The Wisconsin Interest

When government commissions are appointed, they are usually charged with investigating some controversial issue. When Congress passed legislation establishing the 9/11 Commission, it did not charge it with “investigating the malfeasance of the Bush Administration,” nor with “determining why the Clinton Administration was derelict in protecting the nation against terrorism.” Rather, the Commission was instructed to “examine and report upon the facts and causes relating to the terrorist attacks of September 11, 2001. . . .”[i]

But when Wisconsin Governor Jim Doyle set up a Commission on supposed “racial disparity” in the Wisconsin criminal justice system, he not only asserted that the disparity is real (which it is) but that it is undesirable. Indeed, his commission is called “The Commission on Reducing Racial Disparities in the Wisconsin Justice System.” It’s true he directed the Commission to “[d]etermine whether discrimination is built into the criminal justice system at each stage of the criminal justice continuum of arrest through parole.” But then he told it to [r]ecommend strategies and solutions to reduce the racial disparity in the Wisconsin criminal justice system. . . .”[ii]

It might seem, on first glance, that “racial disparity”—and here the issue is that blacks are jailed and imprisoned at a much higher rate than whites—is a bad thing.

But what if the disparity is the result of the fact that blacks commit more crimes than whites? Looking back at the Governor’s charge to the Commission, if it’s not established that the disparities are the result of discrimination, how do we know we want to eliminate them? And what if incarceration in fact serves highly desirable goals of deterring crime and incapacitating the criminals? If so, the Commission is on a fool’s errand, instructed to recommend things that will make the quality of life in Wisconsin worse. And particularly worse for black people.

Racial disparity in incarceration

How does one measure racial disparity in imprisonment? One simply divides the rate of incarceration for blacks (usually per 100,000 black population) by the rate of incarceration for whites (measured similarly). When one does this one will find that, in every state in this union, blacks are jailed and imprisoned at a higher rate than whites. And not just a little bit more. Liberal activist group Human Rights Watch reports disparity ratios ranging from three (in Vermont) to 34 (in the District of Columbia).[iii] Likewise, the equally liberal Sentencing Project shows that nationally blacks are incarcerated at 5.6 times the rate of whites.[iv]

In all of these analyses, Wisconsin ranks near the top in terms of racial disparity. Human Rights Watch shows Wisconsin in fourth place, and The Sentencing Project shows Wisconsin in fifth place, with a disparity ratio of 10.6, right behind Iowa, Vermont, New Jersey, and Connecticut.[v]

But before one gets too upset about the disparity in Wisconsin, one has to ask whether it is actually out of line. It is if we assume that blacks commit crimes at the same rate as whites. But we all know that’s not true.

The first clue about problems with these ratios comes when we look at what states are where in the list. Southern states cluster near the bottom of the rankings. They imprison blacks and whites in much more equal ratios than does Wisconsin, or states like Minnesota, Iowa or Connecticut. One has to wonder whether this has anything to do with the fact that Southern states still have substantial rural black populations, while the Great Migration brought blacks into the central cities of Northern states.

So we created a statistical model of racial disparity, including each state and the District of Columbia. In order to predict the disparity of each state, we simply used the proportion of blacks in the state who live in a central city of a Metropolitan Statistical Area.[vi] It turned out that there is a strong relationship: states with relatively few blacks in the central cities of metropolitan areas had low disparity scores, and those with a heavily urban central city black population showed high ratios. A ten percent increase in the percentage of blacks living in central cities increased the disparity index by about one, and a 30 percent increase kicked it up by about three. We then added, for each state, the percentage of the black population below the poverty level, and the percentage of the white population below the poverty level. We found, as expected, that black poverty drove up the disparity ratio, and white poverty drove it down (as more whites were imprisoned).[vii]

So how did Wisconsin fare in this model? Our model predicted a racial disparity of 12.13 in Wisconsin. But the actual figure was 11.58. In other words, Wisconsin has a bit less racial disparity than one would expect.

One can argue, of course, that this country is just crawling with racists, and that racist cops, judges, prosecutors, and jurors are about as common here as anywhere else. But it’s absurdly simplistic to point out that Wisconsin’s disparity ratio is especially high. Given the state’s population, it is about where we would expect it to be.

Blacks vs. Whites: Length of prison sentence

The fact that Wisconsin’s racial disparity index is about where we would expect it to be doesn’t, therefore, rule out considerable racial discrimination in the system. If blacks are disproportionately incarcerated in Wisconsin, this could be the result of either or both of two factors. First, it could be that blacks are more likely to get incarcerated for committing the same crime that a white commits. Secondly, it might be that blacks, once sentenced for a certain crime, serve longer sentences. If the latter is true, it might be for legitimate reasons—having a prior felony conviction or chronically breaking prison rules, for example. Or it might be for discriminatory reasons. Blacks might simply be given longer sentences for the same crime. Parole boards might be less inclined to let blacks out of prison than they are similarly situated whites.

To examine this issue, we obtained data on every prisoner released from prison in Wisconsin in the years 1998-2002.[viii]

We produced a statistical model to explain the number of years a prisoner serves before being released. Our model controlled for things that should be controlled for: being a male (rather than a female) and having been previously incarcerated for a crime. We are dealing here with FBI “index offenses,” which fall into two categories—violent offenses (murder, forcible rape, robbery, and aggravated assault) and property crime (burglary, larceny-theft, motor vehicle theft, and arson). We also add a third category: drug offenses. It is often asserted (with considerable evidence) that drug crimes are largely responsible for the high levels of incarceration of blacks.[ix]

One might wonder, of course, whether breaking offenses down into narrower categories (rape only, murder only, robbery only, etc.) would show some significant racial effects. We tried doing that, and to no avail. There is simply no evidence that the prison system in Wisconsin keeps blacks behind bars longer than it does whites.

Who goes to prison?

If there is no apparent bias against blacks in the length of sentences, is there a bias in who gets sent to prison?

Any social scientist has some major problems dealing with this, because data isn’t available to do it the obvious way. If we had a database of every crime committed, and the race of the offender, we could see whether blacks who commit a certain crime are more likely to get sent to prison than whites. There has long been such data where murders are concerned.[x] And a new reporting system, which is just beginning to come into use in Wisconsin, has similar information (more on that later).

But in the absence of such information, we have to make do with less than the ideal data. For example, a recent report from the Wisconsin Sentencing Commission examined only offenders who had been charged, gone to court and convicted (or more likely, copped a plea) to see whether blacks were more harshly sentenced.[xi] As the report clearly acknowledges, this fails to take into account several stages in the process, including whether the offender is apprehended, charged, and prosecuted. Yet all these processes could produce racial disparity.

We tackled this problem by using aggregate data. We produced a statistical model of admissions to prison in the years 1998-2002, using all counties in Wisconsin. Our model used the number of Uniform Crime Report offenses in each county to predict the number of admissions to Wisconsin state prisons from each county.[xii] Obviously, there is a strong relationship. The more crimes committed in any county, the more people get convicted and sent to prison.

We then added to our model the number of black citizens in each county. If blacks are disproportionately imprisoned, the greater the black population, the more people will be imprisoned from that county. If blacks are treated like everybody else, race will add nothing to the predictive power of our model. Our findings were as follows.

In short, it seems that blacks are more likely to be imprisoned for property crimes, but less likely to be imprisoned for violent crimes. On net, there is nothing here that suggests that in the aggregate blacks are disproportionately incarcerated. Since prison sentences for property crimes are generally less than for violent crimes (especially for blacks, as we have seen) the net effect is either a wash, or perhaps even less aggregate prison time for blacks.

This sort of highly aggregated analysis certainly has its limitations. For example, it’s perfectly possible that counties with a large black population punish all offenders more severely for property crimes, and less severely for violent crimes. But as we shall see, the evidence tends to converge in support of our aggregate findings.

An alternative approach to bias in prison admissions

An alternative approach to bias in the admission of offenders to prison is to look at the prior offenses of black and white prisoners admitted to the Wisconsin state prison system.

The Department of Corrections Public Information Data File contains information on inmates in state prisons. We analyzed a subset of admissions to Wisconsin prisons between 2001 and 2006.[xiii] For about 39,000 cases, there was information on both the race of the offender (self-reported) and the number of prior offenses.

Prior offenses is a powerful predictor of both whether an offender will be sentenced to prison (versus jail, or parole, or a suspended sentence), and of the amount of time an offender will serve.

Our logic in looking for racial bias was simple. If there is a bias against blacks, it should be “easier” for a black offender to get himself admitted to prison. That is to say, more blacks than whites should be locked up with no prior convictions. Conversely, it should take more prior convictions to get a white sent away. Think of this as a kind of perverse affirmative action program. A bias against black offenders will lead to a “less qualified” (in terms of prior offenses) freshman class of blacks. Perhaps, for a particular offense, a black with no prior convictions will be imprisoned, while a white will not – at least not until he commits another offense. Table 1 contains our results.[xiv]

Contrary to the notion of disproportionate black incarceration, it seems that it is easier for a white person to get sent “up the river” than a black. While 48.2% of white prison admissions were of an inmate who had no prior felony convictions, only 37.5% of black prison admissions were.

This, of course, doesn’t necessarily prove racial bias, since the black and white offenders in the sample may be different in a variety of ways. But it certainly fails to support the notion that blacks are victims of discrimination in incarceration.[xv]

Working with better data: The Milwaukee Police Department incident reporting

A new system of crime reporting, the Incident Based Reporting System, is slowly coming into use across the nation. In Wisconsin, the only city with a large black population that uses this system is Milwaukee.[xvi] The data allow, in theory, researchers to identify both the race of the victim and the race of the offender for any crime that came to the attention of the Milwaukee Police Department, and thus created an “incident report.”[xvii]

The first thing we will look at is the race of offenders in violent offenses. For the 6085 violent offenses in the database, only 12.6% have missing data for the race of the offender. Thus our data is pretty good in terms of completeness.

The “offender” in each offense was coded as black, white or some other race.[xviii]

The key, indeed stunning, thing about this is the nearly six to one ratio of black to white offenders. In the city of Milwaukee, violent crimes are overwhelmingly committed by blacks. This, of course, is a disparity in the raw numbers of offenders. Given that there are 1.338 times as many whites as blacks in the city’s population, this translates into a 7.99 disparity ratio.[xix] A randomly drawn black in Milwaukee is about eight times as likely to commit a violent crime as a white. Ideally, we would like to look at prison admissions for the city of Milwaukee, and see whether this disparity in offenses matches the disparity in imprisonment. Unfortunately, prison admission data are available by county, but not by city. And there is no way to know the racial composition of offenders in the suburbs. Since we lack the necessary data, any precise estimate of racial disparity among people entering prison from Milwaukee isn’t possible. Given the lopsided numbers that we see, however, it seems likely that the disparity in imprisonment mirrors a disparity in offenses.

The incident reports turned out to be less helpful with regard to other kinds of crime. Where property crimes were concerned, for over three-quarters of the cases the race of the offender is not known. This is not surprising. Things are stolen all the time by people not seen. Cars are taken for a joyride and then abandoned, with nobody having any knowledge of the thief.

Where drug crimes were concerned, we faced a very different, but equally nettlesome problem. By a ratio of about three to one, blacks are the offenders in drug crimes in the Milwaukee incident reports. Unfortunately, since this is a “victimless” crime it is a certainty that most offenses go unreported. Further, particularly for white users in suburbs, a drug offense may lead to only a municipal citation, rather than a criminal charge. So again, we can say little about racial disparity.

So what is the conclusion?

Does all this add up to a consistent pattern? Not neatly. But some things can be said. Our finding that blacks are more likely to be sentenced to prison for property crimes, but serve less time for those crimes is partially consistent with the Wisconsin Sentencing Commission study.[xx] Their data shows blacks more likely to be given a prison sentence for property crimes, although it shows no difference in sentence length. Our finding (above) is that their actual time served for a property crime is less than whites.

Then we can throw into the mix the fact that a Vera Institute study of the Milwaukee County District Attorney’s office showed that office a bit more likely to decline prosecution of non-whites accused of property crimes.[xxi]

While it would be interesting to see exactly what is going on here, the evidence does not add up to any general bias against black offenders where property crimes are concerned. Other evidence suggests black offenders are treated more leniently both in general (our prison admissions data), and with regard to violent crimes (our aggregate analysis).

Incarceration is good

A pervasive bias in the literature on racial disparity is the notion that it is a bad thing to lock people up. The simple reality, however, is that incarceration has some extremely good social consequences. First, it deters offenders. Second, it incapacitates people who would otherwise be on the street committing crimes.

The notion that punishment deters crime used to actually be controversial—at least among academics, a group largely immune to common sense. Since the 1970s, however, even the sociologists admit that crime can be deterred by punishing criminals. (The economists believed it all along.)[xxii] There is still some debate about whether the death penalty deters murder better than long prison sentences. But if the issue is whether typical prison sentences deter crime the evidence is clear. Indeed, most of the studies showing no deterrent effect of the death penalty show that prison deters murder.

The data thoroughly debunk the uninformed notion that certain crimes—especially crimes like murder and assault—can’t be deterred because the criminal “isn’t thinking rationally.” The data show these crimes can be deterred.

If prison has a deterrent effect, it also reduces crime by incapacitating criminals who would otherwise be free to prey on victims.

For social scientists, the separate effects of deterrence (people know they are likely to be punished if they commit a crime, and are less likely to do it), and of incapacitation (criminals are kept locked up) are hard to distinguish. If we find that a particular jurisdiction locks up a lot of people, and crime is low, is this deterrence or incapacitation, or both? The answer, when social scientists try to sort these issues out, is “both.”

University of Chicago economist Steven Levitt, for example, says that increased imprisonment is one of the reasons crime decreased sharply in the 1990s:

In another study, Levitt estimated the social benefit of incarcerating one additional prisoner at $53,900 annually, which was well above the average cost of imprisonment.[xxiv]

One of the interesting phenomena in this literature is the “grumpy liberals.” Quite frequently, authors who clearly don’t like incarceration have to admit that it does reduce crime. A 1994 Rand Corporation study, for example, deals with a “three strikes and you’re out” law in California.[xxv] The authors conclude that each five million dollars spent on prisons will prevent one murder, 20 rapes, 55 robberies, 120 aggravated assaults, 110 burglaries of a serious nature, and five arsons.[xxvi] Note that while this certainly seems cost effective, sending felons away for life on the third conviction is probably not the optimal policy. Criminals tend to “age out” of crime, and so a long (but not lifetime) sentence would give most of the benefits of a life sentence, but at less cost.

Likewise, William Spelman examined several studies on this issue, and concluded that each percent increase in prison expenditures likely cut crime by .20% to .40%.[xxvii] He then proceeded to suggest that any of a smorgasbord of nice sounding social programs might “yield benefit-cost ratios much greater than the 1.50 to 2.00 that is the best we can expect from continued prison expansion.”[xxviii] A sensible person would conclude that when these programs start to actually show some reduction of crime rates, we then might consider reducing prison populations. At the moment, such notions are pie in the sky.

Of course, there are some costs these studies typically ignore: the possible earnings of an offender if not imprisoned and the family consequences of having a father locked up and away from the family, for example. But there are also benefits ignored: the ability of citizens to walk down the street, sit in a park, or go to the store without fear, for example.

And these studies show that incarceration has a favorable cost-benefit ratio even if we leave deterrence out of the equation. But in the real world, punishment deters, making incarceration even more cost effective.

Drug crimes

We need to discuss briefly the area that seems to best support the notion that blacks are disproportionately imprisoned: drug crimes. It is common to point out that whites are about as likely to use drugs as blacks, but blacks are more likely to get sent to prison for drug use than whites.[xxix]

And indeed, if we look at the Wisconsin prison admissions, we find that (again using 2001-2006 data) blacks are disproportionately imprisoned for drug offenses. While 17.8% of all white prison admissions were for drug offenses, 38.4% of black admissions were.

It is, however, at least a bit simplistic to treat white drug use and black drug use as equivalent. Whatever the legality and morality of white drug use, it is far less likely to be a threat to the community than black drug use. A private vice is not the same thing as a public problem. Pot smoking suburban teenagers seldom if ever get in gun fights and shoot up innocent children. But that happens with some frequency in the inner city.

And indeed, the black community (as opposed to white liberals and some black activists) seems to favor tough penalties for drug use. As explained by criminologist John J. DiIulio, Jr., the Congressional Black Caucus was fully on board in 1986, when the federal law against crack cocaine was first debated. A few years earlier, these same representatives had pushed President Reagan to create the Office of National Drug Control Policy (more popularly knows as the drug czar’s office). This was something Reagan was not anxious to do. Of course, America’s first “black” president (Bill Clinton) staunchly refused to weaken penalties for drug crimes.[xxx]

And not surprisingly, blacks are as tough in their attitudes toward drugs as are whites. As the Pew Research Center put it:

Some black leaders have criticized mandatory minimums, but blacks and whites see eye-to-eye on this issue and several other drug policy issues.[xxxi]

This is not to say that current drug policy is optimal. But it’s not something that was imposed on the black community by racist whites.

Conclusion

So is the Wisconsin Commission on Reducing Racial Disparities in the Wisconsin Justice System on a fool’s errand? No necessarily. Regardless of what Jim Doyle directed them to do, they might well end up doing some useful work. While evidence of bias against black offenders is thin, we can’t summarily deny that it ever exists, especially in light of the recent Frank Jude case.

But all too easy to ignore, in all this concern for people in prison, is the fact that other people have a stake in this issue. From the Milwaukee incident reports, here is the distribution of offenders and victims for violent crimes

In Milwaukee (and this is true everywhere) the vast majority of the victims of black offenders are themselves black. Of attacks by black offenders, 73.3% were committed against black victims. If we average the table the other way, we find that 97% of all violent crimes against black victims are perpetrated by blacks. Failure to lock up violent black criminals hurts whites only marginally (and for those in the further out suburbs, barely at all). It especially hurts blacks.

Blogger James Harris lives in the Milwaukee inner city. He was out walking his dog one night, and describes what happened to him.

No analysis, no rhetoric and no policy should side, implicitly or explicitly, with the people who shoot little girls in the head.

John McAdams is an Associate Professor of Political Science at Marquette University.

Notes

[i] Public Law 107–306, 107th Congress. Online at: http://www.9-11commission.gov/about/107-306.pdf

[ii] EXECUTIVE ORDER #189. Online at: http://www.wisgov.state.wi.us/journal_media_detail.asp?locid=19&prid=2573.

[iii] Human Rights Watch, “Table 3, Rates of Incarceration by Race, 1996” at http://www.hrw.org/reports/2000/usa/Table3.pdf.

[iv] Marc Mauer and Ryan S. King, “Uneven Justice: State Rates of Incarceration By Race and Ethnicity.” Online at: http://sentencingproject.org/Admin/Documents/publications/rd_stateratesofincbyraceandethnicity.pdf.

[v] Op. cit., p. 12.

[vi] Populations were taken from 2000 Census data, Summary File 3. Ratios of incarceration were taken from “Table 16: Number of inmates in State prisons and local jails” from the web site of the Bureau of Justice Statistics.

[vii] The technical details of the analysis can be found online at: http://mcadams.posc.mu.edu/blog/techreport.pdf.

[viii] This data did not include inmates released from jails (unlike the data for racial disparity, which does count as “incarcerated” inmates in jails). Since jail terms tend to be short, and there is less room for discretion to be used in a discriminatory manner, this should not affect our analysis.

[ix] It is not uncommon for someone to be convicted of multiple offenses. When we categorized the crime for which they were convicted, but used the offense with the longest possible prison sentence, in other words, the most serious charge.

[x] See John C. McAdams, “Racial Disparity and the Death Penalty,” 61 Law & Contemp. Probs. 153 (Autumn 1998).

[xi] Brenda R. Mayrack, Race & Sentencing in Wisconsin: Sentence and Offender Characteristics Across Five Criminal Offense Areas. Wisconsin Sentencing Commission, 2007.

[xii] Unfortunately, drug crimes are not among FBI “index crimes,” and thus we had no data on drug crimes in this analysis.

[xiii] Note that our unit of analysis is an admission, and not a particular inmate. Given a fairly high rate of recidivism, some inmates in our sample were admitted twice during the time included time span. Also, the dataset includes admissions to programs besides prison. We filtered out the latter cases, but not the former.

[xiv] A very small number of Asians, and all inmates whose race is unknown are omitted from the table.

[xv] We wondered whether our findings might reflect the fact that many black offenders are sent to prison from Milwaukee County. Perhaps it’s harder for all Milwaukee offenders to get sent to prison, and our results are merely the result of the disproportion of black offenders from Milwaukee among all black offenders in the state. Our analysis showed, however, that differences in prior convictions followed the same pattern in Milwaukee as the rest of the state, with the differences being a bit larger. See: http://mcadams.posc.mu.edu/blog/techreport.pdf.

[xvi] A detailed technical description of the data can be found on the Wisconsin Office of Justice Assistance web site at: http://oja.state.wi.us/docview.asp?docid=3603&locid=97.

[xvii] Quite obviously, some crimes are never reported. Unfortunately, this underreporting is doubtless not uniform across types of offenses and segments of the population. But fortunately, it’s probably most complete and accurate for the most serious violent crimes (with the possible exception of rape).

[xviii] In cases where there were multiple offenders, where they were all of the same race, they were coded as that race. If there was a mixed race group of offenders, they were coded as “other race.”

[xix] This is based on 2000 Census population.

[xx] Mayrack, op. cit.

[xxi] Vera Institute, “Milwaukee County Declination Decisions: 12 – Prosecution and Racial Justice.” Unpublished tabulations.

[xxii] A very partial list from a voluminous literature includes: Alfred Blumstein, Jacqueline Cohen, and Daniel Nagin, eds. Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates (Washington, D.C.: National Academy of Sciences, 1978); Stephen S. Brier and Stephen E. Feinberg, “Recent Econometric Modeling of Crime and Punishment: Support for the Deterrence Hypothesis?” in Feinberg and Reiss, Indicators of Crime and Criminal Justice, pp. 82-97; Philip J. Cook, “Research in Criminal Deterrence: Laying the Groundwork for the Second Decade,” Crime and Justice, Vol. 2, ed. Norval Morris and Michael Tonry (Chicago: University of Chicago Press, 1980), p. 219; Isaac Ehrlich, “Participation in Illegitimate Activities: A Theoretical and Empirical Investigation,” Journal of Political Economy 81 (1973): 521-565; Maynard L. Erickson, Jack P. Gibbs, and Gary F. Jensen, “The Deterrence Doctrine and the Perceived Certainty of Legal Punishment,” American Sociological Review 42 (1977): 305-317; Matthew Silberman, “Toward a Theory of Criminal Deterrence,” American Sociological Review 41 (1976): 442-461; Charles R. Tittle, Sanctions and Social Deviance: The Question of Deterrence (New York: Praeger, 1980); Steven Levitt, “Why do Increased Arrest Rates Appear to Reduce Crime: Deterrence, Incapacitation, or Measurement Error?,” Economic Inquiry, July 1998; Steven Levitt, “Juvenile Crime and Punishment,” Journal of Political Economy, December 1998; Michael Weinrath & John Gartrell, “Specific Deterrence and Sentence Length,” Journal of Contemporary Criminal Justice, May 2001.

[xxiii] “Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline and Six that Do Not.” Journal of Economic Perspectives, Winter 2004, pp. 178-179.

[xxiv] “The Effects of Prison Population Size on Crime Rates: Evidence from Prison Overcrowding Litigation.” Quarterly Journal of Economics, May 1996, p. 346.

[xxv] Peter Greenwood, et al., Three Strikes and You’re Out: Estimated Benefits and Costs of California’s New Mandatory-Sentencing Law. Santa Monica: The Rand Corporation.

[xxvi] Op. cit., p. 19.

[xxvii] “What Recent Studies Do (and Don’t) Tell Us about Imprisonment and Crime.” Crime and Punishment Vol. 27 (2000), p. 484. Spelman said that any figure between .20 and .40 “could be defended,” but also pointed out that the actual figure could be outside that range.

[xxviii] Idid. P. 487.

[xxix] For example, Graham Boyd “The Drug War is the New Jim Crow.” Online at: http://www.aclu.org/drugpolicy/racialjustice/10830pub20010731.html.

[xxx] “My Black Crime Problem, and Ours,” City Journal, Spring 1996. Online at: http://www.city-journal.org/html/6_2_my_black.html.

[xxxi] “Interdiction and Incarceration Still Top Remedies.” Online at: http://people-press.org/reports/display.php3?ReportID=16.

[xxxii] “Walking in a war zone” The National Conversation, June 15, 2007. http://nationalconversation.typepad.com/the_national_conversation/2007/06/walking_in_a_wa.html.

Content on this page requires a newer version of Adobe Flash Player.

Does Wisconsin Lock Up Too Many Blacks?