How Affirmative Action Falls Short

Tuesday, October 6, 2020
Image credit: 

It has been fifty-six years since the passage of the Civil Rights Act of 1964, legislation that took aim at the systematic forms of legal segregation that had long dominated large segments of American life. It did not take an expert in implicit biases to see the corrupting influence that officially sanctioned racial segregation had on public life, nor did it take a subtle analysis to understand the importance of the Voting Rights Act of 1965 in undoing the exclusion of African-American citizens from their lawful place in society. The effects of these statutory reforms were lasting and profound.

The passage of these landmark statutes did not put an end to racial conflict simply because they ended explicit forms of discrimination. Indeed, one of the toughest issues to resolve was the proper regime for dealing with labor markets. The great mistake of the 1964 Civil Rights Act was to adopt an explicit colorblind standard for employment under Title VII, which had the effect of slowing down the introduction of affirmative action programs that might have led to more African-American employees in the workplace, especially in unionized firms.

Those affirmative action programs received belated judicial approval in United Steelworkers v. Weber (1979), in which Justice William Brennan held that Title VII “does not prohibit such race-conscious affirmative action plans.” In Weber, Justice Brennan upheld a program that set aside 50 percent of the in-plant craft-training places for black workers until they achieved parity to the percentage of black workers in the overall labor force within that community. That decision was the second major piece of Title VII’s employment law regime, following the 1971 decision in Griggs v. Duke Power, which had previously adopted a strict “business necessity” test to justify a disparate impact that any facially neutral test or business practice had on racial minorities. Weber enabled affirmative action programs, while Griggs blocked discrimination against protected minority groups. 

The effects of these decisions were not small. Weber inaugurated what would become a massive movement towards affirmative action programs in the United States, a phenomenon that is radically inconsistent with the claim that structural racism exists. The anti-racist movements that have gained strength in the past several years insist that “structural racism . . . typically instigated or sanctioned by government . . . creates inequality in every aspect of life and puts black people on the lowest rung of the racial hierarchy ladder.”

Notwithstanding Weber and Griggs, such charges might be sustained if there were any public statements or official policies that defended racially discriminatory practices. Fortunately, there are no such declarations to be found today. Faced with that reality, those who claim structural racism is a problem necessarily point to implicit social practices and norms that are said to be racist.

At this point, the basic claim is that any form of disparity in outcomes by race—such as by wages or professional representation—is conclusive evidence of some form of systematic racism. That expansive claim eliminates one of the key elements of a disparate impact claim under Griggs. The baseline for the determination of unequal access under Griggs is not the entire population of the nation, or even that of any state or local area. It is, or at least was, the class of potential workers or students who were in the relevant occupational or educational niche. 

The readiness of individuals for advanced schooling or certain classes of jobs sets the denominator for the overall inquiry. In addition, it is critical to note that the federal courts have been very tough in their examination of general ability or aptitude for entry level positions, as in Ricci v. DeStefano (2009), where by a five-to-four vote the Supreme Court upheld the use of a carefully prepared and vetted entry-level test for police officers, notwithstanding its disparate impact on outcomes. But even if Ricci had come out the other way on entry-level positions, it would have had little or no impact on the distribution of jobs in advanced positions, where the common form of job interview asks applicants to analyze an assignment drawn from the company files. In these cases, business necessity is a given, so that any disparate impact in outcomes is beyond the ordinary field of view. But even while these firms are protected against disparate impact suits, they still engage in extensive outreach programs to include more minority members in their ranks, further undermining claims of structural racism.

Yet there are limits to the extent that these outreach programs are able to reverse the evident level of disparate impact in operation. As economists Edward Lazear and the late Sherwin Rosen have suggested, one way to think of how promotions work in firms is to treat the process as though it were a tournament, in which the first bracket includes all entry level candidates. Under this framework, an explicit affirmative action program at level one guarantees that at level two, which lacks the same affirmative action requirements, not all groups will gain promotions in the same proportion that brought them to the second round.

There are of course many different ways to promote and many different measures of performance that can be used at various stages of the process. But for the general point, the basic tendency remains the same if one assumes that all workers have a constant ability that does not alter over time, without trying to plot variations in individual performance levels—a task better left to the employer.

Under this assumption, if that second round is selected wholly without regard to race (or any other single characteristic), the disparities in representation between the two groups will necessarily increase. Suppose that after an initial round of promotions there are two groups, A and B, each comprising fifty employees, with group A candidates being rated from 21 to 45 and group B candidates being rated from 16 to 40, the ratings being evenly distributed over the respective ranges for both groups. Next, suppose there are only fifty higher level positions available to this combined pool, with promotions being made strictly on the basis of formal qualifications. Recall “business necessity.” All of the group-A candidates with performance scores over 40 will make it through, which yields ten candidates. The next forty candidates selected will comprise twenty employees from group A and twenty from group B, all with scores ranging from 31 to 40, resulting in a final pool of thirty group-A employees compared to only twenty group-B employees.

Repetitive application of these same rules only increases the disparity, eventually converging to a pure ability-based hiring distribution.

The question then arises of what countermeasures can be used to deal with this reality. Still using the above numbers, one possibility that is constantly entertained is to lower the formal cutoff below the 31-point measure, and then base selection on a “holistic” evaluation of the candidates in order to increase the number of group-B employees in the higher tiers. But there are powerful limitations to this strategy, especially in technical and scientific areas, as this approach creates a situation where individuals with supervisory rules have weaker credentials than the next cohort that they supervise. The further along one moves in a hierarchy, the more costly the continuation of the holistic model of selection.

At this point, the only remaining strategy is to get stronger candidates from group B into the initial pool, often achieved by engaging in costly searches and creating better benefit packages to locate and secure the additional talent. But if the identical strategy is adopted by multiple firms, the wage and benefit packages that are offered to group B employees will disproportionally increase relative to group A, which could easily create an uneasy firm dynamic where weaker employees receive greater total compensation than stronger ones.

Systemwide, the problem becomes a social one, centered on increasing the input of talent from disparately impacted groups to avoid the problems of initial hiring and promotion. Thus, the foundational issue focuses on educational policy. The Nobel laureate economist James Heckman has repeatedly stressed the importance for disadvantaged children of developing the skills for success later in life. Unfortunately, state monopolies on public education hamper student performance, as in Detroit. This points to a critical role for charter schools to play in overcoming the many shortcomings of public education. Sadly, working along these lines requires a focus on the long game and a willingness to limit political influences on educational practices.

The shortsighted impulses of the day may produce short-term political gains, but only a committed effort to supply all children with the educational opportunities they deserve will ultimately correct the disparate outcomes seen today.