Shelby County v. Holder: Preclearance may be a blunt instrument, but bailout is a sharpening stone

The following contribution to our Shelby County v. Holder symposium comes from Ryan Eric Emenaker professor of political science at College of the Redwoods, where he serves as chair of the department. His research interests focus on judicial politics especially in relation to federalism and separation of powers conflicts.

Congress’s broad powers to regulate elections derive from Article I, Section 4 of the Constitution as well as the Fourteenth and Fifteenth Amendments. Citing these constitutional provisions, in 1965 Congress passed the Voting Rights Act (VRA); the most recent update to the VRA came in 2006. The VRA represented the most comprehensive measure since 1870 to protect the voting rights of blacks. At the heart of the VRA’s dramatic powers: the requirement, imposed in Section 5, that “covered” jurisdictions seek preclearance from the federal government for all electoral changes. Jurisdictions were originally singled out for coverage based on evidence of voter discrimination in 1964. Preclearance effectively derailed jurisdictions’ ability to stay one step ahead of the next lawsuit by continually deploying new tactics to deny minorities the right to vote.

In the case now before the Court, Shelby County, Alabama, emphasizes that the massive disenfranchisement of African-American voters – which inspired prior Court decisions upholding the constitutionally of the VRA and its preclearance requirement –has significantly improved, negating Congress’s power to impose such federal oversight. Shelby County further asserts that the criterion for determining what jurisdictions are “covered” by the preclearance requirement is based on a “decades-old formula.”  Preclearance requirements based on this formula, the county claims, indiscriminately impose burdens on jurisdictions through an automatic formula that no longer bears any relationship to Congress’s interest in reducing voter discrimination.

The argument is simple and straight-forward; it is also overly simplistic and simply wrong. The formula for determining “covered” jurisdictions is not solely based on a “decades-old formula” as critics assert; instead, the bailout process, included in the VRA, gives jurisdictions without recent violations of minority voting rights the possibility to terminate coverage, thereby providing a dynamic process to limit the reach of preclearance. “Covered” jurisdictions remain subject to Section 5’s requirements not simply because of a “decades-old formulat,” but because they either (1) remain ineligible to bail out, or (2) opt to remain covered.  Because of the possibility of bailout, it is increasingly difficult to argue that Congress’s coverage formula has no rational relation to reducing infringement on minority voting rights.

In Northwest Austin Municipal Utility District No. 1 (NAMUDNO) v. Holder (2009), the Court dodged the constitutional question by expanding the types of jurisdictions eligible for bailout. This resulted in a significant increase in the pace at which jurisdictions terminated coverage. Of the jurisdictions successfully bailing out since 1965, thirty percent did so in the three years since the NAMUDNO decision. 2012 saw the largest number of jurisdictions bailing out in one year, and 2013 is set to triple that record. Because of bailout, the number of jurisdictions subject to preclearance is rapidly shrinking. Thus, the bailout process acts like a sharpening stone to refine the blunt instrument of an automatic coverage formula. Because of the existence, and practical effects, of bailout, the Court should once again uphold the constitutionality of preclearance.

Shortly after the VRA was enacted, South Carolina challenged the law, arguing that the preclearance requirement unfairly burdened the state and ran afoul of federalist principles. In South Carolina v. Katzenbach (1966), the Court upheld all contested provisions of the Act, including the preclearance requirement. The Court recognized that the preclearance requirement was an exceptional statute requiring exceptional justification, but less recognized by legal scholars is the Court’s finding that bailout provided “covered” jurisdictions with a means of protection. In the words of the Court, if the preclearance “formula is improperly applied, the area affected can always…obtain termination.” Termination, like preclearance, placed the burden of proof on the jurisdictions to defend their innocence. Despite this unusual arrangement, the Court asserted, the bailout process is “quite bearable.” The twenty-five-year extension of the VRA in 2006 used the same basic formula to determine “covered” jurisdictions, but it also retained the critical bailout provision.

Despite claims by opponents to the contrary, abolishing the preclearance requirement does not end federal oversight of elections; instead, it simply shifts some voter discrimination challenges from preclearance to after-the-fact lawsuits. Even if an electoral process was pre-cleared, nothing precludes a lawsuit against a jurisdiction for a violation under Section 2 of the VRA or the Fourteenth or Fifteenth Amendment.

Ellen Katz, a law professor at the University of Michigan, led an extensive study of Section 2 violations of the VRA from 1982 to 2004. While Section 5 requires preclearance for voting changes, Section 2 allows a plaintiff to challenge a law after adoption. Importantly, Section 2 applies to all jurisdictions, not just “covered” ones. While not a perfect parallel, Section 2’s universal coverage allows comparisons in voter discrimination between “covered” and “non-covered” jurisdictions. Professor Katz’s study showed some stark differences between “covered’ and “non-covered” jurisdictions. Of the Section 2 lawsuits that ended with merit-based decisions, 46.4% originated in covered jurisdictions. Yet, out of the 89,476 jurisdictions in the United States, only about 12,000 (13.4%) are “covered” jurisdictions. When the percentage of “covered” jurisdictions is factored in, a covered jurisdiction was more than five-and-a-half times as likely to be sued for a Section 2 violation of minority voting rights. Further, published Section 2 lawsuits are only a portion of all Section 2 claims filed; many claims are settled without a published opinion.  When published and unpublished cases are combined, eighty-one percent of the cases that resulted in favorable outcomes for minority voters originated in “covered” jurisdictions. When the percentage of “covered” jurisdictions is factored, a favorable outcome for minority voters in Section 2 claims was nearly twenty-eight times more likely to occur in a “covered” jurisdiction.

It should be remembered that the preclearance requirement blocks much of the voter discrimination that would constitute aSection 2 violation, and it also acts as a deterrent. Many proposed electoral laws are withdrawn because the DOJ would likely refuse clearance. Other changes to voting practices were modified in a dialogue between the DOJ and covered jurisdictions. In the absence of Section 5, covered jurisdictions would likely be subjected to even more voter discrimination suits.

Even with a correlation between voting discrimination and “covered” jurisdictions, coverage determined by a statistical trigger will necessarily be over- and under-inclusive. The bailout process refines coverage. Consider that, since Professor Katz’s above study was completed, more than one percent of all “covered” jurisdictions have bailed out, and more than two percent are pending termination (the entire state of New Hampshire plus a special district in California). To bail out, a jurisdiction must not have any federal violations of minority voting rights in the last ten years. As the percentage of covered jurisdictions decreases, those remaining covered will be those disproportionally engaged in violations of minority voting rights.

The empirical evidence paints a complex picture. Voter registration and turnout are less stratified by race than when the Court first upheld preclearance, and this stratification is not perfectly connected to “covered” jurisdictions. At the same time, there is a high correlation between “covered” jurisdictions and the rate of minority voter discrimination lawsuits. Importantly, this correlation is getting stronger as more jurisdictions bail out.

Critics of the idea that bailout is a sufficient protection against an over-inclusive “decades-old formula” assert that jurisdictions may decline to terminate coverage because of expense and difficulty. J. Gerald Hebert, who has represented the majority of jurisdictions bailing out since 1982, quotes the price of a bailout for small, local governments at $2,500 to $5,000 – hardly a prohibitive expense when each application for preclearance can cost $500. Hebert goes on to note that, as of August 2012, “not a single local government that has sought a bailout has been denied one.”

The high number of jurisdictions that qualify for bailout, the low expense, and the one-hundred-percent success rate indicates some jurisdictions may be making a conscious choice to remain covered. In fact, New York, California, and Mississippi (all of which are covered at least in part by the automatic formula) filed a combined amicus brief arguing that the preclearance requirement should be upheld. As the brief argues, jurisdictions receive benefits from coverage; it grants a measure of protection against lawsuits and provides DOJ feedback on proposed election law changes. Remember, federal oversight of elections continues without preclearance. Compared to the expense of defending against voter discrimination lawsuits, remaining “covered” can be a desirable choice.

Again, the empirical data in no way proves that every “covered” jurisdiction disproportionally engages in voter discrimination. In fact, empirical analysis probably fails to demonstrate that the current preclearance formula is the best conceivable. However, the Court should not hold preclearance to that standard. In City of Boerne v. Flores (1997) the Court indicated that federal legislation could prevent state laws from taking effect if there is a “significant likelihood” that the laws could be unconstitutional. Based on the distribution of Section 2 lawsuits, the blocking and deterrence features of Section 5, and the increased targeting of jurisdictions because of bailout provisions, this standard seems to be met in a reasonably justified way.

In the end, an argument for why the preclearance requirement remains constitutional should focus on bailout. The prospect of termination was cited in Boerne as part of the consideration that determines if a law is “proportional and congruent” to achieving Congress’s goals. The mere existence of termination places the preclearance requirement on firmer footing than other federal laws struck down by the Court (such as the Religious Freedom Restoration Act). Bailout also serves another important function for meeting the constitutional standard established in Boerne: the increasing number of jurisdictions bailing out leaves the remaining targeted jurisdictions as the most deserving of coverage (or those desiring coverage). The jurisdictions now “covered” by preclearance could be argued to be more directly related to the goals of Congress than the “covered” jurisdictions from just a few years ago. Thus, the actual functioning of bailout, not simply its existence, helps ensure that the preclearance requirement is meeting another part of the test under the Boerne standard; preclearance is increasingly confined to those regions where violations of the law are most flagrant.

The standard developed in Boerne is an important consideration when contemplating how the Court might interpret the VRA. However, it is not the only insight into a constitutional challenge. Both the Family and Medical Leave Act and the Americans with Disabilities Act were upheld by the Court after Boerne. They were deemed constitutional because they protected suspected classes or fundamental rights. The VRA does both. Further, Congress can cite Article I, Section 4 for most of the elections it “intrudes” upon, and for others it can cite both Fourteenth and Fifteenth Amendment authority. This seems to indicate that the VRA is at the apex of congressional power. When all these factors are considered, it is a strong legal argument for the Court to uphold Section 5 of the VRA. (I discuss this data and argument in more detail here.)