My initial reaction to the lawsuits—before I had chance to examine the relevant Ohio statutes—was that Republicans should lose the first case and win the second. That is, I entered into the statutory analysis with a thumb on the scale in favor of voter enfranchisement, which could be overcome only by clear statutory language to the contrary or strong competing policy reasons. Eventually, the Ohio Supreme Court, relying on such a canon of construction favoring voters, indeed sided with the voters in both cases.¹

This “Democracy Canon” of statutory construction, as I call it, has long and broad support in state courts, from cases in the 1800s through those decided in the 2008 election season. But it has been ignored by legislation and election law scholars and appears to have no independent vitality in federal courts. Its origins trace back to at least 1885. In that year, the Supreme Court of Texas declared in Owens v. State that “[a]ll statutes tending to limit the citizen in the exercise of [the right of suffrage] should be liberally construed in his favor.”² The Owens court rejected an argument by one of the candidates in an election contest that ballots marked with information such as the name and address of the President and Vice President or the counties in which presidential electors resided should not be counted because they violated a state statute barring the counting of ballots containing pictures, signs, vignettes, stamp marks, or devices.

The Democracy Canon is a legitimate tool of statutory interpretation that should be expanded to federal courts, or at least be accepted in federal courts as a legitimate tool of statutory interpretation by state courts. To be sure, the Canon’s use raises some dangers of exacerbating the actuality and appearance of the politicization of the judiciary and, in some cases, some knotty federalism questions. Nonetheless, state legislatures, rather than federal courts, are the institutional actors best situated to rein in potential state court overreaching.

 
I.
History and Importance

The Democracy Canon is the Rodney Dangerfield of canons.³ Because of its use primarily in state courts rather than federal courts, it is not on canonical lists of statutory canons; nor is it discussed in legislation or election law casebooks. Nonetheless, the Canon has had long and consistent acceptance in state courts. The rule announced in 1885 by the Supreme Court of Texas in Owens has been followed by courts throughout the United States. Interpretations of statutes in favor of a broad right to vote continued to be prevalent throughout the twentieth century, and many of the same themes have carried through into modern cases, with cases as recent as the 2008 election season relying on the Canon.

Since Owens, the Democracy Canon has been applied primarily in three contexts: vote counting cases, in which someone relies upon the Canon to argue, following an election, for the counting of ballots that have not been counted because of minor voter error, election official error, or a disputed reading of a relevant statute; voter eligibility/registration cases, in which someone relies upon the Canon to argue, before an election, that a voter or certain group of voters who have been told they cannot vote should be allowed to cast a ballot that will be counted even though election officials have determined they cannot register or vote because of minor voter error, election official error, or a disputed reading of a relevant statute; and candidate/party competitiveness cases, in which a candidate or political party relies upon the Canon (and particularly upon the voters’ right to vote in a competitive election) to argue, before an election, that a certain candidate or party should be allowed to run in an election or appear on an election ballot, even though election officials have excluded the candidate or party from the ballot because of minor candidate or party error, election official error, or a disputed reading of a relevant statute. Vote counting cases are the most prevalent type of cases relying on the Democracy Canon, but the Canon has been deployed in all three kinds of cases across a number of states over more than a century.

The Canon’s stated purposes usually are described in terms of its role in fostering democracy. Its purpose is “to give effect to the will of the majority and to prevent the disfranchisement of legal voters.”⁴ It plays a role in “favoring free and competitive elections.”⁵ It recognizes that the right to vote “is a part of the very warp and woof of the American ideal and is a right protected by both the constitutions of the United States and of the state.”⁶ Liberal construction of election laws serves “to allow the greatest scope for public participation in the electoral process, to allow candidates to get on the ballot, to allow parties to put their candidate on the ballot, and most importantly to allow voters a choice on Election Day.”⁷

Though the Democracy Canon is usually the result of judicial declaration, it sometimes appears explicitly as a legislatively drafted rule of interpretation. For example, a provision of the Kansas statutes governing rules for regulating elections and voting states that “[t]he provisions of this act shall be construed liberally for the purpose of effectuating its purposes.”⁸ Though the state courts have relied heavily on the Democracy Canon for well over a century, it has been cited much more rarely in federal courts. I have not discovered any federal cases considering whether federal laws governing the casting and counting of ballots—such as the Help America Vote Act (“HAVA”),⁹ the National Voter Registration Act (“NVRA”),¹⁰ or the Uniformed and Overseas Citizens Absentee Voting Act (“UOCAVA”)¹¹—should be liberally construed in favor of the rights of voters.

The Canon likely has not had yet gained independent vitality in federal courts for two reasons. First, since the founding of the Republic, there has been much more state law rather than federal law governing the nuts-and-bolts of voting and registration thanks to the decentralized nature of elections in this country. Thus, federal courts until recently simply have not had the same opportunities to construe election statutes as have states courts; there has not been much federal statutory election administration law to construe. Federal courts have certainly been active in election law cases, especially since the 1960s. But these have been primarily constitutional cases, not statutory cases involving the interpretation of federal statutes governing the nuts-and-bolts of election administration.

Second, the Canon has not spread to federal courts because legislation courses and treatises tend to focus on canons in federal courts. As the Democracy Canon’s widespread and longstanding use in state courts becomes more widely known in the legislation and election law fields, federal courts are more likely to adopt it.

Though the Canon’s use in state courts is longstanding and broad, there is some variation in (1) the scope and reach of the Canon; (2) the strength of the Canon; (3) and when it is triggered. In addition, the Canon’s reach is subject to two important limitations. First, courts will not apply the Canon when there are serious allegations of fraud. Second, in those cases involving voter error or candidate/party error, courts tend to limit the reach of the statute to cases involving minor errors (what the courts often term “substantial compliance” with the relevant statute).

Before turning to a normative defense of the Canon, and a discussion of political and federalism issues surrounding its use, it is worth putting the debate over the Canon’s application in perspective: there is a lot more election law litigation now, and most of it is statutory. In the pre-2000 period, state and federal courts handled an average of about 94 election-related cases per year. In the 2000-2008 period, that number has more than doubled to an average of 237 election cases per year. Considering only state election law cases, statutory interpretation questions arise in the vast majority of cases. In 2008, for example, over 81 percent of cases involved either statutory interpretation questions (70.8%) or a mix of statutory and constitutional issues (10.6%).

 
II.
Normative Defense of the Democracy Canon

The Democracy Canon is a “substantive canon.” Substantive canons “are generally meant to reflect a judicially preferred policy position. [They] reflect judicially-based concerns, grounded in the courts’ understanding of how to treat statutory text with reference to judicially perceived constitutional priorities, pre-enactment common law practices, or specific statutorily based policies.”¹²

Substantive canons are quite controversial.¹³ Justice Scalia, though he sometimes applies them, argues against substantive canons, which he characterizes as “the use of certain presumptions and rules of construction that load the dice for or against a particular result.”¹⁴ In contrast, William N. Eskridge and Philip P. Frickey have defended them as part of an “interpretive regime” serving rule of law and coordination functions.¹⁵ That is, substantive canons can act as gap-filling devices that provide clarity for the law and allow courts to signal policy preferences to legislatures, who may draft around such preferences when desired.¹⁶ Eskridge and Frickey further defend them as “a way for ‘public values’ drawn from the Constitution, federal statutes, and the common law to play an important role in statutory interpretation.”¹⁷

It is not my intention here to provide a general defense of substantive canons. Instead, I argue that the if any substantive canons are going to be used by the courts—and Anglo-American courts have accepted some substantive canons as legitimate for at least 400 years—the Democracy Canon should be.

The Canon serves two important purposes. First, as with some other substantive canons, the Democracy Canon can help protect an underenforced constitutional norm. In this case, the Canon protects constitutional equal protection rights in voting, rights which courts for various reasons have declined to protect directly through constitutional litigation. Second, the Democracy Canon is a preference-eliciting mechanism. A clear statement rule requires the Legislature to take affirmative steps to express its intent to limit voter enfranchisement only when justified by other important interests.

 
III.
Dangers of Actual or Perceived Politicization

Despite its pedigree, controversy has surrounded the Democracy Canon, or at least surrounded the results of the Canon’s application in some recent high-profile election law cases, such as New Jersey Democratic Party v. Samson.¹⁸ In Samson, a unanimous New Jersey Supreme Court relied on the Democracy Canon to allow Democrats to replace the name of U.S. Senator Robert Torricelli on the general election ballot shortly before he was up for reelection to the Senate. The relevant New Jersey statute contained rules for the party to replace withdrawn candidates on the ballot when the withdrawal occurred at least fifty-one days before the election, but Torricelli, facing an ethics scandal, withdrew in fewer than fifty-one days.

Though Samson was controversial as an election law case, it is less controversial when viewed as a legislation case applying a substantive canon. Substantive canons are employed regularly as a tool of statutory interpretation, but the public does not generally pay attention to, much less understand, the prevalence of their use. In the context of a hot-button election law case, a court’s use of a substantive canon may appear illegitimate and result-oriented. Moreover, because of the political stakes, judges may subconsciously rely on the Canon in ways consistent with their political preferences. State legislatures, through the passing of clear rules, are best positioned ex ante to avoid judicial overreaching. Importantly, despite the controversy over the Samson case, the New Jersey legislature has not amended its law to impose a stricter reading of statutory election deadlines.

 
IV.
Federal Courts, State Courts, and the Canon

In Palm Beach County Canvassing Board v. Harris,¹⁹ the Florida Supreme Court relied on the Democracy Canon to, among other things, extend the time for a manual recount of votes during the election protest brought by Al Gore against George W. Bush in the Florida 2000 presidential election. Bush appealed the decision to extend the time for the protest to the United States Supreme Court, which remanded the case for further proceedings to determine whether the Florida court’s reliance on the Canon, embodied in the Florida Constitution, violated Article II of the United States Constitution. The issue reemerged in Bush v. Gore,²⁰ when three concurring Justices determined that the Florida Supreme Court’s interpretation of Florida election statutes in light of the Democracy Canon “impermissibly distorted [the statutes] beyond what a fair reading required, in violation of Article II.”²¹

When a state court construes a state statute to a question in a federal election (as in Samson or Palm Beach County Canvassing Board) it runs the risk of violating either Article II of the U.S. Constitution (vesting in each state legislature the power to set the rules for choosing presidential electors) or Article I, section 4 (vesting in each state legislature the power to set the rules for choosing members of Congress, at least to the extent Congress has not set such rules). In Palm Beach County Canvassing Board, the court left open the issue whether broad interpretations of state statutes involving presidential elections could violate Article II, a point embraced by three concurring Justices in Bush v. Gore. The concurring Justices relied upon their own narrow views of proper interpretation to see a constitutional problem.

Contrary to the position of the Bush v. Gore concurring Justices, use of the Democracy Canon to construe state statutes dealing with presidential or congressional elections does not violate Article II or Article I, section 4. Instead, the long-standing nature of the Democracy Canon and the values it supports give state courts ample authority to construe state election statutes covering federal elections in light of the Canon. Only when a state court relies upon the Canon in a way that counters longstanding practice should a federal court consider intervening in a state court election case on constitutional (likely due process) grounds. For the most part, concerns about overreaching should be addressed ex ante by the legislature: a state legislature concerned about state court application of the Democracy Canon in the context of federal elections can use clear statements to negate its application, as the Samson court illustrated in its opinion.

Acknowledgments:

Copyright © 2010 Stanford Law Review.

Richard L. Hasen is the William H. Hannon Distinguished Professor of Law at Loyola Law School, Los Angeles.

This Legal Workshop Editorial is based on the following Article: Richard L. Hasen, The Democracy Canon, 62 STAN. L. REV. 69 (2009).

Moreover, some scholars hoped the Florida controversy culminating in Bush v. Gore would make lemonade indirectly as well: the attention generated by the Florida debacle—particularly attention directed to problems with election machinery and partisan discretion over the counting of votes—would spur state and federal legislative action to fix the problems. One especially important problem is the conflict of interest created when partisan election officials oversee elections in which their party, or even they personally, has a stake in the outcome.

Now, a little over six years later, Bush v. Gore is dead. The death did not come in the usual way that Supreme Court cases die, through outright or sub silentio overruling in a later case. Indeed, no Court opinion—majority, concurrence, or dissent—has cited the opinion since it was decided. But election law developments in the relatively short time since Bush v. Gore show that conservative federal circuit court judges so far have been able to resist the “lemonadization” of Bush v. Gore. Worse, the Supreme Court’s recent opinion in Purcell v. Gonzalez,² allowing Arizona to implement (at least temporarily) its controversial voter identification law, shows that the Court itself has not understood the problems it caused with its Bush v. Gore opinion. The Court’s decision to quickly issue an opinion in Purcell, the casual empiricism of its unanimous opinion, and its discouragement of pre-election litigation all are exceedingly troublesome.

By stating Bush v. Gore is “dead,” I am not making the claim that the Supreme Court will never rely on the case as precedent in an election administration dispute. I mean instead that the promise of election reform inspired by the case is now dead. Indeed, a case could come along some day reviving Bush v. Gore as precedent. Perhaps it is better to think of the case as dormant as a constitutional precedent. My main point is that we should abandon any hope created by the case that the judiciary would serve as an engine of election administration reform.

Bush v. Gore’s failure has been not just a failure in the courts. Legislative fixes to problems of election administration have fared no better, except in the area of voting technology. The good news is that changes in voting technology, subsidized by the federal government, mean that many fewer votes are now “lost” due to inadequate vote counting machinery. But the rest of the news is bad. States have not learned what is arguably the primary lesson of Bush v. Gore: partisan officials should not run elections because of the obvious self-interest problem. Indeed, election administration has become more, rather than less, politicized. State legislatures have not searched for an honest broker to design and implement fair and impartial electoral rules. Many Democrats appear concerned only about problems of voter “access,” while many Republicans appear to care only about voter fraud or “ballot integrity.” This divide has played out in a number of areas, most importantly in the enactment by state legislatures of voter identification laws supported almost exclusively by Republicans and opposed almost exclusively by Democrats.

Unfortunately, the story is even worse than this description. Bush v. Gore’s main legacy has been to increase the amount of election-related litigation. As election law has become a political strategy, it threatens to further undermine public confidence in the electoral process. No lemonade, only lemons.

The death of Bush v. Gore was not unexpected, but its early demise is still something to be mourned.

I.
Bush v. Gore’s Failure to Ignite Election Administration Reform Through Litigation

Briefly, the Supreme Court in Bush v. Gore voted 5-4 to end the recount ordered by the Florida Supreme Court in the 2000 election contest brought by Al Gore to overturn a narrow victory in Florida by George W. Bush. Seven of the Justices on the Court saw equal protection problems with the Florida Supreme Court’s order mandating a statewide manual recount of “undervotes,” that is, ballots which were classified by vote counting machines as not including any vote for President. The U.S. Supreme Court flagged a number of problems with the Florida court-ordered recount, perhaps most importantly the lack of uniform standards for judging when a ballot classified as an undervote by a vote counting machine should be counted in a manual recount as a valid ballot for one of the candidates (the “dimpled chad” problem).

Two of the seven Justices recognizing equal protection issues, Justices Breyer and Souter, would have remanded the case back to the Florida courts for a recount using uniform standards. The remaining five voted to end the recount, thereby awarding Florida’s electoral votes and the presidency to George W. Bush, on grounds that prolonging the counting would deprive Florida of the chance to have its electoral votes counted without challenge in Congress under the Electoral Count Act. The Court held that the recount standards, through “arbitrary and disparate treatment,” impermissibly “value[d] one person’s vote over that of another.”³ The Court limited its holding, however, with some important language: “Our consideration is limited to the present circumstances, for the problem of equal protection in election processes generally presents many complexities.”⁴

Much ink has been spilled on the question whether the Court’s equal protection rationale was a logical extension of or a break from existing precedent. The subtext of this debate, of course, was whether the Court was consciously or subconsciously making a political decision (with conservatives on the Court backing the legal theories benefiting the Republican candidate and liberals on the Court backing the theories benefiting the Democratic candidate) as opposed to a legal one. Part of that debate too concerned whether Bush v. Gore itself would serve as valid precedent to bring greater equality to the administration of elections, a debate that continues to this day.

Some scholars predicted that the Court would eventually endorse the use of Bush v. Gore as precedent to bring greater equality to the nuts and bolts of election administration. Thus, in a New York Times op-ed written just two days after the Court’s decision, Sam Issacharoff wrote that the Court has “asserted a new constitutional requirement: to avoid disparate and unfair treatment of voters. And this obligation obviously cannot be limited to the recount process alone. . . . The court’s new standard may create a more robust constitutional examination of voting practices.”⁵ Steve Mulroy expanded on this point in a law review article, asserting that, while liberals may have been disappointed with the result in Bush v. Gore, its broadly written equal protection holding meant it was possible to make “lemonade from lemons.”⁶

My own view in 2001 was far less sanguine but not quite as dire as those reading the case as a “one-day-only ticket.”⁷ Though I believed that the Court would ultimately limit Bush v. Gore to its facts, I also thought that “[l]ower courts will first apply Bush v. Gore as precedent to cases coming before [them] . . . [s]o there is at least a window of time in which the case may serve as valid precedent.”⁸ I further thought that lower courts would view Bush v. Gore in Rashomonic fashion, with liberal judges embracing a more expansive equal protection reading of Bush v. Gore and conservatives embracing a more restrictive reading of the case—returning liberal judges to the more familiar position of pushing for an expansion of voting rights through equal protection and conservative judges resisting such expansion. In that window of time, I believed that public interest litigants appearing before sympathetic judges could use the logic of the case to make conditions fairer for voters who, because of intentional election administrator choice or mismanagement, would be much less likely to be able to cast a valid vote than other voters in the jurisdiction.

Indeed, at first some lower courts played this enabling role by reading Bush v. Gore to require greater equality in the administration of elections.⁹ Sometimes the threat of litigation was enough: to avoid a probable adverse judgment, the California Secretary of State settled litigation brought by Common Cause to bar the use of punch card voting machines.¹⁰ But that initial success has fizzled, at least as evidenced by the “punch card” cases. As these cases reached the en banc process in circuit courts, conservative judges have blocked Bush v. Gore’s lemonadization.¹¹

II.
The Rise of Partisan Election Administration Laws and the Troubling Public Confidence Gap

Even without cajoling from the courts, some states took it upon themselves to upgrade their vote-casting and -counting technology. Congress also helped matters along by providing funding for states through the Help America Vote Act (HAVA) to phase out antiquated and unreliable technology such as punch card machines.¹² But outside the area of voting technology, most legislative bodies have done too little to fix problems or, worse, have made changes in their election laws aimed at assuring partisan advantage.

Legislative movement outside the area of technology has been mostly in the wrong direction.

First, states have shown very little interest in fixing ambiguities and gaps in their election codes, despite the fact that Bush v. Gore amply demonstrated that such gaps can lead to great controversy as courts are called upon to interpret the elections code in high stakes litigation over who should be declared the winner of a contested election.

Second, legislation that has been passed has proven ineffective. The U.S. Election Assistance Commission (EAC), formed by Congress as part of HAVA to fix the problems made apparent by the 2000 Florida debacle, has so far proven ineffective and now appears in danger of becoming a new site for partisan stalemate over election reform. Even given the EAC’s ineffectiveness thus far, the National Association of Secretaries of State (NASS), the main body of (mostly partisan and elected) state chief elections officers, has not backed off its resolution calling for the EAC to be disbanded. What little good the EAC can accomplish is being undermined by state officials’ need to protect their turf and by lack of funding from Congress.

More important than this failure to act effectively in the face of obvious problems, however, is the increased partisan divide over election reform legislation. It now is beyond question that there is such a divide, with many Republicans expressing concern about voter “integrity” and the possibility of voter fraud affecting the outcome of elections, and many Democrats expressing concern over voter “access” and the possibility that the government or others will take steps to suppress the votes of the poor, minorities, and others. Some Democrats suspect that the Republican integrity claims are false and are intended to suppress the vote. Some Republicans suspect that Democrats’ concern about access is overblown and is intended to create the conditions where ineligible voters (such as felons or illegal immigrants) are allowed to vote. It is no exaggeration to say that “election reform,” in the sense of making it easier for people to cast a vote, without intimidation, that will be accurately counted, has become an issue for Democrats and the liberal reform community. Republicans, in contrast, have focused their attention on voter fraud. There has been little movement for bipartisan cooperation. The most prominent attempt at such cooperation, the Baker-Carter Commission, got mired in controversy concerning its endorsement, over the dissent of three Democratic members of the commission, of a voter identification card.

The country’s partisan divide has manifested itself in a number of ways. For example, Democrats have pushed for election-day voter registration as a means of making it easier for eligible voters to vote, but Republicans have opposed this change on grounds that it would allow for more voter fraud. The partisan divide has been on display most heatedly in the dispute over state voter identification laws. With the exception of Arizona, which enacted its voter identification law through a voter initiative (aimed more broadly at issues of benefits for illegal immigrants), every state that has enacted or tightened its requirements for voters to show identification at the polls has done so through the support of a Republican-dominated legislature. Democrats have uniformly opposed the efforts to impose new voter identification requirements, as in Pennsylvania, where the Democratic governor vetoed a new voter identification law passed by the Republican-dominated legislature, and in Missouri, where the newly elected Democratic Secretary of State has opposed voter identification laws and argued against them in a report on the 2006 election.

Republicans have defended voter identification laws as necessary to combat voter fraud. But Democrats and civil rights organizations see those voters who are more likely to vote Democratic, have a more difficult time securing voter identification. Poor people tend to drive less (meaning they are less likely to have a driver’s license, which is the most common form of identification), and they may not have the money to secure certified copies of documents, such as birth certificates, necessary to obtain a state-issued voter identification.

The debate over voter identification has generated a great deal of heat, but very little light. At its heart are two separate empirical questions: (1) How much impersonation fraud (where one person shows up at the polls claiming to be someone else) takes place that an identification card would detect or deter? (2) How much would a voter identification law deter eligible voters from voting? There is in fact very little evidence of impersonation fraud (as opposed to absentee ballot fraud) and we have very little understanding of how much voting a strict voter identification law may deter.

There is also suspicion, which definitely deserves further research, that the imposition of voter identification procedures and other similar rules is causing a decline in voter confidence among minority voters. By 2004, it was clear that there was a growing party and racial divide in public confidence in the electoral process. By 2004, 21.5% of Democrats thought the means of conducting the most recent presidential election was “somewhat unfair” or “very unfair,” compared to 2.9% of Republicans. In that same election only one-third of African-Americans called the vote “accurate and fair.”

No doubt, some of the disparity is due to the fact that Democrats were on the losing end of the 2000 and 2004 presidential elections, and winners tend to have more confidence in election outcomes than losers. But the gap in confidence between winners and losers may not explain everything. Just before the 2006 election, when it already appeared that Democrats were likely to do well in the midterm congressional elections, the gap between Republican and Democratic views persisted. Even more troubling, the percentage of African-American voters who were “not too confident” or “not at all confident” that their votes would be fairly counted nearly doubled from 15% in 2004 to 29% in 2006.

When it comes to election reform, many state legislatures have spent more energy debating issues like voter identification laws than considering how to take partisanship out of the process of deciding elections. In 2005, I counted thirty-three state chief elections officers chosen in partisan elections. That number remains the same in 2007. Congress, meanwhile, has not fully funded the electoral reforms it authorized in HAVA, and current proposals for election reform on the federal level are getting mired in presidential politics.

III.
Mislearning the Lessons of Bush v. Gore: The Supreme Court’s Poor Response in Purcell v. Gonzalez to the Rise of Election Law as a Political Strategy

Election litigation has more than doubled since 2000, and it continues to grow. In an earlier study, I showed a large increase in the amount of election law litigation in the courts, from an average of 96 “election challenge” cases per year in the 1996-99 period, to an average of 254 per year in the 2001-04 period.¹³ The high rates continued through 2006.

The rise in election law litigation since 2000 is part of a trend of election law as a political strategy. There seems little question that candidates, parties, and others have lost any inhibition to resort to litigation in the case of a close election. So it is very important for the Supreme Court and lower courts to send appropriate signals about how and when courts should entertain such challenges. Unfortunately, the Court’s first post-Bush v. Gore opinion on an election administration subject, Purcell v. Gonzalez, sends some very troubling signals, apparent from the Court’s decision to issue its opinion quickly, the casual empiricism of its unanimous opinion, and its discouragement of pre-election litigation.

Arizona voters adopted a new voter-identification law in 2004 as part of Proposition 200—a measure aimed primarily at the problem of illegal immigration. Among other provisions (including one that voters provide satisfactory evidence of citizenship at the time of voter registration), the law requires those who vote in person to produce either a photo identification, or two other pieces of identification, showing the voter’s name and address. A coalition of voting rights organizations filed a complaint alleging that the law violated federal election laws and the U.S. Constitution. A federal district court, without providing any reasoning, denied their request to delay implementing the law pending a full trial on the issues in the case. A Ninth Circuit motions panel, also without providing any reasoning, reversed that decision, temporarily halting the voter-identification requirements (as well as its voter-registration requirements). The trial court then belatedly issued its statement of reasons for denying the order.

The State of Arizona then asked Justice Kennedy, who has jurisdiction over emergency appeals from the Ninth Circuit, to stay the Ninth Circuit’s decision. Justice Kennedy referred the question to the entire Court. The Court treated the stay request as a petition for certiorari, granted it, and then reversed the Ninth Circuit in a surprise per curiam opinion on a late Friday afternoon soon before the election, allowing the identification requirements to be put into effect pending a full trial on the merits in the case. The Supreme Court described constitutional interests on both sides of the voter-identification question:

Confidence in the integrity of our electoral processes is essential to the functioning of our participatory democracy. Voter fraud drives honest citizens out of the democratic process and breeds distrust of our government. Voters who fear their legitimate votes will be outweighed by fraudulent ones will feel disenfranchised. . . . Countering the State’s compelling interest in preventing voter fraud is the plaintiffs’ strong interest in exercising the “fundamental political right” to vote. Although the likely effects of Proposition 200 are much debated, the possibility that qualified voters might be turned away from the polls would caution any district judge to give careful consideration to the plaintiffs’ challenges.¹⁴

The Court also signaled its disfavor with last-minute court interventions in the electoral process: “Court orders affecting elections, especially conflicting orders, can themselves result in voter confusion and consequent incentive to remain away from the polls. As an election draws closer, that risk will increase.”¹⁵ The Court seemed especially concerned that the Ninth Circuit gave no reasons for its decision to reverse the district court and stay implementation of Arizona’s voter identification law.

The Court’s decision to reverse the Ninth Circuit is defensible, given the circuit court’s failure to provide any reason for not deferring to a lower court’s decision not to issue the preliminary injunction (though the district court’s failure to timely issue findings of fact and conclusions of law made the Ninth Circuit’s review difficult as well). Justice Kennedy, or the Court, arguably should have granted the stay of the Ninth Circuit’s order, restoring the district court’s opinion to let the election go forward with the identification requirement, pending a trial on the merits.

But in three ways, the Court’s treatment of the Purcell case shows that the Court has not internalized the lessons from Bush v. Gore and has actually made things worse. First, given the opinion’s sloppy reasoning (as detailed below), it is quite plausible that the opinion was rushed. Likely the Court decided to rush the opinion out to send a message to lower courts that they should be less willing to entertain litigation that makes last minute changes in election laws (an issue I return to in the third point below). The Court had two alternatives to issuing the opinion when and how it did. First, Justice Kennedy (or the Court) could have reversed the Ninth Circuit without issuing an opinion. Second, perhaps fearing that this vote would be seen as too political before the upcoming election (or perhaps not, if it were unanimous), the Court could have issued the order, followed by an opinion issued later, after the Court had time to give it more careful attention. Some lower courts have followed this strategy in election law cases.

Second, my most serious objection to Purcell is the Court’s adoption of a wholly unsupported empirical assumption about the effects of possible voter fraud on turnout and how that unsupported assumption must now be balanced in cases challenging the right to vote. As noted in Part II, there are two basic empirical questions for the courts to sort out in these cases. First, is there enough evidence of impersonation fraud to justify such laws, which no doubt place some burden on the right to vote? Second, how onerous are such laws? The empirical evidence on both fronts is still rather scant, but it is being developed.

Once the Court was going to issue an opinion, it should have said that courts must engage in some kind of balancing (whether under strict scrutiny or a less burdensome standard) of the state’s interest in preventing voter fraud with the rights of voters lacking identification (and lacking ease of obtaining such identification) required by the state in order to vote. The Court even could have said that in the absence of any evidence of either voter impersonation fraud or voter deterrence due to a voter identification requirement, a court is within its discretion in allowing the state law to go forward until there is proof the law is deterring voters from voting. That would be the time for an appropriate challenge.

But the Court did not do so. Instead, the Court stated, without any proof whatsoever, that concerns about voter fraud “drive[] honest citizens out of the democratic process and breed[] distrust of our government.”¹⁶ The Court further said that the fundamental right to vote of voters lacking identification must be weighed against the interest of those supposedly disillusioned voters who “will feel disenfranchised.”

The discussion is troubling on a few levels. First, the Court cited no evidence, and I am aware of absolutely no evidence, supporting the view that voters are deterred from voting out of fear that their legitimately cast votes will be diluted by the votes of those committing voter fraud. Moreover, the Court offered no explanation why it is appropriate to balance feelings of disenfranchisement against actual disenfranchisement, whatever the appropriate standard of review.

The final problem with the Purcell opinion is its discouragement of preelection litigation. The Court on this point has it backwards. When courts get involved in election disputes, they run a risk of undermining the public’s faith in the electoral process and in the fairness of the courts. To minimize that problem, it makes sense to encourage litigation well before elections (that is to say, before the winner is known and everyone will question the biases of the judges) and to discourage litigation after the election whenever a suit might have been brought earlier. The risk of confusion as the election approaches should be balanced against the risk of disenfranchisement or other loss of rights that cannot be fixed after an election. If a voter identification law is indeed disenfranchising, there is likely no effective post-election remedy to restore the right to vote. Indeed, in a close election, an unconstitutional law could make a difference in the outcome. If, as Democrats claim, voter identification laws fall more heavily on their supporters, such laws could tip the balance in favor of Republicans. Purcell may have the ultimate effect of moderately reducing the total amount of litigation, which would look like a modest improvement over the status quo. However, Purcell achieves that result only by eliminating cases for which the only viable remedy may come through pre-election judicial review.

IV.
Conclusion

The 2000 election debacle brought well-deserved attention to the serious problems plaguing our election system. Bush v. Gore divided the country at that time, but many hoped that the case and surrounding controversy could spur meaningful reforms to fix serious election administration problems. The case and the Florida controversy have not directly led to meaningful election reform. States and legislatures, aided by congressional funding, have taken steps to improve voting technology, though the transitions to new technologies and election administration incompetence have colored what would otherwise be success on the technology side.

But on the side of reforming elections to create fairer and nonpartisan rules, it appears the country has learned the wrong lessons from the Florida debacle and Bush v. Gore. Election administration has become more, not less, partisan. Public confidence in election administration, especially among African-Americans, is at troubling and embarrassingly low levels. Elections more frequently result in litigation than before 2000. And the courts, especially the Supreme Court, have not been careful in addressing election administration claims. The result is likely to be further contentiousness and growing voter distrust of the system by which we cast and count votes for the foreseeable future.

Now more than ever, the country needs to learn the right lessons from Bush v. Gore. But it appears Bush v. Gore’s moment has passed from public consciousness, replaced instead by partisan recriminations and retrenchment.

Acknowledgments:

Copyright © 2009 Stanford Law Review.

Richard L. Hasen is the William H. Hannon Distinguished Professor of Law at Loyola Law School, Los Angeles.

This Editorial is based on the following full-length Article:   Richard L. Hasen, The Untimely Death of Bush v. Gore, 60 STAN. L. REV. 1 (2007).

Note: The Article was published before the United States Supreme Court decided Crawford v. Marion County Election Board, 128 S.Ct. 1610 (2008), upholding the constitutionality of Indiana’s voter identification law.