Hearing of March 16, 2016 - Public Input into the Proposed Revisions to the EEO-1 Report
Madam Chair, Commissioners Barker, Lipnic, Feldblum, and Burrows and colleagues, my name is David Fortney and on behalf of The OFCCP Institute, I appreciate the opportunity to appear before the Commission today to present our views on the EEOC's proposal to revise the Employer Information Report (EEO-1).
I am an employment attorney and have practiced in this field for more than 35 years. I previously held senior leadership and policy positions at the U.S. Department of Labor's Office of the Solicitor. In my current practice, I represent a broad range of employers and federal contractors, and I regularly advise clients on complying with the EEO-1 reporting obligations and related matters.
In response to the proposed changes to the EEO-1 Report,1 it is my privilege to present the views of The OFCCP Institute ("The Institute") which is a national not-for-profit corporation whose mission is to educate federal contractors as to their affirmative action, diversity, and equal employment responsibilities.2 I am a co-founder of The Institute and I am very proud of our programs that assist hundreds of businesses understand and effectively respond to and comply with the new complex and technology-based affirmative action and non-discrimination compliance obligations through training and education. All of our members are dedicated to the principles of equal employment and fair compensation.
In that regard, I and The Institute fully support the goal of the EEOC to be more effective in enforcing the pay discrimination laws. We were encouraged when, at the behest of the White House's National Equal Pay Task Force, the EEOC called upon The National Academy of Sciences ("NAS") to help it decide what wage data to collect that would be useful to enforcement without creating an unnecessary burden on employers. The NAS Panel report, "Collecting Compensation Data from Employers" ("NAS report")3 included recommendations to collect data on rates of pay, conduct a pilot program, and enhance the federal government's ability to retain highly confidential pay data.
Despite our initial enthusiasm for the Commission's willingness to address this challenging issue as originally conceived, we frankly are disappointed with the results that are included in the proposed revisions to the EEO-1 Report. Below, I detail our specific concerns, and then provide alternatives that we respectfully recommend be considered to address the deficiencies in the current proposed revisions. Additionally, The Institute will be filing more detailed written comments in response to the EEOC's proposal that will include detailed information and data based on surveys of member companies in The Institute that currently are underway.
Based on our careful review, we have concluded that the EEOC's proposed compensation data collection program does not achieve its stated goals. The proposed data reporting does not meet the rigorous utility and burden requirements of the Paperwork Reduction Act4that must be addressed to justify the expanded collection of W-2 income and hours worked data through the EEO-1 data collection process. Moreover, the additional data will not assist either the EEOC or the U.S. Department of Labor's Office of Federal Contract Compliance Programs ("OFCCP") or employers in identifying existing pay disparities or detecting discrimination. As presented, the proposed revisions of the EEO-1 Report should not be implemented.
Perhaps the biggest flaw with the proposal is the EEOC's decision to reject the NAS recommendation and, instead, use W-2 compensation data. The NAS Panel's recommendation, after reviewing all the available options, was to collect data on rates of pay, not actual earnings. Of enormous significance, using rates of pay would allow for meaningful analysis of realistic statistical measures.
There are a number of material shortcomings in using W-2 data. As an initial matter, the W-2 data is over-inclusive.5 It includes items such as reimbursements for travel, transportation, relocation, tuition reimbursements, non-production bonuses, and outplacement services, which have absolutely nothing to do with "pay" as normally construed by the EEOC and the OFCCP. Additionally, many of the elements included inW-2 earnings are solely a function of employee choice, such as voluntary overtime, with no employer involvement, or are solely a function of employer largesse such as are provided through a wide range of benefits, or are unique to a particular circumstance and may or may not be repeated in subsequent years and are variable by definition (i.e., commissions, awards and bonuses).
Moreover, the types of supplemental earnings included withinW-2 data will vary significantly from employer to employer, based on the types of benefit programs offered (e.g., group term life insurance, accrued leave, sick pay) as well as compensation structures (e.g., advance commissions, profit sharing plans, nonqualified deferred compensation plans). As a result, it is more than likely that no two employers will include precisely the same earnings in their W-2's. Thus, the inclusion of these variables will prevent the Commission from making accurate comparisons between similarly situated employees at similarly situated employers. Since that comparison is the agency's primary reason for collecting compensation data, we have difficulty understanding why EEOC has opted for overly inclusive W-2 data when other, superior data sets are available. In sum, the IRS definition of "compensation" includes extraneous variables that will skew results and result in anomalous comparisons.
What is more, the Commission's decision to rely upon "existing" W-2 data will in no way lessen the burden on employers. In fact, this requirement will significantly increase the burden for all employers. The EEOC's EEO-1 survey is generated and submitted at the end of the EEOC's fiscal year, September 30, as opposed to the calendar year, so employers cannot utilize pre-existing W-2 reports and data, which are prepared by January 31st to report on income from the preceding calendar year in accordance with the Internal Revenue Code requirements.6 Instead, the Commission's proposal will require employers to generate "faux" W-2s for each employee using earnings data from a 12-month period which typically will span two separate calendar/tax reporting years, thus magnifying the variability inherent in W-2 earnings data. In our view, the federal government should not impose duplicative W-2 reporting requirements on employers at two different times of the year to meet the different timelines of two federal agencies - that is not a sound policy approach nor one that passes muster under the Paperwork Reduction Act.
The proposal will also require the costly creation of software "patches" to integrate employers' human resources information systems ("HRIS") - which are used to populate the EEO-1 Report currently - and payroll systems which are used to produce W-2 reports. Most HRIS experts will tell you this is no easy feat. In addition to devising a way to import the W-2 data, employers must then match it to their EEO-1 data: two different data sets housed in separate systems. Employers who rely upon outside payroll services will also incur additional annual fees associated with generating W-2 reports mid-year, importing that data to their systems, and then integrating that outside data with their EEO-1 data. Again, in the limited time that the EEOC has provided for comments, we are surveying our members to develop a more detailed understanding of these burdens, but our preliminary investigation indicates that these burdens are significant and are far greater than the projections provided by the EEOC.7
In light of the unavoidable problems W-2 earnings creates (and in light of the NAS recommendation), we question why rates of pay - that is, annualized base pay - was not chosen as the appropriate data to be collected. When base pay is annualized, then potential W-2 disparities among equally compensated employees based on whether a full reporting year in the same position was worked by the employees is eliminated. Thus, the proposed collection of hours worked by each employee would not be necessary if annualized base pay was reported. Further, when employers monitor their pay practices, annualized base pay typically is used for those analyses. Finally, base pay data are available in employers' HRIS systems, which would significantly reduce the reporting burden.8 Although W-2 data may be more comprehensive, that is its chief deficit: the additional data included in the W-2 are not representative of the pay for performance that is at the root of the EEOC's concerns. They are distractions.
Lilly Ledbetter's circumstances, which President Obama again highlighted to announce the expansion of pay data reporting, demonstrate why W-2 data are not what is focused on when discussing fair pay. Ms. Ledbetter complained about her base pay - she did not care about retention bonuses, relocation pay, severance pay, or the other extraneous data that routinely are included on a W-2. Instead, her complaint was that her base pay was lower than her male colleagues, and that her lower pay resulted in the reduction of her benefits and her pension. Base pay was the source of Lilly Ledbetter's pay complaint that the President wants to address, and that should be the focus of the EEOC's data collection efforts. If the Commission wants an effective means of identifying addressing and eliminating pay discrimination, it should focus on base pay - like Lilly Ledbetter did.
And what Lilly Ledbetter knew, the government's reporting fully confirms. The U.S. Department of Labor's Bureau of Labor Statistics has reported that variable pay is a tiny component of pay for the vast majority of the American workforce. Indeed, the EEOC's Pilot Study conducted by Sage Consulting cites Bureau of Labor Statistics data finding that supplemental pay accounts for only 2.4 percent of total compensation for all civilian workers.9 Given the difficulties in making meaningful pay comparisons using W-2 data and the additional burden employers will encounter generating "faux" W-2 data mid-year and then migrating that data to an independent HRIS system, it is simpler and more accurate for the Commission to focus its compensation collection efforts on collecting base pay. Using base pay reporting will account for 97.6 percent of the civilian pay, which is compelling reporting coverage. Why should 100 percent of employers bear huge burdens to capture 2.4 percent of compensation?10
In conclusion, base pay is more precise and more easily accessed than W-2 data. Base pay is a far superior measure of compensation for the purpose of identifying the existence of pay discrimination. The Institute and its members urge the Commission to follow the recommendation of the NAS and focus on rates of pay as collected by employers across the country: annualized base pay.
As briefly noted above, in addition to creating a more accurate data analytical tool, annualized base pay will also be significantly less burdensome on employers. In contrast with W-2 data, annualized base pay typically is already housed in an employer's HRIS system. Our members report that annualized base pay is far easier to generate - without any reference to the daunting challenge of creating "faux" mid-year W-2 earnings reports. Using annualized base pay would allow for a point-in-time report pulled for EEO-1 purposes which would easily include the employees' relevant pay data. Accordingly, the use of annualized base pay solves the IT integration issues as well as a host of issues related to the timing of the reports.
The EEOC's proposal also fails to generate data that can be relied on to identify the potential compensation discrimination that should be addressed. Although the EEOC's proposal concedes that the revised EEO-1 Report is not intended to be an indicator of discrimination, we believe the results are wholly irrelevant to increasing the ability to identify employers with unlawful discrimination. The first problem with the proposed EEO-1 is that salary data will be reported by EEO job categories which are unrelated to any employer's compensation system and unrelated to the way in which employers make compensation decisions. This is so because the job categories are, like the W-2, broadly over-inclusive. Instead, employers use more narrowly defined salary grades/bands or market data. As an example of this problem, look at the EEO job category of "Professionals." This single category might include, for a single employer, such varied jobs as engineers, financial analysts, attorneys, physicians, accountants, and computer technicians. How would such information about such disparate jobs at such disparate pay levels be useful to EEOC in making decisions about pay discrimination?
A related, and most important problem, is that an organization's EEO-1 results using EEO job categories and the 12 pay bands will be error-filled and will not mirror the results of a properly conducted pay equity analysis. To test this assumption, we analyzed the data from a client for whom a compensation assessment recently had been completed, in accordance with Title VII principles. We then used the company's actual workforce data and created an EEO-1 Report, in the proposed format. The EEO-1 Report indicated 11 statistically significant variances in job categories across three race/ethnicity categories. However, 10 of these 11 "flags" based on the EEO-1 pay report were false indicators - there was only one instance in which an actual compensation disparity was accurately predicted. Under this example, if the revised EEO-1 Report was used, and either the EEOC or OFCCP investigated the employer, it would have been a fool's errand, a fruitless waste of both government and employer's resources chasing after false indicators.
The proposed methodology also disserves employers in another way. Given the potentially low validity of the proposed EEO-1 Report, we are concerned that employers might see a "clean" EEO-1 as a sign that they do not have any gender or race/ethnicity pay issues and, then, not conduct the proper research that is needed to determine if there are actual issues. Weak, poorly conceived data creates misleading results for everyone - public and private sector alike.
We respectfully submit that the EEOC should not collect hours worked data, which poses significant burdens for little benefit to the EEOC and OFCCP. The proposal asks that employers report employees' total hours worked for the12-month reporting period. EEOC maintains that doing so "will allow analysis of pay differences while considering aggregate variations in hours."11 However, this "solution" ignores the reality that EEOC will not be receiving an accurate depiction of actual hours worked. The reality is that few employers track the hours of their exempt employees.
Since few employers collect hours worked for significant percentages of their workforces, the Commission will be defaulting to a fictional 40 hours for full-time exempt workers (2080 hours/year) and likely 20 hours a week (1040 hours/year) for part-time exempts. This is no solution. A "default" is not real and, therefore, cannot generate a meaningful analysis, especially for something as important as pay discrimination. Using the estimated data will yield not a functional hourly rate but a fictional one, and, therefore, the data will not be a meaningful measure for the purpose of determining pay discrimination.
The sole purpose of requiring the hours worked to be reported on the EEO-1 Report is to be able to determine an estimated hourly rate for comparison purposes. Adding to the value of the recommendation that annualized base pay should be used instead of W-2 earnings is this: the use of annualized base pay will eliminate the need to include "hours worked" as part of the EEO-1 Report because such comparison can be easily made from the base pay data.
The insertion of Salary Bands into existing EEO-1 job categories further complicates the reporting process. The 10 EEO-1 job categories have no relationship to any employers' compensation system. The addition of the 12 pay bands to each job category adds complication without clarity. As a result of the salary bands, exempt and non-exempt employees, unionized and non-unionized employees, workers paid on commission and those paid by the hour, will all be lumped together. It is impossible to understand what meaningful comparisons among employers can be derived from this system.
We also see a number of issues related to the confidentiality of the data. Again, this was a concern of the NAS report, which specifically recommended improvements. The EEOC's proposal fails to address the NAS report's recommendations.
EEOC personnel are subject to criminal sanctions if they make EEO-1 reports public under 42 U.S.C. § 2000e-8(e), but Title VII's privacy protections do not extend to the OFCCP, which is bound only by the Freedom of Information Act ("FOIA").12 Under current procedures, if the OFCCP receives a request for EEO-1 data, OFCCP provides employers with notice of the FOIA request along with a copy of the specific information being requested. If an employer does not timely respond to the OFCCP, the agency will release the data. If, on the other hand, the employer objects to the release in a timely manner, the OFCCP will determine whether the requested information is protected under FOIA's exemption for trade secrets. This process does not provide the same level of protection for pay data held by OFCCP as is provided by the EEOC. Moreover, we are aware that a different employer's data has been sent in response to a FOIA request; we have received reports that OFCCP has lost and disclosed pay data collected during compliance evaluations from federal contractors. Thus, it is particularly important that confidentiality protections be expanded before the OFCCP receives extensive pay data from all federal contractors.
Once the EEOC collects compensation data as part of the revised EEO-1 Report, the stakes will be even higher. Because of the number of cells and the granularity of the data, there is a distinct possibility of breaching not only confidentiality, but of anonymity, as well - revealing an individual's compensation.
Given these concerns, should the Commission insist on moving forward with changes to the EEO-1 survey, the EEOC and the OFCCP should join with the employer community in urging Congress to extend Title VII's confidentiality protections to the OFCCP and to any other federal agency personnel with whom the EEOC shares EEO-1 data.
It is the Institute's belief that the EEOC's proposal is well intended, but not effective. The data collection process should assist enforcement agencies in identifying pay discrimination. Yet virtually nothing in the proposed procedure can reach that goal. The EEO-1 Report uses inherently flawed data in an artificially created "faux" W-2, all for the purposes of creating distributions of pay on an over-complicated chart with thousands of cells. The entire exercise is for the purpose of identifying variances within and among these cells and, then, between and among different employers using different pay systems. Even the EEOC recognizes that these variances are not related to discrimination.
We believe that because of the substance of the data collected and the manner in which it is analyzed, the EEOC's efforts will be fruitless. As stated above, the proposed procedure is one that cannot achieve the ends for which it is proposed. We urge that the suggestions we have made today - which will also be included at greater length in Comments we plan to submit - will be considered in the spirit with which they are made, that is, in the hope of formulating a pay data collection procedure that will genuinely assist in the fight against unlawful pay discrimination.
1The EEO-1 Report changes were published in 81 Fed. Reg. 5113-5121 (February 1, 2016), and comments are to be submitted by April 1, 2016. We understand that the EEOC's hearing on March 16, 2016 also is part of the public comment efforts, and we request that this Remark be included in the comment record.
3 NAT'L RESEARCH COUNCIL OF THE NAT'L ACADS. COMM. ON NAT'L STATISTICS, Collecting Compensation Data from Employers (2013), available at http://www.nap.edu/catalog.php?record_id=13496.
5 The W-2 form is available on the IRS website, and includes the form and the instructions on how employees should use the form. https://www.irs.gov/pub/irs-pdf/fw2.pdf. Included as an attachment is a copy of the relevant pages within the W-2 instructions that explain the wide range of income that must be reported on the W-2 that far exceeds wages.
8 Some have suggested that many of the problems created by the proposed reporting procedure would be reduced were the report date moved to, say, April 1. In that way, a real rather than a "faux" W-2 could be used. However, that would not address or resolve the fundamental problem that W-2 earnings are so over-inclusive as to undermine the EEOC's efforts. Annualized base pay remains a superior database in every respect: more easily accessed, more representative of actual pay for work, and it provides a common data input for analysis.
9 ''EEOC Pay Pilot Study,'' September, 2015, Sage Computing. Available at: http://www.eeoc.gov/ employers/eeo1survey/pay-pilot-study.pdf. at 7; citing, U.S. Bureau of Labor Statistics. 2015b. "Occupational Employment Statistics: Frequently Asked Questions." Available at http://www.bls.gov/oes/oes_ques.htm.
10 As comprehensive as it is, even the W-2 does not capture the incentive compensation - stock options - that so concern many in the regulatory community. Thus, the W-2 does not even do the job for which it has been picked: being comprehensive with respect to variable pay.