American Economic Association

Is Equal Opportunity Enough? Author(s): Glenn C. Loury Reviewed work(s): Source: The American Economic Review, Vol. 71, No. 2, Papers and Proceedings of the Ninety- Third Annual Meeting of the American Economic Association (May, 1981), pp. 122-126 Published by: American Economic Association Stable URL: http://www.jstor.org/stable/1815704 . Accessed: 26/12/2012 09:07

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The American Economic Review.

http://www.jstor.org

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions ECONOMICS OF AFFIRMA TIVE ACTION

Is Equal Opportunity Enough?

By GLENN C. LoURY*

Affirmative action policies have come in- taking actions to facilitate economic pro- creasingly under attack in recent years. Both gress for minority group members. This in the courts and in public discourse ques- would seem to be the logical first step in tions have been raised about the legitimacy constructing an intellectual basis for affir- of government efforts on behalf of blacks mative action policies. Of course, philoso- and other racial minorities.' The criticism phers and legal scholars interested in theo- seems to have two central themes. First, it is ries of distributive justice have devoted argued that those policies which have been considerable attention to this question in the tried have not had a noticeable effect on the past ten years. (See R. Dworkin and T. economic standing of minority group mem- Nagel.) The approach adopted here differs bers. (See James Smith and Finis Welch.) from these earlier efforts in two ways. First, They thus constitute yet another example of I shall endeavor to meet the empirical argu- costly but ineffective government regula- ment directly, by pointing to evidence which tion, according to this view. The second suggests that significant racial economic dis- theme strikes more deeply at the foundation parity persists. Secondly, I will treat the of these policies. Its adherents argue that philosophical argument in a manner in even if effective programs could be de- keeping with the economist's traditional ap- signed, they ought not be implemented. proach to the question of the desirability of There have been philosophical and empiri- laissez-faire. This approach is based upon cal arguments advanced to support this con- the concept of market failure. Intervention clusion. Essentially, the philosophical argu- is favored over laissez-faire when, because of ment states that it is wrong for government some externality, the market outcome is to intervene on behalf of certain groups inefficient. Below I argue that an analogous (and thus, necessarily, at the expense of "market failure" contributes to the mainte- others); this amounts to reverse discrimina- nance of economic inequality between racial tion-a visiting of the fathers' sins upon the groups in our society. As such, intervention sons.2 The empirical argument concludes which redresses this inequality is warranted. that, moral issues aside, such intervention is unwarranted because the consequences of I. The EmpiricalArgument historical discrimination have been (or will soon be) largely eliminated. (See B. Watten- Since the passage of civil rights legislation berg and W. Wilson.) in the early 1960's there have been profound In this essay I would like to offer a de- changes in the economic experience of ra- fense of affirmative action policies against cial minorities in this country. A number of the second of these thematic criticisms. That analysts have called attention to this change, is, I shall hold in abeyance questions con- observing that traditional discriminatory cerning the efficacy of particular program- practices, such as unequal pay to equally matic efforts, and concentrate instead on skilled workers, have been dramatically re- whether government should in principle be duced. (See Richard Freeman, and Smith and Welch.) Moreover, when the data are *Professor of economics, University of Michigan. 'The arguments of this paper are not intended to disaggregated by cohorts, one finds that the apply to affirmative action for women. disadvantage in wages of younger minority 2This argument is developed at length in N. Glazer. workers is quite small. (See Smith and

122

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions VOL. 71 NO. 2 ECONOMICS OFAFFIRMATIVE ACTION 123

TABLE I -PERCENT OF MALE WORKERS UNEMPLOYED, BY RACE AND AGE, 1970-79

Age 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 16-19:White 13.7 15.1 14.2 12.3 13.5 18.3 17.3 15.0 13.5 13.9 Nonwhite 25.0 18.9 29.7 26.9 31.6 35.4 35.1 37.0 34.4 31.5 20+:White 3.2 4.0 3.6 2.9 3.5 6.2 5.4 4.6 3.7 3.6 Nonwhite 5.6 7.2 6.8 5.7 6.8 11.7 10.6 10.0 8.6 8.4

Source: Economic Report of the President.

Welch.) Thus, as younger workers continue for the currently observed unemployment to enter the labor force and older workers disparity. However, the data on unemploy- retire, differences by race in the wages of ment certainly suggest that the progress of equally skilled workers may be expected to nonwhite workers in the post-civil rights era attenuate. This has led some to question the cannot be accurately assessed by looking at need for affirmative action, since minority earnings alone. Moreover, these data are workers seem to be catching up without clearly consistent with the hypothesis that government help. there exists racial discrimination in employ- This conclusion seems to me premature ment opportunities. In a market char- because it is based on only one aspect of acterized by excess supply and downward economic status- the earnings of employed price rigidity (for example, the market for workers. No observer of the economic expe- young workers with a minimum wage floor) rience of nonwhites in the past decade can buyers must use some device for rationing have failed to notice that unemployment their purchases among the more numerous rates are much higher for minority workers sellers. The possibility that race is among than workers as a whole. Table 1 presents the characteristics influencing a worker's unemployment rates for white and nonwhite position in this job queue ought not be male workers by age for the years 1970-79. ignored.3 It is apparent that nonwhite workers are There is another sense in which compari- unemployed roughly twice as often as their sons of the annual earnings of racial groups white counterparts, and that unemployment incompletely represent their respective eco- constitutes a chronic problem for young and nomic positions. The use of cross-section nonwhite workers. While these aggregate data from a sequence of years does not data do not control for differing individual allow the analyst to discern what happens to characteristics (for example, education) the incomes of particular individuals over which may account for part of this racial time. There is some evidence that patterns disparity, one study of youth unemployment of year-to-year earnings mobility are quite has found that no more than half the racial different for white and nonwhite workers. difference in unemployment rates among For example, using longitudinal data, several young workers can be explained in this way. researchers have found that while the entry (See Martin Feldstein and David Ellwood.) level wages of young black and white male Thus, even if racial differences in the workers of similar skills are now quite close, earnings of similarly skilled employed work- the subsequent rate of wage growth is sig- ers were to disappear in the near future, a nificantly smaller for the black workers. (See continuation of current trends in the unem- Edward Lazear, Saul Hoffman, and Greg ployment experience of nonwhite workers Duncan.) An earlier analysis of occupa- would imply significant economic disparity between the groups. These figures do not prove that minority workers are currently 3Charles Betsy has regressed unemployment fre- discriminated against in employment op- quency and duration measures on a variety of explana- portunities instead of wages, nor do they tory variables, finding quite different coefficients for show that historical discrimination accounts blacks and whites.

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions 124 AEA PAPERS AND PROCEEDINGS MA Y 1981 tional mobility among mature male workers legitimacy of affirmative action. The nature showed blacks to be less upwardly mobile of the ethical problem should be clear: Ra- out of low-paying occupations and more cial minorities are undoubtedly worse off downwardly mobile out of high-paying oc- today by virtue of the historical use of pro- cupations than whites (See Bradley Schiller). cedures which did not respect their liberty. Moreover, there is evidence that black heads Yet, to use the power of the state to "cor- of households in poverty in a given year are rect" history's wrong doing is to condone considerably more likely than whites to re- disregard for the liberty of those citizens not main in poverty in the following year (see so favored. The aphorism "two wrongs don't Lee Lillard and Robert Willis), while black make a right" would seem to apply. families with "high" incomes in one year are One way to think about this problem is to much less likely than whites to retain that inquire whether, in theory, we should expect status in the following year (see my paper the continued application of racially neutral with Jerome Culp). procedures to lead eventually to an outcome It would appear then that, while the na- no longer reflective of our history of dis- ture of economic inequality between the crimination. If the answer to this question races has undergone significant change, the were negative, then adherence to a policy of gap does not appear to be withering away of equal opportunity alone would condemn its own accord. If this conclusion is accepted, those whose rights had historically been a question then arises as to what, if any- violated (and their progeny) to suffer indefi- thing, government should do about it. nitely from what most would regard as ethi- cally illegitimate acts. Since this would (pre- II. The PhilosophicalArgument sumably) be an undesirable state of affairs, a case for intervention would thereby be This question is a crucial one for advo- made. Of course, even if the effects of his- cates of affirmative action. In the period torical discrimination were to eventually be since World War II the principle that an eroded through the application of racially individual's opportunities for advancement neutral procedures, this "correction" might ought not depend on race has come to be take so long as to be of little practical sig- broadly accepted in our society. Given that nificance. The point here is that there are racial discrimination in the private sector is reasons (to be discussed presently) to be- not currently practiced, government efforts lieve that our society operates so as to pass on behalf of minorities would seem to con- on from one generation to the next that tradict this basic principle. Advocates of racial inequality originally engendered by affirmative action now argue that the his- historical discrimination. tory of discrimination has created an The above discussion is intended to per- environment in which equal opportunity suade the reader that a certain aspect of the alone would not permit minority groups to dynamic performance of market economies gain economic parity. Their focus is on the is important in evaluating the ethical legi- results of the income determination process. timacy of affirmative action. The choice be- Critics, on the other hand, note that one tween public policy limited to equal oppor- cannot logically urge the necessity of equal tunity or extended to affirmative action, I treatment while simultaneously demanding submit, should depend upon the extent to special favor. Their focus is on the neutral- which we are confident of the ability of the ity of the process itself. This distinction be- market to naturally erode historically gener- tween the fairness of procedures and the ated differences in status between groups. It fairness of outcomes is a critical one in is in this sense that this choice is analogous social philosophy,4 and in my judgment to the one economists face when consider- constitutes the core of the debate over the ing whether public intervention in the marketplace is desirable. In the latter in- 4See R. Nozick's criticism of "end state" theories of stance the ability of laissez-faire to attain justice. John Rawls adopts an opposing view. an efficient allocation of resources is the

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions VOL. 71 NO. 2 ECONOMICS OFAFFIRMATIVE ACTION 125

crucial issue. Here I suggest that we focus choices, and that an identical distribution of on the extent to which equal opportunity innate aptitudes characterizes each genera- eventually erodes discrimination-induced in- tion of majority and minority workers. Thus, equality when judging the appropriateness in the absence of historical racial dis- of intervention via affirmative action. crimination, we should expect that the eco- This question, like the question of when nomic status of minority and majority group does competition lead to efficient resource members would be equal, on average. I then allocation, is necessarily a logical query inquire whether, in this idealized world, the about the operations of an idealized eco- competitive labor market would function in nomic system. Like the efficiency question, such a way as to eventually eliminate any it may be studied by developing a theoreti- initial differences in the average status of cal model of the social phenomenon at is- the two groups. sue, and seeking conditions within the con- The results obtained depend upon whether text of that model under which the desired only income, or both income and race, in- outcome obtains. The basic human capital fluence the community to which a family theory of earnings determination, extended belongs. In the former instance, with some to allow for intergenerational effects, is well additional reasonable assumptions, one can suited to an investigation of this sort.5 show that equal opportunity always leads Elsewhere (1977) I have pursued this ques- (asymptotically) to equal outcomes. In the tion at some length; however, space limi- latter case, however, it is not generally true tations necessitate that I merely summarize that historical differences attenuate in the that investigation here. A model is developed face of racially neutral procedures. Exam- in which job assignments are made under ples may be constructed in which group conditions of equal opportunity, based solely inequality persists indefinitely, even though on an individual's productive characteris- no underlying differences in tastes or ability tics. However, the individual's acquisition of exist. productive characteristics is favorably in- This last result arises because, when there fluenced by the economic success of the is some racial segregation among communi- individual's parents. Thus, the deleterious ties, the intergenerational status transmis- consequences of past discrimination for the sion mechanism does not work in the same racial minority are reflected in the model by way for minority and majority families. An the fact that minority young people have intragroup externality is exerted, through lo- less successful parents, on average, and thus cal public goods provision, by the (relatively less favorable parental influences on their more numerous) lower income minority skill acquisition processes. Further, the families on higher-income minority families model posits that families are grouped to- of the same community. Because the racial gether into clusters or "communities," and composition of one's community depends in that certain local public goods important part on the choices of one's neighbors, this to subsequent individual productivity (for kind of effect cannot be completely avoided example, education) are provided uniformly by an individual's actions. As a conse- to young people of the same community. quence, the ability of equal opportunity to This provides another avenue by which bring about equal results is impaired by the background influences achievement, since desire of majority (and minority) families to the nature of the community to which a share communities with their own kind.6 family belongs also depends on the eco- Since this social clustering of the races seems nomic success of the parent. In order to pose the question most sharply, 6L. Datcher finds that, for young black males, the it is assumed that all individuals have identi- racial composition of the community in which they cal preferences with respect to economic were raised has a significant influence, other things equal, on subsequent earnings. An increase of ten per- centage points in the fraction white in the community 5Such an extension is provided in Gary Becker and implied an increase in subsequent annual earnings of 3 Nigel Tomes, and my forthcoming paper. percent.

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions 126 AEA PAPERS AND PROCEEDINGS MAY 1981 to be a continuing feature of our society, the employment: What's the Problem," Nat. theoretical analysis leads to the conclusion Bur. Econ. Res. disc. paper, 1979. that intervention may be justified. R. Freeman, "Changes in the Labor Market for Black Americans," Brookings Papers, III. Conclusion Washington 1973, 1, 67- 131. N. Glazer, Affirmative Discrimination, Basic I have argued that current economic dif- Books 1975. ferences between whites and nonwhites are S. Hoffman, "Black-White Life Cycle Earn- such as to obviate the conclusion that the ings Differences and the Vintage Hy- historical effects of discrimination have (or pothesis," Amer. Econ. Rev., Dec. 1979, will soon be) dissipated. Additionally, I have 69, 855-867. suggested some reasons why a laissez-faire E. Lazaer, "The Narrowing of Black-White policy of equal opportunity, but not affir- Wage Differences is Illusory," Amer. mative action, could leave minority group Econ. Rev., Sept. 1979, 69, 553-64. members perpetually constrained by histori- L. Lillard and R. Willis, "Dynamic Aspects of cally practiced discrimination. Thus, the Earnings Mobility," Econometrica, Sept. second thematic argument against affirma- 1978, 46, 985-1012. tive action, mentioned in the introduction, is G. Loury, "A Dynamic Theory of Racial deemed unsatisfactory. Income Differences," in P. A. Wallace, ed., Women Minorities and Employment REFERENCES Discrimination, Lexington Books, Lexing- ton 1977. G. Becker and N. Tomes, "An Equilibrium , "Intergenerational Transfers and the Theory of the Distribution of Income Distribution of Earnings," Econometrica, and Intergenerational Mobility," J. Polit. forthcoming. Econ., Dec. 1979, 87, 1153-89. T. Nagel, Equality and Preferential Treatment, C. Betsy, "Difference in Unemployment Ex- Princeton Univ. Press, Princeton 1977. perience Between Blacks and Whites," R. Nozick,Anarchy, State and Utopia, Oxford Amer. Econ. Rev. Proc., May 1978, 68, 1974. 192-97. John Rawls, A Theory of Justice, Harvard J. Culp and G. Loury, "In Defense of Public Press: Cambridge, Mass. 1974. Policy Aimed at Reducing Racial Eco- B. Schiller, "Relative Earnings Mobility in nomic Disparity," Rev. Black Polit. Econ., the United States," Amer. Econ. Rev., Dec. Summer 1980. 1977, 67, 926-41. L. Datcher, "Effects of Community and J. Smith and F. Welch, "Race Differences in Family Background on Achievement," Earnings: A Survey and New Evidence," mimeo, Univ. Michigan 1980. R-2295-NSF, Rand Corp. 1978. G. Duncan, "The Determinants of Wage B. Wattenberg, The Real America, Doubleday Growth, 1971 - 1976," mimeo., Univ. 1974. Michigan 1979. W. Wilson, The Declining Significance of Race, R. Dworkin, Taking Rights Seriously, Duck- Chicago 1978. worth 1977. U.S. Council of Economic Advisers, Economic M. Feldstein and D. Ellwood, "Teenage Un- Report of the President, Washington 1980.

This content downloaded on Wed, 26 Dec 2012 09:07:18 AM All use subject to JSTOR Terms and Conditions