Wednesday, October 07, 2009

Debating the economic basis for punishment theory

I've argued before that the criminal justice system suffers from inappropriately basing its (often false) assumptions about human behavior on an economic price-theory approach, by which I mean the idea that punishment is a "price" for criminal behavior and so increasing punishment will lower "demand" for crime (i.e, reduce the crime rate). So I was interested when a couple of readers pointed me to this New York Times piece by economist Robert Frank arguing that the price theory model might work better if the likelihood of immediate punishment is increased.

LAW enforcement policy in the United States rests implicitly on the “rational actor” model of traditional economics, which holds that people take only those actions whose benefits exceed their costs.

This model says that crime will be deterred if the expected punishment is strong enough — a prediction that has not been borne out in practice. Although long sentences are now common and the incarceration rate is five times what it was during most of the 20th century, the crime rate is still two and a half times the average of 1950-62.

Mark Kleiman, a professor of public policy at the University of California, Los Angeles, says there is a better way. In a new book, “When Brute Force Fails,” he argues that instead of making punishments more severe, the authorities should increase the odds that lawbreakers will be apprehended and punished quickly.

First, a few background points:

Most crimes in the United States are committed by long-term repeat offenders, a majority of whom are eventually caught. One of every 100 adults in the United States is now behind bars; many are serving lengthy sentences. The crimes they committed clearly did not “pay” in any objective sense of the term.

Why, then, did they commit them? The short answer is that most criminals are not the dispassionate rational actors who populate standard economic models. They are more like impulsive children, blinded by the temptation of immediate reward and largely untroubled by the possibility of delayed or uncertain punishment.

The evidence suggests that when hardened criminals are reasonably sure that they will be caught and punished swiftly, even mild sanctions deter them. But not even the prospect of severe punishment is effective if offenders think they can get away with their crimes.

One way to make apprehension and punishment more likely is to spend substantially more money on law enforcement. In a time of chronic budget shortfalls, however, that won’t happen.

But Mr. Kleiman suggests that smarter enforcement strategies can make existing budgets go further. The important step, he says, is to view enforcement as a dynamic game in which strategically chosen deterrence policies become self-reinforcing. If offense rates fall enough, a tipping point is reached. And once that happens, even modest enforcement resources can hold offenders in check.

Consider violent crimes committed by drug gangs. In many cities, such gangs are too numerous for police to watch them all closely. Knowing that they are unlikely to be caught and punished, members can violate the law with impunity. In such situations, Mr. Kleiman argues, the police can gain considerable leverage just by publicizing an enforcement priority list.

It is an ingenious idea that borrows from game theory and the economics of signaling behavior....

Considerable evidence supports Mr. Kleiman’s emphasis on the efficacy of immediate sanctions. Experimenters have found, for example, that even long-term alcoholics become much less likely to drink when they are required to receive a mild electric shock before drinking. Many of these same people were not deterred by their drinking’s devastating, but delayed, consequences for their careers and marriages.

Several notable law enforcement successes, like a crackdown on gang homicide in Boston and strategic drug market disruptions in High Point, N.C., and Hempstead, N.Y., provide further testimony to the effectiveness of focused deterrence.

I wholeheartedly agree with Prof. Frank that the price-theory-based notion that "crime will be deterred if the expected punishment is strong enough " is a "prediction that has not been borne out in practice." But he and Professor Kleiman want to insist the price theory approach to crime could be valid, if only it could be tweaked. I'm not sure I understand why.

Certainly immediate and certain punishment is more effective, when it's feasible. But without dramatically expanding financial resources dedicated to criminal justice by many orders of magnitude, the suggestion cannot be scaled up to apply to the millions of people processed through the justice system.

And besides, there are MANY other ways besides the lack of price certainty in which economic theory fails to stand up to scrutiny in a criminal justice setting. From Wikipedia, here are the basic assumptions required for classical economic market theory to operate:
  1. People have rational preferences among outcomes that can be identified and associated with a value.
  2. Individuals maximize utility and firms maximize profits.
  3. People act independently on the basis of full and relevant information.
Let's start with the final assumption: Full and relevant information. Who ever has complete information about anything? With 2,324 separate felonies on the books in Texas (and that was before the Lege created 40 new crimes in last spring's legislative session), it's impossible even to know what's illegal, much less what will be the penalty or what the statistical chances are of being caught and facing punishment. Crimes are committed by people with limited and often skewed information.

As for the second assumption regarding maximizing utility or profits, that rarely applies to crimes committed on impulse - from the shoplifter stealing small items at the checkout counter to the cuckolded killer who murders his wife's lover upon finding them together in bed. For that matter, it would be easy to catalog a long litany of offenders whose actions were driven by essentially self-destructive motives, not a desire to maximize their own self interest.

That same observation ties in to the first assumption that people are "rational actors." But as Frank said, people are often more like "impulsive children, blinded by the temptation of immediate reward and largely untroubled by the possibility of delayed or uncertain punishment." Many people make irrational choices driven by addiction, alcoholism, mental illness, pride, anger, lust, or a variety of other non-rational motives. The prisons and jails are full of them.

There are other economic assumptions behind price theory, like a lack of coercion among parties entering contracts, that just don't play out in reality. And for the same reasons that price theory can fail in economic settings when these assumptions aren't true (all the way up to the recent Wall Street fiascoes), traditional punishment theories about deterrence frequently fall apart except under very particular and usually unachievable circumstances.


dirty harry said...

I would suggest the IRS is good at this type of enforcement. Although in reality, if you cheat or make a mistake on your taxes, it will be almost a year before they contact you. But by then, you will also owe interest and penalties on top of what you originally shorted them.

However, just look at the size of the agency itself, compared to the vast population they police. I'd say they do a pretty good job all things considered. Of course, you could say that all the IRS really does is keep honest people honest. If you're a real criminal at heart and don't care to live a normal comfortable life, you can escape the IRS. (Or, if you happen to be Obama's secretary of the treasury, Tim Geithner, or Charlie Rangel, you can also get kid glove treatment by the IRS.)

But, you can't deny the one thing the IRS depends upon for its enforcement. That one thing is fear. The IRS has the working public at large believing that if you don't pay your taxes in full, you will undoubtedly be caught, and the penalty will be swift and severe. Even though in reality, any experienced business owner will tell you this isn't necessarily the case.

PirateFriedman said...

This article uses the term homo economicus instead of "classical ecconomics". Maybe this will be a chance for you to learn: (from another wikipedia article)

Defenders of the Homo economicus model see many critics of the dominant school as using a straw-man technique. For example, it is common for critics to argue that real people do not have cost-less access to infinite information and an innate ability to instantly process it. However, in advanced-level theoretical economics, scholars have found ways of addressing these problems, modifying models enough to more realistically depict real-life decision-making. For example, models of individual behavior under bounded rationality and of people suffering from envy can be found in the literature. It is primarily when targeting the limiting assumptions made in constructing undergraduate models that the criticisms listed above are valid. These criticisms are especially valid to the extent that the professor asserts that the simplifying assumptions are true and/or uses them in a propagandistic way.

and by the way, the government caused the Wall Street problems... accept the truth if you can.

Charlie O said...

I doubt most people view harsh prison sentences as a deterrent to crime. That's a false argument for death penalty. If the death penalty was a deterrent, there'd be far less murders.

My experience from living in both California and Texas (Texas hands out severe prison sentences like candy, CA has severe three strikes law), is that most people (legislators included) view these laws and sentencing policies as revenge and punishment, not deterrent. I firmly believe this is true in Texas. Prison (and the death penalty) are punishment and revenge, plain and simple. Deterrence and economics are not in the equation.

Anonymous said...

Actually, the term punishment as used in the criminal justice context and the scientific definitions in psychology are not at all the same thing. The necessity for punishment to be immediate and contingent on the behavior is what makes it a viable means of altering human behavior. It does work if those conditions are satisfied, unless, of course, the organism can escape the situation. Criminal justice "punishment" is neither immediate nor contingent. Criminal justice "punishment" is probably more accurately referred to as "retribution". That is something different. It is a torture or injury imposed to satisfy the society's or the individual's sense of payback for wrongs.

Not only is criminal justice "punishment" not timely, and thus unlikely to be associated with the behavior which it ostensibly addresses, it is largely not contingent, either. That is, it is not strictly dependent on the behavior. Who is punished, and when, has less to do with what the person does than who they are, and largely depends on chance.

Also, punishment is a technique to alter or change behavior. Criminal justice "punishment" does not seek to change behavior, but to disable and injure the person being "punished". That is, the goal is not to keep a shoplifter from shoplifting again, but to make sure the shoplifter can never work at a lucrative job, can never obtain financial independence, and is destined to "suffer". Scientific punishment has nothing to do with "suffering". Its aim is behavior change. There is no chance for criminal justice "punishment" to change behavior, because behavior is not possible in the same context as it was before the punishment. That is, scientific punishment is a technique employed to change behavior, not a method designed primarily for the purpose of causing damage or injury to the organism, measured by the extent to which observers assume the organism is experiencing "suffering".

Thus, criminal justice "punishment" does not at all meet the definition of the term, whether it is actually a defensible practice or not. In order to evaluate criminal justice "punishment", we need to change the name so it is less confused with something entirely different.

PirateFriedman said...

from the NYT article:

In such situations, Mr. Kleiman argues, the police can gain considerable leverage just by publicizing an enforcement priority list.

What he's basically saying is that police sell a product called public safety, and they need to focus more on marketing. I might support this, if it diverted money from the police hiring.

Gritsforbreakfast said...

11:41, you write that "Criminal justice "punishment" does not seek to change behavior, but to disable and injure the person being "punished".

First, I don't think that's true; the purpose of punishment is many-layered. Second, to the extent you're correct, by that standard the system has been an utter failure. Incapacitation in most cases can only at best be a short-term solution because of limited resources.

Also, MANY theorists have relied on economic price theory and applied its precepts directly to criminal punishment. You can split hairs, but the "harsher punishment = less crime model" is based on assumptions that presume people behave as economic rational actors, and that's often just not true.

Anonymous said...

Gritsforbreakfast wrote, "You can split hairs, but the "harsher punishment = less crime model" is based on assumptions that presume people behave as economic rational actors, and that's often just not true."

Yes, it's not true, as so skillfully elucidated by Nobel Prize winner Daniel Kahneman! I just have trouble fitting what the criminal justice system does with the commonly understood meaning of punishment. The meaning tends toward something that represents an episode in the lifetime of a person which is insignificant in overall time, and its main result is the change in behavior. The idea is that there is a bad behavior, it results consistently in punishment, then the behavior stops. This sort of implies that that there is something left of the person after being punished -- a larger life to be lived of which the punishment episode was a small part that resulted in change. Criminal justice punishment isn't that way. It is a life consuming and radically life-altering event. It isn't an isolated episode which has the artifact of behavior change. It changes life's circumstances and possibilities.

In terms of prison sentences that consume significant percentages of people's available lifetimes, as well as the radically life-altering "class" system that arises from criminal records, the idea that this is simple "punishment" just doesn't seem to fit. Punishment is a brief interlude that results in change. Criminal justice punishment is life-consuming. Rather than being an event in life, it becomes one's life. It defines it.


PirateFriedman said...

Grits just doesn't understand the theory, so he's in no position to critique it.

Immediacy and certainty of punishment can make the cost of committing crimes too high, need proof.. well, serial killers don't commit their crimes in front of uniformed officers. There you go. So the ecconomic basis of punishment holds up.

I wish the author had mentioned gun ownership as a way to provide very immediate punishment for criminals.

Gritsforbreakfast said...

1:35, you make some good points.

Pirate, I'd wager I've spent more time in economics classrooms than you, so why not stick to explaining your own views instead of claiming you know mine.

My original post said, "Certainly immediate and certain punishment is more effective, when it's feasible," so you're just griping for argument's sake, not disputing anything I've said.

David C said...

I don't think you fully comprehend what Mark Kleiman is advocating because I don't think Prof. Frank explained it very well. The most important component is reducing the length of sentences, say by a factor of 10 to 100. Beyond a few days, jail cells are useless as a form of rehabiliation other than keeping the criminal away from others. This reduced expenditure on sentences will also reduce the number of appeals. Fewer people will want jury trials if they entail waiting for even longer periods in a jail cell before receiving the sentences. The savings in this area can be used to redirect money towards enforcement, and particularly towards parole. Since most criminals are re-offenders, having long parole sentences would allow society to focus on them. More parole officers would lower their caseloads. Mandatory immediate punishment would replace arbitrary punishment handed out only to the worst offenders for parole violations. By the way, Mark Kleiman has not just a book, but also a blog called the Reality-Based Community:

Gritsforbreakfast said...

DavidC, maybe I'll have to read Dr. Kleiman's book. Sounds like I might agree with some parts and question others.

I'm in agreement that sentences are needlessly long and with the idea of stronger community supervision as an alternative to incarceration, but I actually think that requires shorter probation/parole stints for non-high risk offenders, not longer ones. (Most reoffense happens in the first 6 months to 3 years.) Probation and parole caseloads are way too high and the two choices are hire more POs or reduce the number of supervised offenders. That's why IMO offenders need ways to earn their way off supervision through good behavior. The alternative is ever-longer terms where successful probationers become cash cows for local government.

Also, after the last few decades' experience with mandatory minimums, we know that when you hear talk about "mandatory immediate punishment," the devil is in the details. Given the present, awe-inspiring scale of the justice system, that tactic may simply not be tenable unless you scale back by orders of magnitude the number of offenders subjected to the punishments.

Bottom line: My point was that even though Kleiman and Frank agree that the orthodox view that economic rational actor theory predicts criminal behavior "has not borne out in practice," instead of examining the problem with fresh eyes they attempt to salvage the theory by propping up one of its most obviously flawed assumptions. But I think there are several other assumptions underlying price theory that also don't apply to the criminal justice arena. At most, economic price theory is an historically important metaphor, but it doesn't have much real-world predictive value.

PirateFriedman said...

Pirate, I'd wager I've spent more time in economics classrooms than you

Perhaps. But I know you are a really old guy. They were still giving Keynesianism intellectual respectability when my father got his MBA in 1983 at TCU, so its hard to say what nonsense you might have been exposed to.

Anonymous said...

To Anonymous 11:41 am :

You are on-the-money, it is both punishment and retribution, but it applies not just to prison but also to probation. The academics call it "delegated" punishment, in combination with other influences such as privitazation of government functions. That's why the community policing and neighborhood watches, having good ideas and reasons for existence - at least ideally, often digress into brown-shirt thuggery so many times as they are used for punishment purposed by the police. More importantly, I think that there is a profit motive rolled in as well that's not really talked about much.

Grits has argued before on this blog that incapacitation is not used with probation, that it is usually handled by prison sentences. But then I found another article where he states that prison isn't meant to be punishment ( implying that the punishment is above and beyond the prison sentence itself). Don't blame Grits for not understanding... most folks won't have a clue unless they have either been through the system, or on the other hand , if they participate in what I call the "punishment matrix". Some of the sadistic bastards call it "The Game".

Unfortunately, Anonymous 11:41 am has the keenest insight into how things work ( at least in Texas, and probably plenty of other places as well). Torture is the best word to describe what goes on for those in prison, since they have to be "punished", and punishment is not viewed, at least at the implementation level of the beauracracy, as the same as the prison term itself. Pretty sick, eh? Probation can be just as bad, since the same thing can be said of the poor (probably unemployed) slob on probation who is stalked day and night by probation surveillance teams and community policing groups. There is a reason for some of these stats that you see. Most offenses take place by repeat offenders? Maybe it's because they are hunted in a sense...stalked. If they can bump you from strike 1 to stike 2, cash rewards and bounties are more lucrative. Stike 1 means a "time out" period of employment, and maybe some community service, where undercover folk always bring up topics such as drinking and drugs, to see if they can get you to stumble. If you make it through this gauntlet, Stike 1 means you get tagged, and become cheap labor after this "time out" period. Such a deal for the small businesses, since you are now lifelong cheap labor. Stike 2 means that now you are running the SERIOUS gauntlet, and the bounty hunters are trying to set you up to stumble into strike 3 so that they can collect their reward. You are messed with both outside of your residence, as well as inside of it...relentlessly, and nonstop. None of the communities want a felon or strike 2 individual in their vicinity, and so they try to drive you out, no matter where you go. Maybe they turn all the AC vents on you at work and try to freeze you out. Maybe they turn your water off, or tap your electricity in your apartment and run your electric bill up several hundred dollars while dumping fumes down your chimney and paying your smoking neighbor to blow smoke in your vents, I don't know. The point is that it's usually dirty, indirect, and underhanded, but very very real. They know that you are defenseless to protect yourself against their constant harrassment...if you fight back, you go to jail. In the meantime, it's nearly impossible to prove, and nearly impossible to find a place to rent an apartment. If you already had a job and apartment before your incident with the law, expect a law-enforcement officer to visit both your place of employment and your apartment manager to let them know THEIR perspective of things...even if it is still before your trial has occured...and all in the name of "public saftey" (a rather elastic term that can be used to justify any draconian action that they wish). You won't find this in your academic stats and books, Grits, but it is true, nevertheless.

Gritsforbreakfast said...

5:41, maybe you can provide some links or something, I don't remember arguing any of what you attribute to me. And I don't think I've ever underplayed - certainly not when my writings are taken as a whole - the significant post-prison costs visited upon ex-offenders from a conviction.

BTW , sometime soon I may blog on a recent academic paper by Bernard Harcourt that touches on the mathematics behind the phenomenon you describe of targeting the "usual suspects" and how it creates unjust outcomes. Interesting stuff, the same guy also has a book-length treatment of the topic.

Miss Daisy said...

Blaise Pascal's aphorism is appropriate for this punishment theory - "Justice without force is powerless; force without justice is tyrannical"

Mark Kleiman said...

Grits, you mean that now that you have given your opinion that my research is wrong, you're considering actually reading my book? What an original idea!

Or you could read the paper in the Proceedings of the National Academy of Sciences (Kleiman and Kilmer, "The Dynamics of Deterrence") instead. I'm not theorizing for my health; the paper and the book set out to explain the actual success of actual programs: Cease-Fire, HOPE, and High Point. See also David Kennedy's book, Deterrence and Crime Prevention.

Anonymous said...

Anonymous 05:41 PM wrote, "That's why the community policing and neighborhood watches, having good ideas and reasons for existence - at least ideally, often digress into brown-shirt thuggery ..."

It sounds like you do know what's going on! Yes, I agree that talk of criminal justice "policy" is purely an academic exercise until extrajudicial enforcement is dismantled. It is largely the extrajudicial community system which controls outcomes in the way you describe, and the actual official policy is irrelevant. Unfortunately, breaking up the extralegal system requires a level of political awareness and organization that was snuffed out in the United States three decades ago, and the population is monitored closely to nip any signs of its return in the bud. I'm not holding my breath!

Gadfly said...

Mark, it's called behavioral economics.

Gritsforbreakfast said...

I didn't say your research was "wrong," Mark, and I'm actually quite aware of all the programs mentioned in your comment, a couple of which have been written up on Grits.

Amd no, I haven't read your book, but I've read your blog, and I've noticed you don't read every author's complete oeuvre before commenting on this or that column referencing their work.

Brian McGiverin said...

I found this conversation to be very interesting. I wouldn't mind seeing a similar conversation comparing the effectiveness of the typical system of punishment in deterring crime to the effectiveness of the attenuated punishments meted out by drug courts.