William D. Araiza*

Abstract:  The United States Supreme Court’s recent decision in Board of Trustees of the University of Alabama v. Garrett further circumscribed Congress’ power to enforce the Fourteenth Amendment. The Court’s recent decisions in this area insist that enforcement legislation be congruent and proportional to the constitutional violations sought to be remedied. The specter of reduced leeway for congressional enforcement authority requires Congress to approach such federal legislation carefully. The Employment Non-Discrimination Act (ENDA), proposed legislation prohibiting sexual orientation discrimination in employment, provides an interesting case study of the Court’s recent Section 5 jurisprudence. This Article, after outlining historic and current Section 5 standards, uses Garrett as a guide to examine whether the Supreme Court would uphold ENDA’s provision allowing state employees to sue their employers for certain types of retrospective relief. The Article both argues that ENDA is a valid expression of Congress’ Section 5 power and provides strategies for navigating the increasingly narrow confines of Fourteenth Amendment enforcement power to provide critical employment protections.

In February 2001, the United States Supreme Court decided Board of Trustees of the University of Alabama v. Garrett,1 holding that Title I of the Americans with Disabilities Act (ADA)2 exceeded Con[*PG2]gress’ power to enforce the Fourteenth Amendment.3 Garrett marked a continuation of the Court’s recent practice of closely scrutinizing federal legislation based on Congress’ power to enforce the Fourteenth Amendment, the so-called “Section 5 power.”4 This line of cases,5 most decided by the same 5–4 majority,6 represents a sustained attempt by the Court to circumscribe what had previously been Congress’ extremely broad enforcement power, reviewed deferentially by courts.7 The cases all feature the Court’s new insistence that such enforcement legislation demonstrate “congruence and proportionality” to the constitutional violations sought to be remedied.8 In turn, those violations must either appear likely to the Court (because, for exam[*PG3]ple, they involve discrimination on some disfavored ground such as race) or because Congress has revealed a pattern and practice of unconstitutional conduct by the states.9

While these cases all point in the same direction with regard to the reduced leeway the Court now gives such legislation, each case deals with the vindication of a right that has a different type and level of grounding in the Constitution.10 City of Boerne v. Flores11 dealt with the Religious Freedom Restoration Act (RFRA).12 RFRA attempted to protect religious freedom, a highly protected right explicitly enshrined in the Constitution and incorporated against the states via the Fourteenth Amendment’s Due Process Clause.13 RFRA attempted to undo the effects of a then-recent Supreme Court decision that identified a lower level of scrutiny as appropriate for free exercise claims.14 The Court struck RFRA down, concluding that the statute was simply an attempt to reinterpret the Constitution rather than protect values the Court had identified in its free exercise jurisprudence.15

Florida Prepaid v. College Savings Bank also considered a statute guarding against state deprivations of a due process right, the property right in a patent.16 In that case, the Court struck the statute down because Congress had not demonstrated a pattern of unconstitutional conduct justifying the statutory remedy.17 The companion case, College Savings Bank v. Florida Prepaid, considered whether the Lanham Act’s [*PG4]prohibition on false advertising bestowed on competitors of the alleged false advertiser a property right to be free from such unfair competition.18 The Court rejected the argument that such a right existed in the Due Process Clause and thus concluded that the false advertising prohibition was not “appropriate” Section 5 legislation.19

Kimel v. Florida Board of Regents dealt with age discrimination, a claim made under the Equal Protection Clause of the Fourteenth Amendment.20 However, the Court has held that age classifications are subject only to the lowest level of equal protection scrutiny—the well-known “rational basis” test—and has never struck down an age classification.21 As Congress did not demonstrate actual instances of state age-related employment discrimination that would rise to the level of constitutional violations, the Court held that the Age Discrimination in Employment Act’s (ADEA) application to the states exceeded Congress’ Section 5 authority.22

Garrett completes the set of cases, as it deals with a right, an equal protection-based right against disability-based employment discrimination, that again receives only rational basis scrutiny.23 Unlike age discrimination, however, disability discrimination (at least mental disability discrimination) has been found unconstitutional by the Court in one case, City of Cleburne v. Cleburne Living Center.24 While Cleburne has apparently not heralded a new day of heightened judicial solicitude for the rights of the disabled, mentally or otherwise,25 there is at least precedent on the books for finding such discrimination unconstitutional. Nevertheless, as in Kimel, the Garrett Court held that Congress had not found sufficient evidence of states’ unconstitutional conduct justifying the broad prohibitions in the statute and struck [*PG5]down application to the states of the Americans with Disabilities Act’s (ADA) employment provisions.26

With the basic doctrine in place (the “congruence and proportionality” standard, the requirement that actual unconstitutional conduct be identified or at least obvious, and, more generally, stricter judicial scrutiny), and with examples of its application to the different types of rights guaranteed under the Fourteenth Amendment,27 eyes should now turn to mapping the terrain of future challenges to Section 5 legislation. The proposed federal ban on employment discrimination based on sexual orientation presents an interesting Section 5 question. Such proposed legislation (in recent years entitled the Employment Non-Discrimination Act, or “ENDA”), has been introduced in Congress every year since 199428 and has gathered more support each time.29 While the attitude of the Bush administration is unclear,30 there is at least a possibility that the bill will be enacted in the next several years.31 Such a law would almost assuredly be constitutional as an expression of Congress’ commerce power, even after the Court’s recent retrenchments in this area.32 As such, it could also [*PG6]constitutionally be applied to the states, under the authority of Garcia v. San Antonio Metropolitan Transit Authority.33 The issue, analogous to those raised in Kimel and Garrett, would be whether plaintiffs (in this case, lesbian and gay state employees) could sue their state/employer for retrospective relief such as a damages award.34 Such retrospective relief against a state cannot be authorized pursuant to a statute justified as an expression of Congress’ Article I powers.35 Thus, in order for such employees to collect damages, the statute would have to be supportable as an expression of Congress’ Section 5 power.36

Thus, the question: Would ENDA be an appropriate enforcement statute under current Supreme Court doctrine? Part I of this Article examines the Supreme Court’s Section 5 jurisprudence, paying special attention to the recent line of cases beginning with City of Boerne. Part II of this Article applies the Court’s current approach to Section 5 legislation to the versions of ENDA introduced in recent congressional sessions.37 It is clear that the current approach requires an examination of the suspicion with which the Court has regarded the underlying conduct being restricted. Applying that test to ENDA will require examining how seriously the Court is concerned about dis[*PG7]crimination against gays and lesbians. This, in turn, leads us to consider Romer v. Evans.38 The task here will be to attempt to “translate” Romer into Section 5 language, much as the Court in Garrett attempted to translate Cleburne.39 After considering whether Romer means that all sexual orientation discrimination in employment is constitutionally irrational, and reaching an equivocal conclusion, Part II continues by examining whether ENDA is sufficiently limited so as to prohibit only that discrimination which the Court would in fact consider unconstitutional. Assuming that ENDA does in fact go beyond what the Constitution commands, Part II concludes by considering whether ENDA constitutes appropriate prophylactic legislation, perhaps going beyond what equal protection requires but sufficiently related to those limits as to satisfy the Court’s test of “congruence and proportionality.” Part III of this Article takes a more prescriptive tone. Since ENDA has not yet been enacted, its proponents in Congress may well wish to consider Garrett and the other Section 5 cases as they draft new versions of the bill. Part III offers some unsolicited advice for those interested in ENDA’s enactment, suggesting ways in which the bill’s drafting, and especially its fact-finding, may maximize chances for its survival as an appropriate expression of Congress’ Section 5 power. In Part IV, the Article concludes by offering some general thoughts on the Court’s recent Section 5 jurisprudence, based on the results of the foregoing analysis of ENDA as legislation designed to ensure the equal protection rights of gay men and lesbians.40

I.  The Supreme Court’s Section 5 Jurisprudence

A.  The Court’s Section 5 Jurisprudence Before City of Boerne

The history of the modern Court’s interpretation of the Section 5 power is a familiar one and will be recounted only briefly here. In Katzenbach v. Morgan, the Court upheld as a valid exercise of the Sec[*PG8]tion 5 power a provision of the Voting Rights Act that allowed graduates of the sixth grade of accredited Spanish-language schools in Puerto Rico to vote, notwithstanding state English-literacy qualification tests.41 Even though the Court had previously held that such literacy tests did not violate the Equal Protection Clause,42 the Court concluded that the statute was an appropriate use of Congress’ Section 5 power.43 Writing for a five-member majority, Justice Brennan enunciated an extremely deferential test for Section 5 legislation. Quoting directly from McCulloch v. Maryland, the Court held that the scope of the Section 5 power was governed by the same broad test applicable to exercises of Congress’ Article I power: “‘Let the end be legitimate, let it be within the scope of the constitution, and all means which are appropriate, which are plainly adapted to that end, which are not prohibited, but consist with the letter and spirit of the constitution, are constitutional.’”44 In a case decided the same year as Morgan, the Court had enunciated that same standard as governing the scope of Congress’ analogous power to enforce the Fifteenth Amendment.45

Turning to the merits, the Court put forth two justifications for upholding the statute. First, providing voting rights for Puerto Ricans might be thought to assist in ensuring equal responsiveness to their concerns from the political process.46 Second, and more controversially, the Court stated that Congress may reasonably have thought that the English literacy test itself violated the Equal Protection Clause, despite what the Court had itself held several years before.47 This potentially revolutionary grant of interpretive power to Congress elicited a sharp dissent from Justice Harlan48 and has been controversial ever since.49 Still, Morgan laid the foundation for two principles of [*PG9]the Court’s Section 5 jurisprudence: deferential review of the appropriateness of Section 5 legislation and a broad conception of what the Section 5 power allows Congress to do.

Retrenchment was quick in coming. Four years later, in Oregon v. Mitchell, a severely fractured Court was unable to agree on a rationale for upholding Congress’ decision to lower the voting age to eighteen in federal elections, while striking down Congress’ decision to do the same with regard to state elections.50 Justice Black, the fifth vote for both of these holdings, tied his Section 5 analysis to the existence of other textual authority for Congress to act in this area.51 Thus, as other parts of the Constitution gave Congress the power to regulate elections for federal offices,52 Justice Black concluded that Congress had greater power under Section 5 to go beyond what he saw as the Fourteenth and Fifteenth Amendments’ overriding concern with race discrimination.53 By contrast, he was not willing to grant Congress such broad power under its authority to “enforce” the Fourteenth or Fifteenth Amendments when it was acting in an area not directly addressed by those provisions.54 Justice Harlan, writing for himself only, would have struck the statute down as applied to both federal and state elections.55 He agreed that the Fourteenth Amendment did not address the issue of voting rights56 and questioned Morgan’s deferential standard for reviewing Section 5 enactments.57 Justice Stewart, joined by two other Justices, also would have struck down both applications of the statute.58 He agreed with Justice Black that the Constitution gavethe statesthe authority to regulateelections and refused to

[*PG10]give Morgan a broad reading.59 Thus, in Oregon, five Justices questioned the scope of the Section 5 power as construed in Morgan.60 Nevertheless, Oregon remains a weak precedent given its fractured nature and the special circumstances of the case, namely, Articles I and II’s arguably clear delegation of the election regulation power to the states and the explicit conclusion of at least two Justices that the Fourteenth Amendment simply had nothing to say about age-based discrimination in voting.61

Ten years after Oregon, the Court in City of Rome v. United States upheld Congress’ power under the enforcement provision of the Fifteenth Amendment to prohibit changes in state election laws that had either the purpose or the effect of diluting black voting strength.62 The same day the Court decided City of Rome, it also held that state action, in order to violate the Fifteenth Amendment, had to evince purposeful discrimination.63 In allowing Congress to prohibit a practice the Court had just held to be constitutional, the City of Rome Court reasoned that Congress might have found that disparate impact suggested discriminatory intent but that such intent might be difficult to prove.64 Thus, according to the Court, Congress had the power to craft a prophylactic rule—what Laurence Tribe has described as “almost a rule of evidence”65—that prohibited electoral changes with racially disparate impact because so many of those changes might [*PG11]have been motivated by racial discriminatory intent.66 This rationale can be understood as a “wider net” theory, in which a wider net of conduct might be prohibited than that which was actually unconstitutional in order to be sure to catch all actual constitutional violations. According to the City of Rome Court, such a rationale derived naturally from Congress’ power to enforce the Fifteenth Amendment.67

Thus, in 1997 when the Court decided City of Boerne v. Flores it had before it a tradition of broad readings of Congress’ enforcement power. Morgan had been the high-water mark, with its intimation that Congress had the power to interpret the Fourteenth Amendment.68 But even leaving that particular holding aside, the general standard of deference enunciated in Morgan and South Carolina v. Katzenbach, and the wider net concept endorsed in City of Rome, all pointed to congressional power to go beyond the actual guarantees of the Fourteenth Amendment when crafting remedial or deterrent legislation and broad discretion to decide on the need for such legislation. City of Boerne, however, ushered in a new era.69

B.  The Change Wrought by City of Boerne

Starting with City of Boerne, the Court began to cut back on the deference it had previously given to Congress’ decisions to use its Section 5 power.70 In City of Boerne, the Court struck down RFRA, which purported to enforce the Due Process Clause’s guarantee of free religious expression, incorporated from the Free Exercise Clause of the First Amendment.71 RFRA accomplished this by prohibiting state and local governments from substantially burdening religious exercise, even via a generally applicable law that did not single out religious expression, unless the burden was justifiable under the strict scrutiny test of narrow tailoring and a compelling government interest.72 The strict scrutiny test imposed by RFRA was similar to the Court’s own [*PG12]rule for deciding free exercise cases between 1963 and 1990, when it replaced it with a test that was much less demanding.73

In striking down RFRA, the Court enunciated principles that it has since applied several times when deciding the scope of the Section 5 power. Most importantly, the Court required, for the first time, that legislation defended as an exercise of the Section 5 power reflect a “congruence and proportionality between the injury to be prevented or remedied and the means adopted to that end.”74 The Court held that RFRA failed this test.75 First, in discussing congruence,76 it concluded that the legislative record revealed no examples of laws enacted because of religious bigotry, which the Court identified as the basic evil sought to be prevented by the Free Exercise Clause.77 The record did reveal examples of laws burdening religious exercise, but these were only tangential to the value the Court had identified in the Free Exercise Clause and thus the value Congress was authorized to protect through its Section 5 power.78 The Court also held that RFRA was not proportional to any violations that might exist, since it was not the case that many actions thereby prohibited would have been unconstitutional.79 Thus, the Court distinguished RFRA from statutes such as the Voting Rights Act, which was upheld in cases such as City of Rome on the “wider net” theory discussed above.80 Given this lack of proportionality, the City of Boerne Court concluded that RFRA was [*PG13]simply an attempt to interpret the Fourteenth Amendment, not to enforce it.81

Thus, City of Boerne establishes several points about the Court’s modern Section 5 jurisprudence. Most basically, that jurisprudence entails much more careful judicial scrutiny than previously needed. In particular, the Court now requires some measure of fit between statutory means and the constitutional violations supplying their underlying rationale. This fit must include at least some evidence of violations of the underlying constitutional right. For example, in City of Boerne the Court noted that Congress did not find examples of states engaging in religious bigotry (the underlying value in the Free Exercise Clause, at least under Employment Division, Department of Human Resources of Oregon v. Smith).82 This “fit” requirement also seems to demand at least some rationale for the scope of the remedial statute that is linked to violations of the underlying right. In City of Boerne, the Court noted that most practices outlawed by RFRA would probably be constitutional and thus criticized the statute for sweeping too broadly.83 These two sides of the “fit” requirement are closely related, of course: If the statute addressed an area where there was little evidence of unconstitutional conduct, it would follow that the statute would also be overly broad in relation to those underlying violations.

C.  Florida Prepaid and Kimel: Variations on the Theme

In Florida Prepaid v. College Savings Bank and Kimel v. Florida Board of Regents, the Court reaffirmed the approach it crafted in City of Boerne.84 The circumstances of each case, however, led the Court to apply that approach in slightly different ways.

[*PG14]1.  Florida Prepaid

The issue in Florida Prepaid was Congress’ amendment to federal patent law, making states liable for infringement suits.85 The Court held that the law was not an appropriate use of Congress’ Section 5 power.86 The Court made explicit what it had suggested in City of Boerne; namely, that Section 5 requires Congress both to identify the conduct actually violating the Fourteenth Amendment (in this case, the Due Process Clause) and to tailor its legislative scheme to remedying or preventing such violations.87 Applying these requirements, the Court in Florida Prepaid concluded that Congress had failed to identify such violations, which the Court defined as not just patent infringements, but patent infringements that deprived the patent holder of property88 without due process.89 Examining the record, the Court concluded that “Congress appears to have enacted [the statute] in response to a handful of instances of state patent infringement that do not necessarily violate the Constitution.”90 The Court also found fault with the proportionality of the statute, concluding that it was not tailored to focus on unconstitutional conduct, but instead made states liable whenever they infringed on a patent, regardless of, for example, whether the infringement was merely negligent (and thus not a “deprivation” of property) or whether the state provided remedies for deprivations (and thus did not fail to provide “due process”).91 Tellingly, the Court distinguished earlier precedent, namely, South Carolina, on the ground that the statute upheld in that case did include various [*PG15]limits and thus was proportional to the violations it attempted to remedy.92

2.  Kimel

Kimel, decided in 2000, was the first case after City of Boerne in which the Court faced a statute that implicated the Court’s three-tiered, class-based analysis under the Equal Protection Clause.93 In Kimel, a state challenged the Section 5 authority for the Age Discrimination in Employment Act (ADEA).94 Again applying the “congruence and proportionality” requirements to strike down the law, the Court began by noting that age classifications challenged as violations of equal protection receive only rational basis review.95 Indeed, the Court noted that it had never concluded that a state’s age discrimination violated the Equal Protection Clause.96 Given its reluctance to strike down such discrimination, the Court concluded that application of the ADEA to the states violated the proportionality requirement.97

But the fact that age discrimination receives only rational basis review by the Court, with the Court having rejected every such age discrimination claim it heard, did not mean the end of the case. Instead, the Court at least considered the plaintiffs’ argument that limitations on ADEA liability meant that the statute prohibited only age discrimination that was so unreasonable as to be unconstitutionally irrational or at least so unreasonable such that the statute could be considered proportional to the underlying constitutional violation.98 The Court rejected this argument, concluding that the ADEA’s exceptions and limitations were sufficiently narrow such that the statute did in fact prohibit a broad swath of conduct that would survive rational basis scrutiny in a constitutional challenge.99 The Court’s rejection of the plaintiff’s argument, however, is less important than the fact that they considered it serious enough to warrant several pages of statutory analysis. Such consideration suggests that the Court seems to have recognized that conduct could still theoretically be unconstitu[*PG16]tional and thus appropriately proscribable by Congress acting pursuant to Section 5, even if it was judged only under the rational basis standard.100

Finally, the Court considered the possibility that the ADEA’s applicability to the states might have been justified as a response to a “difficult” or “intractable” problem.101 The Court seems here to be indicating that unconstitutional age discrimination might be a serious problem, difficult to prove, or otherwise impervious to correction, thus requiring an aggressive legislative response. In considering that possibility, the Court examined the legislative record to determine whether there was in fact a significant problem with states engaging in unconstitutional age discrimination.102 Reviewing the legislative record for examples of such discrimination engaged in by states, the Court concluded that the record did not reveal such a problem, dismissing the plaintiffs’ evidence of congressional concern about age discrimination by state government as “isolated sentences clipped from floor debates and legislative reports.”103

Thus, in Kimel, the Court presented the situation as one combining a deferential judicial review standard for age discrimination with a broad-based legislative prohibition on such discrimination unaccompanied by evidence that Congress perceived a significant problem with state government age discrimination. This picture strongly suggested that the Court would be skeptical of any Section 5 legislation addressing discrimination against groups that do not receive heightened judicial protection, at least in the absence of significant legislative evidence that a constitutional problem does in fact exist. Before that suggestion could solidify, however, the Court had to confront one additional fact pattern.

[*PG17]D.  Garrett and the Problem of Cleburne

Board of Trustees of the University of Alabama v. Garrett completes the set of general situations the Court could face when confronting a Section 5 statute.104 In Garrett, the Court, by the same 5–4 majority that decided Kimel and Florida Prepaid, struck down the application of Title I of the ADA to the states, to the extent that statute authorized private party lawsuits seeking damages against states that had not consented to federal jurisdiction.105 Applying the congruence and proportionality test, Chief Justice Rehnquist began by noting that state discrimination against the disabled was subject only to the rational basis standard, which allowed the disabled to be treated differently if there were any rational reason for doing so.106 So far, this was no different from the analysis in Kimel, which noted that the same deferential review applied to age classifications.107 But in Garrett, the Court confronted a challenge beyond that faced in Kimel. Unlike age classifications, the Court had at least once struck down an instance of disability discrimination as failing the rational basis standard.108 Thus, while the lack of suspect class status (and the attendant heightened review) may have made it very difficult for Congress to justify legislation benefiting that group, the Court in Garrett had to confront the fact that on at least one occasion, it had found discrimination against the group to be so unreasonable as to fail the rational basis test. The question then arose: Given that the Court itself had found such unconstitutional discrimination against the group, would not application of the ADA to the states “enforce” against such discrimination?

The majority’s response to this proposition, however, was not encouraging. While it noted and reaffirmed City of Cleburne v. Cleburne Living Center’s conclusion that negative attitudes or fears alone could not justify government action, the Court in Garrett described its holding in Cleburne as resting on standard rational basis review, as opposed to the less deferential type of review commentators (and other mem[*PG18]bers of the Court) had seen in Cleburne.109 By reading Cleburne as the exceptional situation where standard rational basis review required that a statute be struck down, the Court laid the foundation for concluding that there was no pattern and practice of such unusually irrational discrimination. In turn, that conclusion led it to hold that the ADA was not congruent and proportional to such violations that did exist and thus was invalid as Section 5 legislation.

Again, as in Kimel, the Court considered the evidence Congress had marshaled regarding the constitutional problem posed by disability discrimination.110 In examining the legislative record, Chief Justice Rehnquist started by severely limiting the scope of the data it was willing to consider.111 In particular, the Court insisted on excluding not just examples of private discrimination but also discrimination performed by units of local governments, on the theory that such units did not enjoy the protection of the Eleventh Amendment and thus could be sued for retrospective relief without Congress having to use its Section 5 authority.112

Turning to actual examples of state government discrimination against the disabled, the Court noted that the record included examples of “half a dozen” instances of state government discrimination.113 It found these instances to be insufficient, observing that it was not clear whether those acts of discrimination were unconstitutionally irrational.114 The Court noted that Congress had failed to make a formal legislative finding that states were acting unconstitutionally and similarly failed to state such a conclusion in the committee reports on the ADA.115 It also noted that accounts of state discrimination against the disabled cited in Justice Breyer’s dissenting opinion were submitted not directly to Congress, but to the Task Force on the Rights and [*PG19]Empowerment of Americans With Disabilities, which had also failed to make findings about unconstitutional conduct by the states.116

Moving on to the actual content of the ADA, the Court concluded that the statute’s provisions went beyond what was required under the rational basis standard, rendering the ADA a disproportionate response to whatever constitutional problem may exist.117 It noted that the ADA required employers to make “reasonable accommodations” for disabled individuals otherwise able to perform their jobs118 and concluded that failure to provide those accommodations might be perfectly rational, though cruel—in the Court’s words, “hardheaded” though perhaps “hardhearted.”119 The Court found similar problems with other provisions of the ADA, including its placing of the burden on the employer to show that a requested accommodation would constitute an undue burden (and thus not required under the statute)120 and its prohibition of standards that had a disparate impact on the disabled.121 According to the majority, these provisions, like the “reasonable accommodation” requirement, went beyond what the Constitution required, all in the context of an area where the Court was unconvinced that Congress had demonstrated the existence of a constitutional problem.

The Court concluded by comparing the ADA, once again,122 with the provisions of the Voting Rights Act upheld in South Carolina.123 The Court described those provisions as a “detailed but limited remedial scheme” applicable “in those areas of the Nation where abundant evidence of States’ systematic denial of [constitutional voting] rights was identified.”124 It pointed to the careful examination of the issue that Congress undertook before enacting the Voting Rights Act and to [*PG20]its documentation of “a marked pattern of unconstitutional action by the States.”125

Concurring, Justice Kennedy, joined by Justice O’Connor, lauded the purposes of the ADA, but suggested that state violations of the ADA might not have their base in “embod[iments of] the misconceived or malicious perceptions of some of their citizens.”126 He noted that allegations of unconstitutional disability discrimination had not been made to federal or state courts and agreed with the Court that Congress had also failed to supply such a record.127

E.  The Current Law of Section 5

After City of Boerne, Florida Prepaid, Kimel, and Garrett, certain propositions about Congress’ Section 5 power seem clear. First, the Court is shying away from the idea expressed in Morgan that Section 5 authorizes Congress to engage in its own interpretation of the Fourteenth Amendment and that congressional interpretation is owed deference by the Court.128 Second, Morgan’s general level of deference to Congress’ Section 5 authority—expressed as McCulloch’s broad formulation of judicial deference to legislative judgments concerning the need for and breadth of Article I-justified regulation—may be less secure, in light of both the Court’s careful review regarding whether constitutional violations exist that justify Section 5 legislation and its requirement that such legislation be limited in scope so as to correspond to the constitutional violation it seeks to address.129 The recent cases’ consistent references to South Carolina and their descriptions of the Voting Rights Act provisions upheld in that case as carefully limited suggest as much.130

The Court has also indicated that it will closely scrutinize statutes benefiting groups that do not enjoy heightened judicial protection under the Equal Protection Clause. This seems true even if, as in [*PG21]Garrett, the statute benefited a group that the Court had found in the past to have been the victim of unconstitutionally irrational discrimination.131 Again, this close scrutiny will entail an examination of whether the scope of the statute is limited in a way corresponding to the constitutional violations found.132 On the other hand, the Court has continued to endorse the “wider net” theory of cases like City of Rome, allowing Congress to prohibit broader swaths of conduct than would be forbidden under the Fourteenth Amendment itself as long as there is a sufficient factual record supporting the need for the broader legislation.133 The Court has also continued to insist that it respects Congress’ determinations about what is needed to guarantee Fourteenth Amendment rights.134

As a rough description, then, the current Court requires Congress to do more by way of fact-finding before it will uphold statutes as valid uses of the Section 5 power. The Court also requires Congress to take more care to limit legislation so as to correspond more closely to the constitutional violations it seeks to address. In essence, the careful review the Court now gives to Section 5 legislation shifts the presumption away from Congress, requiring it, rather than the state, to make out a case for its use of that power.135 The next part of this Article applies these observations to ENDA in an attempt to determine whether the Court would uphold its applicability to the states as appropriate Section 5 legislation.

II.  Garrett, Romer, and ENDA

At first glance, it might appear that Board of Trustees of the University of Alabama v. Garrett sounds the death knell for any authorization [*PG22]in ENDA for state employees to sue their employers for retrospective relief.136 In both Kimel v. Florida Board of Regents and Garrett, the Court was quite skeptical of the Section 5 basis for federal legislation benefiting groups that the Court had not previously favored with suspect class status. Garrett further suggests that an occasional decision striking down a statute harming a particular group, on the ground that the statute failed the rational basis test, would not be of much help in saving a Section 5-based statute benefiting that group. This latter fact makes it even more doubtful that the Court would find ENDA an appropriate enforcement of the constitutional rights found to be violated in Romer v. Evans137 given that decision’s well-known ambiguity.138

This part of the Article considers the constitutionality of ENDA as Section 5 legislation. It starts by summarizing ENDA’s most important provisions. It then identifies and describes the constitutional violation resulting when a state engages in employment discrimination against gay men and lesbians. The Article then considers whether there is a pattern of such violations and then asks whether ENDA is directly targeted at those violations or, if not, whether it is a congruent and proportional response to them.


ENDA’s basic provisions are straightforward.139 ENDA would make it unlawful “to fail or refuse to hire or to discharge any individual, or otherwise to discriminate against any individual with respect to the compensation, terms, conditions, or privileges of employment of the individual, because of such individual’s sexual orientation.”140 This terminology tracks closely the anti-discrimination language in Title VII of the Civil Rights Act of 1964, which prohibits employment [*PG23]discrimination on the basis of race or gender.141 In an important contrast to Title VII, however, ENDA does not allow an employee to make out a discrimination claim based on an employment practice that has a disparate impact relative to sexual orientation.142 ENDA explicitly waives state sovereign immunity143 and provides for retrospective relief, although, importantly, it does not allow for back pay awards as a component of compensatory damages.144 ENDA also includes a number of other important limitations. In addition to the prohibition on back pay awards, ENDA limits relief by barring courts, as part of a relief order, from ordering quotas or preferential treatment for gays.145 [*PG24]It “does not apply to the provision of employee benefits to an individual for the benefit of the domestic partner of such individual.”146 ENDA does not apply to religious organizations,147 the military,148 or employers employing fewer than fifteen persons.149 It also allows employers to enforce rules regarding “nonprivate sexual conduct” if such rules apply equally regardless of sexual orientation.150 Enforcement is analogous to enforcement under Title VII.151

B.  Determining the Fourteenth Amendment Violation

In order to determine whether ENDA is “appropriate” Section 5 legislation, the first step is to determine the scope of the constitutional violation inhering in state government employment discrimination on the basis of sexual orientation. As the Court in Garrett noted, under Section 5 “Congress is not limited to mere legislative repetition of [the Supreme] Court’s constitutional jurisprudence,”152 but instead has the power “both to remedy and to deter violation of rights guaranteed [by the Fourteenth Amendment] by prohibiting a somewhat broader swath of conduct, including that which is not itself forbidden by the Amendment’s text.”153 The Garrett Court continued by noting that “legislation reaching beyond the scope of Section 1’s actual guarantees must exhibit ‘congruence and proportionality between the injury to be prevented or remedied and the means adopted to that end.’”154

This statement of the law suggests that different inquiries may be appropriate for legislation that simply repeats the constitutional limits on state action the Court has found inhere in the Fourteenth Amendment and for legislation that prohibits conduct beyond those limits in order to remedy or deter unconstitutional conduct. The first [*PG25]of these conceptions of the Section 5 power is hardly exceptional. If the Section 5 power means anything, it must mean that Congress has the power to prohibit conduct in general that the Court has found, in a particular case, to violate the Constitution, as well as the power to create remedies for such unconstitutional conduct. The narrowest conception of the Section 5 power must include the power to prohibit as a general rule what the Supreme Court has declared to be unconstitutional in the context of a particular case. Indeed, given the Court’s statement in Cooper v. Aaron that a principle of constitutional law enunciated by the Supreme Court is itself the “supreme Law of the Land,” such a power is close to superfluous.155 The congressional power to create remedies for such unconstitutional conduct is only slightly broader. This power too must be uncontroversial, as a matter of textual interpretation, if Section 5’s explicit grant of power to “enforce” the Fourteenth Amendment is to have any meaning.

Thus, logically the task is simply to determine what sorts of sexual orientation discrimination the Court has already found to violate the Fourteenth Amendment and compare that invidious discrimination to the conduct outlawed in ENDA. Because disability discrimination has a constitutional status quite analogous to discrimination on the basis of sexual orientation, Garrett provides a useful guide for how the Court may approach that inquiry.

As with sexual orientation discrimination, the Court has on one occasion, City of Cleburne v. Cleburne Living Center, struck down a government action discriminating against the disabled on the ground that it failed the rational basis test.156 The Garrett Court characterized Cleburne as a case where the Court simply applied standard rational basis review and struck the action down as reflecting “mere negative attitudes, or fear, unsubstantiated by factors which are properly cognizable in a zoning proceeding.”157 The Court then noted that in the ADA Con[*PG26]gress had failed to provide sufficient evidence that states were engaging in a pattern of discrimination motivated by such negative attitudes, divorced from proper government ends.158

Applying this first step of the analysis to the situation posed by sexual orientation discrimination, the challenge will be to characterize the constitutional violation in Romer, the one case in which the Court has found anti-gay discrimination to be unconstitutional.159

C.  What is the Equal Protection Violation in Romer?

Commentators have described Justice Kennedy’s opinion in Romer in a variety of ways: a reflection of an “anti-caste” principle inherent in the Equal Protection Clause,160 a gloss on the Bill of Attainder Clause,161 an example of heightened scrutiny under the more stringent version of the rational basis test,162 and the first step toward granting homosexuality suspect class status.163 Assuming that a future Court dealing with a challenge to ENDA would not read into Romer the broader, more speculative content suggested by the “anti-caste” theory or the quasi-Bill of Attainder analysis,164 and assuming further [*PG27]that it would not use Romer as the jumping off point for declaring sexual orientation a suspect classification,165 the Court would be left with the need simply to take Romer at its word, for what it actually said, rather than for what it implied or for what it should logically lead to.166 Thus, a close reading of Romer seems appropriate.

Romer considered a challenge to Amendment 2, an amendment to the Colorado Constitution enacted by the people of Colorado in a referendum election in 1992.167 Amendment 2 stated, in relevant part:

[N]either the State . . . nor any of its . . . subdivisions . . . shall enact . . . or enforce any statute . . . or policy whereby homosexual . . . orientation [or] conduct . . . shall constitute or otherwise be the basis of . . . any protected . . . status or claim of discrimination . . . .”168

The majority’s analysis of this provision was surprisingly terse, taking up little over four pages in the U.S. Reports.169 The Court cited two [*PG28]reasons for striking down Amendment 2. First, it concluded that Amendment 2 “confounds” normal rational basis review.170 According to the Court, Amendment 2 was different from laws upheld under the rational basis standard because those laws were “narrow enough in scope and grounded in a sufficient factual context for us to ascertain [that there existed] some relation between the classification and the purpose it served,” thus ensuring “that classifications are not drawn for the purpose of disadvantaging the group burdened by the law.”171 By contrast, Amendment 2’s extreme combination of a group, identified by a single trait, which is then denied protection across the board via a broad-ranging disability,172 results in the “disqualification of a class of persons from the right to seek specific protection from the law,”173 making it “in general more difficult for one group of citizens than for all others to seek aid from the government.”174 According to the Court, this sort of broad-based denial of government assistance, based on a single trait, constituted a violation of equal protection in its most literal sense.175

The Court’s second reason for striking down Amendment 2 was much more prosaic and, perhaps for that reason, easier to express. The Court concluded that the broad-based nature of the disabilities Amendment 2 imposed on gays and lesbians exceeded any legitimate purpose the government might have had, thus leaving animus as the only explanation for its enactment.176 The Court noted that Colorado defended Amendment 2 as a means of respecting the associational rights of landlords and employers that might object to homosexuality and as a way the state could conserve its resources to fight discrimination against other groups.177 The Court found these justifications “impossible to credit” given how much further Amendment 2’s burdens ran.178 The Court concluded that Amendment 2 violated a “conventional and venerable” principle of equal protection: Namely, that a [*PG29]law must bear a rational relationship to a legitimate government interest.179

Leaving aside the more speculative theories about Romer,180 for our purposes the unifying theme in the Court’s rationale is the lack of legitimate justification—and thus the irrationality—of the classification given the scope of the burdens Amendment 2 placed on gays and lesbians.181 The Court’s second justification, that Amendment 2 simply failed standard rational basis review, clearly reflects this concern, as the imbalance between Amendment 2’s ends and means raised the suspicion that the provision was motivated by another, illegitimate, motivation. The Court’s first justification also reflects the concern about legitimate justification, as the Court was simply suspicious that any enactment so broad could serve a legitimate government interest.

Thus, in determining whether ENDA simply targets or provides a remedy for discrimination of the sort condemned in Romer, it might be helpful to begin by considering the rationality (in the constitutional sense) of sexual orientation discrimination in state government employment.

D.  Is ENDA Targeted at Unconstitutionally Irrational Sexual
Orientation Discrimination?

The strongest argument distinguishing the ADA from ENDA may be that employment discrimination against gays and lesbians is simply irrational in a way that various types of discrimination against the disabled are not. In both Cleburne and Romer, the Court struck down state action as violating the Equal Protection Clause, even though discrimination against the burdened group—respectively, the mentally disabled and homosexuals--ostensibly received only rational basis scrutiny.182 In both cases, the Court concluded that the government action was motivated by animus or fear—that is, by reasons that are illegitimate for purposes of the rational basis test, which requires a [*PG30]rational relationship to a legitimate government purpose.183 In Garrett, the Court concluded that a state’s refusal to provide the accommodations required by the ADA might not be motivated by such an illegitimate purpose, but instead by a rational (in the Court’s words, “hardheaded”)184 desire to save money, as cruel (“hardhearted”)185 as that decision may be. Although left unsaid, the Court’s analysis clearly implies that such a refusal to take the steps required by the ADA might be irrational in a broader sense—given either the value disabled workers could produce if provided an accommodation or, even more generally, the net social benefit of making it possible for disabled individuals to work and participate in society.186 However, the standard rational basis test leaves to the state the balancing of these costs and benefits.

Would the same analysis apply to employment discrimination against gays? In other words, is there a “hardheaded” cost-benefit balancing that would go into a state decision to engage in employment discrimination on the basis of sexual orientation? There does not seem to be, for the obvious reason that a rule of non-discrimination on the basis of sexual orientation would not necessarily require states to spend funds to modify most workplaces to take account of or accommodate differing sexual orientations.187 The one possible way in which employing gays and lesbians may in fact require extra costs among states would be if other workers’ resistance or hostility to the presence of a gay co-worker required the state to shift personnel or otherwise alter the workplace, thereby increasing the state’s costs.188

Is it otherwise rational to engage in workplace discrimination on the basis of sexual orientation? This is, of course, an important question to ask for the Section 5 analysis. Indeed, it would be dispositive if the answer is “no.” The Court in Garrett stated that “Section 5 legislation reaching beyond the scope of Section 1’s actual guarantees must exhibit ‘congruence and proportionality between the injury to be [*PG31]prevented or remedied and the means adopted to that end.’”189 Thus, congruence and proportionality seem to be unnecessary if the legislation directly aims at actual Fourteenth Amendment violations.190 This Article now considers whether sexual orientation discrimination of the type prohibited in ENDA can ever be rational.

1.  Status and Conduct, “Don’t Ask, Don’t Tell” and Bowers v. Hardwick

Perhaps surprisingly, relatively few courts have considered the constitutionality of government employment discrimination against gays and lesbians, at least outside of the realm of discharges from military service. Some employment dismissals are defended on the theory that the plaintiff’s expression of her sexual orientation, or her acting in a way consistent with it (say, by participating in a same-sex marriage ceremony), constituted disruptive conduct that justified the dismissal completely apart from the plaintiff’s sexual orientation.191 Relatedly, in cases dealing with law enforcement jobs or those involving security clearances, courts have held that government has the power to exclude on the basis of whether the individual has violated the law or lied when asked about his sexual orientation.192 Strictly speaking, these cases are not relevant to the question of whether sexual orientation itself—as a status, divorced from conduct—is a constitutionally permissible ground for firing or not hiring someone. They do, however, thereby raise a subsidiary question: To what extent is sexual orientation relevant in that homosexual or bisexual orientation allows the government rationally to presume homosexual conduct?193 A state [*PG32]whose law criminalizes sodomy might well argue that it is rational for it to engage in employment discrimination against homosexuals because they have a propensity for violating the law, even though such a classification might be seriously under or over inclusive.194

The main battleground on which this status versus conduct argument has been played out has been the military under the “Don’t Ask, Don’t Tell” policy.195 Under that policy, the military can discharge anyone who engages in homosexual acts or who states that they are gay or lesbian unless they can prove that they do not have a propensity to engage in homosexual acts.196 Thus, while seemingly only proscribing conduct (engaging in homosexual acts or having a propensity to do so), the policy, according to some commentators, effectively proscribes status since it is relatively easy for the military to argue that the status of being homosexual indicates a propensity to engage in homosexual acts.197 This logic has led courts to uphold the policy.198

If the Court were to follow such a path with regard to ENDA, it might well strike the statute down as going beyond unconstitutional sexual orientation discrimination on the ground that a state would have a rational reason to discriminate on the basis of sexual orientation to the extent that sexual orientation indicated a propensity to violate the law, that is, to commit a sex crime.199 While a propensity to [*PG33]violate the law might be more relevant to some state jobs than others—for example, more relevant to a state trooper position than to a clerical one—a state’s desire to have law-abiding employees presumably would be considered legitimate. Nevertheless, it should not be taken for granted that a sexual orientation job criterion would be sufficiently linked to that goal as to survive rational basis scrutiny.200 To the extent that state sex crime laws simply prohibit certain types of sex, regardless of the gender of those engaged in it, there is evidence that heterosexuals engage in such illegal sexual conduct at least as frequently as homosexuals.201

Given cases upholding classifications with only the most tenuous relation to legitimate interests,202 the Court would have to apply a more stringent type of rational basis review in order to strike down such a statute on this theory. On the other hand, to the extent that a state’s sex crimes law prohibits such conduct only when engaged in between two persons of the same sex, such a prohibition may itself be constitutionally problematic as a matter of equal protection.203 Without a rational reason for singling out gays and lesbians, Romer provides at least some hope that the current Court would strike such legislation down, or at least not credit it as providing a state with a legitimate justification for discriminating against gays and lesbians in employment.

Thus, a state may not have a good argument that it should be able to engage in sexual orientation-based employment discrimination because sexual orientation is a marker for a propensity to engage in illegal conduct that the state can constitutionally seek to prevent in its workforce. Nevertheless, Hardwick remains a hurdle, requiring the Court not only to disentangle the concepts of status and conduct, but to decouple the two by refusing to credit the argument that sexual [*PG34]orientation is a legally sufficient marker for same-sex conduct.204 It is, however, exactly that sort of careful analysis that would require the Court to engage in a more searching review than suggested by the standard rational basis test.205 To the extent that such unusually searching “rational basis” review is necessary, it might militate against the Court’s willingness to hold that sexual orientation employment discrimination is irrational.

2.  Pure Status Arguments

On the other hand, some courts have focused their analysis squarely on the question whether sexual orientation discrimination in employment is constitutional. For example, in Weaver v. Nebo School District,206 a federal district court granted summary judgment to a lesbian high school teacher/athletic coach who was removed as a coach after she disclosed her orientation in response to a student’s question and thus according to school officials, caused controversy. The court explicitly rejected as insufficient grounds for the school’s action the community’s supposed negative response to her disclosure, “[i]f the community’s perception is based on nothing more than unsupported assumptions, outdated stereotypes, and animosity.”207 According to the court, the only appropriate justification for removing the plaintiff would have been one related to her job performance.208

If community dislike is an inappropriate ground for employment discrimination, presumably it would be insufficient for a state to justify employment discrimination on the ground that co-workers would resist or be hostile to a gay or lesbian colleague. In both situations, the employer would have to incur costs in order to ensure a smoothly [*PG35]functioning workplace: In Weaver, the school would have to spend time, and perhaps money, in responding to parental complaints, educating both parents and students, and generally calming the situation and ensuring physical safety. This is exactly what the state as employer would have to do if the complaining or resisting parties were co-workers.209 If such costs, or the general tension caused by the mixture of gay and homophobic individuals, are not considered rational reasons to allow the state to discriminate in Weaver, they should not be considered any different when the complaining or resisting parties are co-workers. Thus, assuming that there is no other “hardheaded” reason for discriminating against a lesbian or gay person, and further assuming that the removal was not for reasons of conduct as opposed to mere orientation,210 there would appear to be no room for morality-based community disapproval to justify discrimination.211

It should be noted that this argument does not deny the more general role of morality in law. Indeed, as a doctrinal matter, the Supreme Court in Hardwick made it clear that morality is a constitutionally appropriate justification for prohibitions on at least some conduct.212 But, decisions such as Weaver suggest that morality cannot [*PG36]serve as a rational basis justifying different treatment of groups based on status such as sexual orientation. Indeed, combining Hardwick and Romer seems to require that conclusion. In Hardwick the Court explicitly stated that morality may be a valid government interest for banning certain conduct,213 while Romer held Amendment 2 unconstitutional in part because it concluded that it was based on animus toward homosexuals.214 But even if harmonizing Hardwick and Romer does not require this conclusion, it nevertheless surely seems a reasonable reading of the Equal Protection Clause, especially in light of other cases prohibiting the criminalization of statuses some of whose characteristic conduct can be proscribed.215 And if such a conclusion seems reasonable, then there is all the more reason to view ENDA as a confirmation of the equal protection guarantee, rather than a rewriting of it.216

3.  The Significance of ENDA’s Limits

a.  Reading Romer as Unique

The problem with the foregoing analysis is that the Court may simply cut it off by pointing to the uniqueness of the burdens Amendment 2 placed on gays and lesbians. Certainly there are sufficient references in Romer to Amendment 2’s novelty to allow the Court to distinguish it from more targeted discrimination prohibited in ENDA.217 Amendment 2’s uniqueness affected the Romer Court’s analysis in two very different ways. The Court’s first rationale for striking down Amendment 2 did in fact focus on the law’s uniqueness, as [*PG37]the Court held that Amendment 2’s very breadth constituted a literal violation of the equal protection guarantee.218 However, Romer’s second rationale was much more prosaic: The Court held that the law simply lacked a rational link to any legitimate justification.219 Of course that lack of linkage was suggested by Amendment 2’s unique breadth; nevertheless, the principle underlying the Court’s second rationale was, as the Court noted, “conventional and venerable.”220

It cannot seriously be argued that employment discrimination is a burden so broad as to constitute a literal violation of equal protection as was Amendment 2. Thus, in order for Romer to support the proposition that sexual orientation-based employment discrimination is constitutionally irrational, it may be necessary to make the more conventional argument that the sexual orientation-based employment discrimination prohibited in ENDA simply fails the standard rational basis test. Much of this argument has been made above;221 however, one piece remains. This sub-part of the argument considers the limits on ENDA’s anti-discrimination rule and examines whether those limits succeed in narrowing its scope to government conduct that is most arguably unconstitutional.

b.  ENDA’s Limitations

Several provisions of ENDA limit the scope of its prohibition on sexual orientation-based employment discrimination. First, ENDA makes proving sexual orientation discrimination harder than proving discrimination based on other criteria by providing that disparate impact is insufficient to make out a claim of discrimination.222 Not only does this limitation make it harder to prove an ENDA claim than, say, a Title VII-based claim,223 but, more importantly, it links ENDA to the intent requirement of the Equal Protection Clause.224 Even more, making mere disparate impact insufficient makes successful ENDA claims more reflective of the Court’s concern in Romer—namely, that [*PG38]Amendment 2 was motivated by animus, a conscious desire to burden a person based on a particular trait. Thus, while tying the ENDA standard to the violation in Romer makes an ENDA claim harder to prove, the claims that do survive are more likely to reflect the unconstitutional animus that underlay the constitutional problem in Romer.225

Second, ENDA excludes from its purview any claim relating to provision of benefits for spouses or unmarried partners.226 Without this provision, such a claim (whether ultimately successful or not) would flow naturally from a general prohibition against sexual orientation discrimination, since states generally provide benefits only for spouses and children, and since no state recognizes same-sex marriages. By excluding such claims, ENDA focuses more narrowly on employment discrimination concerns completely divorced from any state interest in limiting marriage (or the benefits thereof) to heterosexual couples. The point here is that, unless courts are willing to interpret federal equal protection guarantees as requiring states to make marriage available to same-sex couples, the limiting of marriage to opposite-sex couples will be held, by hypothesis, constitutionally rational.227 If so, then job-related sexual orientation discrimination could also be considered rational if it is justified as a way of preventing gay employees from demanding, as a component of workplace equality, job-related benefits for their partners that “undermine” the special status states accord the marriage relationships the state has reserved for opposite-sex couples. By excluding that issue from its purview, ENDA focuses on the aspect of state conduct—discrimination in the simple act of employment, unrelated to spousal benefits—that does not implicate the state’s interests with regard to marriage.228

[*PG39] Third, other provisions of ENDA, while not directly applicable to states, underscore its overall limited nature. First, ENDA has no application to the military, thus exempting the institution that, because of its uniqueness, presents a different argument for excluding gays and lesbians.229 Second, the statute exempts businesses with fewer than fifteen employees.230 Finally, the statute exempts religious organizations. These latter two limitations reflect a respect for free association and free exercise values embodied in the First Amendment when those values might be most threatened by a non-discrimination rule.231 While a state would not fall under the “small employer” exception and while Establishment Clause concerns presumably mean that a state would not be affected by the exemption for religious organizations,232 these exemptions reflect again the bill’s concern for targeting [*PG40]only that conduct that society considers so worthless as to be irrational. Indeed, these exceptions tie in closely to the justifications Colorado offered in defense of Amendment 2, justifications that, while the Court did not believe (because Amendment 2 went so much further), the Court never said were illegitimate.233

Taken together, ENDA’s limitations narrow the class of prohibited conduct to that which is most arguably unconstitutional. The disallowance of disparate impact claims focuses the statute on conduct that is purposely directed at gays and lesbians and thus most likely to be motivated by unconstitutional animus. The other limitations exclude from ENDA’s purview situations where such purposeful discrimination might be constitutional, either because the state might be thought to have a legitimate interest (limiting its recognition of committed relationships to heterosexual marriages or, in the case of the federal government, recognizing the special needs of the military234) or because a competing private interest has some constitutional stature (such as free religious exercise or a right to associate).

E.  Going Beyond Actual Violations: ENDA as Remedial or
Prophylactic Legislation

Assuming, however, that ENDA does not prohibit only conduct that is unconstitutional—in other words, assuming that states sometimes constitutionally engage in the type of employment discrimina[*PG41]tion ENDA prohibits—the inquiry shifts to whether the statute is a congruent and proportional response to conduct that is in fact unconstitutional. Even under the Court’s newer, stricter test for Section 5 legislation, the Court has continued to accept the appropriateness of legislation that bans a “broader swath” of conduct than that actually prohibited by the Constitution in order to deter violations of the constitutional rule.235 To determine whether such broader legislation represents an appropriate remedy or deterrent, the Court examines whether, first, the legislation is a congruent and proportional response to those violations and, second, whether there is a pattern and practice of such violations that may have evaded the Court’s own eyes.236

1.  Is ENDA Congruent and Proportional?

Under the Court’s post-City of Boerne v. Flores jurisprudence, congruence and proportionality require that there be some relationship between the statute justified under the Section 5 power and the underlying constitutional violation. While this concept is not particularly novel,237 cases starting with City of Boerne have required a tighter fit between the statute and the violation.238 In Kimel, Florida Prepaid v. College Savings Bank, and Garrett, the Court struck down the statutes because they swept too broadly in relation to the constitutional violation that Congress was assertedly enforcing, prohibiting too much state conduct that was in fact constitutional without a showing that states were engaging in significant amounts of unconstitutional conduct.239 In two of the modern cases, the Court contrasted the statutes at issue with provisions of the Voting Rights Act that were upheld because of their limited scope in South Carolina v. Katzenbach.240 These modern [*PG42]cases stand in stark contrast to the deferential standard of review enunciated in Katzenbach v. Morgan, which approached rational basis at least in form and perhaps also in application.241

Is ENDA sufficiently closely related to unconstitutional sexual orientation-based employment discrimination so as to survive this test? In part the answer to this question must derive from the earlier question, whether the discrimination prohibited by ENDA can ever be considered constitutionally rational. If it cannot, that is, if such discrimination is always so irrational as to violate the Equal Protection Clause, then there is perfect congruence between ENDA and the constitutional violations Congress has the power to prevent under even the narrowest reading of Section 5.242 As discussed above, however, it is at least possible that the Court would read Romer sufficiently narrowly (as a holding dictated by the unusually broad scope of Amendment 2)243 and ENDA’s prohibitions sufficiently broadly as to conclude that ENDA did not simply prohibit conduct the Court would itself consider violations of equal protection.244 Thus, this Article now considers ENDA as prophylactic legislation going beyond the actual equal protection guarantee and thereby subject to the congruence and proportionality test.

[*PG43] Most of the ENDA limitations relevant to the congruence and proportionality inquiry have already been discussed above, in the discussion whether ENDA is limited to prohibiting constitutionally irrational government employment discrimination.245 It should be unsurprising that the same factors are relevant to both inquiries: Even if ENDA’s limitations do not exactly track the conduct that the Court would find unconstitutional, those limits would still support an argument that ENDA was carefully tailored so as to remain proportionate to the constitutional violations it sought to remedy. As noted above, the most important of these limits are the prohibition on disparate impact claims and on claims for spousal benefits for same-sex partners. Also significant, if indirectly, are the exemptions for the military, religious organizations, and small employers. Taken together, these limits restrict ENDA’s scope to situations that are quite arguably consistently unconstitutional. But even if ENDA, so limited, continues to prohibit conduct the Court would find constitutional, these limits significantly restrict ENDA’s scope and focus it much more closely on unconstitutional conduct.

In addition to the limits discussed above, two other provisions of ENDA further limit its scope and are especially relevant to the congruence and proportionality inquiry. First, ENDA explicitly refrains from handicapping employers’ enforcement of general codes of sexual conduct that are designed and implemented in a sexual orientation-neutral fashion.246 This provision limits ENDA’s scope by exempting situations where the employer might claim that the adverse employment action was based on sexual conduct, not sexual orientation. By allowing at least some sexual conduct to serve as the basis for an adverse employment action, the exemption goes some distance toward cabining ENDA to cases of pure orientation. This, in turn, keeps the statute responsive (or congruent) to the status/conduct distinction created by the combination of Romer and Hardwick.

Obviously, this provision does not completely track the status/conduct distinction: For a state employer to cite sexual misconduct as the reason for an adverse employment decision, the misconduct must be nonprivate and based on sexual orientation-neutral grounds. Thus, for example, ENDA would still prohibit a state from firing an employee because she had sex with her female partner in the privacy of their home. Indeed, ENDA also would still prohibit a [*PG44]state from firing an employee because the state had a sodomy law restricted to homosexual conduct, since the sexual conduct rule would not be orientation neutral. Thus, the sexual misconduct exemption does not give states complete freedom to cite conduct as the reason for the dismissal. For that reason, the exemption may limit states more than the Constitution (as interpreted in Hardwick) does.247 Still, the exemption serves to keep the statute focused on irrational status-based discrimination and thus proportional to the underlying constitutional violation.

Second, ENDA prohibits affirmative action, quotas, or preferential treatment on the basis of sexual orientation, whether imposed by an employer or a court as part of an order or consent decree.248 While this provision may be a politically savvy response to the argument that anti-discrimination laws confer “special rights,”249 the legal effect of this provision is to prevent a well-meaning employer or court from responding to discrimination by ordering or implementing a remedy that outruns the original statutory violation.250 Thus, not only are ENDA’s limits closely tied to the prohibitions of the Equal Protection Clause, but in turn, implementation of those limits is closely tied to what the statute requires and goes no further.

In sum, several arguments converge to present a plausible argument that ENDA survives the congruence and proportionality test. First, much garden-variety sexual orientation-based employment discrimination has to be considered irrational. Second, ENDA limits its prohibitions to those instances where there is purposeful discrimination against homosexuals and thus a greater likelihood of unconstitutional animus. Third, ENDA limits its scope when it confronts areas where either the state has a potentially legitimate interest (such as “protecting” heterosexual marriage or discouraging certain conduct) or, more indirectly relevant, where private parties may have constitutionally cognizable countervailing interests. Fourth, the statute pro[*PG45]hibits employer or judicial remedies from outrunning the underlying constitutional violation. If a statute may be congruent to constitutional violations in an area the Court itself has not heavily involved itself in, ENDA seems to be it.

2.  Have States Engaged in a Pattern and Practice of Unconstitutional Employment Discrimination Against Gays and Lesbians?

In determining whether a pattern and practice of conduct exists, thus making remedial or prophylactic legislation appropriate Section 5 legislation, the current doctrine demands (1) a pattern of (2) relevant (3) unconstitutional conduct (4) by states.251 In this case, the unconstitutional conduct would be state government sexual orientation-based employment discrimination that is unconstitutionally irrational. Garrett provides the clearest guidance on what the Court might demand. In Garrett, the Court acknowledged the finding in the ADA that discrimination on the basis of disability “continue[s] to be a serious and pervasive social problem”252 and conceded that that conclusion was supported by evidence assembled by Congress.253 However, the Court refused to consider such findings and evidence relevant, since the “great majority” of the incidents cited in the record did not involve states.254 Having limited the field of inquiry to examples of conduct engaged in by state governments, the Court then further limited the field to examples of conduct by state governments in the field of employment.255 Rejecting Justice Breyer’s compilation of state discriminatory conduct against the disabled, the Court stated that “only a small fraction” of the incidents Justice Breyer cited dealt with employment.256 According to the Court, “the overwhelming majority” of those incidents alleged discrimination by states in the provision of public services and public accommodations, areas addressed by other parts of the ADA not challenged in Garrett.257 Finally, the Court questioned whether the examples provided of state government employ[*PG46]ment discrimination reflected unconstitutional conduct on the states’ part.258

After Garrett, then, plaintiffs (and Congress) face a difficult task of compiling a detailed record including examples of the specific violations the statute seeks to remedy. The question, then, is whether the phenomenon of sexual orientation-based employment discrimination is easily susceptible to the compilation of such a record.

The normal way of developing any legislative record would be for Congress to look at the issue, seek out evidence, and find facts based on the evidence it examined. In the case of discrimination, though, the possibility always exists that such an examination will not reveal the true extent of the problem. Indeed, the Court’s Section 5 jurisprudence recognizes this difficulty. Proof problems not only lay behind the Court’s willingness to allow Congress, under Section 5, to prohibit more conduct than required by Section 1259 but also help explain the Morgan Court’s overall deference to legislative judgments regarding the need for particular restrictions on state conduct in order to guarantee Fourteenth Amendment rights.260

It is important to distinguish between these two versions of the proof problem. In Kimel, for example, the Court reaffirmed the concept enunciated in City of Rome v. United States that Congress had the power under Section 5 “to remedy and to deter violation of rights guaranteed [by the Fourteenth Amendment] by prohibiting a somewhat broader swath of conduct, including that which is not itself forbidden by the Amendment’s text.”261 In order to use Section 5 to enact such broader prohibitions, however, the Kimel Court insisted that there be a record of a problem that demanded such a broad remedy and concluded that Congress had created no such record.262 By contrast, the Court in City of Rome reiterated earlier cases’ conclusions [*PG47]that the Voting Rights Act was justified by legislative factfinding about southern states’ attempts to infringe on African-Americans’ voting rights.263 These proof problems come together, though, when considering the appropriateness of prophylactic legislation: If a discrimination problem requires a broad remedy because the discrimination is easy to disguise, it might well be that Congress will have a similarly difficult time uncovering the problem and creating an airtight record of its existence. In such a situation, the Court would be faced with the choice of deferring to Congress’ determination (perhaps without overwhelming evidentiary support) that a problem existed or allowing the very insidiousness of the problem to serve as a shield from remedial legislation.

In the case of age- or disability-based discrimination it is easy to imagine the problem. Such discrimination can be veiled under non-objectionable criteria such as merit or efficiency, thus making the actual violation hard to spot and broader remedial rules acceptable. The problem is magnified if we take a broader view of the harm caused by such discrimination. If that discrimination not only harms those who actually are employed or apply for employment, but also deters many people from seeking work, then the extent of the discrimination may well touch individuals well-hidden from congressional investigation, not to mention as a practical matter unable to sue to enforce their right to equal treatment.264 In such cases, the best factfinding that could be expected would yield only a rough under[*PG48]standing of the scope of the problem based on unreliable and imprecise mechanisms such as surveys (for example, “if you knew that you wouldn’t be discriminated against because of your age/disability, would you seek work?”).265 Nevertheless, the Court in Garrett did not seem moved by these difficulties confronting Congress. In the face of a congressional finding of overall societal discrimination against the disabled266 and the presentation of a large number of instances of their disparate treatment,267 the Court nevertheless faulted Congress for failing to provide specific examples of conduct by states,268 relating to employment discrimination,269 that constituted unconstitutional conduct.270

This evidentiary problem may be even more severe in the context of sexual orientation discrimination. What makes sexual orientation different from age or disability—or almost any other objectionable criterion for discriminating—is that sexual orientation can be hidden. It may be practically impossible for blacks to hide their race, for women to hide their gender, for the elderly to hide their age, or for the disabled to hide their disability, but it is not only possible, but in many situations the norm for gays and lesbians to hide their sexual orientation.271 That fact makes it all the more difficult to compile a reliable record of sexual orientation discrimination, especially in the workplace, where it is normally relatively easy for a lesbian to hide her orientation, and apparently quite common.272 While, for example, a heterosexual woman would not be “coming out” if she complained about discriminatory treatment, a lesbian might well be, and thus would bear an extra burden in disclosing the discrimination, whether [*PG49]to a government investigator, a journalist, or her attorney.273 Nor, contrary to Justice Kennedy’s suggestion in his Garrett concurrence, would prior litigation patterns necessarily reveal the existence or scope of the problem if the offending conduct resulted “only” in a gay person remaining closeted at work.274 Indeed, since only one or two overt acts of discrimination might suffice to keep all gays at that workplace closeted, the scope of the problem might substantially exceed the amount of any litigation.275

Moreover, forcing a gay man or lesbian to stay in the closet at work—on pain of official action such as firing or demotion or unofficial action such as harassment or threatened or actual violence--could very easily be considered discriminatory in itself, to the extent that the workplace tolerates statements of, or openness about, heterosexual orientation.276 Of course, it would be even harder for Congress to construct a record of that kind of evidence, at least in the sense required by the Court up to now—that is, with actual examples of individual violations in which there was a direct causal link between the employer’s conduct and the harm suffered by the employee.277

Because of this hidden nature of the discrimination problem, there would be a stronger argument for deferring to legislative factfinding that was based not on explicit examples of wrongdoing (such as lawsuits or even complaints to EEOC-type offices) but instead on more aggregate studies that document this phenomenon in general. And it is, of course, exactly that type of factfinding that courts [*PG50]are incompetent to perform and where deference to the legislature is thus most appropriate. Whether the Court would be willing to credit such factfinding is unclear, given its recent insistence on actual examples of unconstitutional conduct performed by states.278

In sum, then, litigants may well be able to demonstrate both ENDA’s congruence and proportionality and the existence of the underlying discrimination but only if courts recognize the breadth and inchoate nature of both discriminatory conduct and the harms therefrom. A fair test of ENDA as appropriate Section 5 legislation would also require courts to understand the uniqueness of discrimination based on sexual orientation. These difficulties require that Congress consider carefully its factfinding responsibility. The next Part of this Article evaluates the factfinding reflected in current versions of ENDA and offers suggestions for further investigation and findings of fact that the Court might find useful in its Section 5 inquiry.

III.  (Unsolicited) Advice to Congress

The Court’s new Section 5 jurisprudence clearly requires more from Congress in the way of proof. This requirement applies especially when the statute addresses an area the Court itself has not considered particularly problematic from an equal protection standpoint. City of Boerne v. Flores, Florida Prepaid v. College Savings Bank, Kimel v. Florida Board of Regents, and Board of Trustees of the University of Alabama v. Garrett all criticized the lack of evidence that states were engaging in widespread constitutional violations.279 Garrett also criticized the lack of actual legislative findings of such violations.280 Under these cases, then, drafters of ENDA should consider carefully both the findings ultimately made in the bill and their evidentiary support in the legislative record.

A.  Documentation and Findings of Actual Government Bias Against Gays and Lesbians in Hiring

An obvious place to start with legislative investigation is the existence and scope of state government employment bias against gays [*PG51]and lesbians. This should not be surprising: Statutes justified under Section 5 must identify the unconstitutional conduct triggering congressional action. Still, Garrett and the other recent Section 5 cases make this documentation task more challenging than under earlier cases such as Katzenbach v. Morgan, even assuming that Morgan would have required such evidence.281

Several problems might arise with regard to Congress’ documentation of sexual orientation discrimination in the workplace. First, as noted earlier, the option of the closet makes documentation of actual workplace discrimination harder for sexual orientation claims, as opposed to claims based on other characteristics such as age or race.282 Indeed, the pressure gays may feel to stay in the closet could itself be considered an aspect of workplace discrimination, an aspect which would be even more difficult to document. Thus, the Court’s insistence on the presentation of actual examples of unconstitutional conduct in numbers amounting to a pattern of such behavior by the states seems at first glance inappropriate to the inquiry with regard to sexual orientation.

Still, the examples of intentional discrimination against gays and lesbians are, sadly, sufficiently numerous that a lengthy record of such conduct perpetrated by states could easily be compiled by Congress.283 And since such intentional discrimination will so rarely be motivated by a reason sufficiently rational to satisfy equal protection, most of the examples found will most likely qualify as instances of unconstitutional discrimination and thus directly support ENDA’s validity.

Second, while under City of Rome v. United States government practices having a disparate impact might be viewed as justifying broader remedial legislation if Congress determines that such disparate im[*PG52]pact suggests the presence of intentional, invidious discrimination,284 it might be harder to convince the Court that government actions having a disparate impact on gays hide an invidious intent. As noted earlier, ENDA disallows disparate impact claims.285 Nevertheless, findings of disparate impact could be useful to a court when considering ENDA’s constitutionality, as they might suggest the existence of invidious discrimination.

Still, the Court might have a problem with this latter approach. First, the possibility of a gay or lesbian employee remaining in the closet may lead the Court to wonder whether the asserted disparate impact actually exists. Of course, this is ironic: A homosexual’s self-infliction of the closet would cause the additional harm of casting doubt on the existence of the discrimination. Second, the Court may view homosexuality as more of a cultural phenomenon than, say, racial identity, with the result that the claimed disparate impact may be viewed not as a result of intentional discrimination but instead as a result of employment or career choices made by gays or lesbians.286 At any rate, these factors could easily be expected to hover in the back of the Court’s mind, leading it perhaps to demand more before disparate impact would be seen as persuasive evidence of discriminatory intent.

B.  Documentation of the Relative Likelihood of Sex Crimes by Homosexuals
and Heterosexuals

This may sound like an odd subject with which Congress should concern itself. However, it may be something that is necessary, and certainly useful, in getting ENDA past the hurdle posed by Bowers v. Hardwick. Essentially, evidence that homosexuals engage in illegal sexual activity as much as heterosexuals would call into question the rationality of a state argument (or a court making the state’s argument under the rational basis test) that employment discrimination based on sexual orientation is an appropriate way of ensuring that the [*PG53]state not hire persons likely to violate the law. This sort of factfinding, for which there is empirical evidence in the social science literature,287 is the sort most likely to be accepted by the Court. It deals with social reality, not legal concepts that might lead the Court to suspect that Congress was attempting to interpret the Fourteenth Amendment.288 Moreover, its social science/empirical basis makes it the type of issue where Congress’ competence is most pronounced, relative to the Court’s.

A problem with this analysis is the existence of sodomy statutes that restrict only same-sex conduct. Currently, six states maintain sodomy statutes that apply only to same-sex conduct.289 This fact raises the specter of a court questioning Congress’ finding, since, at least in those states, only same-sex sodomy is a crime. This raises an interesting, potentially circular argument: Gays and lesbians can be discriminated against in the workplace because they are more likely to commit sex crimes, and they are more likely to commit sex crimes because some sex crime statutes are targeted at same-sex conduct. Once again, Hardwick intrudes into the equal protection argument due to the complex relationship between conduct and status. However, there is at least reason to wonder whether sodomy statutes confined to same-sex conduct remain viable after Romer v. Evans. Moreover, this problem becomes less pressing as the number of states with sodomy laws falls, as those laws are either repealed or struck down under state constitutional principles.290

[*PG54]C.  The Irrationality of Sexual Orientation Discrimination

Finally, it makes sense for Congress to discuss in its findings the irrationality of sexual orientation discrimination in government workplaces. Previous versions of ENDA have recited findings about the lack of a relationship between an individual’s sexual orientation and his or her ability to contribute to the economic life of the nation.291 Still, Kimel and especially Garrett suggest that the Court would not consider such a finding dispositive. In Kimel, the Court ignored the legislative findings that employers imposed “arbitrary” age limits on workers.292 Instead, the Court concluded, based on its own investigation of the record, that Congress had not found sufficient instances of age discrimination that were so irrational as to violate equal protection.293 Kimel, however, does not fully resolve the question of the Court’s response to Congress’ findings, since in the ADEA the reference to “arbitrary” age limits may well have been to the simple drawing of a bright line age cut-off rather than an age limit that was necessarily “arbitrary” in the sense of being unreasonable or irrational.

In Garrett, however, the Court faced somewhat more definite congressional findings about the irrationality of the type of discrimination being outlawed. In the ADA, Congress found the “continuing existence of unfair and unnecessary discrimination,”294 which cost the nation “billions of dollars in unnecessary expenses resulting from dependency and nonproductivity.”295 Similarly, Congress found that individuals with disabilities “have been faced with restrictions . . . resulting from stereotypic assumptions not truly indicative of the ability of such individuals to participate in, and contribute to, society.”296 Read broadly, these findings could be taken as legislative uncovering of a significant problem with disability-based discrimination that was irrational in the sense of not being justified on a cost-benefit basis. Indeed, the dissent in Garrett cited these findings in arguing that the ADA was an appropriate response to that problem.297 The majority, however, did not discuss these particular findings. Instead, it focused on the finding that Congress did not make, either in the statute or in [*PG55]the committee reports on the bill: Namely, that states engaged in a pattern and practice of unconstitutional conduct298 with regard to the employment of disabled persons.299

The Court’s refusal to credit Congress’ findings in the ADA suggests several conclusions. First, and most obviously, legislative findings must be quite precise in order to satisfy the Court that the accompanying prohibitions are valid Section 5 enactments. In particular, the findings must refer to states themselves (not just “society” or “employers”), and they must refer to irrational or unreasonable conduct. Second, and more generally, the Garrett Court may have interpreted such findings as attempts by Congress to interpret the Fourteenth Amendment to the extent those findings purported to determine that discrimination against the disabled was often unconstitutionally irrational. In that sense, the Court might have viewed these findings as analogous to RFRA: Just as RFRA was thought to have represented a congressional attempt to reinterpret the Free Exercise Clause, so too these findings might have been seen as attempts by Congress to perform the rational basis review the Court may have thought was its own domain.

The first of these concerns—about the precision of Congress’ findings—is easy enough to address in ENDA, assuming that empirical evidence exists allowing Congress to honestly make the finding. The second concern, however, raises a more theoretical and difficult question: To what extent may Congress cast as “findings” statements that have direct legal significance? For example, if Congress were to “find” that a particular government act (say, an affirmative action program) “was narrowly tailored to meet a compelling government interest,” would the Court accept that “finding” or discount it as an obvious attempt to short-circuit the Court’s role in performing judicial review?300 An analogy might be found in appellate review of lower [*PG56]court factfindings. For example, one appellate court (dealing, ironically, with the question whether gays and lesbians constitute a suspect class) viewed the trial court’s determination of that issue as a “‘constitutional fact’” that was supported by the trial court’s “ultimate facts and interrelated applications of law, sociological judgments, [and] mixed questions of law and fact.”301 For that reason it reviewed de novo the trial court’s finding.302 Full explanations of the constitutional fact doctrine or the scope of the “clearly erroneous” standard governing appellate review of trial court fact findings are well beyond the scope of this Article.303 However, the analogy should be clear: To the extent the fact that is found (either by a trial court or Congress) starts to take on the character of a legal holding, appellate review (or, in the case of legislatively found facts, judicial review in general) may appropriately become less deferential.

Regardless, Congress could probably avoid this problem with regard to ENDA because it should be relatively easy to marshal sociological and other social science data supporting the proposition that sexual orientation has no impact on the factors that are relevant to successful employment.304 The empirical nature of this finding makes it more likely that the Court would credit it as being within Congress’ area of expertise and authority, as opposed to an attempt to engage in law-interpreting as the Court clearly suspected in City of Boerne.305 Put another way, the empirical nature of the facts found leaves courts with the ultimate authority to decide how those facts affect the outcome of [*PG57]a particular legal test, for example, whether a government action was motivated by a compelling interest or was so “irrational” as to violate the rational basis test of equal protection.306

By contrast, one finding that the Court ignored in Garrett and, based on the above analysis, is likely to ignore in ENDA concerns the status of the benefited group as a suspect class. In the ADA, Congress “found” that the disabled had all the indicia of a suspect class: According to Congress, they were a “discrete and insular minority,” “subjected to a history of purposeful unequal treatment,” and “relegated to a position of political powerlessness, based on characteristics that are beyond the control of such individuals and resulting from stereotypic assumptions not truly indicative of the ability of such individuals to participate in, and contribute to, society.”307 These words could have come out of the Supreme Court itself (indeed, they have),308 which may be the reason the Court thinks those determinations belong to it rather than to Congress.309

The upshot, then, is that drafters of ENDA may wish to think about linking findings of the irrationality of sexual orientation-based employment discrimination to empirical evidence of sexual orientation’s irrelevance to successful employment.310 Such evidence should [*PG58]be presented and cited in the legislative history and the bill’s findings themselves, not only because the Court now looks more closely at congressional findings in general,311 but because these findings may be construed as either empirical or partially legal due to their special nature.312

IV.  Thoughts on the Court’s Section 5 Jurisprudence in Light of ENDA’s Possible Fate

The scope of the Section 5 power is a breathtakingly broad topic, involving issues of the separation of powers, judicial supremacy, federalism, and the scope of the individual liberties in both the Fourteenth Amendment itself and, via incorporation, the Bill of Rights. For this reason, some humility may be called for when attempting to discern the proper scope of that power. This concluding part of the Article attempts to offer some modest insights into the Section 5 issue. It proceeds from a proposition broadly accepted by the current Court: that RFRA was unconstitutional. It then considers why RFRA may have gone beyond Congress’ Section 5 power and applies that tentative reasoning to federal laws justified as provisions enforcing the equal protection guarantee. It concludes by considering Congress’ special role in identifying discrimination that society no longer considers reasonable, including sexual orientation discrimination.

A.  A First Cut: Congress’ Use of Legal Terms of Art

In approaching the issue, it may help to start by thinking about the one case of those which this Article has discussed that did not produce a dissent on the Section 5 issue: City of Boerne v. Flores.313 What was it about RFRA that allowed the Court, sharply divided on so many of the issues encompassed within the Section 5 issue, to forge some[*PG59]thing of a consensus that RFRA was unconstitutional? One obvious characteristic of RFRA was that it used legal terms of art, namely, mandating that government not “substantially burden”314 a person’s religious exercise unless the burden furthered “a compelling government interest”315 and was “the least restrictive means”316 of furthering that interest. Initially, our intuition might be that there is something inappropriate about Congress’ use of such “legal” terms. On reflection, though, that intuition reveals itself as incorrect. For example, if Congress thought that any racial disparities in employment were an abomination that should be allowed only if absolutely necessary, there would be nothing wrong with it using these terms in a statute mandating racial proportionality in every workplace in the country unless it was critical that a disparity be maintained. Indeed, using these terms brings the clarity of a common language: Congress knows what those terms mean to the courts, and the courts know that Congress knows, with the result that there is increased predictability and accuracy in the dialogue between the judiciary and the legislature.

B.  A Second Cut: Judicial Supremacy

A more subtle version of our initial intuition might revolve around the kind of issue on which Congress was legislating. On this view, RFRA’s problem was that it used these legal terms in an area subject to the courts’ ultimate authority: the meaning of the Constitution. The sequences of events leading to City of Boerne suggests the problem. In Employment Division, Department of Human Resources of Oregon v. Smith, the Court interpreted the Free Exercise Clause so as to subject government action burdening religion to less stringent scrutiny than the compelling government interest test that had prevailed since Sherbert v. Verner.317 In RFRA, Congress sought to overturn Smith and reinstate a version of the Sherbert test.318 In turn, City of Boerne struck RFRA down as exceeding Congress’ power to “enforce” the free exercise guarantee.319

[*PG60] On the other hand, commentators have argued that this sort of legislative disapproval of a constitutional law standard is unexceptional and not troubling, at least when the legislative response increases, rather than limits, the scope of the right.320 The existence of a tradition of such legislative expansions of rights321 provides a troubling response to the argument that fundamental separation of powers principles, informed by statements of judicial power such as Marbury v. Madison322 and Cooper v. Aaron,323 flatly foreclose any role for Congress in interpreting the Constitution.

Further concerns arise when one considers the difference most commentators perceive324 between Katzenbach v. Morgan’s two rationales for upholding the Voting Rights Act provisions challenged in that case: First, the theory that Congress may have been seeking to protect the rights of Puerto Ricans to equal attention from government, which, in Congress’ view, could best be achieved by ensuring their voting rights; and, second, the theory that Congress could have determined for itself that the inequality suffered by Puerto Ricans constituted invidious discrimination and was thus itself unconstitutional. As the City of Boerne Court conceded, though, “the line between measures that remedy or prevent unconstitutional action and measures that make substantive changes in the governing law is not easy to discern.”325 Indeed, Morgan’s discussion of this second rationale referred to Congress’ institutional competence,326 an interesting observation in light of the congressional law-interpreting function that that rationale is thought to embrace.

[*PG61]C.  A Third Cut: The Nature of the Equal Protection Guarantee

A third cut at our intuition, informed by our progress so far, should also focus on the kind of issue on which Congress was legislating but from a slightly different perspective. This approach focuses on the type of constitutional right Congress is attempting to affect through its Section 5 power. The key distinction here is between equal protection rights and other rights found in the Fourteenth Amendment.

Equal protection is an enormously powerful tool. In contrast to the Fourteenth Amendment’s Privileges and Immunities Clause, which simply protects certain fundamental interests (interpreted quite narrowly since the Slaughter-House Cases),327 and the Due Process Clause, which protects only life, liberty, and property interests,328 the Equal Protection Clause prohibits any government action that discriminates invidiously, regardless of the importance of the interest.329 Thus, the military’s eligibility rules for dependents’ benefits,330 a city’s distribution of contracting business,331 its regulation of advertising on trucks,332 and even its singling out of an individual homeowner with regard to utility connections333 all raise equal protection concerns. While the other Fourteenth Amendment guarantees demarcate particular freedoms or interests government cannot intrude on, equal protection is a broad command of equal, or fair, treatment across every sphere of government action.334

The breadth of the equal protection guarantee suggests that Congress should have broader discretion to “enforce” it compared with other Fourteenth Amendment rights. At base this might have to do with the character of the rights guaranteed. Concepts such as “lib[*PG62]erty interests,” “cruel and unusual punishment,” or “free exercise of religion” have as much legal as empirical meaning. For example, as laypersons we might call many things “cruel” that the Court would not consider “cruel and unusual.”335 By contrast, the constitutional requirement of equal or fair treatment may be more readily comprehensible as a non-legal matter. In other words, whatever one might think about the reasonableness of a distinction between a lay person’s understanding of “cruelty” and what the Court considers “cruel and unusual punishment,” it is surely much more difficult to justify a difference between lay and legal understandings of fairness or equality. Indeed, this difference is suggested by the very vacuousness of the term “equal protection” (vacuous since all legislation classifies) and the strong pull political process theory has had in the equal protection area, both of which suggest that an inherently legal, or specialized, definition of equality simply does not exist.336 To be blunt: “Cruel and unusual punishment” may be a term of art; “equality” is much less so.

In turn, equality’s less specialized meaning potentially justifies a wider berth for legislative assistance in informing the Equal Protection Clause. Thus, while the City of Boerne Court might have been justified in considering RFRA an illegitimate congressional usurpation of the Court’s power to interpret the Constitution, there may be less reason for the Court to reject congressional attempts to give meaning to the Equal Protection Clause’s promise of equality. Congress might not be able to create a privilege or immunity or expand the meaning of free religious expression,337 but Congress still might be able to de[*PG63]termine what constitutes invidious discrimination. Indeed, as noted above, Morgan’s discussion of Congress’ determination that the denial of voting rights to Puerto Ricans constituted invidious discrimination included a reference to congressional competence.338 More generally, Morgan cited a variety of issues or concerns that the Court concluded were for legislative assessment and weighing:

[T]he risk or pervasiveness of the discrimination in governmental services, the effectiveness of eliminating the state restriction on the right to vote as a means of dealing with the evil, the adequacy or availability of alternative remedies, and the nature and significance of the state interests that would be affected by the nullification of the English literacy requirement as applied to residents who have successfully completed the sixth grade in a Puerto Rican school.339

The important point here is that these factors share a common-sense basis in the amount of equality that is appropriate and feasible in a given context. The point is not that the existence of these factors requires a weighing of competing considerations; in general, balancing tests may well include factors that courts are uniquely qualified to judge given the existence of extrinsic aids to identify and consider those factors.340 But what extrinsic aids govern the equality determination made in the statute upheld in Morgan? If there are none, or few, [*PG64]the argument should naturally follow that Congress’ enforcement power should correspondingly expand.

D.  ENDA and Section 5

Thus, the equal protection guarantee is broad and based on a constitutional principle that is less susceptible than most to legal determination and reasoned explication. For these reasons, it is also fluid and especially susceptible to social change. Sexual orientation discrimination provides a striking example. What was once thought to be perfectly reasonable discrimination against “perverts” and, later, people with serious mental illness has come to be understood as inappropriate status-based discrimination, explainable, if at all, only as an expression of moral disapproval irrelevant to the conduct of “an almost limitless number of transactions and endeavors that constitute ordinary civic life in a free society.”341

The fluidity of society’s conceptions of appropriate and inappropriate classifications provides another reason for allowing the legislature a broader scope for enforcing the equal protection guarantee. This is especially the case with sexual orientation discrimination given the complexity of our attitudes toward sexuality and gender roles. Because understanding sexual orientation discrimination requires study and data gathering, and because equal protection may fundamentally be a matter of social consensus on what discrimination is appropriate or fair,342 Congress appears well suited to play a major role in determining the contours of the broad and majestic, yet vague, command of the Equal Protection Clause.

There may be cause for concern if Congress defines those contours too narrowly, so as to violate some judicially recognized floor of protection.343 But that is not an issue with ENDA. Instead, ENDA appears to be a classic example of Congress perceiving a new societal concern with a particular type of discrimination and targeting the broad command of the Equal Protection Clause toward its elimination. It does so narrowly, respecting interests (such as the right to as[*PG65]sociation) that courts have recognized as constitutionally based. If Congress documents the problem of sexual orientation discrimination, the fact that courts have not themselves had significant occasion to consider the issue should not matter.344 To the extent that the Supreme Court’s prior attitude should matter, though, the core concern in Romer v. Evans—that it is unconstitutional to burden a group for no relevant reason—finds a close reflection in ENDA’s findings, prohibitions, and exemptions.

Under this analysis, ENDA should be found a valid expression of Congress’ Section 5 power.


?? ??