The "Pardon Effect" - economic theory applied to the death penalty
January 24, 2002 2:02 PM   Subscribe

The "Pardon Effect" - economic theory applied to the death penalty Pardons increase homicide rates. "The pardon effect is a matter of simple market economics... pardons reduce the chances that convicted killers will pay the ultimate price for committing murder. Thus a pardon represents a decrease in the cost of committing the crime and is accompanied by an increase in the homicide rate. Conversely, if you increase the "cost" of committing murder by executing a larger share of convicted killers, then economic theory suggests that the murder rate will fall" . The chairman of the economics department at the University of Colorado claims to have the data to support the theory from their analysis of all 6,143 death sentences handed out between 1977 and 1997 in the United States.
posted by Voyageman (23 comments total)
Lies, damn lies and statistics.

I know that if I ever had to consider killing someone, heaven forbid, the death penalty would not factor into my thinking at all. Murderers (in the U.S.) never expect to be caught.
posted by stevis at 2:20 PM on January 24, 2002

While I believe in perfect world that the death penalty would be effective it is not because of the economic factor. It is a fact that the death penalty is not distributed evenly & is nearly always used to kill people of lower economic status. Until that is resolved, any analysis will be inadequate.
posted by thekorruptor at 2:28 PM on January 24, 2002

This assumes that criminals weigh up the consequences of their actions, and do some sort of cost-benefit analysis. My understanding was that criminals tend not to think along these lines, i.e. they generally absolutely believe that they will not be caught.
posted by phatboy at 2:47 PM on January 24, 2002

One rule of statistics is that once you've found a relationship in a set, you can't use it to predict outside of your observed domain. The reasoning is that the overall relationship might be a lot more complicated than the relationship of the subset you've looked at.

The implications the article speaks of are way out of the observed domain.
posted by skyline at 2:47 PM on January 24, 2002

We seem to have reached the point where people think that everything can be described in economic terms. Sure, there may be similarities between economic theory and certain social behavior patterns, but human behavior is much more complex than cost-benefit analysis or "maximizing utility". These people don't seem to understand that humans are more than the sum of their economic activities.
posted by jnthnjng at 3:03 PM on January 24, 2002

can you say "ex-post"?
posted by signal at 3:08 PM on January 24, 2002

Yes. How does this jibe with this article from the NY Times on altruistic punishment? The point being that you are sometimes willing to suffer as long as someone else suffers more.

Doesnt anyone here remember playing Monopoly or Sorry with a group of players, being wronged by some other player, and then being determined to take the other guy down, almost irrationally, even if it meant certainly losing the game yourself? What kind of cost-benefit analysis is involved there? Maybe its just me.
posted by vacapinta at 3:12 PM on January 24, 2002

Do pardons increase the number of murder-suicides too? (being a smart-ass)
posted by vacapinta at 3:14 PM on January 24, 2002

If you remove the death penalty the argument about the number of pardons being responsbile for the murder rate disappears. Problem solved.
posted by vbfg at 3:21 PM on January 24, 2002

Sociologist Joel Best has written on bogus social statistics. Unfortunately, since Tuesday night, I can't get on the site here proper (I'm on and, therefore, the links cgi doesn't work--so excuse the format:

One obvious counter to this post's linkt is to examine the murder rates in death penalty states in ratio to executions per state. The executin'est states tend to have the highest murder rates.

The death penalty is not a deterrent but state sanctioned retribution. Victim's families are all for it as it gives them 'closure' but when you factor in the appeals process, it's a very economically expensive form of 'closure' in the form of revenge, considering the taxpayer money spent.

Of course, if the process was shortened, Illinois, among other states, would have already executed several innocent men.
posted by y2karl at 3:24 PM on January 24, 2002

That article is total nonsense. Period. It wouldn't be worth discussing but some idiot out there may think it makes sense and favour death penalty because of it.

One would think BIG risk is often connected to BIG reward, and therefore a murder with a big enough risk of being taken(and killed) would command a very big reward ; no big reward, no big risk.

But many times
1) the monetary ($$) value of the reward isn't a factor ( one hungry person would kill for a dollar worth of food)
2) decisions aren't taken rationally , because of lack of intellectual skills , wrong information or altered mental conditions (drug, alcool, depression etc)

So it's quite obvious that the economic-murder theory doesn't take those two factors in consideration, so that we can say that this theory is only just-another-academic-model.
posted by elpapacito at 3:41 PM on January 24, 2002

Then there is the story of Richard Marquette, convicted of murder and dismemberment in 1961, paroled in 1975, met Betty Wilson, killed and dismembered her. Fear of going to prison certainly wasn't a deterrent, while the death penalty would have deterred his later murders. Neither prison nor the death penalty are deterrents for would be criminals, otherwise all the prisons would now be empty. They are for punishment and the protection of society.
posted by Mack Twain at 4:41 PM on January 24, 2002

I think this theory works not for murderers, but for white-collar criminals such as Marc Rich and potentially Kenneth Lay. It's definitely no secret that if you're rich enough and influential enough, you can squirm your way out of any scandal. The same could be said for pardoned ex-presidents, etc. The theory definitely holds true for crimes such as those.
posted by cell divide at 4:56 PM on January 24, 2002

Most murderers do "expect to be caught", as much as they expect anything to follow from the act. They're not weighing the decision up like you would a house purchase, thinking "should I kill this guy or not? what will it cost me?". Ask a police officer: most murders occur in reasonable hearing or sight of witnesses, or have only one or two possible assailants, and as such they are easily solved. Very few murderers make any rational attempt to "cover up" what they have done, or deny having committed the crime if confronted. Society's attitude to murder affects murderers too.

Why kill people? Many murders occur as a result of violent altercations, like fistfights, domestic violence and robberies that get out of hand. The other common type of murder is vengeance killing, in response to some insult or humiliation, either a single incident (as in road rage pistol-shootings) or a series of incidents over a long time (as in stabbing one's father or wife in the night). Organized crime killings are almost always vengeance killings, although sometimes this is pre-emptive retaliation (get him before he gets me). Under most legal systems, provocation is a defense to the charge of murder, and if the provocation is reasonable, the defendant may instead be charged with manslaughter or negligent homicide.

Self-defence with reasonable force is always a defense to "bodily injury" crimes such as assault, battery and murder; you are allowed, legally, to defend yourself against being killed by killing your assailant. Matters such as the relative size of the combatants, the circumstances (eg night burglary) and past history are taken into account. In the classic "battered wife" cases, the woman is definitely a victim of torture, and probably (and this is the central point) would be killed by her husband at some stage in the future. Accordingly she may legally defend herself by killing him.

This shades into the defense of insanity. A person who is insane is not responsible for their actions; the inverse of this is the legal definition of insanity. If you were responsible for what you did, you can't be insane. Obviously, outrageous behavior such as murder, is not in itself proof of insanity. Momentary insanity--entirely likely under the circumstances--is the hardest to prove of all.

As to how murderers should be punished, personally I don't believe the death penalty is appropriate for murderers whose crime was a one-off incident in an otherwise ordinary life. The battered wife, for instance, only wants her husband dead. Simply "being emotionally capable of murder" is not enough to make a person a danger to society. I would venture to say that almost all males, and most females, are emotionally capable of murder, if sufficiently provoked and under personal threat.

The death penalty, if it is to be used at all, should be used on untreatable sociopaths, the sort of people who are proven to be unable to have a "normal" existence in society. Serial muggers, rapists, vandals, and suchlike. Along with this (polishing my classical liberal memership badge here) I advocate legalization of pretty much all drug use (apart from stuff like crack and PCP that actually does send its users violently insane), and I advocate extensive, free-to-user, mental health services. The death penalty should be applied as an admission of the failure of the criminal to exist in society, and society's failure to cure that criminal's psychological problems.

posted by aeschenkarnos at 5:59 PM on January 24, 2002

I don' have the nimbers, but aren't anti-death penalty types always claiming that it's not a deterrent?

It seems like everyone makes conclusions based on how they feel about murdering someone, not based on the criminal mind.
posted by Yelling At Nothing at 6:02 PM on January 24, 2002

Sheesh, I oppose the death penalty for a number of reasons including but not limited to racism. But, you guys are jumping all over this guy for limited reason. Y2karl, you talk about bogus stats and then mention that states that execute people have higher murder rates? There are a million problems with this, but the most obvious is that greater crime might lead people to adopt harsher laws.

I personally wouldn't be surprised if it had some deterrent power. Even if only a small percentage of criminals are influenced by it, the overall level of crime will be influenced. So counter-examples of undeterred murders are worth what exactly? We bash a guy who at least probably tried to do some statistics with examples and anecdotes?

It seems like the only decent objection that can be made without looking closer at his work is to argue that social statistics are all (or mostly) crap. You could probably persuade me of that. I mean, what can you really establish about the possible "hidden variables". Of course, this leaves you with no way to establish that the death penalty does not have deterrent power.

If indeed, we anti-death penalty folks have been saying that it is not a deterrent we must accept that there could be scientific evidence that it is a deterrent.

(p.s. "The plural of anecdote is not data." --Marjory Blumenthal)
posted by Wood at 6:22 PM on January 24, 2002

A correlation is not the same as causality.

It is interesting that there may be a correlation between pardons and the murder rate. Assuming that the statistics is sound and all, the problem is his theoretic starting point: that the murder rate goes down because the pardon rate goes down. What if there are more pardons when the murder rate is low because the government is feeling charitable? The causality may go the other way.
posted by dness2 at 7:05 PM on January 24, 2002

Just because a criminal doesn't weigh lethal injection versus life in prison when he's deciding whether to commit a homicide, doesn't mean that the availability and vigor of enforcement of the death penalty doesn't contribute, in a broader way, to reducing the overall appetite for risk of the homicidally inclined. Individual actions are by definition statistically insignificant, and the supposition that criminals are not, at a stastical level, rational actors, is simply unsupported.

Correllation should not be confused for causation, true enough, but correllation is always evidence of causation, worthy of being weighed against other evidence and testing as a hypothesis, in drawing conclusions. Many posters appear, far from wishing to see the hypothesis tested, in fact are guilty of a far worst logical fallacy than c=c ... attacking an assertion about the deterrent effect against future homicide of the death penalty with completely irrelevant complaints about who among those who have already committed homicides are selected for the death penalty, i.e., something that has nothing whatever to do with the deterrent effect or lack thereof the death penalty.

Interestingly, if we juxtapose one hypothesis (that the death penalty statistically deters homicides) and another hypotheis (that the death penalty is imposed in a racially biased fashion) then we could draw a second order hypothesis: that the death penalty should reduce the number of homicides committed by those who are targets of the racial bias. Since the overwhelming percentage of homicides are committed by members of one race against members of the same race, we could conclude that death penalty bias actually disproporianately protects members of the supposedly discriminated against race.
posted by MattD at 7:30 PM on January 24, 2002

You did read the link in my comment, Wood?
posted by y2karl at 8:41 PM on January 24, 2002

Yeah, I read it. I don't see the connection. But, upon re-reading, you mention the stat as a "counter" to what the original poster is suggesting. Which is not the same as trying to use the isolated stat to "prove" that the death penalty is not a deterrent. But still, isn't part of the point of your article that isolated stats (like yours) without context or justification should be regarded with suspicion?

By the way, I believe this is the paper. I think we'd have to read it to get any more out of this discussion. Personally, there's no way I'm going to be able to read it now. Who really cares, anyway? I'm against the death penalty for other reasons.

(p.s. From the paper: "To control for the impact of the 1995 Oklahoma City bombing, we included a dummy variable". OK, cool.)
posted by Wood at 9:02 PM on January 24, 2002

Your point is taken, Wood--my comment was written in haste. I can get on MetaFilter except via and it takes so-o-o long to load. If i'd thought it out and had the time, I would have stated it more elelgantly--something along the lines of

As I've read that the states which execute at a higher than per capita rate (for death penalty sates), one would think that there should be some statistical evidence showing a reduction in murder rates. Of which I should assume we would have heard. But they seem instead to have the highest rates for violent crime. Still.

But for a fact, I have written Professor Best to ask his opinion, and if he answers, I'll share it here. The link I provided, upon re-reading, is not the best. Ah, comment in haste, repent at leisure.

But Best is interesting to read about social problems and what people in his field call 'claims makers' or 'moral entrepeneurs'--I'm not saying the professor who wrote the paper in the post is necessarily one, but his paper is going to be ammo for people who are.

I don't favor the death penalty for various reasons, such as that historically it has not reduced the crime rate, like in England during the Industrial Revolution, when harsher and harsher laws were passed and one could be hanged in theory and in fact for stealing a loaf of bread, say, yet the crime rates did not go down. And, I agree with you that that it's applied far too harshly and unequally on raical minorities, and that that alone is reason enough for it to be jettisoned. As if its apparent lack of deterrence were not enough.

I think the death penalty has s social function, just not a very pretty or utilitarian one. I've linked in the past to Girard's scapegoat theory, which is how I see the whole spectacle of crime and punishment played out--as a sort of religious ritual, a drama where crime isn't affected so much as a sort of communal release and reaffirmation is done.

And when I look at the executin'est nations,--China, Afghanistan under the Taliban--I don't see the US in such good company. What the rest of the industrialized world must think of us...
posted by y2karl at 9:48 PM on January 24, 2002

We seem to have reached the point where people think that everything can be described in economic terms.

Well, isn't that the point of Public Choice economics (beyond that, suggesting "We" have gotten anywhere as a collective entity is a bit of a generalization)? The problem with Public Choice is that it's been overextended. Because of its success in certain areas, it's been applied to everything.
posted by yerfatma at 5:02 AM on January 25, 2002

MattD: Individual actions are by definition statistically insignificant, and the supposition that criminals are not, at a stastical level, rational actors, is simply unsupported.

This isn't necessarily so. But I can't seem to prove it other than by pointing to this NYtimes article that I can't link to, because it's too old. You'll have to pay to find out that I'm right. I wonder what rational choice ecomonics has to say about that connumdrum.

The Quirks of Behavior, as Part of Economics

The article, I recall, discusses a lot of ways that economics incorrectly predicts human behavior, since there are so many variables that it's silly to assume that people, even in the aggregate, do a successful cost-benefit analysis of the outcomes.

As this relates to the death penalty, I think that murders are almost always committed by people of less-than-perfect rationality, either in the moment or all the time. But I think the bigger point, about economic theories not being perfect predictors of human action, espeically in 'real world' (and therefore complex) issues, is interesting.

posted by zpousman at 10:16 AM on January 25, 2002

« Older   |   Fireplace log may cause fire. Newer »

This thread has been archived and is closed to new comments