Concept
Shelby County v. Holder
Fifteenth Amendment enforcement power
Shelby County v. Holder
Section 2 of the Fifteenth Amendment gives Congress power “to enforce this article by appropriate legislation,” and in 2013 the Supreme Court drew its sharpest limit yet on how Congress may use that sentence to support the Voting Rights Act.1 In Shelby County v. Holder, decided 5–4 in an opinion by Chief Justice Roberts, the Court held that Congress may not subject some States, but not others, to federal preclearance of changes in their voting laws on the strength of an outdated coverage formula. The Tenth Amendment and principles of equal sovereignty demand extraordinary justification for that kind of disparate treatment, and reliance on decades-old data could not supply it under the Fifteenth Amendment's enforcement power.2 Northwest Austin had warned of this result without reaching it; Shelby County is where the warning became a holding.
The holding concerns the coverage formula, so the first task is to see exactly what fell. Section 4(b) identified which States and subdivisions were covered, and the Court held that its formula could no longer be used to subject jurisdictions to preclearance.3 Section 5, the preclearance requirement itself, remained in the statute. But Section 5 binds only the jurisdictions the formula identifies, as the earlier discussion of the 1965 design showed, so a command that survives without a test for who must obey it reaches no one. Preclearance stopped operating in practice although the Court did not strike down the provision that creates it.
The case also matters for its method. Katzenbach sustained the 1965 regime by asking whether Congress could rationally have chosen its means on the record before it. Shelby County asks whether a burden that falls on only some States still fits present conditions. The subsections below supply the facts and the holding, follow the majority's reasoning in order, mark the holding's limits and the question it left open, and end with the separate opinions, which show what the majority rejected.
Background and Holding
The Fifteenth Amendment authorizes Congress “to enforce this article by appropriate legislation,” and the question in Shelby County v. Holder was whether one use of that power, the coverage formula in Section 4(b) of the Voting Rights Act, remained appropriate in 2013.4 The challenger was Shelby County, in Alabama, a covered jurisdiction. The statute it attacked had been reauthorized in 2006 for another 25 years with the formula unchanged, so coverage still turned on whether a jurisdiction had used a voting test in the 1960s or 1970s and had low registration or turnout at that time.5 This subsection gives the facts and procedural path and states the holding; the later subsections supply the reasoning.
Shelby County sued the Attorney General in 2010 in the federal district court in Washington, D.C. It sought a declaratory judgment that Sections 4(b) and 5 are facially unconstitutional and a permanent injunction against their enforcement.6 A facial challenge asks a court to invalidate a provision as written, in every application, rather than to excuse the plaintiff alone. The posture shapes what victory means, because a ruling that the provision itself is invalid disables it everywhere it would otherwise operate.
The county did not seek bailout, the statutory exit from coverage described earlier in this chapter; it sued instead.7 Its suit arrived a few years after Northwest Austin had construed the bailout provision in a way that left the constitutional issues for another day.8
The county lost in the district court, which upheld the Act.9 A divided D.C. Circuit affirmed. On Section 4 it acknowledged that the evidence for singling out the covered jurisdictions was “less robust” and that the issue presented “a close question.”10 It nevertheless sustained the formula in part on the pattern of successful Section 2 suits, the evidence that Judge Williams, dissenting, would break down State by State.11
Judge Williams's dissent went to the fit between formula and problem. He found no positive correlation between coverage under Section 4(b) and low Black registration or turnout. Disaggregating the successful Section 2 suits by State, he concluded that the five worst uncovered jurisdictions had worse records than eight of the covered ones, and that two covered jurisdictions, Arizona and Alaska, had no successful reported Section 2 suit in the 24 years studied. On that record he would have held the formula irrational.12 The panel majority relied on Congress's judgment; the dissent tested whether the line tracked the problem at all. That second question, whether coverage is sufficiently related to what it targets, is the one the Supreme Court took up.
The Supreme Court heard argument on February 27, 2013, and decided the case on June 25, 2013.13 It reversed, 5–4. Chief Justice Roberts wrote for the Court, joined by Justices Scalia, Kennedy, Thomas, and Alito; Justice Thomas also filed a concurrence, and Justice Ginsburg dissented, joined by Justices Breyer, Sotomayor, and Kagan.14 The Court began from two principles it had stated in Northwest Austin: the Act imposes current burdens and must be justified by current needs, and a departure from the principle of equal sovereignty requires a showing that the statute's disparate geographic coverage is sufficiently related to the problem it targets.15 Applying them, it held Section 4 unconstitutional: its formula can no longer be used as a basis for subjecting jurisdictions to preclearance.16
The holding is narrower than a first reading suggests. The Court struck only the coverage formula in Section 4(b) and issued no holding on Section 5.17 Students often write that Shelby County invalidated preclearance; that is imprecise. Section 5 stays in the statute but binds only the jurisdictions the formula identifies, so with the formula unusable it has none to bind until Congress enacts a new formula grounded in current conditions.18 The next subsections explain how the Court reached that result and what it left in place.
Federalism and Equal Sovereignty
Congress's enforcement power was never in doubt in Shelby County: the Fifteenth Amendment gives Congress power “to enforce this article by appropriate legislation,” and the majority did not question that the grant exists.20 It instead measured the Voting Rights Act against two structural premises about the relationship between the national government and the States, stated before it reached the coverage formula. Both grow out of the principles drawn from Northwest Austin: the first concerns the States' authority over their own elections, the second equality among the States.21 Each supplies a baseline, and the opinion asks how far the Act departs from it.
The first premise starts from the Supremacy Clause, which makes the Constitution and laws of the United States “the supreme Law of the Land,” so that state legislation may not contravene federal law.22 That rule tells a court what to do with a state law that conflicts with federal law. It gives the Federal Government no general right to review and veto state enactments before they take effect.23 The Constitutional Convention considered a proposal to let the national government “negative” state laws and rejected it, choosing to let state laws take effect, subject to later challenge under the Supremacy Clause.24 The choice decides who must move first. Under the adopted design, a State legislates, the law operates, and the party who says it conflicts with federal law must come forward and prove it. Under the rejected design, the State would need approval before its law could operate at all.
Beyond the Supremacy Clause, the Tenth Amendment reserves to the States or the people every power not specifically granted to the Federal Government, and States keep broad autonomy in structuring their governments and pursuing legislative objectives.25 The Court gave the reservation a reason: the allocation of powers preserves the “integrity, dignity, and residual sovereignty of the States,” and diffusing sovereign power secures liberties to citizens.26 Elections are where the reservation has particular force. The Framers intended the States to keep the power to regulate elections, and the cases the majority cited recognize state power to set the conditions under which suffrage is exercised, to prescribe the qualifications of state officers and the manner of choosing them, and to draw congressional district lines.27 The Federal Government keeps significant control over federal elections, since Article I, Section 4 lets Congress set the time and manner of electing Senators and Representatives.28 If each State shapes its own electoral system, then a federal rule that makes a State seek permission before changing that system reaches into a function the Tenth Amendment leaves with the State.
The second premise has a different source. Alongside each State's retained sovereignty, the Court identified a “fundamental principle of equal sovereignty” among the States, tracing it to a 1911 decision that described the Nation as “a union of States, equal in power, dignity and authority.”29 That decision, Coyle v. Smith, concerned the admission of new States, and it called the constitutional equality of the States essential to the harmonious operation of the scheme on which the Republic was organized.30 The logic is plain. If the States are equals in power, a federal burden placed on some and withheld from others alters the relationship among them, because the burdened States hold less practical authority over their own legislation than their neighbors do. A court reviewing such a burden must ask why the line was drawn where it was, which is the question the requirement that coverage be “sufficiently related” to the problem poses.
Against that background the Act departs sharply from basic principles, and it does so twice. The first departure comes from the structure of preclearance. The requirement suspends all changes to state election law, however innocuous, until federal authorities have cleared them, so States must beseech the Federal Government for permission to implement laws they would otherwise have the right to enact and execute on their own.31 The Attorney General has 60 days to object, longer if he requests more information.32 That is the advance review the Convention declined to give the national government, applied to a field the Tenth Amendment leaves to the States. The second departure comes from coverage. The Act applies to only nine States and several additional counties, so one State waits months or years and spends money to implement a validly enacted law while its neighbor can typically put the same law into effect at once.33
The preclearance proceeding also differs from ordinary litigation: it not only switches the burden of proof to the supplicant jurisdiction but also applies substantive standards quite different from those governing the rest of the nation.34 The same statute contains a provision that raises neither problem. Section 2 forbids any voting standard, practice, or procedure that results in a denial or abridgement of the right to vote on account of race or color; it is permanent, applies nationwide, and is not at issue in the case.35 It carries no preapproval requirement, and the onus is on voters or the United States to sue after the fact and prove their claim.36 The majority itself treated preclearance as an addition to that remedy, describing it as operating “subject of course to any injunction in a §2 action.”37 The premises therefore bear on provisions that require advance approval from only some States; they leave the nationwide litigation remedy alone.
The premises set a high bar, not a flat prohibition. In 1966 the Court itself called the Act “stringent” and “potent,” conceded that it “may have been an uncommon exercise of congressional power,” and justified it by “exceptional conditions.”38 A departure from the usual allocation may be justified, but only by a showing proportioned to its size. That is why the equal-sovereignty premise needed room to operate outside the admission context, and the opinion's handling of that point is the most contested step in this part of the case.
The majority conceded that Katzenbach rejected the notion that equal sovereignty barred differential treatment outside the admission of new States, and it relied on Northwest Austin for the proposition that the principle remains “highly pertinent in assessing subsequent disparate treatment of States.”39 Read this way, Katzenbach ruled out an absolute bar and left the principle free to demand justification for a disparate remedy. The dissent read the same decision more narrowly, quoting its holding that the principle “applies only to the terms upon which States are admitted to the Union, and not to the remedies for local evils which have subsequently appeared.”40 On that reading, equal sovereignty has no work to do in reviewing the Act's geographic scope at all.
The dissent's objection was that the majority extended dictum beyond the precedent it came from. Northwest Austin had cited Katzenbach's holding while declining to decide whether the Act was constitutional or even what standard of review applied, so its language could not have silently overruled Katzenbach's limit to the admission of new States.41 By the dissent's account, the majority ratcheted up pure dictum without explaining why it found Katzenbach wrong and without asking whether stare decisis counseled adherence.42 The dissent added two practical worries: many federal statutes treat States differently, such as the Yucca Mountain provisions that name Nevada, and Congress had relied on Katzenbach in every reauthorization of the Act.43 The majority did not say it was overruling Katzenbach, so an answer that reports an overruling misstates the opinion; the better description is that the Court gave equal sovereignty a role outside the admission context that the 1966 decision had not given it. The principle's reach beyond voting remains contested. Some lower courts have held that it does not restrict the commerce power, and the Court's emphasis on the Act's regulation of state choices about their own voting systems suggests that the principle, in its strongest form, may not govern most other congressional powers.44
Together the two premises fix what the Act must show. The federalism premise makes advance federal review of state election laws exceptional, and the equal-sovereignty premise makes its application to only some States a second exception that needs its own justification. Whether the 2006 formula supplied that justification is what the Court went on to decide.
Current Burdens, Current Needs
The enforcement clause of the Fifteenth Amendment authorizes “appropriate legislation,” and Shelby County asks whether a remedy built on that power still satisfies constitutional requirements under present conditions.46 The Court framed the inquiry with a concession and a question. Voting discrimination still exists, and no one doubts that, so the issue is whether the Act's extraordinary measures, including their disparate treatment of the States, continue to satisfy constitutional requirements. Its answer rests on the principle that the Act imposes current burdens and must be justified by current needs.47 Applied to the coverage formula, the principle becomes a requirement about selection. If Congress is to divide the States, it must identify the jurisdictions it singles out on a basis that makes sense in light of current conditions, and it cannot rely simply on the past.48 The test is easiest to learn by watching one formula meet two sets of facts, one from 1965 and one from 2004.
In 1965 the formula fit the problem it was built to solve. The registration gaps in the covered States were enormous: Virginia's gap between white and Black registration was 22.8 points, and Mississippi's was 63.2, with white registration at 69.9 percent and Black registration at 6.7 percent.49 Those gaps traced to the tests and devices States had adopted to keep Black citizens off the rolls, and the formula followed that chain directly. It looked to cause (discriminatory tests) and effect (low registration and turnout) and tailored the remedy to jurisdictions that showed both.50 That is why the Court in 1966 called the formula rational in both practice and theory, and why it saw the stringent remedies as aimed at areas where voting discrimination had been most flagrant.51
By 2004 the same States told a different story. Across the States shown, the registration gap had narrowed to between negative 3.8 and 10.8 points, and Mississippi, the State with the widest gap in 1965, now showed Black registration at 76.1 percent against 72.3 percent for whites.52 Census Bureau data from the most recent election added that Black turnout exceeded white turnout in five of the six originally covered States, with a gap in the sixth of less than one half of one percent.53 Congress itself wrote in 2006 that significant progress had been made in eliminating first-generation barriers, including increased numbers of registered minority voters, higher turnout, and greater representation.54 These figures describe a different Nation from the one the formula was designed for.
The Court did not conclude from these figures that racial discrimination in voting had ended, and an answer that says it did misreads the opinion. The majority acknowledged that voting discrimination persists and credited the improvements largely to the Voting Rights Act itself.55 That success is the changed fact: a remedy justified by exceptional conditions is justified only while those conditions last, and the better the remedy works, the sooner they move. Respondents did not deny the improvements on the ground. They argued that much of the improvement reflected the deterrent effect of Section 5, which would fall away if Section 5 were struck down.56 The Court answered that this argument would make Section 5 effectively immune from scrutiny, because no matter how clean the record of covered jurisdictions, someone could always say that deterrence explained the good behavior.57 A remedy whose success counts as proof that it remains necessary could never be reviewed, so a court applying the current-needs test must look at present conditions on their own terms.
Congress reauthorized the formula in 2006 after forty more years of the Act's operation. In assessing the current need for a preclearance system that treats States differently, that history cannot be ignored: voting tests were abolished, racial disparities in registration and turnout were erased, and Black candidates attained office in record numbers, largely because of the Voting Rights Act.58 Yet the formula Congress reauthorized ignored these developments, keeping its focus on decades-old data relevant to decades-old problems rather than current data reflecting current needs.59
The Fifteenth Amendment explains why the mismatch matters. It commands that the right to vote not be denied or abridged on account of race or color and empowers Congress to enforce the command, but it is not designed to punish for the past; its purpose is to ensure a better future.60 A division among the States that rests on decades-old events punishes the past, while a division that rests on present conditions aims at preventing future violations, which is the aim the enforcement power serves.61 Meanwhile Congress made the regime heavier. It reauthorized the Act for another 25 years and expanded Section 5, so the bar covered jurisdictions had to clear was raised even as the conditions justifying it dramatically improved.62
These steps ask whether the line Congress drew still matches the present. They do not ask whether discrimination persists or whether Congress has power to address it, because the Court accepted both points. The next subsection follows how the Court applied the steps to the formula's two criteria, voting tests and low participation.
Why the Formula Failed
Section 2 of the Fifteenth Amendment authorizes Congress to enforce the ban on racial discrimination in voting by appropriate legislation, and Shelby County decided that a coverage formula resting on decades-old data does not qualify when it singles out some States and spares others.64 The previous subsection stated the governing principle, that current burdens must be justified by current needs. This one follows what happens when the principle meets the formula and then meets each defense offered for it. One idea ties the discussion together: if Congress divides the States, the basis of the division must make sense in light of current conditions.65
Start with what the formula measured. Coverage under Section 4(b) depended on whether a jurisdiction had used a voting test and had low registration or turnout in the 1960s and early 1970s, and the Court found that both criteria had lost their connection to the present.66 Congress last revisited the criteria in 1975, when it extended coverage to jurisdictions with a voting test and less than 50 percent registration or turnout as of 1972, bringing in Alaska, Arizona, and Texas along with several counties elsewhere.67 Since then the criteria have been frozen. The sole factors were whether a jurisdiction had used a voting test in 1972 and whether its registration or turnout was below 50 percent that year, so no later development could count, either to include or to exclude a jurisdiction. Tests have been banned nationwide for over 40 years, so the first criterion no longer separates one State from another; it records only which States once used a device that none may use now. Registration and turnout in the covered States have risen dramatically, so the second criterion no longer signals a racial disparity, the evidence that had justified the formula in the first place.68 In 1965 the formula sorted the States into two groups that really differed. The Nation is no longer divided along those lines, though the Act continues to treat it as if it were.69
The Government's first defense was that the formula was reverse-engineered: Congress picked the jurisdictions it wanted to reach and then wrote criteria that happened to describe them. On that view the criteria need no logical relationship to the reason for coverage, and the formula is sound so long as it captures the right jurisdictions.70 The Court refused to read Katzenbach as approving that approach. Katzenbach called the formula rational because it was relevant to the problem: tests and devices matter to voting discrimination because of their long history as a tool for perpetrating it, and a low voting rate matters because widespread disenfranchisement must reduce the number of actual voters.71 Relevance is the only handle a court has for checking whether the line Congress drew corresponds to the problem Congress was addressing. A defense that disclaims any logical link between criteria and coverage leaves the court nothing to examine. Here the Government's argument made no attempt to show that the formula remains relevant to the problem it targets.72 The Court held that the failure to establish even relevance was fatal in a decision to subject a disfavored subset of States to “extraordinary legislation otherwise unfamiliar to our federal system.”73
The Government's fallback was that relevance in 1965 suffices: the formula may continue so long as some discrimination remains in the States Congress identified then, however it compares with discrimination in States left unburdened.74 The flaw is that the justification for preclearance is comparative. Congress burdens some States and not others, so the question is whether the burdened States still differ from the rest in a way that explains the difference in treatment, and a test that asks only whether they retain some discrimination never reaches it. The argument also leans on a 1965 comparison among the States that reflected the different histories of the North and South, a history the Court had rightly invoked in 1966.75 But history did not end in 1965, and forty more years of change had left that comparison behind.76 The Fifteenth Amendment looks forward, as the previous subsection noted: it is not designed to punish for the past.77
An exam stem may recite that Congress held extensive hearings and compiled thousands of pages of evidence before the 2006 reauthorization, and then ask whether that record saves the formula. It does not. The Court said first that, however the record is read, no one can fairly say it shows anything approaching the “pervasive,” “flagrant,” “widespread,” and “rampant” discrimination that distinguished the covered jurisdictions from the rest of the Nation in 1965.78 The more fundamental problem was that Congress did not use the record to shape a coverage formula grounded in current conditions; it reenacted a formula based on 40-year-old facts having no logical relation to the present day.79 A record helps a statute only when the statute was built from it. The Court said it could not pretend to review an updated statute, or try its own hand at updating the statute, based on the new record compiled by Congress; it was not ignoring the record but recognizing that the record played no role in shaping the formula before it.80 The reason follows from the judicial role. A court that rescued an old formula with evidence Congress never used in drafting would be choosing new coverage criteria itself, and the enforcement clause assigns that choice to Congress.
The dissent's best use of the record was to point to second-generation barriers, which are electoral arrangements that affect the weight of minority votes rather than impediments to casting a ballot. The Court answered that this evidence does not cure the problem. A formula built on voting tests and access to the ballot is a poor instrument for finding vote dilution, so describing preclearance as aimed at dilution only highlights the irrationality of continued reliance on that formula.81 The mismatch has a concrete source. In 2006 Congress expanded Section 5 to reach voting laws that diminish the ability of minority citizens to elect their preferred candidates, a dilution concern, yet the formula that decided who must preclear stayed pegged to 1960s tests and turnout.82 The burden had moved to a new target while the selection criteria stayed behind, so in the majority's view the dissent's evidence confirmed the mismatch and could not repair it. Vote-dilution claims have a home elsewhere in the Act: Section 2 reaches districting schemes and other practices that produce discriminatory results, though it creates no right to proportional representation.83
One last argument concerned who may raise the objection. The dissent used the record to argue that Shelby County, given the voting discrimination found in the county, could not complain about being subjected to preclearance. The Court answered with an analogy: a driver pulled over under a policy of stopping all redheads cannot be told he has no complaint about the policy because his license turns out to have expired.84 The county's claim was that the coverage formula is unconstitutional in all its applications because of how it selects jurisdictions, and since the county was selected by that formula, it could challenge the formula in court.85 A policy that picks drivers by hair color is judged by whether hair color has anything to do with the reason for stopping, and that answer controls the policy's validity for every driver it catches. Whether the county itself had a record of discrimination likewise does not tell a court whether tests and turnout from 1972 are a sensible way to choose which States face preclearance.
The Court drew these strands together in a test of its own. Had Congress started from scratch in 2006, it could not have enacted this formula: it would have been irrational to distinguish between States in so fundamental a way on 40-year-old data when today's statistics tell a different story, and irrational to base coverage on voting tests that had been illegal for 40 years.86 The test asks only what Congress could rationally have done on the facts of 2006, so the formula's earlier enactment and repeated approval do not shield it.87 Congress had also been warned. Northwest Austin raised the constitutional questions in 2009; because Congress did not update the formula, the Court said it had no choice but to declare Section 4(b) unconstitutional and to hold that the formula can no longer be used to subject jurisdictions to preclearance.88 A formula drawn from current evidence about where voting discrimination is occurring is a different matter, and the next subsection returns to what Congress may do. The nationwide Section 2 remedy, which has no preapproval requirement and leaves the onus on challengers, was untouched by the ruling.89
Scope of the Holding
Congress's authority to enforce the Fifteenth Amendment rests on its enforcement clause, which lets Congress enforce the article by appropriate legislation, and Shelby County asked whether one exercise of that authority, the Section 4(b) coverage formula, could still stand.91 The Court held that Congress may not subject some States, but not others, to federal preclearance of changes in voting laws on the strength of outdated coverage formulas.92 That holding is narrower than a quick summary suggests. The earlier subsections explained why the formula failed; this one marks the boundary of the result, covering which provision fell, which provisions stayed, and what Congress and the Justice Department may still do.
The provision that fell is Section 4(b). In 2009 the Court had avoided ruling on the Act's constitutionality and decided the case on statutory grounds.93 Congress could have updated the formula then and did not, and that failure left the Court with “no choice but to declare §4(b) unconstitutional,” because the formula “can no longer be used as a basis for subjecting jurisdictions to preclearance.”94 The opinion also quotes Justice Holmes's description of striking down an Act of Congress as “the gravest and most delicate duty that this Court is called on to perform.”95 The shape of the ruling matches that caution. The Court decided the one question it had to decide, whether the selection of burdened jurisdictions could stand, and it left the rest of the statute alone.
The Court then drew two express limits in a single paragraph. Its decision “in no way affects the permanent, nationwide ban on racial discrimination in voting found in §2,” and it issued “no holding on §5 itself, only on the coverage formula.”96 Those two sentences explain why an exam answer saying the Court struck down the Voting Rights Act is wrong, and why an answer saying it struck down Section 5 is imprecise. Section 5 remains in the statute. What the Court invalidated is the test that identifies who must obey it.
A provision that stays on the books but binds no one is dormant, and Section 5 is in that position. Preclearance applies only to the jurisdictions Section 4(b) picks out, as the description of the 1965 design earlier in this chapter showed, so once the formula cannot be used, the duty has no one to apply to. Justice Ginsburg's dissent made the same point in a footnote: although the Court purports to declare unconstitutional only the coverage formula, “without that formula, §5 is immobilized.”97 The practical consequence is that every State and subdivision is free to change its election procedures without seeking advance federal approval. The legal status and the practical status therefore differ, and the difference matters if Congress acts, because a new formula would give the surviving Section 5 something to operate on again.
Section 2 survives because it has none of the features that drew the Court's objection. It applies nationwide and has no preapproval requirement, so voters or the United States who believe a practice produces discriminatory results must bring suit and attempt to prove the claim.98 The statute also specifies that it does not require proportional representation: “nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion of the population.”99 A Section 2 case follows the challenged practice rather than preceding it, places the burden on the challenger, and reaches every jurisdiction alike. The two problems identified in the earlier discussion of federalism and equal sovereignty were advance federal permission and a burden imposed on only some States, and a provision that involves neither is outside the reasoning that decided the case.
The Act also contains a tool that never depended on the coverage formula. Section 3 contains a bail-in provision, under which the Justice Department may ask a federal court to treat a particular jurisdiction as subject to preclearance, and the decision did not affect it. The jurisdiction may be one the old formula covered or one it never reached. If the Department believes a jurisdiction made a voting change with intent to discriminate and the court agrees, the jurisdiction may not alter its voting rules unless the court or the Department approves the change in advance. The route works one jurisdiction at a time, and it selects on a court's finding about that jurisdiction's own conduct rather than on a list Congress wrote in advance. That difference explains why bail-in survives the ruling: the Court's objection was to selecting States by decades-old data, and bail-in does not select by formula.
Congress also keeps the option to rebuild. The Court said that Congress “may draft another formula based on current conditions,” and it added that such a formula is “an initial prerequisite to a determination that exceptional conditions still exist” justifying preclearance.100 The second statement limits the first. A formula drawn from current evidence would cure the defect in the 2006 version, yet the Court called it only a starting point, so the further question whether exceptional conditions exist would remain open. A new formula is therefore necessary, though not necessarily sufficient.
Justice Thomas joined the Court's opinion in full and wrote separately to say that he would have held Section 5 unconstitutional as well.101 His reasoning was that the majority's own analysis shows Congress failed to justify current burdens with a record of current needs, so the conclusion about Section 5 follows; by leaving it unstated, “the Court needlessly prolongs the demise of that provision.”102 A concurrence states one Justice's view, so that position is not the holding, and the majority's express reservation of Section 5 controls. The disagreement is over how much to decide. Justice Thomas would finish the work the formula holding began, while the majority confined itself to the provision that selected the jurisdictions, leaving Congress free to answer with a new formula.
Unsettled Standard of Review
A court reviewing a statute passed under the Fifteenth Amendment's enforcement clause must decide whether the statute is “appropriate legislation,” and the answer depends on the standard of review.104 Katzenbach answered that question in 1966 with the rational-means standard, and Shelby County v. Holder later held that Congress may not subject some States, but not others, to preclearance on the strength of outdated coverage formulas.105 What the later opinion leaves open is the test by which it reached that result, because Chief Justice Roberts nowhere specifies the standard of review he is applying to the coverage formula. This subsection collects the clues the opinion does supply, sets out the competing readings of them, and ends with a way to handle the ambiguity in an answer.
The first clue is the language of rationality, which points toward mere-rationality review. It returns at the end, where the majority said that if Congress had started from scratch in 2006 it would have been “irrational for Congress to distinguish between States in such a fundamental way based on 40-year-old data,” and equally irrational to base coverage on voting tests that had been illegal for decades.106 One reading therefore treats Shelby County as an irrationality holding. A formula with no logical relation to present conditions fails even deferential review, and the case applies the old test to a record that had aged forty years. Rationality is judged against the conditions Congress faces when it acts, so a measure that passed on the 1965 record can fail on a 2006 record built on 1972 data without any change in the test. The dissent states the premise behind this reading most firmly: Katzenbach set the standard, the Court reaffirmed it in later reauthorization cases such as City of Rome, and today's Court “does not purport to alter settled precedent establishing that the dispositive question is whether Congress has employed ‘rational means.’”107
The majority's reply to the dissent complicates that reading. The dissent had quoted McCulloch v. Maryland and emphasized that means must be appropriate and plainly adapted to a legitimate end. The majority answered that its case turns on the part of the same sentence that the dissent left unemphasized, which asks whether a means is consistent with the letter and spirit of the Constitution, and it recalled that the Court had said in 2009 that the preclearance requirement and coverage formula raise serious constitutional questions.108 That choice of clause locates the majority's concern in consistency with constitutional structure, which is where an equal-sovereignty principle would do its work, in addition to the question whether the means fit the end. The structural concern leads to the second reading.
The second reading takes the equal-sovereignty reasoning at face value. The holding itself rests on the proposition that the Tenth Amendment and principles of equal sovereignty require extraordinary justification for disparate treatment of States, and that reliance on decades-old data cannot supply it.109 On this reading the case sets a limit on the enforcement power itself: a measure that burdens some States and spares others must clear a higher bar than ordinary rationality review, whatever its fit with the end. Some commentators accordingly group the decision with Boerne on the narrow side of a larger debate, as limiting Congress's enforcement power where Katzenbach v. Morgan had read it expansively.110 Others find the question open: whether the Court held the unchanged formula irrational or announced state-sovereignty limits on congressional power under the Fifteenth Amendment is for the future to determine.111 The debate behind that uncertainty pits a narrow view, under which Congress may only prevent and remedy violations the Court recognizes while more governance stays with the States, against a broad view, under which Congress may also interpret and expand rights.112
The test the opinion did not use is easy to describe. In City of Boerne v. Flores, the Court held that Congress may enforce a guarantee but may not determine what counts as a violation, and it demanded congruence and proportionality between the injury to be prevented and the means adopted, a test announced for legislation under the Fourteenth Amendment. The same opinion took pains not to question Congress's authority to enact the provisions of the Voting Rights Act, as reauthorized in 1970 and 1982, as civil rights remedies.113 Shelby County followed that path, since the Court did not apply the Boerne test.114 The two statutes differ in a way that may explain the omission. Boerne involved legislation that effectively redefined a guarantee, while the statute in Shelby County sought mainly to prevent what the Court agreed would be Fifteenth Amendment violations. Whether that difference accounts for the Court's silence is an inference, and an answer should present it as one.
Both readings converge on the requirement that decided the case. If Congress wants to use its Fifteenth Amendment power to prevent future violations rather than punish past ones, it must adjust its measures to reflect current rather than historical voting circumstances.115 Under the irrationality reading, a formula fails when it has no logical relation to the present. Under the state-sovereignty reading, the same defect defeats the extraordinary justification that a geographic disparity needs. The readings diverge only on harder facts. A selection basis with a plausible link to present conditions might survive rationality review and still face the heavier burden that equal sovereignty imposes, while a measure that treats every State alike gives equal-sovereignty reasoning little to work on.
The first step guards against the error students most often make, which is to say that the Court applied the Boerne congruence-and-proportionality test or that it overruled Katzenbach. The majority did neither. It repeated Katzenbach's own description of the Act as uncommon and justified only by exceptional and unique conditions, and it measured the 2006 formula against that justification.116 Because the fourth step produces one answer under either reading, the choice between readings changes the label on the analysis and leaves the result alone, at least on facts like those in Shelby County.
Concurrence and Dissent
The enforcement clause of the Fifteenth Amendment gives Congress power to enforce the Amendment by appropriate legislation, and the separate opinions in Shelby County disagree about who decides when an exercise of that power has outlived its justification.118 The majority held that Congress may not subject some States, but not others, to preclearance on the strength of outdated coverage formulas.119 Justice Thomas joined that opinion in full and would have gone further, while Justice Ginsburg dissented, joined by Justices Breyer, Sotomayor, and Kagan.120 The dissent framed the case as a question of institutional responsibility: whether this Court, or a Congress charged with enforcing the post-Civil War Amendments, decides whether Section 5 remains justifiable as currently operative.121 Each separate writing is a coherent position with its own premise. Reading them that way shows what the majority rejected and where its holding sits between them.
Justice Thomas's concurrence is short because it rests on the majority's own reasoning.122 He accepted the Court's account that the conditions that originally justified Section 5 no longer characterize voting in the covered jurisdictions. He then pointed to what Congress did in 2006: Section 5 now forbids voting changes adopted with any discriminatory purpose, and also changes that diminish the ability of minority citizens to elect their preferred candidates, so Congress enlarged the burden just as the conditions justifying it improved.123 In his view the pre-2006 version had already gone beyond the protection the Constitution itself guarantees, which made the further expansion harder to defend.124 The logic is a matching exercise. If the Act imposes current burdens that must be justified by current needs, and the opinion shows that Congress compiled no record of current needs sufficient for the burdens it imposed, then the conclusion about Section 5 follows from the majority's own premises. For that reason he treated the majority's refusal to rule on Section 5 as leaving the inevitable conclusion unstated and needlessly prolonging the provision's demise.125
The gap between the majority and the concurrence concerns how much to decide, and it has practical weight. Section 5 reaches only the jurisdictions a formula identifies, so the majority's holding leaves Congress free to draft a new formula based on current conditions, which the Court called only an initial prerequisite to showing that exceptional conditions still exist.126 On the concurrence's reasoning, a new formula would likely not rescue the statute, because the objection runs to the burden Section 5 imposes and the list of covered places is a secondary matter. The usual exam error is to attribute that position to the Court. The majority expressly declined to rule on whether Section 5 is constitutional, so the concurrence states one Justice's view and the holding stops at the coverage formula.
The dissent begins from a different premise about institutional roles. Justice Ginsburg read the Civil War Amendments as arming Congress with sweeping power to enact transformative legislation against state abuses, with the remedy placed in the legislature's hands and not left to the courts.127 Summaries of her opinion compress it into two themes: the majority failed to defer to Congress's judgment about the means needed to combat voting discrimination, and the reauthorized Act was better calibrated to current conditions than the majority allowed. The standard of review follows from the first theme. South Carolina v. Katzenbach had held that, as against the reserved powers of the States, Congress may use any rational means to effectuate the ban on racial discrimination in voting, later reauthorization cases such as City of Rome reaffirmed that rule, and the dissent observed that the majority did not purport to alter it.128 Review under that standard asks whether Congress rationally selected means appropriate to a legitimate end, and it does not ask whether Congress chose the wisest means.129 The standard still leaves the Court a job, which is to decide whether the legislative record sufficed to show that Congress could rationally have found its chosen provisions appropriate.130
On that footing the dissent argued that a statute reauthorizing an existing remedy is especially likely to satisfy rational-basis review, and it gave three reasons, each answering a different objection to renewal:
- Congress has already assembled a record justifying the initial legislation and may weigh that record together with the new one, particularly where the Court has repeatedly upheld the statute and Congress has kept to the model the Court approved.131
- Reauthorization is needed only because Congress built a time limit into the Act, which commits Congress to review the continuing need for the statute in light of contemporary evidence after a span of years.132
- A reviewing court should expect a less stark record on renewal than at enactment, because demanding an equivalent record would trap Congress in a catch-22: a working statute produces less evidence of discrimination, while a failing statute produces plenty of evidence but little reason to renew it.133
The third reason does the most work, because it answers the inference on which the majority's analysis turns. The majority treated rising registration and turnout as the changed fact that undercut the formula, and the dissent treated the same improvement as the product of the remedy and as a reason to guard against backsliding.134 A court that reviews renewal by asking for the original degree of discrimination would penalize a remedy for succeeding. The dissent therefore concluded that the rational-means test should be easier to satisfy, and the challenger's burden higher, when the question is renewal of a remedy the Court has previously affirmed and that Congress found from contemporary evidence to be working.135
The dissent's treatment of the record shows what it meant by a remedy that works. Its yardstick was whether preclearance was still blocking discriminatory changes, and Congress found more Justice Department objections between 1982 and 2004 (626) than between 1965 and the 1982 reauthorization (490).136 Congress also concluded that Section 2 litigation could not substitute for preclearance, because litigation occurs only after the fact, when the illegal voting scheme is already in place and individuals have been elected under it, gaining the advantages of incumbency.137 An illegal scheme might run through several election cycles before a Section 2 plaintiff can gather sufficient evidence to challenge it.138 Section 2 also has no preapproval requirement; the onus is on those who object to bring suit and prove their claim.139
A second strand of evidence addressed the formula rather than the remedy. Congress learned of a report, known as the Katz study, that examined Section 2 suits between 1982 and 2004 and found that covered jurisdictions held less than 25 percent of the country's population yet accounted for 56 percent of successful Section 2 litigation, or nearly four times as many successful cases once population was controlled.140 Because the private right of action under Section 2 applies nationwide, the comparison works as a yardstick: the rate of successful suits would be roughly equal in covered and uncovered areas if the difference in risk had disappeared.141 From findings the dissent said the majority ignored, Congress reasonably concluded that the formula still identified the jurisdictions of greatest concern.142 The dissent added that racially polarized voting makes a jurisdiction more vulnerable, because a governing coalition's ordinary incentive to protect its position produces race-specific disadvantages when political preferences fall along racial lines.143
The two strands show where the opinions part ways. For the dissent the question was whether, on this record, Congress could rationally conclude that preclearance was still needed and that the covered jurisdictions were still the right ones, and it answered yes. The majority asked, as the previous subsections showed, whether the basis for selecting the burdened States made sense in light of current conditions, and it held that a basis tied to voting tests and turnout from 1972 and earlier did not, whatever the record showed.144 The dissent's complaint was that the majority never engaged the legislative record and never identified a standard of review.145
The dissent also answered the charge that the formula was frozen in 1965. In its account the Act is a dynamic statute: a jurisdiction can bail out by showing ten years of compliance and efforts to end intimidation and harassment of voters, and a court can bail in a noncovered jurisdiction on finding violations of the Fourteenth and Fifteenth Amendments.146 Nearly 200 jurisdictions had bailed out, the Justice Department had consented to every application from an eligible jurisdiction since the current procedure took effect in 1984, and courts had bailed in several jurisdictions, including New Mexico and Arkansas.147 Those mechanisms bear on the fit of the formula, since bailout lets a jurisdiction with a clean record leave and bail-in lets a court reach a jurisdiction the formula missed. The jurisdictions that stayed covered because of recent noncompliance, the dissent reasoned, confirmed that they belonged under preclearance.148
That reasoning produced the dissent's best-known image. The majority had faulted the formula for resting on decades-old data and eradicated practices, and the dissent replied that one current condition in the covered jurisdictions was the preclearance remedy itself, which Congress designed to catch discrimination before it causes harm and to guard against a return to old ways.149 In the dissent's words: “Throwing out preclearance when it has worked and is continuing to work to stop discriminatory changes is like throwing away your umbrella in a rainstorm because you are not getting wet.”150 A dry traveler shows that the umbrella is working, and the analogy treats the absence of recent discrimination as evidence of the remedy's success. The disagreement over what the umbrella proves is the disagreement over whether the Act's success is a changed fact or a continuing justification.
The same logic answered the majority's statement that it was irrational to base coverage on voting tests that had been illegal for 40 years. The dissent replied that the Act never stopped at the particular tests and devices that happened to exist in 1965, and that the 2006 record showed second-generation barriers, which dilute minority votes instead of blocking access to the ballot, emerging as substitutes for the first-generation barriers that first triggered preclearance.151 Earlier statutes that singled out particular devices failed because States replaced each one as it fell, and the dissent saw the majority repeating the assumption that a problem is solved once specific methods are identified and eliminated.152 The dissent also rejected the demand that Congress start from scratch. Congress's task differed in 1965 and 2006: in 1965 it had to find where the problem lay among a small number of States familiar to it by name, and Katzenbach upheld a formula that also swept in a few places Congress had not studied in detail, while in 2006 the question was whether to continue a remedy.153
Two further objections concerned how the Court reached the question at all. The first was that Shelby County brought a purely facial challenge, which the dissent called the most difficult to mount because the challenger must show that no set of circumstances exists under which the Act would be valid, and because a person to whom a statute may constitutionally be applied will not be heard to argue that it might be applied unconstitutionally to others.154 Under that rule the question was whether Congress could constitutionally subject Shelby County itself to preclearance, and the dissent found the answer hardly contestable: Alabama, home to Selma, had between 1982 and 2005 one of the highest rates of successful Section 2 suits in the country, second only to Mississippi, even while subject to Section 5.155 The majority's reply, discussed in the previous subsection, was that the county attacked how the formula selects jurisdictions, and that attack is not defeated by the county's own record.
The second objection concerned severability. The Act provides that if any provision or its application to any person or circumstance is held invalid, the remainder of the Act and the application of the provision to other persons not similarly situated are unaffected.156 The dissent read that clause as an instruction to sever any unconstitutional applications, such as to Arizona and Alaska, and to leave the Act in force wherever its application is valid, which made facial invalidation of the formula unnecessary.157 It rejected the majority's suggestion that tailoring the Act's reach would have the Court updating the statute itself, noting that a plurality in National Federation of Independent Business v. Sebelius treated a materially identical clause as Congress's own textual instruction to leave the remainder unaffected.158
The dissent's last objection, outlined in the earlier discussion of equal sovereignty, was that the majority left controlling precedent behind without acknowledging it. Katzenbach had confined the equal-sovereignty principle to the terms on which States are admitted to the Union, and a dictum in Northwest Austin could not silently overrule that limit.159 The point has a practical edge for defenders of a statute. Congress relied on Katzenbach in each reauthorization and had every reason to think that the Act's limited geographic scope favored its constitutionality, and it could hardly have foreseen that the Act's limited reach would render it constitutionally suspect.160
The dissent closed by calling the Voting Rights Act no ordinary legislation, because Congress embarked on a long-delayed mission to realize the promise of the Fifteenth Amendment, and by calling the 2006 record extraordinary, the product of exhaustive evidence-gathering and enacted with overwhelming bipartisan support.161 Congress judged that 40 years had not been enough to eliminate the vestiges of discrimination that followed nearly a century of disregard for the Amendment. The dissent said that this judgment, made by the body empowered to enforce the Amendment, merited the Court's utmost respect and that the Court erred egregiously by overriding it.162
Set side by side, the three opinions ask different questions of the same statute:
- The majority asks whether the basis on which Congress selected the burdened States still makes sense in light of current conditions, and it holds that a formula resting on decades-old data does not.163
- The concurrence asks the same current-burdens, current-needs question of Section 5 itself and concludes that the provision cannot be justified.164
- The dissent asks whether Congress could rationally conclude on its record that preclearance was still needed and aimed at the right places, and it answers yes under a deferential standard that is especially forgiving of reauthorization.165