Affirmative Action: Is There a Middle Way?

Glenn Loury
8:17 a.m.  Tuesday  10/8/96 

       It may be useful to take up Tom Wood’s implicit challenge as to whether or not the matter is really as simple as he claims. I think it is not. He offers the example of aggressive advertising designed to ensure that job or educational opportunities reach a wide audience. He says that such efforts may disproportionately benefit a minority group, but would nevertheless be permitted under Proposition 209. What he does not say is that outreach efforts targeted specifically at an underrepresented racial group, but entailing no preferential treatment, would be prohibited by the proposition. The distinction is that, whereas “aggressive” advertising takes no explicit cognizance of racial identity, targeted outreach does. A search committee which says, “We need to seek out more qualified black applicants, because we have too few blacks in our institution,” if it acts on that sentiment, will have violated the law, even if every black ultimately appointed to a position is as well or better qualified than every white rejected. This, in my view, is an extreme, and an unfortunate, position.
       We can distinguish between, on the one hand, judging applicants by the same standards–which is what I mean by the absence of preferential treatment, and, on the other hand, treating the race of an applicant as an irrelevancy–which is what I mean by color-blind absolutism. The aggressive recruitment of black candidates in the face of a manifest imbalance in (say) a police force clearly violates the color-blind standard. Or, in Wood’s terminology, it clearly involves racial discrimination; it can have the consequence of bestowing benefits (lots of job offers) on some persons, and burdens (not being aggressively recruited) on others, based solely upon their racial identity. Yet it is a reasonable thing to do under many circumstances, and there is no good argument–except the quasi-religious one about the necessity to maintain a “separation of race and state”–that I can see for making it illegal.
       In fact, I can see a strong argument for engaging in this kind of color-conscious recruitment behavior as a check on the practice of illegal discrimination against blacks. The enforcement of anti-discrimination laws must, if it is to be minimally effective, rely upon the use of quantitative measures of employment practices to trigger deeper scrutiny when a manifest disparity is observed. (This is routinely done when enforcing the tax laws, the antitrust statutes, and the security exchange regulations–why should employment discrimination be any different?) Given this enforcement practice, a prudent employer will anticipate the likelihood of a manifest imbalance, and seek to avoid it by, for example, targeting his recruitment resources toward sharply underrepresented groups. This permits him to tell an auditor, “Sure, we have few black employees; however, we tried aggressively to recruit them and found none qualified.” But, the result is that the government enforcement agent’s use of statistics to aid in ensuring compliance with anti-discrimination statutes has induced the employer (who well could be another government agent) to depart from color-blind practice in recruitment.
       It is reasoning of this kind which makes me suspicious when I hear the claim that we can somehow do what we all want to do–assure fairness and eliminate discrimination–without any attention to racial identity among employees or students. I seriously doubt that this is possible. Here, I think, is the kernel of truth in the claim of rabid defenders of affirmative action that measures such as Proposition 209 will usher in a return of racial discrimination. Such measures need not have this result; but, to the extent that they seek to excise race from the domain of legitimate considerations by public actors, they well may make it more difficult to identify or forestall discriminatory practices. Michael Kinsley
8:46 a.m.  Tuesday   10/8/96 

       Questions for Edley, Wood, and Taylor, and an answer for our moderator.
       To Chris Edley: Do you (and President Clinton, whose affirmative action task force you chaired) really believe there’s a difference in principle between “rigid numerical straightjackets” and more muddied forms of racial preference? The result is the same: Some number of whites is denied an opportunity they would otherwise get, and an equal number of blacks gets an opportunity they would otherwise be denied. What difference does it make that you don’t know the exact number in advance?
       For Thomas Wood, who says that either you favor allowing the government to give preferential treatment on the basis of race or you don’t: “The issue really is that simple.” What about the selection of Supreme Court justices? President Bush, a government official, obviously took race into account in choosing Clarence Thomas. It wasn’t his only qualification, but he did get preference over any number of white conservative lower court judges because of it. Doesn’t bother me (though the Thomas appointment bothers me for other reasons). Does it bother you? In other words: Is it REALLY that simple?
       Stuart Taylor: You acknowledge that in a society free of all current racial discrimination, in either direction, blacks would not get a proportional share of the meritocratic goodies because of the legacy of past discrimination. Why is it therefore “unfair” to whites (your word) to make up a bit for that legacy? For every black who doesn’t get a plum job because of past discrimination, there’s a white who got it instead. In what sense does that white deserve the job, instead of a black person who–but for the legacy of past discrimination–might well have gotten it?
       To the question posed by Herb Stein about which policies to prefer. This raises my main objection to affirmative action, which is that it is a set of policies aimed at redistributing inequality rather than reducing inequality. It is justice of a sort when a few lucky minority group members get a share of the fantastically valuable broadcast licenses the government has historically given away for free. But it is a better sort of justice for the government (as it is now starting to do) to auction off those licenses so their value can be shared by all the taxpayers. By definition, racial preference only becomes an issue when the government is allotting preferences of some sort. Fewer preferences of all sorts are my preferred policies. Christopher Edley
9:23 a.m.  Tuesday   10/8/96 

       The “democratic deliberation” Glenn Loury sees in California’s Proposition 209 debate is a rather diseased specimen. In a mid-September teleconference call, Gov. Pete Wilson and House Speaker Newt Gingrich pitched the ballot initiative to some 60 business executives. According to press reports, it was unvarnished wedge politics. It was not about competing visions of racial and gender justice, or the complex value choices at stake in how we define and prevent “discrimination.” They touted the supposed benefit of Proposition 209 for GOP candidates, including Bob Dole.
       Yes, we need the elusive national conversation on race, but politics-as-usual is guaranteed to make the exercise a divisive rather than healing one, more likely to mislead than inform. And that’s what we’ve seen. So I’m in search of ideas on how we do better, post-November.
       Finally, I listed some questions to illustrate our differences, but Tom Wood’s statement illustrates still another: We disagree on whether practical considerations of effects, or consequences, should matter much.
       For starters, will critics agree that it’s not enough to abhor discrimination in employment? That we also must have effective measures to remedy and prevent it? I don’t mean to ignore disagreement about complex terms, but would critics be moved by any amount of evidence that classic racial discrimination remains both too subtle and commonplace to be cured with litigation? Thomas Wood
9:47 a.m.  Tuesday   10/8/96 

       I have been misunderstood, at least on one point. When I said that anyone who opposes Proposition 209 thereby adopts the position that it should be permissible for the state to discriminate against at least some individuals or groups on the basis of race, sex, or ethnicity, I did not say–nor did I mean to imply–that the considerations to be weighed in the balance on the issue were themselves simple.
       All I said was that any decision to grant a preference is eo ipso a decision to discriminate. This is simply a truism. Nevertheless, experience in debating the subject has taught me that it is a point that is worth making, because it is often very difficult to get opponents of Proposition 209 to admit that they are advocating preferences, even when it is perfectly clear that they are doing so.
       I think there are several reasons for this. First, there is the obvious political consideration that a mountain of polling data shows with astonishing consistency that Americans oppose racial and gender preferences, even for the purpose of redressing past discrimination. Second, the linkage that undeniably exists between preferences and discrimination makes it very difficult to defend the fairness of preferences. There is, I believe, a third reason as well. Asserting, as opponents typically do, that preferences are necessary to attain racial and gender representativeness in the major institutions of American society seriously undercuts their claim that women and minorities still face massive amounts of discrimination.
       To see why the advocacy of preferences undermines the discrimination claim, imagine a labor market which is divided in equal parts between discriminating employers (call them A) and non-discriminating employers (call them B). Subject to some fairly weak assumptions about this labor market (e.g., that there are no major problems with respect to labor mobility or the information about employers that is available to workers), one would expect to see a kind of hydraulic effect operating: over time, women and minorities will simply move from the employers in set A to the employers in set B. After this has happened, the non-discriminating firms should find themselves, not only at parity, as measured by the percentage of minorities or women in the qualified labor force, but above parity. Yet in American society, there is no dearth of firms (presumably non-discriminating ones) claiming that outright preferences are essential to their progress toward “diversity.”
       This is a purely a priori argument, but it is confirmed by a wealth of empirical research. Out of all the evidence that could be cited here, perhaps the most pertinent to Proposition 209 is a study which the State Personnel Board recently completed. The study found that 97 percent of the job categories covered by the study were already at parity when the comparison is made (as federal law requires) to the qualified labor force. Furthermore, although the study found, on the whole, little underrepresentation, it also found that all racial, gender, and ethnic groups were underrepresented in some way or other, though the patterns varied from group to group.
       None of this means, obviously, that there is no discrimination in American society, or that we should minimize the seriousness of the discrimination that does occur. Laws like the 1964 Civil Rights Act have not eliminated all discrimination against minorities and women, and if Proposition 209 is adopted by the electorate on November 5, it will not eliminate all “benign” discrimination in their favor. But that is not a good reason for opposing either the Civil Rights Act or CCRI. Furthermore, considerations of the above kind strongly suggest that discrimination, though it is real enough, is not the major obstacle to “inclusiveness” and “diversity.” As Stuart Taylor rightly points out, the real problems lie elsewhere.
       Glenn Loury has suggested that while “color-blind absolutism” might lower the amount of discrimination against white males, it could also have the unintended consequence of increasing the amount of discrimination against women and minorities, because color blind absolutism might make it more difficult to identify or forestall discriminatory practices.
       I agree with Loury that statistics have an important role to play in ensuring compliance with anti-discrimination statutes. But this would be a valid reason for opposing CCRI only if CCRI prohibited the gathering and the use of such statistics for the purpose of monitoring employment practices. CCRI contains no such prohibition, provided that the information is used to assist in the detection and eradication of problematic practices and policies, rather than to grant preferential treatment on the basis of race, sex, or ethnicity in pending decisions.
       I confess to being an absolutist (but not a religious or quasi-religious one) on this matter. I understand the temptation to compromise the absolutist principle in the interest of advancing “diversity,” but I also believe that it is a temptation that needs to be resisted. In particular, I believe that it is a fundamental and very serious mistake to sacrifice civil rights principles in order to address problems, like educational or economic disadvantage, that are not in the first instance civil rights problems. Herb Stein
2:47 p.m.  Tuesday  10/8/96 

       I am still waiting for more enlightenment from our panelists about what kinds of policies they would prefer or rule out. Kinsley does not rule out “a bit of social engineering in favor of African-Americans.” But then he asks his own question: What kind of social engineering? What can he conceive of that he would favor? In today’s comment he complains that affirmative-action policies redistribute rather than reduce inequality. Would he favor affirmative-action programs that reduced inequality, between rich and poor, and between blacks and whites? He is skeptical of class-based programs, which might have that effect.
       Edley makes a plea for debating “practical remedies.” What does he have in mind?
       Wood gives an example of neutral advertising, not preferential in its terms, that would nevertheless confer a relative benefit on minorities. Is that an example of a general category of policies that he would advocate, or only of policies that are not ruled out by Proposition 209?
       While waiting for further discussion of policy I would like to ask a factual question: Do the panelists think that discrimination in employment is declining or not? In a general survey of the American economy, The New Illustrated Guide to the American Economy (1995), my colleague, Murray Foss, and I wrote:
       “In 1940 the average black male worker with twelve years of schooling earned about 56 percent as much as a white male with the same number of years of schooling. By 1980 that ratio had risen to 79 percent. There were similar gains for workers with other amounts of schooling.
       “Between 1980 and 1990 the gap between the earnings of blacks and whites changed very little. This is sometimes interpreted as a sign that the process of declining discrimination has come to an end. There is, however, another possible explanation. Since the 1980s the gap between the earnings of less skilled persons–with the same years of schooling and experience–has widened, for both whites and blacks. If, as is probable, the blacks on the average were less skilled than the whites because of earlier qualitative differences in schooling and experience, then the widening of the skill differential would have caused a decline of black earnings relative to whites. The fact that this did not happen suggests that this factor was being offset by a continuing decline in discrimination.”
       There are factors other than government policy that work to reduce discrimination. Discrimination is costly to employers, and increased education and mobility make it easier for workers to escape discrimination. Some view of the trend in discrimination would be relevant to thinking about how important affirmative action is compared to other measures that might be taken to improve the absolute and relative status of blacks in America.
       I will give an example:
       In 1990 about 5 percent of all black people and less than 1 percent of all white people lived in what were called “underclass” areas, as defined by high rates of unemployment, school dropouts, female-headed families, and welfare dependency. Programs to improve schools, provide jobs, and increase law enforcement in those areas would be especially, “preferentially,” beneficial to blacks. They would be, apparently, consistent with Proposition 209. How would their results compare with the results of affirmative-action programs? Does the striving for affirmative-action programs reduce the possibility of getting programs to improve life in the “underclass” areas?