“Component 2” pay data reporting. What a long strange trip it’s been.
The recently released report of the National Acad¬emies of Sciences, Engineering, and Medicine’s Committee on National Statistics (NAS) – Collecting Compensation Data from Employers, Panel on Measuring and Collecting Pay Information from U.S. Employers by Gender, Race, and National Origin (the Report) – suggests the trip will be longer, and end with employers having federal annual pay reporting requirements in addition to current and future state requirements.
Some history is in order. See Circa’s blog, EEOC Pay Data Collection – Part 1: How Did We Get Here? To summarize, the trip began in 2012 when an EEOC-funded National Research Council report “Collecting Pay Data from Employers” recommended that EEOC not require employers to produce pay data until government had a “definitive plan for how the data would be used” and “collect data on rates of pay, not actual earning or pay bands.” Ultimately, however, without a definitive plan for use, the government required employers to produce pay data for the 2017 and 2018 calendar years, using pay information from the W-2 form (i.e., actual wages, not pay rates). In July 2020, EEOC commissioned NAS to “document lessons learned … and identify ways to improve potential future collections”.
From a practitioner’s point of view, Recommendation 5-2 is likely the most important finding in the Report: “Before future collection of Component 2 data, EEOC should conduct a field test to investigate issues of burden, data availability, and questionnaire design.”
The reason being, according to the Report, “important data-quality issues exist in the 2017-2018 Component 2 data”. These include “missing data, response inconsistencies, and implausible extreme values” in the 2017-2018 data set. Other issues include problems with company and enterprise report identifiers, the wide EEO-1 job categories and wide Component 2 and pay bands that mask pay differences. (Bands “equate employees who may be receiving very different levels of pay, particularly for top earners”.) Perhaps the testing that the Report recommends will surface additional quality issues. Recommendation 5-1 does say “… some issues would be best addressed by modifying the basic data collection methodology”.
Nevertheless, the Report notes that “The 2017- 2018 Component 2 data are a potentially valuable resource” (Conclusion 1-1). The Report states that “The data could be used as an initial step in prioritizing investigations and the allocation of resources, including for calculating pay gaps within an establishment under investigation, making comparisons across similar establishments, and for systemic investigations.”
It is remarkable and important that the Report acknowledges that EEO-1 job categories are not useful to study pay equity (Conclusion 3-3) and that the data collected may only be used, after cleaning, to estimate pay gaps (Conclusion 6-1). However, while the Report outlines these and other limitations of the data and some of the concerns to address before attempting to use it, EEOC’s July 28, 2022 announcement of the Report results appears to throws these caveats to the wind. It states that the Report “Concludes that the EEOC’s Pay Data Collection Completed in 2020 Could be Used to ‘Prioritize Investigations and the Allocation of Resources.’”
Given the Report’s holdings, caution is advised before relying on the 2017-2018 Component 2 data collected.
As noted above, since 2012 many experts have suggested that federal pay equity review focus on rates of pay, not actual earnings or pay bands. Employer representatives providing input to the panel in November 2021 also recommended the use of base pay. The Report, however, dispatches this request: “A second alternative measure is annualized base pay, which imposes a relatively low filing burden on employers. However, this measure may be inappropriate for the Component 2 data collection as it fails to capture compensation that comes in other forms” (e.g., bonuses, commissions, and fringe benefits).
Well, the earning detail from Box 5 of the W-2 form recommended by the Report has its own problems. The Box 5 pay reported for individuals who were hired or promoted in November will not align with the earnings of peers who held their jobs and pay all year – however, newcomers and old hands will be reported together. If mostly women who were recently hired or promoted, their reported Box 5 pay could seem low in comparison to males – when all could in fact be paid at the same annualized rate.
At a more technical level, multiple experts and stakeholders have also raised the issue that legitimate compensation factors such as tenure with the company or productivity may impact each component of compensation differently. For example, base pay and tenure usually grow together, but tenure may only impact bonuses through minimum tenure eligibility requirements. The Report does not provide guidance regarding the statistical analysis of W-2 data or any other total compensation concept.
Given that there are challenges with any means of compensation data collection, and that as the Report states, base pay may have a “relatively low filing burden on employers”, let’s hope that the recommended field testing of pay data includes review of the utility and burden of base pay reporting.
One of the NAS Panel charges was “to identify ways to improve potential future collections.”
The panel recommended that to improve future collections more types of data need to be collected. These include detail on employees 40 years of age and older, veteran status, disability status, and additional race detail (asking for the specific races of individuals identifying as two or more races or Hispanic).
Perhaps the biggest element of mission creep is the Report recommendation that rather than aligning employee information into summary pay bands, future EEO-1 reporting provide information on each individual employee. The Report recommends that individual-level detail include educational levels and job experience. The Report assumes, without support, that reporting individual employee-level detail “could substantially reduce respondent burden.”
This assumption cries out for field testing. Will such detail be a breeze to produce? What do employers think about providing such extensive business detail? Will the additional trillions of lines of data facilitate meaningful pay equity review? Seeing that different companies define their employees, their jobs, their levels, and so many of the elements of their pay systems so differently, how could it? Let’s also not forget that HR systems do not commonly record all relevant fields that may impact compensation. For example, educational information is seldom complete in HR systems, and when it exists it may lack important characteristics such as degree and major.
Speaking before a House of Representatives Education and Labor subcommittee in April 2022, EEOC Chair Burrows said that EEOC staff had already “looked at the pay data and are able to say that it would be very useful to us.” Relying on the Report, another point of view is that great caution should be taken if using the collected 2017-2018 Component 2 data.
Certainly, field testing should be done before any further pay data collection is required and the results of such testing should be widely shared and scrutinized. In addition, the Report itself states, “Further analysis is important if Component 2 data are to be used by EEOC for enforcement or research purposes.” There should be concrete evidence of benefit in terms of guiding compliance efforts or benchmarking progress before requiring firms to shoulder the cost and burden of this initiative.
Regular pay data collection is coming. Let’s hope that it does not come before testing is completed and a worthwhile plan for data collection and analysis is in place. This has been a refrain familiar since 2012.