Sometimes we just don't understand the incentive.

As you know, Dominique Strauss-Kahn, the head of the International Monetary Fund, has been accused of sexual assault and is currently being held without bail at Rikers. Mr. Strauss-Kahn is an renowned economist and, until this happened, a possible candidate for the French presidency.

Tyler Cowen at the Marginal Revolution blog, asks a question no one thinks to ask (sort of like we do here), how should a "true economist" should react to the news? Mr. Cowen notes that Mr. Strauss-Kahn has strong incentives not to commit sexual assault but is now accused of committing one. The post questions whether the economic theory that incentives can predict individual behavior is central to economics and whether it should be central.

Andrew Leonard at Salon criticizes the post and suggests that reasonable economists think that incentives only matter in the aggregate and not in individual cases. But that can't be right because if incentives don't matter in individual cases, they can't matter in the aggregate.

It seems to me that incentives can be used to predict individual behavior. The trick is to understand the incentives. If the charges against Mr. Strauss-Kahn are true, it could be that that the incentives to commit the crime are larger than the things Mr. Cowen views as incentives not to commit the crime.

By way of example consider, Peter Elliott. As the Milwaukee Journal-Sentinel reports, Mr. Elliott is a disbarred attorney who was sentenced to ten years in prison for committing fraud involving his trust account. Normally, we would assume that people awaiting sentencing for crimes have strong incentives not to commit additional crimes. If they commit another crime they could be sentenced to even more time in jail.

What we might call the normal incentives (desire to avoid punishment, desire to get out of prison as quickly as possible) apparently did not apply to Mr. Elliot. Instead of keeping his nose clean while awaiting sentencing, Mr. Elliott spent some his time robbing a bank. His robbery of a bank in Wales Wisconsin netted him $9,516.

Anyway, as a result of the robbery, Mr. Elliott is now looking at an additional five years in prison. Mr. Elliott earned $1,903.20 per additional year of prison time. That does not seem like a very good return on his crime.

So what were the incentives that drove Mr. Elliott to commit the robbery? He told investigators that he used $2,600 of the money to get his daughter's vehicle "out of hock." Put another way, he felt his child needed his help. As a 62-year-old man facing a ten-year prison sentence, Mr. Elliot may expect to die in prison. Thus, he may have thought this was the last time he would ever be able to help his child. If we view Mr. Elliott's incentive that way, his decision to rob a bank seems less irrational. Mr. Elliott's incentives were simply not the normal ones we might expect from a criminal awaiting sentencing.

It seems possible that Mr. Cowen and Mr. Leonard have simply failed to identify the incentives that would drive Mr. Strauss-Kahn to behave as he is accused of behaving. I can't identify any incentives either. However, just because I don't recognize the incentives does not mean that none exist.

What do you think Mr. Torvik? Is Leonard right? Is Cowen asking the right questions? Should a non-economist be even thinking about these sorts of things?