Mar.05

electoral-vote.com

Dem 51

   

GOP 49

One Last Look at Michigan

We had a bunch of questions this weekend about the Michigan primary, and what it all means for the presidential race. We set them aside, so we could write a somewhat more thorough assessment of the numbers. Since Super Tuesday is going to sweep aside nearly all thoughts of earlier primaries and caucuses, it's now or never.

There are primarily two subjects that we were asked about: the uncommitted votes for Joe Biden, and the fact that far more Republicans showed up to vote than did Democrats. We'd like to give a data-driven response to those questions as best we can. That said, the amount of relevant data that is available is pretty scant.

What we mean by that is that the only real basis we have for drawing conclusions is past primary and general election results in Michigan. However, the Democrats awarded their delegates via a caucus prior to 2008. And then, in 2008, Michigan pulled a New Hampshire and defied the Democratic Party, scheduling their primary earlier than they were supposed to. Hillary Clinton stayed on the ballot, Barack Obama withdrew. So, this means we really only have three elections where the setup is the same as 2024, plus the fourth, in 2008, that's kinda close. Here are the data:

Again, not a lot of data to work with, but you go with what you've got.

We'll start with the turnout, because that's easier. Clearly, there is no meaningful correlation between how many people vote in the primary and how many vote in the election. The lower-turnout party won twice (and was within a whisker of doing it three times). The only thing the turnout really tells us is that people are more likely to show up when they feel their vote makes a meaningful statement. So, the poorest turnout in the past four cycles was for the coronation of Barack Obama, while the highest turnout was for Joe vs. Bernie.

Keeping that in mind, the 2024 turnout numbers actually look pretty good for Joe Biden. Yes, 100,000 or so people were motivated enough to show up and vote "uncommitted," but 600,000 or so were motivated enough to show up and make clear they are with the president. Compare that to Obama's 2012 tally, and it's practically a Biden wave. Meanwhile, the Haley "making a statement" voters were a much larger group than the "uncommitted" Democratic voters, both as a total number, and as a percentage. If we assume that both presidential candidates have a ceiling of 2.5 million votes or so, then Trump needs more of the apostates to come home than Biden does. And we would guess the Haley voters are less likely to do so than the Democratic uncommitteds, because while both candidates will benefit from a "hold your nose and vote for the lesser of two evils" effect, Biden is also able to make progress on the Gaza situation. There's no obvious "fix" for Trump to pursue.

Moving on to the uncommitted voters, along with the subject of the Gaza situation, exit polls have made clear that the uncommitteds were heavily concentrated in the Dearborn area, which is the center of the state's Muslim and Arab-American populations. Biden also did more poorly than expected in college towns. Clearly, he does have a "Gaza problem." That said, let's not overstate it. Looking at the most comparable election, Barack Obama saw 10.7% of primary voters cast their ballots for uncommitted in 2012 (again, the wonky 2008 election can't really be used here, because of the squabbling over primaries). In other words, it sure looks like Biden inspired a larger-than-usual number of uncommitted votes, but not a huge number more (+2.5% or so). Arab Americans and other Muslims make up about 2% of the population of Michigan, so one might speculate that maybe 10% of Biden's uncommitted vote was the usual for an uncontested primary, 2% or so was Arab Americans and other Muslims, and 1% or so was students angry about Gaza.

Some of those folks will come home to the Democrats, even if Biden does nothing, simply because they view Trump as so much worse, either in general, or on the issue of Israel. Others will come home if Biden is able to improve the situation in Gaza, something that he knows well he needs to do, and is clearly working on. Some won't come home, no matter what. If Michigan is as close as it was in 2016, then those otherwise-Democratic voters could be decisive.

But will it be that close? Well, as you can see from the table, Donald Trump has his usual ceiling. He's going to get somewhere between 47.5% of the vote and 48% of the vote, and we don't believe anything is going to change that. That's his base, and that share of the vote is typical for him in swing states. That means the real question is whether Joe Biden will bleed roughly 3 points from his 2020 total, dropping him below Trump's 47.8% or so. It could happen, if he can't get back any of the Arab-Americans/Muslims and students he's currently struggling with. But that would be a pretty unusual outcome.

One last thing. The polls, which were admittedly limited in number, predicted that Biden would get about 77% of the primary vote in Michigan. They underestimated him by about 4 points. Meanwhile, as we already noted, the polls overestimated Trump's Michigan tally by 11 points. So, a net total of 15 points of error between the two certain nominees. That's... a lot. The polls will get more accurate as it becomes easier to separate likely voters from not likely voters. But nonetheless, when you hear that yet another Siena poll has Trump up 5 in the swing states and/or nationally (see below), you simply have to take that with several grains of salt, given the pollsters' performance so far this cycle. (Z)