Tuesday, December 21, 2004

Race, Intelligence, and Affirmative Action

In "Affirmative Action: A Modest Proposal," I began by writing about the findings of a study which

shows that the income of a Korean orphan who was adopted in the U.S. between 1970 and 1980, through a process of random selection, is about the same regardless of the income of the adoptive parents. On the other hand, the income of the biological children of the same parents is highly correlated with the parents' income; that is, low -income parents tend to produce low-income children, whereas high-income parents tend to produce high-income children....

I went on to say this:

...The obvious implication of these findings is that intelligence (and hence income) is a heritable trait, one that remains differentiated along racial lines (a consistent but controversial finding discussed here, for example). Thus the findings give further evidence, if any were needed, that affirmative action policies -- whether government-prescribed or voluntarily adopted -- tend to undermine the quality of workplaces and educational institutions. (I am speaking here of the quality of effort and thought, not the value of workers and students as human beings.)

A reader objects -- sort of. He begins by saying:

[T]here's a flaw in your guest blogger's logic. He takes the adoption study as evidence that intelligence is a heritable trait and thus passed through racial lines (fine). He then says since affirmative action rewards racial minorities who may be less qualified (fine), that affirmative action tends to undermine quality of work.

[H]is conclusion may be correct, but is only tenuously related to the first premise. he seems to be saying that, on average, if you give preference to minorities, the quality of work will suffer, because on average minorities are less intelligent....

Let's stop right there and take things one step at a time. What I said is that intelligence "is a heritable trait, one that remains differentiated along racial lines (a consistent but controversial finding discussed here, for example)." There is less controversy about the persistence of the racial differential and more controversy about race, per se, being the underlying cause of that differential. For a sample of the controversy, go to the linked article and follow the many links in the article. One of those links leads to a statement by Charles Murray, co-author of the infamous The Bell Curve, who says in a footnote:

Intelligence is known to be substantially heritable in human beings as a species, but this does not mean that group differences are also heritable. Despite our explicit treatment of the issue, it is perhaps the single most widespread source of misstatement about The Bell Curve.

How is it that intelligence is "substantially heritable" and yet "group differences" may not be heritable? Here is Professor Richard E. Nisbett of the University of Michigan, a noted opponent of the notion of inherent racial disparity:

Estimates of heritability within a given population tell us nothing about the degree to which differences between populations are genetically determined. The classic example is an experiment in which a random mix of wheat seeds is grown on two different plots of land. Within either plot, the environment is kept uniform, so the height of the different plants is largely or entirely genetically-determined. Yet the average difference between the two plots is still entirely environmental, because the mix of genotypes in each plot is identical....

In other words, there's a school of thought that a racial group that starts out "behind" because of environmental causes (e.g., nutrition and exposure to education and other experiences that "stretch" the mind) stays behind, even as the average intelligence of all racial groups seems to advance over time (a phenomenon known as the Flynn effect). In any event, inter-racial differences in intelligence seem to be real and persistent, and racially related genetic causes cannot be ruled out. (Again, refer to this article.)

The distribution of those differences does not follow the pattern supposed by the reader, who goes on to say this:

[I]f I understand the studies correctly, they say that each race has members that represent the full spectrum of intelligence, and that it's only on average that the scores are lower.

I'm not sure that the reader correctly understands the distribution of intelligence and its implications for the labor market. Let's say there's a pool of 200 "typical" black applicants and 1,200 "typical" white applicants for a "typical" job that requires an IQ of 100. (I use 200 blacks and 1,200 whites because the 1:6 ratio reflects the relative numbers of blacks and whites in the U.S. I take an IQ of 100 because that's about the mean for whites, whereas the mean for blacks is about 85. IQs are assumed to be normally distributed around those means, with a standard deviation of 15 IQ points.) Now, of the "typical" applicants for this "typical" job, only 32 (16 percent) of the blacks would have an IQ of at least 100, whereas 600 (one-half) of the whites would have an IQ of at least 100. Thus the ratio of qualified blacks to qualified whites would be about 1:19 for the "typical" job.

Bump it up a notch and set the intelligence qualification at an IQ of 115. Then, only 5 (2.5 percent) of the 200 black applicants would qualify, whereas 192 (16 percent) of the 1,200 white applicants would qualify -- a ratio of about 1:38. In other words, it gets harder and harder to find qualified blacks as jobs require more intelligence (not to mention specific kinds of education and training). So, it's irrelevant that there are some blacks at the higher end of the spectrum of intelligence. Why? Because there are proportionately few of them, and fewer still who have the requisite education and training for the kinds of jobs that are associated with high intelligence (e.g., astrophysics, computer engineering, advanced mathematics).

To look at it another way, take 200 randomly selected blacks and 200 randomly selected whites: 100 of the blacks and 168 of the whites would have an IQ of at least 85 (a ratio of 1:1.7); 5 of the blacks and 32 of the whites would have an IQ of at least 115 (a ratio of 1:6.4).

The black-white difference in average intelligence is meaningful, despite what the reader seems to think, because it reflects a significant difference in the distribution of intelligence. University slots and jobs that require at least average (white) intelligence can't be filled in proportion to the number of blacks in the population, or in proportion to the number of black applicants, without tending to dilute the quality of universities and workplaces. (Again, I am speaking of the quality of effort and thought, not the value of workers and students as human beings.)

That leads me affirmative action, about which the reader says:

Thus [because there are some blacks at the high end of the spectrum of intelligence], affirmative action can be structured in such a way as to give special preference to the higher-achieving members of any minority, who face the difficult task of not being stereotyped by the lower scores of their fellow minorities. I.e., If a white person and a black person have the same or nearly the same qualifications, then you pick the black person.

Yes, as I have just shown, there are blacks at the high end of the spectrum of intelligence, and those blacks are courted assiduously by universities and employers. Why? Because universities and employers are anxious to demonstrate their commitment to affirmative action, diversity, racial equality, or whatever you want to call it. What better way to do that than to admit or hire the "best and brightest" blacks, which is a relatively risk-free proposition for universities and employers. What happens to those blacks who aren't in the higher reaches of the spectrum of intelligence? Well, that's where affirmative action, as most Americans know it, kicks in.

Here's how it seems to work at universities: Blacks get preferential treatment for being black, to the extent that universities can concoct and defend affirmative-action plans that allow them to give preferential treatment. Sometimes a university fails (as in Gratz v. Bollinger), and sometimes it succeeds (as in Grutter v. Bollinger). But if there's a prevailing tendency among the left-dominated universities of the United States, it's to allow blacks to meet a lower standard of intelligence, thus displacing some whites who would have made better students and, eventually, better employees. So, at universities, affirmative action isn't just about "picking the black person" who has "the same or nearly the same qualifications."

What about affirmative action in the workplace? Here, I speak from long experience. (See my credentials.) Affirmative action, in theory, is supposed to be about hiring and promoting regardless of race, among other attributes. As an example, here's the Department of Labor's summary of its guidelines for federal contractors and subcontractors:

Each contracting agency in the Executive Branch of government must include the equal opportunity clause in each of its nonexempt government contracts. The equal opportunity clause requires that the contractor will take affirmative action to ensure that applicants are employed, and that employees are treated during employment, without regard to their race, color, religion, sex or national origin....

It doesn't say "If a white person and a black person have the same or nearly the same qualifications, then you pick the black person," as the reader would have it. What it says, in effect, is this: Faced with two equally qualified candidates for hiring or promotion, you can't discriminate against a black person or a person who belongs to any of the other protected groups. To act in the way that the reader suggests would amount to blatant discrimination in favor of black job candidates over white job candidates, and that's facially illegal, even though universities sometimes get away with similar discrimination in the name of "diversity."

Nevertheless, what happens, in practice, is what the reader suggests, and then some: If a black person seems to have something like the minimum qualifications for a job, and if the black person's work record and interviews aren't off-putting, the black person is likely to be hired or promoted ahead of equally or better-qualified whites. Why?

  • Pressure from government affirmative-action offices, which focus on percentages of minorities hired and promoted, not on the qualifications of applicants for hiring and promotion.
  • The ability of those affirmative-action offices to put government agencies and private employers through the pain and expense of extensive audits, backed by the threat of adverse reports to higher ups (in the case of government agencies) and fines and the loss of contracts (in the case of private employers).
  • The ever-present threat of complaints to the EEOC (or its local counterpart) by rejected minority candidates for hiring and promotion. Those complaints can then be followed by costly litigation, settlements, and court judgments.
  • Boards of directors and senior managers who (a) fear the adverse publicity that can accompany employment-related litigation and (b) push for special treatment of minorities because they think it's "the right thing to do."
  • Managers down the line learn to go along and practice just enough reverse discrimination to keep affirmative-action offices and upper management happy.

The following case, about an employee who was victimized by reverse discrimination, illustrates just about everything I've said about the practice of affirmative action in the workplace:

A Federal Aviation Administration employee recently settled an employment discrimination case where he said he was passed over for promotions because of his gender and race.

Michael C. Ryan of Toms River, N.J., who worked at an FAA research and development facility as a GS-14 manager, said that between 1995 and 1997 he was denied eight promotions to GS-15.

After complaining to the FAA, Ryan went to the Equal Employment Opportunity Commission. Nine years later, a formal consent order gives Ryan, a 28-year FAA worker, the managerial and supervisory position he wanted. The order also begins a three-year agencywide policy review intended to reform FAA's affirmative action policies.

Ryan, a white male, said he was qualified for the promotions he applied for at the William J. Hughes Technical Center in Atlantic City, N.J., but was passed over by people with less experience because he was not a woman or a minority. During the trial, Ryan's attorney, Hanan Isaacs, argued that four of the seven minority candidates who were promoted were not selected using merit principles, including one person that Ryan trained who had 13 years less seniority.

According to Isaacs, the 22-day trial showed that the candidates were promoted ahead of Ryan so that minority and women promotion quotas could be met. Isaacs said FAA's 1988 affirmative action plan, which called for "a workforce that looks like America by 2000," started to go afoul when it compared the racial and gender composition of technical positions to the general population rather than to the minority composition of the comparable workforce.

Isaacs said that an unwritten but well publicized "50-50" policy" required FAA managers to promote women and minorities at least 50 percent of the time in order to get career and financial incentives. This type of affirmative action has no end-plan and perpetually discriminates against nonminorities, Isaacs argued.

Ryan was offered a settlement a year ago that would have given him back pay - which could total about $100,000 - and the promotion, but Isaacs said Ryan refused because he wanted to see the agency's policy change.

John G. Larsen, a FAA senior policy analyst, testified during the trial that the agency was not in compliance with the law after 1992 and that its affirmative action program would "almost always come up with the appearance of under-representation."

Larsen, a 36-year FAA employee, said that after a 1995 Supreme Court ruling which found that preferential treatment based on race almost always is unconstitutional, even when it is intended to benefit minority groups that suffered injustices in the past, the agency's affirmative action policies became illegal.

He said the FAA refused to conduct a review requested by the Clinton administration following the ruling that would have brought the agency back into compliance with affirmative action laws. "The culture of the agency was one, in my opinion, that did not entertain challenges or disagreement ... and nothing changed," Larsen said.

The agency did not admit liability in the settlement, but did agree to start a three-step comprehensive review of its programs and policies on hiring and promotion to put them into compliance.

A Justice Department spokesman said the department was happy to resolve the nine-year-old case. He said that with the assistance of the court, the department was able to reach a settlement that is fair to both parties and upholds the FAA's commitment to ensure a workplace free of unlawful discrimination of any form.

That's the real, illegal, world of affirmative action. And here is the price tag:

Because of affirmative action -- and legal actions brought and threatened under its rubric -- employers do not always fill every job with the person best qualified for the job. The result is that the economy produces less than it would in the absence of affirmative action....

[A]ffirmative action reduces GDP by about 2 percent. That's not a trivial amount. In fact, it's just about what the federal government spends on all civilian agencies and their activities -- including affirmative action....

Moreover, that effect is compounded to the extent that affirmative action reduces the quality of education at universities, which it surely must do. But let us work with 2 percent of GDP, which comes to about $240 billion a year, or more than $6,000 a year for every black American....

So, the reader has it about right when he says, in his closing sentence,

It may be true that in practice affirmative action tends to downgrade quality....

But he glosses over the high price we pay for affirmative action, in dollars and divisiveness. And then he closes with this:

...but this [downgrading of quality] doesn't follow necessarily from the heritability premise, and I find [the guest blogger's] attempt to use this to bolster his argument inflammatory and intellectually dishonest.

The downgrading of quality -- and the price we pay for that -- follows directly from the demonstrable premise that affirmative action -- as it is practiced -- puts race ahead of quality in the selection of students and workers. Putting race first affects quality because of the unequal distribution of intelligence between the races, as intelligence is usually measured. The cause of the unequal distribution of intelligence may be controversial, but as far as I can tell there is no settled science in the matter. The notion of inherent racial differences in intelligence is still on the table, and it carries with it stark implications for the long-term success of blacks in an economy that increasingly demands more intellectual skills and fewer physical skills.

Therefore, it isn't "intellectually dishonest" to raise the issue of inherent racial differences in intelligence. Nor is it "inflammatory," except to those who -- unlike me -- are unwilling to review dispassionately the evidence on all sides of the issue. But dispassion is hard to come by in any discussion of race or affirmative action. That is why I offered my "modest proposal" -- which I mean to be taken seriously. It cuts through all the cant and controversy about race, intelligence, and affirmative action. Here it is, again:

...End affirmative action and give every black American an annual voucher for, say, $5,000 (adjusted annually for inflation). The vouchers could be redeemed for educational expenses (tuition, materials, books, room and board, and mandatory fees). Recipients who didn't need or want their vouchers could sell them to others (presumably at a discount), give them away, or bequeath them for use by later generations. The vouchers would be issued for a limited time (perhaps the 25 years envisioned by Justice O'Connor in Grutter), but they would never expire.

That settles affirmative action, reparations, and school vouchers (for blacks), at a stroke....

Favorite Posts: Affirmative Action and Race