Racism and the Death Penalty

Opponents of the death penalty often claim that the death penalty is implemented in a racist manner with the result being that blacks are overrepresented on death row. Well, sort of. Let's take a look at the numbers.

In 2007, about 42% of the prisoners on death row were black, while 56% were white*. Of the 1088 prisoners actually executed since 1984, 374, or about 34%, were black. About 1/8 of the US population is black, so relative to their representation in the general population, blacks are definitely overrepresented on death row, and in executions.

But that's a silly basis for comparison. The question isn't whether a random black person is more likely to be executed than a random white person, but whether a black murderer is more likely to be executed or sentenced to death than a white murderer. Between 1976 and 2005, 52% of all murders were committed by black offenders. So it turns out that blacks are actually underrepresented on death row relative to the rate at which they commit murder.

Granted, there could be other reasons besides an unbiased criminal justice system for the underrepresentation of black murderers on death row. It could be that black murderers are more likely to be tried by predominantly black juries, who may be less inclined than predominantly white juries to impose the death penalty. It could be that whites are more likely to commit the types of murders that invite death sentences. It could be that juries find black victims less sympathetic, perhaps because of racism, or perhaps because they really are less sympathetic (e.g., gang members). But it clearly is not true that blacks are overrepresented on death row in any relevant sense.

Besides, it's not clear to me why biased application of the death penalty should be considered a valid argument against it. If one accepts that executing murderers is a good thing in principle, then who cares that it's more likely to be applied to murderers of one race or another?

Another interesting point: The death penalty really is applied in a profoundly sexist manner. 11% of murders committed from 1976-2005 were committed by women, but only 54 (less than 2%) of the more than 3,000 prisoners currently on death row are women, and only 11 women have been executed since 1976, accounting for just over 1% of all executions. But when was the last time you heard a leftist complaining about the sexist application of the death penalty?

*For all statistics in this post, "white" includes Hispanic mestizos. I haven't been able to find good statistics broken down by race and Hispanic origin.

Share this


Excellent analysis, thank you.

His maths a little off

Did you take a look at his link in the following sentence?

He says: "Of the 1088 prisoners actually executed since 1984, 374, or about 34%, were black."

The numbers are 1,099, and 376. So it's 34.2%.

your reading is off

thats why it says about and not exactly 34%

Percentage of Blacks in America

You are wrong on just about every count. Blacks are not 1/8 of the American population. African-Americans make up 14% of the American population. And the study by professor Baldus is factual. The court system is indeed bias against Blacks.

Between 1976 and 2005, 52%

Between 1976 and 2005, 52% of all murders were committed by black offenders.

Actually, the data you cite doesn't show this at all. It shows that 52% of the people convicted of murder are black. Those are two very different sorts of things, unless of course you believe that everyone convicted of murder is actually a murderer (in which case, I'd invite you to take a closer look at Cory Maye). A huge part of the entire objection is that blacks are more likely than whites to be convicted on shoddy evidence in the first place. IOW, the objection is that race plays a factor in both your numerator and your denominator. Your analysis looks at only one of those factors while blithely assuming that the other is perfectly kosher.

Even leaving that worry aside, the issue is far more complicated than you make it out to be. Merely showing that black murder convicts are sentenced to death at a lower rate than white murder convicts are sentenced to death is really kind of irrelevant. If, for example, white (or mostly white) juries sentence blacks to death at a higher rate than they sentence whites to death, then that would be evidence of race playing a factor. And that could be true even if the overall rate of death penalty sentencing for blacks were lower than that of whites.

I don't claim to know the answer to this particular question. It may well be that there is no systematic racial bias in death penalty sentences. But that's not something that one can prove or disprove with 30 minutes of Googling crime statistics. It would require a pretty detailed analysis of actual jury behavior. And, to be really safe, it would probably require performing said analysis on a state-by-state basis.

The deeper point is that even if it is not happening across the board, there is pretty decent evidence that, at least in some trials in some parts of the country, race does play a factor in sentencing decisions. (Again, see Cory Maye.) Even if that it is not true in all cases, the fact that it is true in some cases makes the actual application of the death penalty morally problematic.

You can infer from victim statistics

While any conviction always carries with it some degree of uncertainty, the identity and therefore the race of the victim is much less uncertain. This can be used to infer the disparity in crime rates among black and whites.

As pointed out here,

Detailed analysis of victimization survey data indicates that violent crime in the United States (robbery, assault, and rape) is intraracial from three perspectives (whites chose other whites as victims, whites were largely victimized by other whites, and blacks were largely victimized by other blacks). However, black offenders were more likely to choose white victims in robberies, assaults, and rapes.

Briefly, whites victimize whites, blacks victimize blacks, blacks victimize whites a little bit, and whites victimize blacks very little at all. So we might reasonably expect the proportion of black criminals to be roughly similar to the proportion of black victims - and if anything, we might expect the proportion of black criminals to be slighly higher than the proportion of black victims because of black-on-white crime.

I grant that the above statistics are not for murder per se. Feel free to dig up murder statistics. Tentatively, though, I think it reasonable to suppose that murder follows the same pattern, and possibly even to a greater degree because of the special nature of murder (the murderer typically gains from murder as a result of a pre-existing relationship with the victim, as opposed to robbery, where the money gained can be spent regardless of whether there was a pre-existing relationship with the victim).

Anyway, here are the victim statistics:

According to the FBI's Uniform Crime Reports, in 2006 about 50% off murder victims were black, 47% were white, and 3% were Asians, Pacific Islander, and Native Americans.

So it is about an equal number of black and white victims. The population stats:

White alone: 74%

Black or African American alone: 13.4%

The ratio of black to white 13.4/74, is .18. Since the victims are about equal in number, the ratio of black murder to white murder is probably somewhere in the vicinity of 1/.18, or 5.5 - that is to say, we might reasonably estimate that blacks are about five times as likely to murder someone as whites. And - keep this in mind - I'm deriving this indirectly from the direct calculation that about blacks are about five times as likely to be murdered as whites.

I realize that my analysis can be criticized for assuming what it purports to prove - after all, in order to calculate the degree to which crime is intra-racial versus inter-racial you need to make an assumption about the race of the criminal. However, the numbers I found were pulled from crimes with surviving victims, and it seems pretty unlikely for the victim to mistake the race of the attacker. Of course, whether this means the data are reliable depends on how, exactly, the data were gathered.

Fair enough

This is an interesting take on the statistics. But, I think, it still doesn't get Brandon off the hook. Here's just a quick rundown of his methodological problems:

  1. The numbers don't account for the fact that some states have the death penalty and others don't. Nor does it control for the fact that many states that have the death penalty don't really use it.
  2. The numbers don't examine the rates at which prosecutors (who have wide discretion in the sentences they pursue) actually seek the death penalty.
  3. The numbers don't account for the various types of murders. Some types simply aren't eligible for the death penalty. There is no examination of the rates of death penalty-eligible cases among races.
  4. The numbers don't control for the extent to which similar cases are treated in a similar fashion.
  5. This one's not a methodology problem so much as a logic problem, but Brandon's entire thesis rests upon a giant division fallacy. Showing that the overall numbers don't indicate a race bias isn't the same as showing that race plays no role in particular cases.

These are just the ones that jump out at me right away. I've no doubt that there are plenty of others. But, you know, it's not as if we have to recreate the wheel here. Oddly enough, the claim that race plays a role in the death penalty doesn't actually rest upon the "silly basis of comparison" that Brandon states. (Indeed, a general rule of thumb is that when you're able to say with great confidence that an argument is "silly" the chances are pretty good that you're jousting with a straw man.)

Among the better arguments, there's the one from Pierce and Radelet looking at 10 years of data in California. And then there's the one from Phillips which looks at data from Houston. There's also the GAO meta-study that looks at 28 post-Furman studies on race and the death penalty. According to GAO, 82% of those studies show that the race of the victim plays a statistically significant role in death penalty cases. Additionally, 75% of them show that, after controlling for all variables, blacks are more likely to be sentenced to death than whites.

Certainly there is room still to disagree, and not all of the studies that GAO examined reached the same conclusion. But Brandon's "look at the numbers" doesn't get us far along this road, largely because, like the leftists he criticizes, he's looking at the wrong numbers. And even if he weren't, he's drawing a fallacious conclusion.

* Full disclosure: I happen to think that capital punishment is morally justifiable, but I have reservations about it in practice. My worries are less about race, though, and more about the fact that juries generally kind of suck. They return wrong results with alarming frequency. I prefer not to kill people until we get a whole lot better at convicting only the guilty.

Showing that the overall

Showing that the overall numbers don't indicate a race bias isn't the same as showing that race plays no role in particular cases.

Right, and I acknowledged that in the fourth paragraph, giving a few possible explanations for the proportion of executed prisoners who are black being less than the proportion of murders committed by blacks, including one that involved racism. Some other possibilities:

1. Black murderers are more likely than white murderers to live in, and be tried in, urban areas. Urban residents skew leftward, and thus may be less likely to impose the death penalty.

2. Because of lower average IQ, murders committed by blacks may skew more towards crimes of passion, while murders committed by whites may skew more towards the premeditated variety. Personally, I don't see why the former should be dealt with less severely, but they are.

My point isn't that racism doesn't play a role in the application of the death penalty, but rather that we can't infer that it does from the fact that more than an eighth of death row prisoners are black. You say this is a strawman, but trust me: When it comes to race and the left, there is no argument so transparently absurd that it can be dismissed a priori as a strawman.

I started this fully expecting to find that the majority of executed prisoners were black, or at least that the proportion of executed prisoners who were black was greater than the proportion of murders committed by blacks, because that's the impression I've gotten from reading left-wing literature on the topic.

Well That Raises the Question

A full 1/3 of the convicted black murderers who we would expect to be on death row got a lesser sentence. So if, according to you, whites are so willing to convict why are they not willing to convict them of a crime that carries the death penalty? Why would a white jury effect the numerator but not the denominator here?
What's your theory? I have no clue.

Is it because where whites are prevalent they tend to vote in the death penalty, and thus face it more often? That would be my uneducated guess. I say uneducated because I haven't really looked at these statistics much.

Poking around I did find what appear to be errors in those charts. For example Brandon's original chart on numbers executed contradicts another chart at the same site. Both have a total of 1099 executions from 1977 to 2007, but the number of whites is 708 on one and 631 on the other, while blacks 376 on one and 373 on the other.

Doesn't inspire my confidence the the abilities of those at the BLS. Of course I never trusted those guys anyway. They calculate the ridiculously low inflation rates.

Brian: There's no error. As

There's no error. As noted in the footnote of the original post, all statistics I listed counted mestizos as whites, and did not break statistics out by Hispanic origin. The chart you found does (note the footnote indicating that the columns other than Hispanic exclude persons of Hispanic origin).

The way government statistics deal with Hispanics is weird. Most Latin Americans are technically Amerindian/white mestizos, but they rarely self-identify as such. For statistical purposes, they're generally regarded as whites of Hispanic origin. This can often be misleading. For example, have you ever heard the claim that the majority of welfare recipients are white? This is only true if you include Hispanics. Non-Hispanic whites account for about a third of TANF recipients.

Heather MacDonald

Heather MacDonald did a series of articles on this subject for City Journal a couple of years ago. Her conclusions were much the same as Brandon Berg's and Constant's.

wait a sec

given that both black ratio, and having the death penalty, vary widely by state, I am skeptical of the meaningfulness of this, without controlling for state. A race that is more concentrated in states w/o death penalties will end up with far fewer representatives on death row.

Baldus Study

African-Americans make up 14% of the American population. Where do you get your information from. The study by Professor Baldus clearly reveals that their is a racial disparity in how the death sentence is implemented.