Kieran Healy

Inequality and Incarceration

Disclaimer: I’m not an expert on criminal justice policy. I don’t even play one on TV.

Jumping around randomly on weblogs (I’m trapped in the library during a bizarre Arizona hail and rain storm), I came across this post by Virginia Postrel:

COLLEGE VS. PRISON: Iain Murray of TechCentral Station has an important piece debunking the common claim, made most recently by the Justice Policy Institute, that there are more African-American men in prison than in college. For some reason, advocacy groups believe that portraying black men as dangerous criminals is good for race relations.

Debunking false claims is the right thing to do, of course, and the JPI people should know better than to try to make political hay out of flawed analogies. But that second sentence seems a bit cheap. My guess is that the JPI is not “portraying black men as dangerous criminals”, it’s complaining that the state is locking too many black men up, which is an entirely different claim. Postrel quotes Iain Murray:

[T]here is a total of 1,216,000 young African Americans [of both sexes] in college, as opposed to 189,000 [of college age, almost all men] in custody. That is a huge disparity which should be very good news.

This doesn’t seem like a helpful way to present the data. The number of young white prisoners in the U.S. is large compared to almost any other prison population in the world. But it’s very small compared to the number of white college students in the U.S. Does this mean we should be happy with the incarceration rate? Murray is rightly criticizing a bad analogy, but then exploits the same analogy to make his own point about the large ratio being “very good news”.

A post I read yesterday by Eugene Volokh is in a similar vein:

In a recent discussion on a lawprofs’ discussion list, I again heard the claim that “the vast majority” of prisoners are not morally culpable because they’re in prison for a drug-related crime. I’d heard this claim before, so I decided to look this up—and it turns out that, at least as of 1997, 24% of the federal and state prisoners had drug offenses listed as their most serious offense. (Source: Bureau of Justice Statistics, “Correctional Populations in the United States, 1997,” tbls. 1.5, 1.14, and 1.15.) … to return to the original claim, I’d say that the vast majority of the people in prison are morally culpable, even if one were to conclude that all drug-related crimes are not morally culpable…

Again, Volokh wants to debunk, and again he’s right. Whoever claimed the “vast majority” were incarcerated for drug crimes hadn’t looked up some easily available information. But like Postrel, Volokh then narrows the question by focusing on the individual culpability of the offenders. (In fairness, Volokh is responding to a claim about culpability, not bringing it up out of the blue. My point is about where both sides end up in this argument.) As with the comparison to college students, the original policy problem gets lost. Twenty four per cent seems like a large number to me, as it does to the Bureau of Justice Statistics. They note that “For the fourth consecutive year, the largest number of admissions to State prison were for drug offenses, 102,600 drug offenders were admitted in 1997”. The moral culpability of prisoners seems to me to push the central questions aside once more. The debate ends up being focused on side-issues and, in these examples, driven by sound bites and a bit of point-scoring.

Postrel rightly points out that “the trick [with data of this kind] is to think about age within the group. Not everybody is college-age, and prison sentences can last longer than four years.” Here’s some data that are are of interest in this context. They come from a paper titled “Black Economic Progress in the Era of Mass Imprisonment” by Bruce Western, Becky Pettit and Josh Guetzkow. (Bruce is a former teacher of mine, and we’ve co-authored papers on a different topic.) They provide some of the best available estimates of how the prison population is stratified, and how that has changed over time. Here is an excerpt from Table 1 of their paper:

Prison and jail incarceration rates for black and white men by age and education, 1980, 1999.

What does this table tell us? The data show that between 1980 and 1999, rates of incarceration increased substantially for men, and that for certain kinds of men they increased dramatically. The incarceration rate for white men goes up by 0.6 percentage points over the period. For black men it goes up by 4.4 percentage points. So whatever has changed is disproportionately affecting blacks over whites. Similarly, the increase in the incarceration rate is disproportionately concentrated in young men, and in unskilled men (measured here as “high-school dropout”). For people who fall into all of these categories—- black, young, unskilled—- the increase in the incarceration rate is enormous. It jumps from 14 percent in 1980 to 41.2 percent in 1999. The authors comment:

After 50 years of relative stability in the proportion of people incarcerated [from 1920-1970], the prison population doubled in size between 1970 and 1982. Between 1982 and 1999, the prison population increased threefold. These broad trends conceal substantial inequality. Prison incarceration rates are about eight times higher for blacks than whites, and high school dropouts are more than twice as likely to be in prison than high school graduates. Consequently, much of the growth in imprisonment in the three decades after 1970 was concentrated among young minority men with little education. By the late 1990s, about two-thirds of all state prison inmates were black or Hispanic, and about half of all minority inmates had less than twelve years of schooling.

What’s driving these changes? The paper should be read in full, (along with the other papers on this topic available on Bruce Western’s homepage), but a central point is that changes in criminal and sentencing policies are an important cause:

The great irony of the prison boom is that it comes at a time when policymakers have set out to roll back the role of government in the lives of the disadvantaged. Despite anti-government rhetoric in policy debates, the government has not regulated the lives of unskilled minority men so intensively since depression or wartime. The policy offensive was wide-ranging, affecting policing, sentencing, prison construction, post-release supervision and a variety of other measures at the state and Federal levels of government. The sheer commitment of public resources is comparable in magnitude to social welfare efforts in the 1960s and 1970s. Unlike anti-poverty policy, however, the punitive trend in criminal justice policy conceals and deepens economic inequality between blacks and whites.

Update: This post has followups here and here.

What does this table tell us? The data show that between 1980 and 1999, rates of incarceration increased substantially for men, and that for certain kinds of men they increased dramatically. The incarceration rate for white men goes up by 0.6 percentage points over the period. For black men it goes up by 4.4 percentage points. So whatever has changed is disproportionately affecting blacks over whites.

Is this reasoning correct? If the changes were race-blind then it seems to me that we would not expect the same net change in incarceration rate, rather we would expect the same percentage change in incarceration rate. If I’m correct about this then the picture becomes quite different: while both races experienced increasing rates of incarceration, in all three categories the percentage change was higher for whites than for blacks.

while both races experienced increasing rates of incarceration, in all three categories the percentage change was higher for whites than for blacks

Yes, that’s right. It depends whether you focus on how the percentage change in the incarceration rate differs between groups, or on how the total percentage of the population incarcerated differs between groups. In the former case, the rate of growth was a bit higher for whites. As Western et al say in their paper, The table shows that the rate of growth in incarceration was slightly faster for whites than blacks. Thus racial disparity in imprisonment, measured by the black-white ratio in incarceration rates, declined slightly from 1980 to 1999. The Prison Boom incarcerated many whites as well as blacks between 1980 and 1999, of course. But I was trying to say that this growth process disproportionately affected blacks because such a higher percentage of them ended up being incarcerated as a result of it. Western et al again: Even so, black men between the ages of 18 and 65 were nearly 8 times as likely as whites to be in prison or jail in 1999. Despite small declines in racial disparity, the largest absolute gains in incarceration were recorded by black men. Racial disparity in incarceration is partly ex-plained by the relatively low level of schooling of young black men compared to young whites. If we look just at high school dropouts, the black-white ratio in incarceration rates is just half as large as for the population as a whole. Even among dropouts, however, blacks remain four times more likely to be in prison or jail than whites. So I think it’s still true to say that whatever drove the increase in the prison population disproportionately affected young, unskilled and especially black men. But you were right to point out that I could have expressed it better (as Western et al in fact did).

A New Analysis of Incarceration and Inequality

I’ve written about the intersection of incarceration, race and the labor market 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 …