SEVEN

A CONSTITUTIONAL AMENDMENT FOR HANDLING FUTURE CRISES

Abolition of the Electoral College would reduce but not eliminate the dangers of a presidential election marred by fraud and post-election chaos. We obviously cannot rule out a national popular vote producing a virtual tie amidst credible allegations of covert manipulation, hacking, or voter-tabulation problems that call into question the accuracy of the count. A major lesson to be learned from the elections of 1876 and 2000 is that we lack a mechanism for reliably resolving such disputes in a way that preserves legitimacy and gives all sides a reasonable degree of confidence in at least the process, if not necessarily the outcome.

The experience of Florida in 2000 also reminds us that numerous things can go wrong. While Florida’s flawed election law produced some of these problems, not even the clearest and most thorough election code can assure that a disputed election will be resolved in a timely and reliable fashion. No legislature, state or federal, can possibly anticipate all of the problems that may arise in an election. Even when problems can be anticipated, their exact nature may remain unknown, preventing them from being easily redressed. Consider malfunctioning equipment, inclement weather, and other logistical problems that cause certain precincts to be shut down for a period of time on Election Day. This is something we know will happen, yet it nevertheless presents judgment calls that will vary from case to case. May (must?) the polls be kept open longer in these precincts? How much longer? Presumably that depends on the cause and severity of the problem. What happens if a hurricane or other natural disaster prevents voting for the entire day in certain locations? If even this foreseeable problem does not lend itself to a perfectly satisfactory solution, imagine the havoc wreaked by the many potential problems that cannot be foreseen.

In part for that reason, abolishing the Electoral College will not guarantee the end of presidential election crises, especially now that hackers in multiple countries are working to infiltrate our voting systems. It is, therefore, imperative that we find a better means for dealing with the kinds of scenarios that arose in 1876 and 2000. To some extent, this has been recognized. Following the 2000 debacle, the Carter-Ford commission was hardly the only effort to explore how to prevent future election crises. Any number of books have been written and legal symposia organized around this very question: How do we prevent another 2000?

Unfortunately, reformers thought too locally. For example, the Help America Vote Act (HAVA) abolished punch-card ballots and requires provisional ballots. But punch-card ballots and voters turned away were only two among many issues in Florida in 2000; Florida was just one state and 2000 was just one election. HAVA’s helpful steps barely begin to guard against future crises. States adopted assorted reforms, many of them clarifying rules governing elections, such as the procedures for voter registration, absentee ballots, and improved voting technologies. These welcome developments, too, failed to address a central issue: How do we respond to another 1876 or 2000? Even if the mechanics of casting and counting ballots were perfected, elections would still be prone to fraud by hackers and others. While doing everything possible to make sure elections unfold properly, we must also position ourselves to handle scenarios in which they do not.¹⁹³

Such efforts are more necessary than ever. We used to worry about conventional fraud by U.S. citizens. In some precincts in Chicago, for example, people, including the deceased, seemed capable of casting multiple ballots. Such chicanery hasn’t disappeared, but the advent of hacking has multiplied the risk of election fraud. It has also multiplied the risk of chicanery that takes longer to detect, something that must be taken into account when we fashion measures for responding to election crises.

While our government should work to develop an impenetrable voting infrastructure system, we need to be realistic, and the history of election technology does not inspire confidence.¹⁹⁴ Who would have thought that butterfly ballots and hanging chads would become decisive features in a twenty-first-century presidential election? While we need to aim for the best, we must prepare for the worst—meaning we must develop a fallback plan in case there is reason to believe that voting systems were hacked (or if, due to another form of fraud or for any reason, the outcome of an election is uncertain).

When we scrutinize past crisis elections, we see three fundamental problems: time pressure, the ad hoc nature of resolving disputes, and partisanship.

The attempt to determine the actual winner was undertaken under serious time pressure. Regrettably, this problem was actually far greater in 2000 than in 1876. In 1876, the nation had until March 4 to determine the winner. That had been the presidential inauguration date beginning with George Washington’s second term, though it was not stipulated by the Constitution. The Twentieth Amendment, ratified in 1933, changed the date to January 20. That change, made to hasten the transition from one president to the next and to reduce the period with a lame duck president, produced an unintended consequence: It left little time to resolve uncertainty over the winner of the election.

With inauguration required by January 20, Congress established January 6 as the day its joint session of Congress counts and records the Electoral College’s votes. The electors meet to cast those votes on the first Monday after the second weekend in December, which in 2000 was December 18. The safe harbor date (guaranteeing that a state’s electoral votes will be accepted) is six days before the Electoral College meets, meaning that in 2000 it fell on December 12—a little more than a month after Election Day. On account of this compressed time frame, the attempt to handle the murky post-election situation in Florida occurred at warp speed. The United States Supreme Court’s dubious determination that Florida law incorporates the safe harbor led the Court to stop the recount on December 12. Justice Stevens opined that Florida had until January 6 to select its electors, whereas Justice Souter and many commentators would have split that difference and given Florida until December 18.¹⁹⁵ For present purposes, what matters is that January 6, the last possible date, was still all too soon.

The resulting problem was twofold. First, while six justices believed that a statewide recount in Florida was called for, only four justices thought there was adequate time. Accordingly, Florida’s effort to determine the statewide (and thus national) winner was deemed doomed for lack of time. Second, everyone involved in the process—the canvassing boards, various state officials including the secretary of state, and the assorted state and federal judges to hear aspects of the case—was forced into hasty action. Fast cases make bad law and bad policy.

A few examples will suffice. As noted, Al Gore’s lawyers requested a recount in four heavily Democratic precincts only, a move that may have cost Gore dearly (by leaving much less time for the statewide recount eventually ordered by the Florida Supreme Court). Asked about that decision, Dexter Douglas, a prominent Florida lawyer who served on Gore’s legal team, explained, “I’m not sure that we ever really had a chance to consider this, we didn’t figure we had time to do [a statewide recount].”¹⁹⁶ Note the double whammy here—Gore’s lawyers lacked time to consider the issue, but to the extent they did consider asking for a statewide recount, they concluded that there was insufficient time. Time after time (so to speak), haste made waste for Gore’s team. Why didn’t they ask a Florida court to establish a statewide standard for the recounts, thereby avoiding the equal protection issue that proved fatal? Douglas claims that he proposed making such a request, but that Gore’s lead lawyer, David Boise, responded, “We don’t have time, the judge [Terry Lewis] has his hands full.”¹⁹⁷

This might be thought sour grapes by the side that lost, were it not for the fact that Bush’s legal team concurs. Barry Richards, a key member of that team, compared the Florida litigation to a “high-speed sailboat race in the high seas in the middle of the storm. . . . So we were making snap decisions as best we could with very little to go on.”¹⁹⁸

It may be that none of this mattered, as perhaps the U.S. Supreme Court was determined to prevent a statewide recount and would have found a way no matter what. But it did not help matters that the Gore team’s counterproductive tactical decisions were dictated less by the law than by the ticking clock. Because the inauguration date of January 20 is established by the Constitution, without an amendment little can be done to expand the post-election pre-inauguration period and reduce the time pressure. Congress can push back the meeting date for the Electoral College (assuming the college is not abolished) and the date Congress accepts the certification of electors, but the actual vote count must be completed by January 20 at the very latest. Realistically, the winner should be determined at least a few days in advance, allowing a little preparation for the inauguration and transition.

A combination of judicial and legislative action could push all the dates back a few weeks, except for the January 20 inauguration. If the dates were pushed back, the next time a Florida-type situation arises, the courts and others won’t be quite as rushed. But a few weeks may not be enough. Accordingly, we need to rethink the January 20 inauguration date. While we’ve become accustomed to January 20, and there were good reasons for moving the date from March to January, we could and probably should amend the Constitution to make it possible for the date to be postponed under extraordinary circumstances. Below I discuss under what circumstances and by whom that determination would be made. For present purposes, we must take notice that lack of time was a big problem in 1876 and even more so in 2000, and likely will be again, unless we take action.

A second problem with the dispute mechanism process in both 1876 and 2000 was its ad hoc nature. In each case, the state and federal governments lacked a clear preexisting system designed for such occasions. In 1876, an ad hoc commission was created to determine the winner of the election. In 2000, for better and worse, no such commission was created. No one knew for sure where to turn. Depending on which expert you believed, the correct venue for resolving the election was the Florida courts, the federal courts, the Florida legislature, or the United State Congress. The uncertainty contributed to the thirty-six days of chaos, culminating in the deeply controversial Supreme Court decision.

The most significant problem in both 1876 and 2000 was that the resolution seemed a function of partisanship rather than neutral principles of law, and was accordingly never accepted by the losing party. While commentators on the 2000 election focus on the outcome-determinative decision by the U.S. Supreme Court, we should not ignore the partisanship of the various state officials whose involvement preceded the Court’s. Virtually Florida’s entire governmental machinery consisted of Republicans with direct or indirect ties to the Bush campaign. That the governor of the state was the brother of the Republican candidate symbolized the problem; the fact that he recused himself from the recount fight in no way eliminated it. Time and again, Secretary of State Harris (who, after all, had been co-chair of Bush’s campaign in Florida) and other Republican officials seemed more interested in carrying water for the Republican candidate than in carrying out their duties impartially.¹⁹⁹ Conversely, the Florida Supreme Court consisted primarily of Democrats and ruled consistently for the Democratic candidate. The partisanship of 2000 echoed that of 1876. The commission that made Rutherford B. Hayes president was composed of eight Republicans and seven Democrats, and marked by 8-7 party-line votes favoring the Republican, Hayes. Government officials in all the disputed states also acted in a matter coinciding perfectly with their political affiliations.

These three problems—lack of time, an ad hoc approach, and partisanship—are interrelated, as the 1876 election demonstrates. Lacking a preexisting mechanism for resolving the election dispute, Congress created an ad hoc commission, but Illinois thwarted the plan by selecting Judge David Davis, expected to be the fifteenth (and one impartial) member of the commission, to the U.S. Senate. Davis’s replacement, like the other fourteen members of the commission, voted along party lines. Despite a genuine effort by Congress to seek a nonpartisan solution, the time pressure and absence of a preexisting commission led to a partisan outcome.

When we focus on the three problems that plagued the 1876 and 2000 elections, the nature of a solution should be clear: a preexisting nonpartisan commission empowered to resolve disputed presidential elections and given ample time to do so.

A PRESIDENTIAL ELECTION REVIEW BOARD

The solution to a disputed presidential election is a permanent nonpartisan commission.²⁰⁰ The commission could be created by statute, but, as noted, a constitutional amendment would be required to make the January 20 inauguration flexible—that is, to allow the commission to postpone it under certain conditions. But we’re getting ahead of ourselves. Before discussing the commission’s powers, we must address preliminary matters about the nature of the necessary tribunal.

History provides useful guidance. We can start by learning from the failure of the commission created during the 1876 imbroglio. As noted, that commission’s failure can be traced to its composition: The commission included more Republicans than Democrats, and almost every determination it made broke down along partisan lines. Members of Congress anticipated this risk, and sought a commission with an equal number of Democrats and Republicans along with the nonpartisan Justice Davis. While their laudable goal was thwarted, the lesson is clear: What I will call the Presidential Election Review Board (PERB) must have an equal number of Democrats and Republicans. As discussed shortly, it should not consist only of Democrats and Republicans.

History also provides an example of a more successful means of resolving a disputed election, one that was used by Minnesota on two occasions: the 1962 gubernatorial race and 2008 U.S. Senate race. A brief review of that history will illuminate how we should prepare for future election crises.²⁰¹

Minnesota’s Success Stories

The 1962 election between the Republican incumbent, Elmer Andersen, and his Democratic opponent, Karl Rolvaag, was so tight that the winner was not determined until mid-March, more than four months after voters cast their ballots. From November 6 (Election Day) until November 20 (the statutory day for certification), the lead seesawed as precincts around the state issued revised counts. On November 20, Anderson led by a few votes, but the State Canvassing Board declined to certify him the winner, because precincts continued to adjust their tallies. On November 26, the five-member board announced that it could not reach a decision. Two Democrats claimed Rolvaag had won, two Republicans favored Anderson, and the fifth board member, a Republican, abstained, claiming that only the State Supreme Court could make the call. The case soon ended up there, and produced a three-two decision (with two Justices recusing themselves because they served on the board), also along party lines, making Anderson the temporary winner.

The loser retained a statutory right to “contest” the election result in trial court. But which court? With the encouragement of the chief justice of the Minnesota Supreme Court, lawyers for the two candidates agreed on a three-judge panel—one judge who had been appointed by a Democrat, one by a Republican, and one who had been appointed to the trial court by a Democrat and the court of appeals by a Republican.

Rolvaag, the one contesting the election, asked for and was granted a statewide recount. Such a recount occurred under the auspices of one hundred recount teams consisting of three members each—a Democrat, a Republican, and a neutral member agreed to by each candidate. The three-judge panel resolved disagreements over particular ballots. The recount, completed on February 5, put Rolvaag back in the lead by 138 votes, but produced 97,000 challenged ballots. The candidates agreed to submit those ballots to ten two-person teams (one Democrat, one Republican) prior to sending them to the three-judge panel if necessary. The bipartisan pairs resolved virtually all of the disputes, leaving only 1,192 ballots for the three-judge panel.

The panel reviewed those ballots and also considered a range of legal claims by the candidates, such as challenges to absentee ballots that did not conform to the statutory standard and challenges to the votes in precincts where alleged administrative failures may have produced tampering. On March 15, the three-judge panel announced that Rolvaag had won by ninety-one votes (out of 1.25 million cast, an insanely thin margin of .0073 percent). The panel claimed that all of its decisions were unanimous. On March 22, Anderson announced that he would not appeal the panel’s determination to the Minnesota Supreme Court. Rolvaag became governor.

Especially when compared to the disputed presidential elections of 1876 and 2000, the 1962 Minnesota gubernatorial race was a staggering success story: There was never a threat of chaos, and the losing party accepted the result. As the Washington Post put it in an editorial, “the state of Minnesota has occasion to congratulate itself,” as its resolution of this ridiculously close election “is a tribute to their political stability and maturity, and to their faith in democratic government.”²⁰²

This successful experience in 1962 proved an invaluable precedent when, almost a half century later, Minnesota experienced another “tie” election, in the 2008 race for a U.S. Senate seat between incumbent Democrat-turned-Republican Norm Coleman and comedian-turned-politician, Democrat Al Franken. On November 5, the morning after the vote, the Associated Press called the election for Coleman, because he led by 762 votes with only a handful of small precincts uncounted. Coleman declared victory; Franken told reporters he expected a recount. Franken noted that this would be the first recount in Minnesota since 1962, “when I was eleven years old. I remember that year very clearly for two reasons. The recount between Elmer L. Anderson and Karl Rolvaag. And the Gophers were in the Rose Bowl that year.”²⁰³ His recollection indicates that the Anderson-Rolvaag election made a deep impression on Minnesotans.

The next two weeks witnessed a public relations battle between the sides about recounts, while precincts continued to adjust the original count. On November 18, the State Canvassing Board (consisting of two state Supreme Court justices, two trial judges, and the secretary of state) met for the first time and heard from the campaigns. The next day the board certified Coleman the winner by 215 votes (out of almost 3 million cast, a margin of less than one-hundredth of a point). Since the margin was less than half a percent, Minnesota law required a recount. Because 10 percent of the vote (almost 294,000 ballots) had been by absentee ballot, and many absentees were rejected, Franken’s team thought they had a legitimate chance of prevailing.

As in 1962, during the recount process the candidates challenged individual ballots, but far fewer—originally 6,689, whittled down (by the candidates’ lawyers withdrawing weak challenges) to just 1,337 sent to the State Canvassing Board. The five-member board reviewed each of these in a process broadcast over the internet. The board, which included two Democrats, two Republicans, and one independent, reached consensus on every ballot. However, state law empowered the board to look only at ballots counted in the original vote, not those disqualified. The latter included a few thousand absentee ballots, and Franken publicly called for inspection of those. (Minnesota election law imposed fairly rigid requirements for absentee votes, but some absentees had been rejected for no apparent reason.) While claiming that it lacked the power to inspect disqualified ballots, the board pointedly called these ballots “uncounted,” rather than “rejected,” and a few days later the secretary of state instructed local boards to identify and create a pile of wrongly rejected ballots.

On December 5, the day the recounts were supposed to be completed, Coleman still led by 192 votes, at least according to the secretary of state’s office—amidst rumors of “missing votes” in mysterious envelopes. Coleman petitioned the Minnesota Supreme Court to declare that uncounted ballots could not be considered. In a weird case of history repeating itself, just as in 1962, two justices recused themselves because they were members of the State Canvassing Board, and on December 18 the remaining five produced a three-two decision along party lines. The Court ruled for Coleman, but with a twist. It agreed with Coleman that votes originally not counted could not be subject to the recount. However, the Court made an exception for ballots that the two candidates and the local election board agreed were wrongly rejected. The Court instructed the candidates’ attorneys to examine ballots in good faith.

That wasn’t good enough for Justice Alan Page, the former star lineman for the Minnesota Vikings of the NFL, whose blistering dissent attacked the “perverse result” in which the Court “has abdicated its role as the defender of the fundamental right to vote.”²⁰⁴ The names had changed, but this was Bush v. Gore all over again, with Page playing the role of Justice Stevens, lamenting that a biased decision not to count votes compromised the integrity of the judiciary and democracy. But Stevens’s lament came at the very end of the case; Page’s disgust with his colleagues proved premature.

By the time the Minnesota Supreme Court decision was handed down, the ongoing recounts had given Franken a forty-nine-vote lead. Suddenly, the two sides changed their positions on the uncounted ballots: Coleman now wanted them counted and Franken did not. Perhaps more surprisingly, their respective attorneys agreed about most of the ballots identified by the local boards as wrongly rejected. The inclusion of these votes expanded Franken’s lead to 225, and on January 5, 2009, the State Canvassing Board announced that he had won. Franken declared victory; Coleman filed an election “contest” which, under Minnesota law, entitled him to a full-blown trial. Meanwhile Franken brought suit in the Minnesota Supreme Court, urging the Court to compel the governor and secretary of state to certify his election in the interim, so that Minnesota would not be without a second senator before the trial was completed. The Court unanimously declined the invitation.

By law, the election contest trial had to commence twenty days after the filing of the contest—January 26. Minnesota law required the chief justice of the Supreme Court to appoint a three-judge panel of district court judges to hear the case. This law was adopted after the 1962 election in order to make permanent the successful approach to that case, but with some tweaking. In 1962, the chief justice urged the parties to agree on a neutral judge, and their attorneys came up with the idea of a three-judge panel and picked the judges themselves. In 2008, the law required the appointment of such a panel, and authorized the chief justice to select it. However, the chief justice was one of the two justices who had recused himself when the case came before the Court earlier, and he decided to remain on the sidelines. The appointing authority fell to the next most senior justice, Alan Page. Page picked the three judges—one of whom had been appointed by a Republican governor, one by a Democrat, and one by an independent, former governor Jesse Ventura.

At the trial, Coleman’s lawyers called for the inspection of roughly 5,000 absentee ballots that had been rejected. The court unanimously rejected the request, ruling that it would inspect an uncounted ballot only if Coleman produced evidence that the ballot complied with Minnesota’s rigid requirements for absentees. When ballots (submitted by both candidates) fitting that description were produced, and evaluated by the panel, they actually extended Franken’s lead to 312.

All things considered, the three-judge panel worked remarkably well, reaching unanimity on virtually all ballots and other decisions. On April 13, more than five months after voters had gone to the polls, the Court issued its final order, accompanied by a sixty-eight-page opinion. Franken remained the winner. Coleman appealed the decision to the Minnesota Supreme Court. On June 30, the Court unanimously affirmed the decision of the three-judge panel. Coleman conceded, and Franken finally assumed office.

There were some complaints in Republican circles. Ben Ginsburg, an election law titan who helped Bush prevail in the 2000 recount and assisted Coleman during the contest trial, was particularly acerbic. He dubbed one key anti-Coleman ruling, handed down by the three-judge panel on Friday, February 13, the “Friday the thirteenth ruling,” implying that it belonged in a horror film, and called another ruling by the panel “unprincipled.”²⁰⁵ When the dust settled a few months later, the Wall Street Journal editorialized that Franken had “effectively stolen an election.”²⁰⁶

But these protests came from outside the state. Coleman himself never claimed unfair treatment, and the major Minnesota newspapers that had endorsed him now praised the process that had produced his defeat. The Minnesota Star Tribune editorialized that even critics of the time-consuming post-election process “have to admire the result: a clear decision, arrived at with thoroughness and care, and with no evidence of fraud. That makes it possible it will be accepted by people of all political persuasions as the legitimate outcome.”²⁰⁷ The Twin Cities Pioneer Press proved the point. Noting that it had enthusiastically endorsed Coleman, it declared that “the system worked. . . . Minnesota is blessed to have an impartial, competent, independent judiciary.”²⁰⁸ Professor Edward Foley, an election law expert who studied the case, succinctly captured the main takeaway from the Minnesota experience: “As in 1962, it was the selection of the three-judge panel that was the most important factor in the state’s successful resolution of the election.”²⁰⁹

If we think of Minnesota’s 1962 and 2008 election disputes as the counterpoint to Bush v. Gore, and ask what lessons can be learned from the state’s successful resolution of “tie” elections, Foley’s conclusion warrants emphasis. It perhaps should go without saying, but an impartial decision-maker is essential to public acceptance of the verdict in a disputed election. A second lesson is that sometimes the process takes time. If one reason courts stumbled in Bush v. Gore was the time pressure, one reason the Minnesota process worked well was the willingness of the decision-makers to take months in order to make sure they got things right. That luxury is less available when it comes to the presidency, but mainly because of constitutional constraints that can be loosened by amendment. A final lesson is the value of transparency: Coleman’s election contest was not only open to the public, but streamed on the internet. Such openness mitigates against bias or corruption and reduces suspicion.

The three-judge panels (selected by the attorneys in 1962 and by the Minnesota Supreme Court in 2008) lived up to the expectation of impartiality and were embraced by the public. However, that will not always be the case with a panel chosen after Election Day. In the midst of a contested election, people on one side or the other will likely be suspicious of whoever does the appointing and therefore of whomever he or she appoints. Minnesota got lucky. As suggested earlier, a preexisting board is the better way to go.

That said, a preexisting decision-making body does not solve all problems, as we shall now see. For, in addition to Minnesota’s positive example of how to resolve a disputed election, we have a negative example on the federal level—not quite in the area of deciding an election, but in something related.

The Unsuccessful Commission on Presidential Debates

The national Commission on Presidential Debates (“CPD”) is an analogue to the Presidential Election Review Board (“PERB”) created by my proposed constitutional amendment. The Commission on Presidential Debates, like the envisioned PERB, was designed to produce presidential elections that conform to democratic ideals. We can think of the CPD and PERB as the alpha and omega of presidential elections: the CPD, a pre-election vehicle to ensure that the electorate can make an informed choice; PERB, a post-election vehicle to ensure that the people’s choice is reliably determined.

The Commission on Presidential Debates arose from a combination of two realizations: 1) televised debates should be a permanent feature of presidential elections; and 2) left to their own devices, the candidates will often avoid such debates. Following the debut of presidential debates in 1960, there were no debates in 1964, 1968, and 1972, and only two incomplete debates were held in 1980. In 1964, 1968, and 1972, the front-runner (Lyndon Johnson in ’64, Nixon in ’68 and ’72) perceived a risk and no advantage from participating in debates. In 1980, Jimmy Carter refused to participate when the sponsor, the League of Women Voters, invited independent candidate John Anderson as well as Carter and the Republican nominee, Ronald Reagan. (Reagan and Anderson did debate once, and Reagan and Carter debated once as well, the latter occurring less than a week before the election when the League disinvited Anderson.)

Recognizing that the candidates held the debates hostage, a private group of influential Democrats and Republicans formed CPD, an independent nonprofit organization. The two major parties gave their blessing to CPD, and in 1988 it successfully sponsored two debates between presidential candidates George H.W. Bush and Michael Dukakis, as well as one between their running mates, Dan Quayle and Lloyd Bentsen. In 1992, CPD sponsored three three-way debates between Bush, Bill Clinton, and independent candidate Ross Perot, as well as one debate among their running mates. However, problems arose in 1996, when the Republican candidate, Bob Dole, believing that Perot had cost the Republicans the election in 1992, did not want him included in the debates. Eventually, the Dole and Clinton campaigns agreed to two debates that excluded Perot, and one vice-presidential debate that excluded Perot’s running mate. George Will remarked that Perot’s exclusion “actually was a deal struck by the Dole and Clinton campaigns,” rather than a determination by CPD.²¹⁰

Although presidential debates have taken place in every debate since, the ’96 scenario exposed two major problems with the CPD. First, the commission is toothless. The Dole campaign simply rejected the invitations by the CPD, and instead privately arranged debates with the Clinton campaign. Ever since, the modus operandi of the national campaigns is to accept conditions proposed by CPD . . . when they feel like it. When it comes to the number and format of debates, as well as which candidates to include, each election produces a debate about debates in which CPD at most plays a mediating role. The commission’s fatal flaw is that it lacks authority. It can only recommend and hope that the candidates accept its recommendations.

The second major problem with CPD is that it is bipartisan but not nonpartisan: It does the bidding of the major party candidates at the expense of third-party candidates and the public that might benefit from more inclusive debates. CPD’s co-chairs and board members have always been Democrats and Republicans (many of them activists, donors, and former officeholders).²¹¹ Its co-founders and long-standing co-chairs, Paul Kirk and Frank Fahrenkopf, were former chairs of the Democratic National Committee and Republican National Committee, respectively. They and other members of the CPD board have been forthright about their hostility toward third parties.²¹²

The only time CPD invited a third-party candidate to debate—Perot in 1992—is the exception that proves the rule: The CPD did not want to invite Perot, and did so reluctantly only because Bush insisted.²¹³ In every election since, CPD has excluded third-party candidates. The 2000 campaign, for example, included two such candidates, Ralph Nader and Pat Buchanan, who had national followings and serious platforms. Many pundits and the public believed that their inclusion in at least one of the presidential debates would have enriched the discourse.²¹⁴ The CPD did not invite them.

The criteria established by CPD for inclusion in the debates are designed to ensure that third-party candidates will rarely qualify. Such candidates must, among other things, receive at least 15 percent of the vote in public opinion polls. This amounts to a self-denying prophecy, since third-party candidates will rarely poll at such levels unless given the opportunity to participate in debates with the Democratic and Republican candidates. But more forgiving criteria (such as a much lower threshold for inclusion in at least the first debate) will not be adopted as long as CPD is of, by, and for the two major parties. Recall that the three-judge panels that worked so well in deciding disputed elections in Minnesota were tripartisan, including someone not considered loyal to either major party.

When creating the Presidential Election Review Board (PERB), we should learn from the experience of Minnesota and CPD. We should prioritize avoiding the two major problems with CPD. Unlike CPD, PERB’s decisions should be given, if not fully binding force, at least a favorable presumption. Also unlike CPD, PERB should include at least one member involved in the decision-making who is not affiliated with either the Democratic or Republican parties. In one key respect, PERB should emulate CPD: It should be established on a permanent basis, to avoid the pitfalls of an ad hoc approach to a disputed election.

Third parties should present little problem for PERB, especially if the Electoral College is maintained: No third-party candidate has captured a state in the last fifty years. Even without the Electoral College, third-party candidates are unlikely to figure directly in post-election disputes. Their disputed votes may affect the outcome, as when Florida’s butterfly ballot gave Pat Buchanan votes intended for Al Gore, but Gore, not Buchanan, was the party adversely affected. Buchanan himself had zero chance of winning Florida. Third-party candidates rarely have a realistic chance of winning any states, much less the presidency.

Nevertheless, we should ensure that PERB is positioned to protect the interests of all candidates, not only the Democrat and Republican. After all, in 1992 Ross Perot received almost 19 percent of the vote, and before he dropped out of the race (only to rejoin it ten weeks later) was actually ahead in some polls.²¹⁵ In 1912, Theodore Roosevelt, running on the Bull Moose ticket, received 27.4 percent of the vote—more than the Republican candidate, Howard Taft. Because we cannot rule out a third-party candidate with a chance to win the presidency, at least one member of PERB should have no party affiliation.

The inclusion of one or more nonaffiliated PERB members will also serve another important purpose: Providing a third member avoids the risk of a voting deadlock without giving the Democrats or Republicans the extra member. But the inclusion of one nonpartisan member should not be used as an excuse to make the other members highly partisan. All members should have a reputation for independence and integrity. Current officeholders should be excluded because of the obvious conflict of interest.

The list of Democrats and Republicans who might pass muster today includes Joseph Lieberman, Jon Huntsman, Sam Nunn, Olympia Snowe, Colin Powell, Bill Richardson, Kay Bailey Hutchinson, William Cohen, Chuck Hegel, and Evan Bayh, among many other possibilities. The list of potential nonaffiliated members today includes James Mattis, Lowell Weicker, Jim Lehrer, David Boren, Richard Lamb, and Michael Bloomberg. (Needless to say, these lists are illustrative and not remotely exhaustive.) Note that many of these people have a legal background, which could be valuable insofar as contested elections often involve questions of statutory interpretation.

There will be presidential elections in which we do not immediately know the winner. That, alone, does not a crisis make. The crisis arises when we do not know who will decide the winner, or how that decision will be made. Such crises are preventable. Consider the following proposed constitutional amendment that simultaneously abolishes the Electoral College and establishes a Presidential Election Review Board.

CONSTITUTIONAL AMENDMENT 28

Section 1 On the Tuesday after the first Monday of November of the presidential election year, voters nationwide shall cast a ballot for the ticket of president and vice president. The vote tabulations will be submitted to Congress by each voting location at the earliest possible date but no later than January 13, unless the date for inauguration is postponed in accordance with the procedures described below. Congress shall tabulate the total vote no later than January 13. The candidates on the ticket receiving the most votes shall be elected President and Vice President, provided that the ticket receives 40 percent or more of the total votes cast. The winners shall take office on January 20 unless the date for inauguration is postponed in accordance with the procedures described below. All reference to “Electors” in Article II is hereby inoperative.

Section 2 Congress shall establish a Presidential Election Review Board (PERB) for the purpose of resolving disputes concerning the winner of the presidential election. PERB shall consist of three members, who shall hold their offices during good behavior. Two of the members shall be registered members of the two political parties with the greatest number of registered members nationwide, with one member each belonging to each such party. A third member shall not be a member of either of those parties. Congress shall select the original three members, with two-thirds of each House required to agree on each member. When death, resignation, or removal creates a vacancy on the Board, the remaining members shall select the person to fill that vacancy. The new member shall have the same political affiliation or lack of affiliation as the person replaced.

Section 3 In the event that no ticket receives 40 percent of the vote, a runoff will take place between the top two finishers on a date determined by PERB but not later than December 1.

Section 4 If at any time prior to January 20, one of the losing candidates petitions PERB, stating under penalty of perjury the basis for a good faith belief that the election results are or will be inaccurate, and the error(s) may be sufficient to affect the outcome of the election, PERB will review the petition. By majority vote, PERB will determine whether the alleged good faith and sufficiency of error exists and requires further action.

Section 5 Upon a finding of good faith and sufficiency, PERB may declare that, in order to have sufficient time to conduct an investigation to determine the winner, the presidential inauguration will be postponed beyond January 20. The incumbent president shall remain in office during the period beyond January 20 and prior to the new inauguration date. Thereafter, PERB will conduct an investigation, which includes full subpoena power, and resolve all disputes necessary to determine the winner of the election.

Section 6 If PERB concludes that the winner cannot be determined, it will explain the reason for that determination and call for a re-vote either of the entire nation or of select precincts in which voter tabulation problems were identified. It will establish the date of the re-vote, though this date may not be less than one week or more than one month from the date of PERB’s determination.

Section 7 If, at some point after the President has been inaugurated, PERB concludes that fraud took place sufficient to have affected the outcome of the election, it will explain the reason for that determination and, absent a determination of unusual circumstances precluding a remedy, shall call for a special election. It will establish the date of the election, though this date may not be less than one month from the date of PERB’s determination. Unless Section 8 of this amendment applies, the ballot shall include the same candidates whose names appeared on the ballot during the previous presidential election.

Section 8 The political parties may petition PERB to replace the name of the candidate in the previous election with the name of a different candidate. Upon a finding of good cause, PERB may grant the request. Good cause requires a finding that the candidate listed is unavailable through death, illness, or other extraordinary circumstance including the candidate’s removal from the ballot by PERB upon a finding that the candidate committed fraud.

Section 9 All determinations by PERB shall be binding and unreviewable in any court. However, Congress may reverse a determination by PERB if two-thirds of the members in each house find that it is arbitrary and capricious.

Section 10 All hearings held by PERB shall be open to the public.

Section 11 The Congress shall have power to enforce this article by appropriate legislation.

Obviously, the precise content of this proposed amendment can be changed in ways large and small.^. For example, while Section 11 specifically empowers Congress to implement and fund PERB, a provision might be added that would insulate PERB from politically motivated de-funding. So too, the provisions dealing with a special election or re-vote might address the treatment of absentee ballots in such elections. And, as long as we’re remaking the presidential election, we should strongly consider tweaking aspects of the system unrelated or only peripherally related to the Electoral College and PERB, such as establishing Election Day as a national holiday or moving it to the weekend, so as to facilitate greater voter participation.

At this preliminary stage, we need not get deep into those weeds. What matters is the basic purpose and substance of the proposed amendment: In addition to replacing the Electoral College with a nationwide popular vote, it would establish a permanent body fully empowered to resolve presidential election disputes. Three features of the proposed amendment may raise eyebrows, and warrant elaboration: 1) the potential remedies of a revote and special election; 2) the provision allowing PERB to push back the date of a president’s inauguration (even in cases not requiring a re-vote); and 3) the provision for Congress to override decisions made by PERB by a two-thirds vote of each House.²¹⁶

Under the status quo, a re-vote is unavailable or at least impractical. After it became apparent that the butterfly ballot in Palm Beach cost Al Gore Florida and thus the presidency, several Palm Beach residents brought suit, urging a re-vote in either Palm Beach or the entire state. A Florida court held that the Constitution forbids such a remedy, and Florida’s Supreme Court affirmed the decision.²¹⁷ Most commentators agreed with the decision.²¹⁸ While there have been re-votes on many occasions on the local level (in February 2019, the North Carolina Board of Elections ordered a new election for a congressional seat based on apparent fraud in the November 2018 election) and occasionally even in statewide elections, the presidential election is different because of the constitutional requirement that everyone vote on the same day. Article II, Section 1, Clause 4 stipulates that the day electors vote “shall be the same throughout the United States.”²¹⁹

Of course, the Constitution is referring to the votes in the Electoral College, not the votes (or appointment process) that would select electors. That distinction may collapse, however, when we consider the situation in 1787. For the founders, where the election of the presidency was concerned, the Electoral College was the main event. Everything that happened before then was mere window dressing: The electors would meet and, exercising independent judgment, cast the only votes for president. Accordingly, the Constitution’s stipulation that electors nationwide meet on the same day conceivably carries over to the votes for president that today are the main event—the vote by the American citizenry (with the Electoral College usually a formality to ratify the people’s choice).

Seen this way, the fact that re-votes have happened in many state or local elections presents no precedent for presidential elections. If a state or municipality holds a re-vote for a statewide or local office, voters still vote on the same day—twice. But if only Florida (or only Palm Beach) re-voted in 2000, that would have violated the notion of everyone voting on the same day. Besides, the Constitution does not prescribe the rules governing state or local elections. It does prescribe some rules governing presidential elections. It is at least a plausible reading of the Constitution that any re-vote has to apply to the entire nation.

There is a policy basis supporting such a requirement. If only one state holds a re-vote, voters of that state will have information that no other voters had, specifically how the other forty-nine states had voted. Had Florida held a re-vote in 2016, for example, its voters would have known that the state’s outcome would determine the nation’s. Presumably many Nader voters would have defected to Gore. Granted, this is simply one of numerous contingencies that would have arisen because of the faulty first vote, and we could not be certain which candidate would fare better in the do-over. Still, Bush would have justifiably felt cheated that Buchanan voters in states won narrowly by Gore would not get the same opportunity for a tactical change of vote.

Ultimately, however, the argument against a national re-vote, based on the Constitution, is unconvincing. It rests on the premise that the clause requiring electors to vote on the same day should today be understood to require that all U.S. citizens vote for president on the same day. But any such requirement has long been ignored by the states when it comes to absentee ballots and early voting. (All fifty states permit the former and most permit the latter.) These days, the exceptions practically swallow the rule, because early and absentee voting have become common phenomena. In 2016, the first votes for president were cast on September 29, six weeks before Election Day. Literally millions of people voted prior to Election Day.

But even if the Constitution prohibits presidential re-votes (whether partial or whole), my proposed amendment undoes the prohibition by authorizing the Presidential Election Review Board to call for a re-vote. This provision serves a major purpose. After the 2016 election, the question was asked: What if it were determined that foreign covert operations produced Trump’s victory? Worse still, what if the Trump campaign knew about or assisted such efforts? Writing for the political website FiveThirtyEight, Political Science professor Julia Aazari addressed whether a re-vote could be ordered as a remedy. Her conclusion: “So, is there a process for dealing with a finding that in essence invalidates an election? When it comes to presidential elections, the answer is: not really. . . . When it comes to the possibility that the winning side colluded with a foreign power to influence the election outcome, the Constitution doesn’t offer much in the way of a plan.”²²⁰

Aazari actually understates the constitutional omission. For one thing, the Constitution doesn’t offer anything in the way of a plan for addressing a post-election discovery of malfeasance, unless you count on electors pledged to the colluding candidate flipping (and this possibility requires that the wrongdoing be uncovered before the Electoral College meets). In addition, the Constitution’s failure is not limited to situations of candidate misconduct, let alone situations involving collusion with a foreign power. The Constitution offers no mechanism for dealing with any difficulties in determining the winner of a presidential election, whether or not involving wrongdoing by the ostensible winner.

Just as the Constitution provides no provision for a re-vote prior to a president’s inauguration, it includes no provision for a special election after a president’s inauguration. A post-inauguration special election threatens governmental stability even more than a pre-inauguration re-vote. Nevertheless, my proposed amendment includes provision for a special election, because there needs to be a remedy if we learn that an election result was inaccurate after the president has taken office.²²¹ As a federal judge put it in a case claiming fraud in New York during the 1976 presidential election, “It is difficult to imagine a more damaging blow to public confidence in the electoral process than the election of a President whose margin of victory was provided by fraudulent registration or voting, ballot-stuffing or other illegal means.”²²² So too, it is difficult to disagree with constitutional law theorist Ronald Dworkin that “it would be wrong to declare, as a flat rule, that no remedy is ever available for demonstrable and grave illegality in the electoral process.”²²³ Exactly so. A presidential re-vote or special election should be a last resort, because finality is important and the last thing we need is defeated candidates routinely demanding a doover. (Our presidential campaigns last long enough as it is!) But there may be circumstances in which no alternative is acceptable.

Under current law, when the discovery of an election-determinative impropriety occurs after the president’s inauguration, impeachment seems to be the only remedy—and wholly inadequate, for several reasons. First, the removal of the president via an impeachment trial requires two-thirds of the senate voting to convict. Since it is extremely rare for either party to control two-thirds of the senate, conviction requires a degree of bipartisanship, precisely what tends to be missing when it comes to crisis elections. Second, as noted above, the flawed election may not result from the president’s wrongdoing. Perhaps the president’s aides participated in fraud without his or her knowledge. Or perhaps the president and presidential staff did nothing wrong, but an outsider hacked the election in their favor. Impeachment makes little sense in that scenario. Third, if the president did commit fraud, impeachment still fails to resolve the underlying problem. As Professor Aazari says, “Even if collusion revelations did lead to Trump’s impeachment and removal from office, the process wouldn’t really address the question of whether his election had been legitimate in the first place.”²²⁴ Indeed, the impeachment and removal of Donald Trump would make Mike Pence president—and Pence, too, would have been a beneficiary of the fraudulent election.

For these reasons, a national re-vote and special election must be available remedies. Anyone who doubts that there should ever be such a remedy ought to consider the hypothetical supplied by election law expert Richard Hasan: “Imagine a terrorist attack on Election Day, where only part of [the country] gets to cast a vote for President, and many people stay home out of fear of additional attacks.”²²⁵ In that circumstance, it seems inconceivable that anything but a re-vote would be considered sufficient. Ditto if a hurricane or other natural disaster made voting impossible for a nontrivial portion of the electorate.

The absence of a remedy to undo an illegitimate presidential election once the president is in office has always been problematic, and has become more so in the age of hacking, which may take months or longer to detect. As the successful recounts in Minnesota in 1962 and 2008 illustrate, even when the problems are detected early, and even if no re-vote is required, the process of figuring out who won a virtual tie election can take time if it is to be done well. (The fiasco in Florida in 2000 reinforces this conclusion by way of negative example.) In part for that reason, my proposed constitutional amendment authorizes PERB to push back the date of the president’s inauguration. As Dworkin observes, “It makes no sense to demand that a breathtakingly close election be finally decided by any magic dates in December in order that a new president be chosen by January 20.”²²⁶

The stipulation in the proposed amendment that the incumbent continue in office until the dispute is resolved mirrors a provision in Minnesota’s constitution that provided stability during the long period of uncertainty following the 1962 gubernatorial race. The Minnesota Constitution provides that the term of the incumbent governor carries over until a replacement is “chosen and qualified.” Accordingly, Governor Anderson served well into March until it was determined that he had been defeated.

Finally, I should say a word about the stipulation in the proposed constitutional amendment that a two-thirds vote of both houses of Congress may reverse a PERB determination. Although the composition of PERB ought to safeguard against partisanship, no governmental entity should be immune from checks and balances. The precise mechanism chosen here, the two-house two-thirds requirement, has precedence in the Constitution, specifically the Twenty-Fifth Amendment, which permits the temporary removal of the president if the vice president and a majority of the cabinet determine that he or she is unable to discharge their duties. The president may subsequently reclaim the powers of the office, unless the vice president and a majority of the cabinet protest and two-thirds of both houses of Congress determine that the president remains unfit. The Presidential Election Review Board’s determinations about an election, like that of the president concerning his own fitness, deserve a strong favorable presumption. Since it is rare for either party to dominate both houses, some degree of bipartisanship will generally be needed for PERB’s determination to be rejected.

While I have coupled my two proposals into a single constitutional amendment, for simplicity’s sake and because they each reduce the risk of a post-election crisis, the two central reforms are not dependent on one another. Abolishing the Electoral College would, by itself, reduce the likelihood of an election called into question by alleged fraud, even if we did not establish something like a Presidential Election Review Board to address such crises as do arise. Conversely, if we abandon the goal of abolishing the Electoral College, the remainder of my proposed amendment, establishing PERB to resolve disputes, would, with some tweaking, stand on its own. Creation of a PERB would be a major step forward whether or not we retain the Electoral College.

Readers may have misgivings about many of the specific provisions I have proposed related to PERB, but there is no good reason to resist the creation of some version of PERB. No one, Democrat or Republican, people in small states or large, should oppose the establishment of a body to deal with presidential election disputes in a fair and timely fashion.