It is important to distinguish “extra innings” as an ordinary part of a well-functioning electoral process from true calamities that deserve to be labeled as such.  To this end, I offer this historical perspective.

I share the view of Doug Chapin, among others, that the media and the public should not be unduly alarmist about the prospect of the presidential election being “too close to call” on Election Night and therefore going into overtime. As I explained in a previous comment, given the “new normal” of provisional ballots as required by Congress in 2002, it is to entirely appropriate (and thus to be anticipated) that a closely competitive election might need to wait for provisional ballots to be verified and counted. Therefore, neither the term “nightmare” nor some similar label should be attached to such circumstance, without more. If one wants to speculate about a genuine nightmare, then one can consider the possible need of a state, or locality, to reschedule Election Day because of “Frankenstorm”—or potential Equal Protection issues arising from the differential treatment of voters within a state in response to a storm, a scenario we “wargamed” in 2008.

In short, it is important to distinguish “extra innings” as an ordinary part of a well-functioning electoral process from true calamities that deserve to be labeled as such. In this regard, I have written an introduction to “elections in overtime” that is a chapter of a newly published book, LAW AND ELECTION POLITICS: THE RULES OF THE GAME (edited by Matthew Streb).

Meanwhile, as part of the effort to explain why “extra innings” in a presidential election is not necessarily a crisis, I offer this historical perspective.

You don’t remember that it took two weeks to decide the presidential elections of 1884 and 1916? Don’t worry, that’s a good thing.

The point is that a presidential election can remain unresolved for a couple of weeks without a crisis ensuing. The Hayes-Tilden election of 1876 and Bush-Gore election of 2000 ended up in severely protracted disputes lasting much longer than two weeks, with the Hayes-Tilden dispute almost becoming a genuinely disastrous constitutional crisis. But both 1884 and 1916 show that a presidential election can be “too close to call” the morning after Election Day and still be resolved without litigation, pursuant to administrative procedures for canvassing returns, and thus not spill out of control into potentially dangerous confrontation.

In 1884, New York was the “tipping point state” that would determine the Electoral College winner, and initial returns showed Republican candidate James Blaine trailing the Democrat, Grover Cleveland, by about only a thousand votes or so in that state (where Cleveland happened to be the incumbent governor at the time). Both sides had battalions of lawyers scrutinizing the New York returns, with Blaine’s supporters saying that they would go to court to overturn a Cleveland victory if they found that fraud had resulted in a stolen election. (Republicans were specifically concerned that year that Democrats might have transported voters from out-of-state to cast unlawful ballots in New York.) Remember, 1884 was only eight years after Hayes-Tilden, so lawyers for both sides knew how to wage all-out electoral war, just as lawyers today have vivid memories of 2000.

But all-out electoral war did not break out in 1884. Instead, the canvass occurred over a two-week period after Election Day, with each side having two witnesses to observe the process in each electoral district in New York. Sure enough, this process uncovered vote-tabulation errors, some favoring Cleveland and others favoring Blaine. One problem that surfaced on Long Island caused the most consternation: it was alleged that votes for one candidate were wrongly attributed to another. But even if that allegation were true—and it was never proved—it would not have been enough to overturn Cleveland’s victory. At the end of two weeks, when the canvass was complete, Republicans conceded that Blaine had lost the election fair and square. In the works of the New York Tribune, which functioned as something of an official mouthpiece of the Republicans at the time, “the canvass of the returns has been thorough, careful and honest, and leaves no room for doubt as to the result.”

Thus, 1884 was nothing like 1876, and almost no one now is aware that it took two weeks to settle whether Cleveland or Blaine had won. As soon as the vote-counting process was complete pursuant to the necessary administrative canvass, without litigation or other form of a protracted dispute, the nation treated the result as entirely legitimate and appropriate, and history has viewed the 1884 as decided without a dispute just as if a winner had been declared the morning after Election Day. 1884, therefore, demonstrates that a presidential election can go into “overtime” without a significant problem developing.

1916 illustrates the same important point. That year Charles Evans Hughes as the Republican nominee had stepped down from the Supreme Court to challenge the Democratic incumbent, Woodrow Wilson. (Hughes would later be reappointed as Chief Justice.) The initial returns on November 7 showed several states too close to call, including California, which ultimately became the decisive one. The result in New Hampshire actually flipped during the time the nation was waiting to hear from California, and Minnesota also was potentially in play depending on what happened in California.

Amidst initial talk among Republicans of taking the election all the way to Congress (where they controlled a majority of the incoming state delegations in the House of Representatives, which is what matters under the Twelfth Amendment), based on allegations of fraud and vote-buying, Hughes decided to wait for the completion of the official canvass of returns in California before making any concession. Wilson was ahead in the state by just a few thousand votes in the initial returns, with remote localities not yet reporting. As in 1884, the canvass was scrutinized closely by attorneys from both sides and turned up errors. But the errors affected both sides in roughly equal proportions and appeared random. No evidence of fraud turned up, despite the Republicans’ efforts to look for it, at least not enough to overturn Wilson’s 3,420-vote margin at the end of the canvass. After California officially certified this result, Hughes conceded on November 22, roughly two weeks after Election Day.

No one today thinks of Wilson’s reelection in 1916 as in any way tainted by the fact that it took two weeks for his victory to be settled. Unlike Bush’s win in 2000, Wilson’s in 1916 was as good as if it had been declared on the morning after Election Day. Waiting for completion of the ordinary administration of the canvass, which leads to an officially certified result, is worlds apart from going to court to contest that certification, as Gore did in 2000.

It is important, too, to remember that Europe was already fighting the Great War in 1916, with the very real fear that the United States would soon enter that war as a central concern during that presidential campaign. Indeed, the national security threat was serious enough that Wilson had made extraordinary contingency plans to let Hughes assume the office of the presidency immediately, if the result of the election showed Hughes the victor. Yet even in that precarious situation, the nation was able to handle two weeks of uncertainty as to which candidate actually won the election.

If the nation could wait two weeks in 1916, it can also wait that long this year to learn whether Obama or Romney wins. Just as completing California’s canvass did nothing to undermine the legitimacy of Wilson’s eventual victory, so too the mere fact that it might be necessary to complete the canvass in Ohio (or elsewhere) this year before knowing which candidate wins would in no way delegitimize the victory of the eventual winner.

To be more specific, suppose it takes two weeks for Ohio to decide whether Obama or Romney wins the state and, with it, the whole election. Suppose that two-week process is orderly, with attorneys for both candidates closely scrutinizing all aspects of the returns, including the validity of provisional ballots, and no major disputes or lawsuits develop despite all this scrutiny. Suppose, then, after two weeks or so, there is an official certification of the result, with an immediate and gracious concession from the losing candidate according to that certification.

I submit that in this situation history would judge 2012 to be essentially like 2004, and 1916, not like 2000. In other words, a concession at the end of two weeks according to this scenario would be remembered by posterity as essentially equivalent to Kerry’s concession in 2004 (which occurred the morning after Election Day), or Hughes’s concession in 1916 (which, like the hypothetical situation this year, occurred after two weeks). It would not be considered equivalent to Gore’s concession in 2000, after the Supreme Court in December ruled 5-4 to stop recounting ballots, as Gore had so fervently sought through litigation even after Bush’s officially certified victory.

2000, like 1876, was a genuinely and deeply disputed presidential election, in a way that neither 1916 nor 1884 had been. This distinction is worth keeping in mind, as we prepare for the possibility that this year’s presidential election may be unresolved on the morning after Election Day.

A contemplation of the possibility that this year’s presidential election extends beyond November 6, and how 2012 might differ from 2000 in this regard.

It is hardly unrealistic to think that on Wednesday morning, November 7, Americans may wake up with the news media reporting that Ohio alone remains “too close to call” and that whichever candidate wins Ohio will cross the finish line of 270 or more Electoral College votes.

Therefore, it is useful to consider what might happen next—how events may unfold after the ballots are cast, so that Americans may judge whether or not the process works properly and leads to an outcome they can accept as correct or at least legitimate.

My crystal ball cannot predict exactly what will happen—there are too many variables and uncertainties—but it helps to begin by imagining one plausible scenario. Other possible scenarios might be considered as well, but this one illustrates the kinds of considerations that may come into play if neither candidate wins a decisive victory Election Night.

Suppose that Romney is ahead in Ohio by about 10,000 votes with 100% of precincts reporting, but there are about 150,000 provisional to review (a conservative estimate given the approximately 200,000 provisional ballots cast in Ohio in 2008). There might be pressure on Obama to concede, especially if Romney is also ahead in the national popular vote. One difference between this scenario and 2000 is that Gore was ahead in the national popular vote while he was fighting to come from behind in Florida. If winning the national popular vote is a kind of “moral victory” even though it is legally irrelevant, it can give a candidate a plausible basis for claiming that he owes it to the electorate to fight on in the still-undetermined “tipping point state” to see if he can achieve the legally controlling Electoral College victory as well as his national popular vote vindication. It is possible, however, that on November 7 Obama—unlike Gore—will be behind both in the national popular vote and the initially reported precinct returns in Ohio.

Still, the way we elect presidents is through the Electoral College, and in this hypothetical scenario Obama could see an Electoral College victory within reach even while being behind in Ohio by 10,000 votes with 100% of precincts reporting. In 2008, Ohio counted about 75% of the roughly 200,000 provisional ballots cast that year. Even if that rate drops to two-thirds this year, it would mean that 100,000 of the 150,000 hypothetically cast would be counted. If Obama “wins” those provisional ballots at a rate of 55% to 45%—a not implausible possibility, if many of those provisional ballots are cast in urban or college precincts (as historically has been true)—then he could see himself erasing a 10,000-vote deficit once the provisional ballots are counted.

Now, I’m not a huge fan of the Electoral College. I find powerful the arguments in favor of replacing the Electoral College with some sort of system that makes the winner the candidate with the most popular votes nationwide. But that would be a reform for the future. As long as we continue to use the Electoral College as our constitutional mechanism for choosing a president, then it seems to me that Obama would be justified in insisting that the nation wait until the process of reviewing provisional ballots in Ohio is complete before either candidate is in a position to make a concession speech. This point, moreover, seems valid even if Obama is behind in the national popular vote while the process of verifying and counting Ohio’s provisional ballots remains underway.

I make this point not to advocate on behalf of any particular campaign or candidate, but as an observer attempting to anticipate before November 6 what might occur afterwards. The same point would apply if on the morning of November 7 Obama is narrowly ahead in Ohio based on initial precinct returns and also leads in the national popular vote. Romney might have a reasonable chance of overcoming Obama’s narrow lead in Ohio based on the possibility that absentee ballots, if mailed on time under Ohio law, may still arrive up to November 16 and still be counted. In this alternative situation, Romney would also be justified in holding out hope of an Electoral College victory and thus waiting until November 16—after all the eligible absentee ballots have arrived and been counted—before deciding whether he or his opponent is the one who needs to deliver a concession speech. (To be technical, under Ohio law the wait for a final count that includes all eligible absentee ballots, as well as verified provisional ballots, would last at least one more day, until November 17, since local boards of elections are not permitted to begin the official canvass of returns before then, one day after the deadline for absentee ballots to arrive.)

There is another point to make about provisional ballots, one which distinguishes this year from 2000. By design, provisional ballots take extra time to count. They are not supposed to be counted on Election Night. Instead, their provisional nature means they need to be evaluated. Indeed, depending on the reason why they were cast as provisional rather than regular ballots, the voters who cast these ballots have a window of opportunity to supply information that shows that their votes are eligible to be counted. For example, if the reason that a particular voter cast a provisional ballot was that this voter did not come to the polling place with the required form of identification, Ohio law gives the voter 10 days—until Friday, November 16—to produce a valid form of ID (which could be a bank statement or utility bill and need not be a photo ID).

Thus, an electoral system that includes provisional voting necessarily commits itself to the possibility of needing to wait until the process of evaluating all the provisional ballots is over before being able to announce the winner of the election. Valid provisional ballots, once verified, count just the same as regular ballots counted on Election Night. Consequently, when a close election potentially turns on the provisional ballots that have yet to be evaluated, the system demands patience sufficient to wait for the completion of the process for evaluating those provisional ballots. It would contradict the basic democratic principle of one-person-one-vote (which applies to Ohio’s electorate in voting for presidential electors even though it does not apply to the Electoral College as a whole) to short-circuit that process and declare a winner before the potentially outcome-determinative provisional ballots have been evaluated and counted.

In this significant respect, the current electoral system is different from the one that the nation used in 2000. Indeed, provisional ballots were mandated by Congress after 2000 as part of the response to the problem of erroneous voter purges that occurred in Florida that year. Therefore, the protracted litigation over so-called “hanging chads” that went all the way to the U.S. Supreme Court in Bush v. Gore had a very different character than simply waiting for all the provisional ballots to be evaluated, as would be necessary in the hypothetical scenario I am describing here.

The “hanging chads” were regular ballots that had been counted, or at least processed, by Florida’s vote-counting machines on Election Day in 2000. Therefore, when Gore pressed for a recount, he was attempting to get a second tabulation of ballots that already had been tabulated once. Indeed, Florida’s ballots were fed through the machines a second time shortly after Election Day. Therefore, when Gore persisted in seeking a manual recount of those same ballots, he was asking for yet a third tabulation of ballots that had already been tabulated twice. Indeed, many of us can remember James Baker claiming on behalf of Bush that after the machine recount it was time for Gore to concede since the ballots had been counted twice.

By contrast, in the scenario I am imagining, during the ten days between November 6 and November 16, the potentially outcome-determinative provisional ballots would not yet have been counted once. Therefore, if Obama were to insist on waiting for a completion of the provisional ballot review process in Ohio, he would not be seeking a “do over” of the initial count, whereas Gore sought a “do over” throughout the long 35-day fight over the hanging chads in Florida.

I know that Americans will want to know who won the presidential election when they tune in to watch the returns on the night of November 6. Particularly because this campaign has seemed so long and hard fought, many Americans will desperately want the race to be over. The idea of prolonging it yet another 10 days, just to wait for the process of evaluating Ohio’s provisional ballots, would seem like agony. (Indeed, as mentioned previously, Americans would need to wait at least one day more—until November 17—for the certified count of the provisional ballots to be announced, according to Ohio law, as part of the official canvass of returns.)

Still, it is important for Americans to understand that provisional ballots are part of the “new normal” in post-2000 elections, and it will be important for the news media to educate Americans that there is nothing inherently wrong if it is necessary to wait 10 days for the evaluation of provisional ballots to be completed. In fact, exactly the opposite is true: waiting 10 days means that the system is working precisely as it is designed to work; the 10-day delay is a built-in feature of the new system.

What would be wrong would be for the media to use the word “recount” to apply to this 10-day period. We would not be in a “recount” situation—and indeed under Ohio law a recount cannot start until the process of evaluating all the provisional ballots is complete (as the state learned from a close congressional race in 2008). We would simply be waiting for the completion of the initial count of ballots cast by all voters entitled to participate in the election. I would hope that, with the media’s help, the public would understand that this wait, however frustrating, would be necessary in order to make the state’s electoral process qualify as democratic. Otherwise, it would be like declaring a winner in a close race when there were still enough precincts left to report that could affect the outcome, or not waiting to count absentee ballots that arrived on time under the law but too late to be included in Election Night returns.

Two further points are worth considering as we contemplate this hypothetical scenario.

First, it is conceivable that Ohio’s legislature might attempt to short-circuit the process of evaluating the state’s provisional ballots. Both houses of the state’s legislature have Republican majorities, and the state’s governor is a Republican. Therefore, reminiscent of what was contemplated by Florida’s legislature in 2000, in this hypothetical scenario Ohio’s legislature might attempt to invoke its authority under Article Two of the U.S. Constitution to appoint the state’s presidential electors directly. But it seems to me that Ohioans would not tolerate the legislature’s taking the power to appoint the state’s presidential electors away from the voters, at least not during the ten-day period between November 6 and November 16 when not all the state’s valid popular votes have yet been counted. It might be a different situation later in the process, if a fight over provisional ballots ends up in protracted litigation, and the judiciary seems unable to handle the dispute fairly. The argument for the legislature coming in to settle the matter definitively would seem more plausible in that later situation. But when the process of verifying provisional ballots is proceeding normally, in accordance with the design of the system, it would seem entirely undemocratic for the state’s legislature to step in to stop that normal ballot-counting process.

Second, there is a variation on this hypothetical scenario that would increase the pressure on Obama to concede before the provisional ballots have been counted. Suppose that everything about the hypothetical is the same except that for Obama winning Ohio would mean, not reaching 270 Electoral Votes, but only a 269-269 Electoral College tie with Romney. This twist is also not implausible, as anyone who has played around with an Electoral College calculator knows. Assuming that the Republicans continue to hold the House of Representatives with numbers roughly comparable to now (a realistic assumption if the presidential race is close enough to result in an Electoral College tie), then Republicans in the House would be in a position to elect Romney president pursuant to their authority under the Twelfth Amendment of the Constitution. If Romney also holds a lead in the national popular vote, that “moral victory” would give the House Republicans a principled (as distinct from a partisan) reason to choose Romney over Obama. Consequently, if Obama is behind by 10,000 votes in Ohio, and hoping for provisional ballots only to get him to a 269-269 tie, he would lack a compelling reason to fight on. Instead, in this specific context, it would be a magnanimous gesture for Obama to acknowledge that Romney deserves to be elected president by the House based on his national popular vote victory. This gesture, which would permit Romney to get on with the task of a presidential transition—and deal with the looming fiscal cliff in as responsible a manner as feasible—would help secure Obama’s legacy in the history books.

Thus, the possibility of a 269-269 tie shows that not every Election Night uncertainty calls for postponing a concession speech until completion of the official count of all provisional ballots. In 2004, too, Kerry thought about waiting for a count of Ohio’s provisional ballots, but in his particular situation the numbers he faced on the morning after Election Day did not make an Electoral College victory possible. Likewise, Obama may face a situation in which he needs to make a judgment of whether there are enough provisional ballots to overcome a narrow lead that Romney holds in Election Night returns. The math might not work for him any more than it did for Kerry, and he may be sensitive to the potential public sentiment that an incumbent president should not appear to cling too long or too hard to staying in office when the overall message of the national electorate is that it is time for him to leave.

There are also other points concerning the evaluation and counting of provisional ballots that one might contemplate. I, for one, am concerned about litigation that attempts to extend the 10-day deadline under Ohio law for this evaluation process. Extending this deadline through litigation runs the risk of bumping up against the Electoral College calendar, which requires the presidential electors themselves to meet on Monday, December 17 (and which also makes the so-called Safe Harbor Deadline six days earlier: Tuesday, December 11). But I leave this and other issues for another day. (There are also potential issues surrounding the counting of absentee ballots that might arise after Election Day, but I leave those aside as well—apart from the passing analogies I have made to absentee ballots while considering the less familiar topic of provisional ballots.)

Here it is enough to say that the uncertainty caused by the existence of provisional ballots, by itself, does not mean that there is anything wrong with the electoral process. Instead, it is simply a sign that the process is working as intended in a very close election. Therefore, if Ohio would give one of the candidates an Electoral College majority of 270 or more (not the 269-269 tie), and if figuring out which candidate won Ohio requires waiting for the evaluating of provisional ballots, then wait the nation must—as patiently as possible.

As a nation, we may well wish to change the system after we are finished with it this year and, with hindsight, we contemplate its flaws. But once we have cast ballots in accordance with the system, we owe it to the voters whose valid ballots may determine which candidate wins the “tipping point state” that determines an Electoral College majority to actually count these valid ballots. Not doing so would seem to make the electoral system not merely flawed but something of a hoax.

Election Law @ Moritz has started to pull together in one place materials that might be useful for anyone interested in the topic of elections going into overtime.

Steve Huefner has already written an explanation of the American Law Institute project that he and I are working on, including the draft versions of the 5-week and 9-week calendars we are developing for the resolution of disputed elections. (Here’s a more recent, simplified description of the 5-week calendar.)

I’ve written a chapter of the newly published LAW AND ELECTION POLITICS: THE RULES OF THE GAME (edited by Matthew Streb). The chapter is an introduction to the topic of recounts and other issues concerning elections in overtime. Therefore, I hope it is useful to anyone who wants an accessible overview.

In October 2011, I gave a presentation to a “federal bench and bar” conference on the possibility that a vote-counting dispute over ballots cast in Ohio might cause problems for this year’s presidential election. Recently, I gave an updated version of the presentation. For whatever it’s worth, a year ago, I guessed that the chances that this year’s presidential election might become mired in a vote-counting dispute over Ohio’s ballots were only about 1 in 10,000. Now, however, based in part on probabilities indicated by Nate Silver on his FiveThirtyEight blog, I believe that the risk of this scenario has risen to perhaps as a high as in 1 in 100—still low, but definitely higher. This presentation offers suggestions on ways that the federal judiciary might manage any litigation that arises over Ohio’s ballots so that the state remains able to meet the congressional Safe-Harbor Deadline (Tuesday, December 11—exactly five weeks after Election Day, November 6) for resolving all such disputes.

Finally, for those who want to dig much deeper into this topic, I have written a trilogy of pieces for the ELECTION LAW JOURNAL on Minnesota’s disputed U.S. Senate election in 2008, between Democratic challenger, Al Franken (who eventually won), and Republican incumbent, Norm Coleman:

The Lake Wobegone Recount

How Fair Can Be Faster

A Tale of Two Teams

The purpose of this trilogy, in addition to telling the inherently interesting story of how this disputed election was resolved, is to assess what lessons it teaches for the disputed elections, presidential or otherwise, that inevitably will occur sometime in the future.

This comment builds on my earlier one on the need for nonpartisanship in the Supreme Court’s decision in this case.  UPDATE: The Court, without dissent, has denied the State’s request for a stay, thereby achieving the unified posture that this comment was hoping for.

the view the Court has plausible grounds under the existing state of law in this area to either grant or deny the emergency stay application filed by Ohio’s Secretary of State, Jon Husted.

The Court could take the view, argued forcefully in Husted’s reply brief, that first, the district court should not have issued the preliminary injunction in the first place (because the Equal Protection claim on which it is predicated is significantly flawed), and second, it is important for the highest Court in the land to remind lower courts not to interfere with a state’s electoral process shortly before voting is about to begin unless there is a compelling need for such interference (which is there is not when the Equal Protection predicate for such interference is weak).

Conversely, the Supreme Court could reach the opposite conclusion, argued cogently by the Obama campaign: one, the district court’s preliminary injunction (whatever its merits) is not harmful to Ohio, as the only thing it does is permit—but not require—local election boards to offer early voting on November 3, 4 & 5 (and therefore cannot interfere with whatever preparations local officials consider necessary to be ready for Election Day itself), and two, there is no urgency to the Supreme Court’s becoming involved at this time with the merits of the underlying Equal Protection claim, which (even if dubious) has enough plausibility under the Court’s precedents that it does not deserve summary reversal when all four judges who considered the claim below sustained its validity.

Since the Supreme Court reasonably could rule either way on the stay application, one might think it naïve to expect all nine Justices—from Ginsburg to Scalia, who disagree about so much else—to end up taking the same position in this case. Nonetheless, I think it important that they do so, even to the point of Justices declining to publish a dissent if they initially vote in a way that does not prevail among the nine.

Why is unanimity more important in this case than in others that come before the Court? My answer is that partisan litigation over casting and counting of ballots in a presidential election, which pits Democrats and Republicans against each other in seeking the Court’s approval for their view of the voting rules that govern the process for determining which side wins the White House, calls for a visibly nonpartisan response from the Court.

Yes, it may be regrettable if the all the Court’s Republican appointees line up on one side of the pending affirmative action case, with the Democratic appointees on the other side. (Justice Kagan has recused herself in that case, meaning that a thoroughly partisan split there would be 5-3 instead of the typical 5-4.) But affirmative action is not election law and thus the litigation itself is not directly a case of Democrats versus Republicans over the rules of the voting process.

Even the 5-4 split in Citizens United, the campaign finance case, is not as troubling as the same 5-4 split would be in Husted v. Obama for America. To be sure, campaign finance is subset of election law, and Citizens United affects presidential elections as well as others. Thus, arguably it is more disconcerting that the Court’s Republican and Democratic appointees would uniformly reach opposite conclusions in Citizens United than in cases involving non-electoral issues, like affirmative action or abortion.

Still, Citizens United is not a lawsuit between Democrats and Republicans over the rules for casting and counting ballots in a particular presidential election. Citizens United concerns campaign practices, rather than voting itself, which is an important distinction. As important as the funding of campaign advertisements may be to the political climate in which an election held, ultimately it is the ballots themselves that determine the winner. Thus, a Supreme Court decision about the rules for casting or counting those ballots has a direct effect on a presidential election in the way that a decision about campaign finance only has an indirect effect.

The Supreme Court’s decision in Crawford v. Marion County Elections Board, the 2008 case involving Indiana’s voter identification law, comes closer to raising the concerns of a non-unanimous decision in the currently pending Husted v. Obama for America. Crawford was far from unanimous. There, the deeply fractured Court essentially divided 3-3-3 over how to analyze the constitutional challenge to the voter ID law. The Court’s only two Democratic appointees at the time (Ginsburg and Breyer) were among the three dissenting Justices (the other being Souter) who would have invalidated the law in its entirety. The Court’s three most conservative Justices (Scalia, Thomas, and Alito) would have categorically rejected any Equal Protection attack on the statute. The balance of power at the Court in the case was held by a trio of three more moderate Republican appointees (Stevens writing, with Roberts and Kennedy), who rejected across-the-board invalidation of the statute but left open the possibility of specific challenges raises by particular voters capable of showing inappropriate hardship as a result of the statute’s applicability to them.

Crawford was decided in the spring of a presidential election year (April 28 to be exact), and a voter ID law certainly concerns the rules for casting and counting of ballots. Moreover, the lawsuit was filed by the Democratic Party in the state, where the officials defending the law were Republicans—and the legislation had been enacted on a party-line vote.

Accordingly, it would have been preferable if the Supreme Court had been able to achieve unanimity in its disposition of that case. But at least the 3-3-3 divide in the case was not the same as all the Court’s Republican appointees lining up against all the Court’s Democratic appointees, or even all the Court’s conservatives arrayed against the Court’s liberals. Justice Stevens joining with Chief Justice Roberts and Justice Kennedy to reject the across-the-board challenge to the law avoided that outcome.

Furthermore, Crawford lacked an element that is present in Husted v. Obama for America. The complaint in Crawford had been filed in 2005, shortly after the enactment of Indiana’s new voter ID law. The litigation clearly was designed to affect all elections, not just presidential ones, and it was hardly foreordained that it would ready for a Supreme Court decision in a presidential year.

Husted v. Obama for America, by contrast, is a lawsuit filed by a presidential campaign, seeking to affect the rules for the conduct of the immediately upcoming presidential election specifically. Consequently, a Supreme Court ruling in this case much more directly involves the Court in playing umpire for the current presidential election than did Crawford.

In this respect, Husted v. Obama for America comes closer to Bush v. Gore itself. To be sure, Bush v. Gore occupies a different point on the spectrum, perhaps the polar point where a partisan divide in the Supreme Court is most troubling. Bush v. Gore involved the counting of presidential ballots after they had been cast, where the Court’s umpiring of the presidential election was as direct as it could be. Husted v. Obama for America does not involve the rules for counting ballots, only when they can be cast. And even though early voting has already started in Ohio, this litigation remains on the “pre-election” side of the line between lawsuits over the rules before ballots are cast, compared to “post-election” lawsuits afterwards, insofar as this lawsuit concerns the rules applicable specifically to the casting of ballots on November 3, 4 & 5—which have not yet occurred.

Therefore, in considering the various points along the spectrum, a 5-4 thoroughly partisan ruling in Husted v. Obama for America would not be as deeply troubling as one in a “Romney v. Obama” that hypothetically might arise after November 6 concerning the counting of ballots already cast—and thus sharing this same feature as Bush v. Gore.

Still, a 5-4 partisan split in Husted v. Obama for America would be troubling enough. It would be the Court’s Republican appointees agreeing with Ohio’s Republican officials who want to limit the availability of opportunities to cast ballots in this year’s presidential election, with the Court’s Democratic appointees at the same time agreeing with the Obama campaign’s effort to extend the opportunities to cast those ballots. That kind of partisan split would cast doubt on the ability of the nine-member Court to be a fair and neutral arbiter of this sort of fight over the rules for determining which party wins the White House.

Therefore, whether the Supreme Court rules in favor of limiting or extending early voting in Husted v. Obama for America, it would be vastly more preferable if the Court’s ruling is not susceptible to the perception of Republican justices simply siding with what Ohio Republicans want, while Democratic justices simply do what the Obama campaign wants.

The Supreme Court managed to avoid such an outcome in Brunner v. Ohio Republican Party, a unanimous October 2008 decision, which much more than the Crawford decision of six months earlier raised the specter of an ugly partisan divide over rules for casting ballots in a presidential election.

Let’s hope the Supreme Court can repeat the same success of Brunner v. Ohio Republican Party.

A unanimous panel of the Sixth Circuit has essentially ruled that the federal Constitution prevents a state from “toss[ing] out” (to use the panel’s own phrase) ballots that are cast in the correct polling location when the defect under state law is that a poll worker caused the voter to cast a “wrong precinct” ballot at the correct location. Although the panel was careful to couch its decision in the language of “preliminary injunction” law, which required the court to find only that a constitutional violation was “likely” on these facts, the panel’s reasoning leaves little doubt that it would reach the same conclusion were the case in the posture that called for a permanent, rather than preliminary, injunction.

Moreover, the panel based its decision on both Equal Protection and Due Process grounds. There had been much discussion in the litigation of this case about whether Equal Protection properly applies, as Ohio law across-the-board invalidates all wrong-precinct ballots (and thus does not differentiate among them). As the panel opinion explains, however, the Supreme Court already had applied Equal Protection analysis to an across-the-board requirement that voters present a specific form of photo identification in order to cast a countable ballot—and, I might add, long ago the Supreme Court ruled invalid under Equal Protection an across-the-board state law that made payment of a poll tax a prerequisite to voting. Thus, the panel’s opinion should serve as a clarifying precedent that the same Equal Protection analysis should apply to any state rule that disqualifies some ballots from being counted. (This is not to say that an Equal Protection claim will win in all such instances, only that the disqualification of ballots always must be justified under Equal Protection principles.)

The panel opinion is especially noteworthy for speaking powerfully on behalf of protecting “the fundamental right to vote” (page 27) in the specific context of correct-location/wrong-precinct issue—or what many in the field of election administration call the “right church, wrong pew” problem. The opinion includes a photocopy of a precinct guide to illustrate “how easily poll workers can make mistakes under the pressures of election day” (pages 19-20). The panel then pointedly observes that to disqualify ballots in these circumstances “effectively requires voters to have a greater knowledge of their precinct . . . than poll workers”—making it unreasonable to “permanently reject their ballots without an opportunity to cure the situation” (page 21). After carefully considering all the justifications the State offered for invalidating these ballots, the panel explained why they are not strong enough to warrant “the summary rejection” of them “where the voter’s only mistake was relying on the poll-worker’s precinct guidance” (pages 23, 27).

Specifically discussing the Due Process claim applicable to this situation, the panel squarely rejected the state’s contention that there can be no Due Process violation predicated on poll-worker negligence. Instead, the panel accepted the plaintiffs’ point—correct in my view—that, even assuming that intentional state conduct is necessary to sustain a Due Process claim—the requisite intent exists in the state’s intentional decision to disqualify the ballots. Here’s what the panel said on this important point: “we find sufficient indicia of purposeful conduct in the State’s intent to enforce its strict disqualification rules without exception, despite the systemic poll-worker error identified in this litigation and others.” Indeed, observing that the Hunter case arising from the 2010 election “shed light on this problem last year,” the panel added “the State [still] persisted in its position” (page 25). Accepting that a Due Process violation exists only when a state’s voting rules are “fundamentally unfair” (page 24), the panel effectively condemned the disqualification of correct-location/wrong-precinct ballots as flunking this elementary standard.

Thus, I concur with Rick Hasen’s assessment of the panel’s decision as a “major victory for voters’ rights.” We may indeed look back on this decision as, in Rick’s words, the “most important” one “of this election cycle,” once it is all over. As of now, this ruling certainly can claim this preeminent status, given its powerful vindication of the basic right to vote. Moreover, it is necessary to note (this being an election case), all three judges on the panel were Republican appointees. Therefore, it is impossible to accuse this panel of harboring a partisan bias against the defense of Ohio’s draconian provisional voting laws by the state’s Republican Attorney General, Mike DeWine. The decision, in other words, is unimpeachably fair.

There are other elements of the decision that deserve mention.

First, the panel left undecided the issue of ballots cast in the wrong polling location. The panel interpreted the district court’s decree as not extending to that issue and, in a footnote, left it to further consideration in the district court on remand. If I were the plaintiffs’ attorney, I would hesitate to press too hard on this issue, given the tenor of the panel’s questioning at oral argument. These three judges likely would see the constitutional calculus very differently in the context of voters showing up at the wrong polling location. (The panel did, however, carefully note that its ruling extended to ballots cast at a county election board’s office, since under state law that location qualifies as a correct polling place.)

Second, the panel rejected the plaintiffs’ constitutional challenge to a different Ohio rule regarding the counting of provisional ballots. That other rule disqualifies ballots when voters have not properly written or signed their names on their provisional ballot envelopes. The panel premised its ruling on the fact that, in contrast to the “right church, wrong pew” situation, voters could protect themselves from disenfranchisement by the simple exercise of writing and signing their names. Even so, the panel was careful to leave open the possibility that a better developed factual record might show the disqualification of ballots in this situation to be unjustified; the plaintiffs simply had failed to meet their evidentiary burden on this issue, and thus the district court’s preliminary injunction could not be sustained on the existing record. Yet I am sure that the plaintiffs’ attorneys are capable of taking a hint and recognizing, at least with respect to this panel, that they face an uphill battle if they wish to attempt to build a record that would make a convincing case on this point.

Third, the panel also considered the relevance of a consent decree that protected from poll worker error voters who used the last four digits of their Social Security Number as their form of voter identification. The panel pronounced that “the consent decree likely violates the equal protection principle recognized in Bush v. Gore” insofar as it mandates the counting of ballots when voters used this permissible form of ID but not, despite being subjected to the same kind of poll worker error, when they used a different—yet equally permissible—form of ID (like a driver’s license). This important Bush v. Gore point applies to two groups of ballots not covered by the panel’s affirmance of the preliminary injunction with regard to the “right church, wrong pew” ballots: (1) ballots cast in the wrong location as a result of poll worker error, and (2) ballots with envelopes having a name or signature problem as a consequence of poll worker error. The panel left it to the litigants and the district court, on remand, to consider which way to solve this Bush v. Gore problem: either extending the scope of the consent decree to cover voters who use different (but equally valid) forms of ID, or curtailing the decree so that the state can disqualify ballots in these two types of situation regardless of the type of ID the voter used. It will be interesting to see how this issue unfolds as the case proceeds on remand.

In the meantime, today is an occasion to celebrate the most important feature of the panel’s decision: the federal Constitution requires the counting of “right church, wrong pew” ballots caused by poll worker error. This result is a vindication of the basic principle that the federal Constitution protects voters from fundamentally unfair disenfranchisement—and all federal judges, regardless of which president appointed them, stand ready to enforce this bedrock principle.

Why one should hope for a unanimous decision, whichever way it comes out.

Ohio’s Secretary of State, Jon Husted, has gone straight to the U.S. Supreme Court, bypassing the full U.S. Court of Appeals for the Sixth Circuit, in seeking to nullify the preliminary injunction to stop the effect of the state’s new early voting rules. The injunction was ordered by a federal district judge and affirmed by a three-judge panel of the Sixth Circuit.

Justice Kagan, as Circuit Justice for the Sixth Circuit, has set a deadline of 7:00 pm Friday (October 12) for the Obama campaign to respond to Husted’s Supreme Court filing. Justice Kagan could grant or deny a stay of the preliminary injunction herself, but she is more likely to refer Husted’s request to the whole Supreme Court, which would have the authority to overturn any decision she made on her own.

I will wait until seeing the Obama campaign’s response before reaching any conclusion on whether the Court should, or should not, grant Husted’s request to block the district court’s injunction (and even then I might remain undecided, a luxury the Court itself does not have). Here I wish only to observe one potential consequence of Husted’s decision to take the case directly to the Supreme Court, rather than seeking review by the full Sixth Circuit.

Whether the Supreme Court grants or denies Husted’s request, there is the possibility of all nine Justices—five Republican appointees and four Democratic ones—agreeing on the outcome. Such unanimity, whichever side prevails, would be far more preferable from a nonpartisan perspective than if the five Republican appointees side with Husted, a Republican, and the four Democratic appointees in dissent express their agreement with the Obama campaign’s position.

Had Husted taken the case to the full Sixth Circuit, however, there was virtually no chance that he could have secured a victory there without provoking a partisan split among those appellate judges. The majority opinion of the three-judge panel that had ruled against him was written by a Democratic appointee, and Husted could anticipate other Democratic appointees on the full Sixth Circuit lining up to support their colleague. Therefore, to win before the full Sixth Circuit, he would have needed the Republican appointees on that court to outvote their Democratic colleagues, who would have vehemently dissented.

That intense partisan divide would have replicated the one that occurred in 2008 when the Ohio Republican Party sued Jennifer Brunner, then Secretary of State and a Democrat, over the state’s voter registration database. In that case, however, the Supreme Court managed to maintain unanimity in ruling in favor of Brunner, granting her request to nullify a temporary injunction that had been entered against her. Husted, accordingly, can hope that the Court will be similarly unanimous in voiding the order currently in force against him. If he does obtain a unanimous Supreme Court victory, then by bypassing the full Sixth Circuit he will have helped the federal judiciary to avoid all the ugliness that an intensely partisan split decision in that court would have produced.

It is not inconceivable that the four Democratic appointees on the Supreme Court would join their five colleagues in ruling against the Obama campaign in this case. Previously, I’ve discussed analytic vulnerabilities in the Equal Protection claim upon which the Obama campaign’s lawsuit is predicated. It is true, of course, that the district court and all three judges on the Sixth Circuit panel ruled in the Obama campaign’s favor, but the district court’s treatment of the Equal Protection claim had its own analytic perplexities, and one judge on the Sixth Circuit panel pointedly distanced herself from the Equal Protection reasoning of the majority opinion (as I’ve noted). Thus, it is possible that the four Democratic appointees on the Supreme Court will decide that the Obama campaign’s Equal Protection claim is untenable, and for that reason would be willing to join in an order nullifying the injunction against Husted and the Ohio early voting rules he is administering.

It is also possible, conversely, that the five Republican appointees on the Supreme Court will join with their four colleagues to conclude unanimously that, regardless of which side has the better of the argument on the merits of the Equal Protection claim, it is not worth it for the Court to intervene at this point to block the injunction ordered by the district court and approved by the three-judge panel. After the panel ruled, but before Husted bypassed the full Sixth Circuit, I explained why a judge who is dubious about the merits of the Obama campaign’s Equal Protection claim might still be willing to let the injunction remain in place just for this year’s election—or, to be technical, might decline to exercise discretionary authority to review the validity of the district court’s temporary injunction. That explanation applies just as much to the Supreme Court as it did to the full Sixth Circuit, if not more so, since the discretionary nature of the Supreme Court’s jurisdiction is widely understood (among attorneys at least) to mean that the Court often declines to review lower court decisions that it considers erroneous. “The Supreme Court is not a court of error” is a refrain attorneys often hear.

Thus, it is plausible that all nine Justices will decide to deny Husted’s request, just as it is plausible that all nine will decide to grant it. The case could go either way, in other words, and my main point here is that—for sake of the system as whole—it would be far better, whichever way it goes, that it go that way unanimously rather than by a 5-4 ruling. If Chief Justice Roberts crosses party lines, as he did in the health care case, to vote with the four Democratic appointees, leaving his four Republican colleagues in vitriolic dissent, I suppose that would be better than a straight party-line divide. But it still would leave one with the uneasy sense that judicial decisions in election-related lawsuits are too much at the mercy of each judge’s partisan background.

In 2008, I advocated the idea of a structurally nonpartisan three-member “amicus court”—one Democrat, one Republican, and the third chosen by the first two—that would issue opinions in high-profile election cases in advance of the Supreme Court (or whatever court actually had the authority to adjudicate the particular case). The “amicus court” could then submit its reasoning to the Supreme Court in the form of an amicus (friend-of-the-court) brief. The virtue of a court that is structured to be evenly balanced between both sides in its nonpartisanship—designed this way specifically for hotly disputed election cases that pit Democrats and Republicans fighting fiercely over the rules of the voting process—is that, because of the mechanism by which the court’s judges are selected, the court’s decision has the greatest possible chance of being accepted by either side as fair and impartial. This virtue applies, most importantly, regardless of which side the court’s ruling favors.

As I contemplate what the Supreme Court might do in Husted v. Obama for America (as the case is captioned there), I find myself thinking that I would trust whatever ruling an “amicus court”—structured in the way I have described—would reach. Alas, however, there is no such “amicus court” assembled this year to assist the Supreme Court with this case. In its absence, the next best thing is a unanimous decision from the Supreme Court because, whichever way it goes, such unanimity is another means by which to obtain an inherently nonpartisan outcome to a high-stakes election dispute.

[NOTE: This commentary was prepared before the Sixth Circuit’s decision today in the seperate litigation over Ohio’s rules for the counting of provisional ballots.  An initial read of that decision does not, in my judgment, affect any of points made above concerning the desirability of unanimity in the Supreme Court regarding the early voting case.]

The full Sixth Circuit should let stand the panel’s decision, which is best understood as an exercise of the court’s “equity” powers regarding temporary injunctions, especially as four federal judges have agreed that a temporary injunction is appropriate and none have disagreed.

Four federal judges agree that for this year’s election Ohio may not deny the same opportunities for in-person early voting to non-military voters that the state grants to military voters. None have ruled the other way. The four are two appellate judges and two trial judges, one sitting by designation on the appeals court. Two are Democratic appointees and two Republican, although one of the latter—Judge Helene White—was originally nominated by President Clinton before being re-nominated by President Bush as part of a bipartisan compromise.

A similar spirit of bipartisanship should let Friday’s essentially unanimous appellate ruling stand. Although Ohio’s Republican Attorney General Mike DeWine might wish to ask the full Sixth Circuit appeals court to review the three-judge panel’s ruling—perhaps even more than Ohio’s Republican Secretary of State Jon Husted (who has distanced himself from more partisan moves when he sees them as contrary to sound practices of election administration)—the better course at this point is to acquiesce in the ruling and figure out how best to implement it. If DeWine does ask the full Sixth Circuit to nullify the panel’s ruling, even Republican-appointed judges who may be dubious about the merits of the ruling should see that a sharply divided reversal (which would be far from the judicial unanimity that has prevailed so far) would not be worth the distrust in the judiciary that such a reversal would fuel.

Although there was much discussion of Equal Protection (as a fundamental principle of federal constitutional law) in what the three judges wrote on Friday, the panel’s decision is best understood as an exercise of “equity” law, an ancient branch of Anglo-American jurisprudence that governs the issuance of court-ordered “injunctions,” which are decrees that require defendants to stop engaging in challenged practices. “Preliminary injunctions” are a distinct subset of a court’s “equitable” powers, designed specifically to be temporary measures to put the contesting parties in the fairest possible position during the time that the court deliberates about the ultimate merits of the matter under the law. Friday’s ruling affirmed the granting of a preliminary injunction against the state’s statutory provisions that had engendered the differential treatment of military and non-military voters during the last three days of early voting before Election Day (November 6), and this ruling can be seen as a judicial effort to figure out what is the most “equitable” temporary situation that should occur in the context of early voting in Ohio for this year’s presidential election, recognizing that the ultimate Equal Protection claim regarding the distinction between military and non-military voters applies to future elections as well and will be decided in due course following full judicial proceedings on the merits of that claim.

The temporary and “equitable” quality of Friday’s affirmance of the preliminary injunction is best captured by Judge White’s separate opinion in support of the district court’s decree. For her, the case in its current posture came down to the facts that (1) Ohio had such a poor history of conventional polling place voting on Election Day in 2004; (2) Ohio had adopted in-person early voting to alleviate overcrowding at the polls on Election Day; (3) Ohio had used the last three days of early voting successfully in 2008 to avoid long lines at the polls on Election Day; and (4) local boards of elections were prepared for a repeat of the successful procedures used in 2008, but not prepared for increased Election Day turnout that might result from the rollback of early voting opportunities available in 2008. Here’s how she explained her reasoning: “The key distinguishing feature here is that Ohio voters were granted [early voting] . . . in response to the unacceptably burdensome situation at many Ohio polling sites during the 2004 election” (page 27). To this she added, citing evidence in the record, “Ohio’s two largest counties . . . have not budgeted or planned for any increase in election-day voting caused by the elimination of weekend and after-hours voting.” Consequently, she believed the best interim solution is to leave early voting in place as it was practiced in 2008, while giving the state’s legislature the opportunity to develop a longer-term approach to early voting that would satisfy Equal Protection and avoid an unwarranted risk of polling place chaos.

It is worth observing how far removed these “key” components of Judge White’s reasoning are from the distinction between military and non-military voters that is at the heart of Equal Protection claim in the case. Her analysis might be vulnerable on this ground were the case at its ultimate conclusion, when it is time for a final adjudication of the Equal Protection claim. But in the context of deciding at “the eleventh hour” (as she put it) what the temporary, interim situation should be regarding this year’s voting, it was appropriate for her to focus on what might actually happen if the last three days of early voting are unavailable in the those localities where “significant numbers of voters . . . have come to depend on” them (page 26). “To [do] otherwise is to ignore reality,” she forcefully proclaimed (page 28).

Indeed, in repeatedly emphasizing that she would not let her decision “be divorced from reality,” Judge White acknowledged that her decision might be different—a “dissent”—if solely before her were an “abstract” Equal Protection issue. She distanced herself from the majority opinion’s Equal Protection analysis. In particular, she explicitly stated: “I cannot agree with the majority’s assertion that ‘Plaintiffs introduced extensive evidence that a significant number of Ohio voters will in fact be precluded from voting without the additional three days of in-person early voting’” (at page 25, quoting Maj. Op. at 12). As Judge White explained, it was simply too much to assert that voters would be “precluded” from voting absentee these three days; the evidence did not support that. Mail-in voting remains an option for all Ohio voters, as do the remaining available days of early voting (which started last Tuesday), in addition to Election Day itself. Thus, the majority overstated the burden that Ohio’s new early regime imposed on voters.

Still, the flaws of the majority’s opinion did not stop Judge White from independently affirming the district court’s preliminary injunction—precisely because it was a preliminary injunction rooted in the “facts” of election administration as practiced “on the ground in Ohio” (as she again put it, on page 24), with the urgent need to avoid last-minute changes as local election officials do their best to prepare for the presidential election (in which voting has now already started). For the same reason, other judges on the Sixth Circuit should decline any invitation for the full court of appeals to engage in “en banc” review. The majority opinion does no permanent damage to the development of Equal Protection law, once it is understood that the panel’s ruling is merely about the appropriateness of a temporary, interim order in the specific context of Ohio’s experience with election administration. There will be other occasions for the Sixth Circuit to address final judgments that rest on Equal Protection grounds, and in those contexts the appellate court can appropriately develop the law of Equal Protection as it applies to voting. Here and now, it is better for the full court to leave well enough alone—the old cliché is apt in this context—and allow Ohio’s Secretary of State and local election officials to get on with the pressing business of figuring out exactly how early voting will work this year during the last three days before November 6.

(To clarify, letting the panel decision stand is not the same is saying that it was correct, even Judge White’s separate opinion.  There is a difference between thinking a panel decision incorrect, or dubious, and believing it should be subjected to full court “en banc” review.  One reason to let it stand is simply a matter of timing; each day brings us closer to November 6, and that fact alone is reason enough to stop further litigation over the last three days of early voting.  Another consideration, as I have indicated, is the consequences of the full Sixth Circuit being sharply divided if it were to take up “en banc” review; better to avoid that messy and counterproductive judicial division.  Finally, the fact that the panel’s ruling can be seen as an exercise of equity power, rather than an Equal Protection precedent, is an additional basis for believing that “en banc” review is not imperative.)

The next move is the Secretary of State’s. According to the Sixth Circuit’s majority opinion, he could leave it to local boards to decide whether and to what extent to hold early voting on November 3, 4, and 5, as long as the same availability applies equally to military and non-military voters. If he does so, however, he will be in the position that he will need to break ties if the local boards split 2-2 along party lines on whether to make weekend early voting available (as they did earlier this year). In order to avoid having to break such ties, Husted issued a directive requiring uniform early voting hours across all 88 counties in the state. He could do so again, but he would still need to decide whether, or the extent to which, early voting should be uniformly open or uniformly closed during those three days.

In making this decision, and without setting hours of his own, Husted could take his cue from Judge White’s separate opinion. She made clear that she thought the district court’s preliminary injunction should be understood as requiring the restoration of early voting to what it had been in 2008. This understanding followed from her analysis (as warranted when “balancing the equities” at the heart of the preliminary injunction inquiry) of the “facts on the ground” concerning the preparation of the local boards for this year’s election in light of Ohio’s history of election administration in the past decade. Therefore, Husted could require all 88 counties to have this year the same schedule of early voting that they had in 2008 for the Friday, Saturday, and Sunday immediately before Election Day.

This directive would have the requisite uniformity: it would be treating all counties the same vis-à-vis the temporary need (in light of the preliminary injunction) to maintain consistency between this year and 2008, and it would treat military and non-military voters equally across the entire state. At the same time, built into this uniform directive would be an accommodation of different past local administrative practices: if some counties held fewer hours of early voting in 2008 during the last three days before Election Day than other counties did, then this directive would not require those “fewer hours” counties to do more than they did in 2008. Since the “key” component of Judge White’s reasoning was to protect the reasonable reliance of voters and local election boards for what had been available in 2008, especially in the state’s two largest counties, it would in no way contradict this reliance interest to permit all 88 counties to do this year whatever they did in 2008 regarding those last three days.

Going back to 2008 practices is obviously not what Ohio’s legislature wanted when it enacted its changes to the state’s early voting rules. But as is happening throughout the country in a variety of different contexts—like in Pennsylvania, with respect to its new voter ID law—state legislatures are learning that they cannot assume that their changes in the rules will have immediate effect for this year’s election, at least not when there are colorable constitutional challenges to the newly enacted rules. Whether or not this kind of “anti-retrogression” inquiry is an appropriate part of Equal Protection analysis to be applied in the final adjudication of these constitutional claims, judges are invoking this idea in the “balancing of the equities” that is the core of preliminary injunction analysis.

When deciding on the fly what should happen temporarily in the interim, while weighty constitutional claims are under litigation, the courts are indicating that the voting process this year should be as it has worked in the past, rather than what the new legislation calls for. That, in essence, is what Judge White said, as did the court in Pennsylvania (insofar as it refused to permit a ballot to be disqualified this year for lack of the newly required ID). That, too, is what should guide Secretary of State Husted concerning the administration of early voting in the remaining days between now and November 6.