It’s May of 2024 and former Vice President Mike Pence appears to be the likely Republican nominee, despite Donald Trump’s victory in the Iowa caucuses only a few months earlier.  He comes out just ahead of the former president as the Republican Party’s establishment manages to rally together around a unifying candidate who talks about something other than stolen elections and Trump’s supporters fail to turn out, facing headwinds voting in a number of states that have made casting a ballot more difficult. Pence begins the gruelling march towards the general election in November, where he will face incumbent President Joe Biden, who cruised to renomination as he was unopposed in the Democratic primaries, many of which were canceled.

As the August RNC, where Pence will officially accept the GOP’s nomination and name former UN Ambassador Nikki Haley as his running mate, approaches, rumors have been building that a disgruntled Trump will counterprogram Pence’s speech with one of his own. And sure enough, on that fateful day in August in a speech at Raymond James Stadium, Trump announces he will run for president on a newly formed Patriot Party ticket with South Dakota Governor Kristi Noem as his running mate. 

This is the scenario I ask you to imagine as we conduct this exercise on an alternate political future, one in which Democratic President Joe Biden is challenged by Republican Mike Pence and the Patriot Party’s Donald Trump, each on the ballot in all 50 states, the District of Columbia, and the five congressional districts that vote independently of their statewide results.1Maine and Nebraska’s congressional districts cast votes for president separately. The last time something comparable happened was 1992, when President George H.W. Bush was challenged by Democrat Bill Clinton and independent Ross Perot. Clinton won the election with 43% of the vote nationally, with Bush at 47% and Perot at 19%. Perot won no electoral votes. But another comparison would be the 1912 election in which incumbent Republican President William Howard Taft was challenged not just by Democrat Woodrow Wilson, but by his Republican predecessor too. Now running as the Progressive (or “Bull Moose”) nominee, former president Theodore Roosevelt decimated Taft’s incumbency advantage and surpassed his vote margin, carrying several states and around 27% of the vote (Taft — the incumbent president, may I remind you again — casrried only two states and just 23% of the vote!) This enabled a blowout Wilson victory, despite the fact he only carried 42% of the vote, and the prospect of a former president challenging his own party makes for a slightly more apt comparison, hinting at the results you might expect.

So in 2024, imagine a former president incredibly popular with his base, an incumbent president carried to victory four years prior on the backs of independents and moderates, and a mainstream GOP candidate divorced from the driving force in the party the last eight years. How does that play out? 

The Model

Our best base-level guidelines for state-level vote share for each candidate in this three-way race are a confluence of six variables:

1. Biden’s 2020 vote share in each state: simple enough, this is the percent of the vote Biden got in every state in the 2020 election.
2. Trump’s 2020 vote share in each state: percent of the vote Trump got in every state in the 2020 election.
3. Biden’s 2020 vote share overperformance compared to Democratic congressional candidates: Biden won nationally by 4.5%, but House Democrats won nationally by 3.1%. This 1.4% margin of overperformance by Biden seems minute, but almost certainly won him the election — Georgia, Arizona, Wisconsin, and Pennsylvania, which all went for Biden (and delivered the election for him), were won by less than 1.4%. Had Biden been a “generic Democrat” he likely would have won the most votes but lost the Electoral College. This also helps us unpack the inverse — are there Trump-sceptical Republicans who voted for Republicans in House races but didn’t support Trump?
4. Partisan lean: the partisan lean of each state and district, reflecting the leaning of a state in a neutral political environment, courtesy of FiveThirtyEight.
5. Trump’s PARP score: this stands for “popularity above replacement president,” a handy metric FiveThirtyEight created to illustrate where the president was under- or overperforming his predicted popularity. For example, Trump was always uniquely weak in states like Kansas and Utah, which are heavily Republican states, but where his net approval was only 2% and 5%, respectively. This gives us a guideline of how many voters in states might be conservative but opposed to Trumpism. This is calculated by taking the net approval of Trump in each state plus the state’s partisan lean. 
6. Republican approval of Trump: an average over 2017-2021 of Republican approval of Trump, this is 86%.

With these six variables we can assemble the model of vote share in each state:

• Biden’s vote share is calculated as his 2020 vote share in each state minus his overperformance compared to the generic Democrat. Critically, Biden is only harmed by this model. He can never get a higher share of the vote than he did in 2020.
• Trump’s vote share is calculated by first finding his average approval amongst Republicans (0.86) plus his PARP, then multiplying that number by his 2020 vote share in each state. Trump has the opportunity to exceed his 2020 vote share if his PARP score in any given state is above 14% (since retaining 86% of his voters is the default assumption based on his popularity among Republicans). He must be demonstrably more popular than expected given a state’s partisan lean to do so, which he is not in any state — he is far more popular than expected in Hawaii, but the PARP there is only around 10.7%, so he’s still losing voters to the Republican candidate overall.
• Finally, the GOP’s share of the vote is calculated by subtracting the 2024 Trump vote share from his 2020 vote share in each state, before adding Biden’s overperformance of 1.4% (as these are presumably voters who picked Biden for president over Trump, but still voted Republican for Congress). The GOP carries more voters when a state is Republican leaning but Trump is far more unpopular than expected in the state, suggesting a larger share of Republicans would prefer a more mainstream candidate.

For the Nebraska and Maine congressional district vote shares, and DC vote share, for which PARP data was unavailable, a PARP of zero is assumed, so Trump carries 86% of his vote share in 2020, with the other 14% going to the GOP candidate.

The Results

It is perhaps not surprising if you know anything about third parties in the US, but Biden easily carries this election because the GOP splits votes away from Trump. You can click on the map to see the results in each state or district: