On February 14, 2010, Mercury News reported that Google, Apple, Yahoo, Oracle and Applied Materials had successfully convinced the U.S. Department of Labor (DOL) not to release race and gender data from their EEO-1 report under the Freedom of Information Act (FOIA) because releasing such data would cause “commercial harm.” CNN made a similar FOIA request on August 18, 2011 for EEO-1 data on 17 technology companies that refused a direct request from CNN for the data. Not all tech companies resisted the requests. Mercury News was able to obtain race and gender data directly from Intel, Cisco Systems, eBay, AMD, Sanmina and Sun Microsystems and CNN was able to obtain data from Dell, Ingram Micro and Intel without FOIA objections being raised. On May 29, 2014, Google dramatically reversed its position and published information about the racial and gender composition of its workforce from its EEO-1. Google states on its website, “We’re not where we want to be when it comes to diversity. And it is hard to address these kinds of challenges if you’re not prepared to discuss them openly, and with the facts.”
The purpose of this article is to explain what an EEO-1 report is, who has to file it, when it has to be filed and what confidentiality protections the government is required to observe with respect to the data provided in the report. The article will also explain what, if any conclusions can be drawn from an EEO-1 about whether discrimination exists in a company as well as the pros, cons and caveats of voluntary EEO-1 disclosure.
Background on EEO-1 Reports
Under Department of Labor regulation 41 CFR 60-1.7, federal contractors or subcontractors with 50 employees and $50,000 in federal contracts, subcontracts or purchase orders and certain financial institutions are required to annually submit an Employer Information Report (EEO-1). The EEO-1 report profiles the race, gender and ethnicity of the companies’ incumbent employees by job group. OFCCP uses this data in the construction of its selection system. OFCCP obtains EEO-1 data from the EEOC. Using different formulas, OFCCP crunches this data to come up with the list of companies to be reviewed for compliance in a given period of time. EEO-1 data is also used from time to time for studies commissioned by the OFCCP such as when the agency was developing the EEO Survey a number of years ago. In the course of compliance evaluations, the agency pulls the last three EEO-1 submissions to determine if there are any trends worth noting. For example, it will look to see if the numbers of minorities and women are increasing or decreasing relative to non-minorities and men.
The Equal Employment Opportunity Commission (EEOC) also requires all employers with 100 or more employees to file the report. The reports generally must be submitted on or before September 30 each year. EEOC is prohibited by law from sharing individualized company information from these reports. An EEOC regulation at 29 CFR 1610.17(e) states, “Each executed statistical reporting form required under part 1602 of this chapter, such as Employer Information Report EEO-1, etc., relating to a particular employer is exempt from disclosure to the public prior to the institution of a proceeding under title VII involving information from such form”. In other words, the EEO-1 of a particular company cannot be released by the EEOC in response to a FOIA request. The DOL is not subject to similar constraints with respect to the EEO-1 data from federal contractors. A FOIA request to the DOL is handled the same as any other FOIA request which means that the DOL will inform the company in possession of the requested information that DOL has received a request. The company then has an opportunity to raise any objections to the release. If DOL agrees that the company’s objections fall within one of the exemptions from the obligation to release the data, DOL can refuse the FOIA request. In these cases where DOL refused to release the EEO-1, the basis for the refusal was DOL’s acceptance of the argument that release of the data would cause “commercial harm.”
The Department of Justice 2004 FOIA Guide provides the following definition of the term “commercial” in the FOIA context,
While not without limitations, the term “commercial” is clearly broader than might be expected by the average lay person. DOL did require that the companies submit documentation to support their claim of commercial harm.
The Google Data Release
The following charts shared on the Google.com website summarize the demographics of the Google workforce by race and gender. Just under a third of the workforce is female. The largest disparities in female employment are in Tech and Leadership categories. The ethnicity figures show that the workforce is predominantly white and that among minorities, Blacks and Hispanics are less prevalent than Asians.
These data have spawned a number of headlines, “Google Employs Fewer Women Than Dozens of Its S&P 500 Peers” (www.businessweek.com, Joshua Brustein, May 29, 2014); “2% of Google Employees Are Black And Just 30% Are Women” (www.forbes.com, Jay McGregor, May 29, 2014); Congress Causes Google’s Diversity Problem” (www.bloombergreview.com, Leonid Bershidsky, May 29, 2014); “Google is Actually More Diverse Than You Think” (www.entrepreneur.com, Laura Entis, May 30, 2014); “Google Has a White Male Problem—Here’s What It Needs to Do to Fix It” (www.Takepart.com, Britni Danielle, May 30, 2014 (as reported by Yahoo News)).
What strikes me immediately is how quickly the authors jump to conclusions about the meaning of the Google data. As noted above, Google’s EEO-1 may be news to the media and the general public but it is not news to the OFCCP. Google and other federal contractors regularly submit this data to the government. The government does not take this data to mean that Google or any other company with a similar profile has engaged in discrimination. OFCCP uses this data in developing its lists of contractors to be scheduled for a compliance evaluation. Once a compliance evaluation has been scheduled, compliance officers pull the EEO-1 data early in the evaluation process to get a general sense of the composition of the company. Compliance officers do not base their discrimination findings solely on EEO-1 data. The “Google Has a White Male Problem…” article states, “The real elephant in the room? The racism and sexism that job applicants encounter, coupled with sheer laziness on the part of hiring managers.” This is quite a conclusion to draw from EEO-1 data. No other support for this conclusion is offered in the piece. OFCCP would not draw this conclusion from a mere review of EEO-1 data.
The article entitled, “Google Employs Fewer Women Than Dozens of Its S&P 500 Peers” is also interesting. The title invites the reader to conclude that all of the S&P 500 are properly to be compared to one another and suggests that Google is a particularly bad actor. The S&P 500 is comprised of companies in a wide varieties of businesses. They would not be considered proper comparators in a compliance evaluation. If a compliance officer were examining how well Google is doing relative to other companies the officer would look at other high tech companies that compete directly with Google, not at, for instance, Tiffany & Company (fine jewelers) or Ralph Lauren (high fashion). The headline on the other hand does not bother with these kinds of distinctions. The article entitled “Google is More Diverse Than You Think” points out that the national workforce is 66% white which is higher than Google’s 61% white workforce. As with the comparison to the S&P 500, the national workforce is really not a good comparator to Google since it includes a variety of non-tech, unrelated enterprises. Finally, the article entitled “Congress Causes Google’s Diversity Problem” focuses on the challenges of getting HB1 visas to bring in more foreign-born minorities into tech positions, OFCCP would not reach this conclusion based on EEO-1 data. All of these articles are in immediate response to Google’s release of the EEO-1 data which gives you some idea of challenges that must be overcome in releasing such data.
Pros and Cons of Voluntarily Releasing EEO-1 Data
As noted at the outset, a number of companies did not resist requests by journalists for their EEO-1 data. At least one of the companies that did not resist was reported to regularly publish this data on their own web page. One thing these companies avoided by releasing this data voluntarily was the media spotlight that was trained on the companies that refused the request. Google’s EEO-1 data would not likely have received the level of attention it received if Google did not initially resist disclosure and then reverse its position. Articles such as, “Jesse Jackson Tells Google It Should Publish Hiring Data and Google Agrees” (www.northstarnews.com, Rosemary Eng, May 15, 2014) which state, “David Drummond, an African American who is senior vice president of corporate development, and chief legal officer of Google, the world’s most-popular search engine, agreed, saying, ‘we are wrong for not disclosing the data. We will do it next month.'” This paints a different picture of Google than would have emerged from a voluntary disclosure.
At this point, pre-Google voluntary disclosure paints the disclosing companies as forward thinking companies with nothing to hide who were seriously working on their diversity challenges. Google’s disclosure paints it as a troubled company that is just now seeing the light and who will be watched closely to see if their newfound understanding actually leads to a more diverse workforce. Post-Google disclosures will likely be viewed as companies succumbing to the pressure created by the Google disclosure to make individual company EEO-1 data public. Post-Google objections to disclosure of individual EEO-1 could possibly create the impression that the company has something to hide. An increase in public pressure to disclose would not be surprising.
As is clear from the immediate responses to the Google disclosure another challenge of voluntary disclosure is education of the media and the public. Companies have to educate the public that a racial and gender imbalance in the workforce is not the same as discrimination in hiring. While the focus is understandably on what companies can do to increase the diversity of their workforces, we have to be careful not to lose sight of the fact that the law requires equality of opportunity not equality of results. Affirmative action requires every good faith effort not a quota of hires based on race, gender or ethnicity.
Disclosing companies will also have to educate the public on how challenging it is to eliminate this lack of diversity. A number of the cited articles assume that the answer to the demographics at Google is readily apparent and can be immediately implemented with results in a fairly short period of time. Some of the suggestions offered included broadening the focus in recruitment beyond computer science graduates, working with educators to introduce the idea of working in high tech to school age children or funding programs that would introduce computer science at lower grade levels, expanding the capacity to increase minority representation through foreign recruitment, providing support programs for minorities and women to reduce the impact of working in a non-diverse workforce and the like. It is worth noting that none of the articles offering these suggestions pointed to actual situations where these approaches made a meaningful difference in the demographic composition of a tech workforce. The articles also did not present evidence that the companies who made voluntary disclosures have actually solved their diversity challenges.
The upside of voluntary disclosure is that it presents the company as being aware of and willing to publicly work on increasing the diversity of its workforce. Disclosure also makes the company somewhat more accountable to the public and to shareholders. On the other hand, the challenge of voluntary disclosure is the risk that the company will be perceived as discriminating and find itself under pressure to hire based on gender, race and ethnicity to handle the public relations challenge presented by its profile. To the extent that such voluntary disclosure of EEO-1 data by companies becomes the new norm, the impact of such disclosures both positive and negative is likely to be less pronounced.