Hearing of November 20, 2019 - Public Hearing on the Proposed Revisions of the Employer Information Report (EEO-1)

PRESENT:

JANET DHILLON, Chair

VICTORIA A. LIPNIC, Commissioner

CHARLOTTE A. BURROWS, Commissioner

ALSO PRESENT:

SHARON GUSTAFSON, General Counsel

KETAN BHIRUD, Legal Counsel

BERNADETTE B. WILSON, Executive Officer

PANEL:

LYNN CLEMENTS, Director, Regulatory Affairs, Berkshire Associates

MICHAEL EASTMAN, Senior Vice President, Policy and Assistant General Counsel, Center for Workplace Compliance

JOCELYN FRYE, Senior Fellow, Center for American Progress

JOSHUA MITCHELL, Senior Economist, Welch Consulting

JESSICA STENDER, Senior Counsel for Workplace Justice and Policy, Equal Rights Advocates

BETSEY STEVENSON, Professor of Economics and Public Policy, The Gerald R. Ford School of Public Policy, University of Michigan

This transcript was produced from a DVD provided by the Equal Employment Opportunity Commission.

PROCEEDINGS

CHAIR DHILLON: Good morning everyone.  The meeting will now come to order.  Thank you all for being here.

Good morning. Welcome and thank you to my fellow Commissioners, to all EEOC staff and members of the public who are here with us in Washington and to all those in our field offices who are watching us online.

I'd like to extend a special thanks to our distinguished panel of speakers for participating in today's public hearing. I appreciate all the time you've spent to put your testimony together and to come here to share your expertise with us today.

This hearing is being called pursuant to Section 709(c) of Title VII of the Civil Rights Act of 1964 and is intended to provide an opportunity to stakeholders to comment on the proposed changes to the Employer Information Report, otherwise known as the EEO‑1 Form.

The proposed revisions were published in the Federal Register on September 12, 2019. The notice of this public hearing was published in the Federal Register on November 5, 2019. Let me briefly explain the procedures for this hearing.

A verbatim transcript will be made of today's proceedings. For that purpose, the hearing is being videotaped. We will then post a transcript on the EEOC website at www.eeoc.gov. The speakers' written statements will be available on our website later today.

As the presiding officer, I am responsible for regulating the course of this hearing and shall dispose of all procedural matters. Each speaker has been allotted the same period of time, approximately 6 minutes, for his or her remarks.

The speakers will be called in the order in which they appear on the agenda.  After the speakers complete their remarks, we will have a 15‑minute break, followed by two rounds of questions from members of the Commission.

While this hearing is open to the public, remarks and questions will not be taken from the audience. With that, I'd like to turn it over to Bernadette Wilson to address some preliminary matters.

MS. WILSON: Good morning, and before I begin, is there anyone in need of sign language interpreter services?

(No response.)

MS. WILSON: Okay, Good morning again Madam Chair, Commissioners, General Counsel, Legal Counsel, I'm Bernadette Wilson from the Executive Secretariat.

We'd like to remind our audience that questions and comments from the audience are not permitted during the hearing, and we ask that you carry on any conversations outside the meeting room, departing and re‑entering as quietly as possible.  Also, please take this opportunity to turn your cellphones off or to vibrate mode.  I would also like to remind the audience that, in case of emergency, there are exit doors to the right and left as you exit this room.  Additionally, the restrooms are down the hall to the right and left of the elevator.

Madam Chair?

CHAIR DHILLON:  Thank you.  I would now like to invite my fellow commissioners to make any opening statements or comments, beginning with Commissioner Lipnic.

COMMISSIONER LIPNIC:  Thank you, Chair Dhillon.  Good morning everyone, and good morning especially to our witnesses, and thank you all so very much for being with us and for the time you've taken to put into your testimony and answering our questions today.

I also want to thank the Chair's staff, my staff, Commissioner Burrows' office, and the Office of Legal Counsel for all of their work in today's Hearing and organizing all of this.  And I also want to take the chance to welcome both our new General Counsel to the EEOC, even though you've been here for a few months, and our new Legal Counsel.

Well, it's deja vu all over again.  Certainly, especially for me and at least one of our witnesses, this has been quite an ongoing experience, but certainly we meet today not drawing on a blank slate.  We are drawing on a ‑‑ quite a very full slate.  And we are also at the point now, I think, where we are not engaging in speculation.  We have a fair amount of experience to draw upon.

I do want to provide just a little bit of context.  Certainly, the idea of having the federal government collect pay data information from employers as a public policy tool to help root out pay discrimination is not a new idea.  This is something for ‑‑ many people in the room may not be aware of this, but this was not a new idea in 2016.

In fact, this is an idea that goes back to 1997 when the first version of the Paycheck Fairness Act was introduced.  And so, while I don't have quite that many years of experience on this issue, I do have the experience dating back to the early 2000s, as at least certainly one member of our panel does, so ‑‑ and since tomorrow's my birthday, I feel free to engage in some ageism on all of this, so.

So when, early in the 2000s, OFCCP created a type of pay data collection that was similar to the situation we find ourselves in today, something that had gone out to the regulated community, at that time, only federal contractors.  They submitted the data.  And when I arrived at the Department of Labor as Assistant Secretary, the question that I asked about it was, does it work?  Does it do what it was intended to do? 

And we did not have any definitive answer to that question, so we funded a study by the Abt organization ‑‑ at least one of our witnesses has referred to that study in his testimony ‑‑ and found out that there were significant false-positive and false-negative problems.

So, that is among the things that I hope that we will be able to get to today.  As I said, we don't draw on a blank slate.  We don't even really have a lot of speculation, but we can certainly take the evidence in front of us and make some good decisions, I hope, going forward, about what we need to do as an agency in consideration of what brings us here today. 

So, I look forward to all of our ‑‑ everyone's testimony today.  And I just note that ‑‑ two points.  One, certainly for purposes of what makes for good regulation, there has to be a balance between what is the burden and what is the utility.  That is what is required of us as policymakers.

And I would also note that I think that everyone here today, we would all stipulate that we all care greatly about equal pay and making sure that we are able to root out pay discrimination and, for the government, have the right tools available to us to help us do that.

So again, thank you all for being with us, and I look forward to your testimony.

CHAIR DHILLON: Thank you, Commissioner Lipnic.  Commissioner Burrows?

COMMISSIONER BURROWS:  Good morning.  Thank you.  And thank you to my staff and to the staffs of the Chair and Commissioner Lipnic, and all of you who spent so much time, and certainly for the Executive Secretariat.

In our 2016 notice announcing ‑‑ authorizing the current ongoing pay data collection from employers, for the very first time, this Commission stated that persistent pay gaps exist in the U.S. workforce correlated with sex, race, and ethnicity.  And that remains true today.

When the Equal Pay Act passed in 1963, women working outside the home made just 59 cents to the dollar compared to White men.  Today, more than half a century since Congress outlawed pay discrimination in employment, the gender pay gap stands at 82 cents to the dollar, not nearly enough progress for five decades.

The gender pay gap is even more significant for women of color.  In 2018, Black  women earned just 16 cents on the dollar, native women ‑‑ Native American women only 57 cents, and Latinas only 54 cents as compared to White, non-Hispanic men.

Significantly, today's hearing to discuss ending the EEOC's pay data collection effort falls on Latina Equal Pay Day, so named because it takes Latina women until late November, almost a full additional year, to make the same amount that non-Hispanic men ‑‑ White men make in the previous year.

Although some of these wage gaps can be explained by occupation, experience or education, a significant portion of these wage disparities cannot be explained by such factors. The EEOC has sought in a variety of ways for a number of years to address this pay discrimination and the job segregation that fuels unjustified pay gaps.

The addition of the Component 2 data collection to the EEO-1 was one effort to tackle that problem.  As a result, EEOC is currently in the process of collecting Component 2 pay data for calendar years 2017 and 2018.  Although the Commission has not yet completed this historic first collection, much less examined its results, we are here to discuss whether to halt future collections.

In doing so, the Paperwork Reduction Act requires we carefully balance both benefit and burden in light of our important mission to prevent and remedy discrimination.  Thus, although the pending proposal to end the pay data collection discussed only the burden, the testimony today appropriately discuss ‑‑ recognizes our duty to look at both benefit and burden.

I take that duty seriously and look forward to our testimony today, and possible alternatives to that effort or to continuing the pay data collection.  Thank you very much.

CHAIR DHILLON: Thank you, Commissioner Burrows.  And I'm pleased now to introduce our panelists for today's hearing. 

Lynn Clements is the Director of Regulatory Affairs, Berkshire Associates.  Michael Eastman is the Senior Vice President of Policy, Assistant General Counsel for the Center for Workplace Compliance.  Jocelyn Frye is a senior fellow at the Center for American Progress.

Joshua Mitchell is a senior economist at Welch Consulting.  Jessica Stender is a senior counsel for workplace justice and policy at Equal Rights Advocates.  And Betsey Stevenson is a professor of economics and public policy at the Gerald R. Ford School of Public Policy in the University of Michigan.

Today's hearing will consist of one panel.  We will then open the floor for questions from members of the Commission.  Panelists, you will each have six minutes to make your oral presentations today.  Your complete written statements will be available on our website and placed in the meeting record.

Please note that we're using timing lights at the center of the console in front of me.  The yellow light will appear when you have one minute remaining on your statement and the red light will appear when your allotted time expires.

We will take a short break after hearing from each of our panelists, and we will then resume with two rounds of questions and closing remarks.  Again, we're very pleased to have such an exceptional panel here today and we thank you for being here.

So we will begin with Mr. Eastman from the Center of Workplace Compliance.  Mr. Eastman?

MR. EASTMAN:  Good morning, and thank you for the opportunity to be here today.  I want to start with Component 1, if I may.  This is a part of the EEO-1 report that has existed in its current form since 2007.  We worked long and hard with the EEOC in a collaborative process to get the form to where it is today, and we strongly support renewal of Component 1.

I do want to make a point about the burdens involved in completing that component of the EEO-1 report.  There seems to be a misconception by some that because most records are stored electronically, that they can simply be printed out and turned in; however, it's important to note that submitting materially false EEO-1 reports is a felony. 

And any employer that takes its obligation seriously is going to carefully check that data.  They're going to take hours, perhaps days, to prepare, to be sure they pull the correct data.  Once they pull it, they're going to do a data quality check and look for missing values, such as missing race and ethnicity information.  You can't have an employee without a value attached to it.

So there are many, many hours involved in just completing Component 1 of the report, and we think the Commission has understated that burden in its proposal, and we encourage you to take another look at that information.

As to Component 2, let me say that the question here isn't whether enforcing pay discrimination laws strongly should be an important priority for the Commission, right?  The question is, is Component 2 an effective tool for the agency to do that?  And we submit, respectfully, that it is not.  And there's two points that I want to make about that today.

The first is, there is some question whether any summary pay data collection tool is going to be able to equip the Commission to effectively combat these kind of pay discrimination issues.  And that's because the data that's collected in that form is too aggregate and influenced by too many variables that don't affect pay.

If we look at the Department of Labor's enforcement statistics and experience collecting actual employee pay data from audited federal contractors, I think that's very telling.  So, I went yesterday to DOL's website and looked at their enforcement database.

Over the last four years, they have conducted 7,352 compliance reviews covering 4.8 million employees.  They found potential contractor violations in 17 percent of those audits. That's 1,257 cases.  And these are conciliation agreements, consent decrees, and financial agreements.  They're not a finding by a neutral fact-finder.  These are settlements, basically.

Now, in all of these cases, they found potential pay discrimination in 60 cases. That's 0.8 percent.  We respectfully submit that, if OFCCP finds pay discrimination in less than one percent of the cases when they've looked at actual pay records for an entire establishment, or an entire workforce, then the summary pay data collection tool is going to be far less accurate.

The other point I'd like to make is, if the EEOC should choose to return to a pay data collection tool, that it return to the 2012 recommendations by the National Academy of Sciences.  This, I think, is a really important step, if the Commission is to return to the pay data collection tool.  And there's two that I'd like to focus on today, of these recommendations.

The first sounds relatively straightforward.  It was the first recommendation made by the Academy.  It was that the Commission and the other enforcement agencies that look at  pay discrimination first come up with a clear policy of how that data will be used in an enforcement regime that is largely built around a complaint process, right?

In 2016, the reasons for the pay data collection and the uses for it seemed to change with each iteration of the proposal.  It was almost like the plan was to first collect the data and then to come up with the policy of how it would be used.  And we encourage the Commission to first very carefully consider what the policy is, before it in fact comes up with the data collection tool.

Secondly, it's really important to have an actual pilot study to test the instrument before it goes out into the field.  Yes, in 2016, there was a publication that said the words pilot study in the title.  But as we all know here, labels are not determinative.

In 2016, there was no attempt to test the data collection tool Component 2 on actual employers and to figure out if it would be of any utility in finding discrimination.  And, of course, there was no attempt to assess the burden using that actual tool.  And if you look at that report, it says in a footnote why that didn't happen. 

And that was because the agency didn't want to go to OMB for approval for a pilot study.  So instead of a pilot study of a hundred or a thousand employers, it pushed out Component 2 to 90,000 employers.  We urge the Commission not to repeat that mistake.

One final point that I'd like to make today in my opening comments is we think the Commission would be correct to go through the Administrative Procedure Act should it choose to return to a pay data collection tool.  It's true we haven't done that in the past with the EEO-1 report, but I think that's because the stakeholders have largely agreed on the type of data collected through the EEO-1 report historically.

With a much more controversial data collection, I think the added protections afforded by the APA would be important, and those include things like judicial review, Regulatory Flexibility Act, and so forth.  Thank you very much.  Look forward to your questions.

CHAIR DHILLON:  Thank you, Mr. Eastman.  Ms. Frye?

MS. FRYE:  Good morning.  Thank you Madam Chair and Commissioner Lipnic, Commissioner Burrows.  Thank you for the invitation to be here today.  I am Jocelyn Frye and I'm a senior fellow at the Center for American Progress.  And as you alluded at the outset, I do have a prepared testimony, and I'm just going to pull from some pieces of it for purposes of this statement.

The Center for American Progress is a non-partisan think tank.  We are committed to improving the lives of all Americans.  And in particular, the Women's Initiative was founded to focus, in particular, on expanding women's opportunities by ensuring their economic stability and security and promoting fair and equitable workplaces.

And it is precisely because of this work that we strongly supported the 2016 revisions to the EEO-1 form to collect pay data.  And it is also the reason that we are opposed to the current proposal that's being discussed today that would effectively discontinue the Component 2 pay data collection. 

We believe pay data collection is critical to enforcement and to addressing issues around equal pay, and to discontinue it is unwarranted and we think would be a step backwards, and so we strongly urge you to reconsider that decision.

I'm going to focus on the notice that the EEOC released in September about the decision to collect only Component 1 and not to collect Component 2, and raise a number of the concerns that we have about that notice, first focusing in particular on the burden and the discussion about the burden.  And the notice goes in detail to explain why it has revised how it is now analyzing burden.

The first concern relates to just the discussion about that decision itself.  There is, in our view, respectfully, not enough detail about how some of the numbers were reached, how some of the calculations were done, and particularly the hours estimates in terms of how much time it takes to actually file a form.

There seems to be an assumption that companies that have multiple establishments, the time would increase.  That may or may not be true for large employers that sometimes have greater technological capacity and ‑‑ to accommodate multiple filings. 

Some of the forms have higher estimated burdens associated with it, even though the number of employees that it would be -- would be calculated are actually smaller.  So, there's just some questions about how some of the underlying calculations were done.

Second is a concern about a built-in assumption that the burden costs would go up and not down in subsequent years.  You know, I think many think that the cost would go down as a company got more familiar with the reporting tool.

The third concern relates to an insufficient discussion about the real world experience with the EEO-1 form.  It is a form that is 50 years old.  There was no discussion in the notice about evidence that indicates that the burden of filing Component 1 data has increased or become more onerous over the past few years.  And so, to suggest that the burden has gone up precipitously seems to be not consistent with what we know.

Moreover, the fact that many of the filers have already submitted the pay data for 2017 and 2018 could mean that, as a practical matter, that much of the initial burden has already been taken care of, and so the subsequent years you would see the burden ease and not grow.

The notice does allude to but doesn't discuss in detail the fact that there are a growing number of technologies that would assist in the employer with filing.  We know that there are many companies out there that already do that work, and we would expect that to continue.  And so, to assume that they wouldn't have that technology available to make the burden less onerous, in their opinion, seems not to be true.

I also want to make sure that we put on the table the burdens of not doing something.  We focus a lot on the burden to employers, but there is also a burden on employees.  There's much that has been written about the wage gap. I won't talk to recount all the details.

I will say that we did an analysis of quarterly median weekly earnings for women and looked at the period 2017 to 2018 and found that a Black woman working full time earned almost $35,000 less than her White male counterpart.  For Latinas, it was almost $40,000.  And on Latina Equal Pay Day, that is a particular note.

And so for us, the burden of not doing anything and not collecting pay data is significant.  I also want to make sure that I do touch on utility and just simply say that the EEO-1 form is a proven document.  We do have 50 years of history and of its effectiveness.

I don't think the measure is, is the data perfect. The only way that I know to do that is to collect all of the pay data from all employers at all times, which is certainly an option before the Commission.  But I think the question is, as Commissioner Lipnic alluded to, can we balance appropriately the interest of employers with the enforcement needs of the agency, and our view is that you certainly can and collect pay data.

In conclusion, I just want to say that we think pay data collection is particularly important for all women, but particularly women of color who have persistently had a more -- a starker pay gap.  And because of that stark pay gap, we think that it's really important for the Commission to move forward in collecting data to really try to discern what's going on there.  And it's an action that is long overdue.  Thank you.

CHAIR DHILLON:  Thank you, Ms. Frye.  Mr. Mitchell?

MR. MITCHELL:  Chair Dhillon and Commission members, it is an honor to appear before you today to discuss the important topic of collecting pay data as part of the EEO-1 Component 2 program.

Before weighing the benefits and costs of continued Component 2 collection, it is helpful to review the elements of a pay equity analysis, the standard statistical framework for estimating gender and race/ethnicity pay differences in the context of litigation, OFCCP audits, or self-audits undertaken by proactive companies.

The analysis usually involves extracting employee-level microdata from a company's human resource information system.  The outcome variable is a compensation measure of interest that is usually expressed as a pay rate, such as an employee's full-time equivalent salary or quarterly bonus earned.

Other key data fields include a job category that meaningfully groups together workers who are similarly situated and other legitimate pay factors that reflect how that particular company sets pay.  Depending on the company, these pay factors might include the level of work being performed, an employee's time in their current role, overall company tenure, relevant outside work experience, educational attainment, certificates earned, performance review scores, et cetera.

The microdata are typically analyzed with a multivariate regression model.  Output from a pay equity analysis reveals whether there is a statistically significant, unexplained gender or race pay disparity among similarly‑situated workers, after accounting for legitimate pay factors.

Turning now to the EEO‑1 Component 2 form, it is important to keep in mind the ultimate objective: what are the EEOC and OFCCP hoping to use these data for?  The most important objective is to enable the EEOC and OFCCP to better target enforcement resources.

In order to meet this objective, it must be the case that an analysis of Component 2 pay data can meaningfully distinguish between companies that are and are not complying with the Equal Pay Act and Title VII of the Civil Rights Act.  This in turn means that an analysis of Component 2 aggregate pay band data must correspond well with what a formal pay equity analysis of the underlying microdata would show.

There are, however, at least six reasons why an analysis of aggregate pay band data is unlikely to yield similar results to a formal pay equity analysis.  First, the EEO job categories are very broad and unlikely to properly group workers together who are truly similarly situated.

A reporting hospital, for example, must categorize such diverse occupations as doctors, nurses, and attorneys in the same professional EEO job category.  

Second, the W‑2 Box 1 annual taxable wage concept is intended for IRS accounting purposes, but it is ill‑suited for comparing employee pay rates because it conflates employer and employee decisions and ignores employment changes that affect compensation. 

Employees who are promoted halfway through the year, for example, will have low annual taxable wages, even if they are receiving the same rate of pay as other similarly‑situated workers at the time of the snapshot.

Third, the 12-pay-band reporting scheme is unlikely to reflect the compensation distribution of most companies' pay structures.  Collapsing the data into pay bands can both exaggerate differences between employees who are paid just above and below an arbitrary cutoff, as well as hide considerable pay differences among employees in the same pay band.

Fourth, the grouping together of non‑exempt and exempt workers and the imputation of 20 hours for part‑time work and 40 hours for full‑time work to all exempt employees is not consistent with many companies' employment practices.

Fifth, the Component 2 form doesn't collect information on any legitimate factors that could explain differences in employee pay.  This means an employee who is just starting his or her career is treated the same as another employee who has worked at the company for 30 years.

Sixth, the kinds of statistical testing available with aggregate pay band data, such as Mann‑Whitney and interval regression, are unlikely to correspond well with the standard statistical testing used in employee‑level multivariate regression analysis.

For all of the above reasons, attempts to detect discrimination with the Component 2 data will likely lead to high rates of false positives, which means many companies will be  flagged for potential violations despite fully complying with the law.  Similarly, analysis of Component 2 data will likely lead to high rates of false negatives, which means companies that are violating the law will not be flagged for further investigation.

How high will these error rates be?  We at Welch Consulting simulated in 2016 the ability to detect discrimination using nationally representative data from the 2013 and 2014 Current Population Survey.  Across thirteen industry and EEO occupation categories, the key finding was that targeting enforcement based on the Component 2 form would do little better at detecting discrimination than selecting firms for audit completely at random.

Earlier research by Abt Associates reinforces this conclusion.  In their study of OFCCP contractors, they compared findings from the EEO survey, which collected summary pay data, to the results of completed OFCCP investigations of the same employers.  As described by the 2012 National Research Council panel, the principal conclusion was that the EEO Survey, quote, did not improve deployment of enforcement resources toward contractors most likely to be out of compliance and did not lead to greater self‑awareness or encourage self‑evaluations.

In conclusion, given the lack of utility of Component 2 data for detecting discrimination and the fact that some additional burden is inevitably imposed on employers ‑‑ and considerably more than was initially estimated ‑‑ it would be difficult to recommend its continued collection under the terms set forth in the Paperwork Reductions Act.  While I am committed to ensuring pay equity, it is clear to me that Component 2 collection does not further this goal.  Thank you for your time and I look forward to your questions.

CHAIR DHILLON:  Thank you, Mr. Mitchell.  Ms. Stevenson?

MS. STEVENSON:  Thank you for the opportunity to speak to you today.  Let me say unequivocally that reducing explicit and implicit gender, racial, and ethnic discrimination requires the collection of pay data.  I want to make four points and then the rest of my testimony will elaborate on these points, which are further elaborated in my written testimony.

First, I realize that the Commission recognizes the essential role of collecting pay data, and therefore discarding the collection of the Component 2 data without an alternative plan to collect data clearly leaves the EEOC in a weaker position to combat discrimination.

Second, alternative approaches that I outline in my testimony likely raise the burden to employers, and we must balance the burden to employers with the usefulness of the data.  The current approach was chosen because it achieved that balance.

Third, the EEOC should take a deeper look at its burden assumptions because there are some highly questionable assumptions and some easy approaches for them to be able to get more accurate estimates.  Finally, the EEOC should evaluate the data it has collected through Component 2 before determining the utility.

And I want to state clearly that the analysis should not look at the role of false positives and negatives only, but whether the collection of data serves the EEOC's purpose, which is to eradicate discrimination, not just to prosecute violators.

So, many employers actively do work to assess their gender and racial wage gaps.  As has been noted, they voluntarily provide microdata to firms that provide compensation analysis for them.  They sometimes do their own ‑‑ use software to do their own analysis internally.  And many employers collect this data, but that does not reduce the usefulness of the data collection done by the EEOC.  But it does actually decrease the burden imposed on the employers collectively.

While a detailed analysis of microdata might prove to be the most illuminating for employers, it is not clear to me that the EEOC should begin requiring firms to report microdata.  That's obviously one option that is available, is to gather this disaggregated data.  Instead, we can think about the usefulness of Component 2, where firms who see problems in Component 2 should ask very important questions.

So, what if a firm discovers that women within a job category are disproportionately in the lower bands of pay?  They need to determine whether they are paying women less for the same work, failing to promote and/or hire women into more senior positions at the same rate as men, failing to retain women, or engaging in occupational segregation practices that leave women in lower-paid positions without access to the career ladders that would help them secure higher pay.

While the data reported to the EEOC will not help either the employer or the EEOC pinpoint the root cause of the problem, it will provide an important signal that the employer needs to discover the root cause of their pay disparity.  And again, thinking about balancing the burden with usefulness, microdata would allow the EEOC to take a deeper role in trying to find the root cause.  But it would do so at the expense of putting an additional burden on employers.

So, it is easy to point to changes that could make the data more useful.  But let me say, clearly this has come up ‑‑ the 2012 National Academy's report said, quote, pay band data are attractive and that they align with the way that human resource managers tend to look at compensation.  They found that was a low-cost, low-burden approach.

And many firms currently have access to software that allows them to do this quite easily.  The growing penetration of highly sophisticated software applications by large employers makes it easier for establishments to provide this data that they're asked for, and these costs appear to be falling quite quickly through time.

The new burden estimates seem to me to be out of line with developments in technology.  And while I hear that there are many employers who are desperately fearful of the sanctions, I'd like to see more data on the sanctions that employers face from erroneously reporting mistakes in their EEOC forms before I would believe estimates of the amounts that they're investing in avoiding those sanctions.

One concern from employers is that they do have to program the software, they may have to link their payroll and HR systems. That ship has sailed.  They had to do that for the 2017, 2018, 2019 data.  And as Jocelyn Frye alluded, we don't need to consider those costs any further.

So, let me address now the usefulness of the data. There seems to be some misconception that binned data cannot be usefully analyzed.  It's simply false.  Researchers often confront income data for individuals, households, or families that fall into income bands, rather than providing a specific estimate.  It's done to preserve the anonymity of respondents, particularly when the data will be made publicly available.

Of course, the plan here is not to make the data publicly available.  But I will note that that is another alternative the EEOC could pursue in line with the approach that the United Kingdom has taken.  Reduce the number of bands, so you're making the data even more aggregated, but then make it publicly available so that workers, consumers, and other employers can assess people's gender wage gaps and put ‑‑ use that kind of market-based pressure as well as EEOC enforcement to push towards a greater gender pay equity.

So, that ‑‑ I would argue that that alternative approach would have a greater burden, but it's certainly an option we could consider, again, reducing the number of bins, but making it publicly available.

Finally, let me say, obviously the microdata, if you're going to actually do the kinds of pilot studies that have been suggested, I do think that that would ‑‑ requires having a pilot study that collects microdata from employers.  What you would do is randomize when ‑‑ which employers had to report the microdata and you could do that in-depth analysis, so ‑‑

CHAIR DHILLON:  Thank you, Ms. Stevenson.  We appreciate your remarks.

MS. STEVENSON:  Thank you very much.

CHAIR DHILLON:  Ms. Clements?

MS. CLEMENTS:  Chair Dhillon and distinguished Commissioners, thank you for the invitation ‑‑

CHAIR DHILLON:  Could you turn on your mic?

MS. CLEMENTS:  I knew I was going to forget that.  Sorry about that.  Thank you for the opportunity to speak with you today.  My name is Lynn Clements, and I'm the Director of Regulatory Affairs at Berkshire Associates.

I want to begin today by emphasizing that, as a female professional and mother of a young daughter who is quickly growing into a strong woman, I support strongly the principles of equal pay for equal work and non-discriminatory pay decisions.

For almost half of my career, I have advanced those ideals as a public servant both here at EEOC and at the Department of Labor.  For the past 10 years, I have helped employers comply with equal employment and affirmative action requirements.

My firm is an HR consulting and technology firm, and over the past several months we have helped employers across this country prepare and file hundreds of EEO-1 reports, both Component 1 and Component 2.  My testimony today focuses on these actual experiences in gathering the data needed for these reports and filing them with the Commission.

First, I want to take a moment to talk about Component 1 because that forms the basis for the Commission's burden estimates for Component 2.  While we appreciate the fact that the Commission revised the burden estimates, we do not believe that Component 1 data can be collected, verified, and reported by merely clicking a button in an HRIS system.

We are not aware of any employer who can file Component 1 in the 45 minutes that was estimated by the Commission.  Quite frankly, it sometimes takes that amount of time to find the login and password to get into the system.

Our experience preparing Component 2 reports has really amplified employer concerns about the burden and lack of utility to employers of collecting this data within broad pay bands and broad occupational categories.  The problems with collecting W-2 income and hours worked data has been well-documented, but a few items bear repeating here.

First, our experiences have shown us that Component 2's requirement to gather and synthesize data over an employee's 12-month period of time is much more complicated than the true snapshot that is used in Component 1.  This small change created a level of complexity that was severely underestimated because it required employers to manually reconcile various data points in order to report this information.

In our experience, this format significantly increased the burden on employers while also simultaneously decreasing the utility of the data because many assumptions had to be made in this manual reconciliation process.

Second, one of the stated goals of the Component 2 collection was to encourage employers to self-monitor their pay practices. However, using annual W-2 earnings did not promote self-analysis.  Time and again, employers expressed frustration to us about gathering W-2 data because the data point was heavily influenced by employee choices rather than employer pay decisions, which should be the focus under Title VII and the Equal Pay Act.

Compensation practitioners who looked at the data often stated that it had little utility in a reliable analysis of their company's pay decisions or systems.  The collection of hours worked data was also proven flawed.  Given the complexity of time-keeping systems and the varied ways they can be used, we found there was little consistency in reporting total hours worked information, despite the EEOC's guidance.

This decreases any utility of the data for identifying trends or patterns among employers.  Many employers also concluded that the two data points, when put together, often bore no reasonable relationship to each other.  For example, the W-2 earnings included wages for time employees didn't work, such as paid time off, while the hours worked figure reported did not include this time.

Perhaps most importantly, Berkshire has never once had an employer ask us to use W-2 wage information when conducting a salary equity analysis.  We are not aware of a single Berkshire client who decided to conduct a more robust analysis of their pay decisions as a result of this filing.

In fact, employers reported to us that in some cases, the substantial amount of time it took directed their efforts away from other pay equity initiatives.  This is the exact opposite of the ‑‑ one of the stated goals of Component 2 data.

Our written testimony provides detailed information about the amount of time it took our clients to file these reports.  Suffice it to say that we have clients who literally spent hundreds of hours collecting and verifying the Component 2 data.

During this process, it has been argued that Component 2 reporting will be less burdensome in future years. I want to make clear that our experience tells us the exact opposite.  Many of the issues that cause significant burdens to employers, such as the large number of employee movements between roles or EEO job categories, changes in exempt to non-exempt status, the impact of mergers and acquisitions, or the use of multiple different systems to record the information being collected ‑‑

CHAIR DHILLON:  Yes.  Thank you, Ms. Clements.  Can you just wrap up briefly?

MS. CLEMENTS:  Of course -- will remain in future filing years.  Despite the significant amount of time, employers will not be able to simply push a button ‑‑

CHAIR DHILLON:  Thank you, Ms. Clements.  We'll have more opportunity to speak with you ‑‑

MS. CLEMENTS:  Thank you.

CHAIR DHILLON:  -- during the questioning.  Ms. Stender?

MS. STENDER:  Good morning.  Thank you for the opportunity to be here today.  My name is Jessica Stender and I'm Senior Counsel for Workplace Justice and Public Policy at Equal Rights Advocates.  We strongly supported the revision to the EEO-1 form approved by OMB in 2016, and I'm here today to respectfully urge you to renew the Component 2 pay data collection.

For more than 55 years, we've had the Equal Pay Act on the books, and yet women ‑‑ and women make up almost half of the workforce; more than half are the primary breadwinners in their households.  And they receive more than half of bachelor's, master's, and doctorate degrees.  Yet U.S. census data reflects that, on average, women continue to earn less than men in every industry and occupation in this country.  And of course, for women of color, this gap is much starker.

One obstacle to effective enforcement is that pay discrimination is often hidden from sight, and thus affected employees are often unaware of the fact that there may be a violation.  The prevalence of pay secrecy policies, where employers either prohibit employees from talking about pay or actively discourage such discussions, prevents employees from even knowing that there may be a violation and taking further action.

Moreover, because some employers do not systematically evaluate their pay systems, they sometimes themselves are unaware of pay disparities or other potentially unlawful practices within their companies and fail to take corrective action.  Tools to promote robust enforcement and employer self-evaluation are therefore critical.

Access to workforce data is an important tool for uncovering potential pay discrimination.  And for this year ‑‑ for this reason, for over 50 years, we have had the EEO-1 reporting rule in place, and we've all acknowledged the importance, and the EEOC in its notice itself acknowledge the importance of that EEO-1 Component 1 data.

Reporting pay data is critical for identifying potentially unlawful pay gaps between employees, as well as occupational segregation, where certain classes of workers are concentrated in specific industries or in specific jobs or pay bands within a company.  Both of these trends contribute to the overall wage gap, and thus need to be addressed.

Using the long-established EEO-1 categories not only helps simplify the process for employers who are used to reporting in those categories, but it also supports efficient use of resources by the EEOC and other enforcement agencies to enforce our anti-discrimination laws.

The specific jobs within EEO-1 categories are often fairly consistent among employers in a given industry.  Therefore, the data enables the agency to identify and track wage trends and the patterns that exist in those trends, and then flag those employers that deviate from the trend in order to better focus their enforcement.

The data may also expose other forms of discrimination.  For example, if an employer's data shows that women or employees of some other protected class, such as race, are consistently over-represented in lower-paid positions within a company, which then leads to consistent wage gaps, either gender or racial, this could constitute a violation of other laws, such as Title VII.

Finally, the pay data will be useful to the agencies in assessing pay discrimination charges that are filed with them, in the initial assessment of a charge and then to help them direct their investigation and understand what pay information may be necessary to continue.

The specific type of data pursuant to Component 2 is also very critical.  Reporting W-2 earnings is important because it reflects more than just base pay, but other types of compensation, such as bonuses, overtime pay, stock options.  This is critical because pay discrimination often manifests in these alternative compensation sources, so employers, if they were only required to report pay ‑‑ base pay, this type of difference would not be reflected.

Finally, reporting the total hours worked can provide relevant context for the data, such as demonstrating when a pay differential may have resulted from differences in hours worked among employees.

Finally, reporting via the EEO-1 categories enables comparisons of pay data broken down by gender, race, and ethnicity, and thus recognizes the reality that many workers often experience discrimination on multiple bases, such gender and race, or gender and national origin.

Some have questioned the utility of Component 2 data because they are not necessarily indicative of unlawful pay discrimination; however, like Component 1, Component 2 data is not intended to be direct evidence of a legal violation but rather one of many tools to help direct the EEOC and the OFCCP's limited enforcement resources to assess and investigate discrimination charges, and help with further robust enforcement of our discrimination laws.

Importantly, the Component 2 reporting will also encourage employers to proactively review their policies and practices.  We've heard today about some of the good-actor employers who already choose to do that on a voluntary basis; however, many do not.

And having to do this type of reporting will require them to take that look under the hood, determine where there may be potential pay disparities, and while that may not be indicative of a violation of the law, it will then prompt them to look further, determine whether there is a justification, and make rectifications in their pay systems when necessary.

Alternatively, if they then discover there are patterns of people of color or women concentrated in lower-paid positions within their company, patterns they may not have even been aware of, this will then prompt them, hopefully, to assess the practices that may contribute to that, related to hiring, promotion, job evaluation, and other internal practices, and make changes to them to ensure better representation at all wage levels throughout their company.

The EEOC itself affirms that Component 1 pay data has already proven its utility to the EEOC's enforcement of employment discrimination laws.  It is precisely because the EEOC has had a chance to analyze and use component data ‑‑ Component 1 data that it has observed its practical utility to the fulfillment of the EEOC's mission.

It is impossible to accurately assess if and how Component 2 data will be useful for enforcement without having adequate time to analyze the actual data collected.  And in fact, as we all know, the current court-ordered pay data collection is still ongoing, and thus not all the data has even been collected.  It is therefore premature for the EEOC to determine that such data is not useful and/or conduct the benefit versus burden analysis.

I'll just close by saying, the gender wage gap has not changed in a statistically significant way for over 10 years.  It is critical that agencies charged with enforcing our equal pay laws have the tools that they need to take proactive steps to better enforce our anti-discrimination laws.

Component 2 pay data collection is critical to this ‑‑ these actions, and we therefore respectfully request that you renew the Component 2 pay data collection.  Thank you.

CHAIR DHILLON: Thank you Ms. Stender, and to all of our panelists.  We'll now take a 15-minute break and we will resume at 10:35.

(Whereupon, the above-entitled matter went off the record for a recess)

CHAIR DHILLON:  If everyone could take their seats, and we'll resume.  All right.  The hearing will now resume, and we will proceed with our first round of questions from members of the Commission to the panelists.  Each commissioner will have six minutes for questions.  We'll do this in two rounds, and I will get us started.

Mr. Eastman, we've heard some talk and some commentary about how the cost to employers of complying with the Component 2 data collection, including the linking of HRIS systems and payroll systems, are kind of sunk costs, and so the cost to complying with the Component 2 data collection in the future would be significantly lower.  Do you have any thoughts about those observations?  Any reaction to it?

MR. EASTMAN:  Yeah.  I guess I would say, first of all, that the obstacles this time around to reporting were considerable.  It's a lot more complicated than simply linking two systems electronically.  Many employers did not have the time or ability to create software solutions for that problem, and so a lot of this was done manually.  So if there were to be another round of Component 2 reporting, I think you would see a lot of the same burdens exist.  Certainly having experience can help, and there will be some efficiencies and lessons learned, but a lot of the problems remain.

CHAIR DHILLON:  Thank you, Mr. Eastman.  And I guess, Ms. Clements, I have a question about the job ‑‑ the EEO‑1 categories and the job categories.  We've heard some talk about how those categories are kind of consistent from employer to employer and then throughout industries.  In your experience, would you agree that there is that kind of consistency?

MS. CLEMENTS:  Once again, sorry.  All employers ‑‑ I also had to raise my seat so I could see the button.  All employers use the same EEO categories, but that doesn't mean that the jobs within those EEO categories are the same within an industry.  And even within an industry, there's a huge variety of jobs within an EEO category.  You know, the example that was given is really a good one.  For example, at Berkshire, we have IT professionals and HR consultants.  They're in the same EEO category, but they're very different jobs. 

CHAIR DHILLON:  Thank you.  Ms. Stevenson, a question around the hours worked portion of the Component 2 data collection.  I guess I'd like your thoughts on whether you think that's an effective measure and whether you can think of any other kinds of alternatives to the hours worked portion of the survey.

MS. STEVENSON:  So I think that it's necessary to collect some information.  And I think that if you didn't have hours worked, one clear argument would be that pay disparities could reflect differences in hours worked.  We hear that all the time when we hear about the gender pay gap.

I have heard, you know, some discussion to the effect the problem with the, you know, W‑2 data is that it reports the total amount earned, which includes paid time off, and the hours worked would not be counted as hours worked in that paid time off.  But that doesn't seem problematic to me.  The thing you'll have to ask is whether there are real reasons for there to be gender disparities in that that are going to ‑‑ that would be misleading, right?

So for instance, if men were more likely to take paid time off than women, then that might suggest that men are receiving a higher pay per hour worked than women.  And that ‑‑ so the two questions then are: is that true that men take more paid time off than women?  I would suspect that that's not likely to be true, but that's an open question.  And then the second would be to say, well, if that's ‑‑ if that is indeed what we observe, that men are taking more time off than women, we might also want to ask why.  What is it ‑‑ does that actually have something to do with the compensation structure that they're giving more paid time off to men.

So the ‑‑ all of this data raises questions that employers have to ask themselves, and that includes the hours data.  The hours data is not precise, but I think that it really is essential in order ‑‑ if you're going to collect aggregated data, you're going to need to have some indication of the aggregate hours worked.

CHAIR DHILLON:  It raises an issue, though, particularly with respect to exempt employees, correct?  We don't typically track their hours, and so the employers, you know, have the option of making an assumption.  Could you ‑‑ have you thought about that?  And is there a better way to do that?

MS. STEVENSON:  Well we don't want to require that they track exempt workers' hours.  I think that would be a mistake.  So the fact that employers have some flexibility in their assumption seems like it is ‑‑ that's a reasonable workaround.  I think ‑‑ you know, my guess would be a lot of employers default by putting 40 hours in for full‑time workers.

Again, so there you might be concerned that that is a greater underestimate of some male hours than female hours, but you do offer employers the opportunity of actually correcting that on their own.  So I think the best you can do is give ‑‑ is tell employers that they should report what is their best guess of a full‑time employee's hours, and I think it would be too burdensome to ask them to track the hours of exempt workers.

CHAIR DHILLON:  Well thank you.  Commissioner Lipnic?

COMMISSIONER LIPNIC:  Thank you.  And again, thanks to everyone for your testimony.  Ms. Stevenson, I just want to make the point that I appreciate the comment that you made about the Equality Act that they passed in the U.K., which is a different ‑‑ entirely different type of reporting scheme.  You suggested that, perhaps, you know, that's something the EEOC could do.

I just want to state for the record, I don't think we could do that just based on our statutory authority, that it's probably something similar to what happened in the U.K., where it took an act of Parliament for that to be put in place.  So I think that there's some legal question as to whether or not that's within our authority.

And I want to come back to the idea of some alternatives, but first I want to ask Mr. Eastman a question. Just a point of clarification.  You said that you looked at the recommendations of the NAS study and you were saying about a pilot study.  And I just want to be clear.  You're suggesting that, if there were something different than what we have all now undergone, which is the Component 2 collection, right, that if there were some other different thing, that that would need to be subject to a pilot study, right?

MR. EASTMAN:  I think it should be, yes.

COMMISSIONER LIPNIC:  Okay.  Because one of the questions that I have ‑‑ and I want to actually ask both Mr. Mitchell and Ms. Stevenson this, as the economists in the room ‑‑ that, would you agree that, now that we have had the collection of the Component 2 ‑‑ and, of course, you know, there are other steps that have to be taken, right, in terms of the data cleaning and all of that.  But that generally estimates about burden may be less important than what the actual experience has been.  In other words, if we can measure something rather than ‑‑ and I know economists love to make assumptions.  But rather than make assumptions that that would be ‑‑ give us all a better understanding of the actual burden.

MS. STEVENSON:  So I actually suggested in my written testimony that you do look at actual measurements.  And one easy way to do that would be to look at what firms charge to file the EEO‑1 for companies.  And that would be an upper estimate of the burden because no firm would pay ‑‑ no business would pay a firm to file the form for them if they could do it for cheaper.  So some will ‑‑ the ones that do it themselves are clearly doing it for cheaper.  The fees that are charged would give you that upper bound.  So I think that that is essential in providing an accurate estimate of the burden that you collect that information and use it.

COMMISSIONER LIPNIC:  Mr. Mitchell?

MR. MITCHELL:  Yes.  And another thing that you could do in terms of ‑‑ say there's another proposal or tweak or serious change to pay reporting in the future is it's possible you could go, say, to federal government agencies, like the Department of Labor, and ask them to prepare a Component 2 form, even though they're not obligated to do so, and have them track the number of hours it took them to do it.

That would be another way of getting an estimate, I think.  But yeah.  In terms of just what the estimate should be if we were to do the same thing going forward, I think market prices are the best way to get at that.

COMMISSIONER LIPNIC:  So if that's an upper‑bound estimate, what would be lower‑bound estimates?  Again, in terms of actually measurement.

MR. MITCHELL:  Yeah.  In terms of ‑‑ I mean I think the only way to do that is to somehow ‑‑ some observational study.  That's what ‑‑ I would recommend actually having folks, say, at the Department of Labor fill out a form and see how long it takes them.

COMMISSIONER LIPNIC:  What about the people who've now submitted it and the firms that have now done it?  Mr. Eastman, how many EEO‑1s, Component 1 and Component 2, did the CWC help employers with?

MR. EASTMAN:  Well, the CWC itself would not actually file those reports, but we have affiliates that do.  And I don't have that number with me, but it was considerable.  We certainly did help a lot of employers with this.  And one thing I would say, I think it's appropriate to look at the cost, perhaps, as one element.

But remember, the employers do a lot of things before they deliver the data to the consultants, and they also have some responsibilities after it's been delivered.

COMMISSIONER LIPNIC:  And Ms. Clements, how many ‑‑ I know it's in your testimony, but how many employers did your firm help both collect the information and file the information?

MS. CLEMENTS:  We filed reports for about 200 employers, covering 930,000 workers.  I would just want to really re‑emphasize that I do not agree that looking at the price that we charged employers for filing the reports actually accurately captures the true cost of Component 1 or Component 2 data.

You have to include in that the cost of collecting the information to begin with.  Race and gender and job category information is primarily gathered for purposes of the EEO‑1 report.  That is something that an employer does before they send the data to a firm like mine.

You have to also account for the cost for Component 2.  We had employers who ‑‑ we filed the reports for them, but they still spent hundreds of hours collecting and verifying the data before it could be sent to us to simply upload it into the system.  Had they not used our services, they would've actually had to create a data upload themselves.  It took us 300 hours to program our system to create the data upload file that would work with EEOC's system.

So there are a lot costs that are beyond just the cost you pay a vendor, and so I would say the cost of a vendor is the bottom, not the top rate, of what it might cost.

COMMISSIONER LIPNIC:  Okay.  Thank you.  Where did my time go?

CHAIR DHILLON:  Commissioner Burrows?

COMMISSIONER BURROWS:  Thank you.  Let me turn my mic on.  I guess I'll start with Professor Stevenson.  In your written testimony, you talked about the fact that ‑‑ and also today, that collecting some form of pay data is really important if we're ‑‑ and essential if we're going to actually get at the root causes of pay discrimination and job segregation, if we're going to reduce that in our economy.  Could you elaborate on that?

MS. STEVENSON:  Yeah.  I think that it's really impossible to combat implicit discrimination without explicit information.  And so pay data helps employers look for ways in which they may be engaged in implicit discrimination. And I think we have seen many employers who ‑‑ you know, as was noted today by several people -- that voluntarily examine their own microdata to try to come up with plans to reduce their discriminatory practices.  They do this because it's good for their business, as well as the fact that they want to ensure that they're not engaged in illegal discriminatory practices.

But it can be very difficult to see subtle biases that are ‑‑ that involve illegal discrimination.  I give in my written testimony an example of an employer who told me that, when he looked at the data and he saw that female salespeople were earning less, had lower W‑2 income, than the men.  And he went to try to figure out why.  He realized that there was a subtle bias in assigning sales territories that led the men to have slightly more profitably territories than if territories had been assigned randomly.  Why?  Because there are more men in the managerial positions and, you know, they reward their buddies.

So this is exactly what we've seen in lots of places where we've analyzed pay data is that, in particular, one of the things that is a cause of the gender pay gap is that men get outsized pay that they don't fully deserve based on their productivity, based on bias, based on relationship, based on friendships.  When we shine light on that data, that causes that practice to stop.  And as a result, you see an improvement in the gender wage gap that comes somewhat from men not doing as well, but also it helps men in general because it reduces inequality in male pay.

So shining light on the problem causes employers to investigate their data further, to identify plans.  And some employers, as I said, voluntarily do this.  But if they don't, they're probably the ones with the biggest problems.  And that's why the EEOC has a charge to collect data and ensure that all employers are actively pursuing policies that will root out implicit discrimination, which is, I would argue, the primary form of discrimination leading to the gender pay gap right now.  And it has proven itself stubbornly difficult to eradicate and there is no way to do that without data.

COMMISSIONER BURROWS:  Thank you.  I would turn to Jocelyn Frye.  You mentioned earlier that the burden ‑‑ there's some ‑‑ you have some real concerns about the way burden estimate was put together for the PRA notice that's the subject of this hearing.  Could you go into that in a bit more detail?  I think most ‑‑ you know, both ‑‑ I guess it's Component 1 and Component 2, but the approach is there in terms of the concerns.

MS. FRYE:  Sure.  You know, I tried to tick through some of the concerns pretty quickly.  One is simply that we don't know enough from the notice itself about how some of the calculations were made around both the cost, the amount of time that was attributed to actually prepare the forms.  And there seem to be at least ‑‑ there's confusion about sort of the assumptions that seem to underlie the burden estimates that are used -- whether or not the hours should go up versus down, whether or not the large employers that have multiple establishments, whether they will have increased time, or whether or not in fact they have greater capacity, and so the time should be smaller.  Whether or not some ‑‑ you know, some of the forms are assigned based on the number of employees, whether you have 50 or more or 50 or less.  But yet the time allotted for those forms seems to not necessarily align with the number of employees.  So there's confusion about that.

There's confusion about the fact that we actually have history with the EEO‑1 form itself.  And so the assumptions that the burden is significantly higher when we have 50 years of data to actually look at whether or not that burden exists is confusing I think to many of us.  And then the other, you know, issue around technology.  I won't, you know, repeat all of it.  But, you know, suffice to say, you know, there clearly is technology out there.  There are a myriad of companies that will continue to be companies to assist with reducing the burden.  The notice alludes to that, but doesn't try to actually document it in some way.  The 2016 notice actually tried to account for it in terms of the cost that it measured.  So some ‑‑ those are some of the concerns along with the larger issue of just looking at the burden on employees themselves.

COMMISSIONER BURROWS:  Thank you very much.

CHAIR DHILLON:  All right.  Thank you.  We'll now go to our second round of questions, and I'll ‑‑ I will lead it off.  Ms. Clements, have ‑‑ now that we are, you know ‑‑ a number of employers have submitted Component 2 data, so they now have that data.  In your experience with your clients, have any of your clients now used this Component 2 data that they've assembled for purposes of conducting a salary/equity analysis within their own establishment?

MS. CLEMENTS:  No.  We have not had a client that we are aware of do a more robust salary/equity analysis based solely on the filing of the Component 2 data.  I think the collection and filing process kind of convinced those employers that this data wasn't useful to understanding the way their pay system actually worked.  And so even if they were going to do an analysis, it would not be with W‑2 wage information or the hours worked information based on the assumptions that were made to file the Component 2 report.

CHAIR DHILLON:  Thank you.  Another question for Ms. Clements.  I know that your firm has clients all over the country and you conduct salary/equity analyses for those clients.  Have you seen an increase in the amount of work of your clients who are seeking to do an internal salary/equity analysis over the past few years?

MS. CLEMENTS:  I think that there is an interest in pay/equity issues in general.  I cannot attribute any of that to this particular data collection.  And as I mentioned in my testimony, many of our clients felt like this data collection took away from other legitimate interests in pay/equity initiatives.  You know, employers have limited time and resources, and so the filing of this report took away time and resources for other more robust analyses that could have been conducted, and that was what we heard from our clients time and time again.

CHAIR DHILLON:  Thank you.  Ms. Frye, I'd be interested in your thoughts around the use of the Box 1 W‑2 measurement as the measurement that was used in Component 2, and whether there might be another form of measurement that would be more effective for use in a pay data collection.

MS. FRYE:  Well, I think, as Ms. Stender said earlier, you know, the value of W‑2 data is that it looks at pay in sort of a broader way.  And the concern of using something that is less rigorous and less comprehensive is that you can miss things.  You can miss things like bonuses and other ways that compensation occurs.

So are there alternative measures?  Maybe.  I'm not familiar with every potential form that could be out there.  But I think the broader point is that if you're really going to measure compensation and look at parity, then you need a measure that actually tries to capture it as best as possible.  And that's why, again, it's a way of doing so.  But you know, certainly could there be other ways to do so that would account for all the different ways that people are paid in their workplace?  Potentially.  That just seems to be a tool that is one that should be familiar to employers.  It's data that they already have and would, you know, be able to capture a lot of the information that we think is important to look at.

CHAIR DHILLON:  Mr. Eastman, based on your experience with your members, what role did the certification of the data play in their efforts, and to what extent do you think, if at all, did the certification add to the burden to employers as they were assembling and analyzing their Component 2 data?

MR. EASTMAN:  Sure.  So certification, first of all, the burden that it establishes may vary depending on what you think you're required to do by certification.  The Commission I think tried to provide guidance that said, look, you're not certifying that every single data point is 100 percent accurate.  We understand this is the first time through this, and you're certifying that you complied with the instructions.  That in itself is ‑‑ does provide additional burdens because you want to be sure that you've done everything right.  If you were to certify that every data point had to be accurate, that would impose significantly more burdens.

CHAIR DHILLON:  Ms. Stender, I would be interested in your thoughts or reactions to some other testimony and other things we've heard about the burden of employers having to take multiple steps to manually reconcile the data.

MS. STENDER:  I am not an employer advisor, and so I don't have direct experience with employers and their experience of how they report their pay data.  But I am familiar with the fact and the statistics that have been mentioned in some of the testimony today and in some of the amicus briefs that were submitted in the case around pay data that's currently in the district court that do cite to the pervasive use of technology.  And so as a non‑data‑specific economist, I would say that, you know, the use of the technology I know has reduced the burden, and again would help people going forward to ‑‑ once they've learned how to use it and learned how report this new pay data would, in my view, lessen that burden on them going forward.

CHAIR DHILLON:  All right.  Thank you.  Commissioner Lipnic?

COMMISSIONER LIPNIC:  Thank you.  Mr. Mitchell, I want to come back to the issue about the false‑positives, false‑negatives, which you talked about in your testimony, and there's been some other testimony about.  Can you elaborate on that in terms of the problems that you see with this data, the Component 2 data that's been collected, and to the extent that you have comments related to the use of that from an enforcement perspective? 

MR. MITCHELL:  Sure.  I think that the primary goal has to be, if you're collecting this data, is it's going to better tailor your resources to targeting those who are in violation of the law.  And so a natural way to look at how well you do that is ‑‑ the gold standard I think would be the Abt Associates analysis, where they actually have the completed investigations on the one hand, and then they back up and said, what if we didn't know what the result was?  Could we have predicted ahead of time who was in violation and who was not?

And so you have on the one hand these sort of false‑positives where you're flagging employers who are nevertheless fully compliant.  And you also have false‑negatives where employers who are in violation are not being detected.  And what we found in our simulation study, and it's consistent with the Abt study, is that these error rates are just so large that you could get the same sort of benefit of deterrence by just auditing companies at random without having anyone submit anything.

COMMISSIONER LIPNIC:  And if we were to look at the information that we've ‑‑ that's been collected, right, and study that, is there ‑‑ what would be like a sample size that would be necessary to, you know, give us some good information about ‑‑

MR. MITCHELL:  Yeah.  So ‑‑

COMMISSIONER LIPNIC:  Obviously there's all of it, right?  But what ‑‑

MR. MITCHELL:  Well, I mean ‑‑ so I mean in some ways what you've done now is you've created the ultimate pilot study, right?  For the last two years.  So you have the ‑‑

COMMISSIONER LIPNIC:  I like to call it "the actual" --

MR. MITCHELL:  Yeah.  The actual.

COMMISSIONER LIPNIC:  ‑‑ at this point.

MR. MITCHELL:  The actual.  So at least for the contractors, you have some of their underlying data from OFCCP audits.  You could link up this data with the Component 2 data and see how well you do.  Sample size, well I mean it really depends on ‑‑ you have to ‑‑ I couldn't give you a number right now, but I'm happy to work with the Commission to try to figure out the power you need to detect discrimination at various levels.

COMMISSIONER LIPNIC:  I want to come back to also the issue about the sunk costs ‑‑ or potentially sunk costs.  And so Ms. Clements, I'll direct this question to you and actually to Mr. Eastman.

Again, in terms of your experience with people who now, you know, went through the collection, collecting the information and then reporting it, and then the potential of, you know, a linkage between the HRIS system and the payroll system, is it your experience that most of the people who have now ‑‑ and your --- with your employers who have now filed actually have those sunk costs or, you know, where they, here's what we've got to do to get this filed.  And if we're going to do some ‑‑ make technological changes in the future, that's going to have to wait.

MS. CLEMENTS:  Most of our employers did not have the time or opportunity given the circumstances under which these reports were filed to create those linkages.  And what we learned, which was even a little bit surprising to us, is that even if those linkages are created, there are additional challenges that will be present every year; the impact of mergers and acquisitions. Something as simple as an employee converting from a non‑exempt to an exempt employee during the 12‑month snapshot that is now being used for Component 2 requires manual calculation.  There's no linkage that allows you to determine what that person's hours worked were when part of their time was spent as a non‑exempt employee and part of their time was spent as an exempt employee ‑‑ or it's very hard to create that linkage.

And the other thing we found is that lots of large employers have very ‑‑ a lot of different systems that would need to be linked.   Most of that work was not done for purposes of filing this ‑‑ the 2017 and '18 reports, and would have to be done in the future.  But even doing that work doesn't mean that you can push a button and out spits your ‑‑ you know, we had an employer whose Component 2 reports were 12,500 pages.  There is no linkage system that is going to push a button and spit out 12,500 pages of accurate, certifiable data.

COMMISSIONER LIPNIC:  Mr. Eastman?

MR. EASTMAN:  I agree.  I would just add one example.  I talked to a member of ours yesterday.  This is a company of about 20,000 employees.  They had to go through the process of manually matching up the demographic information with the W‑2 information in order to be sure people got in the right box, and that's not something that you're going to get an efficiency from next year.

COMMISSIONER LIPNIC:  I'm going to run out of time on what I want to ask next, so I'll yield back the remainder of my time.

CHAIR DHILLON:  Thank you.  Commissioner Burrows?

COMMISSIONER BURROWS:  Yes.  There's been ‑‑ and I guess I would ask Professor Stevenson.  There's been some discussion about this nineteen, 2005 study, which was a very different pay data collection by Abt Associates for OFCCP.  And I think that one of the things that I'm trying to puzzle through here is it's somehow being used, this different study, as a argument, I think ‑‑ or, it at least raises questions about whether or not Component 2 will assist us with, you know, our current enforcement authorities.  Could you speak to that and give your perspective on that?

MS. STEVENSON:  Yeah.  So I was aware of that study and was made further aware of it by reading the National Academies ‑‑ the 2012 National Academies report quite thoroughly.  And they argue there that you cannot rely on that study or that data collection effort to inform ‑‑ to get further information about pay data collection.

And one of the things they argue in the National Academies report is that there was evidence that the data collection was hampered by the fact that employers were aware ahead of time of what was going to be collected and how it was going to be used.  And as a result, many employers had a chance to make sure that the data was not going to yield any information that would suggest that they were engaged in any kind of illegal discriminatory practice.

So in fact, the way to think about it is that, if employers, like, are ‑‑ if employers understand that this is part of a enforcement action, then they can solve problems ahead of time.  That's actually what we want them to do, so that's not a bad thing.  But then that does reduce the ability of the data itself to actually predict problems if employers see the data and say, whoa, I don't want to submit something like that.  I'm going to take steps to reconcile it.

So I would say that the fact that the National Academies study, which is ‑‑ you know, it's been referred to by many people on this panel with many, you know, different views.  It really was a gold standard look at what our efforts should be.  And the fact that they sort of dismiss that as not being informative I think should inform the EEOC.

COMMISSIONER BURROWS:  So in that situation, those folks who turned out not to be false‑positives were actual positives, or the people who found out that they're ‑‑ through the OFCCP's back and forth that they had a problem, and then still didn't fix it.  Not necessarily everybody who had a problem when OFCCP started there, you know ‑‑ or, at some point during the compliance process.

MS. STEVENSON:  That's my understanding, yes.

COMMISSIONER BURROWS:  So I wanted to ask, for those of you ‑‑ just so that I, you know, walk away here with a clear understanding.  For those of you who have made the point here today and in your written testimony that there's an undue burden, shall we say, for the Component 2 in terms of the burden analysis, is there ‑‑ is your position that there's no level of burden that would be acceptable, or is there some level, it's just this isn't it?

I was struggling a little bit because we're ‑‑ we have the task of not just looking at this but trying to figure out, going forward, well what should we do?  And in terms of figuring that out, I'd love your help in understanding that.  And if you're ‑‑ compound question.  Do you have any proposal if you're opposed to it to how we would fix it?  Thank you.  We can start with Mr. Eastman or whoever wants to start.

MR. EASTMAN:  Sure.  That's a hard question to quantify an answer to.  I can tell you if we felt that the burden of the data collection would have utility, then that would be an easier question to answer.  Certainly we think Component 1 is burdensome, but it has utility and we support it.  But it's hard to give you a number for --

COMMISSIONER BURROWS:  Do you have suggestions for the Commission as to what would ‑‑ what a useful data collection would look like, or is your position that there is no such ‑‑ there is nothing that is reasonably burdensome but would also be useful?

MR. EASTMAN:  Right.  So in our testimony, we raised concerns that it may not be possible to create one, but we are certainly willing to engage.  We offered suggestions in 2016 that would've halved the burden, for example, but using pay rate instead of the W‑2 wages.

COMMISSIONER BURROWS:  Mr. Mitchell?

MR. MITCHELL:  My response is a bit similar.  Because I didn't find much of expected utility from the form, sort of, any additional burden is hard to justify; however, if you were going to do some low‑hanging fruit, I would again say that, yeah, using a pay rate would only ‑‑ would not only be more accurate but would also substantially reduce the burden for those companies that have their W‑2s in a completely different system, or even an outside third‑party vendor.  So just switching to a pay rate would be low‑hanging fruit.

COMMISSIONER BURROWS:  By pay rate, you're saying, base pay.

MR. MITCHELL:  Base pay or, you know ‑‑ and you could have some measure of expected, say, commissions if they are sales workers.  The kinds of things that are maintained in a normal HRIS environment.

COMMISSIONER BURROWS:  Your position is not that there's nothing we could do here in terms of data ‑‑

MR. MITCHELL:  There are certainly things you could do to reduce the burden, but again I'm still skeptical that, you know, given that I haven't found utility, that any change on its own is going to affect the overall calculus of benefits and cost.

COMMISSIONER BURROWS:  And lastly for Ms. Clements, any ‑‑ is this an impossible task?  Should we ‑‑ or is that your position that there's nothing that we could do in terms of any data collection?  Or is it just, you know ‑‑ do you have proposals for how we could do that?

MS. CLEMENTS:  I would agree with the comments that have been echoed by my colleagues.  I will say that I have personal experience having been at OFCCP during the time period that some of this pay data was collected.  And I do believe that the experiences of the agency there are instructive about the limitations of a uniform pay data collection tool and the utility of that kind of pay data collection tool.  And it's part of the reason why I believe OFCCP now seeks microdata as part of its enforcement efforts.

And so I think it is a challenging task to come up with a uniform tool that actually will allow you to evaluate contractor pay systems because they're unique.  Do I think it's impossible?  No.  And I welcome the opportunity to continue dialogue with you on that.

COMMISSIONER BURROWS:  Thank you.

CHAIR DHILLON:  At this point, we have completed our questionings, and so each commissioner will have the opportunity to make a closing statement of three minutes or less.  I will get us started.

I want to again sincerely thank all of our panelists who have traveled here today to share your insights and recommendations and respond to our questions.  It's been very valuable and very helpful to us, and I appreciate the level of effort that you obviously expended to provide us with your points of view.  Obviously we will take into consideration both your oral testimony and your written testimony, as well as the comments we receive in response to the 60‑day notice as we consider our next steps.  And with that, I will turn it over to Commissioner Lipnic.

COMMISSIONER LIPNIC:  Thank you.  I again want to thank Chair Dhillon and her staff for putting this hearing together, Commissioner Burrows, and especially our witnesses for your ‑‑ offering us your expertise today.

I also want to acknowledge how much even since 2016, when the EEOC approved the Component 2 and the revisions to the EEO‑1 form, how much the discussion and the landscape, in terms of the public policy world, has changed about equal pay.  And I want to give credit where credit is due, and that really is to the work of the advocacy groups, among those sitting on our panel, who have made this a priority issue, and who have gotten it into the public consciousness, I dare say, in ways that even 20 years ago probably were not the case.  So, and that is certainly a good thing.

And for everyone who is interested and concerned in pay discrimination and equal pay, even five years ago, pay equity was not a term that we were using quite as often as we use today.  I will say though also for those of us who are the public policy makers and in the business of having to expend government resources, enforcement resources are scarce.  They are as scarce for employers, but they are certainly scarce for the Government.  And any enforcement agency ‑‑ and I say this as someone who's been the head of two major enforcement agencies ‑‑ has to think very carefully about what the best use of those enforcement resources are.

I hope that ‑‑ I certainly hope that, going forward, we can come up with some analysis of what we have all now gone through in terms of the Component 2 collection ‑‑ and I do want to add that I appreciate the comments about Component 1 also, both in your written testimony and your oral testimony today.  I think that, you know, an ability to study this and have some common understanding of both the burden and the utility is something that would be very useful to all of us.

And that we ‑‑ again, as I said in my opening, we're past the world of speculation at this point.  We actually are now in the actual.  And so I hope that all of our decisions and all of our concerns about equal pay will be informed by what we now have the ability to study.  And with that, Madam Chair, thank you for this hearing.

CHAIR DHILLON:  Thank you.  Commissioner Burrows?

COMMISSIONER BURROWS:  Yes.  Thank you.  In recent years, EEOC has obtained significant relief for class of female law professors systematically underpaid compared to their male counterparts, African American maintenance worker paid less and given lower raises than his white counterparts, and Chinese auto technicians ‑‑ this always stuns me ‑‑ who are paid just randomly three hours less per ‑‑ $3 less per hour than non‑Chinese workers, among many others.

But we have ‑‑ while resolving these very important cases, including I would note our $1.2 million settlement earlier this month involving African American oil field workers denied higher‑paying assignments, which was more a job segregation kind of case, our work has still been hampered by the lack of information.

We often discover these pay violations by accident in a needle‑in‑a‑haystack search, or  when an employee stumbles on information by happenstance.  Lilly Ledbetter, for whom the Lilly Ledbetter Fair Pay Act was named, only learned she was being paid thousands of dollars less for almost two decades because someone left her an anonymous note because civil rights enforcement is far too important to be left to chance.

In 2016, this Commission adopted Component 2 of the EEO‑1 survey to collect pay data from employers by race, national origin, and sex.  In 2016, we concluded that, combined with a wealth of other information that we received ‑‑ not alone, as has been discussed here today, but together ‑‑ that would be a useful tool for us.

And we stated persistent pay gaps continue to exist in the U.S. workforce correlated with race, sex, and ethnicity.  Workplace discrimination is an important contributing factor to those disparities, and implementing the proposed EEO‑1 pay data collection will improve the ability to efficiently and effectively restructure its investigation of pay discrimination charges.

So we are required to evaluate the benefits and the burdens, again, in light of all the evidence, including that we are currently collecting today's testimony, for which I thank each of our witnesses, and a fair assessment of the evidence that once this ongoing pay data collection is completed.

I would note that the hearing that we had in 2016, we started about the same time, we ended around 4:30.  And I unfortunately have not had ‑‑ you know, I apologize to those witnesses who I didn't get a chance to ask much of because I think that debate would've been helpful.  But I know that we will continue to wrestle with that.  I have a very thick set of questions for each of you, and hopefully I will be able to puzzle through things on my own.

I do look forward to working with my colleagues on this very important issue, and I once again extend my sincere thanks to each of you for the time that you put in.  It clearly shows and it's appreciated.  And also, of course, to the Chair, to Commissioner Lipnic, her ‑‑ their staffs, Exec Sec, and to all who helped make this such a success today.  Thank you.

CHAIR DHILLON:  Well thank you, Commissioner Lipnic and Commissioner Burrows for those remarks.  And again, thanks to our panelists for your participation in today's hearing as well as to all of you who attended or listened in.  And a special thanks to all of our staff who worked so hard to bring this hearing together.  And with that, our hearing is concluded.  Thank you.

(Whereupon, the above‑entitled matter went off the record.)