Abstract
A rumor has been circulating that Glenn Loury has changed his position along the ideological spectrum, moving away from the once iconoclastic but always lucrative camp of the black neoconservatives—Condoleeza Rice, Thomas Sowell, Walter Williams, Alan Keyes, Armstrong Williams, John McWhorter, Shelby Steele, Carol Swain, Clarence Thomas—toward the more conventional center of black American politics. This center is well represented by what Sowell once dismissively called “the Civil Rights Vision” and by the positions generally taken by the NAACP and the National Urban League. From the perspective of public policy—for example, Loury now stands in favor of affirmative action—the rumor may be true. But I remain curious whether there has been a fundamental reorientation in Loury's theoretical approach toward race and society. Has he, in fact, departed in any significant way from the frame of thought that informed his 1976 dissertation at MIT?
I have known Glenn Loury since I first entered the Ph.D. program in economics at MIT in fall 1975. We were part of a comparatively large cadre of black graduate students located there at that time. The group included Sam Myers, Jr., Julianne Malveaux, Ron Ferguson, Ron Mincy, Linda Datcher-Loury, Darius Mans, and Virgulino Duarte. Susan Collins, Eugene Flood, and the late Rhonda Williams came through immediately after I finished in 1978. It is safe to say that we would not have been there in those numbers in the absence of a commitment to affirmative action.
Glenn Loury and Sam Myers finished two years before I did, but we were together long enough to exchange ideas, to observe the development of the premises that formed the basis for our respective dissertations and generally to get a feel for where each one of us “was coming from.” Is Glenn Loury coming from a different place now? My direct answer is “no.” Let me elaborate.
Loury was kind enough to send me a copy of his new book with a generous inscription that intimated that part of his “meditation” in The Anatomy of Racial Inequality was prompted by criticisms I had raised about the statistical theory of discrimination. So I expectantly, and a bit egotistically, opened the book anticipating that my concerns would be cited chapter and verse and would be addressed, even if refuted decisively. But nowhere in the book is there any reference to any of my work on discrimination or stigma, nor is there any reference to the specific ways in which I framed my reservations about the statistical theory of discrimination.
Putting personal pique aside, I must add that despite the growing body of scholarship, both theoretical and empirical, undertaken by black economists, there is no reference to any black economists' work except that of Thomas Sowell. It is perhaps surprising—given the rumor of Loury's transformation—that his reference to Sowell's work is quite enthusiastic in one of the two places where it is mentioned. Regardless, Loury's “meditation” is uniformed by the substantial body of empirical research on discrimination in markets for employment, real estate, housing, or credit, regardless of the race of the researcher.
The work of Sam Myers, Jr., Patrick Williams is just as absent from Glenn's text as that of George Galster, John Yinger, Marlene Kim, Frank Levy, Francine Blau, or Helen Ladd, to name a few. Not even Derek Neal and William Johnson surface, although their well-known Journal of Political Economy paper attempts to provide evidence that is broadly supportive of Loury's position. This led me to speculate that the “mature” Glenn Loury of 2002 still embraces the views of the “young” Glenn Loury of 1976, unaffected by more than a quarter century of empirical studies and debate.
What were the views of the young Glenn Loury circa 1976? I submit they were the following:
Labor market discrimination is comparatively unimportant in explaining racial differences in wages, earnings, occupational status or unemployment in the USA.
“Premarket” factors that dictate the mean human capital or productivity-linked attributes of racial groups are decisive in explaining racial economic inequality.
Racial differences in the accumulation of human capital attributes are due to historical differences in opportunity and deprivation visited upon each group in question. Thus black Americans bear the burden of America's past—the legacy of slavery and Jim Crow—produces cumulative, intergenerational effects that weigh down the status of blacks today. Present-day racism is of a low order of significance in affecting black lives.
Equal opportunity is not enough. Loury's long adventure with the neoconservatives led him, for a politically significant time, to reject affirmative action as an appropriate corrective measure. But he never took the position that corrective measures of all types are unnecessary or unjustified. Schools, neighborhood dynamics and familial practices (“family values”) became his targets for public policy action. However one views Loury's position on the policy mission, he always has said that there is a moral basis for social action to redress racial inequality even if the source of the disparity lies exclusively in pre-market processes.
I further submit that the position above is reproduced in its entirety in The Anatomy of Racial Inequality. The reader must be alert that Loury's inventive rhetorical dichotomies actually reproduce familiar arguments. He uses different ways to say the same thing. Consider, for example, Loury's distinction on pages 93–95 between “development bias” and “reward bias.” The former, “development bias” actually refers to “pre-market” processes that promote or inhibit group human capital acquisition while the latter, “reward bias,” is explicitly designated as in-market discrimination by Loury. Loury (p. 95) says “reward bias” or discrimination largely has evaporated in the USA:
So for both empirical and philosophical reasons I conclude that a continued focus on the classic racial discrimination problem is now misplaced. Anti-black reward bias has declined sharply in the United States over the past half-century. And the normative challenge posed by enduring racial inequality can be fully grasped and effectively met only if greater attention is given to the problem of development bias.
Here Loury's silence on the specifies of the “empirical reasons” for his claim is convenient. Judgment is rendered without making explicit his basis for rejecting studies with findings at odds with his contention that “[a]nti-black reward bias has declined sharply in the United States over the past half-century.” The best inter-temporal studies on discrimination of which I am aware suggest that the most important decline in measured discrimination in employment occurred solely in the decade immediately after passage of the Civil Rights Act of 1964, followed by a persistent, significant and fairly stable level of discrimination in wages and employment thereafter (see the critical surveys in Darity 1998 and Darity and Mason 1998).
Nevertheless, Loury (p. 101) gives particular emphasis to what he describes as a substantial gap in skills between blacks and whites. This skills gap, he also contends, is primarily due to America's past:
[T]he so-called underclass in the ghettos of America is behaving badly, in self-destructive and threatening ways. But those patterns of behavior, embodied in those individuals, reflect structures of human development that are biased because of a history of deprivation and racial oppression (p. 124).
At this stage of his argument, Loury (pp. 95–99) makes an additional rhetorical distinction between discrimination in contract and in contact. The former “is meant to invoke the unequal treatment of otherwise like persons on the basis of race in the execution of formal contracts,” while the latter “refers to the unequal treatment of persons on the basis of race in the associations and relationships that are formed among individuals in social life, including the choice of social intimates, neighbors, friends, heroes, and villains … the private spheres of life” (pp. 95–96). Why the distinction is made, given Loury's prior assertion that discrimination (or reward bias) in all its modes has all but disappeared, is unclear.
Researchers doing empirical work on discrimination have long been aware that what Loury's calls “discrimination in contact,” particularly if it includes the sacrosanct private grounds of “networking” and “informal associations,” can play an exceptionally invidious role in excluding members of subaltern groups. It is at this stage of the text that Loury even references his 1976 MIT dissertation explicitly:
Elsewhere [my dissertation] I have demonstrated (in the context of a theoretical example) that notwithstanding the effective prohibition of discrimination in contract, historically engendered economic differences between racial groups can persist indefinitely when discrimination in contact continues to persist (p. 129).
But is discrimination in contact the only barrier to racial economic equality in the USA? It is possible only if one believes that there is equal pay for equal work, there is no occupational crowding, opportunities for entry and promotion are not racialized, there is no racial steering by realtors, and banks and other lending institutions do not practice redlining. Would Loury's “historically engendered differences between racial groups” also disappear if we could eliminate racialized patterns of social interaction? if is possible only if one believes that there is no black- white wealth gap. Indeed, do we eliminate racialized patterns of social interaction? If we do not, for Loury, is there any remedy to the situation?
Admirably, in 2002 Loury's detailed critique of liberal individualism leads him to reject race blindness in favor of racial egalitarianism. (I prefer to describe the distinction as one between a color-blind and a color-fair society.) A society committed to color-fairness will recognize group identity for the purpose of achieving the just cause of redressing inherited group disparities. Loury's language is incandescent and passionate in this context:
The self-limiting patterns of behavior among poor blacks are not a product of some alien cultural imposition on a pristine Euro-American canvas. Rather such “pathological” behavior by these most marginal of Americans is deeply rooted in American history. It evolved in tandem with American political and economic institutions, and with cultural practices that supported and legitimated those institutions-practices that were often deeply biased against blacks. So, while we should not ignore the behavior problems of this so-called underclass, we should discuss and react to them as if we were talking about our own children, neighbors and friends. This is an American tragedy. It is a national, not merely a communal, disgrace (p. 105).
The rhetoric is also paternalistic. The presumption is that the black poor are defective, albeit their defectiveness is borne of an oppressive history. Still here one can see the basis for a straightforward endorsement of reparations. But Loury is not willing to take that step. For him, the past is “ethically indefensible,” but it does not warrant compensation in the form of reparations. While the “mature” Loury now countenances the legitimacy of affirmative action, he will not endorse efforts to pay the debt associated with an ethically indefensible history. Why not?
He says little about his objections to reparations in this text, apart from the suggestion that the harms cannot be measured or demonstrated systematically. Of course, economists have been computing the magnitude of monetary damages for both psychological and physical harms, even across generations, for many years. Why not for damages inflicted upon African Americans? Loury (p. 128) says “history can be messy stuff. Teasing out causal implications across the centuries of procedural violations is impossibly difficult.”
Difficult, yes. Impossible, no. And why consider “centuries” rather than generations? In a paper published in the American Journal of Sociology in 2001 in conjunction with Jason Dietrich and David Guilkey, I demonstrated statistically using IPUMS data for 1880–1990 that occupational status outcomes for blacks in 1980 and 1990 were weighted downward heavily by the labor market deprivations faced by their forebears 100 years earlier. While Loury has made a philosophical pitch for reparations, he has retreated before jumping over the cliff where some of us already have taken the plunge.
Elsewhere, in a short commentary in Black Issues in Higher Education, (2001) Loury (2001) has offered a compact statement of his objections to reparations. Note that these objections have a fundamentally “practical politics” tone to them rather than constituting a principled rejection. He makes three points:
Recent immigrants will not engage with “a racial reform movement built around the theme of reparations [which] is a backward-looking slogan in what has increasingly become a forward-looking country.”
Reparations are particularistic when the political coalition needed to confront poverty must be universal in scope.
Gaining reparations will mean black Americans can make no further claims on the American polity for any future problems that persist.
Of course, reparations on behalf of any group subjected to a grievous injustice is inherently particularistic, particular to the specific violations of human rights to which members of the group have been subjected. All Americans, including recent immigrants, have to be persuaded of the social value of reparations for African Americans—just as recent immigrants consented to reparations for Japanese Americans incarcerated by the US government during World War II, although recent immigrants were not here. Indeed, an effectively designed reparations program provides an avenue to move forward beyond a persistent history of racial injustice and inequality. It also is an avenue to complete healing and closure; therefore, one of the positive goals of reparations is to bring an end to the necessity of black Americans making further race-based claims for remedies to the American policy.
In my view, both Loury's reward bias and development bias coexist at important levels in the USA for black Americans, a double whammy. This is not unusual. Empirical evidence of both forms of bias also is apparent in other parts of the world as well. For example, the huge extent of historical educational and social deprivation for blacks in South Africa certainly meets Loury's criterion for development bias but the estimated levels of wage discrimination after controlling for schooling levels and other individual characteristics, both before and after the end of apartheid, are staggering signs of reward bias (Allanson, Atkins, and Hinks 2002 and Sherer 2000). Data from Brazil displays a similar combination of adverse effects for Afro-Brazilians (Lovell and Wood).
Conventional economic theory, to which Loury generally appears to subscribe, eradicates reward bias as a persistent phenomenon. Empirical evidence of sustained discrimination creates a conundrum for conventional theory. The most popular avenue for addressing the puzzle in an era where theoretical economics gives pride of place to asymmetric information is the statistical theory of discrimination.
A central difficulty with the statistical theory of discrimination is it is premised on the prevalence of the belief that one group is generally less productive than another. If the theorist posits that the belief is valid, then he or she reverts to Loury's exclusive emphasis on development bias, leading in turn to Barbara Bergmann's wonderfully apt observation that “the statistical theory of discrimination is really human capital theory in drag.” If the theorist posits that the belief is wrong, then he or she has to establish why agents do not learn they are mistaken.
Loury does address this dilemma with the statistical theory of discrimination on page 37 of his text where he takes up the questions I have raised about the theory (e.g., Darity 1998, Darity and Mason 1998, Darity and Williams 1985) without mentioning my work. But it is odd that he is so concerned with preservation of the statistical theory of discrimination, since he deems reward bias unimportant. This also lends a certain deceptiveness to the structure of the text. Pages 36–54 are devoted to a series of examples of why agents need not ever learn the true distribution of abilities between the two groups (whatever it actually is). A series of clever self-fulfilling prophecy effects are central to Loury's answer.
But, again, empirically how often does discrimination take the statistical form? Audit studies that provide direct evidence of discrimination suggest that raw racism is operative, unmitigated by additional information. A correspondence test conducted by Bertrand and Mullainathan (2004) involved sending 5000 letters of inquiry to employers who had advertised positions publicly where the race of the “applicants” was signaled by their names and where artificial curriculum vitae were enclosed designed to make the black and white applicants equivalent in credentials. For example, presumably Lakisha Washington would be assumed to be black, while presumably Caitlin O'Brien would be assumed to be white by an employer. In the first phase of their study, letters with white-sounding names received almost twice the rate of callbacks as letters with black-sounding names.
Most damning for the statistical theory of discrimination are the results from the second phase of their study. Bertrand and Mullainathan sent out a new set of letters with enhanced resumes. They found that the callback rate rose for “applicants” with white sounding names significantly, while the callback rate did not change perceptibly for the “applicants” with black sounding names. More information did not translate into a change in the employers' responses.
The statistical theory of discrimination seeks to explain racist practices without resort to impugning the practitioners with racism. Direct evidence of discrimination of the type generated in the Bertrand and Mullainathan study suggests that racist practices are founded in racism. Throughout the Anatomy, Loury makes great hay about distinction between “stigma” and “racism” preferring, rhetorically, the former. Again, the empirical evidence at hand suggests that no such distinction is needed. The wellspring of racial inequality is racism.