The Trump administration and its supporters have been trying out a peculiar argument to justify several of the administration’s discriminatory policies. The argument goes something like this: The administration did not discriminate against a particular group because the administration did not discriminate against all of the members of that group. (When applied to more discrete actions, the argument is that the administration did not discriminate against a particular individual based on the individual’s membership in a particular group because the administration did not also discriminate against all other members of that group.)

I’ll call this argument the “under-inclusive theory of discrimination.” The theory has little to recommend it. Uncharitably, the argument serves as a way to justify clumsy and poorly executed discrimination. That is, although the Trump administration may be bad at carrying out discrimination (because it has not managed to discriminate against all members of a particular group), it does not follow that the administration is not engaged in discrimination when it disadvantages some members of a particular group.

But back to the “serious” version of the under-inclusive theory of discrimination.  

Here are some versions of the under-inclusive theory of discrimination that the administration and its supporters have made to defend some of the Trump administration’s policies:

(1)  The Muslim ban is not a Muslim ban because it does not discriminate against all Muslim-majority countries. The pro-ban e-mails that impermissibly flooded the Fourth Circuit Court of Appeals stated that the district court made a terrible mistake by “fail[ing] to consider [that] the travel ban addressed only six of forty-nine (12%) of Muslim majority countries.”

(2)  In its briefs to the Fourth Circuit Court of Appeals and Ninth Circuit Court of Appeals, the Trump administration likewise argued that the Muslim ban is not a Muslim ban because it does not discriminate against all Muslims and all Muslim-majority countries. The brief notes that the travel ban order focused on only “a handful of Muslim-majority countries.” And, the brief maintains, the order cannot possibly discriminate against Muslims because the order applies to countries that “represent a small fraction of the world’s 50-Muslim majority nations and approximately 10% of the global Muslim population.”

(3)  ICE officials are not arresting, detaining, and interrogating individuals because they are of Mexican nationality (or because they are DREAM-ers). How do we know this? Because when ICE is arresting those individuals, it is not arresting, detaining, and interrogating other individuals who were also present and who are also of Mexican nationality (or also DREAM-ers).

These arguments all take the following form: A policy (or decision) does not constitute “discrimination” against a particular group unless the policy (or decision) discriminates against all members of that particular group.

The under-inclusive-discrimination-is-not-discrimination argument is wrong.  A policy can still discriminate against a group if the only individuals it disadvantages are members of a particular group, even if the policy does not disadvantage all members of that group.  (The same is true if most or many of the individuals disadvantaged are members of a particular group.)  Likewise, just because an decision-maker has not "discriminated" against some individuals based on a particular trait does not mean that the decision-maker has not discriminated against some individuals based on that same trait (be it sex, race, religion, or something else).

The under-inclusive-discrimination-is-not-discrimination argument is also foreclosed by case law. Below I’ve grouped some cases that have rejected the Trump administration's under-inclusive theory of discrimination into three categories—(1) “facial disparate impact” cases, which involve facially neutral policies that disproportionately burden or benefit individuals who share a particular trait; (2) “facial disparate treatment” cases, which involve policies that facially discriminate against some individuals because of their membership in a group; and (3) “individual disparate treatment” cases, which involve allegations that a particular decision-maker (or decision-makers) discriminated against a particular individual (or individuals) because of their membership in a group.

In all three kinds of cases, it doesn’t matter that a policy may not discriminate against all members of a group, or that a decisionmaker does not discriminate against all members of a group all of the time. Rather, the relevant question is whether a decision-maker or policy selected some individuals for differential treatment because of those individuals’ group affiliation. That is, under current doctrine, what matters is the decision-maker's motives—why the decision-maker(s) have singled out some individuals for disparate treatment, or why they treated particular individuals. That a policy (or decision-maker) did not treat all individuals with a particular trait differently does not answer whether the policy (or decision-maker) treated a particular individual or some individuals differently because of those individuals’ particular trait.

Facial Disparate Impact (facially neutral policies that disproportionately burden or benefit individuals who share a particular trait)

Personnel Administrator of Massachusetts v. Feeney involved a constitutional challenge to a statute that gave veterans preferential treatment in their applications for state civil service jobs. The plaintiffs argued that the veteran-preference rule was unconstitutional because it disproportionately burdened women—over 98% of veterans were men.

If the under-inclusive theory of discrimination were correct, then all Feeney would have had to say is that the veteran-preference rule did not benefit all men and did not disadvantage all women: not all men are veterans and almost 2% of veterans were women. But that’s not what Feeney said. Rather, Feeney (like all of the disparate impact cases) said that the constitutionality of the policy turned on the government’s motives for enacting the policy, and specifically whether the government intended the policy to benefit men and disadvantage women.

Applied to the Muslim ban, then, it’s irrelevant that the policy disadvantages only some Muslims and some Muslim-majority countries. The relevant question is why the policy does so—that is, was one of the motivating factors for the ban the desire to exclude Muslims, or animus toward Muslims?

Rice v. Cayetano involved a constitutional challenge to Hawaii’s policy of allotting voting power in the Office of Hawaiian Affairs. Hawaii had limited voting in OHA to those persons who were descendants of individuals who lived in Hawaii in 1778. The Court held the policy amounted to a classification based on race. Hawaii had argued the policy could not possibly be a classification based on race because it neither included all individuals of particular races, nor excluded all individuals of particular races. Here’s what the Court had to say about this argument:

[T]his too is insufficient to prove the classification is nonracial in purpose and operation. Simply because a class defined by ancestry does not include all members of the race does not suffice to make the classification race neutral. Here, the State's argument is undermined by its express racial purpose and by its actual effects.

Substitute “Muslim” for “Polynesian,” and “religion” for “race,” and Justice Kennedy (the author of Rice) might as well have been explaining why the Muslim Ban is a Muslim Ban.

In Connecticut v. Teal, a group of employees argued a test that determined promotions constituted unlawful discrimination because white candidates performed better than African-American candidates on the test (68% pass rate versus 54%). Although the employer relied on the test, the employer also applied something like an affirmative action policy to determine which employees ultimately received a promotion. The result was that the employer promoted 11 out or 48 African-American applicants, and 35 out of 259 white applicants (22.9% of African-American applicants and 13.5% percent of white applicants). The employer’s defense was that it “compensated for a discriminatory pass-fail barrier by hiring or promoting a sufficient number of black employees to reach a nondiscriminatory bottom line.”

The Court rejected that defense: The employers, the Court reasoned, “seek simply to justify discrimination against respondents on the basis of their favorable treatment of other members of respondents’ racial group …. [I]rrespective of the form taken by the discriminatory practice, an employer’s treatment of other members of the plaintiffs’ group can be of little comfort to the victims of discrimination.”

In other words, in the context of a disparate impact case, where the government adopts a facially neutral policy that disadvantages one group more than others, the relevant legal question is why the government adopted that policy. The government’s policy is not legal simply because it does not discriminate against all members of the group that has felt most of the policy's burdens.

Facial Disparate Treatment (where a policy, on its face, treats some groups differently than others)

Craig v. Boren invalidated a state statute that restricted the sale of non-intoxicating beer to men over 21 and women over 18. The Court described the statute as a classification based on sex, and thus discrimination on the basis of sex. But the statute did not discriminate against all men—men under 18 and men over 21 were treated the same as women of similar ages. The statute only treated men between 18 and 21 differently than women of the same age.

Plessy v. Ferguson addressed the constitutionality of a state statute that directed railroads to have separate cars for white passengers and “colored race” passengers. The statute did not prohibit all African-American passengers from sitting in the car for white passengers.; it excepted “nurses attending children of the other race.” Plessy, of course, infamously upheld the statute on the ground that it did not treat individuals differently because of their race. Does the Trump administration think that Plessy was correct that the statute did not constitute discrimination on the ground that the statute did not prohibit all African-American passengers from sitting in the white car?

Phillips v. Martin Marietta Corp. addressed a challenge to a company’s policy of accepting job applications from only those women who did not have pre-school-age children. (The employer accepted job applications from all men.) The lower federal courts held that the employer did not discriminate against women because “70-75% of the applications for the position … were women; 75-80% of those hired for the position … were women, [and] hence no question of bias against women as such was presented.” The Supreme Court reversed. (The trial court had granted summary judgment to the employer, and the court of appeals had affirmed.) Thus, the fact that a policy does not disadvantage all members of a particular group does not mean the policy is necessarily legal.

Individual Disparate Treatment Cases

Price Waterhouse v. Hopkins addressed a claim that an employer had declined to elevate a woman to partner because she was perceived as “macho” and “overcompensated for being a woman,” among other sex-based reasons. The employer had seven female partners. The Court held that the plaintiff could establish her employer had discriminated against her because of her sex if it had considered her sex in deciding whether to promote her. (Another part of the decision was overruled by the 1991 Civil Rights Act; that portion of the decision held an employer could avoid liability by showing it would have made the same decision about an employee without considering the employee’s sex.)

Furnco Construction Corp v. Waters addressed a claim of employment discrimination. The plaintiffs alleged that two out of three African-American applicants were rejected and pointed to the employer’s history of not hiring African-American applicants. The Court held that a court could consider the overall make-up of the work force to determine whether the employer had discriminated against the African-American applicants, but stressed that the employer’s “overall” treatment of a group was not a defense to discriminate claims: “A racially balanced work force cannot immunize an employer from liability for specific acts of discrimination.”

So, again: The fact that a defendant may not have discriminated against all members of a particular group does not mean the defendant did not discriminate against an individual member of that particular group.

To wrap up: The under-inclusive theory of discrimination is not the law. Courts should refuse the Trump administration’s invitation to make it the law.  And the Trump administration should stop trying to make the under-inclusive theory of discrimination happen. It’s not going to happen.

Disclosure: I am among the counsel to Daniel Ramirez Medina, one of the DREAM-ers who has been arrested, detained, and subject to deportation proceedings.  I am also one of the signatories to an amicus brief in the travel ban cases; Joshua Matz, the publisher of this blog, is counsel to the signatories. Thanks to Helen Klein Murillo for her help with this post.

Tags: