Concept
Gender and Intermediate Scrutiny
Classifications subject to heightened scrutiny
Gender and Intermediate Scrutiny
Gender classifications receive intermediate scrutiny, and the Court has made that middle tier steadily more demanding. Once a plaintiff establishes a prima facie case of gender discrimination, a court evaluates the classification under a standard that has grown more rigorous in recent years, even though it may not yet be quite as strict as the review used in race and national origin cases.1 In its most quoted form, the standard places a heavy burden on the government: parties who defend gender-based action must demonstrate an “exceedingly persuasive justification.”2
The tiers are not interchangeable, so the label matters. Every racial classification imposed by any governmental actor, whether federal, state, or local, must satisfy strict scrutiny: it is constitutional only if it is a narrowly tailored measure that furthers a compelling governmental interest.3 A court that tests a race classification under intermediate scrutiny has therefore used the wrong tier. When the Court announced that rule, the court of appeals below had upheld the challenged programs under intermediate scrutiny without deciding whether the interests were compelling or the programs narrowly tailored, and the Court remanded for that analysis.4 Gender review is equally distinct from the “more searching form of rational basis review” that the Court has applied when a law exhibits a desire to harm a politically unpopular group. Gender lines have a standard of their own.5
The reason sex draws heightened scrutiny differs from the reasons given for other classes. State laws that discriminate against aliens, for example, receive strict scrutiny in part to keep the states from nullifying the federal government’s decision to let those individuals live in the United States.6 Sex classifications raise a different worry. Like classifications based on race and national origin, they are usually based on stereotypes rather than important government interests, and many purported biological differences offered to justify them are stereotypes in reality, as in the cases where women were kept from being licensed as bartenders or from being automatically considered for jury service.7 Sex is also an immutable characteristic, and advocates of strict scrutiny argue that laws burdening people for traits they did not choose and cannot change deserve a strong presumption against them.8 Gender, like race, is immediately visible, and women, like racial minorities, tend to be significantly underrepresented in the political process.9 Those who argue for intermediate rather than strict scrutiny answer with arguments of their own.10
The subsections below trace the middle tier from Craig v. Boren through United States v. Virginia. They then turn to the remedies Virginia requires, to two applications in marital property and jury selection, and to the sex lines that survive review. Each asks the same working question: does the sex line track a real difference, or does it repeat an assumption about what men and women are like?
Craig and the Intermediate Test
Craig v. Boren stated the formula for gender classifications: they “must serve important governmental objectives and must be substantially related to achievement of those objectives.”11 The challengers were a man between eighteen and twenty-one and a licensed vendor of 3.2% beer, who claimed that Oklahoma’s statute violated the Equal Protection Clause. The district court agreed that the gender-based differential denied equal protection and enjoined its enforcement.12 The statute established different drinking ages for men and women.13
The formula came at the end of a short, contested development. Before 1971 the Court had reviewed sex classifications in the same manner as purely economic ones.14 Goesaert v. Cleary, decided in 1948, belongs to that era. Craig disapproved it to the extent it was inconsistent with the new analysis,15 and Virginia later cited it as an example of how sex classifications “once” were used to create or perpetuate the inferiority of women.16
In Reed v. Reed the Court first offered realistic protection against sex discrimination under the equal protection guarantee. Idaho’s statute ranked eligible administrators of an intestate estate in eleven categories by their relationship to the decedent. When two applicants fell in the same category, the male was to be preferred, and the Idaho Supreme Court read the preference as mandatory and as applying without regard to individual qualifications.17 The Chief Justice framed the question in rational-relationship terms, and the Court answered no. It conceded that reducing probate workload by eliminating one class of contests was not without some legitimacy. It held, however, that a mandatory preference for one sex, adopted merely to eliminate hearings on the merits, is the kind of arbitrary legislative choice that equal protection forbids.18
The result is striking because the Court conceded the legitimacy of the State’s objective, and a law with a legitimate objective might be expected to survive conventional rationality review.19 The Court was evidently applying something stricter than its words, and Justice Brennan later found implicit support for treating sex as suspect in Reed’s “departure from ‘traditional’ rational-basis analysis.”20
Frontiero v. Richardson pressed that premise further. Justice Brennan advocated treating gender as a suspect classification, but he obtained only a plurality.21 Joined by Justices Douglas, White, and Marshall, he concluded that “classifications based upon sex, like [those] upon race, alienage, and national origin, are inherently suspect and must therefore be subjected to close judicial scrutiny.”22 His reasoning rested on history:
“[Our] Nation has had a long and unfortunate history of discrimination. Traditionally, such discrimination was rationalized by an attitude of ‘romantic paternalism’ which, in practical effect, put women, not on a pedestal, but in a cage. As a result of notions such as these, our statute books gradually became laden with gross, stereotyped distinctions between the sexes . . . .”23
Because the plurality did not become a majority, strict scrutiny for sex was not adopted.24 Craig then supplied the middle formula, and the dissent in Craig recorded both the move and the objections to it. Justice Rehnquist complained that the Court announced the standard “without citation to any source,” and he read the opinion as a retreat by the Frontiero plurality from its view that sex is a suspect classification.25 He added that “important” and “substantially” were “so diaphanous and elastic as to invite subjective judicial preferences or prejudices” disguised as judgments about legislative objectives.26 A concurring opinion objected from the other direction. It found the classification objectionable because it rested on an accident of birth, because it was a remnant of a now almost universally rejected tradition of discriminating against males in that age bracket, and because, “to the extent it reflects any physical difference between males and females, it is actually perverse.” That opinion also stressed that there is “only one Equal Protection Clause,” which does not direct courts to apply one standard in some cases and another in others.27
Craig itself shows the second half of the test doing its work. Oklahoma justified its line by traffic safety, and the question was whether that justification could make an otherwise offensive classification acceptable.28 The Court held that the State’s showing did not satisfy it that sex is “a legitimate, accurate proxy for the regulation of drinking and driving.” It added that proving broad sociological propositions by statistics is a dubious business, in tension with the normative philosophy that underlies the Equal Protection Clause.29 The dissent read the same evidence differently. For rationality purposes, it argued, a classification does not depend on a statistical “fit” between the class and the trait. It turns on whether the trait occurs sufficiently more often in the included class to justify different treatment, and the evidence suggested clear differences between the drinking and driving habits of young men and young women.30
The disagreement shows what the middle tier changes. Under the deferential baseline a higher incidence of the trait in the burdened class is enough, while under Craig the sex line itself must be a legitimate and accurate stand-in for the conduct the State regulates. The Court later described Craig as invalidating a law although the evidence behind the age differential was “not trivial in a statistical sense,” and it stated the lesson broadly: gender classifications that rest on impermissible stereotypes violate equal protection even when some statistical support can be conjured up.31
Oklahoma also argued that its power to regulate alcohol under the Twenty-first Amendment sheltered the statute. The Court held that the amendment does not save invidious gender-based discrimination from invalidation as a denial of equal protection.32 The Court has never recognized enough strength in the amendment to defeat an otherwise established claim of invidious discrimination, and state liquor regulatory schemes cannot work invidious discriminations that violate the Equal Protection Clause.33 A footnote marks what Oklahoma could still do: it remained free to redefine the cutoff age for 3.2% beer as it chose, provided that the redefinition operated in a gender-neutral fashion.34 A single cutoff age for every buyer classifies by age, while separate ages for men and women classify by sex.
Because the challengers in Craig were men, the case also shows that the test follows the classification and not the claimant’s sex. The dissent objected that men challenging a statute that treats them less favorably than women should not be able to invoke a more stringent standard than most other classifications receive.35 The Court later confirmed the opposite: a state practice that “discriminates against males rather than against females does not exempt it from scrutiny or reduce the standard of review.” It rejected the argument that men deserve no protection because they are not victims of historical discrimination.36 A separate opinion in J.E.B. put the principle in textual terms: the Clause’s neutral phrasing, extending its guarantee to “any person,” shows a concern with the rights of individuals and not of groups.37
Later cases tightened the language that Craig left elastic. Parties who defend gender-based government action must demonstrate an “exceedingly persuasive justification,” an instruction the Court traced to its decisions in J.E.B. and Mississippi University for Women.38 The requirement rests on history, since “our Nation has had a long and unfortunate history of sex discrimination” that warrants heightened scrutiny of all gender-based classifications.39 Even a separate opinion that conceded the intermediate test “may not provide a very clear standard in all instances” observed that the case law reveals “a strong presumption that gender classifications are invalid.”40 The next subsection shows how Virginia turned that presumption into a working burden.
Exceedingly Persuasive Justification
United States v. Virginia states the working standard for a gender classification as a burden on its defender: parties who seek to defend gender-based government action must demonstrate an “exceedingly persuasive justification.”42 The phrase does not convert the middle tier into strict scrutiny. The standard remains intermediate review, increasingly rigorous but not yet quite as strict as review of race and national origin classifications.43 Virginia nonetheless drew sharp disagreement about how demanding the standard had become, and one dissent charged that the Court, while making no reference to the Government’s argument, effectively accepted it.44
The standard has three parts, and any one of them can defeat a defense:
- The classification must serve important governmental objectives, and the discriminatory means must be substantially related to achieving them.45
- The justification must be genuine, not hypothesized or invented post hoc in response to litigation.46
- The justification must not rely on overbroad generalizations about the different talents, capacities, or preferences of males and females.47
The second part closes the opening that after-the-fact lawyering would otherwise create, since a respectable-sounding reason can almost always be composed once a suit is filed. The Court therefore insists that a tenable justification must describe actual state purposes, not rationalizations for actions in fact differently grounded, and that “benign” justifications offered in defense of categorical exclusions will not be accepted automatically.48
The third part targets how a sex line works in practice. A classification built on a generalization applies to every member of the sex, including the individuals who depart from the average, so it removes opportunity from exactly the people the generalization misdescribes. The Court has cautioned reviewing courts to take a “hard look” at generalizations or “tendencies” of this kind, and it has instructed that state actors who control the gates to opportunity may not exclude qualified individuals based on “fixed notions concerning the roles and abilities of males and females.”49 That caution fits the observation that gender classifications are usually based on stereotypes.50 Sex classifications are not forbidden, however. They may compensate women for particular economic disabilities they have suffered, promote equal employment opportunity, and advance the full development of the talent and capacities of the Nation’s people, but they may not be used, as they once were, to create or perpetuate the legal, social, and economic inferiority of women.51
Virginia defended the all-male Virginia Military Institute (VMI) with two justifications. The first was that single-sex education provides important educational benefits and contributes to diversity in educational approaches. The second was that VMI’s adversative method of character development and leadership training would have to be modified if women were admitted.52 The Court took them in turn.
The diversity justification failed the genuineness requirement. Mississippi University for Women was directly in point: the State there defended excluding men from a nursing school as “educational affirmative action,” but a “searching analysis” found no close resemblance between the alleged objective and the actual purpose underlying the classification, and the Court reached the same conclusion about Virginia’s claim.53 History and the State’s own conduct supplied the evidence. When Virginia established VMI in 1839, higher education was considered dangerous for women, and the Nation’s first universities admitted only men; VMI followed the lead of the University of Virginia.54 The Fourth Circuit found no policy of diversity in the movement of Virginia’s other public colleges away from single-sex education, and it questioned how one institution with no authority over others could give effect to a state policy of diversity among institutions. A purpose genuinely to advance an array of educational options is not served by a plan that affords a unique educational benefit only to males: however “liberally” the plan serves the Commonwealth’s sons, it makes no provision for her daughters.55
The adversative-method justification failed as a generalization. It was undisputed that admitting women would require accommodations, mainly in housing and physical training, but it was equally undisputed that “the VMI methodology could be used to educate women,” and the district court allowed that some women would want to attend if they had the opportunity.56 The findings Virginia relied on restated expert opinions about typical tendencies: “[m]ales tend to need an atmosphere of adversativeness,” while “[f]emales tend to thrive in a cooperative atmosphere.” The State’s expert said that educational experiences must be designed “around the rule,” not “around the exception.”57 The Court assumed that most women would not choose VMI’s method but observed that many men would not either, and it framed the question as whether the Commonwealth could constitutionally deny to women who have the will and capacity the training and opportunities VMI uniquely affords.58 A rule designed around the rule necessarily sacrifices the exception, which is why an average, even a true one, cannot carry the justification.
The Commonwealth’s stated goal of producing citizen-soldiers is great enough to accommodate women. It is not substantially advanced by their categorical exclusion, in total disregard of individual merit, from the Commonwealth’s premier citizen-soldier corps.59 Virginia therefore showed no exceedingly persuasive justification, and the Court affirmed the Fourth Circuit’s initial judgment.60
In a VMI-type problem the importance of the stated goal is rarely contested. The fit between the goal and the exclusion is where the state’s proof gives out. The common error is to accept a diversity or tradition rationale because the goal sounds benign, when the standard asks whether that was the state’s actual purpose and whether the exclusion advanced it. The program Virginia offered women as a remedy, the Mary Baldwin VWIL program, did not cure the violation, so the Court reversed the final judgment, and the next subsection takes up that remedy question.61
Separate Programs as Remedy
A remedial decree must closely fit the constitutional violation, and it must be shaped to place persons unconstitutionally denied an opportunity or advantage in “the position they would have occupied in the absence of [discrimination].” In Virginia the violation was the categorical exclusion of women from an extraordinary educational opportunity afforded to men, so any remedy is measured against that opportunity. A proper remedy aims to eliminate the discriminatory effects of the past, so far as possible, and to bar like discrimination in the future.63
The rule takes this shape because the way a court defines the violation fixes what the cure must deliver. When the wrong is exclusion from a particular opportunity, the injured persons are those ready, willing, and able to take it, and a remedy helps them only if it gives them that opportunity or a genuine equivalent. Virginia chose to leave VMI’s exclusionary policy untouched and proposed, for women only, a separate program, different in kind from VMI and unequal in tangible and intangible facilities. Having violated equal protection, Virginia was obliged to show that its proposal “directly address[ed] and relate[d] to” the violation, meaning the equal protection denied to women ready, willing, and able to benefit from educational opportunities of the kind VMI offers.64 Virginia described the program, VWIL, as a “parallel program” sharing VMI’s mission of producing “citizen-soldiers,” and the Court tested that description against the two programs themselves.
The comparison ran along three lines:
- VWIL gave women no opportunity to experience the rigorous military training for which VMI is famed. It deemphasized military education and used a “cooperative method” of education meant to reinforce self-esteem.65
- VWIL students did not live in a military-style residence, were not required to live or eat together throughout the program, and did not wear uniforms during the school day. They missed the barracks life that the district court found to be the most important part of the VMI educational experience.66
- VWIL delivered leadership training through seminars, externships, and speaker series, which lacked the physical rigor, mental stress, and psychological bonding of VMI’s adversative training.67
The resources and standing of the two programs also differed. VMI offers an educational opportunity that no other Virginia institution provides, and its prestige, associated with its success in developing citizen-soldiers, is unequaled. VWIL’s faculty was less impressively credentialed and less well paid, its course offerings were more limited, and it offered fewer opportunities for military training and scientific specialization.68
The differences in method and setting were deliberate. Virginia deemed the barracks experience, the core of VMI’s training, nonessential and even inappropriate for its female citizen-soldiers.69 That choice rested on the same claims about typical female tendencies that the liability analysis had already held insufficient to deny opportunity to women who want it. Purported differences between the sexes are often in reality just stereotypes.70
The Fourth Circuit acknowledged that the VWIL degree lacks the historical benefit and prestige of a VMI degree, that VMI is an ongoing and successful institution with a long history, and that no comparable single-gender women’s institution exists. It nonetheless declared the substantially different and significantly unequal VWIL program satisfactory.71 It reached that result by changing the question. It did not ask whether the remedy placed women denied the VMI advantage in the position they would have occupied absent discrimination. It asked whether the Commonwealth could provide separate and unequal educational programs for men and women, and it devised a “substantive comparability” inquiry that gave the effect of the classification little or no scrutiny.72 The Supreme Court held that deferential review inconsistent with the more exacting standard its precedent requires, and it reversed the final judgment and remanded for further proceedings.73 A separate program is itself a sex-based arrangement, so deference at the remedy stage would let a state undo the liability holding through its choice of cure. The question whether a separate state-created school can stand in for the one a student was denied is an older one; compare Sweatt v. Painter.
A concurring opinion would have defined the violation differently. The majority’s definition of the violation, that concurrence said, necessarily implies that the only adequate remedy is admission of women to the all-male institution. It would instead locate the violation in the maintenance of an all-men school without providing any, much less a comparable, institution for women.74 It still agreed that VWIL fails as a remedy, because VWIL is distinctly inferior to VMI, is a program appended to a private college and not a self-standing institution, and is substantially underfunded.75 That alternative is a minority view. The majority’s definition of the violation as categorical exclusion governs, and both opinions agree that VWIL was inadequate.
The third step does the most work in a problem built on a program like VWIL, because a state can usually show comparable spending or a shared mission statement, and those facts leave the intangible gap untouched. The common error is to treat a program labeled parallel, or equal funding per student, as the end of the inquiry, when the question is whether the excluded persons received the opportunity they were denied.
Gender Rules in Marital Property
Kirchberg v. Feenstra applies the intermediate test to a rule about who controls a married couple’s property. Article 2404 of the Louisiana Civil Code allowed husbands, but not wives, to execute mortgages on jointly owned real estate without spousal consent, and the Court held that the article violated the Equal Protection Clause.77 The article named the husband in its own text, so it classified spouses by sex on its face, and the question was whether it substantially furthers an important government interest.78
History explains why a rule of this kind draws scrutiny. Under the common-law doctrine of coverture, a married woman had no legal identity apart from her husband.79 The doctrine eroded gradually, but invidious sex-based classifications in marriage remained common through the mid-twentieth century.80 The Frontiero plurality described the older legal position of women in just these terms: married women traditionally were denied the legal capacity to hold or convey property or to serve as legal guardians of their own children.81 A statute that empowers the husband alone to dispose of jointly owned property belongs to that family of laws.
The party defending the article did not claim that it served an important governmental interest.82 He argued instead that Mrs. Feenstra could have taken steps to keep her husband from mortgaging the home without her consent. The Court answered that the argument overlooks the critical question, whether the article substantially furthers an important government interest, and that the “absence of an insurmountable barrier” will not redeem an otherwise unconstitutionally discriminatory law.83 The burden remains on those defending the discrimination to make out the claimed justification, and the Court would not speculate about a justification no one had offered.84 The answer follows from where the constitutional injury lies. A default that gives control to the husband unless the wife acts has already sorted the spouses by sex, and a way for the disfavored spouse to avoid the sorting leaves the sorting unjustified.
A defender who fell back on administrative convenience in having one designated manager would meet Reed, which held that a mandatory preference for one sex, adopted merely to eliminate hearings on the merits, is the kind of arbitrary legislative choice that equal protection forbids.85 A concurring opinion reached the same result through the language of similarity: men and women were similarly situated for all relevant purposes with respect to the management and disposition of community property.86
An answer to a marital-property problem should begin by asking whether the rule assigns power by sex on its face. If it does, the defender must answer for the objective and the fit, and the existence of a way out for the disfavored spouse answers for neither.
Gender and Jury Selection
Peremptory challenges based solely on gender violate the Equal Protection Clause. In J.E.B. v. Alabama ex rel. T.B. the Court held that gender, like race, is an unconstitutional proxy for juror competence and impartiality.88 The case arose from a paternity and child-support action that the State brought on behalf of a mother against J.E.B. The panel of 36 potential jurors included 12 men and 24 women. The State used nine of its ten peremptory strikes against men, and the defendant used all but one of his against women, so every seated juror was female.89
The strikes themselves were the classification, since each removed a potential juror because of sex, and the Court applied the heightened scrutiny it affords all gender-based classifications, which demands an exceedingly persuasive justification.90 The justification offered could not carry that burden. The rationale was reminiscent of the arguments once advanced to justify the total exclusion of women from juries, it offered virtually no support for the claim that gender alone predicts a juror’s attitudes, and it assumed that gross generalizations impermissible when made on the basis of race are somehow permissible when made on the basis of gender.91 A grain of statistical truth would not have helped, since gender classifications that rest on impermissible stereotypes violate equal protection even when some statistical support can be conjured up, and the generalization advanced by Alabama was at the least overbroad.92 The defect is the one that sank the proxy in Craig: a strike that assumes how a juror will vote because of sex applies an average to an individual who may not fit it.
Discrimination in jury selection, whether based on race or on gender, harms the litigants, the community, and the individual jurors who are wrongfully excluded.93 The litigants are harmed by the risk that the prejudice behind the discriminatory selection will infect the entire proceeding.94 The community is harmed by the State’s participation in perpetuating invidious group stereotypes and by the loss of confidence in the judicial system that state-sanctioned discrimination in the courtroom engenders.95 The jurors’ own right does the most work in the rule. Striking jurors on the assumption that they hold particular views simply because of their gender is “practically a brand upon them, affixed by the law, an assertion of their inferiority.”96
The right belongs to the potential jurors as well as to the litigants, so the possibility that members of both genders will end up on the jury despite the intentional discrimination is beside the point. Excluding even one juror for an impermissible reason harms that juror and undermines public confidence.97 That individual focus answers the dissent’s argument that the system is evenhanded because, for every man the State struck, the defendant’s lawyer struck a woman.98 Equal numbers of strikes against each sex do not cure the injury to the individuals struck. The right extends to men and women alike, since all persons, when granted the opportunity to serve on a jury, have the right not to be excluded summarily because of discriminatory and stereotypical presumptions.99 A separate opinion put the point in individual terms: a person denied jury service because a peremptory challenge was exercised against her on account of her sex is no less injured than a person denied service by a law banning members of her sex.100
The Court recognized that, whether the trial is criminal or civil, potential jurors as well as litigants have an equal protection right to jury selection procedures free of state-sponsored group stereotypes rooted in historical prejudice.101 A concurring opinion joined the Court but urged that the holding be limited to the government’s use of gender-based peremptory strikes.102 The Court’s own language is broader, since it holds that litigants may not strike potential jurors solely on the basis of gender, and the strikes in J.E.B. itself included those of the defendant.103
J.E.B. leaves the peremptory challenge in place. Parties may still remove jurors they feel might be less acceptable than others on the panel, and gender simply may not serve as a proxy for bias.104 Careful questioning of prospective jurors makes reliance on stereotypes both unnecessary and unwise, because voir dire can reveal actual or implied bias and give the parties a firmer basis for their strikes.105
A jury-selection problem turns on what the strike rested on. A lawyer who strikes a juror because of a tie to a party or a witness has acted on a characteristic other than sex, and J.E.B. does not reach the strike even if the juror removed is a man. A lawyer who strikes the same juror because men are presumed to sympathize with one side has acted on sex itself. The common error is to ask whether the lawyer’s prediction about the juror was accurate, when the question is whether sex supplied the reason.
Gender Lines That Survive
Heightened scrutiny does not forbid every sex classification. Virginia itself lists the uses it permits: sex classifications may compensate women “for particular economic disabilities [they have] suffered,” as in Califano v. Webster, may promote equal employment opportunity, and may advance the full development of the talent and capacities of the Nation’s people. They may not be used, as they once were, to create or perpetuate the legal, social, and economic inferiority of women.107 The decisions below show a sex line surviving when it rests on a genuine difference or a genuine compensatory purpose, and failing when it rests on an assumption about what men and women are like.
A court therefore tests a claimed difference instead of assuming it. Gender classifications are usually based on stereotypes rather than important government interests,108 and many purported biological differences invoked to justify legal distinctions are in reality just stereotypes.109 The survivors are the lines for which the state can show the difference is real and tied to the law’s purpose.
In Michael M. v. Superior Court the Court upheld California’s statutory rape law, which treated the sexes differently, by applying what amounted to rational basis review. The plurality opinion by Justice Rehnquist did not require the state to prove that the gender-discriminatory law was any more effective than a gender-neutral law would be in deterring minor females from engaging in intercourse.110 The State offered two arguments: that the law could single out the participant who suffers fewer of the harmful consequences of such a pregnancy, and that a gender-neutral statute would frustrate enforcement because a female would be less likely to report a violation if she could also be prosecuted.111 Michael M. is a weak survivor, and an answer built on it should state the pregnancy premise and the enforcement reason and should not present the case as showing that any sex line tied to sexual conduct passes.
Nguyen v. Immigration & Naturalization Service allows a third type of gender classification: one benefiting women because of biological differences between men and women. The Court allowed a difference in INS rules that favored mothers over fathers because of the greater certainty about the identity of the mother as compared to the father and the greater opportunity mothers have to establish a relationship with their children.112 Later challenges to gender distinctions in citizenship rules, including Sessions v. Morales-Santana, should be read alongside it.
The second group of survivors compensates for a disadvantage the beneficiaries actually face. In Schlesinger v. Ballard the Court upheld a Navy regulation that required the discharge of male officers who had gone nine years without a promotion but allowed women officers to remain thirteen years. The Court reasoned that men had more opportunities for promotion than women, and it wrote that “Congress may quite rationally have believed that women line officers had less opportunity for promotion than did their male counterparts, and that a longer period of tenure for women officers would, therefore, be consistent with the goal to provide women officers with fair and equitable career advancement programs.”113 The commentary on these cases adds that the Court had not yet considered a challenge to an affirmative action program designed to benefit women. Intermediate scrutiny governs all gender-based classifications, and many of the same issues will arise as in race-based affirmative action, including which interests justify such a program and which techniques are permissible.114
A benign label does not carry a law through. In Mississippi University for Women the State defended excluding men from a nursing school as “educational affirmative action,” but a searching analysis found no close resemblance between the alleged objective and the actual purpose underlying the classification.115 A law that favors women receives no extra credit for its aim. It is judged by the same standard, and the contest is usually over whether the means fit the disadvantage the law claims to address. The same questions carry over to other settings, such as the male-only draft registration in Rostker v. Goldberg and the alimony statute in Orr v. Orr.
The third step decides most problems, because a state can almost always call its purpose compensatory or protective, and the label carries weight only when the record shows a real difference or a real disadvantage in the sphere at issue. Pregnancy and birth supplied that showing in Michael M. and Nguyen, and a documented promotion gap supplied it in Ballard. The nursing school in Mississippi University for Women and the stereotype about preferences in Virginia supplied nothing of the kind.
Every decision in this part involved a law that classifies by sex on its face. The next part turns to laws that do not, as in the Arlington Heights dispute that began when a developer applied to rezone a fifteen-acre parcel for 190 clustered townhouse units for low- and moderate-income tenants.116