Sunday, August 22, 2004

Crime control

This strikes me as simple and right.


Rousseau said...

Uh, this is the kind of elementary economic logic that most social theorists know to run away from. A different way of putting it would be, we're trying to make the costs of doing crime higher than the reward (R), and we can effect the probability of getting caught (P) and the punishment and jailtime(J). We want R< P*J. We can affect P and J, but increasing P is really costly (cops on streets, better prosecutors, etc). So just increase J a lot. 50 years for pickpocketing would deter most pickpocketters.

The problem being that practically the world has a lot of other factors, and when you bring those in, increasing just "the threat" looks pretty bad. Giving police life and death power over any committer of any minor infraction, along with high randomness, starts to make a vicious and corrupt police force. Or making the punishment for any crime super high, gives criminals more motive to resist arrest and less to lose by committing other crimes to get away (like it's be worth it to kill a cop in order to not be arrested for drunk driving).

Sadly our justice system does have to be hindered by the laborious process of doing things right and fairly.

Dennis said...

Rousseau: I think you've kind of missed the point here. The idea seems to be increasing the credibility of threats of prosecution, not necessarily the actual punishment or detection rate. In particular, Kleiman doesn't seem to say we should rachet up jailtime, only that we should find ways to make people believe that the costs of crime are high.

The thing that he misses is subtler: anybody you catch committing a crime has to be arrested and prosecuted, because otherwise you undermine your threat structure. That is, threats of punishment for criminal activity are credible exactly and precisely as much as criminals get caught and punished, so anytime you get a chance to catch a crook, you have to do it. That said, maybe we as a society should hire more police simply to drive around and increase threat credibility, but only after everything we actually catch has been dealt with, which is currently an uphill battle.

The good point he reminds us of is still there though: more prosecutions means prosecutors doing their jobs, yes, but usually also an increasing crime rate, which means the fight is being lost.

Rousseau said...

The article seems either very much based on the flaw I mentioned, or completely non-substantial. Statements like "the purpose of police is crime control" and "maximum of threat with minimum of activity" seem rather focused on the simplistic economic logic I listed above.

If you think all he meant was a focus on arrests and convictions is faulty, well I don't think that's news to anyone. Individual officers and DA's might brag about their records, but society, their politician superiors, and the civic minded public servants these people are, all seem pretty well aware that the goal is to keep crime from happening. I think we'd be hard pressed to find anyone in an important law enforcement role who has demonstrated this flaw that Mark believes is prevalent enough to be a public problem.

You say he misses the subtler thing, that you have to back up your threats and here is how. I don't think he's considering anything anywhere near that line of logic though, as he puts forth no ideas on how you would be increasing threats or reducing costs or what not.

Perhaps this was linked to a news article where some public official was rather egregiously bragging about arrests and prosecution rates. But outside that, the statements seem completely devoid of meaning or suggestiveness in how to run law enforcement. Which is why I thought at least maybe it was just an opening into a discussion about increasing punishment and decreasing arrests.

Max Goss said...

This guy says:

(1) "a really convincing threat will virtually eliminate the activity it's directed at (or crime by the person it's directed to), thus eliminating the need to deliver on the threat."

A common argument against the death penalty is:

(2) The death penalty does not really deter.

Which of these do you reject, Neil? Or do you have a way of reconciling them?

Neil Sinhababu said...

I definitely accept the one about how the death penalty doesn't deter. I'd be willing to support the death penalty if I thought it had a deterrent effect (and if I were confident that it wasn't being used to kill lots of innocent people). I just haven't seen any evidence to convince me of that.

I accept #1 as well. I just think the highest-powered threats in our criminal justice system aren't that convincing. My guess is that most criminals are fairly confident that they'll be able to get away with their crimes, and the rest act irrationally in some way (drug addiction, wild fury, insanity). Kleiman's analysis is, I realize, wrong for the irrational people. It's right about the rational criminals, though it doesn't clearly point to a good solution to that problem. As for the death penalty, it isn't so much scarier than life imprisonment that it'll change behavior.

The real beneficiary of Kleiman's argument, I think, are those late-night shows on the Discovery channel about people using complicated forensic science to track down murderers, and especially to catch husbands who kill their wives. In Rousseau's terms, these shows are how you get P to go up.

(ps - why "Rousseau"? And are you coming to my house this weekend?)

Dennis said...

Rousseau: Yeah, I consider the article pretty close to vacuous. My own personal bias to believe things true/useful if at all possible. I certainly can't see it having, say, any substantial policy implications, where "policy" is understood in the broadest possible sense; my understanding of the article is that it thinks we should increase threat credibility and decrease following through on threats, two goals whose contradictory nature undermines their compatability except for very minor circumstances.

Rousseau said...

Neil: It's a nice fundamental political philosopher, that I feel is somewhat overlooked in the modern state. Also, the context of my blog is about giving more power to direct democracy and less belief in classically liberal rules about to run our society (ie, the Constitution).

Anyway, what's this "if it has deterrent effect and doesn't kill innocent people"? Since the deterent effect (if it exists) is almost certainly greater in preventing murders than the number of innocents who mgiht face it, why does good utilitarian you care?

Neil Sinhababu said...

I actually deny your premise, Rousseau -- I don't see a reason to think the extra deterrent effect of capital punishment saves more lives than the number of innocents that get killed. Life without parole vs. death doesn't seem a big difference in deterrence to me, but not being able to free an innocent who you kill versus being able to free an innocent you imprison is a big difference in rectification. (Perhaps I just don't know enough about criminals and life imprisonment doesn't sound that bad to them -- but I doubt this.) And given the crazy stuff that Gov. Ryan found in Illinois, I'm quite worried about how many innocents we sentence to death.

Rousseau said...

Neil: I agree with you that death penalty doesn't have much of a deterrent effect. The question was, if it DID have any significant effect (and those who endorse it seem to think that the threat of death is a pretty big one), wouldn't that be worth the cost of innocent lives (to you, at least)?

I imagine anything that has significant deterrent effect (say, preventing 1% of murders) would have a much higher effect in terms of lives than the number of innocent people executed.

Also raises the interesting question of what do you mean or care about innocent people getting the ax? A death is a death, with the same dis-utility. If the person's committed murder, just not the one he's been accused of, is killing him any worse or better? (We're assuming here that only the Great Utility Calculator in the Sky knows these things, and that all hypothetical executions are convicted of a crime and have the same deterrent effect). So why do YOU care if an innocent person is executed more than anyone else? (and to what degree do they have to be innocent, of any crime)

Neil Sinhababu said...

There's another purpose of the criminal justice system besides deterrence, and that's removal. Taking bad people off the streets is a good thing, because it prevents future crimes. Taking good people off the streets is a bad thing, because it prevents future good deeds. They both involve a big chunk of badness in causing suffering to that person and in requiring us to pay for the operation of the criminal justice system. I'm willing to let a murderer go free if for some reason it won't reduce the deterrent effect and if he'll do more good than evil out in the world.

Your question does connect to some of the classic counterexamples to utilitarianism -- is it right to hang an innocent scapegoat to save 10 other people from dying in riots? If you fill the example out the right way, you can generate situations where utilitarianism says 'yes' and I do too.

Rousseau said...

Mmmm, you're right, we can make any scenario to get you to say yes. I guess my scenario was tailored to loosely reflect the empirical views many people have about the system. That deterrent effect exists, that many thug defendants are guilty of some crime even if not this one, and that innocent people do eventually get the ax.

So I guess I'm asking, if you had the empirical views of a Southern Sheriff, but the philosophical views of Neil, does it really matter how many innocent people are executed.