Skip to navigation – Site map

HomeIssues1The Voting Rights Act at 50Do The Facts of Voting Rights Sup...

The Voting Rights Act at 50

Do The Facts of Voting Rights Support Chief Justice Roberts’s Opinion in Shelby County?

Morgan Kousser

Abstracts

In June, 2013, a 5-4 majority of the U.S. Supreme Court brought to an abrupt and likely permanent end the most important provision of the most successful civil rights law in U.S. history. Initially passed in 1965, Section 5 of the Voting Rights Act required “covered jurisdictions,” at first in the Deep South and later extended to Texas, Arizona, Alaska, and certain counties and townships in other states, to “pre-clear” any changes in their election laws with the Justice Department or the District Court of the District of Columbia before putting them into effect. Laws that changed the political structure – for instance, redistricting laws, annexations, and shifts from district to “at-large” elections for local governments – were restricted, as well as provisions and practices that directly affected individuals’ rights to vote. While acknowledging the success of the law in greatly increasing the number of African-American and Latino elected officials, Chief Justice John Roberts contended in his majority opinion in Shelby County v. Holder that the problems of 2013 were much less grave than the “pervasive . . . flagrant . . . widespread . . . rampant” voting discrimination of 1965 and that the coverage formula was outmoded because “today’s statistics tell an entirely different story.”1
Neither the Chief Justice nor any scholars or civil rights proponents or opponents have systematically examined the evidence on the entire pattern of proven voting rights violations over time and space. Was the Chief Justice correct in asserting that such violations no longer tracked the coverage scheme in Section 4 of the Act -- that, as he put it, the relationship of the formula to problems of vote dilution was purely “fortuitous?” Had the number of violations diminished so much in the years leading up to the 2006 renewal of Section 5 that Congress should have ended preclearance altogether because discrimination had basically disappeared? If the number of voting rights lawsuits has diminished, why is that so?

Based on the largest database of voting rights “events” – successful lawsuits, Section 5 Justice Department objections and “more information requests,” and consent decrees or settlements out of court that led to pro-minority changes – ever compiled, this paper provides a unique overview of the history of U.S. voting rights from 1957, when the first U.S. civil rights law in 82 years passed, through 2013. It shows that the Chief Justice’s factual assertions were incorrect, that the coverage formula was still congruent with proven violations, and that to the extent that recorded violations had decreased, that was not because problems had ended, but because the Supreme Court had made it more difficult to win lawsuits.

Top of page

Full text

The author wishes to thank Richard Pildes for insightful comments on an earlier draft of this paper, Adam Adler and Matthew Smalley for excellent research assistance, and Caltech for support in creating the database. Matt Smalley created the wonderful maps.

Introduction

  • 2 South Carolina v. Katzenbach, 383 U.S. 301, 308, quoted in Shelby County v. Holder, at 2628.
  • 3 811 F.Supp. 2d 424 (D.D.C. 2011).
  • 4 679 F.3d 848 (C.A.D.C. 2012)
  • 5 Shelby County v. Holder, 133 S. Ct. 2612 (2013) at 2622, 2629.
  • 6 Shelby County, at 2628.
  • 7 Shelby County, at 2626. The Chief Justice did not note that these are self-reported registration ra (...)
  • 8 The comparison between the percentage of Department of Justice objections to changes in election la (...)

1In the Supreme Court’s first consideration of Section 5 of the Voting Rights Act (VRA), Chief Justice Earl Warren began by remarking that “The Constitutional propriety of the Voting Rights Act of 1965 must be judged with reference to the historical experience which it reflects.”2 Although Chief Justice John Roberts quoted the passage from Katzenbach in his 2013 opinion overturning the coverage formula in Section 4 of the Act, he took a very different stance toward history than his predecessor did. Devoting only two short sentences to the painstaking 84-page opinion of federal district court judge John Bates3 and only seven more to the thorough 32-page majority opinion of the Court of Appeals for the District of Columbia by Judge David S. Tatel,4 Chief Justice Roberts dismissed the 15,000-page record compiled by Congress, which the lower courts discussed extensively, as irrelevant because “Congress did not use the record it compiled to shape a coverage formula grounded in current conditions.”5 “History,” Roberts reminded us, “did not end in 1965. By the time the Act was reauthorized in 2006, there had been 40 more years of it.”6 Yet apart from comparing voter registration rates in 1965 and 2004 in six Deep South states7 and making a brief, misleading reference to the rate of Department of Justice objections to election law changes,8 the Chief Justice ignored that history.

  • 9 Shelby County, at 2617.

2What if we delve into the history that Chief Justice Roberts disregarded? What if we look at where proven violations of the VRA and related laws and constitutional provisions actually took place and at the course of those violations over time? How do the geographical and temporal patterns from the years shortly before the passage of the Act in 1965 through the years after its latest renewal in 2006 reflect on the adequacy of the Section 4 coverage formula that the Chief Justice summarily rejected as “based on decades-old data and eradicated practices?”9 What can we learn about how voting rights law has actually worked by arraying the patterns of legal actions involving minority voting rights in maps and charts? Although Chief Justice Roberts’s opinion rested entirely on his assertions that voting rights violations had severely declined and that they were no longer concentrated in jurisdictions covered under Section 4 of the VRA, he only briefly and superficially examined the “historical experience” that he and Chief Justice Warren before him considered key to “the Constitutional propriety of the Voting Rights Act.” When we examine that experience in detail, will we reach the same conclusions that Chief Justice Roberts announced in Shelby County?

3To answer questions about the adequacy of the Section 4 coverage scheme, as well as about the nature of and reasons for the patterns of legal events involving voting rights, I have compiled the largest database on such events in existence. Drawn largely from lists of cases and other actions compiled by civil rights organizations, individual attorneys, and the Department of Justice, the database has been supplemented by ferreting out details and following case citations from published cases, from PACER, and from newspaper articles. The events include any successful or unsuccessful case, published or not, decided or settled; Section 5 objections and “more information requests” by the Department of Justice; and election law changes that manifestly took place as a result of the threat or reality of legal challenges. The sources are described in Appendix A to this paper, and the methods of expanding the information from cases and objections that appeared in the lists are described in Appendix B. Collection and manipulation of the data began in 2009, and it continues. The data, which starts with the first successful modern voting rights law, the 1957 Civil Rights Act, and proceeds with some confidence into 2006 and then no doubt less comprehensively into 2014, is incomplete and contains (unintentional) errors and omissions. Nonetheless, it is far larger than any of the single sources that were presented to Congress during the process of renewing the VRA in 2006 and that were scrutinized in the district and appeals court opinions in Shelby County. For example, Prof. Ellen Katz’s database of Section 2 cases, discussed extensively during the 2006 congressional hearings and included in my database with Katz’s kind permission, contains 324 cases. The total number of cases and other events in which the minority side was successful contained in my database is currently 4173.

Mapping the Voting Rights Act

  • 10 Alaska is so large that including it in anything like its proportional size reduces the detail of t (...)

4The most dramatic way to assess the adequacy of the coverage scheme is to consider a series of maps. Map 1 looks at every county in the 48 contiguous states.10 Each county in which no event took place during the entire 57-year period is colored blue, while the projections extending above every county that experienced a “successful” legal event – i.e., an event in which the minority side won – is proportional to the number of events in the county over the whole period. Red counties experienced the largest number of events, orange counties the second largest, and so on, down to blue-gray, for counties with one or two events. The total number of county-level events was 3839. The 334 events that took place on the state level, such as challenges to state redistrictings, are not included in the county-level maps, but they are taken into account in the graphs.

5The conclusion from Map 1 is simple to state. The number of high-rises over the South, the focus of the Section 4 coverage scheme, dominates the real estate. Outside the South, only New York City, where there were 137 precinct boundary changes in 2001 that were objected to by the Justice Department, and California, where I have included numerous events connected to the 2001 California Voting Rights Act, contain many tall buildings.

  • 11 Of 3669 county-level events that took place before 2009, when large numbers of local jurisdictions (...)

6Map 2 outlines the Section 4 coverage scheme that the majority of the Supreme Court overturned in Shelby County, an outline that will be familiar to most readers of this article. Comparing it to the Map 1 tracing of the counties where an election rule or law was overturned, the coverage scheme looks to have been quite precise. That is, it accurately predicted the locations of the proven voting rights violations over the period, the vast majority of them since 1965, to which Chief Justice Roberts referred. Of the 3839 county-level events, 3519 or 91.7 % took place in covered jurisdictions.11

Map 1: Voting Rights Events by County, 1957-2014

Map 1: Voting Rights Events by County, 1957-2014

Map 2: Counties Covered Under Section 4 at the Time of Shelby County v. Holder

Map 2: Counties Covered Under Section 4 at the Time of Shelby County v. Holder

7But not all counties are of the same size. The 19 events that took place in Dallas County, Texas, with a voting-age population in 2010 of 1,713,876, should perhaps not count as much as the 10 events in Dawson County, Texas, which had a voting-age population in 2010 of 10,411. Map 3 divides the number of events in each county for the whole period by the total population and rescales the color markers to reflect the change. It reflects the same general pattern, but puts much more emphasis on the rural South, which was the principal focus of attention of the 1965 Act.

  • 12 557 U.S. 193, at 203.
  • 13 Shelby County, at 2616, 2628, 2631, 2629, 2630.

8Map 4 changes the observer’s point of view, looking at 57 years of voting rights history as someone from the North might have seen it. From the North, restrictions on minority voting rights look like almost entirely a southern and southwestern problem. In 2009, in Northwest Austin Municipal Utility District No. 1 v. Holder, Chief Justice Roberts pointedly questioned whether the problems that Section 5 had been meant to address in 1965 were still “concentrated in the jurisdictions singled out for preclearance.”12 In Shelby County, he quoted that passage from Northwest Austin and criticized Congress because it had failed to “narrow[] the scope of the coverage formula” since 1965, and because it had neglected to determine “how that discrimination [in covered jurisdictions] compares to discrimination in States unburdened by coverage.” The 1965 coverage formula was outmoded, he said, because “today’s statistics tell an entirely different story.” If not in 2006, then certainly after Northwest Austin, the Chief Justice remarked disapprovingly, Congress should have produced an “updated statute.” “If Congress had started from scratch in 2006, it plainly could not have enacted the present coverage formula,” the Chief Justice announced, and he contended that the relationship of the coverage formula to problems of vote dilution was purely “fortuitous.”13

Map 3: The Number of Voting Rights Events, 1957-2014, Divided by the County Populations

Map 3: The Number of Voting Rights Events, 1957-2014, Divided by the County Populations

Map 4: Northern View of Voting Rights Events, 1957-2014

Map 4: Northern View of Voting Rights Events, 1957-2014
  • 14 Shelby County, at 2630.
  • 15 I have tried to be as fair as I can to the rather vague picture that the Chief Justice painted, bas (...)

9What might the pattern of voting rights cases and other violations look like over time and space if we try to visualize the description implicitly pictured in the Chief Justice’s majority opinion in Shelby County? Because discrimination in the Deep South in 1965 was, as he describes it, “pervasive . . .flagrant . . .widespread . . . rampant,” we should presumably hypothesize that the number of events was large, and that it shrank markedly over time as, in the Chief Justice’s words, “due to the Voting Rights Act, our Nation has made great strides.”14 Even though there might have been an actual increase in events as the Act gave the government and civil rights attorneys more legal tools, and there might have been periodic increases in legal cases and objections challenging the decadal redistrictings, the trend would have traced a linear decline, and the realization that discrimination was no longer so concentrated in the South would have led to an increasing proportion of cases being filed and won outside the South. This hypothetical picture might look like Figure 1, where the area outlined in red represents covered jurisdictions and that in blue, those not covered in Section 4.15

  • 16 The data in Figure 2 includes cases and objections filed on the state level.
  • 17 The cases whose times are noted on the graph will be briefly discussed later in this article.

10How does the hypothetical graph compare to that based on reality? Figure 2 substitutes non-fiction for fiction, and it paints a very different picture from that underlying the majority opinion in Shelby County.16 Before 1965, when we know that discrimination was most rampant, there were very few cases. As Department of Justice and non-governmental civil rights lawyers got more experience with the VRA, as legal cases were decided favorably, and as Section 5 was extended and Section 2 was amended, the number of events mushroomed. When, later, the Supreme Court sharply and suddenly reined in the Act, particularly in a series of cases in the 1990s, the number of events shrank just as markedly as it had grown in the 1970s and 80s.17 But until 2009, when a large number of jurisdictions began to change from at-large elections to single-member districts under the threat of lawsuits under the California Voting Rights Act, included here for completeness, the graph is almost entirely red. Before 2009, 93.8 % of all cases and other events (including statewide cases) took place in covered jurisdictions, and there was only a small increase in the number of cases outside covered areas.

Figure 1: A Hypothetical Timeline of Voting Rights Events as Chief Justice Roberts Might Have Seen Them in Shelby County, 1957-2014

Figure 2: Actual Timeline of Voting Rights Events, 1957-2014, by Covered and Non-Covered Jurisdictions

93.8 % in covered jurisdictions, 1957-2009

  • 18 Note also that I have shifted the scale so as to include less red in the second period.

11Another way to visualize the data is to compare maps of the events by county in different time periods. Maps 5 and 6 summarize voting rights legal actions that took place in the years before Congress clarified Section 2 and effectively encouraged its use in 1982 with actions over a similar number of years from 1982 on, ending in 2005, which was the last year which members of Congress considering renewal would have been able to look at before voting. Perhaps the chief justice was wrong about the shift of voting rights violations from south to north, but right about the downtrend in the sheer number of events. Maps 5 and 6 provide strong evidence against both of the contentions. Although there is a bit of suburban sprawl into South Florida and Southern California, the skyscrapers were markedly higher and denser in Map 6 than in Map 5, and they were still highly concentrated in the neighborhoods redlined by Section 4.18 In the 24 years from 1957 through 1981, there were 807 events (including those at the state level); in the 23 years from 1982 through 2005, there were 3095 events – nearly four times as many as during the earlier period. In the first period, 98.1 % of the events originated in covered jurisdictions; in the second period, 93.1 % of the much larger total.

  • 19 See many of the documents listed in the appendix to this paper.
  • 20 Nathaniel Persily, “The Promise and Pitfalls of the New Voting Rights Act,” 117 Yale L.J. 174 (2007 (...)

12Congress in 2006 was not presented with maps or other documents that laid out the pattern of proven voting rights infractions so starkly, but it received plentiful evidence in the form of lists and discussions of cases that showed that the problems were still overwhelmingly concentrated in the South and that discrimination continued to be widespread.19 A northern senator or member of the House, or a southern member who was sympathetic to minority voting rights, who read the 15,000 pages of evidence, would no doubt form in her mind images looking something like Maps 4-6 or Figure 2. The few cases in non-covered jurisdictions that were noted during the debates20 must have seemed like trivial exceptions to the overwhelmingly dominant pattern, because in fact, they were. And the map would have shown that the number of voting rights infractions had increased, not decreased, compared to the earlier period. It is easy to understand why someone who formed an image like Maps 4-6 – based on the evidence, not on outmoded stereotypes or prejudices against the South – would see little reason to modify the coverage formula and every reason to continue the Act in full vigor.

Map 5: Early Voting Rights Activity: Voting Rights Events, 1957-81

Map 5: Early Voting Rights Activity: Voting Rights Events, 1957-81

Map 6: Voting Rights Activity Still Concentrated in the South and Southwest, 1982-2005

Map 6: Voting Rights Activity Still Concentrated in the South and Southwest, 1982-2005

13The framers of the 1965 Act had only a general idea of where the most serious21 restrictions would take place in the future, and they used a blunt tool, a combination of a history of “tests and devices” for voting and a pattern of low overall turnout, to target those areas. That formula was substantially amended in 1975 to include states and counties that had printed ballots only in English, if at least five percent of the people of the area were “language minorities,” which added Texas, Arizona, Alaska, and scattered counties elsewhere to the coverage scheme.22 In his opinion in Shelby County, Chief Justice Roberts barely mentioned this amendment, which did not include political participation at all in the formula, a surprising omission in light of the fact that over a thousand events, nearly a quarter of the total, originated in Texas alone. When he explicitly discussed the coverage formula, the Chief Justice referred only to the original 1965 formula, not the 1975 expansion. So Section 4 contained two formulas, one of which was based on turnout and the other, on a very small percentage of the population that formed a “language minority” – without any test of their ability to speak or read English. But on the evidence of Maps 1 and 3-6 and Figure 2, it appears that the blunt tools of Section 4 succeeded in accurately homing in on the counties and states where the vast majority of violations would take place during the Chief Justice’s forty (and more) years since 1965. Far from “fortuitous,” the relationship between the adjusted formula and proven discrimination over the period since 1965 was nearly perfect, the coverage about as narrowly tailored and precisely targeted as a large nationwide regulatory scheme could be.

Counterfactuals Show How the Supreme Court Has Shaped Voting Rights Reality

  • 23 Brian K. Landsberg, Free at last to vote: the Alabama origins of the 1965 Voting Rights Act (Lawren (...)
  • 24 U.S. v. Board of Elections of Monroe County, 248 F.Supp. 316 (W.D. N.Y. 1965), appeal dismissed 383 (...)

14There were three basic innovations in Section 5: First, election laws or practices in suspect areas would be frozen at a certain time. Second, any change would have to be approved by an authority outside of the South, either the District Court of the District of Columbia or the Department of Justice. Third, the jurisdiction would bear the burden of proving that it had not acted in a discriminatory manner. These prophylactic devices derived from the experience of attorneys, especially from the Department of Justice and especially during their attempts to litigate cases under the Civil Rights Acts of 1957 and 1960.23 Although the congressional hearings on the 1965 bill spotlighted that experience, Congress did not design the coverage scheme by focusing it only on the states and counties in which there had been legally proven charges of voting rights violations. If it had done so, there would have been many fewer states and counties covered, as Map 7, which details the geographic pattern of successful voting rights cases from 1957 to 1965, shows. There were only 73 successful cases in those eight years, or about an average of ten a year, despite considerable efforts by a talented corps of Justice Department attorneys. All except one, a 1965 challenge to a literacy test in Monroe County, New York,24 were in a pocket of rural counties in Alabama, Georgia, Mississippi, and Louisiana.

Map 7: The Very Small Core of the Coverage Scheme: Voting Rights Events From 1957 Through 1965

Map 7: The Very Small Core of the Coverage Scheme: Voting Rights Events From 1957 Through 1965
  • 25 See, e.g., Steven F. Lawson, Black Ballots: voting rights in the South, 1944-1969 (New York: Columb (...)

15Why were there so few successful cases when the Civil Rights Movement was so active in trying to register voters in the South and when the Department of Justice was for the first time since the First Reconstruction actively attempting to foster black voting rights in the South?25 The obvious answer is that the legal tools and resources that were available were few and unwieldy, federal courts were often hostile, and there simply had not been enough judicial opinions, especially at the Circuit and Supreme Court levels, to settle many questions and smooth the path to victories by the civil rights forces. As has often been noted, the insufficiency of the tools was precisely why the VRA was needed. But that insight, emphasized by the contrast between the generally-recognized level of discrimination in 1965 and the small number and geographical compactness of the violations exposed by Map 7, has not been systematically applied to the history of the VRA after its passage, perhaps because no one before now has charted the whole sweep of successful challenges to anti-minority discrimination in its whole history.

16Applied and expanded, the insight counsels that the number and character of voting rights violations recorded is a function not only of the extent or depth of discrimination, but crucially, it is also a function of the tools made available or restrained, particularly by judicial decisions. Thus, by rendering decisions that make it easier or harder to bring and win voting rights cases or make objections, the Supreme Court can, in effect, manipulate the evidence of discrimination, which it can then use, in a second stage, to justify a decision to further weaken or strengthen the tools. It can create the reality that it subsequently reacts to. The Rehnquist and Roberts Courts have done exactly that.

17The database contains a great many illustrations of this self-fulfilling prophecy or self-generating-evidence hypothesis. Figure 3 shows not only the effect of legal changes and court decisions on the success of cases brought under Section 2 of the VRA, but it also demonstrates, once again, the concentration of discrimination in jurisdictions subject to preclearance under Section 4 before Shelby County ripped the covers off. Section 2 applies nationally, so that if discrimination were equal everywhere, one would expect to see an equal amount of blue and red in the graph. Instead, five out of every six successful Section 2 cases were filed in covered jurisdictions. Only in the foothills of the very most recent period, not in the mountains of the glory years of the VRA during the Reagan and Bush I Adminstrations, does the amount of blue equal or sometimes surpass the amount of red in Figure 3. That is surely less important than accounting for the upthrust of the mountains from 1983 through 1995 and the sudden declension thereafter.

Figure 3: Successful Section 2 Cases in Covered and Non-Covered Jurisdictions, 1957-2014

82.7 % in covered jurisdictions

  • 26 446 U.S. 55 (1980).
  • 27 478 U.S. 30 (1986).
  • 28 509 U.S. 630 (1993).
  • 29 515 U.S. 900 (1995).

18Before 1982, Section 2 was barely used at all, and it was in limbo after Justice Stewart’s decision, for a four-person plurality, in the 1980 case of City of Mobile v. Bolden,26 that to be constitutional under the Fifteenth Amendment, Section 2 had to be interpreted as requiring proof of intentional discrimination. But when Congress amended the statute in 1982 to provide that it was only necessary to prove a discriminatory effect, and when it carefully specified a series of factors that would satisfy the burden of proof, attorneys hastened to file Section 2 cases, judges felt obliged to award victories to plaintiffs who ticked off the items on the Senate’s list, and jurisdictions became eager to settle out of court. The number of successful cases exploded from 3 in 1981 to 121 by 1984. After a majority of the Supreme Court sustained the 1982 congressional amendments and seemingly simplified the evidence necessary for minority plaintiffs to win in the 1986 case of Thornburg v. Gingles,27 the number of minority Section 2 victories rose to 175 in 1988. By contrast, after a conservative 5-4 majority ruled that redistricting which increased the number of minority opportunity districts was suspect in the 1993 Shaw v. Reno case28 and followed that by ruling in Miller v. Johnson29 that a district drawn with a “predominant” (minority) racial intent was unconstitutional whatever its shape, the number of Section 2 cases plummeted from 80 in 1995 to just 6 in 2001. Attorneys rightly feared that Shaw and Miller reflected a general change in the Supreme Court’s attitude toward voting rights, one that would affect all issues, not just redistricting.

19The amount of racial discrimination in election laws did not suddenly skyrocket in 1982, and it did not suddenly abate in 1995. Just as the earlier maps made clear that discrimination continued to be concentrated in covered jurisdictions, Figure 3 demonstrates that the ebb and flow of cases proving discrimination has been largely due to responses to decisions by Congress and the courts.

  • 30 393 U.S. 544 (1969).
  • 31 425 U.S. 130 (1976).
  • 32 For more information, see my discussion in “The Strange, Ironic Career of Section 5,” at 711-12.
  • 33 See Peyton McCrary, Christopher Seaman, and Richard Valelly, “The End of Preclearance As We Knew It (...)

20Figure 4 demonstrates that a similar story can be told about Section 5 as about Section 2. Section 5 matured before Section 2, but only because the Supreme Court decision in Allen v. Board of Elections30 in 1969 allowed Section 5 lawsuits and objections to election laws that diluted minority votes, and not just to those that affected the individual right to vote per se. The number of successful Section 5 cases and objections rose from 4 in 1968 to 76 in 1976. Then in 1976, the Supreme Court clamped down, ruling in Beer v. U.S.31 that not all election law changes that had a discriminatory effect were illegal, only those that actually made minorities worse off than they had previously been. Laws that could be shown to have had a discriminatory intent, usually a more difficult matter to prove, were still deemed violative of Section 5. The number of Section 5 events dropped by half, to 38, by 1981. But Congress in a footnote to the authoritative Senate Report 97-417 on the 1982 amendments to the VRA suggested that the Department of Justice should effectively ignore Beer and refuse to preclear any election law that had a discriminatory effect, not just those that were retrogressive.32 That suggestion, combined with the increasing willingness of the Department of Justice to accept evidence that election law changes had been adopted with a discriminatory intent,33 allowed the number of Section 5 objections and requests to local and state jurisdictions for more information to grow after 1982. They reached a pinnacle of 89 in 1992, the year before Shaw v. Reno.

Figure 4: Section 5 Objections and Cases, 1957-2014

  • 34 Holder v. Hall, 512 U.S. 874, at 945.
  • 35 Shaw v. Hunt, 519 U.S. 804 (1996) decided all three questions the other way.

21What if Allen had been decided the other way and Congress had not acted to overturn or at least undercut it? What if the VRA had always been confined, as Justice Clarence Thomas has explicitly favored,34 only to guaranteeing minorities the right to vote, not to attack election structures that deny them an equal opportunity to elect candidates of their choice in areas where there is racially polarized voting? Or what if Beer had been decided in 1970, before the 1971 round of redistricting, or even earlier? At that point, there were very few African-Americans or Latinos who had been elected to any offices in the South. Any contention that a change in an electoral structure would make them worse off than being able to elect no representatives at all could easily have been dismissed as purely speculative, if not entirely illogical – because it would be hard to retrogress starting at zero. Or, on the other hand, what if Shaw v. Reno had decided that redistricting that took race into account in order to rectify generations of racial discrimination in politics or to comply with Sections 2 and/or 5 of the VRA was entirely consonant with the constitution?35

  • 36 For simple, convenient introductions, see Tim De Mey and Erik Weber, “Explanation and Thought Exper (...)

22Historians and philosophers have often employed such counterfactuals to explore causal statements. When used carefully, a counterfactual can be a powerful mode of argument.36 This dataset makes it possible to answer at least some of the questions about the effects of Supreme Court decisions empirically, thereby casting further light on the contentions of Chief Justice Roberts about the trends in racial discrimination in voting.

  • 37 More information requests are not included in Figure 5.

23Figure 5 charts successful Section 5 objections and lawsuits over time.37 The blue area is the same as in Figure 4, but the red portion deletes objections and lawsuits to enforce Section 5 if those objections or lawsuits concerned election structures – at-large elections, residency requirements, numbered seat requirements, majority-vote requirements, annexations, and redistricting. Most of the remaining objections concern changes in voting laws that might have been contested even if the Allen decision had never been made, such as precinct line changes, alterations in candidate qualifications, voting registration procedures, and purges of voter lists. Only about a quarter of the Section 5 objections and lawsuits, 421 rather than 1523, would have been lodged if vote dilution had been excluded from Section 5. The difference between the red and blue areas illustrates with considerable accuracy the effect of the Allen decision. Moving Beer back to 1971 or earlier would probably have had the same effect as eliminating Allen, because it would have placed the comparison date for a retrogressive effect at a point when there were almost no minority officeholders, making a Section 5 objection to an electoral structure nearly impossible.

Figure 5: Charting the Effect of Allen: Section 5 Objections If No Vote Dilution Mechanisms Had Been Considered Illegal, Compared to What Actually Happened

72.4 % fewer events without Allen (421, instead of 1523)

  • 38 See, e.g., William L. Taylor, “John Roberts: The Nominee,” New York Review of Books, vol. 52, # 5 ( (...)

24We can specify an analogous counterfactual in order to weigh the effect of the Supreme Court’s 1980 decision in Mobile v. Bolden. If Congress had not amended Section 2 in 1982 to reverse Bolden’s intent requirement, then it is logical to assume that no Section 2 cases would have been filed thereafter, because intent cases could just as easily have been brought under the 14th and/or 15th amendments. Assuming that Allen and Beer had actually been decided when they were and just as they were, we can isolate the effect of Bolden by subtracting Section 2 cases after 1982 from the total number of voting rights events. Figure 6 accomplishes this in a fashion parallel to Figure 5. By construction, the red and blue areas are entirely congruent with each other until 1982. After 1982, they diverge, the blue area representing all the actual events, the red, all the events except the 1174 Section 2 cases that took place after 1982. We assume, contrary to fact, that Congress did not strengthen Section 2 by overturning Bolden. (As is well known, the young John Roberts was one of the leaders of the Reagan Administration’s fight against the Bolden amendment.)38

Figure 6: Voting Rights Events if Bolden Had Not Been Overturned

28.4 % fewer total events if Bolden had not been reversed (2999, instead of 4173)

25The effect of the decision, cutting off more than a quarter of all the successful voting rights events, is forcefully portrayed in Figure 6. The mountains are leveled, the valleys deepen, and the whole graph looks much more subdued from 1982 until 1999, after which the red and blue areas converge. It is not unrealistic to postulate that there would have been substantially fewer measurable voting rights events but for the 1982 congressional “Bolden amendment.” Six hundred and eleven of the total 1174 Section 2 events that took place after 1982 were consent agreements or settlements. It is difficult to imagine nearly so many lawsuits being favorably concluded without going to trial if the plaintiffs had had to bear the burden of proof in intent cases.

California Shows That There Is Still a Need for Voting Rights Acts

26If the precipitous decline in the number of voting rights events after 1995 is not enough to convince a reader that the relative dearth of very recent cases and objections is to a considerable degree the result of Supreme Court rulings that make it more difficult for minority plaintiffs to win voting rights cases, perhaps a glance at California will. California is now a majority-minority state, and the Latino population has been increasing since 1970 at a much more rapid rate than the proportion of Latino officeholders has. Hundreds of city councils, school boards, and community college boards throughout the state are elected at-large, but Latinos have not succeeded in being elected to those bodies in nearly the proportions that their co-ethnics comprise in the population or even in the citizen voting-age population.39 Voting rights lawyers, led by former Mexican-American Legal Defense and Education Fund attorney Joaquin Avila, sought for years to attack this problem by bringing lawsuits under Sections 2 and 5 of the VRA, but their success was uneven.40 When an opportunity presented itself, Avila and Robert Rubin of the Leadership Conference on Civil Rights drafted what became the California Voting Rights Act (CVRA), signed into law in 2002.41 Compared to Section 2 of the national VRA, the CVRA reduced the proof needed to establish a violation to evidence that voting was racially polarized in the jurisdiction and, to a lesser extent, to a showing that minorities had difficulty getting elected there. There was no intent requirement, and minorities did not have to satisfy the “Gingles factor” of proving that they could comprise a majority of the voting-age or citizen voting-age population of a prospective district. In practice, another provision of the CVRA turned out to be crucial: If minority plaintiffs win, they and their attorneys may be awarded fees and costs, including expert witness fees, but if they lose, they do not have to reimburse the jurisdiction unless their lawsuit is determined to be “frivolous.” Unlike in federal litigation, plaintiffs may be held to be “prevailing parties” (and thus collect fees and costs) if there is a “causal connection” between their lawsuit and a change in the defendants’ behavior – for example, if a defendant jurisdiction switches from at-large to district elections once a lawsuit is announced, even if only minimal legal paperwork is ever filed.42 This last provision encourages lawyers to file numerous lawsuits without the fear of having to carry them through complicated hearings, depositions, and trials, and it encourages local jurisdictions to settle well before they begin to run up large bills.

  • 43 51 Cal. Rptr. 3d 821 (2006), 145 Cal. App. 4th 660.
  • 44 Michael R. Blood, “Jackpot: Lawyers earn fees from law they wrote,” AP Report, Nov. 15, 2009, avail (...)
  • 45 See, for example, the agenda of the California State Board of Education for March, 2012, which stat (...)

27After Stanislaus County Superior Court Judge Roger M. Beauchesne in 2005 declared the CVRA unconstitutional because it benefitted only minorities, it took more than two-and-a-half years for the California Court of Appeal to overturn the lower court decision and for the California and U.S. Supreme Courts to decline to overturn the Court of Appeal. What happened next considerably amplified the effect of the ruling in Sanchez v. Modesto:43 Avila and Rubin, who had litigated the case, were awarded $ 3 million in legal fees, a number that reverberated loudly in city halls and school board offices just as the Great Recession savaged their budgets.44 Facing large costs, local jurisdictions began to settle cases and to preclude lawsuits by switching to districts before they could be sued, and attorneys who had never filed voting rights cases before began to enter the field. Because school boards and community college boards must seek permission from state governing bodies before changing their electoral structures, we have both a fairly complete, though frequently updated, record of which local bodies planned to abandon at-large elections and proof that they were doing so to avoid or settle lawsuits.45

28Figure 7 shows the initial results of the effort to integrate Latinos into the local governing structures of California by forcing districts to substitute district or sometimes, proportional representation methods of election for the previous at-large rules. There were only 32 successful VRA lawsuits under Sections 2 or 5 in California from 1965 to 2007. After the CVRA pressure ratcheted up and both sides in potential lawsuits had hired expert witnesses or consultants to perform statistical analyses of racial polarization in relevant elections, districts began to settle in droves. This was especially the case with school districts that had majority- or near-majority Latino student populations, but no Latino school board members. One hundred and sixteen city councils, school boards, or community college boards changed their electoral structures as a result of the CVRA, at the time of this writing. Figure 7 shows a massive red spike, representing CVRA cases after 2008, which will probably broaden as more data becomes available.

Figure 7: California Proves That Discrimination Persists If You Have the Tools to Find It

  • 46 Richard Fausset, “Mostly Black Cities, Mostly White City Halls,” New York Times, Sept. 28, 2014, av (...)

29California did not suddenly become more discriminatory after Sanchez v. Modesto. Indeed, it is unquestionably less discriminatory than when in 1994 a majority of voters favored Proposition 187, which sought to deny governmental services, including public schools, to undocumented persons. What actually happened was that the CVRA and the favorable decision in Sanchez gave attorneys the tools with which to expose existing discrimination and to cure it. If attorneys across the nation had similarly sharp tools, not instruments blunted or destroyed altogether by Supreme Court rulings since 1993, they might well uncover discriminatory electoral structures and restrictions on individual voting rights comparable to or, likely, at greater levels than those in California. Of 340 American cities where more than 20 % of the population is black, African-Americans enjoy less representation on the city councils than their proportion in the population in 338.46 Most of these cities are in the South, and at least some have at-large elections or elect officials in off-year elections. Especially in the light of the publicity given to the underrepresentation issue by the example of Ferguson, Missouri, many of these cities would be targets for reform if other states had laws like California’s or if federal court rulings made the lower federal courts more favorable venues for VRA lawsuits.

Alternative Explanations of Comparisons and Trends

  • 47 I owe these suggestions to Richard Pildes.

30Although the geographical and temporal patterns in the data described above seem to make an unanswerable case for the adequacy of the coverage formula and the need for a mechanism to counter continuing discrimination in political rules, there might be other explanations for the patterns. It is conceivable that anomalies in the data, a lack of comparability of covered and non-covered counties, or an exhaustion of targets of discrimination might account for the results.47 Closer examination, however, will dismiss these alternatives.

Did the Data Guarantee Finding the Coverage Scheme Adequate?

  • 48 MIRs were preliminary to objections and involved considerably less time and effort on the part of p (...)
  • 49 2009, it must be remembered, is the date when the CVRA-induced changes began to mushroom.

31As to data anomalies, if there were a large number of “more information requests” (MIRs), these relatively insignificant events48 might overwhelm the rest of the data and therefore bias the results, especially if all such requests were included, and not just those that resulted in changes in questionable Section 5 submissions. But in fact, I included only the 1051 MIRs that resulted in changes in or withdrawals of apparently discriminatory changes in state or local election laws or practices, and those successful MIRs amounted to only 25.2 % of all the 4173 events. If we exclude the 1051 MIRs from the calculations, then the percentage of events that took place in covered jurisdictions before 200949 falls from 93.8 to 91.5 – still an overwhelming concentration in the areas targeted in the pre-Shelby Section 4.

  • 50 Some legal cases involved more than one statutory or constitutional basis, but there were 2512 even (...)

32If we continue along this line and exclude all the events due to Section 5 or MIRs, which could only take place in covered jurisdictions, leaving only Section 2, constitutional, and a small number of Section 203 (language discrimination) cases, still, 83.4 % of the total number of successful events before 2009 took place in covered jurisdictions.50 And this proportion no doubt understates the concentration of discriminatory propensity, for in covered jurisdictions, many discriminatory actions were negated by Section 5 objections or lawsuits, and others were aborted because of the knowledge that such actions would not survive Section 5 review. There was not as great a need to file Section 2 or fourteenth or fifteenth amendment cases in covered jurisdictions, because other, more efficient tools were available.

Do The Patterns Hold in Demographically-Similar Areas?

  • 51 See, e.g., James W. Loewen, Sundown Towns: A Hidden Dimension of American Racism (New York: The New (...)

33But it might be contended that many non-covered jurisdictions were unlikely to produce cases of racial discrimination against minorities, because there were too few minorities to bother to discriminate against (a contention that ignores a great deal of American history)51 or at least, that there were too few minorities to have a sufficient chance to affect elections to give the minorities an impetus to agitate for changes in election laws or to sue. So suppose we compare counties in non-covered jurisdictions with appreciable minority populations to those in (pre-Shelby) covered jurisdictions with similar minority proportions.

  • 52 To maintain comparability across time, I have retained counties in the “covered” classification eve (...)

34There are 3143 counties or county-equivalents in the current United States, if we include the District of Columbia. From 1957 through 2014, 2390 of those counties experienced no voting rights events that I have so far identified. Of those 2390, 2111 were in non-covered jurisdictions and 279 in (pre-Shelby) covered jurisdictions.52 Suppose that we assume that a county in which the non-Hispanic white citizen voting-age population in 2010 was less than 80 % was likely to contain a minority concentration at least somewhere in the county that might seek to elect a candidate of its choice and might therefore be seen as a threat to non-Hispanic white control of politics. To compare demographically similar counties in covered and non-covered jurisdictions, then, let us divide the counties into those with minority citizen voting age populations of 20 % or more, and those where non-Hispanic whites comprised more than 80 % of the CVAP.

  • 53 The percentage of victories in Section 2 cases filed in 20 % nonwhite CVAP counties or states in wh (...)

35Table 1 provides a first pass at the question of whether the amount of discrimination was the same in those covered and non-covered jurisdictions that were demographically similar to each other. The answer is very striking. Nearly five times as high a percentage of covered as of non-covered counties with white CVAPs of less than 80 % experienced at least one voting-rights event over the period from 1957 through 2014 (80.9 % to 17.8 %). Furthermore, minorities won cases in covered jurisdictions in much higher proportions than in non-covered jurisdictions.53 Nearly seven times as high a proportion of counties with white CVAPs of less than 80 % witnessed a successful event in covered as in non-covered counties. The contrast between covered and non-covered counties in the percentage of events in counties that were over 80 % white CVAP is almost exactly the same as in the under-80 % white CVAP counties (38.1 % to 7.3 %).

Table 1: A Much Higher Percentage of Covered Than Non-Covered Counties Experienced Voting Rights Events, Even Controlling for Minority Population Percentages

 % Non-Hispanic White CVAP, 2010

Covered Jurisdictions

Non-Covered Jurisdictions

 % Counties with Any Events

 % Counties with Successful Events

 % Counties with Any Events

 % Counties with Successful Events

>80 %

38.1

36.0

7.3

6.2

<80 %

80.9

80.3

17.8

11.9

36But Table 1 counts only the proportions of counties with any events or any successful events at all. If non-covered counties contained a larger average number of events than covered counties did, then the contrast would be less stark. Table 2 reveals that the reverse is true. Although there was a much larger number of uncovered than of covered counties where the non-Hispanic white percentage of the CVAP in 2010 was greater than 80 %, the covered counties produced nearly five times as many events, and over six times as many successful events as the uncovered counties (first row of Table 2). The success rate in covered overwhelmingly white counties was 94 %; that in non-covered overwhelmingly white counties, only 67 %. Even more impressive is the contrast between covered and non-covered counties in which the white CVAP in 2010 was less than 80 % (second row of Table 2). Nine-tenths of the events in the total of 964 less than 80 % white CVAP counties came from the 661 covered counties, and only one tenth from the 303 non-covered counties. The contrast between successful events was even more pronounced: 92 % of the successful events that originated in less than 80 % white CVAP counties came from covered jurisdictions. The success rate in covered counties was 95 %; in non-covered counties, 76 %.

37To summarize these numbers: Even when we confine our focus to those counties with an appreciable number of minorities, there is still a very pronounced difference between covered and uncovered jurisdictions. The rate of electoral discrimination that has been proven in demographically similar covered counties was several times as high as in uncovered counties. Evidently, there was something special about the counties singled out in the two formulas in Section 4, perhaps a history that trumped or at least supplemented their sociology.

Table 2: More than 90 % the Total Number of Events Came from Covered Jurisdictions, Even in Counties with Significant Minority Percentages

 % Non-Hispanic White CVAP, 2010

Covered Jurisdictions

Non-Covered Jurisdictions

Total Number of Events

Number of Successful Events

Total Number of Events

Number of Successful Events

>80 %

300

283

66

44

<80 %

3415

3236

364

276

Was It Shaw or Had Minority Political Power Maxed Out?

  • 54 This contention reflects my attempt to state as strongly and fairly as possible a view somewhat loo (...)
  • 55 A majority of the Supreme Court in LULAC v. Perry, 548 U.S. 399 (2006), at 423-25, joined numerous (...)
  • 56 Bartlett v. Strickland, 556 U.S. 1 (2009) announced that a minority group could not claim a Section (...)

38A final possible contention is that the temporal pattern of decline in the number of voting rights events after the mid-1990s reflected not the discouraging effect of adverse Supreme Court decisions, but the exhaustion of minority political opportunity.54 According to this line of reasoning, in the 1990s round of redistricting, African-Americans and Latinos had been granted nearly all of the majority-minority state legislative and congressional seats possible, and under the non-retrogression principle of Section 5, subsequent redistrictings would likely maintain those seats. Since voting was racially polarized, minority candidates who reflected the views of their co-ethnics55 were unlikely to win “whiter” seats, and in any event, they had no Section 2 claims except where they could form majorities of the potential voters in a district and no right under Section 5 to any new district, whatever the minority percentage in an area.56 Because the classic voting discrimination restrictions, such as literacy tests and poll taxes, had declined since 1965, leaving redistricting as the principal focus of voting rights lawsuits and Section 5 objections, the number of events was bound to decline after the mid-1990s, whatever the Supreme Court held in its decisions in Shaw v. Reno, Miller v. Johnson, and their progeny.

  • 57 Likewise, in his classic Southern Politics in State and Nation (New York: Vintage Books, 1949), at (...)

39This is a much more nuanced gloss on political and social trends than the Chief Justice’s view in Shelby County, and it is even more markedly different from the popular conservative belief that racial discrimination has been almost entirely extinguished. It recognizes that racial conflict and discrimination may continue in attenuated form, but denies that Supreme Court decisions themselves exaggerated that attenuation or created a new reality. But like the legal and lay conservative views, this more sophisticated scenario concludes that the Voting Rights Act was bound to fall into disuse after the 1990s round of redistricting, and that therefore, the demise of Section 5 was of little moment, the judicial recognition of a fait accompli destined to be brought about by more fundamental political and social forces.57

  • 58 Lani Guinier, “The Triumph of Tokenism: The Voting Rights Act and the Theory of Black Electoral Suc (...)

40This view is misleading for five principal reasons: First, it treats redistricting, especially redistricting for Congress or state legislatures, as more static and more central to the operation of the VRA than it has been. Second, it fails to explain why a decline in the opportunity for minorities to gain seats through redistricting should affect every provision of the Act and every form of discrimination so uniformly. Third, it does not account for the counter-trend in California. Fourth, it neglects to place the inquiry in the larger context of the history of racial discrimination in voting in the United States, a history of a constant, continual search for new and/or different tools to maintain white supremacy. And fifth, it ignores the recrudescence after Shaw of individual-level voting restrictions similar to the so-called “first-generation barriers”58 that were the focus of most pre-1965 voting rights lawsuits.

  • 59 Some cases or objections concerned more than one provision of an electoral structure, such as an at (...)

41First, then, despite their importance and popular prominence, only 21 % of the voting rights events from 1957 through 2014 concerned redistricting. A larger proportion, 27 %, had to do with at-large elections, and three other structural provisions – majority vote requirements, numbered posts, and annexation – accounted for 15 % of the events. Other events were too diverse to characterize easily. But the overall point is that nearly 80 % of the voting rights events – nearly 75 % since 1991 – were not about redistricting.59

  • 60 Alabama, Alaska, Arizona, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Texas, a (...)
  • 61 As the Supreme Court later held in Alabama Legislative Black Caucus v. Alabama, 135 S.Ct. 1257, at (...)
  • 62 After the Supreme Court remanded the redistricting case of Alabama Legislative Black Caucus v. Alab (...)

42Nor was redistricting in the covered jurisdictions as static as the “maxed-out thesis” seems to assume. Population in the ten states that were fully or largely covered under Section 4 of the VRA60 grew by 39 % from 1990 to 2010, which was sufficiently more rapid than the rest of the nation that these states gained 12 new congressional districts to go with the 96 that they held after the reapportionment between states of the early 1990s. Every new seat (or loss of a seat, as in Mississippi in 2000) represented an opportunity to reshuffle power, even if there had been little geographic mobility within each state. Perhaps more important, although the black percentage, 20 %, was almost exactly the same in 2010 as in 1990, the Latino proportion of the population in these ten covered states nearly doubled, going from 10 % to 19 %. And the minority populations were constantly in motion, 49 % of the African-Americans and 57 % of the Latinos in the South (as well as 45 % of non-Hispanic whites) moving their residence from 1995 to 2000, for example. Even if the 1990s redistricting cycle had approached a practical limit on the overall number of majority-minority congressional seats in the covered jurisdictions, population movements and demographic shifts left plenty of room for discriminatory and non-discriminatory choices, especially at the levels of city, county, and state legislative districts.61 And every such choice harbored the possibility of a lawsuit, depending on the legal climate set by the U.S. Supreme Court.62

  • 63 For descriptions of some of the districts, see Order in Perez v. Perry, available at https://drive. (...)

43A paradigmatic example of the continuing complications of redistricting in light of population growth, demographic changes, and partisan struggles, and of the continuing relevance of the Voting Rights Act, was the congressional redistricting drawn by Texas Republicans in 2011 plan in the Dallas-Fort Worth “Metroplex.” To control seven of the eight districts in the area, in which the population was only 50.2 % non-Hispanic white, Republicans packed as many African-Americans as possible into one Dallas district and extended tentacles from suburban/exurban counties into the minority areas of the central cities. In a technological tour de force, the redistricting experts extracted just enough African-Americans and Latinos and paired them with just enough suburban whites to prevent minorities from winning or even seriously influencing any other districts.63 Only the extremely complex litigation under both sections 2 and 5, still ongoing four years later, resulted in a plan in which minorities could elect one more candidate of their choice. Racial redistricting litigation was hardly maxed out after Shaw.

44Even if statewide redistricting for Congress and state legislatures had faded as a possible VRA subject after the 1990s round of redistricting, that alone would not have caused the dramatic drop-offs in voting rights events after 1993, because statewide redistricting never comprised a large proportion of the total events. Of 2608 events involving minority victories between 1965 and 1993, when Shaw was decided, only 88 (3 %) involved statewide redistricting. To be sure, many of these statewide cases were extremely significant for the distribution of political power. But many commentators have failed to appreciate the extent to which almost all VRA politics before Shaw was local.

45The second problem with the maxed-out thesis is that even if redistricting litigation declined because minorities had won nearly all the districts they could at both the state and local levels, that in itself would not explain why litigation and objections would decline simultaneously under other provisions of the VRA or with respect to other discriminatory mechanisms. Indeed, one might have expected that voting rights attorneys freed from the mathematical quagmires of redistricting plans would happily switch to attacking at-large elections, annexations, and other discriminatory mechanisms.

  • 64 The “other structural” mechanisms in Figure 9 were annexations and majority-vote and numbered-post (...)

46But in fact, as Figures 8 and 9 show, there was an across-the-board decline in Section 2 lawsuits and Section 5 objections, as well as in lawsuits attacking every facet of the electoral structure after the mid-1990s. Before Shaw and Miller, the temporal patterns of Section 2 and Section 5 actions were largely independent of each other, reacting, as Figures 3-6 above showed, to separate Supreme Court decisions on each section of the Act. After 1995, as Figure 8 makes clear, actions based on both Section 2 and Section 5 declined in tandem. Even if fewer new discriminatory laws had been passed in jurisdictions covered by Section 5 after 1993, that should have left voting rights attorneys and the Clinton Justice Department more time and energy to file Section 2 cases against existing discriminatory laws or new ones passed in non-covered jurisdictions. And Figure 9 shows that there was a similar pattern of minimal correlation between the number of lawsuits on redistricting and at-large and other electoral rules before 1995 and a parallel decline after 1995, with two exceptions – redistricting cases rose again after the decadal redistricting of 2001, and at-large cases jumped after the California Voting Rights Act became operative in 2007.64 The patterns in Figures 3-6 and 8-9 show that successful attacks on discrimination followed favorable Supreme Court decisions and that unfavorable Supreme Court decisions in the 1990s preceded the only across-the-board decline in actions in the history of the VRA. These patterns strongly suggest that it was these adverse Supreme Court decisions, not a maxing out of political opportunities for minorities, that produced the decline in events that the Chief Justice then used in Shelby to rationalize a further constraining of minority political opportunities.

Figure 8: Shaw and Miller Constrain and Coordinate Sections 2 and 5

Figure 9: Shaw and Miller Depress All Lawsuits, Not Only in Redistricting

47Figure 9 also reminds us of the third problem with the maximum opportunity suggestion – that when it became legally easier to attack at-large elections in California, attacks proliferated. Obviously, opportunities to elect further minorities had not been exhausted. They had been squashed by unfavorable decisions of the U.S. Supreme Court.

  • 65 Kousser, Colorblind Injustice, at 22-38. Note that in the 19th century, laws that discriminated aga (...)
  • 66 Smith v. Allwright, 321 U.S. 649 (1944).

48The fourth and larger problem with the maxed-out thesis is that it disregards the history of racial discrimination in politics in America. When the Fifteenth Amendment made it unconstitutional to ban people from the suffrage explicitly because of their race, the game changed. First, as soon as white supremacists violently took over legislatures during and after Reconstruction, they adopted election structures, such as gerrymandered districts or at-large elections, that disadvantaged minorities as a group. Then, they established rules that facilitated administrative discrimination, such as appointing registration and polling place officials who would bend the law to discriminate. Finally, they made voting participation contingent on correlates of individual racial characteristics, such as literacy and wealth, which brought about widespread disfranchisement.65 The same types of laws and practices continued once African-Americans began to vote in large numbers in the South after the abolition of the white primary in 194466 and especially, after the passage of the VRA in 1965.

  • 67 See City of Mobile v. Bolden, 446 U.S. 55 (1980); Lassiter v. Northampton Election Board, 360 U.S. (...)

49One means of discrimination could substitute for another that had been successfully attacked, and each successive law could be rationalized as having a non-racial purpose. For example, at-large elections could be painted as attacks on political machines or literacy tests as efforts to produce a more civically-conscious electorate.67 Even if African-Americans and Latinos had reached the maximum number of majority-minority legislative seats that their numbers allowed, that would not preclude the adoption by the majority of other methods of racial disadvantage – indeed, a growth in black and brown power would invite political opponents to seek alternatives, as they had repeatedly done since 1870. The genius of Section 5 and to a lesser extent, of Section 2 of the VRA, was that they were flexible remedies, attempts to counter future, unpredictable moves in the game of electoral discrimination. To the extent that the maxed-out thesis or a “decline in racism” thesis assumes that that game has ended, it slights the nation’s history.

50The widely-noted “voter suppression” laws that proliferated in the wake of the weakening of legal barriers to discrimination in Shelby County were actually a continuation of a trend since 2000 of laws that made it harder for individuals to register and vote, and such laws constitute the fifth problem with the exhaustion of opportunity thesis. The two most significant and hotly contested types of laws were those requiring voters to identify themselves at the polls with an often very restricted number of official documents bearing their photographs,68 and the disfranchisement of convicted criminals, often continuing long after their confinement and probationary periods had ended.69 Criminal disfranchisement was an old device clearly adopted with a racially discriminatory intent in the postbellum South,70 which gained power with the explosion of incarceration, especially black incarceration, beginning in the 1980s.71

51In addition, other new laws that were likely to have disproportionate impacts on the ability of African-Americans and/or Latinos to vote included statutes requiring documentary proof of citizenship before one could register to vote; repeals of same-day registration and curtailments of pre-election-day voting, both of which had been used disproportionately by racial minorities; restrictions on private voter registration campaigns, which had registered blacks and Latinos at twice the rate of non-Hispanic whites; and refusals to allow votes cast in the wrong precinct (usually because voters or polling places had moved) to count for higher offices, which had a larger impact on the more geographically mobile minority populations than on the somewhat less mobile non-Hispanic whites. From 2011 through 2014, 13 states passed more restrictive voter identification laws; 3 states made it harder to restore voting rights to those with criminal convictions; 9 states passed laws requiring proof of citizenship to register to vote or restricted registration drives; 8 states cut back on early voting or same-day registration; and North Carolina curtailed out-of-precinct voting on election day. Nine of the 15 states covered in whole or in part under Section 5 before Shelby County instituted new individual-level restrictions on voting since 2010.72

  • 73 Kara Brandeisky and Mike Tigas, “Everything That’s Happened Since Supreme Court Ruled on Voting Rig (...)

52Many of these laws were challenged in court under Section 2 of the VRA, but the lawsuits have been lengthy and very expensive for proponents of minority voting rights, and at the time of this writing, none has been resolved by the U.S. Supreme Court. When the cases are decided, the way may be opened, as it was with Shelby, for the passage or strengthening of many other laws making it more difficult for minority citizens to vote.73 In any event, the proliferation of such laws is one more proof that the establishment of so many new legislative and congressional majority-minority districts in the early 1990s did not toll the end of electoral discrimination or account for the decline in the number of voting rights events. The Supreme Court majority, in its decisions from Shaw through Shelby County, encouraged further legal assaults on minority voters, and those voters continued to resist, even as their legal tools were increasingly blunted.

Conclusion: It Was Still a Congruent and Proportional Remedy

  • 74 521 U.S. 507 (1997).
  • 75 See, e.g., Richard L. Hasen, “Congressional Power to Renew the Preclearance Provisions of the Votin (...)
  • 76 Kristen Clarke, “The Congressional Record Underlying the 2006 Voting Rights Act: How Much Discrimin (...)
  • 77 Richard Hasen, “The Curious Disappearance of Boerne and the Future Jurisprudence of Voting Rights a (...)
  • 78 Shelby County, at 2638.
  • 79 Shelby County, at 2629.

53During preparations for congressional consideration of amendments to the VRA in 2006, commentators frequently warned that the Supreme Court might employ its ruling in City of Boerne v. Flores74 to question whether Congress had assembled enough evidence of discrimination to satisfy a majority of the justices that Section 5 was still “congruent and proportional” to the problem.75 As a consequence, “Boerne” became the abbreviation for the reason for compiling such a huge record for that renewal, a much larger compilation of evidence than for the original passage or the renewals of 1970, 1975, or 1982.76 It therefore came as a surprise to proponents of minority voting rights that in Shelby County, Chief Justice Roberts not only did not discuss the evidence that Congress had so carefully prepared, but also that he did not cite or mention Boerne in his opinion.77 Indeed, the chief justice did not consider the level of scrutiny at all and seemed to propose a “logical” basis test, rather than the traditional “rational” basis test that the Supreme Court had employed in Katzenbach and that Justice Ginsburg explicitly relied upon in her dissent in Shelby County.78 Evidence from vote dilution cases, the Chief Justice asserted, did not justify continuing to use the initial Section 4 coverage formula, because it was “based on voting tests and access to the ballot, not vote dilution.”79 In other words, a coverage formula must not merely isolate a problem; the problem that it spotlights must constitute the formula. If that is really what the Chief Justice meant, then evidence largely from cases concerning electoral structures would be irrelevant to evaluating the strength of his opinion.

  • 80 Allen, 393 U.S. 544, at 583-84.
  • 81 Justin Levitt, “Section 5 as Simulacrum,” 123 Yale L.J. Online 151, at 161 (2013)
  • 82 Georgia v. U.S., 411 U.S. 526, at 531-35 (1973), City of Rome v. U.S., 446 U.S. 156, at 172-87 (198 (...)
  • 83 400 U.S. 379 (1971).
  • 84 539 U.S. 461 (2003).

54But such a reading cannot be justified, because it would implicitly undercut 44 years of judicial and congressional decisions. The Chief Justice did not purport to overrule so many cases or to discount the three rounds of congressional renewal of Section 5 before 2006. When a 7-2 majority of the Supreme Court ruled in 1969 in Allen v. State Board of Elections that Section 5 pertained to changes in electoral structures, as well as to measures touching the individual right to vote, it rejected dissenting Justice John Marshall Harlan’s argument that its interpretation logically and practically disconnected Section 5 from Section 4’s formula.80 Although Chief Justice Roberts did not explicitly mention Justice Harlan’s position, his argument closely paralleled Harlan’s. From 1969 until 2013, both Congress and the Court had repeatedly explicitly accepted the majority’s position in Allen, only Justices Thomas and Scalia rejecting it in Hall v. Holder. In its renewals of Section 5 in 1970, 1975, and 1982, Congress specifically approved Allen81 and relied on evidence from vote dilution cases as a justification for renewal. The Supreme Court also specifically reaffirmed the majority opinion in Allen and accepted that Section 5 covered vote dilution in its three cases previous to Northwest Austin that explicitly considered the validity of Section 5.82 In fact, every major Supreme Court case that has considered Section 5 contentions from Perkins v. Matthews83 and Beer to Georgia v. Ashcroft84 and Northwest Austin has explicitly or implicitly considered that dilutive practices are subject to Section 5. If he had meant to reject all of these precedents on the meaning and constitutionality of applying Section 5 to electoral structures, the Chief Justice would surely have made the point more clearly. We may therefore conclude that the record of VRA cases and objections since 1969 is relevant to assessing the empirical support for Chief Justice Roberts’s opinion in Shelby County.

  • 85 Coyle v. Smith, 221 U.S. 559 (1911).
  • 86 Shelby County, at 2616. On the lack of precedent for “equal sovereignty,” see James Blacksher and L (...)
  • 87 Justice Ginsburg harshly and to my mind, convincingly, criticized Chief Justice Roberts’s novel use (...)

55And that empirical support was critical, because even the theoretical underpinnings of the Chief Justice’s opinion rested on empirical grounds. Despite the fact that Katzenbach had rejected an attack on Section 5 that was based on a case that concerned the admission of new states,85 the Chief Justice in Northwest Austin had asserted that “the fundamental principle of equal sovereignty” of the states called into question any coverage scheme that treated states differently. In Shelby County, he insisted even more strongly on “the tradition of equal sovereignty” of the states, which he did not connect explicitly to any constitutional provision.86 But a departure from that principle or tradition could be justified, as Chief Justice Roberts said it had been at the time of the first passage of the VRA in 1965, by evidence that the problems to be remedied were focused principally on a few states and localities.87 Thus, even the theoretical justification of the Chief Justice’s opinion can be seen to have rested on two empirical generalizations: First, that voting discrimination had declined so much by 2006 that Congress no longer had sufficient justification to address it. Second, that whatever discrimination remained was no longer concentrated in the states and counties singled out by Section 4.

  • 88 Especially convincing were the voluminous discussions of legal cases since 1982 published by the Am (...)

56This paper marshals more evidence, more systematically than any previous research, to assess those two propositions, and it rejects both. An objective observer in 2006 comparing the number and location of all successful voting rights events in the period since the last renewal in 1982 with the events of the years from 1957 to 1981 would conclude that Section 5 needed to be renewed, and that the coverage scheme still fit the problem remarkably well, hitting the target about 94 % of the time. Even among Section 2 cases, which could be filed anywhere in the country, 83.2 % of the successful cases from 1982 through 2005 originated in covered jurisdictions. Although these exact statistics were not available, the interest groups on whose research the statistics are largely based had made an overwhelming showing on the two points.88 To the extent that there had been a decline in the number of cases and objections since the early 1990s, those who followed voting rights activity would have recognized the significance of adverse Supreme Court opinions, which were heavily criticized by civil rights forces. And experienced lawyers, such as most members of Congress, would have recognized that more cases would be filed if precedents were favorable and fewer, if they were unfavorable, and that trends in successful cases over time therefore reflected more than merely the level of discrimination.

57Whatever the degree or version of scrutiny that should have been applied in Shelby County, then – rational basis, congruence and proportionality, or compelling state interest and narrow tailoring – the evidence before Congress in 2006 or available now, especially including the most recent experience under the CVRA, should have supported an opinion upholding the coverage scheme in Section 4 and the continuation of the supervision of voting laws under Section 5. When confronted with the facts of the history of voting rights since Congress began to regulate it in 1957, Chief Justice Roberts’s opinion in Shelby County is found wanting. That history shows that the VRA retained what Chief Justice Warren termed its “Constitutional propriety” in 2013.

Top of page

Appendix

Appendix A: Sources of Voting Rights “Events”

Cases Brought by the United States Department of Justice

Brief of the United States in South Carolina v. Katzenbach, pp. 44-45, n. 33 (cases, 1957-65).

Voting Rights Act, House Hearings, 1965, Tables B2(a), B3(a), B4(a).

Voting Rights Act, House Hearings, 1975, Serial No. 1, Part 1, pp. 198-216 (cases, 1965-75).

Voting Rights Act, Senate Hearings, 1982, vol. I, pp. 1803-05, 1828-29, 1837-41, 1849 (1975- 82 cases).

“Attachments to the Statement of Wm. Bradford Reynolds, Assistant Attorney General, Civil Rights Division, Before the Subcommittee on the Constitution of the Senate Judiciary Committee on Amending the Voting Rights Act, March 1, 1982 (unpublished 12-page memo).

“Voting Section Cases in Which the United States’ Participation Began Since October 1, 1976“ (unpublished 27-page memo in Department of Justice, dated Jan. 17, 2008.

“Section 5 Declaratory Judgment Actions” (unpublished 12-page memo in Department of Justice, dated Oct. 18, 2005).

“All Section 4 Bailout Cases Filed under the Current Bailout Standard through October 17, 2005 (unpublished, undated 2-page memo).

“Table 1: Voting Rights Cases Brought on Behalf of American Indians and/or Interpreting the Voting Rights Act re Indian Interests” (unpublished, undated, 8-page memo).

Cases Brought by Public Interest Groups and Individuals or Settlements Brought About by Such Cases

ACLU, “Voting Rights in the South: Ten Years of Litigation Challenging Continuing Discrimination Against Minorities” (1982).

“Vote: The Case for Extending and Amending the Voting Rights Act: Voting Rights Litigation, 1982-2006: A Report of the Voting Rights Project of the American Civil Liberties Union.” available at. <http://www.aclu.org/​voting-rights/​case-extending-and-amending-voting-rights-act/​>.

ACLU “Promises to Keep: The Impact of the Voting Rights Act in 2006.” “ACLU Voting Rights Project Litigation Docket: Sept. 2006-Sept. 2007,” available at: https://www.aclu.org/​voting-rights/​aclu-voting-rights-project-annual-report-2006.

“ACLU Voting Rights Project Litigation Docket: Jan. 2007-Dec. 2007,” available at: https://www.aclu.org/​voting-rights/​aclu-voting-rights-project-annual-report-2007.

“ACLU Voting Rights Project Annual Report: January 1, 2008-December 31, 2008,” available at: https://www.aclu.org/​voting-rights/​aclu-voting-rights-project-annual-report-2008.

“ACLU Voting Rights Project Annual Report, January 1, 2009-December 31, 2009,“ available at: < https://www.aclu.org/​voting-rights/​aclu-voting-rights-project-annual-report-2009>.

“ACLU Voting Rights Project Annual Report: January 1, 2010-December 31, 2010,” available at: https://www.aclu.org/​voting-rights/​aclu-voting-rights-project-annual-report-2010.

“Voting Rights in Indian Country: A Special Report of the Voting Rights Project of the American Civil Liberties Union” (September 2009).

“Joaquin G. Avila Voting Rights Cases” (unpublished 15-page memo dated March 2009), courtesy of Joaquin Avila.

James U. Blacksher, List of “Dillard v. Crenshaw County”Cases, courtesy of James U. Blacksher, (unpublished memo dated Aug. 3, 2009), courtesy of James U. Blacksher.

James Blacksher, Edward Still, Jon M. Greenbaum, Nick Quinton, Cullen Brown and Royal Dumas, “Voting Rights in Alabama: 1982-2006,” Review of Law and Social Justice, 17 (2008), 249-81.

Chandler Davidson and Bernard Grofman, eds., Quiet Revolution in the South (Princeton: Princeton Univ. Press, 1994). Cases not explicitly referred to in book are available at: https://www.icpsr.umich.edu/​icpsrweb/​ICPSR/​studies/​6646?author[0] =Davidson %2C+Chandler&paging.startRow =1.

Anita Earls, Case Lists for North Carolina, 1982- (unpublished 38-page memo), courtesy of Anita Earls.

“Significant Cases” of Jose Garza (unpublished, undated, 2-page memo), courtesy of Jose Garza.

Stanislaus Anthony Halpin, Jr., “The Anti-Gerrymander: The Impact of Section 5 of the Voting Rights Act of 1965 Upon Louisiana Parish Redistricting,” (Ph.D. thesis, George Washington University, 1978).

Ellen Katz database on Section 2 cases, described in Ellen Katz, with Margaret Aisenbrey, Anna Baldwin, Emma Cheuse, and Anna Weisbrodt, “Documenting Discrimination in Voting: Judicial Findings Under Section 2 of the Voting Rights Act Since 1982: Final Report of the Voting Rights Initiative, University of Michigan Law School,”39 Univ. of Mich. Jl. of Law Reform 643 (2006), courtesy of Ellen Katz.

George Korbel, “Litigation in Texas Relating to A. Statewide Redistricting B. Section 2 litigation against Texas cities and school districts C. Section 5 Objections to Statewide Redistricting D. Section 5 Objections to Statutes relating to the Election Process in Texas #. Litigation filed after the 2011 redistricting on submission to this Court” (unpublished 13-page memo)), courtesy of George Korbel.

Brian K. Landsberg, Free at last to vote: the Alabama origins of the 1965 Voting Rights Act. Lawrence, Kansas: University Press of Kansas, 2007).

Leadership Conference on Civil and Human Rights, “Voting Rights in the States,” (2006), available at

<http://www.civilrights.org/​voting-rights/​vra/​states.html>

Justin Levitt, Database on Preclearance and More Information Requests, 2011, courtesy of Justin Levitt.

Justin Levitt, “Questions for the Record from Senator Charles E. Grassley ‘From Selma to Shelby County: Working Together to Restore the Protections of the Voting Rights Act’, Senate Judiciary Committee, July 17, 2013” (unpublished 30-page memo), courtesy of Justin Levitt.

National Commission on the Voting Rights Act, “List of Successful Section 2 and Section 5 Enforcement Actions in Selected Jurisdictions,” (unpublished, undated 39-page memo) courtesy of Jon Greenbaum.

National Commission on the Voting Rights Act, “Voting Landscape: A State-by-State Overview” (2014), available at: http://votingrightstoday.org/​ncvr/​resources/​state-pages.

Peyton McCrary, “The Struggle for Minority Representation in Florida, 1960-1990,” Florida Historical Quarterly, 86 (2007), 93-111, Appendix, “Cases That Resulted in Elimination of At- Large Elections (Brought by David M. Lipman).”

Peyton McCrary, “Declaration of Peyton McCrary in Shelby County v. Holder” (unpublished 29-page declaration, 2010), courtesy of Peyton McCrary.

“Mexican American Legal Defense and Educational Fund Litigation Timeline, 1968-2002” (unpublished 9-page memo), courtesy of Nancy Ramirez.

“MALDEF Master List 2008, Political Access Program” (unpublished 4-page memo), courtesy of Nancy Ramirez.

“Cases Litigated by Rolando L. Rios & Associates, PLLC” (unpublished, undated 17–page memo), courtesy of Rolando L. Rios.

Section Five Administrative Actions by the United States Department of Justice

U.S. Department of Justice, Section 5 Objection List, 1965-2009, available at

<http://www.justice.gov/​crt/​voting/​sec_5/​obj_activ.php>

Computer file on “more information requests,” which served as the basis for Luis Ricardo Fraga and Maria Lizet Ocampo, “More Information Requests and the Deterrent Effect of Section 5 of the Voting Rights Act,” in Voting Rights Act Reauthorization of 2006: Perspectives on Democracy, Participation, and Power (Berkeley, CA: Berkeley Public Policy Press, 2007), 47-82, courtesy of Luis Fraga. The more information requests that are included here are those that Fraga and Ocampo considered “MIR-Induced Outcomes,” i.e., those that were withdrawn, superseded, or to which the jurisdiction did not respond. See Fraga and Ocampo, Table 3.2, pp. 59-62.

U.S. Dep’t of Justice, Notices of Section 5 Activity Under Voting Rights Act of 1965, As Amended, http://www.justice.gov/​crt/​about/​vot/​notices/​noticepg.php ; U.S. Dep’t of Justice, Archive of Notices of preclearance activity under the Voting Rights Act of 1965, as amended, http://www.justice.gov/​crt/​about/​vot/​notices/​votarch.php. Only those more information requests that resulted in withdrawn submissions, 2006-13, are included. Courtesy of Justin Levitt.

California Voting Rights Act Lawsuits and Changes Brought About by Threats of Lawsuits or Preemptive Action to Avoid Lawsuits

California Board of Education Website

http://www.cde.ca.gov/​search/​searchresults.asp?cx=001779225245372747843:gpfwm5rhxiw& output =xml_no_dtd&filter =1&num =20&start =0&q =at-large %20elections. (Search for waivers under CVRA).

California State Community College Board Website http://californiacommunitycolleges.cccco.edu/​ChancellorsOffice/​BoardofGovernors.aspx. (Search for waivers under CVRA).

“Drawing The Lines: Like it or loathe it, California’s Voting Rights Act is a force to be reckoned with,” California Schools (Spring 2012), 40-47. Available at http://www.csba.org/​Newsroom/​CASchoolsMagazine/​2012/​~/​media/​Images/​NewsMedia/​PublicRelations/​CASchoolsMagazine/​2012/​CaliforniaSchools_spring2012.ashx.

Leadership Conference on Civil Rights, “LCCR Voting Rights Database Master List, CVRA Challenges, Settlements & Voluntary Changes (2006-2013), courtesy of Joanna Cuevas Ingram.

National Center for Education Statistics, Dataset ACS 2007-2011 Profile, California, All Districts (for demographic data on California school districts).

Nielsen Merksamer Website

https://nmgovlaw.com/​articles/​02_-_NMPGL_Voting_Rights_Act_Resume_2013.pdf.

Appendix B: How the Dataset Was Compiled

Since of the case lists, only Prof. Ellen Katz’s was in the form of a database file, my first undergraduate research assistant, Adam Adler, and I created a new Excel form which eventually grew to contain a much larger number and range of categories than Prof. Katz had coded. It began with Katz’s dataset and was expanded to include the lists from the Department of Justice and from the American Civil Liberties Union and Leadership Conference of Civil Rights lists and reports. Gradually, cases drawn from the other lists in Appendix A were added, as well as the Section 5 objections from the Department of Justice website and the list of “More Information Requests” from a dataset kindly provided by Prof. Luis Fraga and Maria L. Ocampo. Only the 1051 More Information Requests that resulted in a pro-minority change in an election law – for example, the withdrawal of a change that the Justice Department thought would hurt minorities or a shift from a less-pro-minority to a more-pro-minority districting plan – were included. This makes the MIR data parallel with the objections under Section 5.

Each case or other event was traced to a specific county, and all counties in the U.S. were entered into the dataset, whether they contained events or not, and given a standard GIS indicator so that they could be coordinated with maps. Every case was counted only once, whether it was appealed or not, and whether the outcome counted as favorable for minorities or not was determined by the decision at the highest level in which a judicial decision was rendered. Information on the ethnic composition of each county’s population, voting age population, and, when available, citizen voting-age population, both for 2000, 2010 and for the date at which a case was concluded or an objection or MIR was made, were matched with each event.

Since there were so many diverse lists, many overlapping, a considerable effort was made to avoid duplication by identifying each event with a specific name, civil action number, and/or case citation. This was often very difficult, since some lists gave only case names, others only case numbers, and others, only published case citations, and since many case names were the same (for example, U.S. v. Texas). Sometimes, it was a matter of judgment whether two cases with the same name should be represented as one case or two. Where citations to published cases were available, an effort was made to read them so that any more information that was available in them could be extracted. For example, attorneys in some of the cases on Prof. Katz’s list had raised constitutional issues, in addition to Section 2 issues, but she had not coded the constitutional issues, because she was focused only on Section 2.

An effort was made to categorize the issues involved in each case and objection (if there was more than one issue, the event was NOT counted more than once), as well as the legal bases for each case. Cases below the state level were categorized by the form of the local jurisdiction (city government, county government, education agency). Attempts were made to identify the attorneys involved in each case by the organization that they represented (U.S. government, ACLU, Mexican American Legal Defense and Education Fund, etc.). Where possible, the ethnicity of the challenging or defending minority plaintiff was noted. Sometimes, this was referred to in the text of cases; sometimes, it was apparent from the case location (if Mr. Jones sued a school board in Mississippi, there was a very low likelihood that he was Latino, Asian-American, or Native American), or from the plaintiff’s last name. Where there were multiple plaintiffs with names of apparently different ethnicity, all of the relevant ethnicities were coded.

Especially for the period before 1982 and after 2006, the footnotes and textual citations to cases were searched for and linked to in order to expand the number of cases at times when the available lists were inadequate. Footnotes and tables in monographs and articles on the VRA were scoured for more case names and descriptions. For CVRA cases, lists from the state board of education and state community college board were supplemented by evidence from newspaper articles and such irregular sources as the websites of consultants who specialize in advising school districts on CVRA matters.

Finally, a rather fruitless effort was made to discover other cases that resulted in consent decrees or other settlements by examining subject categories in PACER and reading the case files of cases linked to “voting” and similar topics.

Top of page

Notes

1 The Chief Justice did not merely differ with the minority in Shelby County on whether it was legally necessary to consider whether the level of voting rights discrimination in non-covered areas continued to differ from that in non-covered areas, a distinction that Richard H. Pildes has called the “legal lynchpin” of the case in “Institutional Formalism and Realism in Constitutional and Public Law,” Supreme Court Review (2013), 1, at 51. Instead, Chief Justice Roberts asserted, with the briefest glance at evidence of such discrimination, that conditions in the covered and non-covered jurisdictions currently differed too little to justify abrogating the “equal sovereignty” of the states. In other words, the Chief Justice did not simply disagree with Justice Ruth Bader Ginsburg’s minority on the proper legal standard – whether a comparison between covered and non-covered jurisdictions was required by the constitution, and he did not merely express the agnostic view that Congress might or might not have found a systematic difference between the covered and non-covered areas if it had focused on that question. He made a positive assertion that the empirical evidence, if examined, would have found little difference in the two sets of states and counties. It is this positive assertion, as well as the Chief Justice’s contention that discrimination has dwindled so much that no invasion of state sovereignty is justified any longer, that this paper addresses.

2 South Carolina v. Katzenbach, 383 U.S. 301, 308, quoted in Shelby County v. Holder, at 2628.

3 811 F.Supp. 2d 424 (D.D.C. 2011).

4 679 F.3d 848 (C.A.D.C. 2012)

5 Shelby County v. Holder, 133 S. Ct. 2612 (2013) at 2622, 2629.

6 Shelby County, at 2628.

7 Shelby County, at 2626. The Chief Justice did not note that these are self-reported registration rates or take into account that political scientists have shown that African-Americans, Latinos, and other disproportionately poorly educated and less wealthy citizens are particularly likely to overstate their registration rates. See, e.g., Robert Bernstein, Anita Chadha, and Robert Montjoy, “Overreporting Voting: Why It Happens and Why It Matters,” 65 Public Opinion Quarterly 22, at 25(2001), reporting that “both minorities and white Anglos are less likely to vote but more likely to overreport as the concentration of minorities increases.” Expressing specific skepticism about the Census Bureau’s finding that African-Americans turned out at higher rates than non-Hispanic whites in 2012 in the Deep South, figures that the Chief Justice relied upon (Shelby County, at 2619), a study for the Pew Foundation pointed out that self-reported turnout was especially exaggerated in the Deep South states. Calculations from an associated table reveal that Census turnout in 2012 was overreported by 5.1 percentage points, on average, in the six Deep South states that were covered by Section 5 in 1965, compared to 1.8 percentage points in the country as a whole. In 2004, the figures had been 3.8 percentage points in the Deep South states and 1.7 percentage points in the nation. See Paul Taylor and Mark Hugo Lopez, “Skepticism about a landmark Census finding,” available at < http://www.pewresearch.org/fact-tank/2013/05/15/skepticism-about-a-landmark-census-finding/May 15, 2013>. In North Carolina, where actual turnout is available by race, I found in research that I did for League of Women Voters v. McCrory that the Census self-reports in 2012 overestimated non-Hispanic white turnout by 1.6 percentage points, but overestimated black turnout by 11.7 percentage points.

8 The comparison between the percentage of Department of Justice objections to changes in election laws that had to be precleared under Section 5 of the Voting Rights Act from 1965 to 1975 with the percentage from 1995 to 2005 is misleading for three reasons. First, very few of the numerous changes in election laws that took place in the South were submitted for preclearance in the first few years of the law’s existence. See Kousser, “The Strange, Ironic Career of Section 5 of the Voting Rights Act, 1965-2006,” 86 Texas LR 667 (2008), at 684. The Justice Department did not even draft a guideline about Section 5 until 1970. During the first four years of the existence of the Act, there were only 251 submissions to the Department of Justice and only one objection. See Allen v. State Board of Elections 393 U.S. 544, 550, n. 5. By 1995, approximately 20,000 submissions were made every year. So the denominators in the two comparison decades were very different. Second, as the Chief Justice noted, Shelby County, at 2626, Section 5 itself inhibited discriminatory changes, making the comparison further evidence of the effectiveness, not the superfluity, of the legal provision. Third, experience both at the national and state and local levels had made the Section 5 process one of bureaucratic regularity by 1995. Extensive litigation and clerical experience had clarified exactly what each bureaucratic player had to do to comply with the law or to discover infractions. The chief justice’s discussion treats Section 5 as high constitutional politics when on a more practical level, it had become largely a rather simple bureaucratic routine.

9 Shelby County, at 2617.

10 Alaska is so large that including it in anything like its proportional size reduces the detail of the lower-48 states far too much for clarity.

11 Of 3669 county-level events that took place before 2009, when large numbers of local jurisdictions in California began to switch from at-large elections under the threat or actuality of suits under the California Voting Rights Act, 3454 or 94.1 % came from covered jurisdictions.

12 557 U.S. 193, at 203.

13 Shelby County, at 2616, 2628, 2631, 2629, 2630.

14 Shelby County, at 2630.

15 I have tried to be as fair as I can to the rather vague picture that the Chief Justice painted, basing the series on some of the same numbers in the actual series, inflating the middle period to allow for the sharper legal tools that the John Roberts opposed as a young member of the Reagan Administration and which he never mentioned in his opinion in Shelby County, and allowing for fluctuations marking redistricting periods. A simpler picture would just draw a straight line from the first date to the last, with the proportion of cases in the non-covered jurisdictions increasing proportionately after about 1970 to almost half.

16 The data in Figure 2 includes cases and objections filed on the state level.

17 The cases whose times are noted on the graph will be briefly discussed later in this article.

18 Note also that I have shifted the scale so as to include less red in the second period.

19 See many of the documents listed in the appendix to this paper.

20 Nathaniel Persily, “The Promise and Pitfalls of the New Voting Rights Act,” 117 Yale L.J. 174 (2007), at 208-11; Samuel Issacharoff, “Is Section 5 of the Voting Rights Act a Victim of Its Own Success ?”, 104 Columbia LR 1710 (2004), at 1728-29.

21 Shelby County, at 2627.

22 See http://www.justice.gov/crt/about/vot/sec_5/about.php for a description of the provisions of Section 4.

23 Brian K. Landsberg, Free at last to vote: the Alabama origins of the 1965 Voting Rights Act (Lawrence, Kan.: University Press of Kansas, 2007).

24 U.S. v. Board of Elections of Monroe County, 248 F.Supp. 316 (W.D. N.Y. 1965), appeal dismissed 383 U.S. 575 (1966).

25 See, e.g., Steven F. Lawson, Black Ballots: voting rights in the South, 1944-1969 (New York: Columbia University Press, 1976); Landsberg, Free at last to vote.

26 446 U.S. 55 (1980).

27 478 U.S. 30 (1986).

28 509 U.S. 630 (1993).

29 515 U.S. 900 (1995).

30 393 U.S. 544 (1969).

31 425 U.S. 130 (1976).

32 For more information, see my discussion in “The Strange, Ironic Career of Section 5,” at 711-12.

33 See Peyton McCrary, Christopher Seaman, and Richard Valelly, “The End of Preclearance As We Knew It: How the Supreme Court Transformed Section 5 of the Voting Rights Act,” 11 Michigan Journal of Race and Law 275 (2006).

34 Holder v. Hall, 512 U.S. 874, at 945.

35 Shaw v. Hunt, 519 U.S. 804 (1996) decided all three questions the other way.

36 For simple, convenient introductions, see Tim De Mey and Erik Weber, “Explanation and Thought Experiments in History,” History and Theory, 42 (2003), 28-38; Richard Ned Lebow, “Counterfactual Thought Experiments: A Necessary Teaching Tool,” The History Teacher, 40 (2007), 153-76.

37 More information requests are not included in Figure 5.

38 See, e.g., William L. Taylor, “John Roberts: The Nominee,” New York Review of Books, vol. 52, # 5 (Oct. 6, 2005).

39 For introductions to the CVRA, see Joanna E. Cuevas Ingram, “The Color of Change: Voting Rights in the 21st Century and the California Voting Rights Act,” 15 Harvard Latino L.R. 183 (2012); Paige A. Epstein, “Addressing Minority Vote Dilution Through State Voting Rights Acts,” Chicago Public Law and Legal Theory Working Paper No. 474 (Feb., 2014), available through SSRN, id2422915.pdf.

40 See, e.g., Gomez v. City of Watsonville, 863 F.2d 1407 (9th Cir. 1988), Gonzalez v. Monterey County, 808 F.Supp. 727 (N.D. Cal. 1992), Aldasoro v. El Centro School District, 922 F.Supp. 339 (S.D. Cal. 1995), Ruiz v. City of Santa Maria, 160 F.3d 543 (9th Cir. 1998).

41 See < http://www.leginfo.ca.gov/cgi-bin/displaycode?section=elec&group=14001-15000&file=14025-14032%3e

42 See Maguerite Mary Leoni and Christopher E. Skinnell, “The California Voting Rights Act,” at 7-8, available at < http://www.cacities.org/getattachment/f736ba74-086a-4f5d-beb7-853d898691d8/LR-Leoni-Skinnell-THE-CALIFORNIA-VOTING-RIGHTS-ACT.aspx%3e

43 51 Cal. Rptr. 3d 821 (2006), 145 Cal. App. 4th 660.

44 Michael R. Blood, “Jackpot: Lawyers earn fees from law they wrote,” AP Report, Nov. 15, 2009, available at <http://www.newsday.com/news/nation/jackpot-lawyers-earn-fees-from-law-they-wrote-1.1590856%3e The major San Francisco law firm that sponsored the case, Heller Ehrman, shared the fee.

45 See, for example, the agenda of the California State Board of Education for March, 2012, which states that “A number of districts in California are facing existing or potential litigation under the California Voting Rights Act of 2001 over their at-large election methods. To help protect themselves from potential litigation, the eleven school districts are taking action to establish new trustee areas and adopt by-trustee-area methods of election for the governing boards. Available at http://www.cde.ca.gov/search/searchresults.asp?cx=001779225245372747843:gpfwm5rhxiw&output=xml_no_dtd&filter=1&num=20&start=0&q=california%20voting%20rights%20act.

46 Richard Fausset, “Mostly Black Cities, Mostly White City Halls,” New York Times, Sept. 28, 2014, available at http://www.nytimes.com/2014/09/29/us/mostly-black-cities-mostly-white-city-halls.html?ref=us.

47 I owe these suggestions to Richard Pildes.

48 MIRs were preliminary to objections and involved considerably less time and effort on the part of proponents of voting rights than fully-litigated cases or even pre-trial settlements. Nonetheless, they should be included as indications of discrimination and as particularly close proxies for the prophylactic effect of Section 5.

49 2009, it must be remembered, is the date when the CVRA-induced changes began to mushroom.

50 Some legal cases involved more than one statutory or constitutional basis, but there were 2512 events that involved Section 5 or MIRs and 1666 that involved Section 2 and/or the 14th or 15th amendments.

51 See, e.g., James W. Loewen, Sundown Towns: A Hidden Dimension of American Racism (New York: The New Press, 2005).

52 To maintain comparability across time, I have retained counties in the “covered” classification even if they “bailed out” of Section 5 coverage. Because some covered jurisdictions that had no events had been excused from Section 5 coverage by the time that the U.S. Supreme Court decided Shelby, the 2013 proportion of non-covered counties with no events is slightly underestimated in the text.

53 The percentage of victories in Section 2 cases filed in 20 % nonwhite CVAP counties or states in which the white CVAP was less than 80 % was 87.1 (926 of 1063) in covered jurisdictions and 60.3 (173 of 287) in non-covered jurisdictions. The winning percentage in Section 2 cases is just one more piece of evidence in a general pattern of contrast between covered and non-covered jurisdictions in this paper, and is therefore not fatally undermined by the criticisms in Adam B. Cox and Thomas J. Miles, “Judging the Voting Rights Act,” 108 Columbia Law Review 1 (2008), at 42-43; Cox and Miles, “Documenting Discrimination ?”, 108 Columbia Law Review Sidebar 31 (2008).

54 This contention reflects my attempt to state as strongly and fairly as possible a view somewhat loosely derived not only from personal communications from Richard Pildes, but also from Samuel Issacharoff, “Comments: Beyond the Discrimination Model on Voting,” 127 Harvard Law Review 95 (2013).

55 A majority of the Supreme Court in LULAC v. Perry, 548 U.S. 399 (2006), at 423-25, joined numerous political scientists in recognizing that it took more than the election of a minority candidate to insure that a minority community of voters was represented.

56 Bartlett v. Strickland, 556 U.S. 1 (2009) announced that a minority group could not claim a Section 2 violation unless it could show that it alone, without any help from members of any other ethnic group, could form 50 % or more of the potential voters in a legislative district. In Reno v. Bossier Parish School Board (Bossier II), 528 U.S. 320 (2000), Justice Scalia announced with his usual verve that minorities could not claim a Section 5 violation even if a jurisdiction openly admitted an intention to discriminate, so long as it was not guilty of an intention to increase the amount of discrimination. For instance, if minorities held two seats on a city council and a redistricting authority announced that in a new redistricting, it would draw no more than two seats that minorities could win, regardless of the growth in the minority percentage of the population, minorities would have no viable Section 5 claim. Congress rejected the Bossier II standard in 2006 and restored the rule that any intentional discrimination would constitute a violation of Section 5. See Kousser, “Strange, Ironic Career of Section 5,” at 734-37, 753-54.

57 Likewise, in his classic Southern Politics in State and Nation (New York: Vintage Books, 1949), at 533 argued that “formal disfranchisement measures did not lie at the bottom of the decimation of the southern electorate [in the late 19th and early 20th centuries]. They, rather, recorded a fait accompli brought about, or destined to be brought about, by more fundamental political processes.” It took a whole book to show that Key’s fait accompli thesis was untrue. See Kousser, The Shaping of Southern Politics: Suffrage Restriction and the Establishment of the One-Party South, 1880-1910 (New Haven, Connecticut: Yale University Press, 1974).

58 Lani Guinier, “The Triumph of Tokenism: The Voting Rights Act and the Theory of Black Electoral Success,” Michigan Law Review, 89 (1991), 1077-1154, at 1093.

59 Some cases or objections concerned more than one provision of an electoral structure, such as an at-large election system with numbered posts. Each is counted separately, and there is thus some overlap in the categories in the text. Although redistricting has loomed larger as a proportion of voting rights events in recent years, it has by no means crowded out other election laws found to have been discriminatory. Since 1981, the percentages of events due to redistricting, at-large, and other structural provisions, were 25 %, 31 %, and 9 % of the total, respectively. Since 1991, the percentages of events due to redistricting, at-large, and other structural provisions were 26 %, 28 %, and 7 %.

60 Alabama, Alaska, Arizona, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Texas, and Virginia.

61 As the Supreme Court later held in Alabama Legislative Black Caucus v. Alabama, 135 S.Ct. 1257, at 1265 (2015), “A racial gerrymandering claim, however, applies to the boundaries of individual districts. It applies district-by-district. It does not apply to a State considered as an undifferentiated ‘whole.’” That view had been at least implicit in Shaw v. Reno, which concentrated on only one congressional district in North Carolina, and almost all subsequent federal court opinions in racial redistricting cases.

62 After the Supreme Court remanded the redistricting case of Alabama Legislative Black Caucus v. Alabama to the district court for further factual considerations, the plaintiffs’ post-remand brief meticulously discussed 36 separate legislative districts for over 60 pages. See http://moritzlaw.osu.edu/electionlaw/litigation/documents/ADCPostRemandBrief06122015.pdf. Each of the 36 asserted violations could, at least in theory, have been a separate lawsuit.

63 For descriptions of some of the districts, see Order in Perez v. Perry, available at https://drive.google.com/file/d/0BxeOfQQnUr_gOEx1X0dxbllTS3VhVmxtRk9aSjJ6QQ/edit ?pli =1, at 33-39, and for a map, see http://txredistricting.org/search/plan+c235. My declaration in the case, “Redistricting in Texas, 2011: Racially Polarized Voting, Racially Biased Population Deviations, and Racially Gerrymandered Maps,” available from the author, discusses the congressional plans at 108-34 and includes detailed district and demographic maps.

64 The “other structural” mechanisms in Figure 9 were annexations and majority-vote and numbered-post requirements. Annexations of non-Hispanic white areas may dilute minority political power in certain jurisdictions. In areas where voting is racially polarized and racial minorities are also population minorities, a dominant group may be able to prevent any minority group choice from being elected if the winner has to obtain a majority of all votes.

65 Kousser, Colorblind Injustice, at 22-38. Note that in the 19th century, laws that discriminated against individual voters were what might be called “third generation” measures.

66 Smith v. Allwright, 321 U.S. 649 (1944).

67 See City of Mobile v. Bolden, 446 U.S. 55 (1980); Lassiter v. Northampton Election Board, 360 U.S. 45 (1959) for examples of the Supreme Court’s credulous acceptance of these rationalizations .

68 A convenient factual overview is at http://www.ncsl.org/research/elections-and-campaigns/voter-id-history.aspx. For a list of links to articles on all sides of the controversy, see http://www.brennancenter.org/analysis/research-and-publications-voter-id.

69 Jeff Manza and Christopher Uggen, Locked out: Felon Disenfranchisement and AmericanDemocracy (New York: Oxford University Press, 2006. It is important to note that governors, most notably former Florida governor Jeb Bush, have often profoundly affected the incidence of disfranchisement by administrative actions. Id., 87-94.

70 Hunter v. Underwood, 471 U.S. 222 (1985).

71 According to the Pew Research Center, the incarceration rate of black males grew by 230 % from 1960 to 2010 and was 6.4 times the rate of white male incarceration in 2010. The rate of incarceration of Hispanic males was 2.6 times as high as that of whites in 2010. http://www.pewresearch.org/fact-tank/2013/09/06/incarceration-gap-between-whites-and-blacks-widens/

72 Wendy R. Weiser and Erik Opsal, “The State of Voting in 2014,” available at: <http://www.brennancenter.org/analysis/state-voting-2014>, for a general description of and statistical summaries of such laws. On the North Carolina out-of-precinct law, see Gary D. Robertson, “New NC law cancels ballots cast in wrong precinct,”.

73 Kara Brandeisky and Mike Tigas, “Everything That’s Happened Since Supreme Court Ruled on Voting Rights Act, Before Shelby County v. Holder After Shelby County v. Holder,” ProPublica, Nov. 1, 2013.

74 521 U.S. 507 (1997).

75 See, e.g., Richard L. Hasen, “Congressional Power to Renew the Preclearance Provisions of the Voting Rights Act after Tennessee v. Lane,” 66 Ohio St. L.J. 177 (2005).

76 Kristen Clarke, “The Congressional Record Underlying the 2006 Voting Rights Act: How Much Discrimination Can the Constitution Tolerate ? “ 43 Harvard CR-CL LR 385 (2008).

77 Richard Hasen, “The Curious Disappearance of Boerne and the Future Jurisprudence of Voting Rights and Race,” SCOTUSBLOG (June 25, 2013), available at http://www.scotusblog.com/2013/06/the-curious-disappearance-of-boerne-and-the-future-jurisprudence-of-voting-rights-and-race, Jack Balkin, “Teaching Materials for Shelby County v. Holder,” at 1 (July 20, 2013), available at < http://balkin.blogspot.com/2013/07/teaching-materials-for-shelby-county-v.html>.

78 Shelby County, at 2638.

79 Shelby County, at 2629.

80 Allen, 393 U.S. 544, at 583-84.

81 Justin Levitt, “Section 5 as Simulacrum,” 123 Yale L.J. Online 151, at 161 (2013)

82 Georgia v. U.S., 411 U.S. 526, at 531-35 (1973), City of Rome v. U.S., 446 U.S. 156, at 172-87 (1980); Lopez v. Monterey County, 525 U.S. 266, at 281-85 (1999).

83 400 U.S. 379 (1971).

84 539 U.S. 461 (2003).

85 Coyle v. Smith, 221 U.S. 559 (1911).

86 Shelby County, at 2616. On the lack of precedent for “equal sovereignty,” see James Blacksher and Lani Guinier, “Free at Last: Rejecting Equal Sovereignty and Restoring the Constitutional Right To Vote: Shelby County v. Holder,” 8 Harvard Law & Policy Review 39 (2014), at 43, n. 21. It should be noted that using “the tradition of equal sovereignty” to assess the VRA coverage scheme turned one of the requirements under strict scrutiny, “narrow tailoring,” on its head. If the law required all areas to preclear their voting law changes, it would not only be administratively unworkable, but it would also be so unfocused on the most problematic areas as to violate narrow tailoring. But under the Chief Justice’s analysis, if it treated different states differently, it might violate “equal sovereignty.” The only way to thread the needle of constitutionality would be to use a formula that targeted precisely, and the Court would have to evaluate the evidence of how well the formula worked or defer to Congress. In Shelby County, the majority did neither.

87 Justice Ginsburg harshly and to my mind, convincingly, criticized Chief Justice Roberts’s novel use of the “principle of equal sovereignty” in her dissent in Shelby County, at 2648-49.

88 Especially convincing were the voluminous discussions of legal cases since 1982 published by the American Civil Liberties Union and the Leadership Conference on Civil Rights, listed in the appendix to this paper. It is hard to imagine reading these tomes and dismissing the problems of voting rights as no longer consequential or as equally prevalent throughout the country.

Top of page

List of illustrations

Title Map 1: Voting Rights Events by County, 1957-2014
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-1.jpg
File image/jpeg, 210k
Title Map 2: Counties Covered Under Section 4 at the Time of Shelby County v. Holder
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-2.jpg
File image/jpeg, 212k
Title Map 3: The Number of Voting Rights Events, 1957-2014, Divided by the County Populations
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-3.jpg
File image/jpeg, 498k
Title Map 4: Northern View of Voting Rights Events, 1957-2014
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-4.jpg
File image/jpeg, 235k
Title Map 5: Early Voting Rights Activity: Voting Rights Events, 1957-81
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-5.jpg
File image/jpeg, 339k
Title Map 6: Voting Rights Activity Still Concentrated in the South and Southwest, 1982-2005
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-6.jpg
File image/jpeg, 309k
Title Map 7: The Very Small Core of the Coverage Scheme: Voting Rights Events From 1957 Through 1965
URL http://journals.openedition.org/transatlantica/docannexe/image/7462/img-7.jpg
File image/jpeg, 335k
Top of page

References

Electronic reference

Morgan Kousser, “Do The Facts of Voting Rights Support Chief Justice Roberts’s Opinion in Shelby County?”Transatlantica [Online], 1 | 2015, Online since 09 January 2016, connection on 28 March 2024. URL: http://journals.openedition.org/transatlantica/7462; DOI: https://doi.org/10.4000/transatlantica.7462

Top of page

About the author

Morgan Kousser

William R. Kenan, Jr. Professor of History and Social Science, California Institute of Technology

Top of page

Copyright

CC-BY-NC-ND-4.0

The text only may be used under licence CC BY-NC-ND 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search