Skip to main content

Affirmative Action: Opening up workplace networks to Afro-Americans

In the workplace, as in many aspects of their lives, Afro-Americans as a group are systematically unconnected to the essential network resources that most other Americans take for granted. Their workplace isolation involves far more than the impenetrable “glass ceilings” associated with the nation’s boardrooms. It is pervasive, hampering working-class laborers seeking well-paying jobs on a construction work crew just as it does upwardly mobile college graduates.

This isolation can be addressed, but only partially, by access to improved educational opportunities. Such opportunities might be able to compensate for network deficiencies in the long run, mainly through introducing Afro-American students to other students with wider network ranges, but the process is painfully slow. For a group as unconnected as Afro-Americans, it may take multiple generations. Speeding up this process requires affirmative action.

How Networks Work

The fundamental insight of network sociologists is that characteristics of a person’s network of ties to others structure the flow of information that both socializes him as he grows up and provides social resources critical for competent functioning as an adult. Economic behavior in particular, as noted by Mark Granovetter in Getting a Job, is “heavily embedded in other social processes that closely constrain and determine its course and results.” Without denying the importance of individual attributes in accounting for success, network sociologists nevertheless consider past and present structures of network ties and their attendant information flows to be of at least equal importance in explaining a person’s achievements. As sociologists Nan Lin, Walter Ensel, and John Vaughn explain, the network approach “challenges the assumption that the labor market is essentially an open and competitive arena where specifications for a job and the necessary skills and competence are easily matched and where information about job and applicant availability is widely diffused.”

Karen Campbell, Peter Marsden, and Jeanne Hurlbert have demonstrated that networks operate as crucial resources for individuals in three ways. First, they help them to find a job, a home, or even a spouse. Second, they provide access to influential people. And, finally, they promote bargaining skills that enlarge the range and influence of networks even more. As Granovetter notes, “ are not made up of random jumps from one job to another, but rather…individuals rely on contacts acquired at various stages of their work-life and before. One important result of this is that mobility appears to be self-generating: the more different social and work settings one moves through, the larger the reservoir of personal contacts he has who may mediate further mobility.”

Building Networks

No ethnic group has ever achieved success in America by relying solely on education. Instead, immigrants—even ethnic groups that have powerful religious and social sanctions against such inter-ethnic bonding—have been able to make inroads into established American networks. Welcomed to the neighborhoods, clubs, and churches of existing groups, newcomers have quickly been able to develop the ties that extend their network ranges, especially in job searches and in getting credit for mortgages and small businesses.


Ethnic intermarriages have also contributed to immigrants’ success, both by extending their range of network ties and by involving them in new dense networks of strong affinal ties. Intermarriage enhances network bargaining skills in two ways: it encourages creative thinking about networking (imagine a Jewish groom’s first Sunday dinner with his Italian wife’s family), and it teaches traditions of network bargaining peculiar to the spouse’s ethnic group. More broadly, intermarriage encourages cultural and intellectual innovation. The explosive growth of artistic, scientific, technical, and entrepreneurial innovations in the United States is certainly linked to the high rate of intermarriage of its ethnic groups.

Afro-Americans, however, have been almost completely isolated from this national process. They are, for example, residentially isolated (see tables 1 and 2). As Douglas Massey and Nancy Denton, authors of American Apartheid, note of the data in table 1, “The isolation indices reveal the full extent of black racial seclusion within U.S. metropolitan areas, both northern and southern. The average value changed little over the decade and remained close to 66% in both regions; and in 1980 no metropolitan area displayed an isolation index under 50%. Despite the legal banning of discrimination and the apparent easing of white racial hostility, blacks and whites were still very unlikely to share a neighborhood within most metropolitan areas. In many cases, the degree of black spatial isolation was extreme.” There are, to be sure, some positive signs. Table 2 indicates modest recent improvement in some cities. And a surprising level of interracial contact, mainly originating at work, has been reported in polls despite the continued spatial segregation. Even so, it is astonishing that after living almost three centuries in this country, Afro-Americans have remained so spatially isolated.

Partly related to the group’s spatial segregation is its marital isolation. The intermarriage rates of Afro-Americans with non-Afro-Americans are far lower than those of other ethnic groups. Researchers often express the extent of intermarriage in terms of the proportion of women in a given group who marry men from other groups. But that measure is misleading because it fails to take account of compositional factors such as the relative sizes of the groups involved. A much better approach, that taken by Stanley Lieberson and Mary Waters in From Many Strands, is to measure intermarriage in terms of the rate that would be expected if spouses were chosen randomly in the population. Column 1 of table 3 shows that Afro-American women have by far the highest in-marriage rate of all ethnic groups. But this figure grossly underestimates the true level of isolation by not taking account of the proportion of all Americans who are Afro-American. Column 2, the proportion of husbands of each ethnic group in the population as a whole, is the proportion of such husbands we would expect to find in each ethnic group under conditions of random marital choice. The proportion of Afro-American husbands in the population as a whole, 7.4 percent, is what the in-marriage rate of Afro-American women would be if marital choices were random. Column 3 shows the ratio of in- to out-marriages (for Afro-American women, 74.4). Here we begin to see the true extent of Afro-American marital isolation. Column 4, the ratio of husbands within each group to husbands outside the group for all women outside the group, indicates that the ratio of Afro-American husbands to non-Afro-American husbands for all non-Afro-American women is .002. Column 5—the odds ratio derived from the ratios of Afro-American to non-Afro-American women’s marital choices—is the best measure of Afro-American marital isolation. It tells us, quite simply, that the odds that an Afro-American woman will marry an Afro-American man are 32,998 times greater than that a non-Afro-American woman will marry an Afro-American man! Compare this ratio with the odds ratios of other ethnic groups.

Further evidence on Afro-American isolation comes from sociological studies of Americans “core” networks—those people with whom they discuss “important matters.” The 1985 General Social Survey—the first nationwide sample of core network data—found striking racial and ethnic homogeneity in these networks. (The enormous liability of Afro-American marital isolation becomes even clearer when we learn that the typical American core network is small and centered on kin.) Significant variations, however, exist across subgroups. Not surprisingly, whites have the largest networks, Afro-Americans the smallest. One troubling finding of this and other studies is that, contrary to the identity rhetoric of Afro-American leaders and the claims of many sociologists, Afro-Americans do not compensate for their isolation from other groups by relying more heavily on kinsmen. To the contrary, as Peter Marsden, a preeminent network scholar, notes, “black respondents cited fewer kin and fewer non-kin than whites did, and their networks have a lower proportion of kin than those of whites.”

Enter Affirmative Action

Affirmative action can best be seen as a medium-term strategy to supplement and expedite the long-term educational solution to the unfair isolation of Afro-Americans. It directly addresses and compensates for their network liability by inserting them into the network-rich educational institutions of the nation and the self-generating career networks at the workplace—networks that Euro-Americans take for granted.

Affirmative action in the educational sector builds networks in several ways. First, it helps youngsters gain skills. Defenders of affirmative action usually emphasize this role, but such a strategy cannot justify why students from disadvantaged backgrounds should be at the nation’s elite colleges if adequate educational alternatives exist for them that are congruent with their educational records. Indeed, as a bigger fish in a smaller pond, a talented Afro-American student may well acquire more purely technical skills at a good state institution than at Harvard.

But while educators are loath to admit it, elite educational institutions serve two primary functions: to recruit and educate the nation’s elite and to provide an environment in which students initiate and extend elite networks and learn network negotiating skills. Living and learning with students from elite environments helps establish wide-ranging links with elite networks, as well as dense links through intermarriages.

Given what we now know about the true nature of labor markets, it should be clear that only some form of affirmative action could have inserted Afro-Americans into the more working- and middle-class jobs on a significant scale. Economists often express bewilderment at the failure of Afro-Americans’ qualifications to match up with their earnings. That a Euro-American construction worker with a high school degree earns more than the typical Afro-American college graduate must be a source of endless disciplinary frustration for those who believe that the price system should match up workers of equal qualifications with jobs paying equal wages. As readers of this essay will know, however, there is an answer to this puzzle: networks.

Mark Granovetter’s description of the utility of networks in Getting a Job provides an excellent rationale for affirmative action intervention in the labor market: “Especially important to recognize, moreover, is the self-maintaining aspect of personal contact systems. Blacks are at a disadvantage in using channels of job information not because they have failed to ?develop an informal structure’ suitable to the need, but because they are presently under-represented in the structure of employment itself. If those presently employed in a given industry or firm have no black friends, no blacks will enter those settings through personal contacts. Once a core of blacks (or whatever group in question) has become established, however, a multiplier effect can be anticipated, as they recruit friends and relatives, who do the same, and so on. Once achieved, this situation is self-sustaining.” Since Granovetter wrote this in 1974, a great deal has been achieved, thanks in good part to precisely this effect of affirmative action. And, in fact, Granovetter offers not only a rationale for affirmative action but a way of judging where and when it has fulfilled its purpose. Affirmative action can and perhaps should be discontinued in any firm that has achieved this self-sustaining recruitment process for minority workers. The self-sustaining group, I would add, need not coincide with the proportion of the disadvantaged group in the population at large.

Another important area of economic justice that can be remedied only by affirmative action pertains to the role of networks in both the internal and external structure of organizations—particularly in middle- and upper-level employment decisions. Organizations, including business firms, rely heavily on informal networks both for the internal structuring of information and decisionmaking and for links with related organizations. At all levels there is not only a tacit culture at play—knowledge of how things really work as distinct from what the job descriptions and work manuals say—but vital network contacts to be made and important network negotiating skills to be learned if one is to effectively mobilize critical network figures and links within and between organizations. Minorities and women are often unconnected to such knowledge and links and may never stand a chance of acquiring them—because they are made, nurtured, and extended precisely in those places where disadvantaged minorities and women are not to be found.

This problem is commonly mentioned, so there is no need to belabor it except to note one important point. The higher up a hierarchy one goes, the more important network factors become in recruitment decisions. An unprejudiced personnel officer bound by a color- or gender-blind code, but alert to the organizational need for persons with the right network skills, will always be obliged to reject applicants from backgrounds with impoverished network resources. The costs are simply too high to train such people. Where then are these skills to be acquired? They can come only from organizations being pressured or provided with governmental incentives to train people in these skills until a critical mass of them exists.

Costs and Responsibilities

All social policies incur costs, and affirmative action is no exception. The costs to individual Euro-Americans have been so thoroughly discussed that there is no need to repeat them here. One way to minimize the costs is to recognize three basic principles emerging from court decisions as identified by Jim D. Newman in 1989. The first is that the right to retain one s job is sacrosanct and cannot be interfered with by affirmative action under any circumstances. The program is best used in recruiting new workers. Thus, one may recruit an equally qualified Afro-American teacher over a Euro-American one on affirmative action grounds but not fire a Euro-American teacher instead of an Afro-American one for this reason. Second, while affirmative action is still needed in promotion decisions, ethnic or gender considerations can come into play only where candidates are otherwise equally qualified. Finally, quotas of any kind are prohibited.

Another important consideration concerns the nature of jobs. It might be helpful, for example, to distinguish between knowledge-intensive and network-intensive jobs. A network-intensive job is one for which employers depend heavily on informal contacts when filling vacancies; it also requires a great deal of network negotiating skill for its performance. A knowledge-intensive job is one where purely technical or cognitive qualifications dominate the recruitment and performance processes. Now, barring sinecures in city halls and family firms around the country, nearly all jobs that are network-intensive also require a minimum level of skills. But many jobs are almost wholly knowledge-intensive. I propose that affirmative action programs exclude such jobs and focus on network-intensive jobs—precisely those that can never be attained through schooling alone. These jobs, incidentally, are to be found at all levels of the occupational hierarchy.

Costs and responsibilities must also be borne by beneficiaries of affirmative action. One danger is that affirmative action may reduce the incentive for beneficiaries to perform at their very best. This troublesome possibility is suggested, for example, by the tendency for Afro-American students to perform below the level predicted by their test scores. I do not know whether the same problem exists at the workplace. One way of addressing or preventing an underachievement problem is for teachers and supervisors to evaluate with scrupulous honesty. Colleges should return to blind grading, and rules should be enacted that forbid all students from revealing their ethnic identities in their examination answers and papers.

The greatest risk of affirmative action is that it might become an institutionalized entitlement program. I strongly suspect that many otherwise sympathetic persons oppose it for precisely this reason. It is not an unreasonable apprehension, given the chronic propensity of government programs to become permanent. I will return to this in my final suggestion.

A related, and equally disturbing, danger is that some Afro-American leaders may conspire with segregationist Euro-Americans to stand affirmative action on its head. As this article has made clear, the main justification for affirmative action is that it compensates for the liability of isolation of Afro-Americans. It is both a medium-term substitution for, and a jump-start promotion of, the long-term solution of full ethnic integration. As integration is achieved, the need for affirmative action should be lessened. In this way, the program becomes self-canceling.

Two developments, however, hint at a possible subversion of this integrative process. One is the Afro-American identity movement, with its explicit rejection of integration and its celebration of ethnic separatism, as in the tendency of minority students to segregate themselves on the nation’s campuses and the increasing propensity for voluntarily segregated neighborhoods among the Afro-American middle class.

The other development is best exemplified by the Atlanta model of ethnic relations. This city, which prides itself on being “too busy to hate,” has fully embraced affirmative action, creating an environment with which the elites of both groups seem very comfortable. The only problem is that Atlanta remains extremely segregated. It seems as if a quid pro quo has emerged between the two dominant ethnic groups in which permanent affirmative action is accepted by the Euro-American elite as the price to be paid for the permanent segregation of the two groups—an adaptation of the old southern racist principle: separate, but now truly equal, at least among the elites. Were this indeed the case—and it is merely one interpretation of what is happening in Atlanta—it would be a disastrous contradiction not only of the ends of affirmative action but of its main justification, at least as I have argued it.

One way of preventing or aborting such an outcome is to place some time limit on affirmative action. The program has already achieved a great deal although Afro-Americans remain badly unconnected and in need of help in overcoming their isolation. I have proposed elsewhere that affirmative action be gradually phased out over 15 years, at which time it would be converted to a class-based program. During this period every effort should be made to break down all remaining barriers to integration, including Afro-American resistance to integrated neighborhoods, intermarriage, and trans-ethnic adoptions.

We could begin immediately to reduce the social cost of the program by removing all categories of disadvantaged persons except Afro-Americans, Puerto Ricans, second-generation Mexican Americans, Native Americans, and Euro-American women. In another five years or so we should remove all upper-middle-class persons, say those from families with incomes greater than $75,000. Ten years from now only working- and lower-class disadvantaged persons should be eligible, and Euro-American underprivileged persons could be phased in. In 15 years all ethnic criteria should be dropped; the resulting class-based program should last as long as poverty and underprivileged classes exist. As we ease out of the existing program and shift toward one based on socioeconomic disabilities, it becomes imperative that we reinforce and vigorously enforce laws against discrimination.

Placing a time limit on affirmative action would in all likelihood blunt the orchestrated politics of controversy that now bedevils it (see box below). And thinking about phasing it into a class-based entitlement program may at long last bring Americans around to a consideration of the growing inequality that threatens the harmony of our democracy far more than the alarmist cry of “racial division.”

Get daily updates from Brookings