Casting blame far and wide

When something goes wrong, do we need to blame someone?

Yes, according to an article in the Journal of Political Economy by researchers from Boğaziçi University, Bocconi University, and the University.

The researchers ran experiments that paired subjects. One subject had to make decisions about how money would be split between a risky and safe investment in an uncertain environment. After the uncertainty was resolved and the payout was determined, the second subject decided how much to reward (or punish) the first.

There were four possible outcomes from the first subject's decision:

1. The person chose the risky outcome and the risky outcome made money (also more than the safe outcome),
2. The person chose the risky outcome and the risky outcome lost money,
3. The person chose the safe outcome and the risky outcome made money (and would have paid more than the safe outcome), and
4. The person chose the safe outcome and the risky outcome lost money.

Although the first subject chose the investment, s/he did not have any influence over how the uncertainty would be resolved.

The researchers set things up so that under some circumstances

1. the risky investment was the better bet (even though there was still a small possibility of a loss),
2. neither the risky investment nor the safe investment was a better bet, and
3. the safe investment was the better bet (even though there was still a small possibility that the risky investment would pay more).

In this set up, there was some scope for blame-worthy decisions. For example, choosing the safe investment when the risky investment had a higher expected pay-off or choosing the risky investment when the safe investment had a higher expected pay-off would each be "bad" decisions a priori.

What was interesting, however, was that the second subjects paid the first subjects less (punished the first subjects) even when they made the appropriate decisions a priori but the luck of the draw led to bad outcomes. For instance, subjects were often punished when they chose the risky investment, even though it had a higher expected pay-off, but the realized pay-off was a loss. That is, the second subjects appeared to shift blame for bad outcomes that were beyond anyone's control onto the first subjects.

Need someone to blame? Chances are you'll find someone.