Gary Becker's "tastes and preferences" approach to labor market discrimination, articulated in The Economics of Discrimination ( 1971), has been the dominant theory of discrimination within mainstream economics. Neoclassical economists have tended to treat Becker's construct as a general theory, a timeless model that can be applied to various forms of discrimination. Becker himself encouraged this interpretation. Instead, we provide an institutional analysis of Becker's work on discrimination that locates his theory within the social, economic, and political context of the mid 1950s, the period in which he wrote. In particular, Becker's focus on a desire for social and psychological distance as the basis for discrimination and his emphasis on job segregation as the critical outcome reflect the way civil rights issues were framed during this period.
Our deconstruction of Becker's theory relies on the assumption, defended in our previous work, that the implicit wage theories guiding economic actors and policy makers interact with the development of academic economic theory (Figart, Mutari, and Power 2002). Similarly, Michael Omi and Howard Winant, in their seminal study of racial formation in the United States, maintained that "[r]acial theory is shaped by actually existing race relations in any given historical period" (1994, 11). The development of race theory in the social sciences is part of the process of racial formation. Science and politics, according to Omi and Winant, are among the historically situated projects (or social practices) through which "racial categories are created, inhabited, transformed, and destroyed" (55). Rereading Becker, therefore, exemplifies the need to think about why certain theories arise within a particular time, space, and place.
There have been many criticisms of Becker's work (see Marshall 1974 for a summary). Our analysis does not comment on the strengths and weaknesses of these critiques unless they echo concerns raised by our contextualized reading. Rather than cataloging all of the strengths and weaknesses of Becker's view of discrimination, we focus on demonstrating the limits of his definition of discrimination. We conclude that the quest to develop a universal model is inferior to an institutionalist methodology. An institutional approach focuses on the diverse manifestations of discrimination in particular social, economic, and political contexts.
General Theory or Race Theory? Becker's Mixed Signals
Becker's pioneering work on discrimination began as a doctoral dissertation completed at the University of Chicago in 1955. He defined discrimination in monetary terms. (1) A person (either an employer, coworker, or consumer) is said to have a "taste for discrimination," just like a [dis]taste for strawberries, if he or she would pay to maintain social or psychological distance from members of a particular group. (Nepotism represents a sacrifice of income to maintain proximity to members of a particular group.) Reflecting on the concept in 2002, Becker succinctly argued that "[d]iscrimination comes from prejudice, and I translate that into a monetary amount--how much you are willing to pay" (Clement 2002). One actually discriminates, as opposed to merely having a taste for discrimination, when one forfeits income in order to indulge this preference. Thus, discrimination is never profitable for the employer, by definition. While Becker claimed to leave the motivation behind such tastes to the province of sociologists and psychologists and to focus solely on the economic consequences of discriminatory preferences, he was clear that he considered the motivation to be nonpecuniary.
The strength of discriminatory preferences can be measured by the amount of income an employer (or co-worker or customer) is willing to sacrifice to maintain distance, which Becker has referred to as a "discrimination coefficient." For an employer with a taste for discrimination, the effective wage (including both nominal and psychic costs) for hiring a member of an undesired group is w(1 + d). The discrimination coefficient (d) is a number between 0 and 1, based on the intensity of the disutility experienced by the individual discriminator. There are no psychic costs to hiring from the desired group, so their effective wage is the market wage (w). Discriminating employers would employ members of the undesirable group only if their market wage is sufficiently lower than that of the desired group to compensate for the discrimination coefficient.
The initial effect is to lower the market wage and employment level of the undesired group. In Becker's scenario, however, this would mean that the employer who hires only desirable workers pays relatively higher wages. The employer is content to do this, since Becker has defined discrimination as a willingness to pay to maintain distance. But the market is less forgiving. With higher wages, the costs of production increase and diminish profits. Conversely, the "nondiscriminating" employer who hires workers from the undesirable group can pay them their market wage, which is actually less than their marginal revenue product. Uninhibited market forces should punish discriminators, according to Becker, since nondiscriminating firms with lower costs would be more profitable. If the industry is...