EEOC Invokes the Information Quality Act for a Third Study of Its Pay Data

The EEOC voted unanimously to defer any use of the EEO-1 Component 2 pay data that it has already collected until 2022 at the earliest. Instead, the EEOC is sponsoring another study of the Component 2 data by a panel appointed by the National Academies of Sciences, Engineering, and Medicine’s Committee on National Statistics (CNSTAT) to “examine the fitness for use of the data, including the utility of pay bands in measuring pay disparities and potential statistical and analytically appropriate uses of the data.”

These data reflect employee job category, pay, and hours for 2017 and 2018. The timeliness of this Component 2 collection was compromised by an August 2017 stay by the White House Office of Management and Budget (OMB). It was only in March 2019, in a suit brought by the National Women’s Law Center (NWLC) and the Labor Council for Latin American Advancement (LCLAA), that a federal court determined that the Trump Administration’s stay was illegal and ordered the EEOC to collect the data. Given the delay, the EEOC had the option of collecting more relevant pay data for 2018 and 2019 but opted to focus on collecting older data. The data collection was not complete until February 2020.

When originally implemented, Component 2 data was envisioned to accelerate and streamline the resolution of complaints, particularly by being able to limit requests for more detailed data from employers by using Component 2 data to establish the scope of investigations.

Component 2 data would support EEOC data analysis at the early stages of an investigation, using statistical tests to identify significant disparities in reported pay. EEOC enforcement staff who conduct these analyses would use them, in the larger context of other available economic data and information, to evaluate whether and how to investigate the allegations of discrimination in more depth.

For example, with Component 2, the EEOC would have been able to see whether a wage discrimination claim was evident in the job category of the plaintiff, and if so, whether there was evidence that the claim extended beyond a given job category. Absent these data and consistent with its mandate to enforce federal civil rights legislation, the EEOC will likely be obligated to continue to make more frequent and expansive data requests of employers charged with discrimination.

Another anticipated use of these data was to allow employers to compare the net impact on pay differences of their hiring, compensation, retention, and promotion decisions to similar firms by industry, geography, size, number of establishments, etc.

[E]mployers would be able to use the summary pay data that the EEOC intends to publish to generally assess their own pay practices.

Employers can now only look forward to being able to compare themselves in 2022 or later to such combined summaries reflecting the pay practices of their peers as they were four or more years ago.

The EEOC has tasked the CNSTAT panel with determining the “quality and utility” of the Component 2 collection. The EEOC’s Chief Data Officer questions the quality of the data despite the data collection being outsourced to an experienced, professional research institution (at the University of Chicago) who performed more than 48 individual validation tests of file format, file structure, and content on every submission.

The EEOC has opted for a panel study rather than the practical use as a means to evaluate the “utility” of the data, that it deemed “Necessary for the Enforcement of Title VII, the EPA, and Executive Order 11246.” This will be the third such study of pay data collection by the EEOC. In 2012, the EEOC commissioned a first study of the issue by the National Academy of Sciences. On the recommendation of that study, the EEOC then commissioned a 2015 pilot study, upon which the Component 2 collection is based.

It is not at all clear how the CNSTAT panel will be able to offer additional insights into the utility of the Comp 2 data, particularly for its intended purposes. If these data had been put to use in active enforcement activities, their utility in practice could have been tested. For instance, the use of Component 2 could have been restricted to a randomly selected subset of discrimination claims. Metrics regarding the timeliness and other elements of resolutions could have been compared in an experimental evaluation between claims where Component 2 was utilized and a control group of claims where those data were not considered.

Similarly, the EEOC could have published summary pay statistics in a subset of industry-specific and or geographic-specific reports and then through subsequent collections measured the impact of those summary reports on reducing pay differences, relative to industries and geographies where summary reports were not published.

It is challenging at the very least to see how the CNSTAT panel will be able to address the value of Component 2 in a context that is abstracted from the primary applications for which these data were intended. The success of the CNSTAT panel will depend on the degree to which it focuses on Component 2 as a tool to assist enforcement efforts in “the early stages of an investigation.” In contrast, the EEOC’s direction to examine “the utility of pay bands in measuring pay disparities” is of concern for at least two reasons.

First, this direction misses the point that Component 2 was never intended or advanced as a substitute for a full pay equity analysis, which would measure pay disparities controlling for legitimate variations in pay. Such a full pay equity analysis is offered by Trusaic’s PayParitySM Complete Pay Equity Audit Services.

Second, it is excessively narrow. Component 2 reveals elements of the compensation distribution by gender and race/ethnicity within job categories. Compensation differences within job categories may be reflective of differences in promotion, retention, or other factors, as well as of pay discrimination. In other words Component 2 would likely have value for assessing discrimination claims beyond a narrowly defined pay discrimination focus.

The EEOC indicates that its decision to further delay the use of these data is in response to the White House’s April 2019 revised interpretation of the Information Quality Act. The Union of Concerned Scientists describes how this act, and the Trump Administration’s revised interpretation of it, have given industries and special interests ample tools to endlessly challenge and forestall federal regulations grounded in empirical studies, all at taxpayers’ expense. In invoking the Information Quality Act as its motivation for commissioning a third pay data study, the EEOC appears to be relying on the act to undermine its earlier commissioned studies, at additional taxpayer expense. Come December 2021, when the CNSTAT panel issues its report, the act may again be used to attack and undermine the findings of the CNSTAT panel itself.

As Trusaic’s PayParity Post has previously noted, the tide towards greater pay equity enforcement is moving in the other direction, internationally and at the state level. For example, California’s senate has passed legislation that would collect Component 2 data from employers in the state. Even at the federal level, the pressure is mounting. First introduced in 1997, the federal Paycheck Fairness Act again passed the House of Representatives and mandates that the EEOC collect Component 2 pay data. Pay equity will also be a top priority for a Biden/Harris Administration.

If your organization wants to get ahead of these developments by becoming informed about potential pay disparities in gender and/or race/ethnicity, experts across the human capital, legal services, and business intelligence industries recommend conducting a pay equity audit.

Not sure how to get started? Let Trusaic provide your organization with a free Pay Gap Risk Assessment, which can be conducted confidentially.

Here is what you will get:

· One-hour consultation with Trusaic pay equity team members including regulatory compliance experts and data scientists;
· Answers to your pay equity questions;
· A pay gap analysis of your workforce.