There's a persistent opinion knocking around that blames the financial crisis on one small molecule: testosterone. It is testosterone, apparently, that makes traders go feral as stocks rise, taking bigger and bigger risks until Lehman Brothers happens, or someone falls out of a window.

Women, then, goes the argument, are safer traders -so let's fill trading floors with the ladies. Naturally cautious, milder creatures, they make better investors. If women ruled the world, continues the logic, getting jovial now, lighting its pipe - there would be no stock-market crashes, and probably no, like, wars, either.

The argument is, on the face of it, pro-women. After all, it's saying women are good at something; they're good at not being overconfident. But is it feminist? Not really. Broad behavioural generalisations based in biology rarely do women much good. Along with "risk aversion" goes "less competitive" and "less confident". Banking may need these traits but they aren't attractive to employers, and it's damaging to saddle an entire gender with them.

So the argument isn't particularly feminist. But perhaps more interestingly, neuroscientists are starting to find that it isn't particularly true, either. It seems risky decision-making has more to do with confidence than gender. Being a woman, evidence suggests, only affects trading behaviour
as far as it's used to deplete your confidence. If you're told your gender doesn't like risk, it's surprisingly hard to go away and bet a million on the stock exchange with the right sort of swagger.

A recent study at Stanford University took 53 men and women - some bankers, some undergraduates - and gave them 14 risky choices. In each choice, a participant could either make a punt for a high sum, with a greater risk of failing, or go for a low-risk, low-pay option. The more frequently someone chose the safe option, the more risk-averse they were judged to be.

When the experimenters brought up gender stereotypes before the trial, women became overwhelmingly more cautious, whereas men took more risks. When such stereotypes were not used, men and women performed almost equally.

For the men, just knowing that women were negatively stereotyped made them confident enough to take riskier gambits. Expose a woman to typecasting and you knock her self-belief enough to produce the opposite effect.

Swat team

How can women avoid the problem? It's quite hard. The experimenters think even the process of shrugging off stereotyping - just thinking, "Nonsense, you total knob" - robs you of just enough energy to affect your decisions. They call it "ego depletion". Swat away a couple of negative comments and your next decision is more likely to be a safe one.

The study is provocative, not least because it invites extrapolation. After all, it's a bit of a coincidence that qualities "natural" to women - trying to please people, avoiding conflict, being self-deprecating - would be equally "natural" to any dominated and therefore less confident group.

I'm tempted to make the link.