Academia. The central assumption of affirmative action programs
is that “underrepresentation” of minorities or women represents employer
“exclusion” rather than different career characteristics of groups or different
choices by the individuals themselves. In the academic world, major intergroup
differences have been found in degree levels, publications, and fields
of specialization-all these differences being to the disadvantage of minorities
and women. For minorities, holding such variables constant reduces, eliminates,
or even reverses salary differentials as compared to white academics. Even
without holding such variables constant, the pay differentials between
minorities and other academics were less than $100 per year before
affirmative action and less than $1,000 afterwards-indicating that both
the necessity for such programs and the effectiveness of them are open
to serious question. For women, holding the same variables constant does
not eliminate salary differentials, but holding full-time employment constant
comes close to doing so, and for those women without marital responsibilities,
sex differentials disappear. Together with much other data, this suggests
that marital status in general and an unequal division of domestic responsibility
in particular explain both differential trends over time and the differences
at a given time between male and female academics.
The term “career characteristics” has been used here, not simply to avoid the emotionally loaded word “qualified,” but because it seems more accurate and germane. Given the enormous range of American colleges and universities, virtually anyone with graduate training is “qualified” to teach somewhere, while only a small fraction of the Ph.D.s are “qualified” to teach at the very top institutions. The question of qualifications therefore amounts to a question of whether a particular individual matches a particular institution, rather than whether he or she belongs in the profession. An institution is not excluding a “qualified” applicant because it hires someone else whose career characteristics fit its institutional needs, even though those not hired have career characteristics which make them valuable to other kinds of institutions. Research universities-a major focus of affirmative action programs-offer a specialized environment which many academics do not want, as well as one for which many do not have the appropriate set of career characteristics. It is as unnecessary to denigrate either individuals or groups as it is to denounce as “irrelevant” the characteristics which research universities seek for their special purposes.
The crucial element of individual choice is routinely ignored in analyses and charges growing out of statistical distributions of people. Women choose to emphasize teaching over research, and this has implications for their degree levels (a Ph.D. is not as essential), the kind of institutions they prefer, and the life style it permits-including part-time and intermittent careers that mesh with domestic life. Black faculty prefer being where they are to such an extent that it would take more than a $6,000 raise to move them, according to a survey of several hundred black academics.
One of the peculiarities of the academic market is its fragmentation or balkanization.1 A particular department typically hires people trained at a relatively small number of other departments. This is due to the high cost of specific knowledge about specific individuals as they emerge from graduate school. At that point, the individual usually has no publications or teaching experience, and the only indications of his intellectual potential are the estimates of professors who taught him or directed his thesis research-and the value of those estimates depends crucially upon the reliability of those professors, which in turn means it depends on how well members of the employing department know members of the department where the applicant was trained. The top departments in many fields typically hire from other top departments in the same fields. This has led to charges of an “old boy network” among the top departments which excludes outsiders in general and minorities and women in particular. But despite loose talk about “recruitment procedures that tend exclusively to reach white males”2 the fact is that (1) most black Ph.D.s were trained in a very few highly rated, predominantly white departments3 and (2) a slightly higher proportion of female doctorates than of male doctorates received their Ph.D.s from the top twelve universities.4 In short, whatever the merits or demerits of the “old boy network,” as high a percentage of minority and female Ph.D.s as of white male Ph.D.s are inside its orbit.
Affirmative action practices ignore both choice and career characteristics by the simple process of putting the burden of proof on academic institutions to explain why their percentages of minority and female faculty do not match the kinds of proportions preconceived by governmental authorities. Career characteristics are accepted as mitigating factors only when job criteria have been “validated”-which is virtually impossible. The statistical “validation” process, as developed for written tests in education, involves prediction for a very short span of time on a very limited number of variables, such as grades and graduation. To extend the “validation” concept to the whole hiring process for complex professions with many dimensions is to demand mathematical certainty in areas where good judgment is the most that can be expected. In such circumstances, where “validation” amounts to convincing government officials, it means convincing people whose own career variables-appropriations, staff, and power-depend upon not believing those attempting to convince them. General findings of reasonable hiring decisions would be a general sentence of death for the agency itself. More basically, this situation replaces the principle of prescriptive laws with ex post administrative determination of what should have happened, combined with never-ending burdens of proof as to why it did not.
A mitigating factor (in the opinion of some) is that the ultimate sanction of contract cancellation is not actually invoked. But this means that the real penalty is having to repeatedly devote substantial institutional resources to producing the pounds of paper which constitute an affirmative action report-and this penalty falls equally on the just and on the unjust. Even aside from the disturbing moral implications of this, it means that the effectiveness of the penalty is reduced when a discriminating employer has little to gain by becoming a non-discriminating employer, in a society where the career characteristics of the target population ensure that he will never be able to fill affirmative action quotas anyway. There is truth in the bitter comments from both sides of the affirmative action controversy that (1) colleges and universities are under unremitting and unreasonable pressures and that (2) virtually nothing is actually being accomplished for minorities and women. An even weightier consideration is that the appearance of massive benefits being conferred on minorities and women undermines the very real achievements of minorities and women themselves, often made at
great personal sacrifice-achievements whose general recognition would be a very healthy influence on society at large.
There are a number of ways in which affirmative action programs hurt the academic world without benefiting minorities or women:
The Economy. To what extent can the patterns found in the academic
profession be generalized to the larger society? That question can be answered
only after applying a similar approach to the economy as a whole-that is,
going beyond the global black-white or male-female comparisons to comparisons
of segments carefully matched for the relevant variables. For women, such
matching eliminates sex differentials among continuously employed single
individuals.7 Among blacks, college-educated men had achieved
starting salary equality by 1970,8 with “virtually all of the
improvement in relative income” occurring “after passage of the 1964 Civil
Rights Act”9 but before affirmative action quotas under Revised
Order No. 4 in 1971. For black male workers as a whole, firms with government
contracts showed a larger increase in the earnings of black workers relative
to the earnings of whites than firms without government contracts, but
this difference “accounts for only about 6% of the overall change in the
relative position of black workers.”10 In short, it was antidiscrimination
or equal opportunity laws, not goals or quotas, that made the difference.
Another way of looking at this is that blacks achieved when given
equal opportunity, and were not passive beneficiaries of conferred gains.
While only segments of minority and female populations have achieved income equality with their white male counterparts, the differences between these segments and other segments of the same populations give clues as to the causes of the remaining inequalities. For example, as noted above, marital status is as crucial a variable among women in general as it is among academic women in particular. Among married women, labor force participation declines as the husband’s income rises, both in the general economy and in the academic world.11 Children have a negative effect on work participation for women in general as well as for academic women.12 As for trends over time, there has been a generally declining trend in the proportion of women in various high-status occupations since around 1930, coinciding with earlier marriages and the baby boom,13 but this trend began to reverse-before affirmative action. For example, the proportion of “professional and technical workers” who were female was 39.0 in 1950, 38.4 in 1960, and 39.9 in 1970.14 The proportion of “college presidents, professors and instructors” who are female was 31.9 percent in 1930, falling to a low of 24.2 percent in 11360, and rising slightly to 28.2 percent in 1970.15
Among blacks, income parity has been achieved not only by college-educated men (and slightly more than parity achieved by
college-educated black women)16 but also by young (under thirty- five) intact husband-wife families outside the South.17 For the latter, this parity was achieved in 1971, but this could hardly have been a result of goals and timetables formulated in December 1971 and implemented the following year. Another way of looking at the still substantial black-white income inequalities is that these inequalities exist among older blacks, blacks in the South, in “broken” families, and among the less educated. There remains a substantial agenda for further progress, but the record shows that the progress that has already occurred was the result of anti-discrimination or equal-opportunity pressures which allowed blacks to achieve sharply rising income relative to the income of whites in a few years, after decades of stagnation in the same relative position.18 The ratio of black family income to white family income reached a peak in 1970-before affirmative action-and has declined slightly in 1971 and 1972.19 It is unnecessary to blame affirmative action for the decline. It is enough that there is no evidence that goals and timetables produced any further advance, but only cast doubt on, and caused interracial bitterness over, what blacks had already achieved themselves without quotas.
Policy. The long and virtually complete exclusion of outstanding
black scholars from all of the leading universities in the United States
until the past generation20 suggests that market forces alone
were not enough to open up opportunities in this nonprofit sector. Indeed,
economic principles would suggest that nonprofit sectors in general are
less likely than other sectors to reduce discrimination in response to
economic forces alone21-and this includes government, both local22
and national.23 The question is not whether there is a legitimate
role for government to play in reducing discrimination, but how government
should carry out its responsibilities. Affirmative action came along after
a series of antidiscrimination laws and a change of public opinion. It
must be judged against that background, not against a background of uninhibited
discrimination in earlier eras, as its proponents like to judge it.
The crucial issue of principle is whether the focus of governmental efforts shall be statistical categories or individual rights. The crucial, practical issue is who shall bear the burden of proof-the government or those subject to its power?
Categories and statistics are a bottomless pit of complications and uncertainties. For example, an economics department with a job opening is not looking for an “economist,” or even for a “qualified” economist; it is looking for an international trade specialist with an econometrics background or a labor economist familiar with manpower programs, et cetera. Statistics on how many “qualified” minority or female “economists” in general are “available” are meaningless. Neither minorities nor women are randomly distributed by field or within fields. Female economists, for example, are not distributed the same way as male economists among specialties.24 Even to define the relevant pools for purposes of realistic goals and timetables is impossible, even if all the statistics on the profession are at one’s fingertips and completely up to date-as they never are. No department can predict in which sub-specialty its vacancies are going to occur, for that involves predicting which particular members of its own department will choose to leave-and in an era of retrenchment, vacancies have more effect on hiring than does the creation of new positions.
Statistical “laws” apply to large numbers of random events. But universities do not hire large numbers of random academic employees; departments each hire small numbers of specialists within their respective fields. To establish numerical goals and timetables for such small-sample unpredictable events is to go beyond statistics to sweeping preconceptions. Nowhere can one observe the random distribution of human beings implicitly assumed by affirmative action programs. Mountains of research show that different groups of people distribute themselves in different patterns, even in voluntary activities wholly within their control, such as choice of card games or television programs, not to mention such well-researched areas as voting, dating, child-rearing practices, et cetera.
The American system of justice puts the burden of proof on the accuser, but this principle has been reversed in practice by agencies administering affirmative action programs. Those subject to their power must prove that failure to achieve the kinds of employment proportions preconceived by the agency is innocent in general, and in particular colleges and universities must “validate” their job criteria-even if the government administrators could never do the same for their own jobs. No proof-or even hard evidence-was necessary for the agencies to demonstrate that the academic situation involved individual discrimination rather than statistical patterns reflecting general social conditions outside the institution. Any policy which is to claim respect as a prescriptive law must put the burden of proof back on the government, where it belongs.
A change from categorical statistical presumptions to evidence on individual cases requires a knowledge of academic norms and practices going well beyond the expertise of nonacademic government officials. The lack of such knowledge by those ad- ministering “guidelines” for higher education has been a bitter complaint among academics.25 Certainly it is revealing when J. Stanley Pottinger can refer to the university “personnel officer” as hiring agent,26 when faculty hiring is in fact done by individual departments, with the candidates having little or no contact with “university” officials before being hired. In any event, if professional judgments are to be subject to review in cases where discrimination is charged, that review requires at least equally qualified professionals as judges. Since academic disciplines have their own respective professional organizations-the American Economic Association, the American Sociological Association, et cetera-these organizations could readily supply panels of experts to review the reasonableness of the decisions made in disputed cases. If academic freedom and faculty self-governance are to be maintained, such a review must determine whether the original hiring decision fell within the reasonable range, not substitute the choice of the panel for the choice of the department.
The great problem with individual case-by-case adjudication is the backlog that can be generated-to the detriment of all and perhaps fatally so for the effectiveness of the program. There are some countervailing factors in the case of judgments by a panel of experts. First of all, the panel can quickly dismiss frivolous claims-especially where the claim must be based on demonstrable evidence of superiority of the candidate rejected over the candidate actually hired. Second, to go before such a panel risks public confirmation of the opposite by leading scholars in one’s field. Finally, the mere fact that such a program is based on professional criteria rather than nebulous presumptions must have an inhibiting effect on claims without substance.
Remedies for demonstrable discrimination must hit those responsible, not be diffused over a sprawling entity such as a large research university. A history department which discriminates against minorities or women is unlikely to be deterred by the medical school’s possible loss of a government contract. But there is nothing to prevent the government from levying a stiff fine on the specific department or other academic unit that made a discriminatory hiring decision-and taking that fine out of that department’s or unit’s budget for salary and research, without interruption of contracts and the often vital work being produced elsewhere in the university. Indeed, such a fine is a more credible threat, for the government and the public would often lose heavily if some university contracts were cancelled. Contract cancellation is like a nuclear weapon that is too powerful to use in any but the most extreme cases and so loses much of its apparent effectiveness. Fines are a more conventional deterrent and can be invoked whenever the occasion calls for them.
Between the original concept of affirmative action and the goals and timetables actually imposed lies an ill-conceived mixture of unsupported assumptions and burdensome requirements which remain ineffective because of their indiscriminate nature-their failure to distinguish discriminators from nondiscriminators, or to give anyone an incentive to change from one of these categories to the other. Inescapable burdens do not cause change but only bitterness. That bitterness not only has been directed against those administering affirmative action programs, but has inevitably affected the perception and reception of minorities and women in the academic world-and beyond.