Fairly interesting paper...
law.duke.edu
Cited: 61 Law & Contemp. Probs. 153 (Autumn 1998)
[*pg 153]
RACIAL DISPARITY AND THE DEATH PENALTY
JOHN C. MCADAMS*
I. INTRODUCTION
II. THE "MASS MARKET" RACIAL DISPARITY ARGUMENT
III. ASSESSING RACIAL UNFAIRNESS
IV. THE "SPECIALIST" RACIAL DISPARITY ARGUMENT
V. RACE-OF-SUSPECT EFFECTS
VI. THE CONSEQUENCES OF RACISM
VII. CONCLUSION
FOOTNOTES
I. INTRODUCTION
We should, generally, want fairness in all areas of public policy. We should especially want fairness with regard to the death penalty, since the stakes are so high. Yet the opponents of the death penalty make a most peculiar argument about fairness. They argue that if the death penalty is not administered fairly, and especially not administered with racial fairness, it must be abolished.1
Nobody would even think of trying to apply this principle in a consistent way. If we find that black neighborhoods get less police protection than white neighborhoods, would we withdraw cops from both black and white neighborhoods? If banks are discriminating against black home buyers in mortgage lending, would we demand they stop all mortgage lending? If we find the IRS discriminating against middle-class and poor taxpayers, would we want to abolish the IRS?
The first question, of course, is whether the death penalty is administered unfairly. Among right-thinking, politically correct people, the phrase "racial unfairness" is a mere tautology. It is difficult to imagine that anything in American society is administered with racial fairness. But, in fact, the opponents of capital punishment have both a "mass market" version of the racial disparity argument and a "specialist" version,2 and the two versions are flatly contradictory. In what follows, I will first examine the rhetoric and data supporting the "mass market" version of the racial disparity thesis. I will then look at the best studies of racial disparity, which lead to a discussion of the (empirically superior) "specialist" version of the thesis. Next, I will present a new statistical analysis (based on previously published data) that attempts to [*pg 154] sort out various kinds of possible racial bias. Finally, I will consider the normative implications of the empirical findings.
II. THE "MASS MARKET" RACIAL DISPARITY ARGUMENT
The "mass market" racial disparity argument is that the criminal justice system is tougher on black offenders than white offenders, and particularly is more inclined to execute blacks than whites. This argument is simple, palatable, and easy to sell. It flows effortlessly from our generic presumption that black people will be treated unfairly by pretty much any American institution, and from the ease of focusing on one particular class of black people: those accused of crimes. It is also supported by casual empiricism. As Frank Chapman put this issue, "[f]or 48 percent of the death row population in our country to be lack is clearly practicing genocide when you consider that Afro-Americans are only 12 percent of the population."3 Presumably, this disparity is the result of racist prejudice on the part of prosecutors, juries, or the voting public to whom judges and prosecutors are responsible in a democracy. Leigh Bienen argues,
The criminal justice system is controlled and dominated by whites, although the recipients of punishment, including the death penalty, are disproportionately black. The death penalty is a symbol of state control and white control over blacks. Black males who present a threatening and defiant personae are the favorites of those administering the punishment, including the overwhelmingly middle-aged white, male prosecutors who -- in running for election or reelection -- find nothing gets them more votes than demonizing young black men.4
As with any mass market commodity, this one is popular. For example, Jesse Jackson, in a book titled Legal Lynching, claims
[n]umerous researchers have shown conclusively that African American defendants are far more likely to receive the death penalty than are white defendants charged with the same crime. For instance, African Americans make up 25 percent of Alabama's population, yet of Alabama's 117 death row inmates, 43 percent are black. Indeed, 71 percent of the people executed there since the resumption of capital punishment have been black.5
It is also common to treat this issue in the historical context of the Slave Codes, which imposed particularly harsh punishments on blacks and lynchings.6
It is indeed the case that blacks are over-represented on death row. The Sourcebook of Criminal Justice Statistics shows 41.7% of the death row population to be black, and 38.7% of all prisoners executed since 1977 have been [*pg 155] black.7 But then suppose we look a bit further. We come face to face with the reality shown in Table 1.
TABLE 1
MURDERS AND NONNEGLIGENT MANSLAUGHTERS KNOWN TO POLICE
UNITED STATES, 1995
--------------------------------------------------------------------------------
Race of Offender
Race of Victim White Black Total
--------------------------------------------------------------------------------
White 4,124 699 4,823
Black 281 4,422 4,703
--------------------------------------------------------------------------------
Total 4,405 5,121
--------------------------------------------------------------------------------
SOURCE: SOURCEBOOK OF CRIMINAL JUSTICE STATISTICS (1996). Omitted are cases where either the victim or the offender was of the "other" racial classification, or was unknown.
The data in Table 1 shows that forty-eight percent of murder victims are black. It also shows that the vast majority of murders are intraracial not interracial. Among murders involving black and white persons, ninety percent involve a white killing a white or a black killing a black. Almost three-quarters of the rest involve blacks murdering whites, and only a small handful involve whites murdering blacks. Knowing this, the number of blacks on death row and the number of blacks executed do not look far out of line.
To get a solid assessment of bias, however, we need to go beyond eyeballing numbers. We have to control for factors that might legitimately result in more or fewer severe sentences. In arguing that racial bias exists in the system, the opponents of the death penalty have actually cited the fact that blacks who murder whites are treated more harshly than blacks who murder blacks. Unfortunately, the probability of black-on-white murders being comparable to black-on-black murders are about zero.
So the key is to compare how black murderers are treated relative to white murderers who commit comparable murders. Likewise, we must analyze what sort of punishment is meted out to those who murder blacks as compared to those who murder whites under similar circumstances. In the following section, I will review the literature on race and the death penalty, with particular attention to studies that control for the factors that might legitimately result in a more severe sentence -- such as the murder being connected with the commission of a felony, the presence of multiple victims, and a prior conviction for a capital felony.
[*pg 156]
III. ASSESSING RACIAL UNFAIRNESS
In attempting to assess whether racial bias exists in the application of the death penalty, there are about as many methodological problems as in any other policy area of social science research. First, there is the question of what historical era we are studying. The watershed is the Furman v. Georgia8 decision, in which the Supreme Court required that the application of the death penalty meet certain standards of supposed procedural fairness. This decision might not have been a watershed, however, because it is possible that the sources of unfairness in the system were simply not capable of being remedied by the reforms put into place after Furman. Indeed, it is quite possible that such reforms made the system less fair.9 If so, this would not be the first time that reform produced unintended consequences.10 Be this as it may, post-Furman data must be presumed to be a better indicator of the way the system operates today.
Another methodological problem with assessing racial bias in the death penalty system concerns state-to-state and rural-versus-urban differences. An analysis that is disaggregated along these lines may very well show important variation across jurisdictions. Such variation may shed light on important dynamics. For example, David Baldus and his colleagues found that rural and urban areas in Georgia had radically different patterns of racial disparity.11 However, the more one disaggregates the data, the smaller the number of cases in the relevant categories, and the less our chance of establishing facts with any level of statistical confidence. We may quickly find ourselves trying to make sense of variations across jurisdictions that are mere statistical phantoms.
Finally, there is the question of identifying our dependent variable. Because we are talking about the death penalty, it seems sensible to study who has been executed. Unfortunately, it is not so simple. Executions are so few as to make statistical inferences difficult. Ironically, the same reality that makes it difficult to show that capital punishment deters murders also makes it difficult to show racial disparity.12 So it is common in the research to look at death sen- [*pg 157] tences (regardless of whether they were carried out) or the prosecutors' decision to ask for the death penalty (regardless of whether it is received). This methodology achieves plausible results, but they are hardly ironclad. For example, if blacks are more likely to be sentenced to death, but are also more likely to attract post-sentencing counsel from among the ranks of political activist anti-capital punishment lawyers, a racial disparity in sentencing might be counterbalanced by an opposite disparity in the appeals process.
One important study that spanned the pre- and post-Furman eras was conducted by Gary Kleck.13 Kleck looked at the imposition of death sentences from 1967 to 1978 and at actual executions during 1930 to 1967. He found that, outside the South, black homicide offenders have been less likely than white offenders to be sentenced to death or executed. In the South, a pattern of apparent discrimination against black offenders in the early data disappeared in more recent data. Between 1950 and 1967, black offenders and white offenders seemed to be executed at identical rates in the South.14 Lawrence Greenfeld and David Hinners used a mode of analysis similar to Kleck's, but with more recent data. They reported that for each thousand blacks arrested for murder or nonnegligent manslaughter, 11.6 were sentenced to death.15 Among whites, the rate was 15.8 per thousand.16 The relatively lenient treatment of black homicide offenders is striking, but neither of these studies controlled for the race of the victim, which is a critical factor.
Most of the best studies have been state-specific. Sheldon Ekland-Olson, for example, examined data from cases in Texas brought between 1974 and 1983.17 The choice of Texas was significant, since the death penalty statute in Texas allowed less discretion than statutes in other states, such as Florida and Georgia.18
Ekland-Olson showed, first, that the death row population overrepresented offenders who killed strangers and underrepresented offenders who killed acquaintances -- a fact that makes clear the uselessness of comparing interracial murders with intraracial murders,19 and, second, that offenders who killed whites were overrepresented, and those who killed blacks or Hispanics were underrepresented on death row.20 Ekland-Olson also showed -- as we would [*pg 158] expect because the vast majority of murders are intraracial -- that whites were overrepresented on death row.21
Raymond Paternoster, using data from South Carolina on the probability that prosecutors will seek the death penalty in homicide cases, likewise found, even after the imposition of a fairly complete set of control variables, that cases where whites killed whites were much more likely to result in a death penalty request than cases of a black killing a black.22 Interestingly, Paternoster looked for effects of the race of the accused, but failed to find any that showed statistical significance.23
Using data from Florida, William J. Bowers found that blacks who killed whites, and whites who killed whites were substantially more likely to be indicted for and convicted of first-degree murder than blacks who killed blacks.24 The data showed that blacks who killed whites and whites who killed whites were more likely to be sentenced to death, but Bowers could not establish this with the degree of statistical confidence that applied to the other two findings.25 This study is interesting because the evidence of racial disparity in indictment and conviction is better than the evidence of disparity in sentencing. Yet it is only the latter that is addressed by the abolition of the death penalty.
Michael Radelet in his study also used Florida data, looking at 637 homicide indictments. He looked for race-of-victim and race-of-defendant effects on both indictment for first-degree murder (as opposed to a lesser charge) and the imposition of a death sentence. He used a single variable to control for the heinousness of the crime: whether the defendant and the victim knew each other. He made no serious attempt to look for racial disparity in "primary" murder cases (in which the victim and defendant were acquainted) since the death penalty was so rarely imposed in such cases. Among the other cases, in spite of an unfortunately convoluted statistical analysis, he found robust race-of-victim effects for both indictments and sentences.26 However, a study by Linda Foley and Richard Powell that included many of the same cases and had a more complete set of statistical controls for the severity of the case showed no statistically significant race-of-victim effects for prosecutors' decisions to seek the death penalty nor on juries' decisions to levy a death penalty.27 They did show such effects for sentences imposed by trial judges, however.28
[*pg 159]
Gennaro Vito and Thomas Keil studied all persons indicted for murder in Kentucky between December 22, 1976, and October 1, 1986.29 They looked for racial disparity in the decision of a prosecutor to ask for a death-qualified jury.30 They found that offenders who had murdered whites were more likely to face a death-qualified jury than offenders who had murdered blacks.31
Perhaps the most elaborate study was conducted with data from Georgia by David Baldus and his colleagues.32 A key element in this study was the prosecutors' decision to seek the death penalty. After extensive controls, clear race-of-victim effects emerged -- although at the middle levels of aggravation. Specifically, where there were three or four statutory aggravating factors, prosecutors were far more likely to seek the death penalty when the victims were white.33
Not all important death penalty studies come from southern states. For example, Leigh Bienen and her colleagues examined homicides in New Jersey committed during 1982 to 1986, 703 in all.34 As with several other studies, the dependent variable was not execution, nor even a death sentence, but rather a prosecutor's decision to seek the death penalty.35 The study found that a higher percentage of white defendants were charged with capital murder.36 However, this was apparently because white defendants usually had killed white victims, since the victim's race had strong effects on the charge.37 Forty-three percent of defendants who had killed a white were charged with capital murder, as opposed to twenty-eight percent who had killed a black, and nineteen percent who had killed an Hispanic.38 These racial disparities held up in the presence of elaborate statistical controls.39
[*pg 160]
IV. THE "SPECIALIST" RACIAL DISPARITY ARGUMENT
The "specialist" version of the racial disparity argument is that the criminal justice system discriminates against blacks by its lack of concern for black victims. As Anthony Amsterdam has claimed, "although less than 40 percent of Georgia homicide cases involve white victims, in 87 percent of the cases in which a death sentence is imposed, the victim is white. White-victim cases are almost eleven times more likely to produce a death sentence than are black-victim cases."40 Randall L. Kennedy, describing the same Baldus study, lamented what he described as the fact that "in Georgia's marketplace of emotion the lives of blacks simply count for less that the lives of whites."41
In a considerable display of chutzpah, the opponents of capital punishment, who first adopted an offender-centered concept of justice (when it was convenient to argue that too many blacks were being executed), simply turned on a dime and adopted a victim-centered concept of justice (when the data in fact showed that too many whites are being executed). Had they maintained philosophical consistency, the opponents might have lauded the current situation on the ground that a sort of affirmative action was being employed. That is, against an historic background of harsh treatment of black people in most aspects of public policy, black offenders are getting favorable treatment from the criminal justice system. Instead, the opponents decided that victims matter after all.
Even worse is the opponents' tendency to adopt vague rhetoric about the evil of racism in an attempt to cover up the fact that they have turned on a dime. It is also troubling to find articles from authors who clearly have data that might show discrimination based on the race of the offender, but who do not present it and choose instead to discuss the race of the victim. For example, Sam Gross and Robert Mauro present a table containing the following data:42
[*pg 161]
TABLE 2
PERCENTAGE OF DEATH SENTENCES BY RACE OF SUSPECT
AND VICTIM, AND FELONY CIRCUMSTANCES, FLORIDA
--------------------------------------------------------------------------------
Felony
Circumstances Nonfelony
Circumstances
--------------------------------------------------------------------------------
Black kills white 28.8% 2.5%
32/111 2/79
White kills white 26.9% 1.4%
63/234 17/1187
Black kills black 6.0% 0.3%
7/116 4/1414
White kills black 18.2% 1.9%
2/11 1/53
White victim 27.5% 1.5%
95/346 19/1272
Black victim 7.0% 0.3%
9/128 5/1468
--------------------------------------------------------------------------------
NOTE: Among the states analyzed in the body of the book, I have chosen to report their data on Florida because it had the largest absolute number of death sentences. Other states showed a similar pattern.
Quite clearly, murders with a white victim were more likely to lead to a death sentence than murders with a black victim.
But what do Gross and Mauro not show the reader? I have derived the following table from their data:
TABLE 3
PERCENTAGE OF DEATH SENTENCES BY RACE OF SUSPECT
AND FELONY CIRCUMSTANCES, FLORIDA
--------------------------------------------------------------------------------
Felony Circumstances Nonfelony Circumstances
--------------------------------------------------------------------------------
White Suspect 26.5% 1.4%
65/245 18/1240
Black Suspect 17.2% 0.4%
39/227 6/1493
--------------------------------------------------------------------------------
The data are striking. White suspects are treated more harshly than black suspects, and this is especially true under nonfelony circumstances. So just what is going on here? Given the studies we have reviewed showing that the system is lenient on those who kill blacks, and knowing that the vast majority of murders are intraracial, we might conclude that white suspects get tougher treatment because they have overwhelmingly killed whites. This conclusion [*pg 162] would be correct, but then we must ask whether the race of the suspect has any effect after we have controlled for the race of the victim.
V. RACE-OF-SUSPECT EFFECTS
The published literature shows some race-of-suspect effects after controlling for the race of the victim, but they are few and lack statistical robustness. Among the studies we discussed above, Bowers reported no race-of-defendant effects.43 Radelet looked for race-of-defendant effects, and found none neither for first-degree murder indictments, nor for death sentences.44 Likewise, Foley and Powell found none for prosecutors' decisions to seek the death penalty, nor for juries' decisions to levy the death penalty, nor for the actual sentences imposed by trial judges.45 Paternoster, as we have already noted, found no significant race-of-suspect effects.46
Bienen and her colleagues found no race-of-defendant effects with regard to the prosecutors' decision to plea bargain rather than go to trial. The pattern of effects for the prosecutors' decisions to "serve a notice of factors" (claim aggravating factors) is complicated, but there seems to be no net race-of-defendant effect.47 Gross and Mauro claim to have found statistically significant race-of-suspect effects in Illinois.48 Unfortunately, they had no prior expectation that such effects would occur in Illinois (but not in the other seven states in their sample). So this looks like yet another example of the fact that, if you estimate a large number of coefficients, some will turn up "significant."
Given the appeal of the "mass market" version of racial unfairness in the system, many people will have difficulty accepting that there is no bias against black defendants.
In an attempt to answer this difficulty, we can use Gross and Mauro's data and construct a statistical model of death sentences. The unit of analysis in the Gross and Mauro data is homicides, and the dependent variable is a dichotomy based on whether a particular homicide resulted in a death penalty. Note that this variable aggregates over a whole series of decision points: Is the suspect charged? Is he charged with first-degree murder? Does the prosecutor ask for the death penalty? Does the jury convict and levy a death sentence? This approach has the virtue of producing a "bottom line" assessment of bias in the process, with the proviso that actual executions are not addressed. Gross and Mauro present data from eight states with a relatively large number of homicides,49 and include all homicides reported to the FBI from 1976 to 1980. The [*pg 163] independent variables used here include both the race of the victim and the suspect, and a key control variable, the number of aggravating circumstances associated with the crime. Aggravating circumstances can include whether (1) the murder was associated with another felony, (2) the victim was a stranger, and (3) there was more than one victim. Gross and Mauro found this control variable to be a powerful predictor of the likelihood of the imposition of a death sentence.
TABLE 4
DEATH SENTENCES IN HOMICIDE CASES LOGISTIC REGRESSION
(standard errors in parentheses)
--------------------------------------------------------------------------------
Equation
Independent Variable 1 2 3 4
--------------------------------------------------------------------------------
Aggravating
Circumstances 1.75
(.065) 1.60
(.066) 1.59
(.068) 1.58
(.069)
Black Suspect -.714
(.113) .066
(.134)
Black Victim -1.56
(.154) -1.60
(.176) -1.61
(.169)
Black on White
Murder .144
(.136)
White on Black
Murder .937
(.419)
Constant -4.80 -4.54 -4.54 -4.55
-2 log Likelihood 3469.8 3469.9 3469.9 3469.8
--------------------------------------------------------------------------------
The first equation models death sentences as a function of aggravating circumstances and the race of the suspect. It shows a powerful and statistically robust effect of the number of aggravating circumstances. It also shows that, after the aggravation factor is taken into account, black defendants are treated more leniently than are white defendants. This is the same reality shown in Table 3 above. The second equation models death sentences as a function of aggravating circumstances and the race of the victim. Again, aggravating circumstances show a large and robust effect, and now the race of the victim shows a large effect. A white victim has an effect almost precisely equal to the effect of one aggravating circumstance.
Suppose we take both the race of the victim and the suspect into account? These two variables are so highly correlated that we would suspect that coline- [*pg 164] arity between them would give us mush for estimates. Yet, in fact, the third equation in the table is a pleasant surprise. Our standard errors have increased only moderately. Our estimates show that the race of the victim and not the race of the suspect matters. Our standard error for the race of the suspect is small enough that any large effect would be statistically significant. Again, murdering a white has an effect almost identical to one additional aggravating circumstance. That is, killing one white person is treated about the same as killing multiple black people, and killing a white nonstranger is treated about the same as killing a black stranger.
A final question remains about this data: Is there an interaction between the race of the victim and the race of the suspect? Given that people who murder whites are treated more harshly than people who murder blacks, are blacks who murder whites treated more harshly than whites who murder whites? And is there a racial disparity in the treatment of people who murder blacks?
Equation 4 deals with these issues. The interaction of race-of-victim and race-of-suspect are modeled with a series of dummy variables. White-on-white murders are the excluded reference category. Black-on-white murders seem to be treated no differently than white-on-white murders. In other words, net of aggravating circumstances, blacks and whites are treated about equally if they kill a white. As expected, people who murder blacks are treated much less harshly than people who murder whites. However, our white-on-black variable estimates the difference between the treatment received by blacks who murder blacks and whites who murder blacks. It seems that whites who murder blacks are treated more harshly than blacks who murder blacks. Does this suggest an anti-white racial bias in the system? Maybe and maybe not. It could be that most such murders happen in urban areas, and that biases of blacks on juries are the key factor. Or it might be that white-on-black murders are different in some way that our scale of aggravating circumstances does not pick up. Finally, white-on-black murders may be viewed as "hate crimes" and therefore particularly dangerous to society.
Some idea of the magnitude of these effects can be found in Table 5. There the estimates from Table 4 are used to generate probabilities that homicides that fall into various categories will result in a death sentence. As we might expect from the estimates, a black-on-black murder with no aggravating factors is extremely unlikely to result in a death sentence. About two per thousand of such murders produce this result. On the other hand, while a white who murders a white is about five times as likely to get the death penalty, it is still unlikely. Among both blacks and whites, aggravating factors radically increase the probability of a death sentence, but all three aggravating factors are required to work against a defendant (an exceedingly unlikely event) in order to increase the probability of a death sentence in a black-on-black murder to the level of a white-on-white murder with only two aggravating factors. If only two aggravating factors are present in a black-on-black murder, the probability of a death sentence is only about five percent.
[*pg 165]
TABLE 5
PROBABILITY THAT A HOMICIDE WILL RESULT IN
A DEATH SENTENCE
--------------------------------------------------------------------------------
Race of Victim Race of Suspect Number of Aggravating Factors Probability of a Death Sentence
--------------------------------------------------------------------------------
Black Black 0 .0021
Black Black 1 .0102
Black Black 2 .0476
Black Black 3 .1955
Black White 0 .0054
White White 0 .0105
White White 2 .1996
--------------------------------------------------------------------------------
This rather simplistic model is hardly the last word on this issue, but it does have very real virtues. It includes data on over 14,000 murders, separates race-of-victim and race-of-suspect effects, and uses a robust measure of the severity of the crime for a control.
Thus, the "specialist" version of the racial disparity argument appears to be correct. The data strongly suggest that the criminal justice system undervalues the lives of black victims. The findings presented here are consistent with those of Baldus and his colleagues who used extensive controls for the nature of the murder and the number of aggravating factors present, and still found murderers of white victims to be more likely to be sentenced to death.50 Baldus is the preeminent scholar in this area, and he recently described the current state of the literature as follows:
[W]hat do the data tell us about differences in discrimination in the pre- and post-Furman periods? There are significant differences in race effects, both across and within states. There are differences in the magnitude of race effects at different decision making levels in the states-i.e., prosecutorial decisions to seek the death penalty and jury decisions to impose death. There are also differences that correlate with culpability. The risk of race effects was very low in the most aggravated capital cases; however, in the mid-range cases, where the "correct" sentence was less clear; and the room for exercise of discretion much broader, the race disparities are much stronger. Whereas the overall average disparity for the two groups (Black v. White) tends to be six to eight percentage points, in the mid-range cases the disparities are typically two to three times that large [twelve to twenty-four percentage points].51
Baldus then goes on to describe some reactions to his findings:
There is much anecdotal evidence from lawyers who represent capital defendants. Many of them seriously question the validity of statistical studies that do not reveal disparities based upon the race of the defendant. It is possible that there is such dis- [*pg 166] crimination, but that it is not sufficiently large and systematic to be picked up by the data.52
Baldus, perhaps out of politeness, does not note that lawyers are in the business of producing "anecdotal evidence" to support their clients' position, and that those who represent capital defendants are a highly self-selected and hardly unbiased group.
I believe that, based on the analysis above, stronger statements than those of Baldus are warranted. There is a general and quite robust bias against black victims, and there is no general bias against black defendants.53
VI. THE CONSEQUENCES OF RACISM
So is the system racist? Yes, but we need to be extremely careful, because nothing is more dangerous in politically correct sectors of society than loose talk about racism. Such talk is a balm, spreading a warm glow of self-righteousness over those who use it. And it is an opiate, stifling critical thought. It is striking that the most straightforward hypothesis about racism -- the "mass market" version -- simply is not true. Why a bias against black victims -- and potential victims -- and not black offenders? It is the latter, after all, who are caught in the maw of the system.
Scholars and activists have devoted a huge amount of attention to the notion that juries may be racist, and therefore particularly harsh toward black defendants.54 Are we now going to posit that, lacking empathy for black victims, they are less willing to punish black defendants? This seems implausible. The jury has no direct control over the fate of any victim. Imagine a jury consisting of members of the Ku Klux Klan. They doubtless have little empathy for the black victim, but presumably even less for a black murderer. So why treat a black offender, who is within their power, leniently? Of course, one could posit racist jurors who are malevolent, extremely shrewd and insightful, and understand that in the long run more damage can be done to the black community by letting black criminals off easy. I think this is insufficiently plausible to merit further discussion.
However, the notion that prosecutors and judges are less willing to expend the scarce resources of the criminal justice system to convict and execute the murderers of blacks is all too plausible. In fact, it is clearly the case. Although this situation certainly needs to be remedied, it is not clear how abolishing the death penalty is any part of the remedy. Merely doing away with the death penalty does not guarantee that everybody will be treated fairly in murder cases. It merely guarantees that they cannot be executed.
[*pg 167]
While we seek to remedy inequities in the American criminal justice system, we would do well to remember the words of Ernest van den Haag:
Guilt is individual. If guilty whites or wealthy people escape the gallows and guilty poor people do not, the poor or black do not become less guilty because the others escaped their deserved punishment. Whether due to willful discrimination, capriciousness, or unavoidable accidental circumstances, some people will always get away with murder. Is that a reason to deny the justice of the punishment of those guilty persons who did not get away? Their guilt is not diminished by the escape of the others, nor do they deserve less punishment because others did not get the punishment they deserve. Justice involves punishment according to what is deserved by the crime and the guilt of the criminal -- regardless of whether others guilty of the same crime escape.55
The opponents of capital punishment are pushing an argument they would never use in other contexts. The fact that you parked illegally and did not get a ticket does not relieve me of paying the fine when I do get a parking ticket. The fact that you bamboozled the IRS does not make it acceptable for me to cheat on my taxes, and it certainly does not require the abolition of the IRS.
VII. CONCLUSION
Given that there is a real racial disparity here, how do we remedy it? Note that if one is already against the death penalty, the obvious remedy is abolition. On the other hand, applying the death penalty fairly is just as much a remedy. Interestingly, polls show a clear majority of black people favoring capital punishment. One Gallup Poll, for example, found that fifty-nine percent of blacks nationwide favored the death penalty, and that only thirty-one percent opposed it.56 In another study, Linda Lichter commissioned a nationwide poll of blacks that contained 600 respondents, more than in any poll that fails to oversample blacks. Among those respondents with an opinion, fifty-five percent favored the death penalty and forty-five percent opposed it.57 Even sixty percent of the upscale readers of Ebony said capital punishment is "the only just punishment" for murder.58
Now of course, a majority of black Americans can just as easily be in the wrong as a majority of white Americans. These data do show, however, the hollowness of the claim of the anti-death penalty forces to speak for black Americans.
In saying that we should seek to eliminate inequity in the application of the death penalty, I am not asserting that such a thing will be easy, or even possible [*pg 168] in the short or intermediate term. That inequity in the application of the death penalty requires its abolition is an argument that will be made only by people who do not like it for other reasons. The inequity argument is a pretext, not a reason. Between an inequitable death penalty and no death penalty, I would prefer an inequitable death penalty, just as I would prefer an inequitable tax system to no tax system, and inequitable policing to no policing. I say this because I think it likely that executions deter murders,59 and it is clear that a majority of Americans -- white and black -- think that justice requires executions for the most heinous crimes. If someone says that capital punishment does not deter murder and that justice does not require it for the most heinous crimes, my response is that those are the real reasons they oppose the death penalty -- not racial inequity. Just as black neighborhoods would be made no better off by the withdrawal of police from white neighborhoods, black victims and potential victims would be made no better off by our refusal to execute murderers, notwithstanding that we are more likely to execute murderers of whites.
Indeed, there is little reason to believe that a system that cannot equitably execute those who murder blacks will somehow impose equitable noncapital punishments. What people pushing abolition of the death penalty must be hoping for is that abolition will reduce the level of punishment meted out to (mostly white) murderers of whites, while leaving intact (save for the elimination of a rare execution) the level of punishment meted out to murderers of blacks. More likely, what we would have is a general ratcheting down of punishments, with the inequities remaining.
Scholars who study the death penalty often also study several other decisions in the process that might theoretically lead to execution. What they almost invariably find is large-scale bias in these earlier decisions, including decisions that would continue to be made if the death penalty were abolished. One particularly interesting study (although pre-Furman) was conducted by Franklin Zimring, Joel Eigen, and Sheila O'Malley, and examined the records of 245 persons arrested for homicide in Philadelphia in 1970.60 Of the 245, 170 were eventually convicted of some charge.61 Sixty-five percent of the defendants who killed a white person received either life imprisonment or a death sentence, [*pg 169] while only twenty-five percent of those who killed a black person did.62 Since these murders produced only three death sentences (all imposed on blacks who killed whites),63 most of the apparent racial unfairness involved life imprisonment. Alfred Blumstein, in a study of the racial disproportionality of prison populations, found that in 1991, blacks were underrepresented among prisoners convicted of murder.64 There were many limitations to Blumstein's study, including failure to control for aggravating circumstances, and a research design that leaves possible racial discrimination in arrests entirely out of account. But his results strongly imply that the system does for imprisonment what it does with regard to executions: underpunish those who kill blacks.
William J. Bowers, as we have already discussed, found that defendants who killed whites were more likely to be indicted for first-degree murder, rather than a lesser charge, and more likely to be convicted of first-degree murder than defendants who killed blacks.65 Along similar lines, Michael Radelet, in a study of indictments for murder in Florida, found that eighty-five percent of the killers of white victims were indicted for first-degree murder, while only 53.6 percent of the killers of black victims were.66
Leigh Bienen and her colleagues, in their study of New Jersey homicides discussed above,67 examined the issue of whether a particular case was plea bargained or went to trial. Cases involving white victims were found to have gone to trial more often than cases involving either black or Hispanic victims.68
One particularly interesting study involved prosecutors' decisions to upgrade or downgrade a homicide charge.69 An upgrade involved a prosecutor charging a felony connected with the homicide when no such felony was mentioned in the police report. A downgrade involved a prosecutor not charging a felony even though the police report indicated its commission. A statistical model that controlled for the circumstances of the crime and of the offender showed that murders involving white victims were more likely to be upgraded than murders involving black victims.70
All of this makes perfect sense. If the system is biased toward punishing those who murder whites, it is implausible indeed that decisions leading up to sentencing are made with strict racial fairness, and only the imposition of a death sentence is racially biased. If people want to punish those who murder whites more harshly than those who murder blacks, this is likely to be reflected in prosecutors' decisions to move ahead with a case, in decisions about whether [*pg 170] to plea-bargain, in the allocation of staff to a particular case, in the decision to indict on more or less serious charges, and in jury verdicts. Even in sentencing, abolition of the death penalty does not eliminate racial unfairness; it only narrows the range of possible punishments. While not all decision points have been studied equally well, theoretically the pervasive undervaluing of the lives of black victims would be expected to be reflected at every point that involves the exercise of discretion.
Interestingly, a very large literature testifies to the racial unfairness of the criminal justice system below the level of capital crimes and capital punishment, although a lot of this literature supports the "mass market" version of the disparity argument. For example, Joseph Gastwirth and Tapan Nayak have presented strong evidence that drug offenders who are black are more likely to be sentenced to life imprisonment than similar white offenders.71 One interpretation of this, however, is that the system is in fact being responsive to the black community, which feels threatened by drugs in a way that few if any white localities do. The evidence, however, is mixed. Blumstein, in the study discussed above, compares the prison population to the population of arrestees, on the assumption that any disproportion will reflect treatment within the criminal justice system. He finds blacks and whites treated equally in robbery cases, blacks treated more leniently in murder cases (as we have already seen), and blacks treated more harshly in other kinds of cases, especially in drug cases.72 The disproportionate treatment of blacks in some kinds of cases might be racial discrimination or instead might be the result of legitimate factors not controlled for.73
The relevance of studies dealing with offenses below the level of murder is questionable, however. Eliminating the death penalty does not really make murder just like robbery, assault, or drug dealing. So the best evidence comes from studies explicitly dealing with how murderers are treated. Studies of racial unfairness with regard to lesser crimes only serve to make us cautious about assuming that the system is unfair only with regard to executions.
Thus, we have a criminal justice system that punishes those who murder whites more severely that those who murder blacks. Abolishing the death penalty would eliminate the most visible outcropping of injustice, and perhaps make us feel better, but it would not aid the law-abiding blacks who are the chief victims of the system. Most importantly, it would not address, in any substantial way, the total structure of unfairness.
FOOTNOTES
Copyright © 1998 by Law and Contemporary Problems
This article is also available at law.duke.edu. * Associate Professor of Political Science, Marquette University.
1. See, e.g., Callins v. Collins, 510 U.S. 1141, 1143 (1994) (Blackmun, J., dissenting from denial of certiorari to Callins v. Collins, 998 F.2d 269 (5th Cir. 1993)); Furman v. Georgia, 408 U.S. 238, 257 (1972) (Brennan, J., concurring); id. at 314 (Marshall, J., concurring); American Bar Ass'n, Resolution of the House of Delegates (Feb. 1997), reprinted in Appendix, 61 LAW & CONTEMP. PROBS. 219 (Autumn 1998); Leslie Harris, The ABA Calls for a Moratorium on the Death Penalty: The Task Ahead-Reconciling Justice with Politics, FOCUS L. STUD., Spring 1997, at 2; Christine Wiseman, Representing the Condemned: A Critique of Capital Punishment, 79 MARQUETTE L. REV. 731 (1996).
2. See John McAdams, Wisconsin Should Adopt the Death Penalty, 79 MARQUETTE L. REV. 719 (1996).
3. Frank Chapman, The Death Penalty, USA: Racist and Class Violence, POL. AFF., July 1987, at 17, 19.
4. The Death Penalty: A Scholarly Forum, FOCUS L. STUD., Spring 1997, at 1, 7 (remarks of Leigh Bienen).
5. JESSE JACKSON, LEGAL LYNCHING: RACISM, INJUSTICE, AND THE DEATH PENALTY 100 (1996). However, Jackson also admits that "the relationship between race and capital punishment is much more complex than most people suppose." Id.
6. See, e.g., RAYMOND PATERNOSTER, CAPITAL PUNISHMENT IN AMERICA 116-19 (1991).
7. See SOURCEBOOK OF CRIMINAL JUSTICE STATISTICS 556 (1996); see also U.S. DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS BULLETIN-CAPITAL PUNISHMENT 1995, at 11 (1996) (listing statistics on executions and persons under sentence of death in 1995).
8. 408 U.S. 238 (1972).
9. One reform that looks particularly suspect is the use of a bifurcated trial, with the jury deciding the sentence in what amounts to a second trial. These proceedings often degenerate into blatant attempts to manipulate the emotions of the jury.
10. For example, the imposition of tough mandatory sentences for certain offenses has caused offenders to demand trials rather than plea-bargain (New York's "Rockefeller drug laws"), nearly doubling the time necessary to dispose of an average case. In Michigan, a tough bill imposing a mandatory sentence for possessing a firearm while committing a felony led to judges reducing the original sentence for the original felony to compensate for the mandatory firearm possession sentence. See JAMES Q. WILSON, THINKING ABOUT CRIME 133-36 (rev. ed., Vintage Books 1985) (1975).
11. See infra note 32 and accompanying text.
12. Arnold Barnett, for example, reviewed Passell's 1975 study which had found no deterrent effect. See Arnold Barnett, The Deterrent Effect of Capital Punishment: A Test of Some Recent Studies, 29 OPERATIONS RES. 346, 353 (1981). Barnett found that the model estimated by Passell had a prediction error-in terms of the number of homicides-of 1,635. See id. at 354. But there were only 44 persons executed for murder in 1960. See SOURCEBOOK OF CRIMINAL JUSTICE STATISTICS 462 (1973). If each of these executions had deterred five murders, which would be an excellent "return" in terms of deterrence, the reduction of 220 murders would not have been discernible in Passell's data within accepted standards of statistical reliability. See id. at 355.
13. See Gary Kleck, Racial Discrimination in Criminal Sentencing: A Critical Evaluation of the Evidence with Additonal Evidence on the Death Penalty, 46 AM. SOC. REV. 783 (1981).
14. See id. at 798-99.
15. See Lawrence Greenfeld & David Hinners, Capital Punishment 1984, U.S. DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS BULLETIN 1, 9 (1985).
16. See id.
17. See Sheldon Ekland-Olson, Structured Discretion, Racial Bias, and the Death Penalty: The First Decade After Furman in Texas, 69 SOC. SCI. Q. 853, 854 (1988). It is important to note that Ekland-Olson studied death sentences rather than actual executions.
18. See id.
19. See id. at 861.
20. See id.
21. See id.
22. See Raymond Paternoster, Race of Victim and Location of Crime: The Decision to Seek the Death Penalty in South Carolina, 74 J. CRIM. L. & CRIMINOLOGY 754, 767 (1983).
23. See id. at 766.
24. See William J. Bowers, The Pervasiveness of Arbitrariness and Discrimination Under Post-Furman Capital Statutes, 74 J. CRIM. L. & CRIMINOLOGY 1067, 1074 (1983).
25. See id. at 1085.
26. See Michael L. Radelet, Racial Characteristics and the Imposition of the Death Penalty, 46 AM. SOC. REV. 918-27 (1981).
27. See Linda A. Foley & Richard S. Powell, The Discretion of Prosecutors, Judges, and Juries in Capital Cases, 7 CRIM. JUST. REV. 16 (1982).
28. See id. at 21.
29. See Gennaro F. Vito & Thomas J. Keil, Capital Sentencing in Kentucky: An Analysis of the Factors Influencing Decision Making in the Post-Gregg Period, 79 J. CRIM. L. & CRIMINOLOGY 483 (1988).
30. A "death qualified jury" is one in which each juror has been told that the death penalty is possible in the case and that has not demonstrated that it is entirely unwilling to impose a death sentence. See Vito & Keil, supra note 29, at 493; see also WELSH S. WHITE, THE DEATH PENALTY IN THE EIGHTIES ch. 8 (1987).
31. See Vito & Keil, supra note 29, at 498-99.
32. See David C. Baldus et al., Comparative Review of Death Sentences: An Empirical Study of the Georgia Experience, 75 J. CRIM. L. & CRIMINOLOGY 661 (1983) [hereinafter Baldus, Comparative Review]; see also DAVID C. BALDUS ET AL., EQUAL JUSTICE AND THE DEATH PENALTY (1990).
33. See Baldus, Comparative Review, supra note 32, at 709.
34. See Leigh Bienen et al., The Reimposition of Capital Punishment in New Jersey: The Role of Prosecutorial Discretion, 41 RUTGERS L. REV. 27, 37 (1988).
35. See id. at 165.
36. See id. at 170.
37. See id. at 192.
38. See id. at 172.
39. See id. at 226, 235.
40. Anthony G. Amsterdam, Race and the Death Penalty, 7 CRIM. JUSTICE ETHICS 2, 84 (1988).
41. Randall L. Kennedy, McCleskey v. Kemp: Race, Capital Punishment, and the Supreme Court, 101 HARV. L. REV. 1388 (1988).
42. See SAMUEL R. GROSS & ROBERT MAURO, DEATH & DISCRIMINATION: RACIAL DISPARITIES IN CAPITAL SENTENCING 47 (1989).
43. See Bowers, supra note 24, at 1073, 1079, 1084.
44. See Radelet, supra note 26, at 923-24.
45. See Foley & Powell, supra note 27, at 21.
46. See Paternoster, supra note 22, at 770.
47. See Bienen et al., supra note 34, at 226, 235.
48. See GROSS & MAURO, supra note 42, at 68-69.
49. Arkansas, Florida, Georgia, Illinois, Mississippi, North Carolina, Oklahoma, and Virginia.
50. See Baldus, Comparative Review, supra note 32, at 708-10.
51. The Death Penalty, supra note 4, at 7 (remarks of David Baldus).
52. Id. at 8.
53. Saying "no general bias" does not deny that there might be bias in some places, under some circumstances.
54. See, e.g., GREGORY D. RUSSELL, THE DEATH PENALTY AND RACIAL BIAS ch. 5 (1994).
55. Ernest van den Haag, Murderers Deserve the Death Penalty, WORLD & I, Nov. 1989, at 189, 193.
56. See Alec Gallup & Frank Newport, Death Penalty Support Remains Strong, GALLUP POLL MONTHLY, June 1991, at 40, 43. This was true even though 73% of black respondents agreed that "[a] black person is more likely than a white person to receive the death penalty for the same crime." Id. at 45. This shows the popular appeal of the "mass market" view of racial disparity discussed above.
57. See Linda S. Lichter, Who Speaks for Black America?, PUB. OP., Aug.-Sept. 1985, at 41, 43.
58. Ebony's Annual Readers Poll, EBONY, Nov. 1994, at 43. This poll was far from scientific, but it is interesting both for the large number of respondents (9,657) and for their relatively affluent nature.
59. It is not the case that all or even a majority of studies show a deterrent effect of executions. But several studies, including some of the most sophisticated ones, do. See, e.g., Dale O. Cloninger, Capital Punishment and Deterrence: A Portfolio Approach, 24 J. APPLIED ECON. 635 (1992); Isaac Ehrlich, Capital Punishment and Deterrence: Some Further Thoughts and Additional Evidence, 85 J. POL. ECON. 741 (1977); Isaac Ehrlich, The Deterrent Effect of Capital Punishment: A Question of Life and Death, 65 AM. ECON. REV. 397 (1975); Stephen K. Layson, Homicide and Deterrence: A Reexamination of the United States Time-Series Evidence, 52 S. ECON. J. 68 (1985); David Lester, Executions as a Deterrent to Homicides, 44 PSYCH. REP. 562 (1979); David Lester, Deterring Effect of Executions on a Murder as a Function of Number and Proportion of Executions, 45 PSYCH. REP. 598 (1979); Radelet, supra note 26; Kenneth I. Wolpin, Capital Punishment and Homicide in England: A Summary of Results, 68 AM. ECON. REV. 422 (1978); James A. Yunker, Testing the Deterrent Effect of Capital Punishment, 19 CRIMINOLOGY 626 (1982).
60. See Franklin E. Zimring et al., Punishing Homicide in Philadelphia: Perspectives on the Death Penalty, 43 U. CHI. L. REV. 227, 229 (1976).
61. See id. at 229.
62. See id. at 232.
63. See id. at 233.
64. See Alfred Blumstein, Racial Disproportionality of U.S. Prison Populations Revisited, 64 U. COLO. L. REV. 743, 751 (1993). Note that Blumstein's 1979 data did not show this disproportionality.
65. See Bowers, supra note 24, at 1073, 1084.
66. See Radelet, supra note 26, at 922.
67. See Bienen et al., supra note 34.
68. See id. at 226.
69. See Michael L. Radelet & Glenn L. Pierce, Race and Prosecutorial Discretion in Homicide Cases, 19 L. & SOC'Y REV. 587, 599-605 (1985).
70. See id. at 608.
71. See Joseph L. Gastwirth & Tapan Nayak, Statistical Aspects of Cases Concerning Racial Discrimination in Drug Sentencing: Stephens v. State and U.S. v. Armstrong, 87 J. CRIM. L. & CRIMINOLOGY 583 (1997).
72. See Blumstein, supra note 64, at 747, 751.
73. See id. at 746. |
