The Frontal Cortex

Excessive Risk-Taking

I’m morbidly fascinated by the massive losses recently incurred by the French Bank Societe Generale. My fascination is partly rooted in the sheer scale of the disaster, a scale that’s essentially incomprehensible. (I have no idea what a $7,000,000,000 loss really means.)

But I’m also interested in how, exactly, a trader could lose so much money and not get noticed. It now appears that the risk-taking culture of Societe Generale is partly to blame:

The 144-year-old bank allowed a culture of risk to flourish, creating major flaws in its operations that enabled the rogue trader’s activities to go undetected, according to bank officials, investigators and traders who worked with Mr. Kerviel.

Far from being discouraged from placing big bets, Société Générale traders were rewarded for making risky investments with the bank’s money. It was not uncommon for traders to briefly exceed limits imposed on their trading before pulling back, despite controls meant to prohibit this.

This description neatly parallels some recent findings in neuroeconomics. In a 2005 paper published in Neuron, neuroscientists at Stanford University illuminated the delicate equilibrium between our lust for gains and our aversion to risks. When the system gets out of whack, we start making bad decisions. The experiment was straightforward: a subject was given a small amount of money to invest and three different investment options. The first two options were stocks. One of these stocks was randomly designated a “bad” stock by a computer (it was more likely to lose money), while the other was deemed a “good” stock, and was more likely to make money. The third option was a safe, low-yield bond. The game was played for ten rounds, and subjects got to keep their winnings.

At first, the subjects had no idea which investments they should choose. Both of the risky stocks seemed equivalent, and it was impossible to tell if the low-yield bond was actually a prudent alternative. (This phase of the experiment isn’t too different from the state of a poker game before the community cards are dealt. In other words, nobody knows very much.) However, after just a few rounds, most “investors” had deciphered the situation and correctly identified the most lucrative option. Whenever these people saw the “good” stock, a part of the brain that processes rewards – the nucleus accumbens (NAcc) – was activated, sending a signal that this was the best investment. The “bad” stock, in contrast, triggered the insula, which is responsible for generating negative emotions. As a result, these people made the “rational” investment more than 75 percent of their time.

The scientists, however, were more interested in what happened when people chose the wrong financial options. Why did some investors consistently choose the low-yield bond and the dangerous stock? Which brain areas were responsible for such decision-making mistakes?

At this point, the experiment took a slightly unexpected turn: the same network of brain areas was activated regardless of whether subjects chose the “good” stock or the “bad” stock. Wise decisions and foolish decisions seemed to come from an identical place. However, when the scientists looked closer at the brain scans they discovered a telling difference. The reason some people made financial mistakes was that they seemed to be feeling too much, as their emotional areas went into overdrive. For example, if the NAcc became too active, then the subject almost always made a “risk-seeking” decision. This involved picking the stock with a checkered history, on the slim chance that the gamble might pay off. On the other hand, if the insula was disproportionately excited, then the subject was most likely to invest in the low-yield bonds. Their excessive worry made them excessively risk-averse. As the scientists note, “Financial decision making may require a delicate balance–recruitment of distinct circuits may be necessary for taking or avoiding risks, but excessive activation of one mechanism or the other may lead to mistakes.” The people who made the best investment decisions listened to their “emotional brain” but never let any single feeling spiral out of control.

What does this have to do with Societe Generale? The job of a bank or hedge fund or investment firm is to help their employees maintain their mental balance. In other words, don’t create an atmosphere where people chase after their NAcc. As Camelia Kuhnen, one of the lead authors remarked in 2005:

Institutions encouraging either “risk-seeking” behavior–such as casinos–or “risk-averse” behavior–such as insurance companies–understand how excitement and anxiety influence people’s decisions. “You go to Vegas, and you are surrounded by all these rewards–free food, potential prizes,” Kuhnen said. “We know from past studies that sights of potential rewards can activate your NAcc. What this study shows is that when the NAcc is activated, you tend to be more risk seeking. You tend to choose the stock more often and you tend to choose it when you shouldn’t.” An insurance company appeals to opposite emotions, Kuhnen added. “If you go to an insurance office, people will show you pictures of crashes or tell you how you could be hurt. We know that such a stimulus might activate your anterior insula. Our study shows that high insula activation is a predictor of you making a risk-averse choice.”


  1. #1 Dunc
    February 5, 2008

    But I’m also interested in how, exactly, a trader could lose so much money and not get noticed.

    You might be interested in this article: How to lose $7.2bn with just a few Basic skills. Executive summary: all these big finance institutions are essentially running off cobbled-together macro-stuffed Excel spreadsheets, and access control is a joke.

  2. #2 Rachael
    February 5, 2008

    Cool study.

    So, potentially high rewards generally have really low probability. Maybe we unintentionally assume the inverse is true – low probability = high reward? Some of that could be a cultural thing that we learn from organized gambling (eg, when the powerball gets really high, everyone buys tickets even though the odds are super low).

  3. #3 Anne-Marie
    February 6, 2008

    I wonder if this data could be used by politicians to bias the electorate to their cause, for example a candidate perceived a more risky (or inexperienced) can underscore the potential benefits if they do decide to vote for them. The conservative candidate can use fear to people less inclined take a chance on someone new. Maybe it is being used and I am just to brainwashed to know…..

    This may also be the phenomenon that underlying the idea of priming suggest by Malcolm Gladwell in his book Blink.

    Love your blog!!

  4. #4 Larry Smith
    February 19, 2008

    Anne-Marie, Clinton vs Obama sounds like it.

  5. #5 güzel sözler
    February 28, 2009

    This may also be the phenomenon that underlying the idea of priming suggest by Malcolm Gladwell in his book Blink.

    Love your blog!!