Why stockpickers keep losing

Every year S&P Dow Jones Indices, the company behind the S&P 500, publishes a scorecard comparing actively managed funds with the indices they try to beat. Every year the results are grim for the managers. In 2025 they were grimmer than usual. Some 79% of active funds investing in big American companies did worse than the S&P 500, the fourth-worst showing in the scorecard’s 25-year history. In 2024 the figure was 65%. Halfway through 2025 it had been a more respectable 54%; a strong second half for the largest companies did the damage.

First half of 202554

202465

202579

050100

Source: S&P Dow Jones Indices, SPIVA scorecards

Why do professionals, with their analysts, models and Bloomberg terminals, struggle to beat a list of companies chosen by a committee? The answer has less to do with skill than most people think.

The arithmetic

In 1991 William Sharpe, a Nobel-prizewinning economist, published a two-page article that remains the most important thing ever written on the subject. Divide all investors into two groups, he said: those who hold the market index and those who do not. Index investors, by definition, earn the market’s return before costs. Since the two groups together own the whole market, the active investors must also earn the market’s return before costs, on average. After costs, which are higher for active managers, the average active dollar must therefore do worse than the average passive dollar. This is not an empirical finding. It is a matter of addition and subtraction.

So some underperformance is guaranteed. The interesting question is whether some managers are good enough to win consistently. Eugene Fama and Kenneth French, two economists, studied the returns of American mutual funds and concluded that, after fees, the spread of results looks roughly like what luck alone would produce. Jonathan Berk and Richard Green offered a kinder explanation. Skilled managers do exist, but investors pile money into them after a good run, and a bigger fund finds it harder to beat the market. In equilibrium, skill shows up in the size of the fees managers earn rather than in the returns investors receive.

Why 2025 was so bad

Last year added a specific handicap: concentration. Active managers tend to own more mid-sized and smaller firms than the index does and to hold less of the giants, partly because rules on diversification limit how much any one fund can put into a single stock. In 2025 the S&P 500 beat an index of mid-sized companies by about ten percentage points and one of small companies by about 12. The ten biggest companies in the S&P 500 now account for about 39% of its value. A fund that did not own enough of them almost certainly lagged.

There is a deeper reason that stockpicking is hard. Hendrik Bessembinder, a finance professor, studied every American listed share from 1926 to 2016 and found that about four in seven did worse over their lifetimes than one-month Treasury bills. The best 4% of companies accounted for all the stockmarket’s net wealth creation. Picking stocks means trying to find those few winners in advance. Owning the index guarantees that you hold them.

Where the money has gone

Investors have drawn the obvious conclusion. Passive funds overtook active ones in American equity funds in 2019, and across all American funds at the end of 2023, when index funds held about $13.3trn against $13.2trn in active funds. Active American equity funds suffered net withdrawals of about $32bn in 2025.

That does not mean active management is pointless. Markets need investors who research companies and trade on the results; if everyone indexed, prices would stop reflecting information, and the rewards for doing the research would rise until some investors returned to it. Active managers also do better in some years and some corners of the market than in others. But for most savers, the lesson of the scorecard has hardly changed in a quarter of a century. Choosing a manager who beats the market is far harder than it looks, and the fees are certain while the outperformance is not.