For the third time in the last 10 years, on August 24, 2018, the Office of Federal Contract Compliance Programs (OFCCP) revamped its guidance on compensation investigation and enforcement by issuing Directive 2018-05 and publishing frequently asked questions (FAQs) on the new directive. This internal agency directive, while not having the force of law that a court holding or regulation would have, does provide federal contractors with some information on how OFCCP may evaluate their compensation in OFCCP audits. This directive will apply to compliance evaluations scheduled on or after August 24, 2018, and may be used by OFCCP in earlier, open cases to the extent it doesn’t conflict with prior guidance or procedure.

On Which Pay Cases Will OFCCP Focus?

In the directive, OFCCP offers general guidance on the legal theories under which it will pursue compensation discrimination and how those align with program priorities. Stating that compensation disparities can be created through inequities in monetary compensation, inequitable training or advancement opportunities, and assignment/placement differences, OFCCP reiterated its reliance on statistical analyses reflecting a statistically significant difference of two standard deviations or more as well as non-statistical (or anecdotal) evidence of discrimination. (Anecdotal evidence could be testimony regarding biased statements or remarks, examples of differential treatment, testimony about individuals who were given misleading or contradictory information about employment or compensation practices, etc.)

Signaling a potential shift in agency practice, however, the directive notes that “OFCCP will be less likely to pursue a matter where the statistical data are not corroborated by non-statistical evidence of discrimination unless the statistical evidence is exceptionally strong.” (OFCCP does not define the phrase “exceptionally strong.”) Moreover, a footnote explains that OFCCP may pursue a case without anecdotal evidence if the agency detects patterns of discrimination over several years or at multiple contractor establishments. Finally, OFCCP may conduct comparative analyses of small groups of similarly-situated employees to determine if pay differences due to discrimination exist.

What Pay Data Will OFCCP Review in Compliance Evaluations?

• “Item 19” data: Contractors must provide individual-level pay data to OFCCP in response to the itemized listing that accompanies a compliance evaluation scheduling letter (item 19 of that listing requests pay data). The directive further instructs agency compliance officers to ensure that the pay data provided is complete and acceptable; if not, the compliance officer must quickly request that the contractor provide it, in electronic format if it is maintained in that fashion. Failure to provide such information will result in the issuance of a Notice to Show Cause, which could lead to litigation for failure to provide access to pay data. OFCCP gives high priority to so-called “failure of access” cases and will initiate litigation quickly in the appropriate situation.
• Compensation hierarchy/job structure: The directive encourages contractors to provide their compensation hierarchies and job structures with the Item 19 data so that OFCCP can attempt to design a pay analysis based on those structures. The directive is silent, however, on how OFCCP defines (or the conditions under which it accepts) a contractor’s compensation hierarchy and job structure. Most important, OFCCP does not commit in the directive to design a pay analysis based on the contractor’s structure!
• EEO-1 categories/AAP job groups: If OFCCP finds that the contractor’s compensation hierarchy or job structure is unreasonable, cannot be documented through written policies, and/or if the pay groupings do not lend themselves to a meaningful systemic statistical analysis, the agency will conduct a preliminary pay analysis based on EEO-1 categories or affirmative action program (AAP) job groups. Unfortunately, these types of broad groupings may include employees who do not do similar work as measured by Title VII of the Civil Rights Act.
• Other aggregations: If the contractor’s AAP job groups are unreasonable; do not represent jobs with similar content, wage rates, and opportunities; and/or are not of a sufficient size to conduct a meaningful systemic statistical analysis, OFCCP will create an analysis model based on information developed during its investigation.
• Additional pay data requests: OFCCP will notify contractors, in writing, of the “general nature” of any preliminary compensation disparities that trigger further information requests or an on-site review. This is a nod to OFCCP’s recent promise to be more transparent with contractors as compliance reviews unfold.

How Will OFCCP Group Contractors’ Pay Data?

Identify Similarly-Situated Employees

According to the directive, “similarly-situated employees” are “those who would be expected to be paid the same” (not the Title VII standard) based on:

• job similarity (such as tasks performed, skills required, effort, responsibility, working conditions, and complexity of duties) and
• other objective factors (such as minimum qualifications and certifications).

Develop Pay Analysis Groupings

As it did under previous guidance, OFCCP will develop Pay Analysis Groupings (PAGs) of “comparable” (not “similarly-situated”) employees. According to FAQ 7, a PAG could aggregate employees from one or multiple job titles, units, categories. Generally, according to FAQ 14, OFCCP prefers each PAG to contain at least 30 employees under a similar pay system who perform broadly similar job functions.

Statistically Control for Differences

Depending on information provided to OFCCP by the contractor, OFCCP will statistically control for additional differences among PAGs, such as division, business unit, product line, location, company tenure, prior experience, education, full-time status, grade level or range, etc.   Disappointingly, this process enables OFCCP to continue its Obama-era practice of aggregating potentially dissimilarly-situated employees in a statistical analysis.  OFCCP will try to ensure that there are at least 10 observations (or employees) per control variable. For example, if a PAG has five control variables (e.g., sex, years in job, other years with company, years of prior experience, and required certification), under FAQ 14 and FAQ 15, that PAG would ideally have at least 50 employees.

How Will OFCCP Analyze Contractors’ Pay Data?

For the first time, OFCCP provides a possible road map of OFCCP’s internal statistical analysis of pay data. Directive 2018-05 provides as follows:

• OFCCP will separately analyze base pay and total compensation.
• OFCCP may analyze other elements of compensation, such as bonuses, commissions, overtime payments, and shift differentials.
• OFCCP will use a predictive analysis called a “multiple linear regression analysis” to evaluate and explain pay information. Under this analysis, OFCCP
  • will transform salary to the log of salary in the regression model (so that salary differences are reported in percentages instead of dollars);
  • will initially evaluate the effect of sex or race in separate regression models;
  • will control for
    • sex;
    • race/ethnicity;
    • components of employee tenure separately;
    • education categories, if provided by the contractor;
    • prior experience, if provided by the contractor;
    • job level or grade, if provided by the contractor; and
    • performance ratings or rankings, where feasible;
  • may control for job title if pay legitimately varies by job title;
  • will not automatically include squared tenure terms in preliminary analyses so as to account for the possibility that the relationship between salary and other controls is not a straight line;
  • may use age as a proxy for prior experience in preliminary analysis;
  • will evaluate market salary surveys to determine if they would be useful to the analysis;
  • will test all variables for neutrality—in other words, OFCCP will test the control variables to confirm that they are predictive of pay and neutrally applied with respect to sex, race, and/or ethnicity; and
  • will omit any variables that are tainted by the contractor’s discriminatory practices or lack of practices or procedures (but does not explain how OFCCP will determine if a variable is “tainted”).
• OFCCP will analyze statistical outliers for indicators of potentially inappropriate PAGs.
• According to FAQ 9, OFCCP may conduct comparative analyses of small similarly-situated groups to determine if pay differences are due to discrimination.
• OFCCP will provide to contractors individual-employee-level data necessary to replicate the PAGs and regression results.

If a contractor cannot rebut OFCCP’s statistical and/or anecdotal findings of pay discrimination, OFCCP may pursue various remedies through conciliation, including back pay and benefits, interest, salary adjustments, and monetary relief, such as modifying policies and procedures or providing training opportunities, work assignments, promotions, and/or job placements. Should conciliation fail, and, according to FAQ 22, where OFCCP “finds strong evidence of discrimination,” OFCCP can pursue legal action and may seek sanctions to include cancellation of current federal contracts and debarment from entering future contracts.

Key Takeaways

While Directive 2018-05 provides high-level direction on the statistical methods and controls OFCCP may use, it also provides broad flexibility to the agency in analyzing pay data. The language of the directive could signal minor shifts or major shifts in practice, depending on the agency’s practical implementation. In the meantime, contractors should consider the following proactive steps to evaluating and defending potential compensation issues:

• Establish clear and defensible compensation policies and pay processes. When deviation from these policies and/or processes is necessary, contractors may want to document the decision/deviation. Contractors may also want to periodically review compensation policies and processes—at least annually—to ensure that they are useful and that minimal deviation occurs.
• OFCCP regulations require contractors to perform in-depth analyses of their total employment processes, including an analysis of their compensation systems to determine whether there are gender-, race-, or ethnicity-based disparities. FAQ 5 further recommends that “self-audits” of compensation should occur periodically and that results should be reported to management, neither of which is required under OFCCP regulations. Nonetheless, written, specific, and current information on  job content, skill levels, essential functions, performance requirements, reporting structure, etc., can assist with establishing—and defending—pay. Although not required, job descriptions that outline the position’s required or preferred knowledge, skills, and abilities are useful in establishing and justifying the starting salary of new and current employees.
• In assessing pay equity, relevant factors in determining similarity may include tasks performed, employees’ skills, effort, levels of responsibility, working conditions, job difficulty, minimum qualifications, and other objective factors.
• Evaluating base pay is a central component—but not the only component—of any pay analysis. Bonuses, stock programs, benefits, reward programs, and overtime pay are all forms of compensation that can be evaluated, as appropriate. In addition, performance-related factors can be reviewed to ensure that opportunities (such as sales territories, production incentives, promotion potential, etc.) are equitably distributed and compensated.
• Compensation decisions are challenging to defend if accurate records do not exist that document why decisions were made. Clear documentation of compensation policies and decision making is helpful to evaluating pay equity. Each element used to establish compensation for a given employee can be recorded in the system, to the extent possible. Common examples of data points that could be useful in explaining pay include time in position, education level, years of relevant experience prior to joining the contractor’s workforce, certifications, knowledge and/or skills, and applicant salary requirements/negotiations.
• Performance ratings often have a substantial impact on compensation, in both the short term (bonuses and pay increases) and long term (total pay, total benefits, and promotion potential). In this regard, it may be helpful to conduct a privileged analysis of performance review scores and anticipated merit increases prior to implementation.