Response to: Adam B. Cox & Thomas J. Miles, Judging the Voting Rights Act, 108 Colum. L. Rev. 1 (2008).
In Judging the Voting Rights Act, Professors Adam B. Cox and Thomas J. Miles report that judges are more likely to find liability under section 2 of the Voting Rights Act (VRA) when they are African American, appointed by a Democratic president, or sit on an appellate panel with a judge who is African American or a Democratic appointee. Cox and Miles posit that their findings “contrast” and “cast doubt” on much of the “conventional wisdom” about the Voting Rights Act,1 by which they mean the core findings we reported in Documenting Discrimination in Voting: Judicial Findings Under Section 2 of the Voting Rights Act Since 1982,2 and a related study by one of us, Not Like the South?: Regional Variation and Political Participation Through the Lens of Section 2.3
This assertion is puzzling given that Cox and Miles’ findings do not conflict or cast doubt on ours. Our studies found that obstacles to minority political participation remain more prevalent in “covered” jurisdictions (i.e. places that, under section 5 of the VRA, must obtain approval from federal officials before changing their voting laws and procedures) than in non-covered regions.4 We found that plaintiffs bringing section 2 claims have been more likely to succeed, and courts hearing these claims more likely to document a history of official discrimination, extreme racial polarization in voting, and a lack of success by minority candidates in covered jurisdictions than in non-covered ones.5 These findings bolster the argument that the need for preclearance persists and support Congress’s 2006 decision to reauthorize section 5.
Cox and Miles do not dispute our finding that section 2 plaintiffs have been more likely to prevail in covered jurisdictions than elsewhere. Nor do they contest our finding that courts hearing section 2 claims in covered jurisdictions have been more likely to make certain subsidiary findings—such as finding extreme racial polarization in voting and a lack of minority electoral success—than were courts in non-covered jurisdictions. Cox and Miles do not focus on section 2 judgments or the findings underlying these judgments, and instead direct their attention to the varied votes individual judges cast in the course of section 2 litigation. Using this lens, they observe that the individual votes judges cast in section 2 cases in covered jurisdictions were no more likely to favor liability than the votes they cast in non-covered ones. Based on this observation, Cox and Miles conclude that coverage does not matter.6
The observation about votes is correct but the conclusion is not. To be sure, the votes Cox and Miles counted do not, standing alone, show that covered jurisdictions “still have more voting rights problems” than non-covered ones.7 But no good reason exists to suspect that they would. Counting votes by individual judges rather than examining final judgments may well illuminate a number of issues, but as a lens through which to compare covered and non-covered jurisdictions, counting votes promises at best a skewed vision of “voting rights problems” in these regions.
Votes and Judgments: Lawsuits challenging electoral practices under section 2 might produce a single liability-stage vote, three such votes if heard by a three-judge trial court, or four or more such votes, depending on the nature of the appeals that follow an initial judgment. An appellate court unanimously affirming a trial judge’s violation vote might signal a particularly egregious underlying practice, or it might represent a close call affirmed based on the standard of appellate review. An unappealed trial judge violation vote might represent a violation so patent that appeal would have been futile, and yet it yields but one vote compared with the four produced by a closer case affirmed on appeal. The mere number of votes a lawsuit generates tells us nothing definitive about the nature of the underlying practice or the region in which it operated.
Judging the Voting Rights Act finds that some judges were more likely to find liability than others. In particular, Cox and Miles find that African American and Democratic judges were more likely to favor liability than their white and Republican colleagues. These proclivities, however, do not render the observed effect of coverage a mere byproduct of race and party affiliation. That is, they do not explain the greater proportion of successful section 2 claims found in covered jurisdictions. Race and party could have such explanatory power only if the judges most likely to favor liability were disproportionately located in the covered regions. But Professors Cox and Miles do not argue that they are so concentrated, and, in fact, they are not.8
Trials and Appeals: Judging the Voting Rights Act reports that trial judges were more likely to vote for liability in covered jurisdictions than in non-covered ones, but that appellate judges were not. Cox and Miles seem to think that the difference they observed between trial level and appellate votes is noteworthy, that it offers a “contrast” with our findings, and reveals “complexity” and “nuance” our studies did not acknowledge. But no contrast or unexpected complexity is operating here.
To the extent that minority voters confront greater obstacles in covered jurisdictions, one might perhaps reflexively assume that judges at all levels would be more likely to cast votes for liability than their counterparts in non-covered jurisdictions. But appellate judges differ from trial judges in important ways. Trial judges are fact finders. Insofar as minority voters confront greater or distinct obstacles to equal political participation in covered jurisdictions, trial judges adjudicating claims in these jurisdictions might well be more likely to make factual findings to that effect and to reach judgments based on these facts than would trial judges in non-covered regions. Judging the Voting Rights Act suggests that trial judges may have done just that.
Opportunities for appellate judges to cast liability-stage votes hinge on a host of factors not tied directly to coverage or its absence. For example, fifty-six trial votes produced no published appellate opinion. To the extent that trial judgments most vulnerable to challenge are the ones in fact appealed, appealing these fifty-six votes would plausibly have yielded three additional votes affirming the underlying vote in each case. Had these appeals materialized, covered jurisdictions would account for a greater proportion of appellate votes finding liability than non-covered regions.9 That these appeals did not materialize hardly suggests that covered jurisdictions have fewer voting rights problems than non-covered ones. Instead, it highlights why the number of appellate votes cast in a case reveals little about the underlying claim and the region in which it originated.10
Affirmances and Reversals: Counting and examining votes rather than judgments nevertheless reveals two insights about the Voting Rights Act that Cox and Miles do not discuss. First, examining appellate votes to affirm or reverse suggests that minority voters confront obstacles of even greater severity and scope in covered jurisdictions than our original studies suggest. Appellate courts in covered jurisdictions were both more likely to reverse denials of liability and less likely to reverse violations than were courts in non-covered regions.11 In other words, defendants were more likely to win on appeal in non-covered regions, while plaintiff-appeals were more likely to succeed in covered regions. This suggests that trial judges in covered jurisdictions, if anything, appear to have read section 2 too restrictively, and that the violations identified in covered regions are more clear and less vulnerable to challenge than those found elsewhere.
The Number of Votes: Examining section 2 votes rather than judgments also offers a means to assess section 5’s deterrent effect in covered jurisdictions. Judges cast many more votes in non-covered jurisdictions than in covered ones. They have done so because three-judge trial panels decided more than four times as many section 2 cases in non-covered jurisdictions than in covered ones.12 Three-judge trial panels are convened to hear challenges to the constitutionality of statewide or congressional apportionment plans, and such challenges were linked more often with section 2 claims in non-covered jurisdictions. These challenges materialized more often in non-covered jurisdictions precisely because section 5 does not operate in these regions.
Only covered jurisdictions must obtain federal approval or preclearance before implementing electoral changes. This requirement has blocked implementation of numerous electoral practices in covered regions, and thus eliminated the need for plaintiffs to challenge these practices under section 2.
For example, since 1982, the Justice Department has denied preclearance to dozens of districting plans of the type that, if challenged under section 2, would have been most likely to be heard by a three-judge panel.13 Covered jurisdictions adjusted many others in order to meet the section 5 hurdle,14 which had been interpreted until 1997 to require that all proposed changes comply with section 2.15 As a result, section 5 objections and adjustments made in anticipation of such objections vastly reduced the likelihood that separate section 2 challenges would follow.16 No such screening occurred in non-covered jurisdictions; hence many more three-judge trial panels were convened.
Section 5’s screening effect produced the lopsided number of votes cast by three-judge panels in non-covered jurisdictions, and, more generally, the disproportionate number of votes cast in non-covered regions. All else being equal, proportionally fewer section 2 violations should be found in covered jurisdictions than in non-covered ones where, by definition, preclearance does not operate. But this is not the case.17 Section 2 plaintiffs were more likely to succeed and in fact succeeded more often in covered jurisdictions.
CONCLUDING REMARKS
Our studies on section 2 litigation suggest that covered jurisdictions indeed “have more voting rights problems” than do non-covered ones. Our studies, of course, cannot prove definitively that this is so. Section 2 litigation offers a lens through which to examine opportunities for minority political participation in covered and non-covered jurisdictions, but this lens—as we have repeatedly recognized18—is imperfect. It requires that cases be brought, resources devoted to their prosecution, and merits decisions be both reached and published for review. To the extent that any of these factors varied systematically between covered and non-covered jurisdictions, a distorted portrait of political participation would have emerged from published section 2 decisions. Cox and Miles, however, have identified nothing that suggests any such distortion. If anything, a careful examination of votes by individual judges reveals section 2 litigation to be a more precise lens through which to study minority political participation than even we had suspected.
APPENDIX
For accompanying tables, please see PDF version at http://www.columbialawreview.org/Sidebar/volume/108/23_KatzBaldwin.pdf.
---
*Professor of Law, University of Michigan Law School. Thanks to Emma Cheuse, Daniel Halberstam, Bill Miller, J.J. Prescott, Scott Shapiro, Mark West, and Chris Whitman for comments, and to Charles Doriean for statistical work.
**J.D., Mich. 2006.
1. See Adam B. Cox & Thomas J. Miles, Judging the Voting Rights Act, 108 Colum. L. Rev. 1, 5 (2008).
2. Ellen D. Katz et al., Documenting Discrimination in Voting: Judicial Findings Under Section 2 of the Voting Rights Act Since 1982, Final Report of the Voting Rights Initiative, 39 U. Mich. J.L. Reform 643 (2006), available at http://sitemaker.umich.edu/votingrights/files/finalreport.pdf [hereinafter Katz et al., Documenting Discrimination].
3. Ellen D. Katz, Not Like the South? Regional Variation and Political Participation Through the Lens of Section 2, in Voting Rights Act Reauthorization of 2006: Perspectives on Democracy, Participation and Power 183–221 (A. Henderson ed., 2007), available at http://www.law.berkeley.edu/centers/ewi-old/research/votingrights/vra/
ch%208%20katz%203-9-07.pdf [hereinafter Katz, Not Like the South].
4. See Katz et al., Documenting Discrimination, supra note 2, at 655–56; Katz, Not Like the South, supra note 3, at 187, 210–13.
5. See Katz et al., Documenting Discrimination, supra note 2, at 655–56; Katz, Not Like the South, supra note 3, passim.
6. See Cox & Miles, supra note 1, at 5, 47 (arguing that coverage “is not a strong predictor of liability in most section 2 cases,” and that the “estimated impact of section 5 coverage is small . . . and statistically insignificant”).
7. See id. at 5.
8. African American judges cast 6.4%, or 16, of the 251 votes cast in covered jurisdictions, and 6.1%, or 19, of the 310 votes cast in non-covered jurisdictions. Democratic judges cast 41.8%, or 105, of the 251 votes in covered jurisdictions, and 43.2%, or 134, of 310 votes in non-covered regions. These proportions are statistically equivalent. See infra Table 1.
Among trial judges, Democratic appointees cast 41.9%, or 39 of the 93 trial votes cast in covered jurisdictions, compared with 44.2%, or 69 of 156 votes in non-covered regions, and African American judges cast 8.6%, or 8 of the 93 covered trial votes, compared with 5.8%, or 9 of the 156 votes in non-covered jurisdictions. These proportions are statistically equivalent. See infra Table 1.1.
9. Thirty-five of the 56 unappealed individual trial votes found a section 2 violation, 22 of which came from covered jurisdictions, 13 from non-covered. Translating these votes into appellate affirmances adds 66 additional appellate votes finding liability in covered regions, and 39 in non-covered. Translating the 21 unappealed trial votes that found no violation into appellate affirmances adds 33 votes against liability in non-covered regions, and 30 in covered. Adding these votes to the tally would mean that 43.7% of the appellate votes cast in covered jurisdictions would have favored liability, compared with 38.5% in non-covered jurisdictions.
10. Our claim here is not, as Cox and Miles believe, that they “were wrong” to count appellate votes in their study on judicial propensities. See Adam B. Cox & Thomas J. Miles, Documenting Discrimination?, 108 Colum. L. Rev. Sidebar 31, 33 (2008), http://www.columbialawreview.org/Sidebar/volume/108/31_CoxMiles.pdf. Instead, we take issue with the use Cox and Miles makes of the count they produced. Simply put, there is no reason to assume that tallying appellate votes on liability will produce the same liability patterns observed when counting votes by trial judges, or, more importantly, that it will mirror the proportion of overall judgments that find liability under section 2. Our claim has consistently been that these judgments, and the findings that support them, matter far more that the number of times individual judges happen to cast liability votes in the course of a section 2 lawsuit.
11. See infra Table 2.
12. See infra Tables 1, 3.
13. See U.S. Dept. of Justice, Civil Rights Division, Voting Section, About Section 5 of the Voting Rights Act (Jan. 2, 2008), at http://www.usdoj.gov/crt/voting/sec_5/obj_activ.htm (listing objections to various redistricting plans by state, including 1 objection in Alabama, 1 in Arkansas, 1 in Arizona, 2 in Florida, 4 in Georgia, 2 in Mississippi, 1 in New Mexico, 2 in New York, 2 in North Carolina, 1 in South Carolina, 2 in Texas, and 2 in Virginia).
14. See Luis Ricardo Fraga & 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, supra note 3, at 47, 56; Pamela S. Karlan, Section 5 Squared: Congressional Power to Extend and Amend the Voting Rights Act, 44 Houston L. Rev. 1, 23–24 (2007).
15. Ellen D. Katz, Mission Accomplished?, 117 Yale L.J. Pocket Part 142 (2007), at http://thepocketpart.org/2007/12/10/katz.html [hereinafter Katz, Mission Accomplished?].
16. See Ellen D. Katz, Congressional Power to Extend Preclearance: A Response to Professor Karlan, 44 Hous. L. Rev. 33, 37 n.12 (2007); see also Karlan, supra note 15, at 31; Katz, Not Like the South, supra note 3, at 209–10 & nn.128, 135.
17. See Katz, Mission Accomplished?, supra note 15.
18. See Katz et al., Documenting Discrimination, supra note 2, at 734; Katz, Not Like the South, supra note 3, at 214; Katz, Mission Accomplished?, supra note 15.
------------------------
Preferred Citation: Ellen D. Katz & Anna Baldwin, Documenting Discrimination, 108 Colum. L. Rev. Sidebar 23 (2008), http://www.columbialawreview.org/Sidebar/volume/108/23_KatzBaldwin.pdf.
