Six years ago, Ron den Braber was working at Royal Bank of Scotland in London when he became worried that the bank's models were underestimating the risk of credit products. But when the Dutch statistical expert alerted his bosses to the problem, he faced so much disapproval that he eventually left.

"I started off saying things gently . . . but no one wanted to listen," Mr den Braber recalls. The reason, he believes, lay in "groupthink . . . and pressure to get business done" - as well as a sheer lack of understanding about how the models worked.