Deterrence and the Death Penalty

by Kieran Healy on March 24, 2005

Eugene Volokh quotes extensively from a new paper by Cass Sunstein and Adrian Vermeule that presents an argument for the death penalty. It begins by reviewing recent studies that find the death penalty has a deterrent effect on potential murderers. In particular:

Disaggregating the data on a state by state basis, Joanna Shepherd finds that the nation-wide deterrent effect of capital punishment is entirely driven by only six states … [The states] showing a deterrent effect are executing more people than states that do not. In fact the data show a “threshold effect”: deterrence is found in states that had at least nine executions between 1977 and 1996. In states below that threshold, no deterrence can be found. This finding is intuitively plausible. Unless executions reach a certain level, murderers may act as if the death is so improbable as not to be worthy of concern. Her main lesson is that once the level of executions reaches a certain level, the deterrent effect of capital punishment is substantial.

This is an elegant idea, but trouble with it is that only few states execute anyone in a given year. Most execute no-one. A tiny few—notably Texas—kill a lot of people in some years. As a result, evidence for a threshold deterrent effect depends on a very small number of observations. In a nice analysis of state-level data from 1977 to 1997, Richard Berk shows that just eleven state-year observations out of a thousand drive the deterrent effect. It’s possible to mess around with the specification a bit to get a less strongly skewed measure (by standardizing the number of executions by the number of death sentences, say) or making the data more fine-grained so that you have more observations (using county-quarters as a unit, for instance), but in the end its hard to escape the worry that about 1 percent of the observations are behind the results.

We’re probably witnessing the birth of a dubious stylized fact about deterrence and the death penalty. I don’t doubt that the Sunstein and Vermeule paper raises a bunch of interesting questions, but the empirical results they rely on just don’t seem that robust. This is a bit ironic given their argument that “The widespread failure to appreciate the life-life tradeoffs involved in capital punishment may depend on cognitive processes that fail to treat ‘statistical lives’ with the seriousness that they deserve.” One of these processes is the tendency to latch on to a cool finding a bit too quickly. Negative results (like the ones reported in Berk’s paper) are just not as interesting, unfortunately.

{ 1 trackback }

Crooked Timber » » Violent Societies
03.24.05 at 5:15 pm

{ 36 comments }

1

bad Jim 03.24.05 at 3:27 am

The results of statistical analyses performed in the course of America’s fitful progress toward the abolition of the death penalty are unlikely to be enlightening, unless they can offer such results as that vegetarians are highly likely to be deterred, meat-eaters considerably less so, and smokers hardly at all.

Some of us have long been persuaded by cruder observations, such as the fact that, back when theft was punished by public hanging, the crowds watching the execution were at risk of having their purses snatched.

For those who believe in the death penalty, the idea that it may have a deterrent effect is but one dubious arrow in a quiver full of rationalizations.

2

anatoly 03.24.05 at 3:31 am

Well, at least it’s an argument. Most people I’ve seen argue against death penalty treat its supposed failure to deter as an article of faith.

3

Chris Bertram 03.24.05 at 3:39 am

The argument depends crucially on accepting the premise that, where the state is concerned, omissions and acts are morally on a par. We’ll see how far any rightists who jump on this paper are willing to endorse that principle! (For example, given the very high rates of premature mortality among African-Americans … and the poor generally … are they willing to hold the US government responsible for those deaths and hold that it is morally required for the government to put in place the kind of health care system, infrastructural investment etc etc that would reduce them? No? I thought not.)

The slippery slope section of the paper is amazing, for it concedes that _if the facts were known_ the argument might justify killing the innocent, torturing murderers, (torturing murderers’ children? etc. But the facts aren’t know so that’s alright then! But surely, the authors shouldn’t have stopped there because the government surely has (on their view) a moral duty to find out what the most effective means of deterrence are, and to carry out surveys, experiments, trials etc to discover this? True, (on this view) if it turns out that torturing murderers’ children isn’t an effective means of deterrence then such methods shouldn’t be employed, but not even to research into whether it might be? The height of irresponsibility!

Truly a vile paper and potentially a very damaging one, since we (and the paper’s authors) know full well that it will be seized upon by many who favour the conclusion (and the “Study Shows Death Penalty Morally Obligatory” headline) despite the fact that they would not, in fact, endorse the underlying principles consistently.

4

bad Jim 03.24.05 at 4:27 am

Please tell me that this was written with tongue in cheek:

Even if we reject strong versions of the Precautionary Principle, it hardly seems sensible that governments should ignore evidence demonstrating a significant possibility that a certain step will save large numbers of innocent lives.

The same logic would suggest that it was necessary to invade Iraq to save Western Civilization, or that it would not have been insane to have bombed the Soviet Union, as well as, more conventionally, the contrary propositions.

5

Tom Maguire 03.24.05 at 7:14 am

Interesting point. How would we relate the Precautionary Principle to the issue of global warming, or preparing for an outbreak of avian flu?

Anyway, good job, Kieran – I never post on guns, but today I may just remind myself why that is.

6

goesh 03.24.05 at 8:01 am

An execution will prevent as many as eighteen murders? Talk about cooking the numbers! I actually spewed my coffee when I read that. With so many potential Ted Bundys running around we need to start hanging petty thieves again, don’t we?

7

Luc 03.24.05 at 8:15 am

As there are always a few politicians over in Europe willing to reinstate the death penalty, I thought let’s try to read it. But luckily we’re saved from this silly argument.

The authors state that a “swedish-style welfare state” might reduce murders even more, but implementing that would not be “feasible”. While the death penalty is “feasible” in the US. Thus the death penalty it should be. As the situation is reversed in the EU, that’s the end of the story.

“Political restraints will rule out some policies …”

No morality involved at all.

8

Steve LaBonne 03.24.05 at 8:59 am

luc, didn’t you know that moral judgements are to be reserved for things of which conservatives disapprove? ;)

9

Matt 03.24.05 at 9:37 am

In general I find myself very disapointed w/ Cass Sunstein. He seems quite smart, and I assume he is. But he writes (or has his name attached) to so much each year that it seems quite impossible for me to believe that it’s all clearly thought out and worked through. And, since he’s so well known and influential among lawyers, this strikes me as quite dangerous and irresponsible. If Kieran is right (I don’t doubt it, but can’t competantly judge myself) then this is a really bad example. Someone ought to get him to slow down and work his ideas over a bit more before spewing them out half (or less!) baked.

10

P ONeill 03.24.05 at 9:38 am

Of course the fact that Volokh is dealing with this issue wouldn’t be at related to Bush getting a new round of questions about why exactly, post Schiavo law, he supports the death penalty, would it? McClellan has been trotting out the deterrent rationalie in his usual robotic way but even the loony Rick Santorum seems to recognise that that dog won’t hunt and is showing signs of packing in his support for the death penalty. In the end the deterrent proponents are always reduced to “well, if we really had a death penalty, the deterrent effect would be easier to spot in the data.”

11

Thomas 03.24.05 at 10:09 am

Chris, it might interest you to know that Vermeule is thought of as being on the right, while of course Sunstein is on the left.

12

spencer 03.24.05 at 11:47 am

I read the other references and the ones debunking the study sem to have the much stronger case. I would just make one suggestion.

If you accept the premise that each excution reduces the number of murders by to 18 how do you explain that the conclusion does not seen to apply to Texas. If the study is valid shouldn’t Texas have the lowest murder rate.

13

Jack 03.24.05 at 12:01 pm

Why restrict to US data? Is force the only language Americans understand?

I’m sympathetic to Kieran’s argument about the data but isn’t this a bit like the Lancet affair? That’s all the data there is and it suggests what it suggests.

Of course this is nitpicking. Shepherd’s paper is far more informative than the Sunstein one although she cites John Lott without comment and the sophistication of the analysis of non-death penalty factors is much less than that of the death penalty factors. In particular she points out that in 39 states there is no deterrent effect.

That makes the Sunnstein paper, which quotes it, ridiculous. All that is left is an ” if executing a murderer saved 18 innocent lives it would be worth doing” which is kind of trivial. It’s chief content is the “lesson” drawn from Shepherd’s paper. We might draw the opposite conclusion if the lesson they chose to draw was instead that although in 6 states there was a deterrence effect, in 13 states executions actually increased murder rates?

Even given the conclusion there are plenty of other fallacies to overcome. For example the idea that deterrence is some kind of universal constant so that if you have a low murder rate it would be lower if there were more frequent executions. Would Japan have an even lower murder rate than it does if it killed people more regularly? Would Ireland if it introduced the Death Penalty. Is Europe missing out?

14

Nicholas Weininger 03.24.05 at 12:07 pm

Well, I’m glad, in a sort of perverse way, to know that I find Sunstein’s utilitarian realism no less repulsive when directed toward rightist ends than when directed toward his usual leftist ones.

15

Jack 03.24.05 at 12:12 pm

Spencer, I imagine that the 18 lives is for the first execution over the threshold. The Sunnstein paper actually has a footnote pointing out that a thousand isn’t necessarily ten times better than a hundred. Could this be to justify not accepting that the extreme morality of Chinese capital punishment is a good reason for Europe to drop its embargo on arms exports to China or US laxity in only selling three times as much as hte EU?

16

Sebastian Holsclaw 03.24.05 at 1:07 pm

“The slippery slope section of the paper is amazing, for it concedes that if the facts were known the argument might justify killing the innocent, torturing murderers, (torturing murderers’ children?”

Thank goodness we have a retributive element to the punishment system which wouldn’t allow killing the innocent or torturing murderers’ children.

We had an interesting recent post decrying the retributive element in the punishment system, perhaps we should revisit the issue.

17

Katherine 03.24.05 at 1:18 pm

“Thank goodness we have a retributive element to the punishment system which wouldn’t allow killing the innocent”

maybe not on purpose, but as long as we don’t mean to it’s apparently just dandy.

18

Sebastian Holsclaw 03.24.05 at 1:29 pm

True in all systems of punishment.

So I take it you are accepting the retributive parameters on punishment by your comment? Fantastic!

19

Katherine 03.24.05 at 2:15 pm

The Texas death penalty system is recklessly or knowingly killing the innocent.

“So I take it you are accepting the retributive parameters on punishment by your comment? Fantastic!”

uh, I said I did last week. More or less on the same page with Mark Kleiman’s excellent post. “Just desserts” as an upper bound, and all that jazz.

Kleiman’s position is NOT the same as Volokh’s position. The thing about punishment as moral condemnation versus sheer vengeance: you can’t commit the same act you’re supposedly condemning. You don’t condemn torture by torturing, rape by raping, or killing the innocent by killing the innocent. I tried to explain this to Volokh but he didn’t really get it.

20

Sebastian holsclaw 03.24.05 at 2:30 pm

Good, unfortunately that doesn’t put you in the majority of CT commentors (and of those who have commented on it, CT contributors).

21

micah 03.24.05 at 2:42 pm

I think what Chris says above is exactly right. The discussion of objections to consequentialism is appalling. Sunstein and Vermeule offer a feeble reference to Rawls’ Two Concepts of Rules and invoke rule-consequentialism as a defense. But then maybe they should address Rawls’s later rejection of that position? As it is, the claim that deontologists would/could go along with them is laughable.

But set that aside for the moment. One might think that retributivists like Volokh would be up in arms over this paper. Sunstein concludes by arguing that deontological moral intuitions are caused by the irrational application of moral heuristics to complicated moral problems. And for that reason, such intuitions are not reliable. Since retributivists like Volokh ignore the value of statistical lives, they, too, can be dismissed as delusional.

Volokh is happy with the results of Sunstein and Vermeule’s, but he shouldn’t be. He should be worried (or outraged), as should those who reject the death penalty on deontological grounds, about the underlying argument, namely, that anyone who isn’t a consequentialist is either stupid or crazy.

22

Jake 03.24.05 at 2:43 pm

With respect to what Jack wrote, why indeed is the argument restricted to the US? How come Australia, for instance, has a murder rate a third of the US’s, without that cool deterrent effect?

23

Katherine 03.24.05 at 2:44 pm

remember my whole thing about how confusing original sense and original application led conservatives to wrongly argue that both were binding and liberals to wrongly argue that neither is?

I would say the same thing is going on with confusing justified retribution & vengeance.

And, actually, with confusing euthanasia & assisted suicide with the right to refuse medical treatment.

24

Sebastian Holsclaw 03.24.05 at 2:49 pm

Btw, I’m not on Volokh’s side about torturing child molestors. I just think trying to reject his argument by saying that retributive justice isn’t a large factor in the criminal justice system (as attempted in the earlier threads on the topic) is both false and dangerous.

25

spencer 03.24.05 at 4:11 pm

If you read the referenced Berk & Goertzel papers it becomes obvious that this is a classic case of someone torturing the data until it confesses to a crime it did not commit.

I will not accept any paper or study on the deterent impact of the death penalty that does not explain why Texas has such a high murder rate and excution rate. If you believe the death penalty had a detertent impact you have to explain why it does not work in Texas.

26

Sebastian Holsclaw 03.24.05 at 4:45 pm

“If you believe the death penalty had a detertent impact you have to explain why it does not work in Texas.”

Do you believe that everywhere in the world has the same murder rate if you factor out the detterent effect of capital punishment? Everywhere in the US?

27

Phillip J. Birmingham 03.24.05 at 8:06 pm

Do you believe that everywhere in the world has the same murder rate if you factor out the detterent effect of capital punishment? Everywhere in the US?

No, but you’d think that if the death sentence were all that effective a deterrent, Texas would have a much lower murder rate than New York, since it has the death machine cranked up to 11. Instead, New York pegged a 5.0, and Texas, 5.9.

An argument that boils down to, effectively, “imagine how much worse my cancer would be if I weren’t taking all this Laetrile,” is a pretty weak one.

28

Michael Cross 03.24.05 at 8:08 pm

Canada had a much lower murder rate than the US to start with. That murder rate has dropped steadily since capital punishment was abolished more than 3 decades ago (although there had not been an execution in Canada, even when capital punishment was legal, since 1962). Canada is a distinct society but one that shares popular culture, media and economic structures with the United States. It would seem, intuitively, that social, economic and cultural developments do much more to explain rates of murder than the presence or absence of capital punishment.

29

MattXIV 03.24.05 at 8:36 pm

cathrine,

“you can’t commit the same act you’re supposedly condemning. You don’t condemn torture by torturing, rape by raping, or killing the innocent by killing the innocent.”

Wouldn’t that rule out punishment altogether? The reason and actor are as important as the nature of action. If not, what separates fines and taxes from theft, imprisonment from kidnapping, etc.

On a somewhat related point, I wonder if the death penality itself is even the main source of violence in the law-enforcement process. I’d guess that the probability of being shot by the police in the process of committing a crime seems much higher than the probability of being formally executed by the state – see here to get a quick look at the relative numbers.

30

Katherine 03.24.05 at 9:06 pm

has anyone tried comparing the rates of murder to the rates of other violent crime?

31

Sebastian Holsclaw 03.24.05 at 10:38 pm

“No, but you’d think that if the death sentence were all that effective a deterrent, Texas would have a much lower murder rate than New York, since it has the death machine cranked up to 11. Instead, New York pegged a 5.0, and Texas, 5.9.”

Isn’t that why you compare murder rates over time? And isn’t there a period where the Supreme Court outlawed the death penalty to help out as a pseudo-control?

32

Phillip J. Birmingham 03.24.05 at 11:15 pm

Isn’t that why you compare murder rates over time? And isn’t there a period where the Supreme Court outlawed the death penalty to help out as a pseudo-control?

Okay, murder rate in Texas, 1972 to 1976
12.4,12.8,13.7,13.4,12.2

1977 to 1981
13.3,14.2,16.7,16.9,16.6

From 1995 on, when Dubya cranked it up to 11, the numbers were lower:
9.0,7.7,6.8,6.8,6.1,5.9

Sadly, each of these years, W-less New York suffered under these crushing rates:
8.5,7.4,6.0,5.1,5.0,5.0

Throughout the other periods I’ve shown, NY has had generally lower murder rates.1 When Texas wasn’t killing anybody, NY was lower. When Texas was half-heartedly frying a few, NY had a lower rate. When Texas was enthusiastically needling people, its Governor cackling all the while, NY still had a lower murder rate. Maybe it’s something in the water, but it sure doesn’t seem to have anything to do with how many people Texas has snuffed.

Face it, the Texas Death Machine is the criminological answer to the magnetic copper anti-arthritis bracelet.

1. I think, but did not take the time to compare year-by-year, that it is actually lower eand every year, but I do not assert so. Check for yourself at http://www.disastercenter.com/crime/nycrime.htm if you like. Links to other states can be found at bottom.

33

Katherine 03.25.05 at 2:41 am

““you can’t commit the same act you’re supposedly condemning. You don’t condemn torture by torturing, rape by raping, or killing the innocent by killing the innocent.”

Wouldn’t that rule out punishment altogether? The reason and actor are as important as the nature of action. If not, what separates fines and taxes from theft, imprisonment from kidnapping, etc.”

The reason and actor and legal sanction matter, but torture and torture and rape is rape. You can’t torture or rape to incapacitate or in self-defense or to pay for shared social services. It’s pain for the sake of pain.

34

spencer 03.25.05 at 12:01 pm

phillip Birmingham — thanks for the link to crime data .

Do you have a link to excutuion data?

35

Phillip J. Birmingham 03.25.05 at 1:25 pm

Do you have a link to excutuion data?

Unfortunately, no. I know a few data points — 119 executions in Texas during the Bush regime, none during the Furman period of 1972-76, but not much more.

To be honest, I didn’t look too hard. I don’t have the sociological chops to answer questions about things like lag between cause and effect, so I didn’t think I’d be able to make use of it to add more light than heat.

36

JStat 03.30.05 at 4:26 am

As usual , we see the incorporation numbers into ideology to give it the trappings of objectivity. Assuming that this is true there are many questions still to be answered. First why is that that on average , the south as a region has much higher violent crime rates than the north and also has more executions? What factors are possible determinants of murders and if these are known were they controlled for or dealt with in some manner? Throwing murder rates and presence of absence of death penality into some time series or logistic regression analysis may show some relationship, but technically speaking I can generate a random binomial and normal variable and I will get a correlation and sometimes a significant coefficient. What is the strength and direction of the relationship? And last, and most importantly, why is the US not compared with other economically developed countries that do not have the death penalty? Following Shepherd’s logic, the death penalty is some kind of universal deterrent and thus her analysis should hold in all “similar” countries, which I’m sure it does not. But what does that imply? Does the death penalty “deter crime” only in the US and if so why? Finally, I am a statistician and fully appreciate statistics and probability theory, but to use non-causal and theoretical models based on “lives saved” to determine if someone should live or die is not only poor judgement but shows a lack of nuanced understanding of statistics and the possible amount of information that can be derived from it.

Comments on this entry are closed.