The Economics of Discrimination WARREN L. COATS, JR. ECONOMIC DISCRIMINATION, to distinguish it from other types, generally refers to discrimination in the market place and in particular in labor markets. Racial and sexual discrimination in labor markets can lower wages or raise unemployment or both. In his important seminal work on the subject, Gary Becker treated discrimination as follows: If an individual has a taste for discrimination, he must act as if he were willing to pay something, either directly or in the form of reduced income, to be associated with some persons instead of others. When actual discrimination occurs, he must, in fact, either pay or forfeit income for this privilege. This article examines some of the wage and employment implications of economic discrimination by discussing the extent to which discrimination has affected Negro and female employment and wages, by discussing the role of free market capitalism in discrimination, and by discussing the major stumbling blocks to the full economic participation of blacks and women in American society. Although most references are to blacks and women the analysis is meant to apply to any groups experiencing discrimination. I IN THE 1960 censuses, the latest for which these figures are available, average hourly earnings of nonagricultural employed persons show that male wages averaged 65 percent higher than female wages (or $2.79 to $1.70) while white wages averaged 60 percent above those of nonwhites (or $2.58 to $1.61).* Interesting regional differences emerge from these figures, contrasting the South with the rest of the country. The figures show that average hourly earnings in the South were 25 percent lower than elsewhere. More to the point, the average hourly earnings of nonwhites were about 60 percent greater elsewhere than in the South.a These figures suggest a substantial amount of the discrimination against nonwhites and females generally, with a more serious degree of racial discrimination in the South than that exhibited in the rest of the country. Although these results accurately reflect actual wage differentials, the figures greatly overstate the degree of discrimination. To prefer a white worker who is more productive over a nonwhite one who is less productive is no more to discriminate in our sense than to prefer a white worker 64 Winter 1974
who is more productive to one who is less productive. Discrimination exists when between two workers of equal productivity one is systematically preferred by some because of race or some other characteristic not directly related to the worker's productive contributions. The preference for a pretty secretary over a homely one, other things equal, is an example of such discrimination. If blacks predominantly hold lower pay. ing jobs because they possess fewer skills, their average earnings would be lower than that of whites even in the absence of discrimination. Discrimination exists when someone is paid less for doing the same job as someone else. The relative earnings fig ures reported above can be adjusted or standardized to remove the effect of differences other than race in average earnings. Victor Fuchs, from whose study of census data the earlier figures were taken, standardizes with respect to age, sex, education and city size. This means that earnings for whites versus nonwhites are compared for groups of the same age, sex, education and city size. Such standardization reduces the 25 percent average hourly earnings differential between the South and non-south by two-thirds but only reduces the comparable 60 percent differential between Southern and non-southern nonwhites by one-half.' Similar results are achieved with the total U.S. average hourly earnings for males versus females and for whites versus nonwhites. In the South the average hourly earnings of whites were 78 percent greater than for nonwhites and elsewhere were 32 percent greater. After standardization for differences in skill and city size the whitenonwhite earnings differential in the South was cut from 78 percent to 10 percent and was actually reversed in the rest of the country. The difference between male and female average hourly earnings was also eliminated in all region^.^ Harry Gilman conducted a similar investigation of white and nonwhite unemploy ment rates.6 He found that the average rate of nonwhite unemployment from 1953 to 1961 was double the average rate for whites (9.8 percent for nonwhites versus 4.8 percent for whites). However, if nonwhites are disproportionately represented in jobs or industries with above average unemployment rates even for whites, to that extent nonwhite unemployment rates would be higher than for whites, even if there were no discrimination. Gilman finds that by standardizing for these industrial and educational differences roughly half of the nonwhite-white unemployment difference is eliminated.6 He also found (a rather surprising result) that the unemployment differential was substantially less in the South than in the non-south. That is, nonwhite unemployment rates were greater than white unemployment rates by a substantially smaller margin in the South. Although there is considerable evidence that blacks are more discriminated against in the South than elsewhere' as seen in the standardized figures showing lower relative wages for nonwhites in the South, the unemployment figures seem contradictory, showing considerably lower unemployment differentials in the South than elsewhere, which is somewhat of a surprise. Hopefully what follows will shed light on this paradox. A proper evaluation of the significance of the economic impact of discrimination requires an understanding of the role of the capitalist system in generating the wage and employment differentials just discussed. Is capitalism somehow the cause of market discrimination or is it simply the conduit through which tastes for discrimination are transmitted and transformed into observable consequences? If the latter, how is this transformation influenced by the capitalist system? Does it make the life hfodern Age 65
of discriminators and discriminated against easier or more difficult? The answers to such questions should importantly influence our attitudes towards capitalism as a system to which minorities can turn in their search for economic progress. Gary Becker has provided a framework in which such questions can be answered. A man who discriminates racially can effectuate his tastes (or prejudices if you prefer) only by restricting his options. If he prefers not to work with blacks, the number of jobs open to him will be smaller than otherwise. As put by Becker, he indulges his tastes only at the expense of reduced income to himself. Employers who prefer not to hire blacks likewise thereby reduce the supply of labor available to themselves which will generally increase the average cost of that labor. Indulging one s tastes for racial discrimination incurs personal costs just as indulging any other tastes incurs costs. It does not follow, however, that all of the costs of racial discrimination fall on the discriminator. If large enough numbers decide not to buy Frank Sinatra s records, his income will suffer. So also, if large enough numbers decide that they prefer not to work with blacks or other minority groups or women, these groups incomes will suffer. Discrimination against blacks means that the demand for their services will be lower than otherwise. With an unchanged supply of black workers the lower demand means that blacks can only be fully employed at lower wages than whites. Lower wages are, in fact, the market s way of compensating for and overcoming white resistance to working with blacks. Racial discrimination in a free enterprise economy will lead to lower wages for nonwhites but will not generally affect their employment. However, in competitive free markets even discrimination s dampening effect on nonwhite wages will tend to be small. Employers whose tastes for discrimination are weaker than the community average or who even prefer to hire blacks will enjoy a competitive advantage. They will tend to hire all blacks because they are relatively cheaper and will find that they can more successfully undercut their competitors prices. This will expand their share of the market relative to those who discriminate, which also expands the demand for nonwhite labor, thus reducing the tendency for discrimination to depress nonwhite (or female) wages. I1 So FAR I have discussed racial discrimination of the sort consisting of firmly held unshakable preferences of association. This is the sort of behavior economists are inclined to call irrational, meaning simply that such behavior cannot be explained by profit-maximizing economic motivations, but then that is true of all tastes. The point is important, however, because it emphasizes that racial discrimination is not the result of capitalism but rather of the tastes of men. Furthermore the capitalist system can be seen to minimize the economic consequences of these unfortunate tastes. The clear implication is that those discriminated against have an interest in strengthening the freedom of competitive markets. There is another reason for lower Negro and female wages for equally productive workers which may be more important than the ones discussed so far. Information is itself a scarce commodity; information is not free. A woman or a Negro may be viewed as a more risky prospect than a male caucasian. A wage differential would have to emerge to compensate the employer for this greater risk if women or blacks are to receive an equal chance at jobs. More specifically, assume that of those 66 Vinter 1974
workers possessing a particular skill, we somehow know that female applicants are on average just as productive in every relevant sense as are male applicants. However, assume also that the performance of any individual woman is less reliable in the sense that the dispersion (variance) of female performance around their average is wider than for men. Some women will be far more productive than the average while others will be far below it due to, say, less regular working habits (because of sickly children or other domestic obligations, time off for pregnancy, having to move away to follow a husband s career etc.). Other things equal an employer will then find a female applicant a more risky proposition and at the same wage (if the employer is risk averse) will prefer a man. This will tend to depress the wages of women relative to men, even for the same work. Once again, however, the wage differential for women (or blacks) will compensate the employer for the extra risk, making women fdy competitive with men. In this case, however, the wage differential tends to be temporary. The lower wage will enable a woman to get the job in the first place and on-the-job experience will provide the employer with the information of whether this particular woman is above or below the average. If she is equal to or above the average the employer will then be willing to pay her more to keep her in his employ. Her pay will tend to rise. In these cases the wage differential is perfectly rational, profit-maximizing behavior for the firm and serves the very useful function of allowing the employer to acquire the information needed to ultimately increase the wage. The initial wage differential does not reflect any taste for discrimination on the part of the firm. However, such differentials may reflect prejudice in the sense of initially faulty information on the part of the firm about the pro- ductive potential of a given class of workers; but once again the wage differential allows the on-the-job exposure which will ultimately overcome and correct the faulty information. Clearly then Fair Employment Legislation and particularly equal pay for equal work movements (by women s liberation and civil rights groups) are very harmful for minority and female employment and economic advancement. Their chief effect is to convert the impact of discrimination and imperfect market information from creating wage differentials to creating unemployment differentials for these groups. As observed by Gilman: Given a preference for discrimination, the greater the pressure in an occupation or region for nonwhite-white wage equality, the greater will be the gap between equilibrium and actual wages, and the greater will be the reduction in employment opportunities for nonwhites relative to white w0rkers.o It is not a very pleasant fact that blacks can only be fully employed at lower relative wages, but of course neither is the existence of the discrimination which is responsible. If I were a young black being asked to accept a lower wage for the same work I would not expect myself to be very patient with such a state of affairs. Yet I believe that it is critically important, given the existence of prejudice, to choose lower relative wages and fuller employment rather than the reverse. Among the many important virtues of holding a job is the possibility that on-the-job contact and exposure will reduce racial prejudice-especially where it is the result of faulty information. According to Gilman : there has been a secular increase in the activities of organizations dedicated to the elimination of discrimination. If such activities do not parallel changes Modern Age 67
in tastes for discrimination, they will tend to make market adjustments more di6cult; they may thus limit rather than increase the employment opportunities of nonwhites. Increased effectiveness or coverage of minimum wages or unions would tend to have the same effect.1 Following Gilman s suggestion, minimum wages and unions may be important in understanding the wage and employment figures cited earlier. Employers will not hire workers whose wages exceed their productive contribution to the firm. Minimum wage laws therefore prevent firms from employing those low skilled persons whose productive abilities are below the minimum. As pointed out by George Stigler, one of the principle effects of minimum wages is to injure some of the lowest paid workers by forcing them into even lowerpaid occupations exempt from the act, one of which is unemployment. 11 But as noted before, since nonwhites and females are over-represented among the least skilled workers they bear a disproportionate share of the harm from such laws. Specifically the minimum wage tends to force Negro relative wages still lower for those who can find employment uncovered by the law, while increasing Negro unemployment by making some unemployable. In addition to the increasingly comprehensive federal minimum wage law, over half of the individual states have broadened the covered areas still further with minimum wage laws of their own. Colin and Rosemary Campbell, among others, have conducted a careful study of the effects of these state laws on state unemployment rates. They found that from 1950 to 1965, the rate of unemployment in major labor market areas with state minimum wage laws was on the average about 0.6 percentage points above that in areas without such laws. 2 Specifically the average unemployment rate for states with such laws was 5.2 percent while 4.6 percent for those states without a minimum wage law.i8 Although a part of this unemployment differential was explained by greater unionization in minimum wage states, not all of the difference could be explained that way and seemed instead to result from the law itself.i4 Edward Banfield in his book, The Unheavenly City, suggests that several additional factors aggravate the tendency of minimum wage laws to create unemployment rather than lower wages in uncovered employment. First, affecting the demand for labor, he cites the general tendency to over-price labor even in areas not covered by the law. This appears in the form of a general social concensus about what the lowest respectable wage ought to be, creating an informal minimum wage constraint. Contributing to this problem is the pressure on government agencies at all levels to set an example of enlightened wage policy, that is, to pay labor more than it is worth. 15 Second, affecting the supply of labor, Banfield cites the tendency of many lower class workers with exaggerated expectations to scoff at working for peanuts opting instead for unemployment and social workers. This problem is particularly prevalent among teenaged males. B Gary Becker s analysis of discrimination in the market place also suggests that the economic consequences of discrimination are potentially worse in monopolized sectors of the economy. The potentially is added because this result is contingent on monopolists having a taste for discrimination. Some labor unions are among the most monopolized sectors of the economy, although the general importance of unions can be easily overstated. Those unions that succeed in increasing wages for their members do so by making the supply of that 68 Winter 1974
type of labor artificially scarce. The workers thereby displaced swell the ranks of nonunion labor driving those wages down relative to union wages. More importantly, however, this means that unions by their very nature must discriminate. The union replaces the market s impartial docation of jobs and workers with its own selection of which workers will get the artificially scarce jobs. Naturally this discrimination frequently takes a racial or sexual form. Among the other factors contributing to the wage and employment differentials cited, a relatively lower average wage level for women is no doubt a result of excluding women from many types of jobs thus swelling their supply to the jobs they are allowed to hold. This situation is aggravated by unions. Happily some women s liberation organizations are asking for the repeal of restrictive labor practices laws (Le. fair labor practices laws ) which have arbitrarily restricted the job opportunities of women, such as the 40 hour week maximum prohibition against night work, limits on weights lifted etc. Practically everything the government has done to regulate behavior in labor markets, such as occupational licensure and the entire Fair Employment Practices Act, Civil Rights Acts etc. has in fact restricted competition in such markets and made nonwhites and women economically worse off. The women s Equal Rights Amendment is a welcomed step in the other directiop. I11 RETURNING to the earlier paradox of Southern wage and employment differentials, we can now see why nonwhite relative wages are lower in the South, where discrimination is greater, while nonwhite to white unemployment differentials are greater in the non-south. The reason is that the employment compensating effects of lower wages for nonwhites is allowed to work more freely in the South than elsewhere. Union strength is weaker in the South, few Southern states have state minimum wage laws and a large fraction of production in the South is carried on in firms uncovered by the Federal minimum wage law. Most important, however, may be the fact that social pressures for equal pay for equal work, in the context of the white-nonwhite comparison, are less great in the South than in the non- South. All of the above-named factors suggest a greater ability to discriminate on the wage side in the South than elsewhere.... The lower wage ratios (NW/W) in the South are probably due to a regional difference in the pressure toward nonwhite-white wage equality. Thus, smaller unemployment-rate differentials in the South than non- South are consistent with the popular concept of greater discrimination in the South if the differences between equilibrium and actual wage ratios (NW/ W) are smaller there than in the non- So~th?~ There would be great benefit in removing the government s perverse influence and increasing the employment of those with very low skills even if at very low wages. According to Banfield : The main beneficiaries, however, would be the workers themselves. A job can be much more than a source of income. It can be an opportunity to learn things, to test one s ability to stand up to strains, to get out of the house and away from home, and to feel that one is a part, however small, of a worthwhile undertaking. Especially for a male, the job (along with other things) helps establish one s identity and supports one s self-respect?8 If for some workers such wages yield in- Modem Age 69
adequate incomes, that problem should be handled in a way that does not interfere with the wage and price structure, for example with a negative income tax. Economic discrimination against blacks undeniably influences their incomes and economic status, but due to our competitive enterprise system the market consequences of that discrimination are modest. Black incomes tend to be lower on average because blacks have less on average to offer an employer. To improve economically blacks must increase their skills and improve their work habits. Government s many attempts to help with training programs, wars on poverty, and protective laws have more often failed or hindered. Government s greatest contribution would be to remove itself from the picture. Although other types of racial discrimination surely share much of the blame for the lower level of Negro skills and although an increasing social consciousness of the immorality of racial discrimination must be encouraged and allowed to erode the remaining areas of discrimination, I do not believe that the Negro s greatest hope lies in any of these. In my judgment, economic equality for our black brothers will come only when they themselves acquire those middle class American values of self-determination and self-reliance, stop pointing at and to others as their oppressors and/or saviors and step out on their own to learn and work and save, and when the government steps out of the way and lets them do it. The Economics of Discrimination, University of Chicago Press, 1957, p. 6. Victor Fuchs, Differentials in Hourly Earnings by Region and City Size, 1959, N.B.E.R. Occasional Paper 101, 1967, p. 3. Ibid., p. 7. Ibid., p. 20. Sbid., p. 20, Table 11. Harry Gilman, Economic Discrimination and Unemployment, American Economic Review, Dec. 1965, pp. 1077-1096. Ibid., p. 1085. Becker, The Economics of Discrimination, Chicago, 1957, Chapters 7-8. ugilman, op. cit., p. 1091. lbid., p. 1093. llquoted in Edward Banfield s The Unheavenly City, p. 96. C. D. and R. G. Campbell, State Minimum Wage Laws as a Cause of Unemployment, South. ern Economic Journal, April, 1969, pp. 331332. lbid., p. 326. YW., p. 331. -Banfield, p. 99. Edwin Harwood, Youth Unemployment-A Tale of Two Ghettos, The Public Interest, Fall, 1969, p. 87. Gilman, op. cit., p. 1092 and 1093. -Banfield, op. cit., p. 103. Winter 1974