Did you know that many scholars still believe the death penalty deters crime? I didn’t.

Among others, Richard Posner, an influential appeals court judge and legal scholar, and Nobel laureate Gary Becker have both recently endorsed scholarship suggesting that the death penalty deters homicide.

But a very useful article by John Donohue and Justin Wolfers in The Economist’s Voice demolishes that claim (registration required to view PDF). Reanalyzing the most prominent recent paper in the field, which claims to find that each execution deters 18 homicides, they find that standard econometric adjustments for the structure of the data yield a confidence interval ranging from “119 lives saved per execution to 82 lives lost” — essentially, a null result. (And there are a number of other, more technical problems with the paper in question — see also this followup.) Donohue and Wolfers then reanalyze a superior dataset and find that most model specifications suggest an increase in executions is associated with a limited increase in the number of homicides.

In short, there does not appear to be any convincing empirical evidence that the death penalty deters homicides. Some may support the death penalty for moral reasons alone, but for most Americans, that is not enough. I suspect that the combination of this finding and the threat of executing innocent people will cause more states to roll the policy back in the future.

Update 7/18 10:46 AM: Here’s the Stanford Law Review paper on which the Donohue and Wolfers article referenced above is based (registration required to view PDF).