skip navigation

RILEY v. KENNEDY (No. 07-77)
Reversed and remanded.

HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Stevens, J., dissenting



on appeal from the united states district court forthe middle district of alabama

[May 27, 2008]

    Justice Stevens, with whom Justice Souter joins, dissenting.

    Voting practices in Alabama today are vastly different from those that prevailed prior to the enactment of the Voting Rights Act of 1965 (VRA), 79 Stat. 437, as amended, 42 U. S. C. §1973 et seq. (2000 ed. and Supp. V). Even though many of those changes are, at least in part, the consequence of vigorous and sustained enforcement of the VRA, it may well be true that today the statute is maintaining strict federal controls that are not as necessary or appropriate as they once were. The principal events at issue in this case occurred in the 1980’s, when the State’s transition from a blatantly discriminatory regime was well underway.

    Nevertheless, since Congress recently decided to renew the VRA,1 and our task is to interpret that statute, we must give the VRA the same generous interpretation that our cases have consistently endorsed throughout its history. In my judgment, the Court’s decision today is not faithful to those cases or to Congress’ intent to give §5 of the VRA, §1973c (2000 ed.), the “broadest possible scope,” reaching “any state enactment which altered the election law of a covered State in even a minor way.” Allen v. State Bd. of Elections, 393 U. S. 544, 566–567 (1969) . I think it clear, as the Department of Justice argues and the three-judge District Court held, 445 F. Supp. 2d 1333 (MD Ala. 2006), that the Alabama Supreme Court’s decision in Stokes v. Noonan, 534 So. 2d 237 (1988), caused a change in voting practice that required preclearance.


    Section 5 preclearance is required “[w]henever a [covered] State … shall enact or seek to administer any voting qualification or prerequisite to voting, or standard, practice, or procedure with respect to voting different from that in force or effect on November 1, 1964.” 42 U. S. C. §1973c. The critical question in this case is whether the procedure for selecting Mobile County Commissioners arising out of Stokes—gubernatorial appointment—is a “change” under §5.

    As an initial matter, the language of §5 requires that the practice be “different from that in force or effect on November 1, 1964.” It is undisputed that the practice in force or effect in 1964 was gubernatorial appointment, see Ala. Code §12–6 (1959); the practice of calling a special election to fill midterm openings on the Mobile County Commission was not introduced until the passage ofAlabama Act No. 85–237 (1985 Act), 1985 Ala. Acts no. 85–237.

    The argument that a return to gubernatorial appointment will never require preclearance under §5 because gubernatorial appointment was the practice in effect in 1964 is neither persuasive nor properly before the Court. Appellant expressly abandoned any such argument in his briefs to this Court. See Reply Brief 8 (“Our contention, as we have already said, is not that the Court needs to rethink prior dicta suggesting that, despite its language, §5 operates like a ratchet to subsume newly-precleared practices … . That question is not before the Court, and we take no position on it”). Further, appellant did not raise the argument in either of his trial briefs to the District Court. Governor’s Trial Brief in Kennedy v. Riley, Civ. Action No. 2:05 CV 1100–T (MD Ala.); Governor’s Supplemental Trial Brief in Kennedy v. Riley, Civ. Action No. 2:05 CV 1100–T (MD Ala.).

    Appellant’s decision not to challenge the preclearance requirement on this ground was no doubt because of the settled law to the contrary. Reflecting the fact that Congress certainly did not intend §5 to create a “safe harbor” for voting practices identical to practices in effect in 1964, the settled understanding among lower courts and the Department of Justice is that §5 operates instead as a ratchet, freezing in place the most recent voting practice. See Brief for United States as Amicus Curiae 16–18 (collecting cases); 28 CFR §51.12 (2007). Furthermore, Congress has reauthorized the VRA in the face of this understanding without amending the relevant language of §5. See Voting Rights Act Reauthorization and Amendments Act of 2006, 120 Stat. 577; ante, at 3, n. 1 (describing the history of renewals and extensions of the VRA). Thus, the inclusion of the date 1964 in the language of §5 poses no obstacle to my conclusion that Stokes—even though it returned to gubernatorial practice—implemented a change in voting practice that required preclearance.


    Whether a voting practice represents a change that requires preclearance is measured against the previously precleared “baseline” practice in force or effect. Young v. Fordice, 520 U. S. 273, 282–283 (1997) ; Presley v. Etowah County Comm’n, 502 U. S. 491, 495 (1992) . The baseline is the practice actually in effectimmediately prior to the putative change, whether or not that practice violates state law. In Perkins v. Matthews, 400 U. S. 379 (1971) , for example, we held that the baseline practice was not at-large elections, even though at-large elections were required by a 1962 state statute. Because the city had never implemented that statute, we held that the practice actually in force or effect on November 1, 1964 was ward elections, despite that practice’s illegality under state law. Id., at 394–395.

    The situation was similar in City of Lockhart v. United States, 460 U. S. 125 (1983) . There we considered whether the practice of using numbered posts for elections was in force on the relevant coverage date and concluded that despite the possibility that this practice was illegal under Texas law, the numbered-post system could serve as the baseline. Id., at 132, and n. 6. We emphasized once again that “[s]ection 5 was intended to halt actual retrogression in minority voting strength without regard for the legality under state law of the practices already in effect.” Id., at 133.

    In Young v. Fordice, 520 U. S. 273 (1997) , our most recent case deciding whether a voting practice was a baseline under §5, we concluded that the registration procedure at issue was not “in force or effect” and therefore could not serve as the §5 baseline. In 1994, Mississippi began modifying its registration practices in an attempt to comply with the National Voter Registration Act of 1993, 107 Stat. 77, 42 U. S. C. §1973gg et seq. (2000 ed. and Supp. V).In late 1994, the Mississippi Secretary of State proposed a series of changes and assumed that the Mississippi Legislature would adopt those changes. The Secretary of State told at least one election official to begin registering voters under the new plan. The proposed changes were precleared, and about 4,000 voters were registered. The legislature failed to adopt the proposal, however, and the registrants were notified that they were not, as they had thought, registered to vote in state or local elections. Fordice, 520 U. S., at 277–278. We held that the provisional registration system was not the baseline because it was never in force or effect.

    An ordinary observer asked to describe voting practice in Alabama with respect to the method of filling vacancies on the Mobile County Commission would no doubt state that before 1985 the practice was gubernatorial appointment, between 1985 and 1988 the practice was special election, and beginning in 1988 the practice changed to gubernatorial appointment.

    In the face of this history, the Court comes to the startling conclusion that for purposes of the VRA Alabama has never ceased to practice gubernatorial appointment as its method of selecting members of the Mobile County Commission. But under our case law interpreting §5, it is clear that a change occurred in 1988 when Stokes returned Alabama to gubernatorial appointment.2 This represented a change because the relevant baseline was the special election procedure mandated by the Alabama Legislature’s enactment of the 1985 Act, which was precleared by the Department of Justice in June 1985. Pursuant to that law, the Governor called a special election when a vacancy arose in 1987. The vacancy was filled and the newly elected commissioner took office in July 1987 serving, by way of his election, until September 1988.

    It is difficult to say that the special election practice was never “in force or effect” with a straight face. Jones was elected and sat on the three-member Mobile County Commission for approximately 14 months. During those 14 months, the County Commission held dozens of meetings, at which the Commission exercised its executive and administrative functions. During the time he served as a result of the special election, Jones was central to actions having a direct and immediate impact on Mobile County. For example, at a meeting held on October 13, 1987, the Commission considered 25 agenda items, one of which was paying claims and payrolls of over $1 million. Minutes from Meeting Oct. 13, 1987.

    The differences between this case and Fordice are legion. In holding that the provisional registration system in Fordice did not constitute the baseline by which to measure future practices, we emphasized that the plan was abandoned as soon as it was clear that it would not be enacted, the plan was in use for only 41 days, and only about one-third of the election officials had even implemented the proposal. 520 U. S., at 283. Further, the State rectified the situation far in advance of any elections; there was no evidence that anyone was prevented from voting because of reliance on the rejected plan. Ibid.

Fordice was in essence a case of “no harm, no foul.” Here, of course, the special election did take place and the elected commissioner held his post for 14 months, voting on hundreds of measures shaping the governance of Mobile County. While the voters in Fordice could be reregistered under the new procedures, Jones’ election to the Commission and his 14-month service cannot be undone.

    The majority seems to acknowledge that Fordice is distinguishable, stating that if “the only relevant factors were the length of time a practice was in use and the extent to which it was implemented, this would be a close case.” Ante, at 15. The Court relies, however, on the “extraordinary circumstance” that the 1985 Act was challenged immediately and that the 1987 election was held “in the shadow” of that legal challenge. Ante, at 15–16. But a cloud of litigation cannot undermine the obvious conclusion that the special election practice was in force or effect. That practice, therefore, is the practice to which gubernatorial appointment must be compared.

    The majority makes much of the fact that to adopt the view of the three-judge District Court would make the question whether a voting practice is “in force or effect” turn on whether the circuit court happened to get the law right in time to stop the election. Ante, at 17. But the majority’s approach turns instead on whether Alabama possesses highly motivated private litigants. If Stokes had not challenged the election until it had already taken place (or had failed to appeal), the election would be in force or effect under the majority’s view. Nothing in the VRA or our cases suggests that the VRA’s application should hinge on how quickly private litigants challenge voting laws.

    Our decisions in Perkins and Lockhart give no indication that if a citizen in Canton, Mississippi or Lockhart, Texas had challenged the legality of the ward elections or the numbered-post system, the illegality of those practices under state law would have been any more relevant to their status as the relevant baselines. This case calls for nothing more than a straightforward application of our precedent; that precedent makes clear that the special election procedure was the relevant baseline and that gubernatorial appointment therefore represents a change that must be precleared.


    The VRA makes no distinction among the paths that can lead to a change in voting practice, requiring preclearance “whenever” a State seeks to enact “any” change in voting practice. 42 U. S. C. §1973c. And changes to voting practice can arise in at least four ways: (1) legislative enactment; (2) executive action; (3) judicial changes, either by a proactive judicial decision (e.g., redistricting) or, as in this case, through judicial interpretation of state law; or (4) informal abandonment or adoption by election officials.

    The majority does not dispute that a change in voting practice wrought by a state court can be subject to preclearance. See ante, at 11 (citing Branch v. Smith, 538 U. S. 254 (2003) , and Hathorn v. Lovorn, 457 U. S. 255 (1982) ). But the majority falters when it treats the change effected by Stokes differently for §5 preclearance purposes than it would treat a newly enacted statute or executive regulation. The majority finds it “anomalous” that Alabama might be bound “to an unconstitutional practice because of a state trial court’s error.” Ante, at 17. The clear theme running through the majority’s analysis is that the Alabama Supreme Court is more deserving of comity than the Alabama Legislature.

    Imagine that the 1985 Act had been held constitutional by the Alabama Supreme Court in Stokes, but that in 1988 the Alabama Legislature changed its mind and repealed the Act, enacting in its place a statute providing for gubernatorial appointment. Imagine further that the Department of Justice refused to preclear the practice (as it no doubt would); if Alabama wanted to fill an open seat on the Mobile County Commission it would have to administer its former special election practice even though that law had been repealed. It is not clear to me or to the United States, see Brief as Amicus Curiae 25–27, why effectively requiring a State to administer a law it has repealed is less offensive to state sovereignty than requiring a State to administer a law its highest court has found unconstitutional. The VRA “by its nature, intrudes on state sovereignty.” Lopez v. Monterey County, 525 U. S. 266, 284 (1999) .

    The majority attempts to portray the circuit court judge’s decision as so far outside the bounds of Alabama law, see ante, at 17, that allowing it to effectively establish the special election practice as a §5 baseline would be intolerable. I am certain, however, that the two Alabama Supreme Court Justices dissenting in Stokes would disagree. 534 So. 2d, at 239 (opinion of Steagall, J., joined by Adams, J.). The dissenting Justices argued that the 1985 Act was sufficiently “amendatory” to avoid the requirements of Peddycoart v. Birmingham, 354 So. 2d 808 (Ala. 1978), because it merely amended the 1957 Act creating the Mobile County Commission. The Circuit Court Judge followed similar reasoning, citing Alabama Supreme Court precedent stating that “[i]t is the duty of the courts to sustain the constitutionality of a legislative act unless it is clear beyond a reasonable doubt that it is in violation of the fundamental law.” Stokes v. Noonan, CV–87–001316 (Mobile County, May 19, 1987). Nothing in the circuit court judge’s decision indicates that this case calls for anything other than a straightforward application of our precedent.


    Finally, the history of the voting practices that the VRA sought to address, especially in Alabama itself, indicates that state courts must be treated on the same terms as state legislatures for §5 purposes. Specifically, the history of Alabama’s voter registration requirements makes this quite clear.3 Alabama’s literacy test originated in a constitutional convention called in 1901 “largely, if not principally, for the purpose of changing the 1875 Constitution so as to eliminate Negro voters.” United States v. Alabama, 252 F. Supp. 95, 98 (MD Ala. 1966); see also M. McMillian, Constitutional Development in Alabama, 1789–1901, pp. 217–232 (1955); Hunter v. Underwood, 471 U. S. 222 (1985) .4 Not wishing to run directly afoul of the Fifteenth Amendment , delegates at the convention devised a poll tax and a literacy test in order to disfranchise African-Americans. The effects of the new Constitution were staggering: In 1900, 100,000 African-Americans were enrolled as voters in Alabama. By 1908, only 3,742 African-Americans were registered to vote. Alabama, 252 F. Supp., at 99; V. Hamilton, Alabama: A Bicentennial History 96 (1977).5

    The Alabama Constitution provided for judicial review of contested registrar decisions, see §186 (1901), but that review provision was rendered all but useless by the Alabama Supreme Court’s adoption of both a strong presumption that the Board of Registrars’ decisions were valid and stringent pleading requirements. For example, in Hawkins v. Vines, 249 Ala. 165, 30 So. 2d 451 (1947), the Alabama Supreme Court rejected a petition from a denial of registration because the petitioner averred that he “is a citizen of the United States,” “is able to read and write,” and “is over the age of twenty-one years,” rather than expressly stating that he met those requirements at the time he attempted to register. Id., at 169, 30 S. 2d, at 455 (emphasis deleted; internal quotation marks omitted). In Hawkins the Alabama Supreme Court also reaffirmed its previous holding in Boswell v. Bethea, 242 Ala. 292, 296–297, 5 So. 2d 816, 820–821 (1942), that the decisions of the Board of Registrars are “presumptively regular and valid and the burden is on the one who would attack the order to show error.”249 Ala., at 169, 30 So. 2d,at 454.

    Alabama’s literacy test was later amended via the “Boswell Amendment” to include a requirement that voters demonstrate that they were able to “understand and explain any article of the constitution of the United States in the English language.” Ala. Const. §181 (1901) (as amended in 1946 by Amdt. 55). That amendment was held to be unconstitutional in Davis v. Schnell, 81 F. Supp. 872, 881 (SD Ala. 1949). Not easily deterred, the legislature responded with a new amendment, ratified in December 1951, which provided that the Alabama Supreme Court would promulgate a uniform questionnaire to be completed by all applicants. Ala. Const. §181 (1901) (as amended in 1951 by Amdt. 91); see United States v. Penton, 212 F. Supp. 193, 204, 205 (MD Ala. 1962) (reproducing questionnaire in App. B).

    During the period from 1951 to 1964, the Alabama Supreme Court rendered the questionnaire more and more complex. In 1960, in response to the efforts of African-American organizations to educate voters, the questions were arranged in different sequences for different questionnaires. B. Landsberg, Free at Last to Vote: The Alabama Origins of the 1965 Voting Rights Act 19 (2007). These new questionnaires had the effect of blocking the registration of thousands of African-American voters. For example, as a district court in Alabama found, between 1954 and 1960 only 14 African-Americans were registered to vote in Dallas County—a county with approximately 15,000 African-Americans. See United States v. Atkins, 323 F. 2d 733, 736 (CA5 1963). Among the African-Americans denied registration were two doctors and six college graduates. Ibid.

    The Alabama Supreme Court responded to the litigation surrounding its questionnaire by drafting a new questionnaire in 1964; that questionnaire had a literacy and civics test on which questions were rotated, resulting in 100 different forms of the test. E. Yadlosky, Library of Congress Legislative Reference Service, State Literacy Tests as Qualifications for Voting 19 (1965). The tests contained questions such as “Ambassadors may be named by the President without the approval of the United States Senate. (True or False),” and “If no person receives a majority of the electoral vote, the Vice President is chosen by the Senate. (True or False).” Ibid. (internal quotation marks omitted).6 These tests were finally put to rest throughout the country in the VRA, which mandates that “[n]o citizen shall be denied, because of his failure to comply with any test or device, the right to vote.” 42 U. S. C. §1973aa.

    In sum, prior to the VRA, the Alabama Supreme Court worked hand-in-hand with the Alabama Legislature to erect obstacles to African-American voting. While I do not wish to cast aspersions on the current members of the Alabama Supreme Court or the court that decided Stokes v. Noonan, 534 So. 2d 237, the history of the Alabama Supreme Court’s role in designing Alabama’s literacy test provides a vivid illustration of why voting changes wrought by state-court decisions must be treated on the same terms as those brought into effect by legislative or executive action.


    There is simply nothing about this case that takes it outside the ordinary reach of our VRA precedents. Because the 1985 Act was precleared and put in effect during the 1987 election, the practice of special elections serves as the relevant baseline. With the correct baseline in mind, it is obvious that the gubernatorial appointment put in place by Stokes is a practice “different from” the baseline. Because gubernatorial appointment represents a change, it must be precleared, as the three-judge District Court correctly held.

    I therefore respectfully dissent.


1 Fannie Lou Hamer, Rosa Parks, and Coretta Scott King Voting Rights Act Reauthorization and Amendments Act of 2006, 120 Stat. 577. The Act passed the Senate by a vote of 98 to 0. 152 Cong. Rec. S8012 (July 20, 2006).

2 Even the majority cannot escape this conclusion, stating that “[t]he State’s reinstatement of th[e] practice [of gubernatorial appointment] did not constitute a change requiring preclearance.” Ante, at 12 (emphasis added); see also, e.g., ante, at 7, 12. Of course, if there was no change, then there was nothing to reinstate.

3 The NAACP Legal Defense and Educational Fund’s amicus brief provides a history of the role that Alabama courts played in promoting and retaining discriminatory voting practices.

4 The spirit of the Constitution’s registration provision was captured by the statement of Delegate Heflin: “We want the white men who once voted in this State and controlled it, to vote again. We want to see that old condition restored. Upon that theory we took the stump in Alabama, having pledged ourselves to the white people of Alabama, upon the platform that we would not disfranchise a single white man, if you trust us to frame an organic law for Alabama, but it is our purpose, it is our intention, and here is our registered vow to disfranchise every negro in the State and not a single white man.” 3 Official Proceedings of the Constitutional Convention of the State of Alabama, May 21st, 1901, To September 3rd 1901, p. 2844 (1941).

5 Provisions following the lead of the 1890 “Mississippi Plan” were enacted in other State Constitutions, with similar results. See C. Zelden, The Battle for the Black Ballot 17–18 (2004) (describing similar changes to registration practice in Mississippi, South Carolina, North Carolina, Louisiana, Alabama, Virginia, Texas, and Georgia and their effects on registration); C. Woodward, Origins of the New South 1877–1913, pp. 321–349 (1951) (describing effect of Mississippi Plan on the States that adopted it). While poor white voters were also disfranchised to a significant degree, these provisions fell most heavily on African-American voters. See id., at 342–343 (demonstrating that between 1897 and 1900 in Louisiana registered white voters dropped by about 40,000 and registered African-Americans dropped by approximately 125,000).

6 Some of other questions were “Are post offices operated by the state or federal government?,” “When residents of a city elect their officials, the voting is called a municipal election (True or false),” “Of what political party is the president of the United States a member?,” and “What is the chief executive of Alabama called?” United States v. Parker, 236 F. Supp. 511, 524, 525, 528 (MD Ala. 1964) (reproducing the questionnaire).