same opportunities. Yet as Medallion scaled from USD 20 million to USD 10 billion, returns showed no meaningful deterioration. This points to strategies with genuine informational or execution advantages, not simply capacity-constrained arbitrage.
The trading edge is deceptively modest. Robert Mercer, one of Medallion’s key investment managers, stated the fund was correct on approximately 50.75% of its trades — barely above chance. Cornell uses this to make a precise point: the source of returns was not a large per-trade advantage but the discipline to execute millions of trades with a consistent, marginal edge while keeping transaction costs — which would be substantial at that volume — sufficiently low. The reported gross returns are calculated after trading costs, which makes the numbers more remarkable, not less.
The public funds comparison closes the case. Renaissance did launch funds open to outside investors — the Renaissance Institutional Equities Fund and Renaissance Institutional Diversified Alpha. Neither follows Medallion’s strategy, and neither comes close to Medallion’s returns. Cornell reads this as evidence that the strategy generating Medallion’s results has a scale ceiling, and that whatever proprietary advantage Renaissance possessed could not simply be transferred to a larger, more open structure.
Cornell stops short of claiming to have solved the puzzle. He raises the possibility that Medallion operated as an exceptionally efficient market maker across millions of short-term positions — but acknowledges the returns are large enough to strain even that explanation. The paper ends by framing Medallion as a Michelson-Morley moment for finance: a result so inconsistent with the prevailing theory that the theory must be questioned, even if no replacement is yet ready.
Reference
Cornell, B. (2020). Medallion Fund: The ultimate counterexample? The Journal of Portfolio Management, 46(5), 1–4. https://doi.org/10.3905/jpm.2020.1.110