The need to link hard sciences with soft sciences became very apparent in the reports that come out on the trading problems of French Société Générale.  The bank had all kind of control systems in place to monitor their risk exposure and one much assume that these were technically sound. One can put a question mark at the fact that Jérôme Kerviel could produce fake emails and that those were not detected. This might be partly weaknesses in the systems, partly in humans supposed to be critical on what they are supposed to do.

 

But the biggest issue on the human side is this: the trading position at the end of the year was (admittedly on paper) about 1.4 bn in profit (half the total profits of the investment bank). Instead of unwinding the deals and declaring the profit, Jérôme Kerviel continued with new deals to hide the success he had. Unwinding would have exposed him and his practices.

 

He apparently felt he could not own up for his success and inform his superiors on what he had achieved. Without further knowledge, a complicated situation to assess, but it only shows that control systems are as good as the human interaction between people that operate them.