Following up on a post I wrote in 2020, there’s been some discussion that the 2024 presidential election could be thrown to the House of Representatives.

But how does this work? And how likely is it?

It happens when no candidate wins a majority of the Electoral College. The House (the new House, which is sworn in January 3, then joins the Senate to count electoral votes on January 6) then chooses among the top three vote-getters, with each state casting one vote. Vermont gets one vote; California gets one vote. It takes a majority (here, 26 states) to choose the next president. As occurred historically, a tie vote in a state’s delegation (say, one state with 26 representatives is 13-13) did not count toward that majority.

A so-called “contingent election” is rare. It happened in 1800, when it took 36 votes in the House before Thomas Jefferson was chosen as President. Since the Twelfth Amendment, it happened only once, after the 1824 election, when Andrew Jackson won a plurality of the electoral vote but John Quincy Adams carried the House vote in the contingent election. (The vice presidential election of 1836 was also thrown to the Senate.)

So, how can it happen? Briefly, there are four ways to look at it (with some sub-scenarios noted, and some even more unusual outliers omitted).

First, there’s a tie in the Electoral College. The 12th Amendment requires that the president receive “a majority of the whole number of electors appointed.” If two candidates tie at 269-269, the House chooses between those two candidates.

When did this last happen? Since the Twelfth Amendment, which separated presidential and vice presidential votes, never. (A tie did happen in 1800 under different presidential election rules.)

How is this scenario possible? Presidential electors are based on each state’s Senate and House delegation. The Senate always has an even number of Senators, because each state receives two. The House is currently fixed at 435 members, an odd number. The District of Columbia is guaranteed a number of presidential electors, but no more than the least populous state (which, at the moment, is three, two senators plus one representative). Even number (Senate) + odd number (House) + odd number (DC) = even number. (We could increase the size of the House to 436….)

What are the odds? Very low. Nate Silver, for instance, puts the odds of no one winning a majority (which includes a tie scenario and others, below) at less than 1 in 200.

Second, three or more candidates divide up presidential electoral votes on Election Day, and no one gets a majority.

When did this last happen? In 1824, four presidential candidates divided up the electoral votes. The top three vote-getters were then sent to the House. While Andrew Jackson had 99 electoral votes to John Quincy Adams’s 83 (but neither crossed 131, the number needed for a majority), the House ultimately backed Adams.

How is this scenario possible? It requires one or more strong third-party candidates. Because all states award their presidential electors on a winner-take-all basis (except for Maine and Nebraska, where it’s winner-take-all for two statewide, and winner-take-all, or really winner-take-each, per congressional district), a third-party candidate would need more votes than both the Republican and Democratic candidates in a state, and end up depriving those candidates of a majority in the Electoral College. Strong third-party candidates like Ross Perot in 1992 won zero electoral votes. Others, like George Wallace in 1968, won a substantial number of electoral votes but did not deprive a candidate of the majority in the Electoral College.

What are the odds? Very low—and here I’ll add, exceedingly low. I’ve seen no state polling to suggest that any third-party candidate is close to exceeding even 5% of the statewide popular vote total in any state. That is particularly true since the once-viable Robert F. Kennedy Jr. has withdrawn from the race.

Third, faithless presidential electors deprive the apparent majority-winner of electoral votes, and no one gets a majority.

When did this last happen? Never.

How is this scenario possible? When we vote for the president and vice president on Election Day, we’re actually voting for a slate of presidential electors who later formally vote for president and vice president. Those electors are typically “faithful” to the popular will of the people of the state. Sometimes, however, electors vote for someone else, behaving as “faithless” electors. Historically, however, faithless electors have been rare. They’ve been rarer still to defect from the majority vote-getter to someone else. A few years ago, the Supreme Court in Chiafalo approved of state moves to cabin the discretion of presidential electors. And the faithless electors in 2016 likely have prompted the parties to look closely at how they vet electors in 2020. So there aren’t many faithless electors in the first place, and by law there cannot be in some places.

What are the odds? Very low, likely exceedingly low. It requires a narrow electoral vote victory (certainly plausible), and faithless electors denying a majority (not just faithless electors generally). If the margin is 275-263, for instance—a seemingly-close election—you’d need six faithless electors from the winning candidate to defect to someone else. And Chiafalo reduces those odds further.

Fourth, the House rejects some number of electoral votes that denies any candidate of a majority.

When did this last happen? Never.

How is this scenario possible? There are a few ways it could occur, because Congress formally counts the vote. A state might not send electors to Congress (which seems unlikely in the present day). It could send a slate of electors, but Congress refuses to count them (like it did in Louisiana in 1872)—perhaps because Congress deems the votes not “regularly given.” It could choose between multiple slates of electors (which has only happened once since 1876, Hawaii in 1960), and end up with a tie (which has never happened). One wrinkle is that the winning candidate under the Twelfth Amendment is among those from “a majority of the whole number of electors appointed” (emphasis added). If Congress throws out votes, it depends on whether it includes those votes in the denominator of “electors appointed.” (In other words, if it throws out 20 electoral votes, it might go from a majority requirement of 270 of 538, to 270 of 538—with 20 votes not counting—or, instead, and easier to meet, 260 of 518.) In short, there are a number of complicated things that might happen in Congress, which would be unusual and historic decisions in Congress.

Congress has also clarified under the Electoral Count Reform Act that is a state refuses to send its electors to Congress or if its votes are not “regularly given,” those votes also come out of the denominator. In other words, it becomes harder to “throw” the election to the House on this basis, because you would need to have a much larger margin. It is instead more likely that one candidate or another simply holds a majority (unless they happen to tie). Throwing out the winning candidate’s votes because Congress concludes the candidate is not qualified, however, creates the greater risk of an election being sent to the House.

What are the odds? Well, again, likely on the verge of exceedingly low. It takes a majority of both houses of Congress to vote to throw out votes. Congress has been reluctant to throw out electoral votes since Reconstruction. Recent attempts to do so in Congress—Florida in 2000, Ohio in 2004, and a slate of states in 2016 and in 2020—lacked even modest support in Congress (only Ohio 2004, and Arizona and Pennsylvania 2020, went to a congressional vote, where objections were soundly defeated and the votes were counted). It all depends on the complexity of the decisions left to Congress and Congress’s trust in the electoral outcomes—all with the added condition that these decisions ultimately deny any candidate a majority.

*

In short, the 2024 presidential election could end up in the House, as it could have done for any presidential election in the last 200 years. There are plausible reasons to think up any one of these scenarios. But, I think, it’s worth thinking through, with some precision about these scenarios, and how they may or may not occur.