A New Analysis of Incarceration and Inequality

by Kieran Healy on July 17, 2004

I’ve written about the intersection of “incarceration”:http://www.kieranhealy.org/blog/archives/000087.html, “race”:http://www.kieranhealy.org/blog/archives/000096.html and “the labor market”:https://www.crookedtimber.org/archives/000386.html several times in the past. In the United States, the remarkable expansion of the prison system over the past thirty years, despite generally falling crime rates, has had far-reaching effects on large segments of the population, but especially amongst unskilled black men. A striking way to characterize the depth of this change is to make a comparison to rates of participation in some other institution — say, for instance, that more black men have been to jail than are in college. But, as a lobby group found out last year, these comparisions are quite tricky to make properly, because the populations are different (all black men vs college-age black men, for instance).

But one of the many good reasons we have sociologists and demographers is to work out those numbers properly. A “new paper”:http://www.asanet.org/pubs/ASRv69n2p.pdf [pdf] by Becky Pettit and Bruce Western[1] does a terrific job of estimating how the effects of mass incarceration are distributed across the population. They estimate the risk of imprisonment for black and white men of different levels of education.[2] The paper is worth reading in its entirety, both to see how the findings might be understood and to understand how one goes about estimating these numbers in the first place — it’s not at all trivial to calculate them. Two core findings — bearing in mind these are the best available estimates — are remarkable:

* Among black men born between 1965 and 1969, 30.2 percent of those who didn’t attend college had gone to prison by 1999. A startling *58.9 percent of black high school dropouts born from 1965 through 1969 had served time* in state or federal prison by their early 30s.

* “Imprisonment now rivals or overshadows the frequency of military service and college graduation for recent cohorts of African American men. *For black men in their mid-thirties at the end of the 1990s, prison records were nearly twice as common as bachelor’s degrees.*” In the same cohort, “imprisonment was more than twice as common as military service.”

Interestingly, racial disparity as such has not grown in sentencing: the rates and risks of imprisonment are 6 to 8 times higher for young black men compared to young whites in both the ’45-’49 and ’65-’69 cohorts. But class inequality has increased. So while lifetime risk of imprisonment nearly doubled between 1979 and 1999, “nearly all of this increased risk was experienced by those with just a high school education.” Incarceration is now the typical life-event for young, poorly-educated black men.

fn1. Full disclosure: Becky’s a friend of mine and Bruce was one of my Ph.D advisers.

fn2. To forestall any misinterpretation, note that “risk” is a technical term here meaning roughly “the probability of being observed as ‘incarcerated’ during the period under study.”



Dr. Weevil 07.17.04 at 4:15 am

You mention “the remarkable expansion of the prison system over the past thirty years, despite generally falling crime rates”. Isn’t there a huge problem in the word “despite”? After all, a conservative might argue that you have it backwards and that we have had “generally falling crime rates” BECAUSE OF “the remarkable expansion of the prison system over the past thirty years”. Even if prison fails to deter future crime, it certainly cuts down on the opportunities to commit most types of crime as long as the criminal is in jail, however he behaves once he’s out.


eudoxis 07.17.04 at 4:51 am

Those are astounding numbers and they portend devastating effects for black families and communities in the near future. I wonder if compulsory education can be combined with incarceration. First time incarcerations appear to be slowing down for black men. This is hopeful news. Recividism continues to rise. The proclivity toward massive incarceration is a travesty.


Kieran Healy 07.17.04 at 5:42 am

we have had “generally falling crime rates” BECAUSE OF “the remarkable expansion of the prison system over the past thirty years”.

The link between crime rates and incarceration rates is by no means clear-cut, because a whole bunch of things intervene, such as shifts in what counts as serious crime worthy of jail time, substantial differentials in the enforcement of laws on the books, and — probably most important — changes in sentencing policy. So beginnging with the assumption that there’s a direct link between crime rates and imprisonment rates begs some of the central questions at issue. But in any event, in this paper, Pettit and Western aren’t making an argument about the causes of increased incarceration, they’re giving us information its radically disproportionate effects on particular groups. As they say themselves:

bq. The high imprisonment risk of black noncollege men is an intrinsically important social fact about the distinctive life course of the socioeconomically disadvantaged. Although the mass imprisonment of low-education black men may result from the disparate impact of criminal justice policy, a rigorous test demands a similar study of patterns of criminal offending. Increased imprisonment risks among low-education men may be due to increased involvement in crime. If patterns of offending follow economic trends, declining wages among non-college men over the last 20 years may underlie the growing risk of imprisonment. Researchers have examined the consequences of race differences in offending for official crime and imprisonment, but relatively little is known about educational differences in offending within race groups. To determine whether the shifting risks are due to policy or changing patterns of crime, we thus need to develop estimates of crime rates for different race-education groups. … More strikingly than patterns of military enlistment, marriage, or college graduation, prison time differentiates the young adulthood of black men from the life course of most others. Convict status inheres now, not in individual offenders, but in entire demographic categories. In this context, the experience of imprisonment in the United States emerges as a key social division marking a new pattern in the lives of recent birth cohorts of black men.


Lance Boyle 07.17.04 at 10:46 am

The ache behind the numbers is hard to talk about, much less quantify. It’s kind of taboo to even suggest, but I am suggesting that there’s more to the profile of imprisoned black men than lack of education. The education is offered by the system. The system is racist. The system is less racist than it once was, but it is still a racist system.
And something more. There’s a defiance that’s being weeded out. Education has a lot to do with submission. A lot more than most educated people are comfortable admitting.
For a lot of us the educational experience was more about learning to obey authority than it was about gaining knowledge of the world. Mixed in among the uneducable and the uneducated are bright young black men who rejected the domestication education seemed, and took the consequences.


Brett Bellmore 07.17.04 at 7:24 pm

“Because” may or may not be a stretch, but “despite” certainly carries some baggage you didn’t even begin to justify.

But I’ll agree that these statistics do say something really ugly… about the choices young black men are making. I scarcely think the answer is to let young black criminals remain on the street. I don’t know what the answer is, or if there even IS one that’s acceptable in a country where it’s hands off of popular culture.


Sebastian Holsclaw 07.17.04 at 7:44 pm

Why do we call it ‘risk of incarceration’ as if it were like a risk of inadvertantly contracting tuberculosis?

The terminology might be correct when talking about innocent people in jail, and there might even be an interesting racial factor to consider in that kind of statistic. It would even be worth talking about when comparing the chances of going to jail when convicted of a crime (e.g. the risk of incarceration for an African-American convicted of petty theft is 1.3 times greater than that of an Asian-American convicted of petty theft).

Treating it like a disease risk seems highly misleading. A more useful approach (and an approach the left could take without betraying its ideals) would be to ask: Why are so many African American males dropping out of high school? Why are so many of those choosing to commit crimes? What can we do to keep them in high school and/or keep them from turning to crime?


djw 07.17.04 at 10:07 pm

Sebastian, one good reason to use this approach is that it allows the investigators to remain agnostic (for the purposes of a limited study) on all the issues you raise. It’s fun for journalists, bloggers, op-eds, politicians, etc. to get all holistic when they approach this stuff, but for researchers trying to work with variables, it’s really hard. It helps to limit the scope of what you address in one project to see what you find. And I don’t see any particular reason to assume the methodology is about remaining true to left principles. Hard as this may be to believe, sociologists and social scientists of all political stripes make these kind of limiting choices all the time.

(I wouldn’t claim that the political views of the investigators doesn’t effect the choice of topic under study, but I wouldn’t assert that they do, either.)


Sebastian Holsclaw 07.17.04 at 11:45 pm

I’m saying that calling it ‘risk of incarceration’ is a political choice if you try to analyze it as if human beings had little or no choice in the matter.


djw 07.18.04 at 12:03 am

But analyzing impacts regardless of the how we might or might not approach the role of individual decisions in individual cases is not the same thing as “analyzing it as if human beings had little or no choice in the matter.” The question why has the rate of incarceration gone up? is a different research question altogether.


h. e. baber 07.18.04 at 1:42 am

I’d bet some of the discrepancy in who gets prison time and how much then get comes about because whatever they did or didn’t do young, lower-class black males look like the sort of people who should be in jail since–after all most of that sort have done prison time. Similarly, Martha Stewart, whatever she did or didn’t do looks like the sort of person who shouldn’t be in jail. We like things to look right.

Students like professors to look professorial–course evaluations reflect that–and in hiring and tenuring we look for “institutional fit.”


Chuck Culhane 07.18.04 at 6:07 am

The primary (but not the only) reason for the ten-fold increase in prison and jail populations in the last forty years (racism assumed) is the so-called “war on drugs.” In the current issue of The Black Commentator there’s an article (second of a four-part series) which makes the link between the “war on crime” and the “war on drugs” and the disparate and harsher treatments accorded to Blacks and Hispanics.

The previous comments that dropping crime rates are consistent with rising imprisonment rates is simplistic if not disingenuous. The logic of that fails when the phenomena occurs year after year. The fact is that draconian drug laws, mandatory minimums, enhanced sentencing schemes for second offenders, three-strikes laws, increased technical parole violations and a decrease in parole releases, among other things, explain the apparent paradox. Crime is not so much a function of individual bad behaviour anymore; now it is more a predictable consequence of racial and social oppression.


Sebastian Holsclaw 07.18.04 at 8:09 am

I agree entirely that the drug laws are inappropriate. But the rest of the comment doesn’t make sense.

“The previous comments that dropping crime rates are consistent with rising imprisonment rates is simplistic if not disingenuous. The logic of that fails when the phenomena occurs year after year.”

Incarceration rate goes up year after year AND the crime rate goes down year after year. There isn’t a failure in logic to think about linking the two. And what is the paradox? You might as well complain about an apparent paradox between declining food production and increasing starvation.

“Crime is not so much a function of individual bad behaviour anymore; now it is more a predictable consequence of racial and social oppression.”

And this silly interpretation is precisely what I was worrying about above.


h. e. baber 07.18.04 at 7:26 pm

The rooster crows day after day and the sun comes up every morning…

Even if the overall crime rate is down I’d be curious whether criminal drug use is down. If the increased incarceration rate is a consequence of the war on drugs–stricter enforcement, harsher penalties and longer sentences for drug related offenses, the the hypothesis that “getting tough” on crime in general would predict that drug use, drug trafficking and related offenses would be down. Is it?


bryan 07.18.04 at 7:40 pm

‘“Because” may or may not be a stretch, but “despite” certainly carries some baggage you didn’t even begin to justify.’

When does because move to despite? If this were the situation, or something similar to this, I would say it is a despite:

Crime is at an all time high, so we increase the level of incarceration, crime increases even more we increase incarceration, crime increases we increase incarceration, crime decreases slightly, we say that this is because finally our stern response to crime is paying off, since it is now paying off we increase incarceration, crime decreases some more, the rate of crime decrease is in fact equal to the rate of incarceration increase! Overjoyed we double our incarceration rate, the crime rate decreases by the same rate as the year before, we increase incarceration, crime decreases very slightly, we publish a study showing that there is a clear correlation between increase of incarceration and decrease of crime, we get funding and increase incarceration, crime holds steady, we increase incarceration, crime increases, we increase incarceration, crime increases, we increase incarceration, crime decreases significantly, we announce that this shows our method works, we increase incarceration.

If it was something like that I might worry that we suffered from some sort of obsessive compulsive need to incarcerate, does anyone have any clear data on correlations between these two things?


Lance Boyle 07.18.04 at 8:57 pm

Laws are an attempt. Education is an attempt. The systems they’re embedded in are so huge now it’s easy to see them as realized, as finished versions of what was attempted. Thus products of either system can be seen as identifiable by the institutions. An educated man, an uneducated man. A criminal, a citizen.
The “crime rate” is not a directly parallel indicator of the amount of bad things that are happening in a society. A falling crime rate doesn’t mean the citizenry is necessarily safer, more protected from bad things. It measures broken laws, and the incarceration rate measures apprehended, and convicted and sentenced, criminals. The social benefits, of safety and preservation of public good, depend entirely on the moral accuracy of the laws “on the books”.
Again, these laws are attempts at finding a code that will preserve what needs preserving, and protect what needs protecting.
The priorities of the laws reflect the priorities of the law makers.
In a healthy democracy the lawmakers reflect, not just the contemporary will of the people, but a kind of overall regard for their continuing well-being.

This is no longer the case in the United States, if it ever was.
The grotesque absurdity of Brown vs. Board of Education being even necessary. The death of Martin Luther King, and the consequent cover-up and obfuscation by law enforcement agencies, and many other less dramatic events, need to be considered if you’re going to talk about the incarceration rates of actual living human beings who happen to be black men in the United States.
If you’re going to talk about them as if they were abstract markers on some game board, then of course you’re free to make up whatever rules you want.


dave heasman 07.19.04 at 12:13 pm

Dr Weevil suggests that “Even if prison fails to deter future crime, it certainly cuts down on the opportunities to commit most types of crime as long as the criminal is in jail, however he behaves once he’s out.”
I doubt he’s asked many prisoners about this. In the UK prisons are hotbeds of all sorts of crime. It would be surprising if they weren’t. Of course the victims can be and are ignored, until they die, and often then too.


Dr. Weevil 07.20.04 at 1:52 am

Dave Heasman should learn to read and write more carefully.

1. His reading: I neither said nor implied that there is no crime in prison, I very carefully wrote that prison “cuts down on the opportunities to commit most types of crime”, and that is a fact. For instance, the number of bank robberies, auto thefts, home invasion robberies, stock frauds, and embezzlements committed in prison is minuscule. As for crimes that are common both inside and out, I do not know whether prisoners commit more or fewer murders, assaults-and-batteries, and rapes than unapprehended criminals on the outside. (It is safe to say that rapes of women are far higher on the outside, of men far higher on the inside, but how the totals compare might be difficult to calculate.)

2. His writing: When he writes “hotbeds of all sorts of crime”, does he mean “a wide variety of crimes” or “every variety of crime”? The first is true, the second false, and “all sorts” is so ambiguous that it’s impossible to say which he means.

Next time, D.H., spend a little more time reading, thinking, and writing before you press the Post button.


John Quiggin 07.21.04 at 3:44 am

On the technical use of the term “risk”, economists who analyse data on unemployed workers routinely model the “risk” of getting a job. AFAIK, the statistical techniques in question were originally used in demography and epidemiology, where the terminology is a bit more natural.


Anthony 07.23.04 at 3:51 pm

John Quiggin: why do demographers prefer the word “risk” to the word “probability”?

Lance Boyle: aside from the fact that most young criminals (black or white) weren’t born when Martin Luther King Jr. (and Malcolm X) were assassinated, and most are only dimly aware of events before they were born, your point about the “moral accuracy of the laws on the books” is blunted, because much of the discussion of crime rates is a discussion of violent crime rates, since those are a) what people worry most about – nobody is afraid to walk down certain streets because they may be the victim of insider trading – and b) violent crimes are more consistently reported to law enforcement. Unless your moral compass doesn’t tell you that violent crimes are important to the health and well-being of society, your abstract point is largely irrelevant to the real world.

There is a way in which your point is relevant, at one remove: laws against “victimless crimes” like drug selling and prostitution attract people with criminal dispositions into those fields of enterprise, and subsidise criminal activity.

Comments on this entry are closed.