Race and The Law
To the Editor:
James Q. Wilson’s review of Randall Kennedy’s Race, Crime, and the Law [September] is, as usual, subtle and persuasive. As Mr. Wilson notes, Kennedy’s book makes some very good points. We have, however, an important caveat concerning one issue with which Kennedy deals and where Mr. Wilson seemingly accepts his argument.
Relying on the work of David Baldus and his colleagues, Kennedy concludes that, while in certain homicide cases (those dubbed either highly aggravated or highly mitigated) the race of the victim is not significant, in other, so-called “middle-range” cases a perpetrator is more likely to receive the death penalty if the victim is white than if he is black.
This argument is not persuasive. Kennedy and, apparently, Mr. Wilson completely ignore the criticisms of Baldus’s work by Joseph Katz. Katz’s is not an obscure technical critique concerned with minor statistical issues. Rather, it cuts to the quick, and, if true, it seriously compromises Baldus’s authority.
In 1983, in a review of the literature, the National Academy of Sciences concluded that according to the most reliable and most carefully researched studies, the link between race and sentencing in the application of the death penalty was very weak. Baldus’s work was designed to overcome that conclusion. It was used to appeal the death sentence of Warren McCleskey, a black man who had shot and killed a white police officer in the course of a robbery in Georgia. To demonstrate that McCleskey was probably the victim of discrimination, Baldus developed a regression equation of 238 variables that, according to him and his colleagues, proved a pattern of discrimination in applying the death penalty in middle-range cases in Georgia. McCleskey’s homicide, Baldus claimed, was just such a case.
The state called on Joseph Katz, a statistician, to review the Baldus data for the prosecution. Katz demonstrated that Baldus’s complicated equation yielded only a 6-percent increase in the probability of a death sentence if the victim were white: even if the Baldus model were perfect, in other words, it did not explain much. But in addition, as Katz pointed out, Baldus made all sorts of untenable assumptions about the raw data, including assigning arbitrary values to items for which he had no information.
Furthermore, many of Baldus’s findings were themselves peculiar. According to his equations, for example, the probability of a death sentence in Georgia was raised if the perpetrator had been coerced into commit-ting the crime; raised if the perpetrator was unaware of a partner’s intent to use force; reduced if he took a victim hostage; and reduced if the murder was committed in the course of a felony. Unless the citizens of Georgia had lost all their marbles, it is highly unlikely that these variables capture the reasons the state courts imposed the death penalty in some cases and not in others.
Katz also pointed out that Baldus’s methodology was unreliable—as he put it, the “importance of each variable changed dramatically from formula to formula depending upon the specific combination of factors . . . included”—and he offered to produce “thousands of formulas in which the effect of the racial variables in forming the ultimate sentence [would] disappear completely.” Katz provided several alternative regressions, using approximately 40 variables as examples, in each of which the race of the defendant and the race of the victim proved insignificant.
Finally, on the issue of whether the McCleskey homicide was a middle-range case and hence at higher risk for discrimination, Katz found that, to the contrary, it was highly aggravated, even by Baldus’s categories. Randall Kennedy himself concedes that the case seems to fit the category of highly-aggravated murders, though he raises no further questions about the reliability of the Baldus study.
Based on what we have learned from our own research (see “Execution by Quota?,” Public Interest, Summer 1994), we believe greater skepticism about Baldus’s study is warranted than either Randall Kennedy or James Q. Wilson allows.
Center for the Study of Social
and Political Change
To the Editor:
In his review, James Q. Wilson writes: “Of course, the chances of an average black male assaulting or robbing anyone are very small. But they are six times higher than in the case of an average white male.”
This understates the seriousness of violent racial crime committed by blacks. In 1992, according to the Justice Department, almost one million whites were murdered, robbed, assaulted, or raped by blacks. In that same year, 132,000 blacks were the victims of similar crimes committed by whites. This means that with just 12 percent of the population, blacks commit more than 50 times as many violent crimes as whites.
In view of these figures, it should not be surprising that so many ordinary citizens are fearful of young black males.
Valhalla, New York
To the Editor:
James Q. Wilson justifies the police practice of considering race in deciding whether to question and arrest people they see on the street, on the ground that young black males are more likely to be engaged in criminal activities. He is, of course, right that the police do take race into consideration and it is entirely sensible for them to do so, even though this results in seriously unequal treatment of blacks as compared with whites.
One could take the position that, despite the logic and reasonableness of this police behavior, it is unconstitutional, and therefore should be enjoined by the courts. However, Mr. Wilson does not take this position and, regardless of my liberal tendencies, I would not take it, either. But this means that Mr. Wilson and his fellow-traveling conservatives should have to account for this “discrimination” against blacks in formulating their notion that the Constitution demands a color-blind society. As this illustration demonstrates, the Constitution demands no such social norm when there are important and rational reasons for deviating from it.
The obvious question is whether the same rule should not apply to other government agencies or institutions in society. For example, why cannot school systems consider race in deciding that diversity is an educational goal important enough to admit blacks with lower test scores than whites? The only basic difference I can see between the two examples—police discrirninating against blacks in making decisions whether to arrest and colleges “discriminating” against whites in choosing students and faculty—involves the people who are being hurt. In the first case, innocent blacks are being hurt and in the second case, innocent whites. But clearly that cannot justify allowing discrimination in one situation and not the other.
Indeed, there is far more reason to permit discrimination in favor of blacks. It is blacks who were subjected to slavery and then to a long history of blatant discrimination that is still not entirely ended. It is blacks who were supposed to be the beneficiaries of the Fourteenth Amendment and other post-Civil War amendments. It is blacks, as a small minority, who need the protection of the Constitution and the courts—not the white majority, which can more than adequately protect its interests in the voting booth.
The simple fact is that we are not a color-blind society and that race matters a great deal in numerous contexts and will continue to do so for the foreseeable future.
Bruce J. Terris
James Q. Wilson writes:
Stanley Rothman and Stephen Powers correctly describe the criticisms made of David Baldus’s essay on race and the death penalty. But they are mistaken in assuming that I accepted the study. I tried to say only that Randall Kennedy had approached the Baldus work with care. Beyond that I did not wish to go. If I inadvertently used a phrase that led them to think I had gone farther, I apologize.
Frank Messmann is wrong to think that blacks commit 50 times more violent crime than whites. We do not know how many whites were murdered, raped, robbed, or assaulted by blacks or how many blacks were victims of the same crimes at the hands of whites. We can guess at these numbers by using victimization surveys or inferences from data about convictions. But they are only guesses. Here are the estimates for adults that I derive from the victimization surveys, based on data from 1985 through 1988.
Rape: 10 percent of all rapists are whites who rape blacks, and 18 percent are blacks who rape whites.
Robbery: 6 percent of all robbers are whites who rob blacks, and 40 percent are blacks who rob whites.
Aggravated assault: 8 percent of all assaulters are whites who assault blacks, and 17 percent are blacks who assault whites.
Murder: there are no victim surveys for murder—the victim can’t talk!
Apply these rates of interracial crime to the number of persons arrested, and you find about 22,680 whites arrested for raping, robbing, or assaulting a black and about 60,064 blacks arrested for raping, robbing, or assaulting a white. The ratio between these two numbers is about three to one. Thus, not counting homicide (which is primarily an intraracial crime), blacks are three times (not, Mr. Messmann, 50 times) as likely as whites to hurt a person of the opposite race.
Bruce J. Terris thinks that if the police can take race into account in stopping and questioning blacks, other government agencies—schools, for example—should be able to take race into account in employment and admissions policies. To him, these two cases are alike in that people are hurt by race-conscious decisions yet they differ only in race.
There are two responses to this argument. First, the Supreme Court has never authorized the police to use race as a basis for arresting someone, and I strongly doubt that it ever would. Similarly, the Court has refused to allow race to be used as an explicit basis for admitting students.
Second, the Supreme Court has not blocked the use of race as one among several factors that induce police to question (not arrest, just question) a person on the street. Mr. Terris may think there is no important difference between questioning and then (unless there is material evidence of a crime) releasing a person, and admitting a person to a college, law school, or medical school. But it seems to me that there is a vast difference. Questioning a person on the street involves a minor inconvenience and no significant harm and may lead to the solution of a serious crime; allowing a person to enter a college will alter his life and deprive the person passed over of the same chance.
Mr. Terris would be right to think that the Constitution cannot be made colorblind if the Supreme Court permitted race-based arrests but denied race-based admissions. But it has never come close to doing this.