Trump defied parallels to the politics of the 1890s. Like Nixon and Henry Clay, he did best in his third try.

One of the most impressive and surprising things about the scale of Donald Trump’s victory in 2024 is that he managed to win over a bunch of voters who had not voted for him in two previous runs for president. For me, and for many other critics of Trump, running him again and expecting a different result from the 46 percent of the vote he garnered in 2016 and 2020 seemed to fit the proverbial definition of insanity. Indeed, there were historical precedents for divisive populist candidates subtracting more and more people from their coalitions over time. The obvious distinction this time was simply that Democrats drove those voters away. My own theory all along was that neither Trump nor Joe Biden could win — the best either could do was benefit from just being there while the other blew the election (a dynamic that wasn’t really changed when Kamala Harris was substituted for Biden without distancing herself from his record).

The weakness of Biden and then Harris is why there was not a lot of surprise by the time Trump won, notwithstanding the issues that had limited the reach of his appeal in two prior bids. But carrying the popular vote by 2.28 million votes (nearly 5.5 million outside of California) was a stunning turnaround from his previous races. How big? The raw numbers are one thing: Trump in 2016 won 62.9 million votes, barely more than George W. Bush twelve years earlier and the fourth straight election in which Republicans won between 59.9 and 62.9 million. But he then grew his electorate to 74.2 million in 2020 and 77.3 million in 2024. The 23 percent growth since his first race and 27 percent growth in the Republican vote total since Mitt Romney isn’t a record: George W. Bush grew his vote total 23 percent from 2000 to 2004, and 58 percent compared with Bob Dole in 1996.

But moving from vote totals to shares of the voters and of the voting populations, I decided to look at the prior subset of candidates in the popular-vote era (since 1828) who ran, failed to win a national popular majority, and ran again. Trump is the 13th of those, of whom five ran three times. I used two metrics. One is growth in the percentage of the vote. The other is candidate share of VEP (voting-eligible population), for which I take Professor Michael McDonald’s collection of data on the share of all eligible voters who turn out in an election and multiply it by the candidate’s share of that vote. So, measured by VEP, we can see that Trump got 27.6 percent of all eligible voters in 2016, 31 percent in 2020, and 31.8 percent in 2024. If you’re wondering, that 31.8 percent was the twelfth-highest since 1920 (when women’s suffrage doubled the size of the electorate) and the sixth-highest for a Republican in that period, behind the two Dwight Eisenhower campaigns, Richard Nixon in 1972, Herbert Hoover in 1928, and Ronald Reagan in 1984:

Overall, Trump places 42nd on the all-time list, given that voter turnout was highest in the era that began with the mass political parties in 1840 and concludes with first the imposition of Jim Crow and then the expansion of the franchise to women. Ulysses S. Grant in 1868 was the all-time winner of the biggest share of all eligible voters, followed closely by William Henry Harrison in 1840, Samuel Tilden (the loser of the 1876 race), and Abraham Lincoln in 1864. Of course, both Lincoln in 1864 and Grant in 1868 benefited from the fact that seceded Southern states were out of the electorate in those two elections.

Relatedly, there are some methodological problems in comparing elections over time when they are not conducted on a nationwide basis. Lincoln, for example, won 39.65 percent of the national popular vote in 1860, but he wasn’t even on the ballot in most of the states that seceded rather than accept him as president. Adjusting for those, he won 48.48 percent of the votes in 1860 in the states that stayed in the Union and voted again in 1864. So, while he expanded his share of the overall vote by 38.8 percent, if you run something like an apples-to-apples comparison by excluding the states that dropped out in 1864, the growth of the Lincoln vote is an impressive but less eye-popping 13.5 percent.

I used a similar adjustment for William Henry Harrison because of the Whigs’ idiosyncratic and never-repeated strategy of running favorite-son candidates in multiple states rather than a single candidate everywhere. Harrison, their lead candidate, was on the ballot in 15 of 25 states. While he got 36.59 percent of the national vote, he got 48.76 percent across the 15 states where he ran, and he got 52.72 percent of the vote in those states in 1840, such that his share of the vote improved by 8.1 percent. Finally, Henry Clay’s first presidential campaign (in 1824) was conducted before every state kept popular vote tallies, so I only compared his second race in 1832 and his third in 1844.

If we exclude those three candidates, the biggest improvement by a candidate who started without majority support was Nixon in 1972. Trump falls closer to the middle of the pack by improving his vote share by 8.2 percent, although he ranks higher in terms of expanding his share of all eligible voters by 15.2 percent between 2016 and 2024 — higher than Nixon’s 7.9 percent increase from 1960 to 1972 and much higher than Bill Clinton’s 2 percent growth from 1992 to 1996, but behind George W. Bush’s 17.8 percent improvement from 2000 to 2004. Much of that reflects the fact that both Bush and Trump originally won low-turnout elections and then later won much-higher-turnout elections, whereas Nixon did the opposite, and Clinton ran both times in three-way races, with turnout down significantly in 1996.

Among the 13 candidates who ran multiple times after starting without majority support, only four — Nixon, Lincoln, George W. Bush, and William Henry Harrison — ever went on to capture a popular majority, and as noted, Lincoln’s majority was the artificial product of the opposing party’s geographic base trying to leave the country. Trump, like Clinton or Grover Cleveland or Woodrow Wilson, didn’t quite make it. But five of these candidates never even matched their initial share of the vote in their first run: Cleveland, Benjamin Harrison, William Jennings Bryan, Tom Dewey, and Adlai Stevenson. Trump not only avoided that fate, he stood it on its head, defying parallels to the politics of the 1890s. Like Nixon and Clay, he did best in his third try. It’s a good coalition — if he can keep it.