Knowledge And Decisions

The point here is not that all is well. Far from it. The point is that both causal determination and policy prescription require coherent analysis, rather than gut feelings garnished with numbers. Many of the hypotheses behind “affirmative action” are not unreasonable as hypotheses. What is unreasonable is turning hypotheses into axioms. The preference for intentional variables (“discrimination”) has virtually excluded systemic variables (age, location, culture) from even being considered. The practical consequences of this arbitrary theoretical exclusion extend far beyond the middlemen — employers — to much larger and more vulnerable groups, notably ethnic minorities themselves. Every false diagnosis of a condition is an obstacle to improvement. When recent studies show the still substantial black-white income differences to reflect conditions that existed before the younger age-cohorts ever reached the employer — reading (or nonreading) habits in the home, education, etc.¹²¹ — this has implications for the effectiveness of programs which (1) postulate that discrepancies discovered at the work place are due to decisions made at the work place, and (2) establish legal processes centering on the work place.

The effect of “affirmative action” programs is viewed as axiomatically as its original process. In fact, however, studies have found little or no effect from affirmative action in advancing ethnic minorities, in either incomes or occupations.¹²² In some particular places — prominent firms, public utilities, and others especially in need of appeasing federal administrative agencies — there have been some changes. But overall, the economic position of minorities changed little since “goals and timetables” (quotas) became mandatory in December 1971.

The ineffective record of “affirmative action” policies is in sharp contrast with the record of “equal opportunity” laws in the years immediately preceding. After passage of the Civil Rights Act of 1964 — and before quotas in 1971 — black income as a percentage of white income rose sharply, with blacks in white collar occupations also rising, along with rising proportions of blacks in skilled and professional jobs.¹²³ One reason for the difference was the different set of incentives presented by the two policies. “Equal opportunity” laws provided penalties for specifically proven discrimination. “Affirmative action” laws penalized numbers that disappointed administrative agencies, and made defenses against “rebuttable presumptions” costly and uncertain.

It might appear at first that “affirmative action” penalties — costs — were “stronger” (higher), but not when costs are recognized as opportunity costs, the difference between following one course of action rather than another. The general unattainability of many quotas means that penalties fall equally on discriminatory employers and nondiscriminatory employers. A discriminatory employer therefore has little to gain by becoming a nondiscriminatory employer, when the characteristics of the target population (age, education, etc.) insure that he will be unable to fill quotas anyway. Moreover, the ease with which a discrimination case can be made makes minorities and women more dangerous employees to have, in terms of future prospects of law suits if their subsequent pay, promotions, or other benefits do not match those of other employees or the expectations of administrative agencies. As in the case of other groups with special rights, as noted in Chapter 5, these rights have costs to the recipients themselves. In short, “affirmative action” provides opposing incentives to hire and not hire minorities and women. It is not surprising that it has been less effective than “equal opportunity” laws which provide incentives in only one direction.