I’ve posted on SSRN a substantially revised version of this article, which will be published in the University of Illinois Law Review as part of a symposium issue. The revisions reflect the many extremely helpful comments I’ve received on earlier drafts as well as my latest thinking on what forms of ranked-choice voting and other electoral methods would be best for electing US senators, governors, and other statewide offices. I very much continue to welcome comments as, not only will this article undergo additional revisions during the editing process, but also it serves as a springboard for additional work on this topic, including an eventual book. Here’s the current abstract for this piece:

The standard system of statewide elections for governor and US senator, among other offices, deserves a thorough overhaul. The collection of signatures by candidates to qualify for the ballot, currently confined to antiquated pen-and-paper technology, can be modernized and put online, so that it can function as a kind of “approval voting” system that yields a reasonable number of candidates (five, for example) for a primary election ballot. Likewise, online party conventions can enable parties to endorse candidates before the primary occurs and to have their endorsed nominees qualify for the government’s “All Qualified Candidates Primary” ballot. Moreover, innovative forms of Ranked Choice Voting—like “Optimal Tournament Voting”—can be used to identify the two candidates on the primary ballot most suitable to advance to the general election, with suitability for this purpose determined by which candidates are most representative of the whole electorate.

Alternatively, even without ranked-choice ballots, the mathematical principles and procedures of Optimal Tournament Voting can be used to create a “top three” general election, in which voters directly express their preferences between each pair of the three candidates who advance to the general election based on the use of approval voting in the primary (or online signature-gathering). Moreover, whether the general election has two or three finalists on the ballot, “fusion voting” can be employed in the general election to permit parties whose nominees are not one of the finalists to renominate whichever finalist they prefer. States should experiment with these innovative alternatives and other variations along the same lines so that elections for statewide office, like governor or US senator, will produce winners who are the candidates most preferred by a majority of the electorate’s voters.

Link to video announcement. Is he referring to possible No Label’s presidential ticket, accentuating the concern of a potential spoiler effect? Or is he referring to a broader effort to create some sort of centrist third-party movement, of the type that Mitt Romney said he’s talked to Manchin about (according to the new biography of Romney)? Whichever it is, the conditions of polarization (and fragmentation) between (and within) the two major parties has reached a point that I don’t think has existed for many decades. What will be the upshot of this dynamic is anyone’s guess.

As readers of this blog know well, the potential role for an effort to sustain a centrist third-party movement depends greatly on what particular rules govern the operation of an electoral system. Without a move to ranked-choice voting, or fusion voting, or some other type of electoral reform, it seems unlikely for a third-party movement to have a lasting effect except perhaps to contribute to some sort of realignment involving the two major parties (either the same two, or perhaps the replacement of one or both, as when the Republicans replaced the Whigs–consider also the effect of the Progressive Party on the changing natural of electoral competition between Democrats and Republicans during the Progressive Era).

The recent polling by Siena College, in partnership with the New York Times, has highlighted the risk that Robert Kennedy, Jr. will play the role of a spoiler next year–and that’s even without the addition of a potential No Labels candidacy. The polling shows Kennedy with almost a quarter of the vote in the six key battleground states–and also showed that these voters, if faced with only a binary choice between Biden and Trump, would end up voting for one or the other rather than abstaining. As the Times summarized its analysis of the numbers:

“The polls found that he pulled similar numbers of voters from Mr. Biden (21 percent) and Mr. Trump (23 percent), but the percentages varied by state. In narrowly divided contests, Mr. Kennedy could have the potential to swing the outcome. Mr. Kennedy’s presence helped Mr. Biden in Nevada and Pennsylvania, but aided Mr. Trump in Georgia, the polls found.

“While Mr. Trump beat Mr. Biden in a two-way contest in Arizona and Pennsylvania, those states were a tie when the polls asked voters to also consider Mr. Kennedy. Mr. Trump’s lead in Georgia increased by a percentage point with Mr. Kennedy in the race, but in Wisconsin, Mr. Biden’s advantage remained the same — two points — when Mr. Kennedy was included.“

Given the uncertainty over whether Kennedy hurts Trump more than Biden or vice versa, I wonder whether there is any chance for new laws in these battleground states that would eliminate the risk of a spoiler effect either way. While Ranked Choice Voting would be the most obvious and straightforward way to eliminate the risk of a spoiler effect in the appointment of a state’s presidential electors–as both Maine and Alaska have done–it’s not the only way to do so, as I discuss in my book Presidential Elections and Majority Rule. It would be possible, for example, for states to hold a preliminary general election vote after the conventions but before November (in September, for example), which would have Kennedy and other third-party candidates on the ballot along with the major-party nominees, but then narrow the field to only two finalists for the November vote (in the same way that California does for other offices in its “top two” system). This would be essentially equivalent to how France conducts its presidential elections, in a two-round system, and would give voters the chance to express their preference for Kennedy (or other third-party options) before having to decide between, from their perspective, the lesser of two evils in November.

Another idea, which would be similar to Ranked Choice Voting but not exactly the same, would be to let voters in November indicate both a first and a second choice. If their first choice didn’t win, their second choice would be counted instead. Given the willingness of most Kennedy supporters to indicate a backup preference between Biden and Trump, enabling these voters to cast a ballot expressing this backup preference would go a long way to preventing a spoiler effect from a Kennedy (or other third-party) candidacy.

While it seems hard to imagine that any of the battleground states would be willing to change their laws for appointing their presidential electors in order to eliminate the risk of a spoiler effect next year, the issue is important enough that it is worth pursuing if there is any chance of this reform at all. The idea that a candidate can win all of a state’s electoral votes as a result of winning less than 50% of the state’s popular vote is antithetical to Jeffersonian revision of the Electoral College system adopted in the Twelfth Amendment, as I explain in Presidential Elections and Majority Rule. More importantly, the ability of a candidate to win an Electoral College majority as a result of sub-50% popular-vote victories in one or more states has led to presidencies that have significantly altered the course of history without receiving the kind of majority support from the electorate that the Twelfth Amendment was designed to achieve. For those unfamiliar with the presidential election of 1844 and its consequences, as I was before researching and writing Presidential Elections and Majority Rule, it is a very sobering story–and a warning of what might happen if the winner in 2024 achieves an Electoral College majority only as a result of a spoiler effect occurring in one or more states.

Adding briefly to what Derek already has blogged about today’s decision by the Minnesota Supreme Court, I think the distinction that the court drew between primaries and the general election is important. Since writing on this topic in the Washington Post a few months ago, I have thought it unnecessary (even if desirable) for the disqualification issue to be settled before the primaries. On the other hand, I’ve thought it crucial that the U.S. Supreme Court settle the issue before the November general election–and, as a practical matter, that should mean before the Republicans hold their nominating convention in July.

Today’s decision rules that the issue of whether Trump can be on the general election ballot in Minnesota is not yet ripe because Trump has not yet been nominated by Republicans as their candidate. What is potentially worrisome about this ripeness ruling, as Derek suggested, is that the court might not consider the issue ripe until after the GOP’s nominating convention. But delaying the decision until that point would be problematic because it would make it more difficult for Republicans, as a party, to deliberate about what to do if Trump were disqualified after the convention but before the November general election ballots are printed.

It would not be impossible for the party to act during this period. Parties have procedures to replace nominees on an emergency basis if something happens after a convention but before the election–like, tragically, the death or incapacitation of the candidate. In 1972, for example, after the Democratic convention nominated Thomas Eagleton for Vice President, Eagleton withdrew when it was disclosed that he has received psychiatric treatment for depression. The party quickly substituted Sargent Shriver as its VP nominee.

Nonetheless, it would be far better to give Republicans the chance to deliberate at their convention about who should be their presidential nominee if Trump is disqualified. Thus, hopefully, the Minnesota Supreme Court could be persuaded to consider the issue ripe well before the convention, in order to give the U.S. Supreme Court time to rule on the issue also. Today’s ruling, which mostly focuses on the reasons why the Minnesota Supreme Court does not need to decide the disqualification issue before the primaries, need not be read as foreclosing the possibility of a pre-convention resolution of the issue–especially if the plaintiffs refile their case, as today’s order permits them to do, with a persuasive explanation of what an an appropriate pre-convention timetable for decision should be.

Of course, it is possible that the Minnesota Supreme Court may construe existing state law as precluding any pre-convention resolution of the issue. While that would be a shame, it would be an example of what I warned about in my Washington Post piece: the possibility that existing state law may not be well-suited to the adjudication of a disqualification issue under section 3 of the Fourteenth Amendment and thus it may be necessary for a state legislature to amend existing state law in order to provide an explicitly appropriate procedural vehicle for settling the matter before the GOP convention. If Minnesota is unable to provide this procedure, perhaps some other state can.

Rick linked earlier today to a new analysis of the risk that a No Labels presidential ticket could cause no candidate to win a majority of electoral votes, triggering the Twelfth Amendment’s antiquated “contingent election” procedure (last used in 1824). I’m inclined to think it unlikely that the No Labels ticket would win any electoral votes; after all, Ross Perot didn’t in 1992 despite winning almost twenty percent of the national popular vote. Much more likely, as been widely discussed, is that the No Labels presidential candidate will take enough votes from Biden to cause Trump to win an outright Electoral College majority when otherwise Biden would have.

Still, the risk of a contingent election is not zero, and it’s worth considering what might unfold if it were to occur. Most commentators who have considered this assume, as does today’s new analysis, that Trump would win the contingent election in the House because the Twelfth Amendment provides that each state’s House delegate has one vote in this anomalous procedure, and Republicans in the House can be expected to hold a majority of seats in a majority of state delegations (26), which is the numerical threshold that the Twelfth Amendment requires. This assumption is hardly unreasonable and indeed the most probable result if the House were to required to go down this road. But there are a number of uncertainties that cloud any attempt at prediction.

For one thing, Democrats might be the majority party in the House on January 6, 2025, even Republicans have the majority of seats in 26 state delegations. As the majority party, Democrats might be able to establish rules for conducting the contingent election that could complicate the proceedings. For example, suppose the House adopted a rule that required each state delegation to vote by secret ballot when making its choice among the three candidates in contention: Biden, Trump, and the No Labels candidate. If No Labels nominates a non-Trump Republican for the top its ticket, as they have indicated they will, it’s possible that this secret ballot vote could fracture Republicans in the House in the same way that House Republicans have been fractured in their secret ballot votes for a new speaker. If House rules require a candidate to receive a majority within a state delegation in order to receive that state’s single vote, this kind of fracturing could prevent any candidate from reaching the constitutionally necessary majority of states–in which case the Twentieth Amendment would provide that whomever the Senate elected as Vice President would serve as Acting President until 26 state delegations in the House were able to agree upon a candidate. By contrast, if House rules permitted a candidate to win a state delegation’s single vote by receiving a merely a plurality, and not a majority, within the state delegation, then it’s at least conceivable that Biden might be able to win the contingent election in the House in the same way that Hakim Jeffries would be the new House Speaker if only a plurality rather than a majority were required (because of the same kind of fracturing among House Republicans).

What is more, suppose the U.S. Supreme Court hasn’t definitively settled the merits of the issue whether or not Trump is disqualified under section 3 of the Fourteenth Amendment. Suppose, again, that the Democrats are the majority party in the House. Would they, by House rule, be able to disqualify Trump as one of the candidates entitled to be considered in the contingent election even if the Senate did not concur in this judgment of disqualification? The text of the Constitution is unclear on this point, and the new Electoral Count Reform Act does not purport to address this particular situation. Would the Supreme Court adjudicate this constitutional question, or invoke the “political question doctrine” to rule it beyond its authority and within the exclusive control of Congress itself?

These uncertainties are just some of the potential pitfalls that could plague a contingent election if it happened. The group Unite to Protect Democracy has produced a report that discusses some of the other dangers that could arise. One is that some House seats necessary to determine which party has a majority could themselves become mired in contestation over which candidate won, preventing organization of the House and election of a new Speaker–and thus lead to the incapacity of the House to conduct the contingent election at all.

The basic point is that seeking to provoke a contingent election, as No Labels now apparently wishes to do, is truly playing with the proverbial fire. We can speculate but cannot really know what would happen. In addition to all the uncertainty in Congress itself, there would also be possibility of severe turmoil in the streets.

It is certainly understandable why No Labels would want to field a third presidential ticket next year, especially with Biden and Trump as the two major-party nominees. But the unavoidable reality, as I’ve said many times before, is that our presidential election system is not currently equipped to handle more than two candidates. Unless and until states introduce ranked-choice voting, or some other form of majority-winner requirement, for the appointment of presidential electors, the only realistic role for third or additional candidates is to undermine the capacity of the system to make an accurate determination of which major-party candidate is truly preferred by a majority of voters in enough states to form an Electoral College majority. Given the stakes of the 2024 election, next year would be the worst possible time imaginable for a third-party ticket to undermine the existing system in this way.

In light of the removal of McCarthy as Speaker, last week I invoked my previous advocacy of a majority winner requirement for congressional elections. This week I want to highlight my more recent advocacy of the need to avoid plurality winners in partisan primaries, a point also made by Yuval Levin this morning in the New York Times. Moreover, FairVote has posted an analysis showing that six of the eight Republicans who voted to vacate McCarthy from the speakership won their first partisan primary election with less than a majority of votes. Gaetz himself barely won more than a third (36%) in his first primary.

As Levin argues, one way to ameliorate this for for the Republican Party itself to adopt ranked-choice voting for its own primaries, as it did in the primary that nominated Glen Youngkin for Virginia’s governor (a general election he subsequently won, as we know). Whether states, or Congress, can mandate political parties to use RCV for their primaries is an open question under the U.S. Supreme Court’s First Amendment jurisprudence in light of cases like California Democratic Party v. Jones, although a lower federal court rejected the Republican Party’s challenge to Maine’s requirement that the party use RCV for its primaries. But without violating the First Amendment, Congress could require members of Congress to win either a primary or general election by means of a majority rather than plurality vote, leaving to the states and the parties whether they wanted to enforce this rule in the contexts of primaries or general elections, or both, and whether they wanted to do so by means of a traditional separate runoff or instead a form of ranked-choice voting. That would provide the maximum flexibility for state-by-state experimentation while still avoiding the “double whammy” of a “double plurality winner system” as I describe in my Washington Post piece last year.

A couple of astute readers of my earlier blog post today asked me whether Georgia’s version of a runoff system would satisfy my proposed majority-winner requirement for congressional elections. The simple answer is no, for a reason that is explained in the law review article containing the text of the proposed statute, but which I should have made clearer in the blog post itself.

My proposed statute would not permit a regular runoff after Election Day in November, the way Georgia does it. Instead, it would require the majority-clearing vote to occur on Election Day itself, either in the form of a ranked-choice ballot or as the second round of a two-round system, like California’s. Although mathematically Georgia’s post-November runoff is equivalent to California’s second round in November, the posture of the parties and their nomination process changes if the government requires the majority vote in the second round of a two-round process. 

Under my proposed statute, any candidate regardless of party would need to participate in the first round of a two-round system, if the state chose that approach instead of a ranked-choice ballot. Therefore, both the ultra-extreme candidate and the less extreme candidate would need to be on the first-round ballot in order to be eligible to qualify for the second round in November. If it wanted to, a party could still hold some sort of primary election or other nomination process in advance of the first round of this two-round system, and thus could nominate the ultra-extreme candidate in that preliminary process. But it wouldn’t block the less extreme candidate from being able to pick up votes across the entire electorate.

To be sure, even under this system, depending on the composition of the specific electorate, there is the possibility that the less extreme conservative will come in third, behind both the ultra-extreme candidate and the Democrat. This is true for both California’s “top two” system, and Alaska’s “top 4” with its “instant runoff” method of RCV. To avoid that problem, a state would need to adopt a “Condorcet-compliant” electoral system, which I think would be better for reasons I’ve elaborated in various pieces, including a new co-authored one. But I don’t think it would be appropriate, or politically feasible, for Congress to mandate a Condorcet-compliant electoral system. Instead, I think the most Congress could, and should, do is mandate a majority-winner system along the lines I’ve suggested and then hope that some states are willing to try a Condorcet-compliant version of RCV as a way to best comply with the congressional mandate. 

To be clear, however, just because some House districts would need a Condorcet-compliant system in order to avoid the election of the most extreme candidate, not all House districts would. It depends on whether a less extreme conservative can come in second at least, leaving the Democrat in third place–as was true in Lisa Murkowski’s win under Alaska’s new system (Murkowski actually came in first, barely ahead of her Trump-endorsed opponent, with the Democrat trailing far behind).

Finally, here’s the text of the proposed statute:

Section 1: For any election to a seat in Congress, the winning candidate shall receive a majority of votes in the November election.

Section 2: To implement the requirement set forth in section 1, states may choose either (a) to use conventional single-preference ballots, on which voters mark only a choice for their most preferred candidate, and to hold the election in more than one stage, in order that the final stage is confined to two candidates from whatever larger set of candidates appeared on the ballot at any earlier stage; or (b) to hold an election, of however many stages the state chooses, using ballots that permit voters to rank their preferences among candidates, from which a final result may be calculated mathematically that yields a single winning candidate preferred by a majority of voters. 

As I’ve watched this week’s removal of McCarthy from the Speakership and the ensuing sense that the House may be ungovernable because of the fragmentation within its Republican conference, I return to the idea that I first advocated shortly after the January 6 insurrection: a simple statute requiring majority, rather than plurality, winners for congressional seats.

In the deep-red districts that produce members of Congress like Matt Gaetz, along with the others who joined his motion to vacate the Speakership and who keep the chamber from being functional, a simple majority-winner requirement potentially could make a big difference. A conservative candidate who was not as extreme as Gaetz or his cohort of ultra-extremists would not need to win a Republican primary in this kind of deep-red district in order to be competitive in the general election. Indeed, this less-extreme conservative would need only to come in second in the general election (in terms of first-choice preferences among the voters) in order to force some sort of runoff, “instant” or traditional, and then add support from voters whose first choice was a Democrat (or other further left candidate) in order to win the runoff. The deep red district would still get a very conservative Representative in the House, aligned with the preferences of the district’s median voter–but it wouldn’t be an ultra-extremist like Gaetz who is even further to the right of the district’s median voter but wins the plurality-winner general election after winning the GOP primary. Because there is no way to force a runoff (again, instant or traditional) in a plurality-winner system, it is useless for the less extreme conservative who better represents the district’s median voter to run in the general election. The voters are then faced with the choice between the ultra-extremist (like Gaetz) and a Democrat (or someone else too far to the left for the district’s median voter), and so the ultra-extremist goes to Congress instead of the less-extreme conservative whom the district voters actually would prefer.

A simple majority-winner requirement would let states decide for themselves how best to implement it. They could chose the kind of “top 2” system that California and Washington State use. They could also embrace Alaska’s new “top 4” system that uses an instant runoff in its general election, or the “final 5” variation on that same system, as Nevada may do if the reform measure its voters approved last year gets a second vote of approval as required by the state’s constitution. Or states could experiment with other innovative forms of Ranked Choice Voting that arguably are even more effective at combatting polarization and extremism than the particular “instant runoff” procedure that Alaska employs.

But whatever particular form of a majority-winner system that a state adopted, it would eliminate the critical defect of the existing plurality-winner system and enable less extreme candidates have a fighting chance in the general election.

To be sure, states can enact a majority-winner system for congressional seats on their own, without a congressional mandate. That, after all, is what the above-mentioned states like California, Washington, and Alaska have done. And given the current dysfunctional nature of Congress itself, it would seem foolish (or at least naive) to expect Congress to enact even a simply remedy that would ameliorate its own dire condition.

Still, it’s worth contemplating the idea of a discharge petition in the House, supported by Democrats and those nervous Republicans who represent districts that Biden won, requiring a vote in the House on a super-simple, one-page statute (see the example on page 399 of this law review article) that would mandate majority-winner elections for congressional seats. (Maybe even if the Speakership remains vacant long enough, interpretation of House rules would permit the Speaker pro temp to entertain a discharge petition of this type.) Even just having a debate on this proposal would focus attention on the best way to get out of the mess that the House currently is in. And who knows, if the Senate picked up the idea as well, it is the kind of reform–because it permits such wide variation of choices for states to adopt–that conceivably could reach the 60-vote threshold for cloture there.

When Americans focus on the distinction between majority and mere plurality winner rules, they understand why it’s desirable to require a majority rather than just a plurality in order to win an election. If it were possible to have a meaningful debate in Congress itself on the desirability of this simple proposal, it is likely that it could gain traction.

The current crisis is so severe, it seems an idea worth considering rather than giving up as hopeless.

Scott Anderson of Lawfare led a thorough and well-organized discussion of the various procedural hurdles and obstacles associated with litigation over section 3 of the Fourteenth Amendment, with Derek and me as participants. I think it would be particularly useful for those who’d like to follow the ins and outs of this convoluted topic.