Tracey E. GEORGE, et al., Plaintiffs–Appellees, v. Tre HARGETT, et al., Defendants–Appellants.
In November 2014, Tennessee voters approved an amendment to the Tennessee Constitution making clear that the Constitution is not to be construed as securing or protecting a right to abortion or requiring funding of an abortion. Understandably, the amendment was a matter of no small controversy. In fact, more votes were cast in favor of and opposition to the amendment than were cast in the governor's race during the same election. In this litigation, the controversy is still kept alive. That is, since federal court jurisdiction was invoked three years ago, it remains to be determined whether the Tennessee electorate did in fact amend their constitution. This litigation, however, is only marginally related to the public policy controversy, focusing on what might be viewed as much more pedestrian questions, such as whether the votes were counted incorrectly, and whether the vote-counting method impermissibly infringed some voters' rights. We answer “no” to both questions and thus give effect to the express will of the people.
Plaintiffs in this case, Tracey E. George, Ellen Wright Clayton, Deborah Webster–Clair, Kenneth T. Whalum, Jr., Meryl Rice, Jan Liff, Teresa M. Halloran, and Mary Howard Hayes, are eight individual Tennessee voters. In the November 2014 election, plaintiffs voted against Amendment 1, an abortion-related proposed amendment to the Tennessee Constitution. After the election, Tennessee government officials determined that Amendment 1 had passed. Plaintiffs brought suit in the Middle District of Tennessee, asserting claims under 42 U.S.C. § 1983 against the Governor of Tennessee, the Secretary of State, the Coordinator of Elections, the Attorney General, and the State Election Commission and its members (“the State officials”) in their official capacities. Underlying plaintiffs' claims is the theory that, in counting the votes on Amendment 1, the State officials incorrectly interpreted the requirements of Article XI, Section 3 of the Tennessee Constitution, which prescribes Tennessee's constitutional-amendment process. After a bench trial, the district court entered judgment in favor of plaintiffs and issued an injunction requiring the State officials to recount the votes on Amendment 1 in accordance with plaintiffs' proposed interpretation of Article XI, Section 3. This appeal followed. The district court stayed the injunction pending appeal. For the reasons that follow, we reverse the district court's judgment.
I. BACKGROUND 1
A. Factual Background
On the ballot in the Tennessee general election in November 2014 were the governor's race and four proposed amendments to the Tennessee Constitution. Only Amendment 1 is at issue here. It provides:
Nothing in this Constitution secures or protects a right to abortion or requires the funding of an abortion. The people retain the right through their elected state representatives and state senators to enact, amend, or repeal statutes regarding abortion, including, but not limited to, circumstances of pregnancy resulting from rape or incest or when necessary to save the life of the mother.
R. 119, Findings and Conclusions at 5, Page ID 3016.
Amendment 1 was proposed by the Tennessee legislature pursuant to Article XI, Section 3 of the Tennessee Constitution, which prescribes two methods of effectuating an amendment: the legislative, or “referendum,” method and the convention method. Article XI, Section 3 provides that, for amendments going through the legislative method,
if the people shall approve and ratify such amendment or amendments by a majority of all the citizens of the State voting for Governor, voting in their favor, such amendment or amendments shall become a part of this Constitution.
Tenn. Const. Art. XI, § 3. This provision, though succinct, is not eloquent and may be understood in more ways than one. Prior to the November 2014 election, State officials had consistently interpreted Article XI, Section 3 as follows:
Counting the Votes
In order for the amendment to pass and become part of the Constitution, two things must happen:
1. The amendment must get more “yes” votes than “no” votes; and
2. The number of “yes” votes must be a majority of the votes cast in the gubernatorial election.
To determine the votes needed, all votes for all candidates for governor are added together. This number is divided by two or halved. The number of “yes” votes must exceed that number. If the number of “yes” votes exceeds the number, the Constitutional amendment passes and becomes part of the Constitution.
Despite the fact that the number of votes cast for governor is used to determine the outcome, it is not necessary to vote in the governor's race in order to vote on the Constitutional amendment. Likewise, it is not necessary to vote for an amendment in order to vote in the governor's race.
R. 119, Findings and Conclusions at 10–11, Page ID 3021–22 (quoting language previously posted on the Secretary of State website). In fact, no evidence was presented to the district court suggesting that Article XI, Section 3 had ever been interpreted and applied differently by State officials. Id. at 9, n.4, Page ID 3020.
Prior to the November 2014 election, the State officials received inquiries from members of the public, as well as one from the Davidson County Election Commission, regarding how the votes on the proposed amendments would be counted. Specifically, the inquiries sought guidance as to whether Article XI, Section 3 required voters to vote for a gubernatorial candidate in order for their votes to count on the proposed amendments. In accordance with the above interpretation consistently applied by officials in the past, the State officials responded to the effect that voting in the governor's race was not a prerequisite to casting a valid vote on an amendment. R. 119, Findings and Conclusions at 9, Page ID 3020. A spokesperson for the Secretary of State's office also explained this response to the media, which was made known to the public through newspaper articles. Id. at 12, Page 3023.
Supporters and opponents of Amendment 1 then began using the published interpretation of Article XI, Section 3 to enhance or diminish the likelihood of the amendment's approval. Applying the published interpretation, both supporters and opponents surmised that the greater the number of votes cast for governor, the greater would be the ultimate number of votes for Amendment 1 needed for its approval and ratification. Hence, supporters encouraged voters to vote for the amendment and abstain from voting for governor, and opponents urged voters to vote against Amendment 1 and cast a vote for a candidate for governor. The largest campaigns, noted the district court, were mounted by supporters of Amendment 1. Id. at 13, Page ID 3024.
The State officials tabulated the votes after the November 2014 election. A total of 1,430,117 individuals voted, of whom 1,353,728 voted for a gubernatorial candidate. Even more votes—1,386,355—were cast on Amendment 1. For the first time since Article XI, Section 3 was amended in 1953, more votes were cast on a proposed amendment than were cast for governor. Id. at 16, Page ID 3027. The voters favoring Amendment 1 numbered 729,163 and those opposed numbered 657,192. The State officials announced their preliminary determination that Amendment 1 had passed because a majority of the votes cast on the amendment were “yes” votes and because the number of “yes” votes was also greater than a majority of the votes cast for governor. On December 8, 2014, the State officials certified the election results, including the vote on Amendment 1. Id.
B. Procedural Background
Three days after the November 4, 2014 election, plaintiffs filed a complaint in the district court, which they later amended. The amended complaint alleged that the State officials, by their erroneous interpretation of Article XI, Section 3, employed a vote-counting method that (1) violated plaintiffs' due-process rights by creating a fundamentally unfair voting system, and (2) violated plaintiffs' equal-protection rights by diluting their votes. R. 51, Amended Complaint, Page ID 596–600.
Plaintiffs sought a judgment declaring that Article XI, Section 3 required the State officials to count only the votes of those who voted in the gubernatorial race, that the State officials' method of tabulating votes violated the Due Process and Equal Protection Clauses of the Fourteenth Amendment to the United States Constitution, and that the election results were void. Plaintiffs also sought an injunction requiring a recount of the vote on Amendment 1, as well as costs and attorneys' fees.
In response, the State officials moved to dismiss the complaint for lack of jurisdiction and for failure to state a claim upon which relief can be granted. They also asked the district court to either decline to decide the case under the Pullman abstention doctrine or certify the question of how to interpret Article XI, Section 3 to the Tennessee Supreme Court. In July 2015, the district court denied the motion to dismiss and refused to certify the question. The parties began preparing for a bench trial, which was conducted in April 2016.
Meanwhile, in September 2015, the State officials filed a complaint in state court, Hargett v. George, Williamson Cty. Chancery Ct. No. 44460, seeking a declaratory judgment that the vote-counting method they had used represented a proper interpretation of Article XI, Section 3. Named as defendants were the eight individual plaintiffs in the present suit, who moved to dismiss the state court action on various grounds, including that the State officials were seeking an impermissible advisory opinion. Their motion was denied in December 2015. Appellants' Suppl. Br. Ex. A, Hargett v. George, No. 44460, Slip Op. (Dec. 9, 2015).
On April 21, 2016, the state court granted the State officials' motion for summary judgment and issued a declaratory ruling, concluding that Article XI, Section 3 does not require an individual voter to vote for governor as a condition of casting a valid vote on a proposed constitutional amendment. Appellants' Suppl. Br. Ex. B, Hargett v. George, No. 44460, Slip Op. (April 21, 2016). The court concluded its 22–page opinion as follows:
The court hereby declares that Article XI, Section 3, requires that a proposed amendment be first approved by receiving votes constituting a majority of the votes cast on the amendment and then be ratified by receiving votes constituting a majority of the total votes cast in the gubernatorial election. Article XI, Section 3, does not restrict or precondition the right of a citizen to vote for or against a constitutional amendment upon that citizen also voting in the gubernatorial election.
Id. at 22.
The district court entered its judgment the very next day, reaching the opposite conclusion about the proper interpretation of Article XI, Section 3. R. 119, Findings and Conclusions, Page ID 3013. The district court essentially concluded that the plain language of Article XI, Section 3, given its ordinary meaning, requires that a voter who would cast a valid vote for a constitutional amendment must, as a precondition, also vote for governor. Id. at 29–32, Page ID 3040–43. The court further concluded that the State officials' implementation of a contrary construction during the 2014 election resulted in violation of plaintiffs' due process and equal protection rights. Id. at 51, Page ID 3062.2 Accordingly, the district court issued an injunction requiring the State officials to recount the votes on Amendment 1 to determine whether it “passed by a majority of those who voted in the governor's race.” R. 118, Order at 2, Page ID 3011. The court deferred ruling on plaintiffs' request that the election results be voided, however, because of the possibility that the recount results would render further relief moot. This timely appeal followed, and the district court stayed its injunction pending appeal.
A. Preclusive Effect of State Court Ruling
Before addressing the merits of the district court's judgment, we consider a threshold issue raised by the State officials. They argue first that we should give preclusive effect to the declaratory judgment of the state court because the state court's interpretation of Article XI, Section 3, having not been appealed, has since become a final judgment entitled to preclusive effect between the parties.3 Because the state-court judgment became final only after this appeal had been filed, we may consider its preclusive effect in the first instance. See Gooch v. Life Investors Ins. Co. of America, 672 F.3d 402, 418–19 (6th Cir. 2012).4
The preclusive effect of the state court's decision in this federal litigation is governed by Tennessee law. Anderson v. City of Blue Ash, 798 F.3d 338, 350 (6th Cir. 2015). Issue preclusion, also referred to as collateral estoppel, is a judicially created doctrine that “promotes finality, conserves judicial resources, and prevents inconsistent decisions.” Mullins v. State, 294 S.W.3d 529, 534 (Tenn. 2009). Collateral estoppel not only reduces unnecessary litigation and fosters reliance on adjudication, but also promotes comity between state and federal courts. Allen v. McCurry, 449 U.S. 90, 95–96, 101 S.Ct. 411, 66 L.Ed.2d 308 (1980). The doctrine generally bars parties from relitigating issues that have been decided in prior litigation between the same parties or their privies. Mullins, 294 S.W.3d at 534–35.
To prevail with a collateral estoppel claim, the party asserting it must demonstrate (1) that the issue to be precluded is identical to an issue decided in an earlier proceeding, (2) that the issue to be precluded was actually raised, litigated, and decided on the merits in the earlier proceeding, (3) that the judgment in the earlier proceeding has become final, (4) that the party against whom collateral estoppel is asserted was a party or is in privity with a party to the earlier proceeding, and (5) that the party against whom collateral estoppel is asserted had a full and fair opportunity in the earlier proceeding to contest the issue now sought to be precluded.
Id. at 535. The State officials contend that all five of these requirements are met by the state court declaratory judgment ruling. Plaintiffs disagree. We consider the five elements in order.
First, the issue decided in the state court, concerning the meaning of Article XI, Section 3 of Tennessee's Constitution, is an integral component of the district court's ruling that the State officials' implementation of Article XI, Section 3 violated plaintiffs' due process and equal protection rights. The district court arrived at its interpretation of Article XI, Section 3 before the state court's one-day-earlier contrary interpretation became final, but our ruling comes after the state court's interpretation became “final” and, to date, the best state-law authority on the meaning of Article XI, Section 3. This particular issue is common to, and identical in, both the state and federal court proceedings.5
Second, although plaintiffs argue the point, there is no serious contention that the meaning of Article XI, Section 3 was not actually raised, litigated, and decided on the merits in the state-court proceeding.
Third, plaintiffs concede that the state-court ruling became final no later than August 2016, over a year prior to our addressing the issue.
Fourth, plaintiffs do not deny that the same parties are involved in both actions.
Fifth, plaintiffs maintain that preclusive effect should be denied the state-court judgment because they were not granted a full and fair opportunity to contest it. Their argument is premised on the state court's denial of their motion to conduct discovery before “rushing” to judgment as a matter of law on the meaning of Article XI, Section 3. Yet, the state court's declaratory judgment ruling clearly represents a judgment as a matter of law based on the language of Article XI, Section 3 and its legislative history. Plaintiffs have not explained how any expected fruits of discovery would have been material to the court's ruling. Nor, tellingly, did they appeal the state court's ruling.
“The courts have not devised a precise definition of what constitutes the sort of ‘full and fair opportunity to litigate’ that will support the invocation of the doctrine of collateral estoppel.” Mullins, 294 S.W.3d at 538. The question is one of fundamental fairness. Id. Generally, a defendant in the matter “must have had notice of the claim and an opportunity to be heard.” Id. Plaintiffs herein, who were defendants in the state court action, certainly had notice and opportunity to be heard. They challenged the state court's jurisdiction to proceed and vigorously opposed the declaratory relief sought by the State officials. Plaintiffs fall far short of demonstrating how the denial of discovery could be deemed to have resulted in fundamental unfairness. If it had, their recourse lay in appellate review. Plaintiffs have not argued that their right to appellate redress was somehow hindered. If any unfairness resulted, it was not for lack of a full and fair opportunity to litigate.
Accordingly, all five factors of the Mullins standard are satisfied and the state court ruling on the meaning of Article XI, Section 3 should be deemed entitled to preclusive effect in our analysis of the constitutional issues presented in this appeal.
Still, plaintiffs object. They contend that the Mullins factors presume a final judgment by a court of “competent jurisdiction.” See Gibson v. Trant, 58 S.W.3d 103, 113 (Tenn. 2001). Although they acknowledge that the Tennessee Declaratory Judgments Act is “to be liberally construed and administered,” they contend that a Tennessee court does not have competent jurisdiction to issue a declaratory judgment unless presented with a justiciable controversy. See West v. Schofield, 460 S.W.3d 113, 129–30 (Tenn. 2015). To be justiciable, a controversy must present a real question regarding a legally protectable interest, not one that is dependent on a future or contingent event and not one that is theoretical or hypothetical. Id. Because the state-court declaratory relief sought by the State officials regarding their interpretation of Article XI, Section 3 would have prospective effect only, affecting only future contingent elections, plaintiffs contend that the controversy before the state court was not justiciable.
The primary purpose of the Tennessee Declaratory Judgments Act is “to settle and to afford relief from uncertainty and insecurity with respect to rights, status, and other legal relations.” Id. at 129 (quoting Tenn. Code. Ann. § 29-14-113). The Act “should be liberally construed in favor of the person seeking relief in a proper case to the end that rights and interests be expeditiously determined.” Tennessee Farmers Mut. Ins. Co. v. Hammond, 200 Tenn. 106, 290 S.W.2d 860, 862 (1956) (emphasis added). To this end, Tennessee courts have “very wide” discretion in deciding whether to exercise jurisdiction over a complaint for declaratory judgment. State v. Brown & Williamson Tobacco Corp., 18 S.W.3d 186, 193 (Tenn. 2000). An appellate court will disturb the trial court's decision whether to exercise jurisdiction only if arbitrary. Id. at 193–94 (affirming refusal to issue declaratory judgment as not arbitrary); State ex. rel. Earhart v. City of Bristol, 970 S.W.2d 948, 954 (Tenn. 1998) (reversing refusal to grant declaratory judgment as an abuse of discretion).
The State officials respond that the pendency of this federal litigation, which precipitated their state-court action, demonstrates that their declaratory-judgment action presented a real, contemporaneously existing controversy. Indeed, the individual voters' federal claims that the State officials' manner of interpreting and applying Article XI, Section 3 resulted in violation of their federal constitutional rights necessarily presented a real controversy arising under Tennessee law. At the heart of the voters' federal claims is the theory that the State officials misinterpreted Article XI, Section 3. The answer to the question whether the State officials' interpretation was proper under state law would not, as the district court noted, necessarily dictate the outcome of the voters' federal constitutional claims, but the voters' claims were undeniably premised on the charge that the State officials had misinterpreted state law. See R. 62, Memorandum at 18–21, Page ID 781–84. Because this question of state law was integral to the federal litigation, the State officials proceeded with their state-court action for declaratory relief only after the district court had denied their request to abstain or certify the question to the Tennessee Supreme Court. See id.
Moreover, the State officials named the same individual voters who brought the federal action as defendants in the state-court action so that they would have the opportunity to contest it. There was no mystery about the factual background or concreteness of the dispute between the parties. Whatever other significance might be found in the intertwined relationship between the two actions, it can hardly be denied that there was a real live controversy, not simply a hypothetical one, regarding the meaning of Article XI, Section 3. Yes, the controversy stems from the 2014 election, but the meaning of Article XI, Section 3 remained in question, and authoritative resolution of this state-law question would invariably impact the State officials' continuing and future discharge of their duties.
Nor did the justiciability question elude the state court's attention. The defendant voters expressly challenged the state court's jurisdiction to proceed under the Tennessee Declaratory Judgments Act. Rejecting the notion that the court was being asked to issue an advisory opinion, the court observed:
Providing the state officials responsible for the conduct of elections with authoritative interpretation of their legal duties under State law is not, in this context, a hollow exercise.
Appellants' Suppl. Br. Ex. A, Hargett v. George, No. 44460, Slip Op. at 23 (Dec. 9, 2015) (emphasis added). The “context” referred to by the state court was the defendant voters' pending federal civil rights challenge to the manner in which the State officials had interpreted and applied Article XI, Section 3 in the November 2014 election, which had called the State officials' reading into uncertainty, despite its consistency with longstanding practice. With reference to Buntin v. Crowder, 173 Tenn. 388, 118 S.W.2d 221 (1938), where, under analogous circumstances, a justiciable declaratory judgment action was held to exist, the state court further observed:
The Plaintiffs [State officials] are interested in construing Article XI, Section 3 in the manner they have customarily applied it, the Defendants are interested in making the Plaintiffs apply Article XI, Section 3 in the manner they prefer.
Appellants' Suppl. Br. Ex. A, Hargett v. George, Slip. Op. at 24. The state court thus determined that it had jurisdiction to proceed under the Declaratory Judgments Act.
The defendant voters (i.e., plaintiffs herein) did not appeal the state court's jurisdictional determination or its final decision and we are not in a position, of course, to review either of the state court's rulings. Yet, cursory examination of the state court's 26–page opinion denying the defendant voters' motion to dismiss reveals that the court did not exercise jurisdiction lightly or carelessly. The state court has not been shown to have so manifestly abused its discretion as to persuade us to hold that it acted without competent jurisdiction.
Plaintiffs argue that the justiciability of the state-court action should be evaluated without regard to their federal action. If we indulge the fiction that the controversy surrounding the 2014 election does not exist, as they propose, then it becomes clear that the State officials' complaint for declaratory relief pertains only to future contingent events. The argument is based on the notion that the declaratory judgment could not include relief affecting the outcome of the 2014 election; therefore, the declaratory relief afforded could be prospective only, affecting only future applications of Article XI, Section 3. Although plaintiffs' position grows stronger if the inquiry whether there is a real controversy between the parties ignores the real controversy between the parties, we find no merit in such an artificial approach to justiciability.6
Further, recognizing that declaratory relief afforded by the state court would not affect the outcome of the 2014 election does not dictate the conclusion that the controversy spawned by that election is not real. Until plaintiffs filed their federal civil rights action, State officials had consistently interpreted and applied Article XI, Section 3 in what appeared to be a reasonable manner without controversy. Only after that filing did the controversy emerge. Nor does the necessarily prospective nature of declaratory relief mean that the controversy at stake is too contingent and hypothetical. Article XI, Section 3 is part of the Tennessee Constitution which the State officials are charged with administering when a proposed constitutional amendment makes its way onto the ballot. The controversy surrounding the 2014 election is the real event regarding a legally cognizable question that affords the occasion for judicial clarification of the state law that the State officials are charged with administering.
But if their federal civil rights action is to be recognized as creating the controversy, then, plaintiffs contend, it should also be deemed to render declaratory relief in a different forum inappropriate. They point to two cases where the Tennessee Supreme Court upheld lower court refusals to entertain actions of declaratory judgment. In Nicholson v. Cummings, 188 Tenn. 201, 217 S.W.2d 942 (1949), the court held that refusal of declaration was proper where the facts were disputed, the declaration would not have terminated the controversy, and the declaration “could have been no more than a stepping-stone to further litigation.” Id. at 943. In Burkett v. Ashley, 535 S.W.2d 332 (Tenn. 1976), the court affirmed the denial of declaratory relief where the plaintiff failed to allege sufficient facts to establish a real controversy and where the declaration had been sought to circumvent an alimony obligation imposed in another court. The court observed that “a declaration will not be given in aid of another proceeding then pending.” Id. at 333.
Both opinions are short and both upheld the lower court's exercise of discretion in sensible, straightforward rulings. Each case involved the pendency or prospect of other related litigation, as here. Yet, the “other related litigation” in this case presented federal constitutional claims creating a real controversy of state law ripe for declaratory relief, distinguishing our case from Burkett. And the State officials' request for declaratory relief to resolve this unsettled question of state law was based on undisputed facts, distinguishing our case from Nicholson.
Moreover, the question addressed to the state court, integral to the pending federal civil rights claims, is a question the district court had declined to certify to the Tennessee Supreme Court for an authoritative state-court interpretation of the Tennessee Constitution. Under these circumstances, we find it highly doubtful that the Tennessee Supreme Court—if it had been asked to review the state court's decision to adjudicate the state law question, which it was not—would have held the decision arbitrary and vacated it for lack of jurisdiction. Rather, it is far more likely that the Tennessee high court, consistent with its own precedents, would have upheld the trial court's decision to exercise jurisdiction as being within the trial court's “very wide discretion.” Cf. State ex. rel. Earhart, 970 S.W.2d at 955 (“Where there is presented a significant issue that needs resolving, as in this case, refusing to issue a declaratory judgment cannot be excused on the basis of discretion.”).
Accordingly, we cannot conclude that the state court was without competent jurisdiction to address the State officials' complaint for declaratory relief. No party having appealed the state court's ruling, it has become final and binding between the parties to that action, who are also the parties to the instant appeal. Mullins, 294 S.W.3d at 535. It follows that the state-court declaratory judgment on the meaning of Article XI, Section 3 is now entitled to conclusive effect, and the parties are barred from relitigating its meaning. Id. at 534. This result is entirely consonant with the purposes of collateral estoppel: promoting efficiency, judicial economy and comity, and preventing inconsistent decisions. Allen, 449 U.S. at 95–96, 101 S.Ct. 411; Mullins, 294 S.W.3d at 534.
In other words, that portion of the district court opinion devoted to discerning the meaning of Article XI, Section 3 is, for purposes of the instant dispute between these parties, effectively supplanted by the state court's contrary interpretation, which is binding on the parties. The district court's determination that Article XI, Section 3 means that “voters must vote for governor in addition to voting on a proposed amendment” (i.e., to cast a valid vote for the amendment), R. 119, Findings and Conclusions at 29, Page ID 3040, is supplanted by the state court's declaration “that Article XI, Section 3, requires that a proposed amendment be first approved by receiving votes constituting a majority of the votes cast on the amendment and then be ratified by receiving votes constituting a majority of the total votes cast in the gubernatorial election.” Appellants' Suppl. Br. Ex. B, Hargett v. George, Slip Op. at 22.
This means that the method of counting votes employed by the State officials in the 2014 election was faithful to the actual meaning of Article XI, Section 3. Although this undermines the district court's analysis of plaintiffs' civil rights claims, it does not necessarily dictate a different outcome. We next consider the impact of this clarification of state law on the merits of plaintiffs' due process and equal protection claims.7
B. Due Process
1. Standard of Review
On appeal from a judgment following a bench trial, we review de novo the district court's conclusions of law. Beaven v. U.S. Dep't of Justice, 622 F.3d 540, 547 (6th Cir. 2010). The district court's findings of fact are reviewed for clear error. Id. The scope of the injunctive relief ordered by the district court is reviewed under the abuse-of-discretion standard. Lee v. City of Columbus, 636 F.3d 245, 249 (6th Cir. 2011).
Plaintiffs here allege that the State officials' manner of counting the votes in the 2014 election impermissibly infringed or burdened their right to vote. Indeed, the right to vote, being fundamental, is afforded special protection by the courts. Ohio Democratic Party v. Husted, 834 F.3d 620, 626 (6th Cir. 2016). Yet, common sense dictates that substantial regulation of elections is required if they are to be fair and honest. Id. “Federal law thus generally defers to the states' authority to regulate the right to vote.” Id. When these interests come into tension and ripen into a constitutional challenge, we evaluate the State's actions under the so-called Anderson–Burdick framework, derived from the Supreme Court's rulings in Anderson v. Celebrezze, 460 U.S. 780, 103 S.Ct. 1564, 75 L.Ed.2d 547 (1983), and Burdick v. Takushi, 504 U.S. 428, 112 S.Ct. 2059, 119 L.Ed.2d 245 (1992). In Ohio Democratic Party, we described the framework as follows:
Though the touchstone of Anderson–Burdick is its flexibility in weighing competing interests, the “rigorousness of our inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights.” Burdick, 504 U.S. at 434, 112 S.Ct. 2059. This flexible balancing approach is not totally devoid of guidelines. If a state imposes “severe restrictions” on a plaintiff's constitutional rights (here, the right to vote), its regulations survive only if “narrowly drawn to advance a state interest of compelling importance.” Id. On the other hand, “minimally burdensome and nondiscriminatory” regulations are subject to a “less-searching examination closer to rational basis” and “ ‘the State's important regulatory interests are generally sufficient to justify the restrictions.’ ” Ohio Council 8 Am. Fed'n of State v. Husted, 814 F.3d 329, 335 (6th Cir. 2016) (citing Green Party of Tennessee v. Hargett, 767 F.3d 533, 546 (6th Cir. 2014), and quoting Burdick, 504 U.S. at 434 [112 S.Ct. 2059]). Regulations falling somewhere in between—i.e., regulations that impose a more-than-minimal but less-than-severe burden—require a “flexible” analysis, “weighing the burden on the plaintiffs against the state's asserted interest and chosen means of pursuing it.” Hargett, 767 F.3d at 546.
834 F.3d at 627. The district court recognized and purported to apply the Anderson–Burdick approach. It concluded that the vote-counting method employed by the State officials imposed a more-than-minimal burden on voting rights and therefore warranted a flexible weighing of the state's asserted interest against the burden on plaintiffs' rights. We consider each of plaintiffs' three theories of relief in order.
2. Plain Language Theory
Plaintiffs' due process claim is two-pronged. First, they contend the State officials' vote-counting method was fundamentally unfair because it was not in conformity with the plain language of Article XI, Section 3. The district court agreed with plaintiffs and found in their favor on this theory.
The relevant language of Article XI, Section 3 bears repeating:
[I]f the people shall approve and ratify such amendment or amendments by a majority of all the citizens of the State voting for Governor, voting in their favor, such amendment or amendments shall become a part of this Constitution.
Tenn. Const. Art. XI, § 3. The district court adopted plaintiffs' literal reading of this language as meaning that only the votes of those persons who cast votes for a gubernatorial candidate can count toward the approval and ratification of an amendment. The court rejected the State officials' position that this provision requires only that an amendment be approved by a number of “yes” votes exceeding the number of “no” votes, and be ratified by a number of “yes” votes at least equal to a majority of the number of citizens who cast votes in the governor's race, without reference to who, personally, cast the votes.
The state court reached a contrary conclusion, holding that, although the proposed literal reading of Article XI, Section 3 was plausible, it was not reasonable because it would have the effect of disenfranchising—by placing extra qualifications on—voters who wished to vote for an amendment but who did not wish to vote for a gubernatorial candidate. The state court thus held that the State officials' interpretation is the only reasonable interpretation. Appellants' Suppl. Br. Ex. B, Slip Op. at 21. As explained above, this interpretation of Article XI, Section 3, has become final and binding on the parties and is preclusive of any inconsistent interpretation in this litigation.8
Accordingly, insofar as the district court's analysis was premised on its determination that the State officials' vote-counting method was fundamentally unfair because inconsistent with Article XI, Section 3, it must be vacated. To its credit, however, the district court recognized that plaintiffs' claims were viable irrespective of the interpretation that ultimately prevailed. The court's opinion thus includes assessment of plaintiffs' due process and equal protection claims as though the State officials' interpretation was correct.
3. Vote–Dilution Theory
In their second due-process theory of relief, plaintiffs contend that even if the State officials' vote-counting method is accepted as a reasonable implementation of the language of Article XI, Section 3, it still results in a fundamentally unfair infringement of their rights. The district court determined that this burden on plaintiffs' rights warranted “mid-level scrutiny.” R. 119, Findings and Conclusions at 38, Page ID 3049. The court did not, however, identify what the burden is. Even though the “rigorousness of our inquiry” under Anderson–Burdick depends on the extent to which the challenged vote-counting method burdened plaintiffs' rights, the district court did not identify the burden. The court characterized the burden as falling squarely between the extremes of a severe burden and a minimal burden, but did not say how plaintiffs' voting rights were so burdened as to deny them due process.
Plaintiffs contend that the State officials' vote-counting method had the effect of according greater value to the votes of those persons who voted in support of the amendment but did not vote for governor than the votes of those persons who opposed the amendment but who voted for governor. This effect is said to have been fundamentally unfair in that it diluted the value of some voters' votes in relation to others' votes. Concomitantly, plaintiffs argue, to avoid this vote-dilution effect, they felt compelled to vote for governor. Plaintiffs' argument fails to identify how any voter's right to vote was burdened by governmental action, apart from the alleged disparate treatment, a concern more properly addressed under their equal protection theory, discussed infra.
The district court seemed to accept at face value plaintiffs' argument that “different voters were treated differently” as substantiating the existence of a more-than-minimal burden. Yet, as the district court itself recognized, quoting Obama for America v. Husted, 697 F.3d 423 (6th Cir. 2012), “[i]f a plaintiff alleges only that a state treated him or her differently than similarly situated voters, without a corresponding burden on the fundamental right to vote, a straightforward rational basis standard of review should be used,” id. at 429, not intermediate, or “mid-level,” scrutiny. There is no basis in the district court record for finding that any particular plaintiff's, or any particular voter's, right to vote for or against Amendment 1, or right to vote for governor or not, was hindered or burdened (or even treated differently, for that matter) by any actions of the State officials. It appears that every Tennessee voter was free to vote his or her conscience on the amendment and for governor. Plaintiffs have thus failed to identify a more-than-minimal burden on their right to vote that would warrant more rigorous examination than rational-basis scrutiny.
A state's election process may be found fundamentally unfair only in the “exceptional case,” such as where “a state employs ‘non-uniform rules, standards and procedures' that result in significant disenfranchisement and vote dilution ․ or significantly departs from previous state election practice.” Warf v. Bd. of Elections of Green Cty., 619 F.3d 553, 559 (6th Cir. 2010) (quoting League of Women Voters v. Brunner, 548 F.3d 463, 478 (6th Cir. 2008)). None of these irregularities is presented in this case. State officials did not apply the process required by Article XI, Section 3 non-uniformly or in a manner at odds with previous practice. Rather, it is undisputed that the State officials' interpretation of Article XI, Section 3 was consistent with established practice and consistent with the interpretation that had been published to the citizenry through the media prior to the election.9
An actionable due process claim may also be implicated where a state's election process impaired citizens' ability to participate in state elections on an equal basis with other qualified voters. See Phillips v. Snyder, 836 F.3d 707, 716 (6th Cir. 2016). Here, the State officials' vote-counting method did not impair any voter's freedom and ability to participate equally in the election. Still, plaintiffs maintain that the State officials' vote-counting method was susceptible to manipulation by campaigns for and against Amendment 1. Because adoption of the amendment required both approval by majority vote and ratification by votes equal in number to a majority of the votes cast for governor, campaigns for and against the amendment encouraged their supporters to either vote for governor, or refrain from voting for governor, in order to influence the likelihood of ratification and adoption. These efforts, plaintiffs speculate, may have affected the outcome.
Yet, as the district court observed, such strategic choices by interested groups of private citizens “were entirely within their prerogative and not this Court's concern.” R. 119, Findings and Conclusions at 45–46, Page ID 3056–57. The due process analysis is concerned rather with whether plaintiffs were subject to fundamentally unfair treatment as a result of governmental action. The record discloses no grounds for holding that the State officials' management of the 2014 election, in accordance with the Tennessee Constitution, deprived plaintiffs of equal freedom and ability to participate or otherwise burdened their right to vote.
Any “burden” imposed on plaintiffs' rights as a result of Tennessee's approval-and-ratification process was no more than “minimal.” The record shows that the purpose of the ratification threshold in Article XI, Section 3 is to ensure that proposed amendments to the constitution enjoy considerable statewide support. This objective of preventing a too-easy amendment of the Tennessee Constitution by small well-organized interest groups is undeniably a legitimate governmental purpose.10 See Gordon v. Lance, 403 U.S. 1, 5–8, 91 S.Ct. 1889, 29 L.Ed.2d 273 (1971) (recognizing that state-law departures from strict majority rule, making certain kinds of governmental action more difficult, are legitimate as long as they do not discriminate against any identifiable class based on extraneous conditions like race, wealth, tax status or military status). And Article XI, Section 3, albeit inartfully drawn, as implemented by the State officials, is nondiscriminatory and rationally related to the purpose. This is sufficient to satisfy rational-basis scrutiny; it is irrelevant that the state might have chosen a better method of furthering its purpose. See Taxpayers United for Assessment Cuts v. Austin, 994 F.2d 291, 298–99 (6th Cir. 1993). Accordingly, we find no cognizable infringement of plaintiffs' due process rights. This is not the “exceptional case” that warrants federal intervention in a lawful state election process. See Warf, 619 F.3d at 559. The district court's judgment on this count must be reversed.
C. Equal Protection
The State officials also challenge the district court's ruling on plaintiffs' equal protection claim. The district court identified plaintiffs as members of a class consisting of voters who both cast a vote on Amendment 1 and cast a vote in the gubernatorial race. The court held that, under the State officials' vote-counting method, members of the plaintiffs' “class” were subject to unequal treatment in comparison with voters who voted on Amendment 1 but not for governor. As a result of the Article XI, Section 3 ratification requirement, the value of plaintiffs' votes against Amendment 1 was held to be diluted by their votes for governor in comparison with the value of votes against Amendment 1 by voters who refrained from voting for governor. Consequently, the court concluded, voters who opposed Amendment 1 were unfairly compelled to vote for governor (to maximize the value of their vote against Amendment 1) whereas voters who supported Amendment 1 were not subject to the same compulsion. R. 119, Findings and Conclusions at 43–46, Page ID 3054–57.
“Equal protection of the laws” means that “[h]aving once granted the right to vote on equal terms, the State may not, by later arbitrary and disparate treatment, value one person's vote over that of another.” Bush v. Gore, 531 U.S. 98, 104–05, 121 S.Ct. 525, 148 L.Ed.2d 388 (2000). Voting rights can be impermissibly burdened “by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise.” Id. (quoting Reynolds v. Sims, 377 U.S. 533, 555, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964)). “Our Constitution leaves no room for classification of people in a way that unnecessarily abridges this right [to vote].” Reynolds, 377 U.S. at 559, 84 S.Ct. 1362 (quoting Wesberry v. Sanders, 376 U.S. 1, 17–18, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964)). “[A]ll who participate in the election are to have an equal vote—whatever their race, whatever their sex, whatever their occupation, whatever their income, and wherever their home may be.” Reynolds, 377 U.S. at 557–58, 84 S.Ct. 1362 (quoting Gray v. Sanders, 372 U.S. 368, 379, 83 S.Ct. 801, 9 L.Ed.2d 821 (1963)). Thus, “a law that would expressly give certain citizens a half-vote and others a full vote” would be violative of the Equal Protection Clause. Wesberry, 376 U.S. at 19, 84 S.Ct. 526 (quoting Colgrove v. Green, 328 U.S. 549, 569, 66 S.Ct. 1198, 90 L.Ed. 1432 (1946)).
On the other hand, broadly applicable and nondiscriminatory laws are generally presumed to pass muster under equal protection scrutiny. Ohio Democratic Party, 834 F.3d at 631. Any minimal restriction of some identifiable class of voters' right to vote is ordinarily deemed justified by the state's “important regulatory interests.” Id. (citing Obama for America, 697 F.3d at 433–34); see also Northeast Ohio Coalition, 837 F.3d at 631.
The State officials contend that the district court erred in concluding that plaintiffs are members of a class of voters who were subject to disparate treatment during the 2014 election. They maintain that their implementation of Article XI, Section 3 was nondiscriminatory and that all citizens' votes, whether on Amendment 1 or for governor, were weighted equally. They acknowledge that a two-step process determined the ultimate outcome on the adoption of Amendment 1. That process involved both an approval step and a ratification step. According to the State officials, “approval” of the amendment required more “yes” votes than “no” votes; i.e., at least a majority of the vote. “Ratification” required that the number of votes for the amendment be at least equal to a majority of the number of votes cast in the governor's race. As the State officials explained, the outcome of each step was a function of the number of votes cast, not a function of who had cast them. Every vote cast—on the amendment and in the governor's race—was accorded the same weight.
The district court's analysis was driven by concerns stemming from the interplay of the two steps in the process. And these concerns were derived from plaintiffs' emphasis on who voted how in both steps. From the beginning, plaintiffs' civil rights claims have been premised on a rigid, literal reading of Article XI, Section 3 that would have allowed the counting of votes for and against Amendment 1 only of those voters who had voted for governor. This interpretation has now been conclusively rejected by the state court declaratory judgment ruling. Yet, the notion that the State officials' method of counting the votes on Amendment 1 should have been linked to those voters who voted for governor has continued to pervade plaintiffs' arguments. That is, despite the state-court declaration that the State officials' method—of simply counting the numbers of votes, irrespective of who cast them—was reasonable and faithful to the meaning of Article XI, Section 3, plaintiffs have maintained, and the district court agreed, that this method effectively treated a class of voters disparately and unfairly, depending on how they voted.
Again, any “disparity” was the result, not of Article XI, Section 3, or of the State officials' vote-counting method, or of any other actions by the State officials, but of strategic choices made by members of the voting public to maximize the impact of their votes. Plaintiffs' suspicion that their opposition to Amendment 1 might have been disadvantaged by the choices of voters who supported Amendment 1 simply does not make out a cognizable claim for denial of equal protection. Neither the language of Article XI, Section 3, nor the State officials' implementation thereof resulted in “classifications” of voters. The record does not support a finding that plaintiffs or their votes were treated unequally because of their race or sex or occupation or income or place of residence or any other characteristic. Their votes were counted in the same manner as all other voters'. The record is devoid of evidence that the State officials treated plaintiffs' votes in opposition to Amendment 1 any differently than they treated others' votes in support of Amendment 1. Nor is there evidence that the State officials treated plaintiffs' votes in the governor's race any differently than Amendment 1 supporters' votes in the governor's race.
Plaintiffs' arguments amount to little more than a complaint that the campaigns in support of Amendment 1, operating within the framework established by state law, turned out to be more successful than the campaigns against Amendment 1. That private-citizen supporters of Amendment 1 may have endeavored to lower the ratification threshold by refraining from voting for governor does not support a finding that State officials' actions had the effect of “diluting” the value of plaintiffs' votes in opposition to Amendment 1. Nor does it support a finding that State officials' actions “compelled” plaintiffs to vote for governor in order to raise the ratification threshold. Quite to the contrary, as the district court noted, plaintiffs enjoyed the same “freedom” as their adversaries to operate within the established framework to promote their opposition to Amendment 1. They were equally free to campaign for greater voter turnout in the governor's race in order to increase the chances that Amendment 1 would fail. The Equal Protection Clause simply does not authorize federal courts to correct a perceived unfairness in state election processes that results from voters' choices rather than from governmental action. If a generally applicable nondiscriminatory law turned out to be susceptible to unfair manipulation by members of the voting public, this would be a matter for state authorities to remedy, not for federal intervention.
Again, plaintiffs have failed to identify how their right to vote was burdened by disparate treatment as a result of the State officials' actions implementing Article XI, Section 3. To the extent that the vote-counting implementation of the ratification requirement is alleged to have disadvantaged plaintiffs in their opposition to Amendment 1, any such “burden” was reasonably justified by the State's interest in ensuring that a proposed constitutional amendment enjoy widespread support as a prerequisite to adoption. We therefore conclude that the district court's judgment in favor of plaintiffs on their equal protection claim must also be reversed.
The subject matter of Amendment 1, touching on abortion rights, is politically sensitive and controversial. The amendment's adoption was closely contested in the 2014 election and, though it appeared to have been approved by the electorate (approximately 53% of the votes cast), its status has remained unresolved pending this appeal. In this litigation, treating matters of process, the issues raised have been fully aired and vigorously disputed. Although the subject of abortion rights will continue to be controversial in Tennessee and across our nation, it is time for uncertainty surrounding the people's 2014 approval and ratification of Amendment 1 to be put to rest.
For the reasons set forth above, we conclude that the vote-counting method employed by defendant State officials in the 2014 election was reasonable and true to the meaning of Article XI, Section 3 of the Tennessee Constitution, as confirmed by the state-court's declaratory judgment ruling. Further, we hold that the State officials' actions did not result in any infringement of plaintiffs' voting rights. The district court's judgment in favor of plaintiffs on their due process and equal protection claims must therefore be REVERSED. The case is REMANDED for entry of an order VACATING the district court's injunctive order and entry of JUDGMENT in favor of defendant State officials.
1. This background summary is drawn principally from the district court's opinion, which, in this respect, is not objected to by the parties.
2. The district court acknowledged the state court's contrary ruling on the meaning of Article XI, Section 3, but held that the declaratory relief awarded by the state court was prospective only, was not a final judgment, and did not address the claim that the November 2014 application of that interpretation had resulted in violation of some voters' due process and equal protection rights under the United States Constitution. Id. at 25–26, Page ID 3036–37.
3. The impact of the state-court ruling was not addressed by the parties below because it issued after briefing and oral arguments had been completed in the district court proceedings and just one day prior to the district court's declaratory judgment ruling. Nonetheless, the district court was cognizant of the state-court ruling and did not disregard the state-court ruling in its opinion, as discussed infra at n.5.
4. In Gooch, the court went so far as to consider sua sponte the preclusive effect of a state-court judgment that had not become final until after district court proceedings had been completed. Because the judgment had since become final, it was deemed entitled to respect, and because its preclusive effect posed a purely legal issue that was presented with sufficient clarity and completeness in the parties' briefs, the court held that sua sponte consideration was appropriate. Gooch, 672 F.3d at 419. See also MSI Regency, Ltd. v. Jackson, 433 Fed.Appx. 420, 430 (6th Cir. 2011) (same).Here, in comparison, consideration of the preclusive effect of the state-court judgment is even more appropriate: the state-court judgment became final only after this appeal commenced; the preclusive effect is a purely legal issue; the issue is not addressed sua sponte, but has been raised and fully briefed by the parties; and, moreover, the court has requested and received supplemental briefing.
5. The district court recognized the commonality of this issue in both cases. R. 119, Findings and Conclusions at 24–26, Page ID 3035–37. In declining to abstain from exercising jurisdiction, the district court gave three reasons for not abiding by the state court's interpretation: (1) the state court's ruling purported to grant only prospective relief and did not address the propriety of the State officials' implementation of Article XI, Section 3 in the 2014 election; (2) the state court's ruling did not address the propriety of the State officials' actions in the 2014 election under the United States Constitution; and (3) the state court's ruling was appealable and therefore not a final judgment at the time of the district court's ruling. Id.
6. Moreover, we note that the State officials' resort to state-court declaratory relief represented a response to the district court's denial of Pullman abstention. Notwithstanding the district court's refusal to hold proceedings on the § 1983 civil rights claims in abeyance pending clarification of the unsettled question of state law, per Railroad Comm'n of Texas v. Pullman Co., 312 U.S. 496, 61 S.Ct. 643, 85 L.Ed. 971 (1941), the State officials pursued state-court clarification concurrently with the ongoing federal-court proceedings. If abstention had been granted, then, to properly reserve disposition of the federal claims in federal court, per England v. Louisiana State Bd. of Med. Examiners, 375 U.S. 411, 421, 84 S.Ct. 461, 11 L.Ed.2d 440 (1975), it would have been obligatory to “inform” the state court of the outstanding federal claims, per Gov't and Civic Employees Organizing Committee, C.I.O. v. Windsor, 353 U.S. 364, 77 S.Ct. 838, 1 L.Ed.2d 894 (1957). It was thus appropriate, analogously, for the State officials to so inform the state court in the concurrently proceeding declaratory judgment action, and for the state court to consider the information in assessing the justiciability of the claim before it.
7. The State officials have also challenged the district court's refusals to abstain and to certify the question of state law to the Tennessee Supreme Court. Our ruling on the preclusive effect of the state-court declaratory judgment effectively renders both challenges moot. Because the state court acted so promptly and its ruling became final so quickly, the resulting state-court confirmation of the meaning of Article XI, Section 3 produced the very benefits (i.e., comity and avoidance of inconsistent rulings), by way of issue preclusion, that the State officials had endeavored to achieve though Pullman abstention or certification of the question.If the district court had abstained, the parties would have pursued clarification of the unsettled issue of state law in state court, while reserving adjudication of residual questions of federal constitutional law in federal court, per England, 375 U.S. at 421, 84 S.Ct. 461. Similarly, if the unsettled question of state law had been certified to the Tennessee Supreme Court (an option potentially even more efficient than, and therefore preferred over, Pullman abstention), the federal court would have postponed further proceedings on the federal claims pending state-court clarification. See Jones v. Coleman, 848 F.3d 744, 749–50 (6th Cir. 2017). When the district court declined to employ either option (on which we express no opinion), the State officials pursued a third course, which, albeit unorthodox, turned out to be an efficient and fruitful substitute.
8. If the state-court ruling were not binding, and we were called upon to review the district court's interpretation of Article XI, Section 3, we would hold, consistent with the state court's ruling, that the State officials' actions interpreting and implementing the provision, viewed through the lens of plaintiffs' due process and equal protection claims, were not unreasonable.In addition to the reasons cited in the state court's opinion, we note that plaintiffs' preferred reading of the text of Article XI, Section 3, while not implausible on its face, would be patently unreasonable in effect. Not only would their proposed construction—requiring a voter to vote for governor as a prerequisite to casting a valid vote on Amendment 1—contravene longstanding practice and pre-election instructions published to the public, and effectively nullify the votes of thousands of citizens; it would also conflict with another provision of the Tennessee Constitution. Article IV, Section 1 prohibits the imposition of any additional qualification to vote, beyond age, U.S. citizenship, state residency, and registration. To adopt plaintiffs' proposed interpretation would be to run afoul of our obligation, in construing state law, “to avoid constitutional difficulty” when fairly possible. Green Party ofTenn. v. Hargett, 700 F.3d 816, 825 (6th Cir. 2012); Davet v. City of Cleveland, 456 F.3d 549, 554 (6th Cir. 2006). See also Estate of Bell v. Shelby Cty. Health Care Corp., 318 S.W.3d 823, 835 (Tenn. 2010) (“No constitutional provision should be construed to impair or destroy another provision.”). Because the State officials' interpretation of Article XI, Section 3 was and is reasonable, this potential “constitutional difficulty” is easily avoided.
9. If, instead, the State officials had altered or departed from the established practice prior to the 2014 election without giving adequate notice of the change to the citizenry, then a stronger due process claim would be made out.
10. It is, in fact, precisely the sort of safeguard that plaintiffs and other opponents of Amendment 1 would ordinarily be expected to wholeheartedly endorse. Their grievance in this case thus appears to be driven by regrets, not so much that the State officials' actions infringed their rights, but that their “adversaries,” supporters of Amendment 1, may have campaigned more effectively than did opponents of Amendment 1.
McKEAGUE, Circuit Judge.