I used to know an investment banker who went around saying he was writing a novel. Obviously, this was self-marketing, but he also genuinely believed he could write a better novel than actual novelists. After all, he reasoned, he must be smarter than they were, because he earned more money.

This man was suffering from the Davos Fallacy: the notion that very high earners are also very clever. But at the top end of incomes this isn’t true, according to a new study of 59,000 Swedish men. The highest-earning 1 per cent and the brainiest 1 per cent seem to be two largely separate groups, with little overlap. If that’s so, how should we treat each elite?

The study, “The plateauing of cognitive ability among top earners”, by sociologists Marc Keuschnigg, Arnout van de Rijt and Thijs Bol, uses an unusually rich dataset. When military service in Sweden was compulsory, and almost all native-born men enlisted aged 18 or 19, they were tested on cognitive ability. There were “separate paper and pencil tests for verbal understanding, technical comprehension, spatial ability, and logic”. The paper analyses the future earnings of men tested from 1971 through 1977, and from 1980 until 1999.