Forum: This paper was originally a professional publication. It was submitted to Enlightenment in November, 2000, as an expression of interest in presenting a paper at, and in partial fulfillment of the requirements for participation in, the First Annual Enlightenment Meeting.
The renewed debate over affirmative action programs makes appropriate a re-examination of the philosophical issues involved in this issue. This paper is concerned with those affirmative action programs which involve preferential hiring, promotion, and university admission. These programs give some degree of preference based on race, gender, or ethnic background to members of certain historically disadvantaged groups. Such groups include blacks, women, and Hispanics.
The proposal of such programs is that selection for desirable jobs not be done on the basis of merit alone, but that some weight be placed on the racial, sexual, or ethnic characteristics of the candidates. This means that we should select the minority candidate or the woman where his or her qualifications are equal to or only somewhat below those of majority candidates (typically white males).
The first position I would like to consider is that preferential treatment is morally required as redress for past injustices. In the past, white males have enjoyed unfair advantages in employment, promotion, and university admission. Minority members and women have often been denied fair access simply because of their race, gender, or ethnicity. In order to redress this past injustice, wouldn’t it be fair to give members of groups previously discriminated against advantages equivalent to those historically enjoyed by white males, and to impose this cost on white males, who, as a group, have benefitted unfairly by discriminatory practices? This would simply serve to make up for the unjustly produced distribution of benefits in the past.
But of course the problem with this line of reasoning is that it ignores the fact that the membership of both the advantaged and disadvantaged groups changes over time. The benefits of such programs do not flow primarily to those who bore the bulk of previous discrimination. And the cost of such programs does not fall on those white males who received the primary and most direct benefit from prior bias, but rather on new entrants into the job or admissions market.
Instead of attempting to justify preferential programs based on past inequities, perhaps we would do better to consider affirmative action programs as corrective of and compensatory for present injustices (some of which, of course, have resulted from past discrimination). It is not because female and minority applicants suffered discrimination that provides justification for giving preference to current applicants. Rather, the justification for present preferences must derive from the claim that women and minority members suffer present injustice.
It is argued that present injustice takes several forms. First, it is plausible that though most overt racial and sexual discrimination is outlawed, covert and subtle discrimination is still prevalent. Where selection is not made on the basis of objective criteria (such as test scores), the conscious and unconscious biases of the selecting officials may very well give rise to significant amounts of discrimination.
Affirmative action serves as a counterbalance to and compensation for this putative tendency to underrate women and minority members. According to this line of argument, such programs would not involve any unfairness to white males, for it would merely deprive them of the relative overrating they tend to receive.
Second, even where rational selection criteria are applied with scrupulous equality, we might hold that the selections produced are unjust because minority members and women have been unfairly disadvantaged in acquiring the qualities that would enable them to compete effectively with white males. A disproportionate number of blacks and Hispanics are poor, and have attended worse schools than whites. To the extent that this relative poverty is a result of racial discrimination, and to the extent that economic disadvantage makes it more difficult to acquire traits relevant to job performance, the distribution of benefits (such as jobs and university admissions) produced by even objective and culturally unbiased tests is unfair, the argument goes, for the differentials are in part artifacts produced by unfair discrimination.
Additionally, women and minority members bear social and psychological costs of discrimination which make them less able to compete. Racial and sexual discrimination tend to rob their victims of self-confidence and self-esteem, and to sap the motivation of those who suffer them. Certainly qualities such as self-confidence, self-esteem, and a high level of motivation are relevant to one’s ability to compete.
Do these arguments, based on current discrimination and current effects of past discrimination, give us sufficient justification for affirmative action programs ? The argument makes use of two central empirical claims: 1.) Some people are likely to be underrated as a result of bias; and, 2.) Some people have been denied equal opportunity to acquire those characteristics which would enable them to compete successfully.
For an affirmative action program to match the justification given, it would have to, as a minimum, correctly target those who suffer either of these disadvantages. Obviously, a program which involved giving preference to persons not disadvantaged in either of those ways could not be justified by appeal to those disadvantages.
This is not to say that we require a perfect correlation between those who are in fact disadvantaged (in ways 1 and 2 above), and those who receive the benefits of affirmative action. Practical considerations would make this impossible. But we should require a reasonable match. We should require that affirmative action programs not give preference to large numbers of people not disadvantaged in the ways specified, and that such programs not, in large numbers of cases, impose costs on those who are not unjustly advantaged.
Unfortunately, affirmative action programs which use racial, ethnic, and gender distinctions fail in critical ways to meet this requirement. While it is true that many women and minority members suffer the effects of discrimination, it is not the case that all or even almost all do. A significant number of people designated as disadvantaged by affirmative action programs are not in fact disadvantaged.
This is made critical by the fact that the proposed preferences accrue disproportionately to just those women and minority members who are actually advantaged, or who are least disadvantaged. We relied in part on the reasonable assumption that there is some correlation between social advantage and the possession of traits relevant to selection for jobs and university admissions. Those women and minority members who received the best educations, who came from the best social backgrounds, and who suffered the least from discrimination will tend to have those characteristics which advantage them over the more oppressed members of the same groups. When we use preferential hiring to increase the proportion of women and minority members, we of course choose from among the most able women and minority members we can--that is, from among those who tend to have the most socially advantaged position within that group. Notice also that the costs of affirmative action tend to fall not on the most advantaged white males, but rather on the lowest ranking white males who would have been selected if not for affirmative action. If our assumption that there is a correlation between social advantage and the possession of desirable competitive traits is correct, the white males displaced by affirmative action will tend to be those least advantaged among that group.
The problem is that racial, ethnic, and gender distinctions simply do not correlate sufficiently with the distinction between those unfairly advantaged and those unfairly disadvantaged. And affirmative action magnifies this lack of sufficient correlation by its tendency to favor the most advantaged minority members and women, and to hurt the most disadvantaged white males.
The traits needed to compete successfully for jobs or university admissions seem much more highly correlated with social class than with race, ethnicity, or gender. The poor of any race or ethnicity tend to receive poorer educations, have less access to cultural enrichment, and have fewer reasons for believing they can compete successfully. But even class distinctions fail to correlate adequately with the kinds of advantages and disadvantages that are relevant to competitive success. Many middle-class and upper-class children are abused and neglected, for example. Many suffer misfortunes or injustices quite unrelated to social class, race, gender, or ethnicity, but which do hinder their development in ways similar to the ways in which racial or ethnic oppression tends to handicap their victims.
Far from being required by justice, affirmative action based on race, gender, or ethnicity does not even seem consistent with justice. By imposing costs disproportionately on disadvantaged white males, affirmative action creates precisely the sort of injustice correctly judged morally repugnant in the case of women and minorities.
Perhaps instead of appealing to standards of justice, we might instead appeal to considerations of social utility. Perhaps affirmative action could be justified on utilitarian grounds. The failure of affirmative action to meet the conditions of justice might be outweighed by the important social good it produces.
What then are the consequences of these programs? What i wish to suggest is that the benefits are smaller than anticipated, ant the costs are excessive. Part of the reason for this is that many of the most important personal benefits which go with holding a high-status job or attending a prestigious university can’t be transferred simply by giving someone a place in the institution. Most social and psychological rewards derive from the perception that people in such positions have a high degree of relative merit--that they have earned their positions by talent and achievement. This is certainly true of internal psychic rewards, such as self-confidence and self-esteem (which having a high-status position usually brings a person). It is also true of the respect and status which society confers on people in such positions.
This suggest that many of the important social and psychological costs of discrimination born by many women and minority members cannot be effectively relieved by affirmative action. Women and members of minority groups have been denied access to appropriate social status and respect, and discrimination is certainly destructive of self-confidence and self-esteem. Unfortunately, affirmative action is ill-designed to correct this.
More unfortunately, affirmative action tends to cause an undervaluation of the genuinely earned achievements of women and minority members. There is a tendency for every promotion, hiring, or admission of a woman or a minority member to become suspect. Neither we nor the candidate can be certain whether his or her success is the result of ability and achievement, or is due to preference given because of race or sex.
Not only does this damage high-achieving women and minority members. It damages society as a whole. An important mechanism for reducing racial and sexual bias is the perception by bigots of examples of high ability and achievement by women and minority members. Awareness of clear-cut cases of achievement by women and minority members helps undercut traditional biases about the abilities of women and minorities. By making us uncertain about whether the successes of women and minority members are based on merit, affirmative action robs us of the sorts of clear and manifest evidence of minority ability which could serve to reduce prejudice.
A second important cost is the creation of racial and gender resentment among which males who feel unjustly treated by affirmative action. In fact, the resentment produced seems disproportionately large relative to the actual levels of discriminatory effect against white males. For, many white males who would not have succeeded anyway blame their failure on affirmative action. Insofar as the advancement of minority interests depends upon the goodwill of the majority, this backlash resentment constitutes an important cost of affirmative action.
There is one final difficulty I wish to consider. A fundamental presumption of affirmative action is that it is desirable to increase the representation of women and minority members in desirable occupations and institutions. But how large should this increase be? What percentage of engineers, physicians, or philosophy professors should be women or black? We need at least an answer in principle. We need some general guideline in order to tell where affirmative action is needed and to tell us when such programs are no longer needed.
One frequently mentioned standard is that minorities and women should have the same percentage of desirable positions as the percentage of the population they comprise. It is this standard I shall directly criticize, though I will show that objections to it apply to any similar numerical standard. We need some standard to serve as a yardstick; yet, any such standard seems arbitrary.
The problem is this: the motivation for affirmative action derives from a desire to eliminate imbalances in occupational distribution due to discrimination. But not all differences in occupational distribution are caused by discrimination. Some are the result of different distributions of career preferences among people of different cultural backgrounds.
Given that minority members come from cultures and social backgrounds which differ from those of the majority population, we should expect these cultural and social differences to be reflected in different distributions of career interests. Why would we assume, for example, that if two percent of white males want to be biochemists that the came percentage of blacks, women, or Hispanics would also prefer that profession?
To use straight proportional representation as our standard is to be insensitive to the sociological fact that distribution of career interests varies by cultural background and socio-economic group. The standard puts pressure on minorities, in the form of economic and social incentives, to alter career preferences to match those of the majority. It seeks a cultural homogenization of occupational preference. It aims at changing the interests of women and minority members, not just at promoting them.
There is no inherent undesirability or injustice in the fact that different groups are disproportionately represented in various occupations or institutions. Only such distributions as result from arbitrary discrimination (or some other artificial impediment) are objectionable. Affirmative action tends to treat all differences in occupation distributions (at least in “high status” occupations) as objectionable.
It is not only that there is no compelling reason to seek a homogenized pattern of career distribution. Because career distribution withing a cultural group depends in part on the values embodied in that culture, the attempt to get proportional representation (or any other particular pattern) is an unwarranted imposition of majority (or male) values on minorities and women. It is an insensitive intrusion. To use a standard of desirable occupation distribution for women and minority members which is derivative from the pattern of white males commits us to the implicit claim that the career distribution among white men is somehow better than alternative distributions. This assumption is simply unfounded, and is disrespectful of cultural differences which manifest themselves in different patterns of career preference.
The problem is not that there are “too few of group X represented in profession y”. The problem, insofar as there is a problem, is that some people are denied appropriate access to certain professions or to the resources necessary to enable them to compete successfully for positions in those professions, because of their racial or gender characteristics. And the problem for affirmative action programs is that they are incapable of distinguishing between differences in occupational distribution due to discrimination and those which derive from different patterns of career preference among members of different groups.
This problem for affirmative action programs is not solved by changing the numerical standard (representation by population percentage) to some other standard. Whatever standard we choose embodies a claim about what the occupational distribution among women and minority members ought to be. But how could we know this? I assume the standard we would want is one which would produce just the occupational distribution which would have existed if racial and sexual discrimination did not (and had not) existed. But what would that distribution be? Any actual occupational distribution is a function of a wide variety of factors, including shifting economic and social conditions, perceived opportunity of access, and other psychological and cultural factors which enter into the decisions by individuals to pursue or not pursue a given career. These factors seem quite resistant to any adequate prediction.
There is just no reliable way of determining in advance what the occupational distribution in a non-discriminatory environment would look like. Unfortunately, affirmative action programs involve the imposition of some standard of what such a distribution should look like, and in doing so have the effect of rewarding women and minority members for altering their preferences to match whatever standard of adequate representation is used. This is, I take it, an illegitimate goal (if intended), and an undesirable consequence (if unintended) of affirmative action.
Given these considerations, I cannot see that affirmative action is productive. And if it can’t be justified on utilitarian grounds, and if it also can’t be justified by appeal to justice, then how could it be justified?