Plausibly, Trump is likely locked in to around 46.5 - 47.0 percent of the voters in the 2024 general election. There are a lot of inputs for that: Trump’s popular-vote percentages in 2016 (46.1) and 2020 (46.8); his rolling aggregate in RealClear Polling (vacillating between 44.6 and 47.8, with the mean for the year around 46.5); July 2024 polls via FiveThirtyEight sticking firm around 43% - 44%.
Now, I’m aware that’s popular vote, not Electoral College counts. I’m merely using the popular-vote poll results to demonstrate that Trump’s popular-vote percentage is apparently baked-in at a certain level. In popular-vote polling, he never quite arm-wrestles Biden’s hand down to the table – he never jumps out to a 55% poll result, say. Trump’s support, plausibly, is firmly-buttressed but inelastic.
Now, fair or not, I’m going to look at different polling figures for Biden – battleground state-level polls that FiveThirtyEight incorporates into their state-level predictions (because fewer state-level polls are done, I have to go back to the beginning of June to have a decent sample). A common theme is that the state polls have something like 8-15 percentage points dedicated to either other third-party candidates or to “undecided”. Trump’s battleground-state percentages,meanwhile, are about in line with his national polling
Pennsylvania
Michigan
Wisconsin
Even in battleground states that FiveThityEight currently projects Trump to carry, Trump’s poll results aren’t better – Biden’s are just lower and the “third-party/undecided” percentages are higher:
Georgia
Arizona
Nevada
North Carolina
Now. To me, it’s plausible that there’s still an underlying “never Trump, but not Biden yet” component in the polling. That is, Biden’s poll numbers aren’t capturing some percentage of voters that will still vote “Biden” when push comes to shove in November. In answering a poll, it costs nothing to beg off of Biden. Once it’s November 5th … you’ve made the effort to get to the voting booth, the curtain’s drawn, you know you’re in a battleground state, you already know you very much don’t want Trump in office – the calculus is very different.
I believe it’s plausible that the last-minute calculus shifts end up favoring Biden because it’s demonstrated to my satisfaction that his support is more elastic than Trump’s. Or it can be viewed from the opposite perspective – that Biden’s non-support is mushier.