Wednesday, April 15, 2009

Rational stupidity

I think a more realistic, but less catchy, title would be "rational stupid actions. And of course, if a politician changes their mind, they will be jumped on by their opposition, and accused of lying."

http://stumblingandmumbling.typepad.com/stumbling_and_mumbling/2009/04/rational-stupidity.html

It can be quite rational to behave stupidly. That’s the claim of this new paper (pdf) by Anne Sibert and Hamid Sabourian - which might explain why highly paid bankers made such a horrible mess.
To see what they mean, start from the premise that bankers are paid not by results, but by perceived competence.
Now, put yourself in the position of a banker who has gone long of mortgage derivatives - toxic assets, to use a phrase I hate - in the belief that the US housing market will continue to boom. You then get a signal that the boom might turn to bust. What do you do?
The obvious thing is to sell your position. But this might not be rational. If you do this, you incur a certain loss. Your bosses think: “Hey, this guy doesn’t know what he’s doing. One moment he tells us to go long of CDOs, the next he’s selling. We don’t need to give this putz a big bonus; who’d be daft enough to hire him?” This certain loss is offset by only a possible gain - the possibility that the market really will bust before the bonus season.
The desire to appear competent - to stick with your previous position - can therefore lead rational people to stick with assets that may well lose money. This happens because our trader is concerned to maintain his reputation, as this determines his salary. Indeed, say Sibert and Sabourian:
It is possible for all experts to continue to predict an event that they know is virtually certain not to occur.
Worse still, the better our trader has been in the past, the more likely he is to be wrong. This is because his response to signs that his original position might be mistaken would be: “There are always noisy signals that I might be mistaken. I’ve been right to ignore them in the past. What’s different now?” As Sibert and Sabourian say: “more competent experts do not necessarily make better predictions.”
There are, at least, two takeaways here. First, bad incentives can cause rational people to look as if they are acting stupidly. To an outsider, our trader looks like he’s committing some combination of the confirmation bias or over-confidence. But in fact, he’s acting rationally, maximizing his expected utility.
It’s not, of course, just in banks where this problem arises. Politicians might stick with bad policies, because the cost of doing a “U-turn” and signalling incompetence would outweigh the benefit of adopting the right policy.
Second, all this might explain why it is banks, rather than hedge funds, who have suffered most. In both, payment depends partly upon actual results and partly upon perceived competence. But in banks, it’s the latter that is (a little?) more important. And the desire to preserve perceived competence might have led to rational irrational behaviour.
==============================================
From a comment to this article

The great George Polya made an important related point. He said that mathematicians and scientists instinctively looked for cases that disproved their theories while other people tended to look for cases that confirmed them.

1 comment:

mamvas said...

Interesting work, Patricia
I will make the effort to read the original document (pdf) by Sibert and Sabourian

Greetings

Post a Comment