The Michigan Civil Rights Initiative ("MCRI") amended the Michigan Constitution to provide that public universities, colleges, and school districts may not "discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity, or national origin in the operation of . . . public education." We argue that, in addition to prohibiting the overt use of racial preferences in admissions, the MCRI also prohibits using racial proxies such as socioeconomic status or a "Ten Percent Plan" that aim to prefer minorities in admissions.
Though the MCRI does not expressly say so, we stipulate for this paper that its language prohibits universities from overtly considering race in any way when making admissions decisions. It remains unclear, however, what other conduct might be barred by the MCRI. In particular, it is unsettled whether the MCRI would prohibit using proxies or other race-neutral criteria designed to aid minority candidates in admissions. Would a plan similar to the "Texas Ten Percent Plan," which accepts all in-state students who graduate at the top of their high school classes, violate the MCRI as a proxy for racial preference? What about giving applicants a bonus for socioeconomic disadvantage or overcoming adversity? Does purpose matter? These questions are vital to universities trying to achieve racial diversity in a post-MCRI world. The answers will depend mostly on how state courts interpret the MCRI's language on "discrimination" and "preferential treatment."
We contend that using proxies for race would violate the MCRI. Like Justice Souter (dissenting in Gratz v. Bollinger), we believe using disguised proxies creates a system in which "the winners are the ones who hide the ball." To demonstrate how legal challenges to the use of proxies might be resolved, we first set out the analytical framework courts should use when approaching these questions. We then explain why the use of racial proxies is impermissible under the MCRI.
I. Choosing the Right Standard
Because the MCRI is new law, courts need to give its language interpretive meaning. As a practical matter, they must decide whether the MCRI modifies an existing body of law (e.g., state civil rights law or equal protection jurisprudence), and to what extent they must fashion a new body of law to give it force. We consider and reject one possible rule that sees the MCRI as a limited modification to equal protection jurisprudence under the Fourteenth Amendment. As explained below, we believe the MCRI modifies existing state anti-discrimination law. Therefore, courts should import Michigan law's disparate treatment and impact standards to admissions policies challenged under the MCRI.
The first possible rule, which we reject, sees the MCRI as a limit on state action otherwise authorized by Grutter v. Bollinger. Under this view, courts would interpret the MCRI as a slight modification to the Fourteenth Amendment. Michigan could no longer argue that the use of racial classifications, a prima facie violation of the Equal Protection Clause, should nevertheless pass strict scrutiny if they are narrowly tailored to the compelling state interest in diversity. This interpretation would put Michigan after the MCRI in the same position as Texas after the Fifth Circuit's decision in Hopwood v. Texas, which invalidated affirmative action on equal protection grounds. But this interpretation, by relying on the Fourteenth Amendment's asymmetrical treatment of minority and majority racial groups, would permit the purposeful use of proxies or something like the "Texas Ten Percent Plan," whose constitutionality the Court acknowledged in Gratz.
The MCRI's text does not support this interpretation. First, there is always an obvious peril in allowing the meaning of a state constitutional amendment to be dictated entirely by its catalysts rather than its words-an interpretive tool without limits. But more importantly, we believe the MCRI's language supports applying the basic framework of Michigan civil rights laws as they now exist. The MCRI speaks not of equal protection, but of "discrimination" and "treatment"-the vocabulary of anti-discrimination laws. Also, critically, its remedies section provides as follows:
The remedies available for violations of this section shall be the same, regardless of the injured party's race, sex, color, ethnicity, or national origin, as are otherwise available for violations of Michigan anti-discrimination law.
This direct reference to state civil rights law (the Elliot-Larsen Civil Rights Act) and the civil-rights language militate against reading the MCRI as a narrow provision overruling Grutter. Instead, it creates a new state constitutional right to be free from state discrimination. And through its juxtaposition of favorable and unfavorable discrimination and promise of equal access to remedies, the MCRI requires symmetrical treatment for all racial groups.
II. Disparate Impact and Proxies for Race
Established Michigan law already prohibits many kinds of discrimination against racial groups through the Elliot-Larsen Civil Rights Act, which-like its federal counterpart-recognizes two theories of recovery for racial discrimination in employment and education: "disparate treatment" and "disparate impact." Michigan courts often interpret the state act as conforming to the federal version, and thus we will treat both as relevant and persuasive authority.
The same analytical framework applies to both state and federal anti-discrimination law. A plaintiff seeking to prove discrimination through "disparate treatment" in the employment context must show that an employer's facially disparate treatment of employees of different races is motivated by a discriminatory purpose. "Disparate impact" requires the plaintiff to show that the consequences of a particular policy fall more harshly on a protected class, and that the employer cannot justify the policy by business necessity. Under Michigan law, proof of an impermissible purpose is not required to recover on the disparate impact theory.
Given this framework, one could argue that MCRI only bans policies that facially prefer candidates by race, because its language only refers to "preferential treatment" (we call this the "Facial Only Reading"). If the word "treatment" must have meaning, its meaning must be that-in contrast to discrimination-the only preferences MCRI prohibits are those manifested through disparate treatment. This construction would achieve the same result as Hopwood, allowing proxies for race like the "Ten Percent Plan," because these admissions policies are facially neutral.
The fundamental flaw with the Facial Only Reading is that it tries to draw a novel distinction between preferential treatment and discrimination. While giving each word of a statute independent meaning is an ordinary principle of statutory construction, it is inappropriate here, where both terms already have a settled meaning. Preferential treatment has always been discrimination. The Supreme Court noted in Griggs v. Duke Power Company that "discriminatory preference for any group, minority or majority, is precisely and only what [civil rights legislation] has proscribed." Thus, trying to give separate meaning to these two terms is a fruitless exercise that needlessly complicates the interpretive undertaking. Rather, the outcome of a challenge under the MCRI turns on three questions: (1) whether the MCRI distinguishes between majority and minority groups; (2) how a prima facie case of discrimination in admissions may be shown; (3) how that prima facie case may be rebutted. Because, as noted above, we do not believe the MCRI distinguishes between racial groups, we turn to the latter two questions.
A plaintiff may establish a prima facie case of discrimination under the MCRI by showing that a particular admissions has a disparate impact on his racial group. The MCRI prohibits "discrimination against" racial groups, and its symmetry requirement ensures that this protection is available to all applicant-plaintiffs. Because the MCRI does not say "discriminatory treatment" but merely "discrimination," we can infer that it imports the disparate impact theory of recovery for discrimination currently available under state law. This theory would apply to suits by whites challenging the use of race-neutral admissions plans.
We also believe that it would be appropriate for Michigan courts to require proof of a discriminatory motive in a disparate impact case, much like in equal protection jurisprudence. This requirement would permit universities to use good-faith admissions criteria that are genuinely race-oblivious without having courts decide what admissions factors are really important, and which are too trivial to justify discrimination.
Under both state and federal anti-discrimination laws, once the plaintiff establishes a prima facie case of a disparate impact, the burden would shift to the defendant university to prove that the challenged practice is justified by business necessity-a kind of heightened scrutiny. How might a court apply this scrutiny in the MCRI context?
When rebutting a prima facie case, a defendant university might argue that its use of racial proxies passes heightened scrutiny because it has the legitimate purpose of achieving diversity by favoring minorities in admissions. This argument is consistent with traditional interpretations of civil rights laws that also permitted overt racial preferences. But because we do not find in the MCRI any basis for a diversity-seeking exception, we think a proxy clearly motivated by racial diversity will not justify discriminatory preference.
First, Congressional intent, the typical basis for permitting a goal of diversity under Title VI, in inapplicable to an MCRI analysis. Proponents of the Facial Only Reading would look to the definition of "discrimination" and accompanying standard of review under Title VI of the Civil Rights Act, which applies to university admissions. As an initial step, this is perfectly logical. The Supreme Court has held that affirmative action in university admissions does not violate Title VI of the Civil Rights Act unless it also violates the Constitution because "discrimination" under that Title is coterminous with the Equal Protection Clause. Leeway for disparate treatment in favor of minority groups under Title VI goes to constitutional limits. But the basis for this interpretation-which comes from Regents of the University of California v. Bakke and which the Court reaffirmed in Grutter-is that Congress intended discrimination under Title VI to mean "constitutionally impermissible discrimination." This logic cannot be a sound method for interpreting "discrimination" under the MCRI because, quite obviously, Congress did not write the MCRI. Its basic purpose-to end the use of otherwise constitutional racial preferences-is much different than Title VI's aim: to enforce existing constitutional limits. There is no reason to import those exceptions to outlawed discrimination that are based solely on legislative intent to a popular referendum. Further, defining MCRI discrimination by the very cases it overrules-Bakke and Grutter-would be untenably perverse.
Moreover, in case you were hungry for reductio ad absurdum, using the "constitutional" standard to inform the interpretation of "discrimination" in the MCRI would vitiate its purpose. If discrimination means "unconstitutional discrimination," why wouldn't preferential treatment mean "unconstitutional preferential treatment?" And since limited racial preferences in admissions don't violate the constitution, they shouldn't violate the MCRI either. We know that can't be right, so we must look elsewhere for a standard of review.
Unlike the federal Civil Rights Act, which is separated into different Titles for education, employment, race, gender, etc., each with its own developed and often different jurisprudence, the MCRI makes a blanket statement. By banning discrimination or preference in "public employment, public education, or public contracting," the MCRI most plausibly intends that there be one standard for all of these areas. Thus, even if federal law were an appropriate place to look for a standard of review, it is not obvious that Title VI (as opposed to Title VII, for example, which deals with employment discrimination) would be the right source. Once outside the safe harbor of Title VI's "limits of the Constitution" standard, affirmative action programs encounter a tougher review from courts, especially where their sole purpose is to achieve and maintain racial diversity. For example, in Title VII's employment context, when a prima facie case of "reverse discrimination" has been made, the Third Circuit has held that an employer's affirmative action program is invalid under Title VII where it aims to promote diversity rather than to remedy discrimination. Even the exception for remedial programs under Title VII is largely based on Congressional intent, which is absent from the MCRI.
We note the example above from the Third Circuit not because we think Title VII is obviously the right source from which to import a standard of review for the MCRI. In fact, it could not be: remedial affirmative action is unconstitutional in state school admissions unless the institution itself has a history of discrimination. Rather, we find the example illustrative because it demonstrates the slim margin by which universities can use racial preferences. In a recent opinion terminating an injunction delaying the enforcement of the MCRI, the Sixth Circuit observed that using racial preferences in admissions is something universities only "narrowly may do" under the Constitution. The MCRI is much broader than the narrow language necessary to overrule Grutter ("means of achieving the goal of diversity must be facially neutral"). It is far too blunt an instrument to make so fine a cut: carving out the barely constitutional practice of candidly preferring racial minorities in admissions, but nothing more. Instead, it creates a new, colorblind regime for state involvement with race in Michigan. That has consequences beyond simply doing away with overt preferences.