Imperfect Signalling, Affirmative Action, and Black-White Wage Differentials

Abstract

The existence of black-white wage differentials has received considerable attention in the economics literature. Imperfect information provides the underpinning of several models which attempt to explain such differentials [1; 2; 3; 4; 11]. Arrow's work, for example, shows that, when screening is costly and expectations of productivity differ across groups, wages may differ between individuals with identical productivity [2; 3]. Borjas and Goldberg point out that biased screens may lead to wage differentials even when the two groups have identical distributions of productivity [4]. It is important to note that these differentials arise in the absence of discriminatory tastes among market participants. Other researchers have investigated the impact of affirmative action programs on minority-majority wage differentials. Welch indicates the kinds of wage differences which can arise under certainty about productivity levels [12]. Lazear demonstrates that the apparent narrowing of blackwhite differentials under affirmative action may well be an illusion due to the failure to measure non-pecuniary income in the form of on-the-job training [7]. While these models provide an appealing explanation of the main outlines of minoritymajority wage differentials, they do not explain another stylized fact which has been observed recently. A study of academicians revealed that black males with a number of publications earned more than their white counterparts, whereas black male academics with no publications earned less than comparable whites [5]. Olneck showed that, in terms of occupational status, the return to college graduation was higher for non-whites than for whites, but the return to secondary education was lower [6, 159-90]. Similarly, while the income of college-educated black males rose relative to that of whites of similar qualifications between 1967 and 1978, the income of less educated black males fell relative to that of their white counterparts over the same period.' Sowell has argued that this skewing of the relative distributions is a result of affirmative action programs [9; 10]. Although the quotas seem to increase the demand for minority workers, their legislative framework increases the