The Volokh Conspiracy

More Prisons, Less Crime?

As Orin points out in a recent post, one out of every hundred adults in now behind bars in country. The source for this factoid is the Pew Center on the States, which released its report "One in 100" to much fanfare. The Pew Center claims that we are not really getting anything in return for the moneys spent on prisons. But curiously, despite the claim that this expenditure is "failing to have a clear impact either on recidivism or overall crime," the study never attempts to assess the impact on overall crime.

As a quick way of assessing the return is interesting to compare the moneys spent on prisons over the last twenty years(collected in the Pew report) and crime rates over the same period of time. This chart plots one against the other:

[click to enlarge]

As can be seen, significant increases in spending on prisons has coincided with significant reductions in crime. Of course, proving causality would require a more sophisticated analysis. But it would be remarkable to think that the prison growth has had nothing to do with the fact that violent crime rates have reached their lowest point in recent years, according to the Dept. of Justice Bureau of Justice Statistics.

To be clear about my own views, we should be smart about who we incarcerate. And of course we should explore alternatives to prison. But all too often factoids like the Pew Center's are used to urge bans on further prison construction without considering the very tangible harms to crime victims that such an approach would produce.

eyesay:
Prison growth has had nothing to do with the fact that violent crime rates have reached their lowest point in recent years.
3.3.2008 1:13pm
Parker Smith (mail) (www):
How remarkable that you should think that...
3.3.2008 1:19pm
Mahan Atma (mail):
"Of course, proving causality would require a more sophisticated analysis."


This is exactly right; yet you appear to pay very little mind to this critical point.
3.3.2008 1:22pm
liberty (mail) (www):
1. Who is imprisoned (what crimes) is likely to have a big effect too.

2. Many more factors could come into play - causality could be very very difficult to show.

3. If we finally were to determine causality, this would not automatically indicate that we should expand (or not shrink) prisons. 100% incarceration would also mean lower crime-- actually...

4. Does the crime number include crimes within prisons?
3.3.2008 1:27pm
ejo:
correlation does not always equal causality. however, it doesn't hurt, particularly given that we know that criminals tend to commit multiple crimes before apprehension, prosecution and imprisonment. that doesn't need very much sophistication. what I think people mean by "sophisticated" analysis in this case is an analysis which will show no causal connection between incarceration and crime. good luck on that one, opponents of prisons.
3.3.2008 1:31pm
Richard Aubrey (mail):
Correlation does not prove causation.
It does, however, require explanation.

Hm. Scans wrong. Oh well, could still be correct.
3.3.2008 1:31pm
KevinQ (mail) (www):
Another issue is scale. From 1991 (the peak of the blue chart) there has been a drop in crime of roughly 36%. But spending over the same time has increased by 54%. (The numbers are rough, and based on guestimate readings of the chart.)

So, even if there is a causal connection, a spending increase of just more than 1/2 has resulted in a decrease in just over 1/3. Sounds like we're not really getting our money's worth.

K
3.3.2008 1:34pm
DanG:
Why did you post this? You can tell just from eyeballing the graph that the correlation is weak. Prison spending goes straight up, at somewhat varying rates of change. Crime goes up and then down, also at varying rates of change. Two inflections points in one variable, none in the other.

And yes, proving causation would require more sophisticated analysis. But so would proving correlation. You can't just put two lines on a graph that sort of move in opposite directions and say, "voila--correlation." Can you tell us what r is?
3.3.2008 1:36pm
frankcross (mail):
Kind of interesting to combine this with the thimerosol/autism post. Both represent junk science. It is certainly possible, indeed probable, that imprisonment reduces crime, but this sort of uncontrolled historical comparison doesn't really show much. This isn't even a good correlation
3.3.2008 1:41pm
b.:
Seriously: prison growth has had nothing to do with the fact that violent crime rates have reached their lowest point in recent years.

Or, at least, that's our presumption until we see evidence to the contrary.

Otherwise, why should we not always presume causation absent proof, or even argument?
3.3.2008 1:45pm
Thomas_Holsinger:
Les Weidman, then Stanislaus County Sheriff, told me about four years after enactment of California's three-strikes law that it had a remarkable effect on criminals. He said the first question asked by hard-core criminals during booking had become how many strikes they had, and that many said then that they planned on leaving the state once they got out.

But there is and was tremendous variance between California counties on use of the three-strikes law. Some, such as Los Angeles County, charge every possible strike regardless. Most apply it with discretion against repeat offenders, in particular against violent repeat offenders.

EVERY study I've read of the effects of sentencing practices says that getting repeat offenders off the street for long periods does more to reduce crime rates than any one other prosecution/sentencing policy.
3.3.2008 1:46pm
Sean O'Hara (mail) (www):
The 1% bit is what's getting all the attention, but I think how we stack up compared to other nations is more significant. We have 800,000 more people in prison than China, and 122 more per 100,000 than Russia. Sure, crime rates are down, but we're doing it by having higher incarcerations rates than thugocracies, and our crime rate is still high compared to other Western countries. I think we're doing it wrong.
3.3.2008 1:49pm
J. F. Thomas (mail):
You'll also notice that if there is a correlation it is certainly one of diminishing returns. Over the last four or five years, prison spending has continued to climb without any seeming impact on the crime rate (which has levelled out). Also, if I am not mistaken, the crime rate rose quite sharply in 2007.
3.3.2008 1:52pm
tdsj:
The graph starts in 1987, and thus reflects only one boom-bust cycle in the crime wave (the huge spike in the late 80s and early 90s, followed by the huge drop since).

I think if you went back further, you'd see less correlation. I haven't looked at the numbers for a few years, but as I remember, for most of the 20th century we had pretty stable (and relatively low) prison populations, and crime rose and fell from time to time (with a large non-cyclical increase beginning in the late 60s).

In the last 20 years, we've seen a massive non-cyclical increase in the prison population. If you just take one crime wave and map it on to that increase, of course you'll see a correlation.
3.3.2008 1:56pm
Armen (mail) (www):
FWIW, I just roughly approximated the data points and got a -0.84 r.
3.3.2008 2:01pm
Kazinski:
Sean,
Comparing US incarceration rates against the Chinese rate is misleading, China executes 8,000 people per year, 20 times the rate of the rest of the world combined. While adding that 8000 people per year to the incarceration total wouldn't explain the gap, maybe the deterent effect would.

But it may be that your point still holds: that China's penal system is more effective/efficient than the US system. Maybe we do need to substantially increase execution rates and see if it will help lower incarceration rates.
3.3.2008 2:03pm
Steven Joyce (mail):
Here's the best study I an aware of that goes beyond correlations to estimate the causal effect of prison population on crime rates:

Steve Levitt, "The Effect of Prison Population Size on Crime Rates: Evidence from Prison Overcrowding Litigation." The Quarterly Journal of Economics, 1996, 111(2), pp. 319-51, Link.

The paper uses overcrowding litigation as an (admittedly imperfect) instrument for prison population, and finds the the raw correlation actually understates the extent to which prisons reduce crime.
3.3.2008 2:06pm
Adam J:
Good idea Kazinski, lets follow China's example, we could sure learn alot from them...
3.3.2008 2:12pm
Thomas_Holsinger:
BTW,

California's enormous incarceration rates and staggering prison population are due as much or more to our being stuck with a true "gangster" union in the prison guards' association than to any other, more mundane, causes.

In addition to the usual public employee scam of selling campaign contributions to the legislature in exchange for hefty pay raises and construction of new prisons to provide still more campaign contribution opportunities, the prison guards' association has engaged in outright criminal conduct. One example was armed pursuit of a guard who reported misconduct INTO AN FBI OFFICE where the guy was then talking to FBI agents about it.
3.3.2008 2:14pm
deenk:
The causes of the drop in violent crime in the 1990s are subject to a much controversy. Conservatives tend to credit increased incarceration rates, while liberals tend to credit the favorable economic conditions. Neither factor explains most of the drop in crime.

Most developed countries experienced significant drops in crime in the 1990s. Canada, in particular, functions as an excellent control; Canada did not increase incarceration rates and did not undergo as much economic growth in the 1990s as the USA. Nevertheless, Canada experienced comparable drops in crime rates to those in the USA.

The major factors seem to be demographic. For an excellent discussion of this, see The Great American Crime Decline by Franklin Zimring.

Incarceration is incredibly expensive (~$35,000/prisoner/year in CA, plus the loss of the prisoner's contribution to the economy). Therefore, finding the best incarceration rates for crime reduction while minimizing spending should be a high public policy priority. Unfortunately, most political discourse on this subject is driven by anecdote and emotional reaction.
3.3.2008 2:17pm
One of the 99%:
Nobody here seems to be trying very hard to discuss possible causes of these trends, which is understandable given the complexity of the issue. As I see it, there are two basic arguments to be made regarding the relationship between imprisonment rates generally, and crime rates generally. Things get a lot more interesting when you start talking about specific types of crime and specific types of imprisonment.

First, one could argue that it's obvious that if you remove the people who commit crimes from the population, you not only prevent them from committing more crimes, but you scare all their criminal friends into obeying the law as well. This explanation is simple, straightforward, and intuitive. It's also backed by at least some anecdotal evidence. It's not, however, well supported (or refuted, necessarily) by the overall trends in our history, as was pointed out by tdsj.

The main opposing view, I think, is that putting people in prisons inculcates criminal cultures, and makes small time criminals into professional criminals. There are plenty of stories of people who went in as pot dealers, and came out as coke dealers. Imprisoning large parts of a community (like, for example, black urban males) breaks up family structures within that community, prevents members of that community from finding gainful employment (nobody wants to hire the guy who just got out of prison) and disenfranchises that community. Large imprisonment rates foster criminal cultures, destabilize communities, and alienate those communities from society at large. Hence, vicious cycle and more crime.

The second argument isn't quite as intuitive, and it's also not proven conclusively by the data we have, but I think it's at least as plausible as the first.

Also, if you really want to find a metric that provides a good correlation with the crime rate, try the unemployment rate. Those two lines will match up beautifully.
3.3.2008 2:19pm
darelf:
When I saw that graph I thought that the line going up was "number of violent video games on the market"..... Since it would be about the same graph. (the game "Doom", the first highly touted "super-violent" video game, came out in 1991)
3.3.2008 2:21pm
Frog Leg (mail):
Of course there is a correlation. If we were all in prison, there would be no crime.
3.3.2008 2:24pm
frankcross (mail):
Levitt's done probably the best analysis of this in the Journal of Economic Perspectives. He attributes the biggest reductions in crime to legalized abortion and higher rates of imprisonment. Still, I would say that our conclusions can be only tentative.
3.3.2008 2:47pm
Malvolio:
Dilbert: I oppose putting career criminals in jail for life. Theres no evidence that longer sentences reduce crime.

Dogbert: So your theory is that when career criminals are in jail, other people commit more crimes to keep the average up.
3.3.2008 2:50pm
Kazinski:
I think the debate misses the whole point. We don't put people in prison to meet some quota for the incarceration rate, we put people in prison because they have committed a crime and been sentenced to prison. Nor should people be let out of prison to conform to some quota. If someone does not want to be in prison, there is a guidebook called the penal code which lists all the proscribed actions that will land you in prison, 99.9999% of the people being incarcerated knew they were risking jail time when they committed their crime.

The reason prison spending has gone up is public intolerance of crime. While interest groups like public employees unions have lobbied for spending increases, there is a reservoir of public support that provides a conducive environment for the lobbying effort.
3.3.2008 2:55pm
Orielbean (mail):
Is it cheaper to stick people in the jails, or to let the crime (violent or otherwise) occur?

Like if the cost of the crime is 20k, and it costs 80k to "punish" the criminal (court fees, legal fees, prison fees), where should we draw the line?

Are there more cost-effective punishments like slave labor or wage garnishment that would get us more bang for the buck?

I would love to see a study on that!
3.3.2008 2:56pm
RPFN (mail):
1. I bet we can reduce crime even further if we execute all criminals. But I doubt many would find that desirable. We tend to believe that the punishment should fit the crime. More than half of those imprisoned are there for nonviolent crimes.

2. I suspect that there is a correlation between money spent on prisons and an aging population.
3.3.2008 2:58pm
J. F. Thomas (mail):
I think the debate misses the whole point. We don't put people in prison to meet some quota for the incarceration rate, we put people in prison because they have committed a crime and been sentenced to prison.

Well of course this is a very simplistic way of looking at it. When someone commits a crime and is caught, how much time they spend in jail (or even if they go to jail at all) depends on a myriad of factors beyond the simple facts of the case. One of the most important is how good a lawyer they can afford. But even if you take that variable away, just simple economic circumstances, where you live, where you are tried, and unfortunately still, the color of your skin, make huge differences in sentencing.
3.3.2008 3:30pm
Virginian:
I am a big supporter of "three strikes" laws. Maybe I am being scientifically illiterate, but I find it hard to believe that incarcerating career criminals for life would not reduce the crime rate.

Why in the world should we tolerate career criminals? Is it really too much to ask that you not commit three felonies?

I have absolutely no sympathy for someone who gets convicted of three felonies (in reality, probably having committed many more than three) and gets thrown in jail for life. We as a society are much better off.
3.3.2008 3:32pm
Thoughtful (mail):
Isn't the dramatic increase in prison population a direct consequence of the war on "drugs" (more precisely, on drug consumers, producers, middlemen)? I think the graph on prision population would be much more interesting, and possibly more enlightening, if it was converted into two curves, one for violent and one for victimless crimes.
3.3.2008 3:39pm
Buckland (mail):
I think it's interesting how many causes try to take credit for crime going down over the last few decades.

Bill Clinton claimed credit for his economic management and "hundred thousand cops on the beat" initiative that reduced crime. Giuliani claimed credit for changes in policing methods. Federal law enforcement points out that the crime rate fell when the country came out with serious penalties for crack cocaine. Economist Steven Levitt of Freakonomics fame credits abortion becoming legal roughly 20 years before the rate started falling, weeding out proportionally more kids from bad homes.

That's just the few claims I can recall in a few minutes. I'm sure there are others out there. It's interesting how many causes claim credit for when a good trend occurs but are nowhere to be found when it reverses.

What's the old saying -- Success has a thousand fathers, failure is an orphan.
3.3.2008 3:39pm
Virginian:

Isn't the dramatic increase in prison population a direct consequence of the war on "drugs" (more precisely, on drug consumers, producers, middlemen)? I think the graph on prision population would be much more interesting, and possibly more enlightening, if it was converted into two curves, one for violent and one for victimless crimes.


While drug use may be a victimless crime, are drug dealing and production really victimless crimes?
3.3.2008 3:44pm
Temp Guest (mail):
Oreielbean:

The studies that you would love to see done have been done by, e.g. Zedlewski and Kleiman and Cavanagh. On average it costs an order of magnitude less to imprison an offender than the crime that offender commits would cost society if he were not imprisoned. On the margin (the worst offenders) the benefits of imprisonment are orders of magnitude greater than the costs.
3.3.2008 3:46pm
Temp Guest (mail):
Regarding the Clinton COPPS program. About five years ago, as aprt of a consulting contract, I did an unsophisticated study of trends in police hiring before and during COPPS and found that it had essentially no impact at all on police hiring. As might be expected, police departments pocketed the money and used basic accounting tricks to claim it was used to create new police positions that -- the data suggest -- were already planned.
3.3.2008 3:51pm
ejo:
drug dealing and production are only victimless to the extent that the drug users are irrelevant in your moral equation. in addition, it discounts the people robbed/burgled/home invaded etc by those who need money for drugs as well as the folks who get smashed in the face by some junkie not being a rational actor.

factor in those costs-I bet it's scary.
3.3.2008 3:52pm
Malvolio:
Is it cheaper to stick people in the jails, or to let the crime (violent or otherwise) occur?
I don't think there is any question that your average working criminal can do in damage many times what it would cost to keep him incarcerated.

Like if the cost of the crime is 20k, and it costs 80k to "punish" the criminal (court fees, legal fees, prison fees), where should we draw the line?
You are doing the math wrong. The question is not what is the cost to society of the crime he actually committed, but what would the cost be of the crimes he would commit if not either deterred or detained.

None of these considerations, of course, apply to people whose "crimes" generate value for society: hookers, drug dealers, loan sharks, immigrant-traffickers, and bookies. (Yes, I realize you might not value the services of these entrepreneurs, but their customers clearly do.)
3.3.2008 3:55pm
Kazinski:
When you try to calculate the cost of "the" crime, you need to consider a few other factors, most criminals commit more crimes than they are caught and charged for. The cost to society of deterring crime; go into some high crime neighborhoods and try to count up the costs of burglar bars, extra deadbolts, alarm systems etc. and the cost of crime goes up by much more than the lost goods, or lost work of a crime victim. Not to mention the lost hours of playtime for children not allowed to play outside because of fear of crime.

Just last week I was at a gas station and saw the aftermath of an assault. An older Asian man was walking out of the store with a bottle of coke, a young punk walked up to him punched him in the nose and ran off with the coke, leaving the man with a bleeding and possibly broken nose. Whats the cost of that crime, a coke and a doctors bill? Its got me considering a CCW permit, and of course a snub-nose .38 to accessorize the permit. What's that going to cost?
3.3.2008 5:07pm
Mark Jones:
While drug use may be a victimless crime, are drug dealing and production really victimless crimes?

That depends on the circumstances, I suppose. Is quietly growing marijuana (or poppies, for that matter) inherently violent? Not to my knowledge, though of course the stories of lethal mantraps around (or lethal violence committed upon visitors to) marijuana farms are because growing the plants is a crime so the owners can't protect them any other way.

Ditto for dealing. Is selling marijuana or cocaine to someone who shows up, cash in hand, eager to buy an inherently violent act? I don't think so. Is selling booze to an alcoholic a crime? Should it be?
3.3.2008 5:12pm
Public_Defender (mail):
The graph seems to correlate to the number of people reaching young adulthood, when they are most likely to commit crime. Compare Professor Cassell's chart to this birthrate chart. The drop in crime almost exactly parallels the drop in 20 somethings.

Plus, the chart appears to be all corrections spending, not just prison spending. Corrections spending includes all sorts of non-prison sanctions, so the graph really doesn't compare crime rates to incarceration rates.
3.3.2008 5:32pm
callao (mail):
Minimum mandatory sentences are outlandish. Judges rightly recoil from sentencing some "mule" with little intelligence and no prior record to ten or fifteen years because they've been caught carrying the threshold quantity for someone else.

Also, why sentence white collar offenders with no prior record to lengthy prison sentences? Better to heavily fine people like Conrad Black or Martha Stewart than put them in jail.

Prison should be mostly for those who commit crimes against the person.
3.3.2008 5:38pm
J. F. Thomas (mail):
Also, why sentence white collar offenders with no prior record to lengthy prison sentences? Better to heavily fine people like Conrad Black or Martha Stewart than put them in jail.

Why should white collar offenders be excused from jail. Martha Stewart stole from the persons who she sold that stock to just as surely as a shoplifter steals a coat off a rack.

As for Black, he probably stole tens, if not hundreds of millions of dollars. If someone drilled into a vault of a cusino (a la Oceans Eleven) and stole $10 million and was caught for the first time, would you say they didn't deserve jail time?

Cusino (spelled properly) is blacklisted--who'd a thunk!
3.3.2008 5:48pm
Habeas Clerk:
Minor point of correction:

"Martha Stewart stole from the persons who she sold that stock to just as surely as a shoplifter steals a coat off a rack."

IIRC, Martha Stewart was convicted and sentenced for lying to federal investigators and obstructing justice. Though I agree with your general point.
3.3.2008 6:22pm
Elliot Reed (mail):
Also, the difference between the deterrent effect of jail time and the deterrent effect of fines is probably much larger for white collar criminals. The average person who would be tempted to commit a white collar crime is vastly better off and has a much stronger contact network than the average blue collar criminal. Your better-off white collar criminal can pay a very large fine without it affecting their standard of living at all and while it will probably (but not certainly cost them their job, I don't think committing a white collar felony makes you permanently unemployable. Going to prison, where you have to wear a uniform, shit in plain sight, eat in a cafeteria where they at won't even cook your steak to order and don't have so much as a Cabernet, and take orders from people you wouldn't hire as a janitor: that's different.
3.3.2008 6:35pm
Richard Gould-Saltman (mail):
My casual reading of Ed Tufte and, decades ago, of "How to Lie With Statistics" led me, before even seeing the units and placement of axises (axes?) on the chart, to ask "What's down there to the left of the graph?" My gut sense is that the red line shows a steady increase across much of the 20th Century, while that blue line bounces up and down in no particular relation to the red line, over time.
3.3.2008 6:36pm
juswonderingby:
The reactions to this post are so telling. Sure, correlation doesn't prove causation - but heck, everything is correlation (after all, smoking is *only* correlated with lung cancer...)

And for the person asking for the R2 - yeah, that would be interesting, but just looking at the graph, you can assume it would be large (now, we're only looking at 2 variables here, so perhaps once more were introduced into the mix, it would be different)

I agree that we need to be smart about how we spend our CJ dollars and that there's something sad about having so many people incarcerated. But, this post does seem to be on to something... (I expect Paul Cassell to be crucified for this post. Too bad).
3.3.2008 7:04pm
juswonderingby:
PS: Judge, can you upload a better copy of that graph please?
3.3.2008 7:04pm
RPFN (mail):
How did Martha steal from the buyer of her stock? For all we know, the buyer could have flipped it immediately for a profit. IMCL did close up that day.
So considering the fact that the buyer could have made a profit thanks to Martha, how did Martha steal from the buyer?
3.3.2008 11:37pm
Barry P. (mail):
I read a paper that posited that the total rate of institutionalization (in prison or mental asylum) was about the same in 1975 as in 2005, but the numbers had gone from 90:10 ssylum:prison in '75 to almost exactly the reverse in '05.

I can't see how warehousing the mentally ill in prisons really helps them or society.
3.4.2008 12:06am
Clayton E. Cramer (mail) (www):

Good idea Kazinski, lets follow China's example, we could sure learn alot from them...
That was not his proposal. He was just pointing out a DRAMATIC difference between China and the U.S. that might explain the relatively small number of Chinese in prison compared to the U.S.


I read a paper that posited that the total rate of institutionalization (in prison or mental asylum) was about the same in 1975 as in 2005, but the numbers had gone from 90:10 ssylum:prison in '75 to almost exactly the reverse in '05.

I think you may be referring to Harcourt's work, but I don't think it was quite dramatic of a change.


I can't see how warehousing the mentally ill in prisons really helps them or society.
It doesn't help the mentally ill. To the extent that it reduces violent crimes committed by the mentally ill, it helps society. It is a very clumsy and crude "solution," however. The decision to deinstitutionalize the mentally ill meant that people who in 1960 might have been institutionalized before they got around to murder, rape, kidnapping, or mayhem, now have to pretty near kill someone before they get any attention.

Los Angeles County Jail has effectively the largest mental hospital in the country now because of the decision to largely destroy the public mental hospital system, starting with the Community Mental Health Centers Act of 1963.
3.4.2008 12:42am
Dave N (mail):
I am a career prosecutor (and have handled death penalty cases both in my state Supreme Court and the 9th Circuit, so I am no soft-on-crime liberal).

That said, I agree with Public Defender. We are going from the "baby boom" to the "baby bust"--and I believe there is a strong corrolation between an aging population and a drop in crime.

I remember my very wise Criminal Procedure professor (who moonlighted as a Federal Magistrate Judge) saying, "The greatest indicator of recidivism is age."

What he meant, and which I have observed over the years to be true, is that if you walk into an average state courtroom on arraignment day, most of the defendants are under 30. If you see a 40-something charged with burglary or robbery or drug offenses, you can almost guarantee that person is in not in court on their first offense.

By contrast, if you see someone in their 30s or 40s in court on their first offense, it will likely be alcohol related (DUI or related crime); a sex crime involving a child (or related crime); a crime of violence against a family member; or a "paper" crime (a la Martha Stewart).

All that said, I am strongly opposed to mandatory three-strikes laws. All they do is transfer long-term medical costs to the Department of Corrections and away from Medicaid or other sources. The unintended consequence of California's three-strikes law is that we will have hospitals that are officially "prisons" with an ever increasing prison budget dedicated to meeting geriatric prisoners' medical needs.

Don't get me wrong. There are people who NEVER deserve to breathe free air for even a second--I don't care how senile or feeble that person might be. But as prison beds need to be rationed, I would rather reserve them for people who need to be removed from society NOW, not those who were moronically stupid before outgrowing the need to commit crime.
3.4.2008 1:47am
Dave N (mail):
I re-read my post and forgot to note that there is a sharp drop-off between 20-somethings commiting crimes and 30-somethings committing crimes. It is as if a lightbulb goes off in many criminals' heads around the age of 30 and they think:

"This is really stupid. I don't like having people watch me take showers and tell me what to do 24/7. Prison and jail sucks."

Of course, there are those who do not outgrow their need to commit crimes--and I have nothing against protecting society from them.
3.4.2008 1:53am
Elmer:
Habits are often persistent, which is one reason stores put items on sale. We are more likely to try something if it costs us little, and there's a fair chance we'll pay full price later for the same thing. In the US we have a similar system for crime. There's a sale that lasts until you turn 18. Later, first offenses are discounted relative to later ones. My guess is that authoritarian countries mirror that discount structure to some extent, but that the slope of the penalty vs. offences curve is much steeper than ours. Facts to replace my guesses would be welcomed.

Another, probably greater factor is the incentive structure outside of the criminal justice system. The Amish have low rates of incarceration and abortion, and have many children. From some of the correlations previously cited, that should make it dangerous to walk through Amish country at night, yet I don't think that is the case. Changes in that incentive structure may well be the best correlate to changes in crime rate, though terribly hard to measure.
3.4.2008 3:04am
UWV (mail):
I would like to see a similar chart showing that as global temperatures rise the crime rates fall. Same sort of logic.
3.4.2008 5:18am
Public_Defender (mail):
Career prosecutor Dave N wrote:


I re-read my post and forgot to note that there is a sharp drop-off between 20-somethings committing crimes and 30-somethings committing crimes. It is as if a light bulb goes off in many criminals' heads around the age of 30 and they think:

"This is really stupid. I don't like having people watch me take showers and tell me what to do 24/7. Prison and jail sucks."

Of course, there are those who do not outgrow their need to commit crimes--and I have nothing against protecting society from them.


Most people would be surprised how often line prosecutors and line defense attorneys agree about who needs to be hammered, who needs a few years in prison, and who needs a second (or even third or fourth) chance.

Keeping every single defendant in prison for life for committing three felonies before age 30 makes no sense. For the vast majority of these guys, we are wasting money by incarcerating them past their 35th birthday.

A three-strikes law might make sense if it gave the judge a choice of the standard penalty or incarceration until the 35th birthday, whichever is later, and as long as it had an exception for the treatable mentally ill.
3.4.2008 7:04am
ruralcounsel (mail) (www):
Here's another factor ...

The criminal laws are not a static system. I suspect if we could somehow create a metric for the number and types of behaviors that are criminal, we'd see that metric rising as we become more "civilized" and society decides it needs to try and control more peoples' behaviors. Thus, over time, it becomes more likely that some behavior is criminal that previously wasn't. Drug laws strike me as a prime example.

And I agree that there shouldn't really be a distinction between mental health facilities and prisons/jails, for the purposes of this causation discussion. They tend to fill the same purpose for slightly different classes of individuals. While it may make a large difference to the individual which one they are held in, as far as society goes, both remove people who functioned in some unacceptable manner from our midst.

Arguably a large percentage of the people passing through our criminal justice system are folks who are self-medicating for some mental or behavioral problem, and the self-medication got out of hand or landed them within the grasp of the police somehow. That's one reason why recidivism is so high ... addiction isn't something that responds well or quickly to cost/benefit arguments.

Perhaps another reason why recidivism drops off at about age 30 is that by that time, people tend to get better control of their addictions (if they've survived them that long). They learn how to "handle" their drugs; know what they can take, how much, what kind of employment and relationships they can expect to have if they use, where to crash.

Presumably, people fit some sort of Gaussian curve of behavior, and somewhere 2 or 3 standard deviations below average there is a point at which behavior becomes "criminal". That line can move some (my first point), and the Gaussian curve is also a dynamic ... as is the population. Individuals may shift their position within the curve, either instinctively or rationally.

Logically, we'd expect both the potential gain from criminal behavior, severity of the penalties, and the chance of getting caught to be factors for the rational changes.

Economics can shift people into the criminal zone ... even by a rational process. Drug dealers may conclude that the chances of getting caught are low compared to the chances of making a lot of money.

But people are not always rational. Particularly the mentally ill or folks with behavioral problems (maybe the one is a subset of the other, I don't know). So some portion of our criminal population doesn't likely respond much to penalties and probabilities. As our population grows, so will the absolute number of those people.

And, as we treat prisoners with more civilized approaches, we can expect the per capita cost of prisons to grow rapidly. Yes, kind of like college costs!
3.4.2008 10:07am
David M (www):
The Thunder Run has linked to this post in the - Web Reconnaissance for 03/04/2008 A short recon of what’s out there that might draw your attention, updated throughout the day...so check back often.
3.4.2008 1:50pm
IndomitableSploog (mail):
There's a wonderful book called "How to Lie with Statistics" and this graph should be in the appendix of the next edition. If you want to understand the relationship between the two variables, plot correlation, ok? It looks to me like the mean instantaneous correlation between these two terms is as close to zero as can be.
3.4.2008 2:08pm
Clayton E. Cramer (mail) (www):

And I agree that there shouldn't really be a distinction between mental health facilities and prisons/jails, for the purposes of this causation discussion. They tend to fill the same purpose for slightly different classes of individuals. While it may make a large difference to the individual which one they are held in, as far as society goes, both remove people who functioned in some unacceptable manner from our midst.
The distinction between the two is a matter of choice. We assume that a person who is mentally ill does not understand what they did was wrong, or did not know the difference between right and wrong. I don't think that either system needs to be unnecessarily brutal, but a mental hospital is not about punishment, but about treatment.

Another difference: at one time, people were committed because they were mentally ill, and might not have even committed a serious crime. They were hospitalized because they needed treatment--and while there was some reason to suspect that some of them might become a criminal matter on the outside, this was by no means certain. This was the medical model of commitment, and from what I can find, it worked for the benefit of both the society, and many of the mentally ill. At least mentally ill inmates of a hospital weren't freezing to death, dying of pneumonia, or begging for money to buy alcohol.

Today we operate (in most states) on a public safety model. A mentally ill person who commits a very serious crime will be committed or imprisoned (and most likely the latter). A mentally ill person who starves himself to death while family, friends, and police look on helplessly, as in Oregon--well, that's THEIR CHOICE, to hear some ideologues tell the story.
3.4.2008 2:12pm
Clayton E. Cramer (mail) (www):

Keeping every single defendant in prison for life for committing three felonies before age 30 makes no sense. For the vast majority of these guys, we are wasting money by incarcerating them past their 35th birthday.

A three-strikes law might make sense if it gave the judge a choice of the standard penalty or incarceration until the 35th birthday, whichever is later, and as long as it had an exception for the treatable mentally ill.
I agree with you. Unfortunately, California adopted Three Strikes because people like Richard Alan Davis demonstrated that judges didn't show much wisdom in using their discretion. Sometimes the only way to get the attention of the elites in our society is to whack them hard with a big wooden club.

As for "treatable mentally ill": in most states, you can't treat a mentally ill person against their will, with limited exceptions for those who are a imminent danger to themselves or others if not treated. And you can't hold a mentally ill person against their will if they aren't being treated, except as punishment for a crime. See Donaldson v. O'Connor (1975). Some state legislatures, with the assistance of a few courts, have managed to carve out some exceptions here and there, but the current system doesn't work. It just imprisons mentally ill offenders--usually after they have become at least a local headline.
3.4.2008 2:19pm
Clayton E. Cramer (mail) (www):

I would like to see a similar chart showing that as global temperatures rise the crime rates fall. Same sort of logic.
Most unlikely. Temperature increases are associated with increases in violent crimes. The causality relationship to murder isn't too hard to figure out; people get short-tempered when they are hot and miserable. Rape goes up as well--perhaps because of the increase in open windows means an increase in opportunities.
3.4.2008 2:25pm