The Yale Law Journal


Mapping a Post-Shelby County Contingency Strategy

07 Jun 2013

Many supporters of section 5 of the Voting Rights Act fear (and both supporters and opponents of section 5 expect) that the Supreme Court will invalidate the provision on constitutional grounds in its forthcoming opinion in Shelby County v. Holder.1 Some voting rights activists are predicting a post-Shelby County apocalypse if the Court strikes down section 5. They believe that state actors, especially in jurisdictions currently covered by section 5, will once again engage in rampant racial discrimination against voters of color.

These predictions of doom and gloom are understandable for at least two reasons. First, as a litigation strategy, predictions of doom and gloom might encourage judicial restraint and respect for the work of a coequal branch of government. Second, given congressional polarization and gridlock, supporters of section 5 cannot be optimistic that Congress would enact a Shelby County fix. There is a significant likelihood that the Court’s word here will be the last on this issue for a long time to come. If one believes that section 5 remains a valuable tool for preventing the implementation of discriminatory voting rights laws by covered jurisdictions, a decision concluding that section 5 is unconstitutional would be deeply problematic on that ground alone.

Although we do not think the Court should strike down section 5—because we think it is readily apparent that Congress has the power to continue to reauthorize the statute2—we are also of the view that supporters of section 5, and the voting rights community in particular, ought to more explicitly and forthrightly map out a post-Shelby County strategy. It is understandable that voting rights activists would, and perhaps should, fight aggressively to preserve hard-won gains secured, literally, through blood, sweat, and tears. And yet, as politically incorrect as it may sound, we should all be sober-eyed about the need to secure meaningful voting protections in the face of adverse changes. More specifically, we should be prepared—because of developments in constitutional law, or politics, or political practice—for a future in which section 5 is not part of the voting rights landscape.

The twenty-first century presents voting rights activists and scholars with two different frameworks for securing and protecting voting rights. The first framework is essentially the centralized regulatory structure that is quite familiar to voting rights activists and scholars. For ease of exposition, we term this framework “the public protection model.” Under this model, Congress identifies both violators and violations. More specifically, it deploys positive law and uses the courts to closely monitor violators and prevent or remedy violations. Also, the primary actors are public ones: government officials, courts, and law. This is the world within which section 5 currently operates and the world that some voting rights activists are trying to preserve.

The public protection model has many advantages, chief among them being its ability to regulate uniformly and generate broad compliance. This advantage is so critical, in fact, that one might think of this model as election law of the first best. But the public model also has its weaknesses. Chief among these weaknesses is the model’s static regulatory structure. Because of its stasis, it is a less effective regulatory framework when it is regulating a dynamic process—and electoral politics is nothing if not dynamic.3

The second framework, which is admittedly incipient and certainly underdeveloped, relies upon an emerging and fragile ecosystem of private entities, non-judicial institutions, and organized interest groups of various stripes, which together are willing and able to mimic the elements that made section 5 an effective regulatory device for protecting the rights of voters of color. In this model, the primary actors are private or civic institutions. We term this “the private protection model.” In other contexts, these civil society or third-party groups might be identified as nongovernmental entities. They include public-interest groups, advocacy organizations, political parties, political committees and the like. For ease of exposition we broadly identify them here as institutional intermediaries.

This private model also has some disadvantages. In the private protection model, different parts of the ecosystem might perform differently. To the extent that the private model may not be able to generate broad compliance or uniformity and to the extent that we value uniformity in coverage and compliance, one might think of the private protection model as the election law of the second best. The model may indeed produce a fair amount of variance among institutional intermediaries, but variation may be one of the strengths of the model as well. Because the ecosystem is varied, it is dynamic and adaptable. Institutional intermediaries are unlikely to respond similarly, but we may not want them to.

In order to make sense of these two frameworks, consider the run-up to the 2012 presidential election, when many states enacted laws requiring voters to present identification prior to voting. Many of these laws were fiercely challenged, both in the legislature and before the courts. In South Carolina, for example, the United States Department of Justice interposed an objection to the state’s law—yet the District Court for the District of Columbia ultimately ruled that the law could go into effect in 2013,4 so long as election officials adhere to what the court called an “extremely broad interpretation” of the contested provisions.5

This response is an example of the public protection model. It uses formal law, interpreted and enforced by public officials, to respond to voting rights issues. But this need not be the only response.

By contrast, in many other states, civic, partisan, and ideological organizations mobilized to support minority voters in their quest to acquire the necessary photo identification to comply with legislative restrictions.6 Consider another example. When a county in Arizona sent a postcard to Spanish-speaking voters that contained inaccurate information about the 2012 election, civic and ideological organizations swiftly engaged in a campaign to provide voters with accurate information.7 Similarly, in North Carolina, the Southern Coalition for Social Justice—a Durham-based nonprofit organization that “partners with communities of color and economically disadvantaged communities in the south to defend and advance their political, social and economic rights”8—has played an important role in fighting recent attempts by the state to, among other things, limit early voting, promulgate a voter identification requirement, and eliminate same-day registration.9 One might also consider the broader context of the 2012 election. President Obama’s campaign, supported by the Democratic Party, assumed both the cost and the responsibility of protecting voters of color, in particular, against what the campaign viewed as attempts by various states to disenfranchise them. The campaign filed lawsuits, educated voters, mobilized its supporters, negotiated with election officials, and did whatever was necessary to ensure that these core voting blocs were able to register and vote. Most importantly, they engaged voters themselves in the task of protecting their own rights.

The appeal of this approach is obvious. It treats voters as central democratic agents. It does not view the task of democracy as simply showing up on Election Day and pulling a lever. While bearing some of the costs of collective action, the private protection model also mobilizes voters to become more engaged citizens, in the manner that classical democratic theory endorses.

Due in large part to the political equality made possible by section 5, which helped remove first-generation barriers to political participation,10 we are living in a different institutional and political ecosystem than the one that existed in 1965. This is not simply because state actors in covered jurisdictions may be less racist or because white voters in covered jurisdictions are less racist and may be willing to support candidates of color. Rather, the ecosystem has changed because we have non-judicial, civic, ideological, and partisan institutions that have both the incentive and the capacity to protect voters of color against state action that would have a disparate impact on their voting rights. These institutional intermediaries may be well positioned to address the vexing questions posed by voting rights policy in a period of transition: How much racial discrimination do we have? Who are the bad actors? Should we have a universal voting rights policy, a race-based one, or one that protects against disenfranchisement motivated by partisanship and ideology?

As voting rights activists plot a post-Shelby County contingency strategy, they should both account for institutional intermediaries and think about the types of changes that could enhance the ability of these groups to better protect voting rights, now and in the future. It may be the case that, if the Court strikes down section 5, we would see a retrenchment on voting rights. It may also be the case that by focusing less on the courts, and the Supreme Court in particular, we might better account for the importance, adaptability, and responsiveness of third-party groups. In turn, this might ameliorate, if not completely blunt, the retrogressive impact of a world without section 5.

I. Five Critical Functions of Section 5

To understand why institutional intermediaries might be equally if not better suited to protect voting rights, we must first briefly rehearse the critical functions or elements that have made section 5 an effective regulatory framework.

Section 5 of the VRA is widely recognized as the most effective provision in one of the most, if not the most, effective civil rights statutes that Congress has ever enacted.11 The regulatory framework is an ingenious scheme of interlocking measures that contain at least five critical properties which, working together, have contributed to the Act’s legendary effectiveness. First and most obviously, the Act uses race to cabin voting rights violations. The VRA is not a statute that protects broadly against voting violations, but a statute that protects specifically against discrimination against racial or language minorities in voting. The VRA was enacted against a background of rampant racial discrimination and political exclusion. There was no political competition for the black vote—in fact, there was very little political competition at all in the one-party South, where the Democratic Party reigned supreme. Voting discrimination was a subset of racial discrimination, which pervasively infected all elements of black life.

Second, sections 4 and 5 of the VRA used geographic targeting to distinguish the jurisdictions that practiced the most egregious forms of racial discrimination in the political process. The identities of the jurisdictions that engaged in impermissible discrimination were largely self-evident and widely known, and the purpose of sections 4 and 5 was to rein those jurisdictions in. The default impulse of these jurisdictions was to discriminate on the basis of race.

Third, by freezing in place the electoral practices of the covered jurisdictions, section 5 imposed a very strong status-quo bias. This bias in favor of the status quo was reinforced early in the VRA’s history by the Court’s decision in United States v. Beer,12 which interpreted section 5 to prevent retrogressive changes. This meant that covered jurisdictions could not make voters of color worse off through changes in their voting laws, though they did not need to make them better off.

Fourth, by requiring covered jurisdictions to preclear voting changes, section 5 created an information-eliciting mechanism. The preclearance requirement forces the institutions with the best information about potentially discriminatory practices to share that information with third parties. Preclearance thus facilitates monitoring through disclosure. And, recognizing that disclosure alone is insufficient, preclearance maximizes the Act’s effectiveness by preventing disclosed changes from going into effect.

Moreover, because central regulators are always at an epistemic disadvantage vis-à-vis local officials, they need local proofers who can confirm the information provided by a jurisdiction’s officials. In this case, the proofers are the local minority community, whose assent to or dissent from the submission by the covered jurisdiction will affect the preclearance chances of the submission. As Pam Karlan has argued, this is a positive externality; it provides minority voters a bargaining chip in their dealings with local officials with respect to any electoral changes that are subject to preclearance.13

Lastly, section 5 employs a burden-shifting device, which transfers the cost of epistemic uncertainty to the state. Under section 5, covered jurisdictions are required to show that their electoral changes are not discriminatory in their purpose or effect. As some members of the Court have complained, this puts a significant burden on covered jurisdictions, as it requires them to prove a negative.14 But the justification for compelling covered jurisdictions to bear the cost of epistemic uncertainty follows from the operating assumption that the covered jurisdiction is likely to discriminate. Thus, when we are unsure whether an electoral change is discriminatory—that is, when we are operating under conditions of epistemic uncertainty—we can assume that the jurisdiction is engaged in discrimination and we will be right more often than not. The burden-shifting framework reinforces the status-quo bias and evinces our deep distrust of the covered jurisdictions.15 In addition, section 5 reflects a regulatory scheme where we are more concerned with false negatives than false positives.16 Because false negatives may be more costly than false positives, section 5 uses the burden of proof to account for this asymmetry.

These five elements have been crucial to the success of the VRA. They have proven to be relatively stable—until recently. Indeed, an underappreciated condition of the VRA is stasis. The VRA’s general orientation is to be leery of change. Historically, change meant discrimination. A justifiable distrust of electoral dynamism is therefore built into the DNA of the VRA. Sections 4 and 5 have generally targeted roughly the same jurisdictions, using roughly the same formula, pursuant to the same regulatory structure, for well over forty years. Stasis is the general orientation of the VRA’s regulatory framework; this fits well with a centralized regulatory mechanism, which is very effective in a static environment.17

However, like waves crashing against large rocks in the ocean, the dynamic political process repeatedly crashes against an inert section 5. This perpetual clash of kinetic politics against static regulations has contributed to the destabilization of the underlying properties that once made section 5 an unassailable regulatory framework. The current political and constitutional controversy over section 5 is fundamentally a reflection of the fact that each of the properties that supported the section 5 framework has become unstable, and that the background assumptions supporting these properties are much less tenable today than they once were.

Start first with racial delimitation. In 1965, voting discrimination was racial discrimination. States intentionally targeted Black voters and excluded them from the political process. In the current era, although we may still have a problem with racial discrimination in voting, we cannot say with any amount of certainty that the central problem of voting is race.18 When racial discrimination manifests itself as intentional discrimination, we have adequate tools for dealing with it, specifically the Fourteenth and Fifteenth Amendments, which were not sufficiently enforced in 1965. The issue then is how to deal with laws that have a disparate impact on the basis of race. The problem with the current framework is that there is a lack of consensus about how to categorize such laws. Should they all be suspect under either section 5 or section 2? Are only some of them suspicious? If so, how do we distinguish among them?

Second, with respect to geographic targeting, when Congress enacted section 5 in 1965, it was very clear which jurisdictions were the most egregious offenders. Moreover, that determination was fairly static. For example, Mississippi had long attempted to deny African Americans the right to vote and its resolve was not going to shift from election to election. Its discriminatory methods were expected to shift, but that was the reason to put the preclearance requirement in place. Today, however, we are less certain which jurisdictions are engaging in, or are more likely to engage in, racial discrimination. Ought we be more worried about Ohio (not covered) than Virginia (covered)? How different are Arizona (covered) and Indiana (not covered)? However one resolves the coverage question, it is clear that we are operating on a narrower scale in 2013 than we were in 1965. In 1965, the geographic differences were truly differences in kind and not degree. Today, they may be fairly characterized as differences in degree and not in kind. History may not inform our calculations the way it once did.

Relatedly, the status-quo bias that was justified by the default assumption of discrimination is less stable today than it was in 1965. When we could safely assume that a covered state was much more likely than not to discriminate against voters of color, it made sense to shift the cost of epistemic uncertainty to the state. More importantly, it made sense to do so by reinforcing the status quo, which the pro-civil rights bar could prefer over an electoral change whose implications would be unclear to those that favored voting rights but presumably clear to the state. Notwithstanding the fact that covered jurisdictions continue to engage in racial discrimination, we cannot say that the default mode of the state is to discriminate. Indeed, even where we have instances of discrimination, the status-quo bias is difficult to justify as long as the discrimination is episodic.

This is why evidence of current discrimination by covered jurisdictions is not responsive to today’s debate over section 5. The question is not whether there are instances of discrimination; rather, the question is whether racial discrimination is frequent enough that we can assume that suggested changes are more likely to be discriminatory than to be racially neutral. Is there a sufficient probability that any electoral changes are motivated by racial discrimination or will have a discriminatory impact? This is a much harder inquiry today than it was in 1965.

Third, at the time the VRA was enacted, the information-eliciting mechanism furnished by preclearance was vital. This is because the covered jurisdictions had access to information about their electoral changes—including their motivations for implementing such changes—that were less accessible or simply inaccessible to anyone else. Moreover, covered jurisdictions were almost always in the best position to understand the likely impact of their actions. They possessed the relevant data and they were intimately familiar with their local jurisdictions. If the United States was going to monitor discrimination, it had to get the information from the parties best positioned to provide it. Additionally, because the federal government reasonably assumed that the default impulse of the state was to discriminate, it was justified in making its discovery request as broad as possible to elicit the most information.19 The information that section 5 elicited, primarily in the form of preclearance requests submitted to the Department of Justice, was manageable because it was only coming from a few jurisdictions.

Today, the information-eliciting function of preclearance is the least stable of the five elements of the VRA. This is, of course, because, as Shelby County v. Holder shows, preclearance is viewed, rightly or wrongly, as the most aggressive mechanism specified by the VRA for eliciting information. The irony here is that the information-eliciting feature of section 5 is more valuable than ever.20 This is because, with respect to section 5, we currently find ourselves in a period of deep uncertainty. We have certainly made a lot of progress in eliminating racial discrimination in voting. But have we made enough progress to declare victory, withdraw the troops, and go home? Accurate information is the most relevant and valuable commodity needed to make that assessment. How bad are the remaining bad actors? How relentless are they? Will they return if we withdraw? These are the types of questions that we need answered; but, unfortunately, we are operating from an epistemic deficit when we attempt to do so.

If the Court strikes down section 5, it will do so in the formal language of constitutional law. But the underlying reason for section 5’s demise will be that the properties that have sustained section 5 have become too unstable to support its continuing viability.

II. Third Party Institutions and a Post-Shelby County VRA

The big question, of course, is this: if the Court strikes down section 5, what is the future for voting rights policy? Among supporters of section 5, the prediction is doom and gloom. But the story of doom and gloom relies upon a view of the Supreme Court as a singularly consequential actor whose decision will facilitate or impede racial equality in voting. From our perspective, this story is too court-centric. More importantly, it fails to fully account for the increasing importance of institutional intermediaries. These intermediaries are capable of offering—and in fact do offer—effective solutions to problems in law and politics. While courts remain important institutions, and serve as complementary partners to institutional or organizational intermediaries, courts are not the only game in town.

Building upon the new institutionalism literature,21 we explore here the potential for institutional intermediaries to effectively mimic the section 5 regime because of their adaptability as dynamic institutions. Institutional intermediaries operate in a competitive environment where they vie with each other to achieve particular objectives, whether civic, ideological, or partisan. To borrow from Heather Gerken and Michael Kang,22 institutional self-interest fuels a vibrant institutional ecosystem, which facilitates competition by institutional intermediaries for the attention of the electorate. We use this framework to offer an alternative narrative and to sketch a way forward in a post-Shelby County world.

As we explained above, section 5 relies upon a centralized regulatory mechanism to: (a) identify the problem (racial discrimination); (b) identify the bad actors (primarily Southern states); (c) make change costly (by freezing in place electoral practices); (d) elicit information (through the preclearance device); and (e) address the problem of epistemic uncertainty (through the burden-shifting device). Instead of the centralized regulatory mechanism of section 5, institutional intermediaries rely upon the evolutionary dynamism of institutions fueled by political self-interest to perform the same functions. But because of their dynamism and adaptability, organized interests have the potential to perform those functions much more efficiently than centralized regulation. If organizational intermediaries are to survive in the political marketplace, they need to be able to perform these five functions, which were critical to the success of section 5.

Institutional intermediaries, which are generally adaptive, seem poised to do exactly these things. They are like amoebas, in that they have a remarkable ability to change their shape in response to the structural challenges of the political process. If they cannot make credible representations to their constituents and provide them with valued services, they will not survive for long. Ultimately, interest groups must deliver something to their constituents. If they cannot deliver, they will fade away.

To be sure, not all groups will survive and not all will be equally effective. For example, some will be wrong about the extent of racial discrimination in voting. Some might make a wrong bet on the location or identity of bad actors. Some might not be skilled at acquiring relevant electoral information and disseminating that information to their constituents. Those that are ineffective will be ignored or will not survive. But the larger point is that politics can be used to help fix politics in a post-Shelby County world in which section 5 is no longer available.

Consider some examples. Prior to the 2012 election, a number of states passed laws requiring voters to provide photo identification at the voting booth. These states included localities that are covered by section 5, such as Alabama, Mississippi, South Carolina, Texas, and Virginia, as well as non-covered states, such as Kansas, Pennsylvania, Rhode Island, and Wisconsin.23 In addition to voter identification laws, many states passed laws limiting early voting, limiting same day registration, and making it harder for third parties to conduct voter registration drives to register eligible voters.24 Many civil rights activists predicted that these and similar laws, either singularly or in combination, would disenfranchise otherwise eligible voters of color.

As it turned out, however, the 2012 election saw record turnout rates among voters of color.25 For example, Black turnout in 2012 equaled and likely surpassed that of white voters.26 Moreover, although African Americans constituted twelve percent of the population, they were thirteen percent of the voting public in 2012.27 Similarly, Latino voters turned out to vote in record numbers in 2012 and constituted ten percent of the voting public.28 About fifty percent of eligible Latino citizens voted in the 2012 election. Asian American voters also increased their turnout to a record three percent in 2012.Taken together, the 2012 electorate was the most diverse in American history. Voters of color constituted twenty-eight percent of the total turnout in 2012. To the extent that the electoral changes prior to the 2012 election were intended to deter voters of color, they certainly did not have the impact that their supporters hoped for and their opponents feared.

What accounted for the failure of the predictions that these laws would suppress the votes of people of color? One reason is certainly the fact that some of these laws were challenged in courts, both federal and state, and some courts enjoined their enforcement or declared them illegal. The DOJ and the preclearance mechanism also played a role, particularly in South Carolina and Texas, where the federal government initially enjoined the state’s voter identification law from going into effect.29 But other laws went into effect, including some that were precleared by the DOJ, which did not end up having the disparate impact predicted. An important part of the story is the fact that various types of institutions, partisan and civic, filled the vacuum and spurred voters to action.

In the Latino community, for example, a diverse collection of institutional actors led the voter mobilization effort. From the National Council of La Raza (NCLR) to the United We Dream network and Promise Arizona, institutions took it upon themselves to ensure that the recent suppression efforts would not have a deleterious effect on Latino voting. For example, NCLR launched the “Mobilize to Vote” campaign with the goal of registering new Latino voters.30 Their strategies included door-to-door canvassing in Latino neighborhoods as well as working in high schools, community colleges, and universities.The NCLR Affiliate Network of service-providers and advocacy organizations also took part in the registration effort by helping to register their clients to vote.31 Closer to Election Day, NCLR contacted registered Latinos to encourage them to cast their votes.

The case of Arizona, and of Maricopa County in particular, is illustrative. In the months before the 2012 election, Latino groups had been actively registering new voters and encouraging those who had already registered to request early mail-in ballots.32 This level of engagement was part of a larger response by Latino groups to what they viewed as anti-Latino policies, such as Arizona’s notorious immigration law, SB 1070, and the actions of Maricopa County Sheriff Joe Arpaio. Two weeks before the election, it came to light that Maricopa County had printed the wrong date for Election Day on Spanish language materials. This error galvanized Latino activists. According to Randy Parraz, head of Citizens for a Better Arizona, “That’s definitely helped . . . . People are more inclined to vote now.”33 Similarly, Petra Falcon, the head of Promise Arizona, said that “[w]hen SB 1070 hit, we saw a lot of energy. Now, with this error, you are seeing the energy going up again.”34 Election Day turnout reflected these diverse efforts and influences, as Latinos in Arizona voted in record numbers. After the election, and as the fight moved from mobilizing turnout to ensuring that the votes were in fact counted, these groups shifted their energies accordingly.35

Unsurprisingly, a similar dynamic took place in the Black community, where the NAACP, the National Urban League, churches, and community leaders played a significant role in driving turnout.36 For example, in early 2012, the National Urban League launched its Occupy the Vote campaign, an effort designed to fight the new spate of restrictive election laws across the country while at the same time educating, registering, and motivating voters in danger of being disenfranchised by these laws.37 This effort deployed volunteers to knock on doors and coordinate community events to educate and register voters. Forty days before the election, the League intensified its efforts, unveiling video, radio, and print advertisements featuring celebrities from the Black community.38

Similarly, the NAACP engaged in an aggressive registration effort entitled “This is My Vote!” that culminated in the registration of 432,935 voters.39 This was the largest registration total in the group’s history. It accomplished this feat by making use of a wide range of media, from an online registration website and a voter registration hotline to a robust mail program. According to NAACP President Benjamin Todd Jealous,

  We did it brick-by-brick . . . . [W]e made sure that our folks used that database to target people who needed to be signed up to vote. We went out there with a plan that we had written a year ago for how we are going to move voter registration rolls in the Black community up in every single state. We even had a target for Alaska. [W]e have registered 3.5 times as many people this year as we did in 2008. And today, we moved 2.5 times more people this year than we did in 2008—despite voter suppression, despite voter intimidation, we met the challenge of community that was ready to be mobilized.40  

Once its registration program had taken place, the group turned its sights to ensuring that these voters would turn out on Election Day. In the words of Marvin Randolph, the NAACP’s Senior Vice President for Campaigns, “our job will not be finished until the last call is made, the last door is knocked, the last ride to the polls is provided, and every polling place is closed.”41 Their efforts were rewarded with record levels of Black voter turnout.42

Consider also, and more generally, the Brennan Center’s role as a producer of the information necessary for monitoring state actors.43 Among its many tasks, the Brennan Center monitors state legislatures and reports on proposed legislation that would make it harder for people to vote.44 The Brennan Center publishes on its website “a regularly-updated, comprehensive roundup of introduced, pending, active, and passed voting laws.”45 Note how the Brennan Center’s research mimics the information-gathering function of preclearance. Institutions like the Brennan Center understand that information is an important commodity. They view it as their institutional mission to gather that information and share it with the general public. Here, their mission is aligned with voters’ interest in learning about potentially discriminatory laws before they are enacted, and mobilizing against state laws that impede their right to vote.46

These groups also fill a significant electoral vacuum. They provide information about voter registration requirements where such requirements have changed from earlier election cycles. They provide information on early voting. They inform voters about the types of identification that they need to satisfy the requirements of voter identification laws old and new. Voting is not an easy practice, and in this new climate, many states are not interested in lowering the barriers of entry—instead, they are demonstrably interested in doing precisely the opposite. As a result, we see the rising costs of turning out to vote. It is precisely here where these groups fit in. The fight over the right to vote is now becoming a grassroots fight.47 These groups respond by helping bear the cost of providing information to voters, ensuring their eligibility, and turning them out to vote. More importantly, they support mass democratization by mobilizing voters and engaging them in the nitty-gritty process of democratic self-government.48

Conclusion: Election Law of the Second Best?

In a recent essay, Pam Karlan argues that an important task for scholars is to “develop a more affirmative vision of the right to vote, one in which the government takes an active responsibility for ensuring that all citizens have full access to the political process.”49 Taking voter registration as one example, she describes the process in the United States, which “uses a decentralized system that places the burden on individual citizens to register and leaves to individual states the responsibility for updating voting rolls to respond to changes in address among a highly mobile population.”50 She then contrasts that process with its counterpart in Canada, where “the national government for many years conducted a ‘door-to-door enumeration’ before every federal election, to make sure that all eligible citizens were able to participate; it moved away from this system only once it had developed a national database with systematic updating.”51

Professor Karlan presents us with a telos for election law, practice, and scholarship. Perhaps that telos represents the ideal aim for these undertakings. But it is not without a competitor. Election law scholars and voting rights activists in the twenty-first century may be presented with two different visions of the future of election law practice, doctrine, and scholarship. The first vision is the public protection model, in which Congress or the Court is relied upon to vigorously protect voting rights. In the election law of the first best, for example, Congress promulgates universal registration laws and updates section 5 of the VRA as needed.

In this Essay, we offer an alternative vision, which we term the private protection model. This model relies on civic, partisan, and ideological groups competing for votes, attention, dollars, and affection to vigorously enforce voting rights. These groups will register voters, drive them to the polls, and assist voters in overcoming barriers to political participation, such as obtaining proper identification. They will provide voters with the information needed to hold their elected officials accountable. In sum, they will mobilize them towards political action.

It is clear that the private protection model is not perfect. For example, it is not systematic or comprehensive, which means that some voters may fall through the cracks. But it is the perfect response to an unpredictable and dynamic political process. Unlike centralized, top-down regulation, third-party institutionalism is both top-down and bottom-up. It engages voters in the democratic process and it empowers them to protect their rights.

There is much more to say and to map out, and we expect to say more about this in the near future. We have sketched out a descriptive argument here, which needs more empirical support and leaves important normative and doctrinal questions for later exploration. While we have not yet worked out our normative positions on the private protection model and institutional intermediaries, we are attracted to the institutional intermediary story because it treats voters of color as autonomous democratic agents and not as passive or inert democratic actors. The institutional intermediaries model presents and preserves the possibility of democratic change through democratic action—social movements and civic and political organization.

Section 5 of the VRA was once the most effective civil rights statute in our nation’s history. But it is increasingly clear to us that section 5 is not long for this world. Whether it is felled by the Court in Shelby County or made increasingly obsolete by voter suppression strategies that have become far too sophisticated, a new solution will be necessary sooner or later. In a post-Shelby County world where the Court has struck down section 5 of the VRA, organizational intermediaries present a tantalizing option for protecting our voting rights. We have provided a sketch in this Essay that illuminates the potential power and influence of these intermediaries—and in doing so, we hope to have opened up a new line of inquiry on the future of voting rights after Shelby County.

Guy-Uriel E. Charles is the Charles S. Rhyne Professor of Law, Duke Law School. Luis Fuentes-Rohwer is Professor of Law and Harry T. Ice Faculty Fellow, Indiana University Maurer School of Law. We are grateful to James Gardner, Heather K. Gerken, Richard L. Hasen, Michael S. Kang, Ellen Katz, Justin Levitt, and Richard H. Pildes for extremely useful comments on this Essay. We are also thankful to Ben Eidelson, Bridget Fahey, and Ryan Thoreson of the Yale Law Journal for superb editing.

Preferred citation: Guy-Uriel E. Charles & Luis Fuentes-Rohwer, Mapping a Post-Shelby County Contingency Strategy, 123 Yale L.J. Online 131 (2013),