How We Act Affirmatively

All federal contractors and subcontractors subject to EO 11246 have requirements related to affirmative action (job advertisement taglines, employee / applicant / union notice, demographic data collection).   Those actions are primarily designed to provide notice that the contractor is committed to equal opportunity and to maximize the number of people whom that notice reaches (sounds good).  All construction contractors and larger non-construction contractors also have some outreach and results-awareness components to their affirmative action obligations - both a subjective determination of the effectiveness of a contractor’s process as a whole and each of the steps in the process (sure thing), an objective quantifiable determination of the available local labor compared to incumbent employees (you bet), and if not sufficiently balanced, taking actions to review the contractor’s processes (of course) and setting participation rate goals and placement rate goals, respectively, and comparing results to those goals (whoa, you sure about that boss?).  While the agency determines compliance by commitment to the process and not just by achieving participation or placement rate goals,1 there is certainly pressure on a contractor to achieve them.2  The difference between a goal and a quota can be a fine line, and can incentivize the types of preferences which undermine, not support, equal opportunity.

While data analysis is an invaluable tool in decision making, and being aware of one’s shortcomings is the first step to correcting them, setting a goal to achieve a number can’t help but put a thumb on the scale.  Recognition of a shortfall and a commitment to the affirmative action process together, guarantee a response.  Using subjective criteria for the good faith nature of that response still allows the agency to do its job (the agency already does this); using data analysis to show where improvement might still be needed is part of determining shortcomings, but using that data to set a quantified goal and giving any weight to that goal, is something else entirely.  It provides an incentive to focus more on the results than on ways to improve the process that produces the results, which increases the danger of a contractor giving, and the agency tacitly approving or even encouraging, racial / gender preferences.  Accordingly, the affirmative action regulations and agency guidance that discuss goal setting (and achieving) contain more disclaimers, warnings and limitations than a box of firecrackers, grain alcohol and unfiltered cigarettes given to a 12-year-old.  Specifically, the regulations point out that goals are not intended to achieve equal results, yet the use of them and the standards required for setting them imply precisely the opposite of that, and often have contractors working to achieve that implied meaning.3

The original use of “affirmative action” was in the predecessor EO (10925, 1961) to EO 11246, which was largely mirrored by EO 11246, including the phrase affirmative action, and including the same specific actions a contractor must take to ensure equal opportunity (see list above).  The inclusion of a results-awareness requirement by the new OFCC (authorized by EO 11246) appeared in their first regulations published in May 1968 - annual subjective review of: minority representation in all job categories, outreach efforts and hiring practices.  The Nixon administration began requiring the use of a quantifiable goal, first in 1969, the Philadelphia Plan (applicable to federally assisted construction contractors in Philadelphia, later applied in other cities as well) set participation rate goals which acted essentially as quotas, and then in the 1970 AAP regulations which required a comparison of available local labor to the contractor’s workforce and the setting of goals and timetables.  The goals and timetables requirement grew out of a time when the agency was putting its thumb on the scale to make incumbents and available labor look the same, and sooner than later; when the agency was putting contractors under its thumb to achieve this, almost at any cost, and using agency-imposed quotas which were openly at odds with equal opportunity.4  While that posture changed over time, the goals and timetables remained for 30 years.  In 2000, the timetable component was removed from the AAP regulations, but the goals were retained, now existing as annual goals.5  When the regulations abandoned timetables in 2000, there was no discussion by the agency of the propriety of continuing to use goals.

The decision to require goals and timetables (1970 regulations) was also made before the U.S. Supreme Court (in Griggs, 1971) recognized disparate impact as a method of proving discrimination.  The comparison of availability to incumbency is now partly included in disparate impact analysis, which compares applicants to incumbents.  While the agency allows using the 4/5ths rule or 2 standard deviations for determining if goal setting is required (rather than any difference at all between incumbents and available labor), it doesn’t shout that from the rooftops; and, if goal setting is required, the regulations require the goal to be at least equal to availability, not within 4/5ths or 2 standard deviations.6  They also require that goals not be a floor or a ceiling, and that achieving goals does not justify discriminating against another race / gender.  The safest way to try and reconcile those requirements is to try and make the workforce look as close to available labor as possible, which looks a lot, and can act a lot, like a quota.  The list of affirmative actions in the EO are designed to ensure that all potential applicants / employees feel equally welcome to seek hiring / advancement.  The contractor’s hiring decisions and, if hired, equal treatment, are the domains of equal opportunity.  If requiring goals at all, comparing availability to applicants would shed light on whether the local labor force feels equally welcome.  That should be affirmative action’s boundary, equal opportunity (disparate impact) can take it from there.

141 C.F.R. 60-2.35, 41 C.F.R. 60-4.2(d)(2) the content of this post omits (and does not apply equally to) the AAP requirements under Section 503 and VEVRAA because not being disabled or not being a veteran is not protected, while any race / gender is protected.

2“While serving in the Labor Department, I helped devise minority employment goals for government contractors. I now realize that the distinction we saw between goals and timetables on the one hand, and unconstitutional quotas on the other, was not valid. Our use of numerical standards in pursuit of equal opportunity has led ineluctably to the very quotas, guaranteeing equal results, that we wished to avoid.” The Road to Racial Quotas, Laurence Silberman, Wall St. J., Aug. 11, 1977; and more recently, “If the penalty for certain “bad results” is sufficiently severe, then people will tend to say that a rigid quota had been imposed. If penalties for bad results are minimal, then the people will tend to say that a flexible goal has been adopted. Clearly, this difference is one of degree, not of kind.”  Affirmative Action and Its Mythology, at page 3, Roland G. Fryer, Glenn C. Loury - Journal of Economic Perspectives, 2005.

341 C.F.R. 60-2.16(e)(3)

4The Philadelphia Plan was challenged for violating Title VII among other things, and upheld by the 3rd Circuit 442 F.2d 159 (3d Cir. 1971), cert. denied.  See also, United Steelworkers of America v. Weber, 443 U.S. 193, Dissent of Justice Rehnquist, at p. 226.

5Commentary in NPRM 65 Fed. Reg. 26087, at 26098

6The final rule commentary of the amended AAP regulations in November 2000 discusses the standards for determining if goals are required - the same standards as disparate impact analysis (65 Fed. Reg. 68021, at 68033), but those standards do not appear in the regulations.  41 C.F.R. 60-2.16(c) requires the goal to be at least equal to availability.