Behavioural finance experts and Nobel laureates Daniel Kahneman and Richard Thaler have long argued that all kinds of cognitive biases bedevil human judgment.

Sceptics disagree, saying laboratory experiments are an unreliable guide to real-life behaviour. In real life, they say, people will get it right if the stakes are high enough.

Not so, according to a new experimental study involving 1,236 “analytically capable” economics students in Nairobi. Harvard researchers devised various tests to see if large financial incentives – equivalent to more than a student’s full monthly income – would eliminate cognitive biases. The money caused them to try harder; response times to questions increased 40 per cent. However, performance “improves either very modestly, or not at all”.

The results are hardly surprising; just think of human decision-making during housing bubbles. As Richard Thaler notes, the idea that people get smarter if you raise the stakes is “one of the most annoying comments in the economics literature”.