On Friday August 24th, the Office of Federal Contract Compliance Programs (OFCCP) issued three new directives, the most important being the replacement of Directive 307 which covered how compensation reviews during audits should be conducted (Directive 2018-05). OFCCP first laid out guidance in 2006 as to how the agency would approach compensation reviews. It was highly systematic and followed equal opportunity law. In 2013, the agency issued Directive 307, which laid out a more comprehensive, yet unclear and opaque approach to compensation reviews. 
Under Directive 307, coupled with changes to the scheduling letter, contractors were required to provide individual employee data. However, the OFCCP took the view that they could group the data anyway they saw fit, pushing the boundaries of case law, to determine compensation inequities and force settlements. OFCCP used Pay Analysis Groups (PAGs) as opposed to case law which referenced “similarly situated employees groups” or SSEGs. Using pooled regression (combining various groups which arguably are connected yet have different factors), OFCCP would many times have false positives, and in a number of cases, spend an inordinate amount of time proving the false positive to the detriment of the contractor. It was costly on both sides.
Therefore, a number of the “compensation” cases were reframed hiring cases, called steering cases. In these situations managers who knew a position, such as material handler, which had a high turnover of women, would “steer” women to jobs known to have less turnover. Yet the outcomes of settlements at times were a little confusing. In one case, G&K, men and women were both steered to lifting jobs – men to heavier lifting, women to lighter lifting jobs – and both were identified as discrimination victims, although heavier lifting jobs were higher paying.
Because of a change of administrations, OFCCP was concerned that contractors lacked the clear guidance Directive 307 intended to give. OFCCP believes that fulsome guidance will further support contractors' ability to conduct meaningful self-audits so that they can proactively identify and address issues with their compensation practices. The new directive still requires OFCCP to analyze pay using Pay Analysis Groups. In addition, the focus on base pay, total compensation; and, if necessary, components of compensation such as bonus, commission, and shift differentials remains.
However, if the compliance officer, or other responsible OFCCP staff member, determines that the submission does not satisfy the requirements of the Itemized Listing data request, the compliance officer must contact the contractor to discuss the discrepancies. Moreover, the compliance officer or other responsible person must send a written request to the contractor seeking production of the compensation data as described in the Itemized Listing within seven business days of receipt of the letter. The seven days may not be enough time to produce the missing data, but OFCCP will likely discuss timelines with contractors. Any contractor that fails to produce the data will be subject to a Show Cause Notice.
The Compensation Directive outlines its approach to statistical modeling and the use of control variables. For instance, the OFCCP clarifies that it uses a favored-group analysis for analyzing compensation by race/ethnicity. It also advises that it may rely on age as a proxy for experience only during the desk audit but will seek to obtain actual prior experience. It will also evaluate market studies on a case-by-case basis if the information is provided and will not automatically include “squared-tenure” terms in the preliminary desk audit analysis. Specifically, OFCCP will generally conduct statistical analysis using among other things:
• Use of multiple linear regression analysis to evaluate information during or subsequent
to the desk audit to minimize false positive and false negative results
• Separate analysis of base pay and total compensation, and, if necessary, components of
compensation (e.g., bonus, commission, overtime, shift differentials)
• The log of salary in the regression model to account for potentially different pay distributions within PAGs
• Analysis of statistical outliers for indicators of potentially inappropriate pay analysis
groupings
OFCCP will focus more on what the contractor has as its salary structure as opposed to redefining the structure to prove discrimination. For example, the directive states that “[i]f a contractor provides its compensation hierarchy and job structure in the submission to the Itemized Listing, OFCCP will attempt to design its analysis based on that structure.”
Also, OFCCP will be driving transparency in the audit. The OFCCP will notify the contractor of the general nature of the disparity under review by explaining that the agency is reviewing, for example: “Compensation policies and practices with respect to women in production, sales, and management.”
Moreover, if OFCCP makes a preliminary finding of discrimination, the agency will disclose “the individual-level data necessary for the contractor to replicate the PAGs and regression results” in electronic format.
Finally, OFCCP will also include representatives from its Branch of Expert Services (e.g., labor economists or statisticians) in the conciliation process, if necessary, to provide clarity around the Agency’s statistical methodology and findings.
The Compensation Directive applies to any reviews scheduled after August 24th and to any open reviews to the extent that the analysis does not conflict with prior guidance.
Source: OFCCP 8/24/18