On July 21, 2014, President Obama issued an amendment to Executive Order 11246 adding “sexual orientation” and “gender identity” to the list of protected bases covered by that order. The Department of Labor was directed to prepare regulations within 90 days to implement the amendments.
The Department notes, in a fact sheet on the new amendments, that many contractors already have voluntary, internal prohibitions against discrimination on the basis of sexual orientation and gender identity. The fact sheet states that,
- Many Federal Contractors Already Have Policies on LGBT Workplace Equality. Of the largest 50 federal contractors, which represent nearly half of all federal contracting dollars, 86% prohibit sexual orientation discrimination and 61% prohibit discrimination based on gender identity. In addition, the five top federal contractors, which receive nearly a quarter of all federal contracting dollars, already bar discrimination based on both sexual orientation and gender identity.
One might get the impression from this, and other statements concerning the expansion of coverage, that the implementation of the new requirements is no big deal. One might assume that the new expansion is simply codifying what is already common practice and that it will be easy to comply. One might even wonder if the new expansion of coverage is more symbolic than anything since voluntary programs are already in place to prevent most federal contracting dollars from going toward discrimination on the basis of sexual orientation and gender identity.
It is often said that the devil is in the details. It would be a mistake to under estimate the potential impact of the new expansion of Executive Order 11246 coverage. This article will discuss how trying to fit the new categories into the current Executive Order 11246 enforcement model is like trying to fit a square peg in a round hole. The entire Executive Order model for both non-discrimination and affirmative action enforcement is built on a statistics driven foundation. This statistical foundation is virtually non-existent in the areas of “sexual orientation” and “gender identity” for the applicant pools, incumbent workforce and the qualified available workforce in the reasonable recruitment area. In other words, the basic data sets that drive Executive Order enforcement are not there.
The method used to amend the Executive Order itself was simply to insert the words “sexual orientation” and “gender identity” into the list of covered bases. This will not work as a method for amending the Executive Order regulations. If such a method is used to revise the Executive Order regulations, federal contractors will find themselves needlessly and expensively retrofitting systems to include categories that are not meaningful in actually addressing the problem of discrimination on the basis of “sexual orientation” and “gender identity.” Executive Order style enforcement is the wrong model for addressing the discrimination problems in this area.
Internal Data Collection
As typically enforced by OFCCP, both non-discrimination and affirmative action Executive Order enforcement are heavily reliant on statistical analysis. The first issue then is how do we populate these two new data categories. In order for enforcement statistics to be meaningful, the contractor will need data on both the preferred group and the disfavored group.
Under 41 CFR 60-1.12(c) contractors are required to be able to identify the gender, race and ethnicity of each employee and, where possible, each applicant including Internet Applicants. Applying a similar requirement to sexual orientation, for example, would require categorization of the federal contractor’s entire workforce and its applicant pools by whether they are homosexual, heterosexual or any other category present in the applicant or incumbent pools that could be reasonably read to constitute a sexual orientation.
The typical way contractors determine applicant and incumbency demographics under the Executive Order is through voluntary invitations to self-identify. Contractors are expected to defer to the individual’s self-identification and not challenge it,
- OFCCP’s policy is that deference should be given to an individual’s self-identification and it should not be questioned or overridden by an employer based on the employer’s visual observation. (Frequently Asked Questions section of the OFCCP web page, hereafter OFCCP, FAQs).
This principle would most likely be applied identically to declinations to self-identify sexual orientation.
What has not been an issue in voluntary invitations to self-identify race, gender, and ethnicity is the issue of confidentiality. OFCCP requires confidentiality in the disability context, but does not have a similar requirement in the Executive Order context. However, some individuals may not be comfortable self-identifying their sexual orientation without some guarantee of confidentiality. Since self-identification is voluntary, the absence of such protections may lead to more applicants and employees opting out self-identification.
It is not clear that principles such as confidentiality of self-identification information can simply be imported from the Rehabilitation Act into the Executive Order regulatory amendments. The language of the new amendments appears to simply extend Executive Order 11246 provisions to new categories. On its face, it does not appear to authorize the importation of protections or processes that are atypical of the Executive Order.
Aside from the potential confidentiality concerns, applicants and employees may opt not to identify their sexual orientation because it is none of the employer’s business. If a substantial number of individuals, both among applicants and incumbents, do not self-identify it will impact the contractors’ ability, and obligation, to carry out traditional affirmative action and non-discrimination analysis under the Executive Order.
Where applicants decline to self-identify race, gender and ethnicity, OFCCP provides the following advice,
- If self-identification is not feasible, post-employment records or visual observation may be used to obtain this information. Contractors should not guess or assume the gender, race or ethnicity of an applicant or employee. (OFCCP FAQs)
Post employment records and visual observation are not likely to be of much use in determining an individual’s sexual orientation. Where all of the described methods of identification fail contractors will presumably not be allowed to guess or assume the sexual orientation of an applicant or employee any more than they are allowed to guess or assume membership in the other categories. Unknowns are typically removed at this stage of the analysis by specific race or gender. Indeed, OFCCP states,
- OFCCP would not hold a contractor responsible for applicant data when the applicant declines to self-identify and there are no other acceptable methods of obtaining this information.
Depending on the volume of unknowns, the absence of applicant and employee data on sexual orientation could significantly impact the statistical underpinnings typical of Executive Order analyses.
The invitation to self-identify “gender identity” presents a different set of problems. The motive for the addition of “gender identity” to the Executive Order was to address discrimination on the basis of transgender status. Unfortunately, it does not actually use the term “transgender status.” An invitation to self-identify “gender identity” will be indistinguishable from the current invitation to self-identify by gender. If the purpose of an invitation to self-identify “gender identity” is to isolate for analysis those who are transgendered and those who are not, simply asking for your gender a second time provides you no information to go on for purposes of analysis. Likewise post employment and visual observation will, at best, simply assign a gender but will not address any issue related to transgender discrimination since the records will not show who is, or is not, transgendered. Even if OFCCP required an invitation to self-identify as transgendered, the issues of confidentiality discussed earlier, as well as the small numbers of transgendered applicants or employees you are likely to have, is insufficient to feed the statistics based Executive Order enforcement model.
External Data Collection
In addition to the internal data collection challenges, contractors will face external data collection challenges. To determine whether an affirmative action goal should be established, contractors are directed to compare their incumbent workforce demographics by job group with the demographics of the available qualified workforce in the reasonable recruitment area. If there is underutilization, they are expected to set affirmative action goals and to implement action oriented programs to attain the expected utilization.
U.S. Census and other data are typically used to conduct this analysis. It is not at all clear where federal contractors are expected to obtain the data to conduct a similar analysis with respect to sexual orientation. As for gender identity, even if OFCCP requires an invitation to identify as “transgendered” it would be hard pressed to find appropriate employment statistics that could be used to support the establishment of affirmative action goals in the manner contemplated by the Executive Order.
The Executive Order is not the best enforcement model for the new categories of “sexual orientation” and “gender identity.” The potential data gap has significant implications for an Executive Order style affirmative action program and for typical statistics driven discrimination analysis.
The entire affirmative action model under the Executive Order is based on the presence of meaningful statistics. The Executive Order regulations 41 CFR 60-2.10 describes the required contents of an affirmative action program. These include an organizational profile, job group analysis, placement of incumbents in job groups, determining availability, comparing incumbency to availability, placement goals, identification of problem areas, development of action oriented programs, etc. These elements all assume the presence of a meaningful volume of statistical data. There is added cost to retrofitting computer systems and internal processes to include information on the two new categories, there are costs even to include columns for new categories. Unless there is a reasonable hope that going to this added expense will actually impact the issue of LGBT equality, this is a waste of time and money.
Similar problems arise in discrimination analysis. Most Executive Order discrimination cases are pattern and practice cases or disparate impact cases, both of which require statistical significance. The main types of discrimination cases that may be suited to Executive Order style enforcement for these new categories will be either cases against a discriminatory policy, such as different or lesser benefits for same-sex married couples than for heterosexual married couples, or not allowing a person access to a restroom consistent with their gender identity, or one-on-one disparate treatment with evidence of intent. None of these depends on the collection of data required under the current Executive Order regulations.
The Executive Order amending Executive Order 11246 requires the Department of Labor to prepare implementing regulations in 90 days from the date of the Order and will apply to contracts entered into on or after the effective date of those regulations. It should be noted that a distinction can be drawn between an instruction to prepare regulations within 90 days and an instruction to implement regulations within 90 days. Compare this language to the language used concerning the promulgation of Americans with Disabilities Act regulations in 1990:
- Sec. 12116. Regulations
Not later than 1 year after July 26, 1990, the Commission shall issue regulations in an accessible format to carry out this subchapter in accordance with subchapter II of chapter 5 of title 5.
When we were drafting ADA regulations, that language served as a hard deadline that had to be, and was, met. A direction to simply prepare regulations is softer, more flexible language. The Department could literally satisfy the requirement to prepare regulations by completing the initial draft for review by the various entities whose approval is required to actually get the regulations issued. So contractors may or may not see regulations issued in late October when the 90 days would be up.
The Executive Order was an expedient, but less than ideal avenue, for extending affirmative action and non-discrimination protections to the LGBT community. What would be more appropriate is a non-statistics driven model of enforcement. Elements of such a model could be derived from Section 503 (disability) and Section 4212 (veterans) protections since these categories also lack a significant body of reliable employment statistics. However, since both disability and veterans’ regulations are grounded in statutory law which is separate and different from the Executive Order, it might be a stretch to import a non-Executive Order enforcement scheme into the Executive Order at all or to import them just for the two added categories.
There are aspects of the Executive Order, such as the protection against discrimination on the basis of color and religion, that are not included in the data collection requirements of 41 CFR 60-1.12(c). However, there are also precious few OFCCP cases that involve either of these two bases.
It will be interesting to see whether the way in which the Executive Order was amended, simply adding these two terms, limits the agency’s regulatory options. This is a matter that may have to be decided in litigation. It is no secret that the signing of the Executive Order was prompted by, among other things, the inability to get legislation passed addressing this issue. In the current political climate, attention may be focused on the extent to which agency regulation appears to venture into traditionally legislative territory.
What should not happen is the implementation of significant requirements for data collection and reporting that add costs to contractors with no clear identification of how these additional costs will have any meaningful impact on eliminating discrimination against the LGBT community. It would be counterproductive to divert contractor time and resources from LGBT protection programs that are working, to bureaucratic exercises that may not make a difference.