Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Philip Seager is head of portfolio strategy at Capital Fund Management (CFM).
Everyone thinks they can pick stocks. The fund manager with 30 years of experience, the retail investor glued to Reddit, the uncle at Christmas dinner who “got in early on Nvidia” — all share the quiet conviction that they possess some edge over the market. We decided to test this belief in the most humbling way possible: we replaced them all with monkeys.
The concept of monkeys as random generators has a distinguished pedigree. The French mathematician Émile Borel introduced the idea in his 1913 work on statistical mechanics (Mécanique Statistique et Irréversibilité), imagining monkeys typing randomly on typewriters to illustrate principles of probability.
Inspired by Burton Malkiel’s famous barb that “a blindfolded monkey throwing darts at a newspaper’s financial pages would select a portfolio that would do just as well as one carefully selected by experts”, our version swaps typewriters for dartboards, but the point is the same.
The experiment is deliberately simple: take the S&P 500, randomly select 50 stocks, equal-weight them, rebalance quarterly, and see what happens. Then repeat 300 times, each simulation running from 2000 through 2025 to capture the dotcom collapse, the Global Financial Crisis, a pandemic, and the largest AI-driven rally in market history.
The results, at first glance, look surprisingly encouraging for the dart-throwers versus a benchmark index earning the risk-free rate.
The grey fan of simulated paths is wide, and many of the monkeys have done rather well. All of them beat the risk-free rate, some spectacularly so. A casual observer might look at this chart and conclude that stockpicking — even random stockpicking — is a viable strategy. They would be wrong.
What the fan chart is really showing isn’t stock selection skill. It’s showing beta. The returns of any individual stock split into a systematic market component (the tide that lifts or sinks all boats) and an idiosyncratic residual (the part that is genuinely unique). With 50 names in a portfolio, the systematic component dominates overwhelmingly.
Apparent variation across our random portfolios is driven almost entirely by small differences in average beta, compounded over 25 years. Instead of beating the market our monkeys have become the market, with minor random tilts.
Let’s strip out the market component entirely and look only at the residual. This should surface genuine stockpicking ability.
Once returns from the equally weighted S&P 500 index are subtracted, everything now centres on zero, with the median random portfolio ending almost exactly where it started. Some monkeys got lucky; others didn’t. But the mean outcome is, to a close approximation, zilch.
And this is a mathematical inevitability. In a universe of correlated instruments driven by a common factor, randomly selecting a subset and hoping it will systematically differ from the whole is a game played against arithmetic.
Moreover, the room for differentiation shrinks as portfolios grow larger. We repeated the experiment with bundles of portfolios holding just five stocks, then ten, then twenty — all the way through to the full 500 — and measured the annualised tracking error of each random portfolio against an equal-weighted benchmark.
Some content could not load. Check your internet connection or browser settings.
With five stocks, the annualised tracking error exceeds 15 per cent — plenty of room to look either brilliant or disastrous. By 50 stocks, it falls to roughly five per cent.
At 200 stocks, tracking error is below three per cent. And at 500, the random portfolio is indistinguishable from the benchmark itself.
The diversification that investors and regulators rightly demand is the very thing that makes outperformance mechanically difficult. The more stocks you hold, the more you converge towards the market, and the less room there is for idiosyncratic returns, good or bad, to compound into anything meaningful.
This is the fundamental tension at the heart of active equity management. A concentrated portfolio has the capacity to deviate from the benchmark, but that deviation is symmetric: it is just as likely to destroy value as to create it. A diversified portfolio has the stability investors want, but at the cost of becoming, for all practical purposes, a very expensive index fund. The narrow space in which a manager holds enough stocks to be prudently diversified, yet few enough to express genuine conviction, is far smaller than the industry’s marketing materials would suggest.
None of this means that skilled active management does not exist. It does. But the arithmetic of correlated returns guarantees that the average active manager, before fees, will match the market, and after fees, will trail it. The exceptional managers who do generate persistent alpha are, by definition, exceptions. For the rest, the dartboard is not a bad approximation. And the dartboard, as our 300 monkeys can attest, delivers precisely zero.
