John Quiggin has been arguing it is. That's why he wants to add America's higher incarceration rate to its unemployment statistics when comparing us to Europe.
To my surprise, when I proposed this theory to Mark Kleiman a while back, he disagreed. Crime is only very weakly correlated with changes in the labor market. It spiked during the golden age of unskilled employment in the 1950s and 1960s, and then fell for no particular reason during the poor labor market of the early 1990s. Crime is a labor market outcome in the sense that people with poor impulse control gravitate towards a "job" that requires little in the way of gratification delay, not in the sense that people who end up in jail literally had no alternatives, or even no better alternatives. In the normal operation of the American economy, most people who want a job can find something. Given the low probability-weighted returns to crime, often even something better than sticking up 7-11s.
There are other problems with the theory. Even if crime were a labor market outcome, incarceration is a policy outcome, not a labor market outcome, because incarceration has increased even as crime has fallen. Furthermore, what correlation there is between crime and the economy is to property crimes--burglary, etc. Violent crime, which accounts for more than half of America's incarceration rate, and virtually all of the change in our incarceration rate since 1980, isn't clearly related to the economy. In theory, being laid off might make you more prone to bar fights or beating on your girlfriend. In practice, it doesn't seem to show up in the numbers.
Now, one could argue that high incarceration rates are supressing the unemployment figures. But America's employment-to-population ratio is still higher than Europe's, though a number of individual European countries do better than we do.
That is not to defend American incarceration policies, which are lunatic, as is the drug war which contributes to them. Mark Kleiman has some very good ideas on how we might lower those rates by using targeted intensive surveillance of those on probation and parole. I'd like to lower it even more by legalizing drugs and eliminating the black market profits that fund today's gangs. But a preference for fewer prisons doesn't require me to believe that someone who rapes a stranger is just a victim of a weak job market.
Copyright © 2009 by The Atlantic Monthly Group. All rights reserved.
Sigh. Not true. Here is the bet:
And here is John's interpretation of the numbers:
And here now is what John is saying, and what Megan is linking to:
Really, this is journalism and scholarship of the most rudimentary kind.
Yeah, I've always been surprised by this, but it seems from my research that economic downturns don't have a consistent effect on crime rates, violent or non-. It's casually assumed to be the case, because it makes a certain amount of intuitive sense, but it seems like it's not true.
I once read something that distinguished economically oriented crime, which is always about some kind of gain and may or may not involve violence, with random violence that has no economic purpose. The writer claimed that the US actually has lower rates or economically oriented crime than many other countries commonly cited as having "less crime."
But the US (and certain identifiable locations within the US in particular) has much higher levels of random violence than many other countries. The author attributed that to an "honor" culture in some parts of the US. This tends to lead people in those areas to resort to violence in response to offenses that other people (including other Americans) would shrug off or address in some other way.
But honor crimes aren't random, are they? If Mr. X claims that I enjoy taking the submissive role in a male homosexual act, and if I perceive this as an insult to my honor that demands a violent response, the violent crime I commit against Mr. X isn't random at all.
A random crime might result if, while driving by Mr. X, I shoot at him and miss, instead killing Ms. Y who was standing nearby. But that would be a random crime regardless of my motive: in other words, it would be random if I were trying to shoot Mr. X for breaching a contract to supply drugs (an economic crime), and it would be just as random if I were trying to shoot Mr. X for insulting me (an honor crime).
It's not random from inside the perpetrator's head, but the entire incident (including the original insult) is random from society's perspective. It's people spouting off for random psychological reasons, not stealing or selling contraband for reasons that might make economic sense even to someone who didn't share the same psychological profile.
The fact that many criminals come from groups with low job prospects does not mean that they become criminals because of low job prospects. Instead, it could be that whatever it is that caused them to have low job prospects also caused some of them to choose crime.
In other words, correlation does not imply causality. And, in fact, it sounds like the evidence refutes causality.
(Earnest Iconoclast): "...correlation does not imply causality."
If correlation does not imply causality, what does?
"Cause" may be tangled and complicated. I suggest an indirect connection between labor markets and crime. Political manipulation of labor markets by politically adept groups (e.g., unionized manufacturers' support for minimum wage increases, public employee organizations' support for expanded compulsory attendance at school) reduces options for members of the less-adept groups. In addition...
1) an abusive childhood is a fairly strong predictor of later abusive behavior,...
2) slavery is abusive,...
3) compulsory, unpaid labor is slavery,...
4) students labor, uncompensated, as window-dressing in a massive make-work program for dues-paying members of the NEA/AFT/AFSCME cartel.
(Scent of Violets): "You don’t have to buy Becker’s assumption that criminality is a “rational” choice, to agree that it is a choice and that choices reflect the attractiveness of the available options."
I dispute that the choice to commit violence reflects --current-- options. Perhaps the more important (that is, more predictive) variable which influences the choice to commit violent crime is not current employment options but past experience.
Compulsory schooling is abusive. You cannot eat a transcript. Training an artistically or mechanically inclined child for an academic career using a transcript as the incentive is like teaching a cat to swim using carrots at the reward.
The greater the mismatch between the culture of the school and the culture of the home, the fewer rewards schools offer students.
From: Hyman and Penroe, Journal of School Psychology.
"Several studies of maltreatment by teachers suggest that school children report traumatic symptoms that are similar whether the traumatic event was physical or verbal abuse (Hyman, et.al.,1988; Krugman & Krugman, 1984; Lambert, 1990). Extrapolation from these studies suggests that psychological maltreatment of school children, especially those who are poor, is fairly widespread in the United States...."
...
"While 1% to 2% might not seem to be a large percentage of a school-aged population, in a system like New York City, this would be about 10,000 children so traumatized by educators that they may suffer serious, and sometimes lifelong emotional problems (Hyman, 1990; Hyman, Zelikoff & Clarke, 1988). A good percentage of these students develop angry and aggressive responses as a result. Yet, emotional abuse and its relation to misbehavior in schools receives little pedagogical, psychological, or legal attention and is rarely mentioned in textbooks on school discipline (Pokalo & Hyman, 1993, Sarno, 1992)."
"As with corporal punishment, the frequency of emotional maltreatment in schools is too often a function of the socioeconomic status (SES) of the student population (Hyman, 1990)."
...
"Research on victims of violence at school suggests that repeated victimization has detrimental effects on a child's emotional and social development (Batsche & Knoff, 1995; Hoover, Oliver, & Thomson, 1993; Olweus, 1993). Victims exhibit higher levels of anxiety and depression, and lower self-esteem than non-victims (eg., Besag, 1989; Gilmartin, 1987; Greenbaum, 1987; Olweus, 1993). Karen Brockenbrough, Dewey G. Cornell, Ann B. Loper, "Aggressive Attitudes Among Victims of Violence at School", Education and the Treatment of Children, V. 25, #3, Aug., 2002.
"Results showed that the over-representation of Black males that has been cited consistently in the literature begins at the elementary school level and continues through high school. Black females also were suspended at a much higher rate than White or Hispanic females at all three school levels." Linda M. Raffaele Mendez, Howard M. Knoff, "Who Gets Suspended From School and Why: A Demographic Analysis", Education and the Treatment of Children V. 26, #1, Feb. 2003.
"The failure to provide education to poor urban children perpetuates a vicious cycle of poverty, dependence, criminality, and alienation that continues for the remainder of their lives. If society cannot end racial discrimination, at least it can arm minorities with the education to defend themselves from some of discrimination’s effects." Justice Clarence Thomas,
ZELMAN, SUPERINTENDENT OF PUBLIC INSTRUCTION OF OHIO, et al. v. SIMMONS-HARRIS et al., Concurring.
"Criminal violence emerges from social experience, most commonly brutal social experience visited upon vulnerable children, who suffer for our neglect of their welfare and return in vengeful wrath to plague us. If violence is a choice they make, and therefor their personal responsibility, as Athens demonstrates it is, our failure to protect them from having to confront such a choice is a choice we make, just as a disease epidemic would be implicitly our choice if we failed to provide vaccines and antibiotics. Such a choice-to tolerate the brutalization of children as we continue to do-is equally violent and equally evil, and we reap what we sow. ..." Richard Rhodes, Why they Kill: The Discoveries of a Maverick Criminologist.
Who gets the wretched schools in the US? Blacks and Hispanics. Who's overrepresented in US prisons? Blacks and Hispanics. Who gets the wretched schools in Hawaii? Hawaiians and Samoans. Who's overrepresented in Hawaiian prisons? Hawaiians and Samoans.
I got these charts form a statistician in the office of the Attorney General, State of Hawaii.
I read the phrase "If correlation does not imply causality, what does?". Then I read it again. Then I facepalmed.
Direct evidence of causation implies causation. This means that your original event (cause) must directly lead to another event (effect) via some identifiable mechanism. Simply saying, "Event A occurred before Event B, therefore A caused B," is not enough.
Who has the lowest ownership of luxury goods? Blacks and Hispanics. Who is over-represented in US prisons? Blacks and Hispanics. The obvious solution, therefore, is to buy every person of color a Lexus, a Brooks Brothers suit and a Rolex watch, thereby raising their self-esteem levels and allowing them to present themselves for high-end executive interviews without feeling inferior.
(Quanticle): "Direct evidence of causation implies causation".
A=>A. Right. 'Scuse me, but I thought we were discussing the real world, not tautological truth.
Here's Russell on Hume (from __A History of Western Philosophy__): "Hume's scepticism rests on his rejection of the principle of induction. The principle of induction, as applied to causation, says that, if A has been found very often accompanied or followed by B, and no instance is known of A not being accompanied or followed by B, then it is probable that on the next occasion on which A is observed it will be accompanied or followed by B. If the principle is to be adequate, a sufficient number of instances must make the probability not far short of certainty. If this principle, or any other form from which it can be deduced, is true, then the causal inferences which Hume rejects are valid, not indeed as giving certainty, but as giving a sufficient probability for practical purposes. If this principle is not true, every attempt to arrive at general scientific laws from particular observations is fallacious, and Hume's scepticism is inescapable for an empiricist."
"We must therefore ask ourselves: What sort of thing is it reasonable to believe without proof? I should reply: The facts of sense experience and the principles of mathematics and logic -- including the inductive logic employed in science." Bertrand Russell, __The Quotable Bertrand Russell__ (ed. Lee Eisler, Buffalo, NY: Prometheus, 1993), p. 253.
(Malcolm): "Who gets the wretched schools in the US? Blacks and Hispanics. Who's overrepresented in US prisons? Blacks and Hispanics. Who gets the wretched schools in Hawaii? Hawaiians and Samoans. Who's overrepresented in Hawaiian prisons? Hawaiians and Samoans."
(Quanticle): "Who has the lowest ownership of luxury goods? Blacks and Hispanics. Who is over-represented in US prisons? Blacks and Hispanics. The obvious..."
"Obvious"? You do not really think so, do you? Why the strawman? Do you deny that abuse generates hostile responses? How would you like to receive the treatment we routinely dole out to young adults in school? If you tried to treat 25 mature adults the way we routinely treat kids, you'd be lucky if all they did was punch you out.
"...solution, therefore, is to buy every person of color a Lexus, a Brooks Brothers suit and a Rolex watch, thereby raising their self-esteem levels and allowing them to present themselves for high-end executive interviews without feeling inferior."
I supplied real world evidence. What you got?
That proposition is worth a closer look. For example, suppose a man has both a poor instinct for delayed gratification and a lazy temperament toward work. He may decide to hold up an occasional liquor store instead of looking for legitimate work because the holdup usually procures $350 in five minutes.
In that case, his poor employment prospects (lazy workers prone to five-finger discounting of merchandise or supplies usually don't do too well), and his decision to engage in crime, would be statistically correllated but the correllation is due to a common third cause.
A great number of the incarcerated were employed prior to being arrested, it is just that their crimes were their jobs.
It seems odd to use incarceration rates as a proxy for crime rates, given that incarceration rates are largely dependant on political policy, posturing politicians and the lobbying of prison guard unions.
Additionally, the theoretical implications of a relationship between unemployment and crime are utterly different from those pertaining to a relationship between unemployment and incarceration. The former directly has to do with whether being unemployed makes one more likely to resort to criminal activity. The latter has to do with whether political incentives exist for governments to manage the labor market by removing certain people from it in times of economic stagnation.
It spiked during the golden age of unskilled employment in the 1950s and 1960s, and then fell for no particular reason during the poor labor market of the early 1990s.
Um. Isn't crime mostly correlated with age. I was taught in a sociology class in the 70s that most crimes were done by males from 15-25 years in age, which is also when car insurance rates are highest. The professor predicted that as our cohort got older, crime rates would fall. By the 1990s, the baby boom was moving to over 25. The gang wars for control over crack distribution had a distorting effect on violent crime rates for a while.
As a non-economist, could someone tell me whether "outcome" is strictly defined? That is, can it mean correlation or is it required to be direct causation? When Earnest I. argues:
The fact that many criminals come from groups with low job prospects does not mean that they become criminals because of low job prospects. Instead, it could be that whatever it is that caused them to have low job prospects also caused some of them to choose crime.
This distinction would matter a lot more if the relation between labor market and outcome is necessarily causal and one-way than if labor outcome it is taken in some more general sense. On a related note, I'm don't understand how something that has as much impact upon the labor market as incarceration rates could be seen as "political policy" rather than some mixture of both. (I.e., what does the political/economic distinction do here?)
So I don't even know what is correlating any more... we have crimes committed, crimes discovered, people incarcerated, people employed in legitimate jobs, people employed under the counter, and people "employed" in crime. So how do these all correlate to job prospects/education/family structure/etc...?
It could be that there is no correlation between a life of crime and poor job prospects but that people who have poor job prospects because of some kind of incompetence are also bad criminals and get caught while people who would have excellent job prospects are also excellent criminals and never get caught. I don't know what the statistics or demographics are on undetected criminals...
All else being equal, a lack of correlation does imply a lack of causality. But correlation could be an accident or caused by another factor that is causing the two things you are measuring. Once you establish a correlation, you need to establish causality by either changing on variable and showing that the other responds or by discovering the mechanism. In social sciences, though, the "mechanism" is often speculative and it's hard to change variables in a controlled fashion. That's why I'm an engineer...
The trouble comes up when people start trying to be social engineers.
If you take a group of middle-aged people, you will probably find that the shorter ones have more hair on the tops of their heads and rather less growing on their faces. Along come the social engineers to say that shortness is bad, and to try to fix it by shaving people's heads and pasting on fake beards.
Which is ridiculous, of course, because the people in question are shorter because they are female.
If you want to start tinkering with mechanisms, you really, really have to know what they are.
I don't actually believe that there is one mechanism that produces criminality. Sometimes it's economic and can be prevented by opening up legitimate opportunities. Sometimes it's just youthful stupidity, which societies typically address by keeping young men very, very busy until they are old enough to settle down. Sometimes it's individual psychology, sometimes it's about copying other people in the environment... which is why you typically need to do a lot of different things to keep crime low in a given society. Or, to put it another way, a functioning society has to get a lot of different things right, not just one or two.
YAY!!!!!! HUZZAH!!!! HUZZAH!!!! HUZZAH!!!! HUZZAH!!!!
I've finally - finally - managed to impart some small part of my wisdom on to my students.
Hey, I take what I can get.
I am not one of your students and I knew that before I ever heard of you. My parents don't even try to take credit or my (lack of) accomplishments.
As I recall, you at one point seemed to rather strenuously reject that idea. Shrug.
One thing I would like to see analyzed is the effect of felony convictions on repeat offenders. I know that many jobs require a background check these days, and that a felony conviction is going to show up rather strongly on any such background check. So, once someone has a felony conviction, their chances of gaining future employment are going to be adversely affected. Would this increase the probability that the person will reoffend?
The experiment design I had in mind would use the disparity in drug laws between different jurisdictions in the country. If one could select cohorts from the same socio-economic group in two different jurisdictions, where a crime is a felony in one jurisdiction but a misdemeanor in another, one could set up a natural experiment in analyzing whether felony convictions affect reoffense rates.
This is not an original thought (I'm not sure I'm capable of one, but if it happens, I'll cut and paste a flashing Drudge light), but when you say "incarceration has increased even as crime has fallen," it's kind of like saying "my intake of green vegetables has increased even as my health has improved." Doesn't it make sense that locking up those who are prone to crime (and what better predictor than the fact that they've committed crimes) would cause the crime rate to drop?