You would assume that in an educational system in which exam results are presented as crucial by schools and parents, the highest achievers would be happiest and have the best self-esteem. This is far from true.

One study looked at the relationship between the average ability-level, across several differing schools, and self-esteem. Esteem was found to be lower in schools with higher-ability pupils: if you are surrounded by very able people, it tends to lower your self-valuation. Another study found that achievers in top American universities have lower career aspirations and self-regard than achievers from less exalted universities.

Speculating on the evolutionary purpose of depression may help to explain this. Originally, according to British psychologist Paul Gilbert, it was a way for subordinate members in a hierarchy to indicate to superiors that they posed no threat. By placing a low value on their personal worth, they could signal to more powerful members that they accepted the status quo. The alternative was violent death.

The theory predicts that on the whole, subordinated members of society, such as women and the poor, would suffer more depression than superior ones - as they do. But under some conditions, even the most successful members of society can also become depressed.

John Price, a key theorist in the field of evolutionary psychology, wrote: 'Not all people who are in subordinate positions ... suffer from depression, while people in a high social position (such as a head of state) may do so. What is critical is what people perceive their social status and power to be and what they believe to be the critical level below which it must not be allowed to fall.'

Academic high-flyers perceive themselves so negatively because they compare with others of their standard, losing sight of how well they are doing compared to the vast majority. Their success leads to further subordination, because having become one of the biggest fish in a pond, they are moved to a bigger one where they are tiddlers in a shoal of equally high achievers.

That is the likely explanation for the finding that, in a definitive study of two large samples of British 15-year-olds, middle-class, high-achieving girls were much more likely to suffer depression or anxiety than working-class ones (38 per cent versus 24 per cent in 1996). A more in-depth study of two small samples of middle- and working-class girls from ages four to 21 powerfully bore this out. Valerie Walkerdine, the researcher who did the study, told me: 'A working-class girl who did well would be held up as a good example by friends and family, whereas the talents of middle-class achievers were largely unsung.'

This is the fate of so many high achievers: death by social comparison. If our self-esteem is contingent on external standards we run a huge risk of feeling like failures, because, even if we succeed in these terms, there will always be someone better than us.