E-mail response to Steven Landsburg

I responded to Mr. Landsburg's first e-mail:

Mr. Landsburg:

Thank you for reading and responding to my blog. With a total readership of four, I am needless to say quite surprised that you saw my comments at all.

Let me take your points in something approaching reverse order.

> > If Mr. Landsburg were to accept that capital punishment
> > deters no one, do you think he would change his stance?
>
>I'm not sure what stance you mean. My stance in the
>column was that *if* executing vermiscripters has a
>sufficiently large deterrent effect compared to executing
>murderers, then executing vermiscripters is better
>policy than executing murderers. Obviously if capital
>punishment deters nobody (including vermiscripters)
>then neither policy is of any value.
>
>I did say quite unambiguously in the column that
>deterrence is, to my mind, the only possible justification
>for capital punishment. If you're asking whether I
>meant that, the answer is: Of course. Why else would
>I have said it?

I wasn’t happy with the way I phrased that question even when I wrote it. I didn’t mean it to be an ad hominem attack in any way. Inelegantly, I was attempting to point to your statement that the deterrent effect is the only rationale for capital punishment, and then ask what would happen if you were to discount a deterrent effect of capital punishment on homicide.

> > My beef? He bases the value of executing murderers
> > on a "consensus" figure that estimates that executing
> > a murderer deters ten murders. The problem is, it has
> > not, to my knowledge, been empirically shown that the
> > death penalty deters murders at all.
>
>You seem to have completely missed the point here, which
>is that I was trying to *over*estimate the deterrence
>effect on murderers in order to bias things *against*
>my desired conclusion. If the deterrent effect on
>murderers is less than I said it was, then the argument
>for preferring a vermiscripter-execution to a
>murderer-execution gets stronger, not weaker.
>
>In addition to being wrong on the logic, you're wrong
>on the facts. The econometric literature on the
>death penalty overwhelmingly supports numbers in the
>vicinity of those I quoted. The biggest name in this
>field is Isaac Ehrlich (a passionate *opponent* of
>capital punishment, by the way) who, in 30 years of
>working on this subject, consistently gets numbers
>between 8 and 20 for the number of murders deterred
>by an execution. For a balanced overview of the
>literature, you might want to look at
>
>http://www.cjlf.org/deathpenalty/DPDeterrence.htm
>
>The article by Levitt, et.al. finds very little
>deterrent effect and should not be ignored, but it
>also has to be weighed against literally hundreds
>of other articles, almost all of which get numbers
>in the Ehrlich range.


I looked at some of the articles on the overview you sent, although some of them were only abstracted (including the Ehrlich piece). I will attempt a few critiques without the benefit of an education in econometrics or statistical analysis.

First, the writers identify a general correlation between homicide and punishment to which they impute a deterrent effect. I won’t dispute that punishment has a deterrent effect on crime generally, or on homicide in particular. However, I don’t see any examination over the different deterrent effects of a prison sentence versus capital punishment. I suspect that any would-be murderer who is exercising enough discretion to consider punishment at all isn’t going to rate the possibility of life in prison much more harshly than death. That is to say, a “reasonable” would be murderer will consider either penalty so harsh that they would be deterred equally, or nearly so.

Second, logic defies the math in some of these cases. I have read previously, and see here (as in the Sorenson article), that studies of contiguous death penalty and non-death penalty states fail to show increased deterrence within the death penalty jurisdiction. More broadly, the homicide rate seems to increase or decrease roughly proportionally in death and non-death states without reference to capital punishment. Logically, if the death penalty had a much greater deterrent effect than life in prison, we would see a much lower homicide rate in jurisdictions that execute murderers.

That said, my original post was really only tangential to your article, which I found to be quite funny. Your analysis brings to mind why I enjoy reading the pragmatic and well-reasoned appellate opinions of Judge Richard Posner.