Change of Course? OFCCP Issues Long-Awaited Revised Compensation Guidelines

In a highly anticipated move, the Department of Labor’s Office of Federal Contract Compliance Programs (OFCCP) issued its new compensation directive on August 24, 2018. Directive (DIR) 2018-05, Analysis of Contractor Compensation Practices During a Compliance Evaluation, replaces the Obama-era compensation guidance DIR 2013-03, Procedures for Reviewing Contractor Compensation Systems and Practices (referred to as Directive 307). OFCCP also included a list of 22 Frequently Asked Questions (FAQs) with DIR 2018-05.

Unlike the Bush-era 2006 Compensation Standards and Voluntary Guidelines (Voluntary Guidelines), DIR 2018-05 was not approved through notice and comment rulemaking and, therefore, does not qualify as an official OFCCP regulation. Contractors should nonetheless consider this new guidance, along with OFCCP’s Federal Contract Compliance Manual, the standard operating procedures for compensation-related compliance evaluations, as compliance officers (COs) can be expected to reference these two materials going forward. DIR 2018-05 states that it applies to reviews started after August 24, 2018 as well as “open reviews to the extent that [the procedures] do not conflict with OFCCP guidance or procedures existing prior to the effective date.” As such, contractors in the midst of open reviews or administrative proceedings should have DIR 2018-05 in mind.

Overall, DIR 2018-05 is a mixed bag. It suggests some helpful moves towards transparency for contractors. But the revised provisions do not solve some of the more contentious issues between contractors and OFCCP, as detailed below. And the directive continues to suggest that the Agency will insist that every AAP location at every contractor that it audits can be properly evaluated using an aggregated statistical model, regardless of the differentiation among the contractor’s workforce. Indeed, DIR 2018-05’s strong emphasis on statistical analysis while almost ignoring Title VII law could be seen as a troubling sign for contractors expecting a complete rejection of Directive 307. Thus, while contractors should be mindful of the views and approaches OFCCP has articulated in the new guidance, they should remain insistent that compliance audits are conducted consistent with Title VII, which the Agency continues to acknowledge furnishes the substantive law that OFCCP enforces.

What are the main features of the new directive?

Directive’s Goals and Disclosure of Information

DIR 2018-05 begins with background section listing the previous compensation guidance issued by the Agency. OFCCP then expresses concern that Directive 307 failed to provide contractors with clear guidance, and states that the new guidance is intended to better enable contractors to conduct meaningful proactive audits. DIR 2018-05 further touts the Agency’s goals of added transparency, compliance through proactive self-analysis, and efficient and consistent compliance audits.

DIR 2018-05 emphasizes that the Freedom of Information Act (FOIA) allows disclosure of contractor data submitted to the Agency, but underscores that contractors can object to the release of any confidential data. In addition, contractors may submit data without identifying information if confidentially concerns exist. While contractors usually provide masked data to OFCCP, the reference to FOIA is a useful reminder to contractors of OFCCP’s tendency to disclose information in a number of ways, including by posting conciliation agreements on its website.  This is contrary to the Equal Employment Opportunity Commission (EEOC), which generally keeps employers’ information and details regarding its investigation more private.

Legal Background

The attachment to DIR 2018-05 purports to include a recitation of the law regarding disparate treatment and disparate impact, and reaffirms the Agency’s intent to analyze potential compensation discrimination in accordance with Title VII. Unfortunately, it cites only Int’l Bhd. of Teamsters v. United States, 431 U.S. 324 (1977) and Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971). Neither case discusses pay discrimination and OFCCP neglects to cite any Title VII compensation case law. Later in the directive, OFCCP cites to the EEOC Compliance Manual. Accordingly, contractors should refer to that guidance to see how OFCCP interprets Title VII.

The Agency notes the ways in which unequal compensation can result, including disparities in base pay, bonus, commissions, differences in training opportunities, and unequal assignments (“channeling”). Contractors should query the extent to which any particular circumstance encompasses an allegation of true compensation discrimination—i.e., paying one worker less than another performing substantially similar work due to the disfavored worker’s gender, race, national origin, etc.—as opposed to some other act allegedly contributing to unequal wages within the organization.

Desk Audit Requirements

DIR 2018-05 lays out the desk audit procedures that the Agency can be expected to follow on receipt of a contractor’s affirmative action plan. It states that a contractor can submit its data in a manner that is consistent with its compensation structure, and requires the CO to refrain from beginning the desk audit analysis until the contractor’s submission is “complete and acceptable.” Where the CO believes that the data submitted by the contractor is incomplete, DIR 2018-05 provides that the CO must request additional or revised data within 7 business days of the contractor’s submission.

Employee Groupings

DIR 2018-05 addresses how the Agency identifies employee groupings for its pay analysis. The new directive notes that job similarity as well as other objective factors may dictate whether the Agency will group employees into the same Pay Analysis Group (PAG). Contractors might thus have felt heartened that the directive signals a renewed commitment by the agency to look at actual job content before uncritically determining that given groups of employees are proper comparators (and, in turn, that any pay differences between them are potential indicia of discrimination). However, DIR 2018-05 goes on to suggest that OFCCP can create groups of “comparable” employees by using statistical controls, thereby permitting the Agency to include employees from different structural units in the same aggregated statistical analysis. The Agency states that its goal is to “mirror” a contractor’s compensation system, but that where it is unable to do so, it will turn to AAP or EEO-1 groupings even if those broad categories unquestionably include employees who are not doing similar work within the meaning of Title VII (e.g., all “Professionals”).

Statistical Analysis

DIR 2018-05 purports to provide more detailed information and transparency regarding the statistical analyses that the Agency will use to assess whether contractors are engaged in compensation discrimination. These additional details include:

  • the Agency’s use of multiple liner regression analysis;
  • its intent to analyze distinct pay elements (base, bonus, etc.) separately;
  • its recognition of the need to appropriately evaluate outliers;
  • its reliance on age as a potential proxy for relevant prior experience;
  • its willingness to evaluate whether market surveys should be an element of the analysis; and
  • its approval of controlling for performance rankings.

DIR 2018-05 further declares, however, that the Agency will generally “[t]est all variables for neutrality” and “omit any variables that it determines from its evaluation are tainted by discrimination.” It does not clarify how or if the Agency will determine (or in what cases it will simply assume) that a variable is “tainted.”

Investigative Findings and Conciliation Procedures

OFCCP echoes Directive 2018-01, underscoring that the Agency will now issue a Predetermination Notice (PDNs) for preliminary discrimination findings prior to issuing any Notice of Violation (NOV). The Agency also indicates that the PDN will include individual level data and other information necessary for contractors to replicate the PAGs that the Agency used and the regression results that it obtained. This additional detail should be a welcome change for contractors who may previously have found themselves tasked with responding to vague suggestions that the Agency has found “indicators” of discrimination without the detail needed to fully understand and address the allegations.  DIR 2018-05 also indicates that the Agency’s statisticians will participate in conciliation proceedings and be available to clarify Agency findings and answer contractor’s questions.

How does the new guidance compare to Directive 307?

As an initial matter, the new directive is unclear as to the status of the prior Directive 307. On the one hand, DIR 2018-05 states that it was issued to “update the guidance set out in Directive 307.”  But then, OFCCP states that it is “rescinding” Directive 307. Notwithstanding the “update” language, our interpretation is that contractors should consider Directive 307 dead and take their guidance from DIR 2018-05.

Overall, DIR 2018-05 provides less detail regarding each step the Agency will take in conducting its compensation analyses, from receipt of the initial compensation data to the issuance of findings. But the directive and FAQs do provide additional detail and transparency regarding OFCCP’s statistical analyses and approach.

On the positive side for contractors, DIR 2018-05 recognizes that the Agency will place less of a priority on cases where its statistical findings are not accompanied by non-statistical (i.e., anecdotal) support. Further, by providing contractors the ability to submit data regarding its own groupings and suggesting that COs consider those groupings, DIR 2018-05 may encourage COs to more readily consider and track the contractor’s compensation structure in their evaluations. Directive 307, by contrast, had underscored OFCCP’s autonomy to create its own PAGs. In addition, DIR 2018-05’s outline of the Agency’s statistical approach and the types of control variables it will generally employ provides greater insight for contractors than did the Agency’s “black box” approach under Directive 307.

On the negative side, DIR 2018-05 dials back the CO’s obligation in Directive 307 to investigate whether a difference in compensation exists, whether the difference is among comparable employees in the context of the contractor’s salary system and whether legitimate reasons explain salary differences. DIR 2018-05 does not overtly contain a similar CO requirement. Second, large contractors will likely be disheartened to hear that “exceptionally strong” statistical cases remain a priority for the Agency. This is because OFCCP’s broad employee groupings and its statistical methods often lead to statistically significant results—and, in turn, allegations of compensation discrimination—driven by the effect of large sample sizes.[1] DIR 2018-05 also retains the PAG rubric from Directive 307, rather than the “similarly situated” employee grouping standard from the 2006 Voluntary Guidelines. The PAG structure under Directive 307 often did not conform to Title VII’s “similar situated” standard and appeared to embolden COs to create broad employee groupings that resulted in comparison of dissimilar individuals and, as a result, “false positive” findings of pay discrimination. DIR 2018-05 also does not make clear what information regarding the contractor’s compensation structure or job classification system would be acceptable to the Agency, such that it would actually lead COs to tailor their standard approaches to fit particular companies. Finally, DIR 2018-05 states that OFCCP can, in some circumstances, use EEO-1 and AAP groupings to structure its analyses. But these groupings, on their face, fail to conform to Title VII requirements which requires that comparator employees be similarly situated and thus will almost certainly lead to unreliable findings of compensation discrimination, with the Agency assuming discriminatory pay where, in fact, the pay differences are driven by dissimilar work.

How does the new directive impact contractors’ obligation to self-monitor?

DIR 2018-05 does not expressly state that it is articulating any new obligations regarding contractors’ obligations to self-monitor. OFCCP attempts to push the envelope here, however. FAQ 5 states that OFCCP’s regulations “require that, at a minimum, all self-audits specifically assess whether race or gender-based compensation disparities exist, that self-audits occur periodically, and that results are reported internally to management.” But this FAQ arguably attempts to impermissibly expand upon the actually regulatory obligation contractors must meet. Rather, the regulations at 41 C.F.R. § 60-2.17 (b)(3) state that contractors “must evaluate … compensation system(s) to determine whether there are gender-, race-, or ethnicity-based disparities.” The only reference to any “audit” requirement appears in subsection (d) of that same regulation, and states that contractors must “develop and implement an auditing system that periodically measures the effectiveness of its total affirmative action program” including, inter alia, “[m]onitor[ing] records of all personnel activity, including … compensation, at all levels to ensure the nondiscriminatory policy is carried out.” “[E]valuations of compensation systems” and “auditing” one’s affirmative action program, however, are not necessarily synonymous with “self-audits,” at least in the sense the Agency appears to intend. Contractors should nonetheless be on the lookout for more aggressive Agency requests for contractors’ internal “self-audits,” and expectations on the part of COs that a contractor’s regulatory compliance program will have taken some particular, prescribed form.

How will the new directive change contractors’ interactions with OFCCP during an audit?

It is too early to tell whether DIR 2018-05 will change the dynamic between OFCCP and contractors under review. Should OFCCP’s regional offices follow DIR 2018-05 in the spirit in which it was written, it should allow for greater contractor input and transparency in the process.

DIR 2018-05 should first affect contractors at the desk audit phase. Contractors should be able to respond to the desk audit request with data that explains their compensation structure and systems before the Agency attempts to develop PAGs in a vacuum. The requirement that OFCCP request additional data within 7 days should also eliminate the significant delays contractors have experienced receiving feedback from OFCCP after their initial desk audit submissions. Also, the increased transparency regarding statistical approaches and the requirement that OFCCP provide data allowing the contractor to duplicate its analysis should allow contractors to cut through the methodological questions that often plagued previous reviews. On the other hand, the Agency may well now provide less detail regarding the overall compensation analysis process and insist on a “self-audit” requirement, such that contractors may find reviews less transparent and more frustrating in some regards. Further, the accompanying FAQs state that the Agency will bring enforcement action where “strong evidence of discrimination” exists. (FAQ 22)  While this may seem like the Agency will pass on borderline cases, the lack of consistency in regional offices coupled with the Agency’s continued heavy reliance on statistics should not put contractors at ease in predicting which cases will fall into the category of having “strong evidence of discrimination.”

It remains to be seen how DIR 2018-05 will be implemented in the field, at the national office, and (ultimately) by Administrative Law Judges. As such, contractors should be vigilant in their compensation practices and ensure that they have experienced OFCCP counsel involved in each step of their reviews.


[1] See, e.g., Ronald L. Wasserstein & Nicole A. Lazar (2016) The ASA’s Statement on p-Values: Context, Process, and Purpose, The American Statistician, 70:2, 129-133, DOI: 10.1080/00031305.2016.1154108 (criticizing overreliance on “statistically significant” findings because “[a]ny effect, no matter how tiny, can produce a small p-value if the sample size … is high enough”).