Do we understand elections?

The media is full of analyses as to why Donald Trump defeated Kamala Harris in the recent election. At various times, I’ve mentioned factors like voter frustration over high inflation, illegal immigration, and woke excesses on college campuses.

The more I think about the election, however, the less confidence I have in any single explanation. This is especially true in a close election. And while Trump had a comfortable majority in the Electoral College, if just 1% of the electorate had uniformly swung from Trump to Harris, she would have won both the popular vote and the Electoral College.

Consider the following thought experiment. The popular vote margin went from Trump trailing by roughly 4.5% in 2020 to winning by 1.5% in 2024.  You can think of that as 3% of the electorate switching from the Democrats to the Republicans.  If only 2% had switched toward Trump, he might well have lost.  This means that almost any factor that moved an additional one percent of the electorate might legitimately be seen as decisive.  Thus if (relative to 2020) 5 different issues each moved 1% of the electorate toward Trump, and 2 single issues moved 1% of the electorate toward Harris, that could explain this year’s result.  In that case, any single one of the 5 issues favoring Trump could be seen as decisive.

Here’s Bloomberg:

Among the moves [Trump] pledged—all of which are up to Congress, not him—were to extend the 2017 tax cuts that largely benefitted corporations and the rich (price tag: $4.6 trillion); remove taxes on tipped wages for service workers ($250 billion); increase the child tax credit from $2,000 to $5,000 ($3 trillion); and eliminate taxes on Social Security benefits ($1.8 trillion). But Republicans can’t possibly deliver all of this, or even most of it, despite having full control of Washington. 

That’s an impressive list, but it doesn’t even include Trump’s promise to bring back the SALT deduction, which is a hugely important issue to many voters in states like New Jersey and New York (two states where Trump did much better than expected.)  Nor does it include Trump’s proposal to abolish taxes on overtime pay.  But I almost never see these tax plans discussed as the reason why Trump won, by pundits of either party.  Most of the analyses have focused on other issues.  It’s almost as if there is something slightly disreputable about speaking of election outcomes in crude financial terms.

Perhaps pundits believe that most voters didn’t decide to vote for Trump on the basis of these promises.  But that’s not the issue at stake.  The question is not how “most voters” vote, the question is whether a promise to boost the child tax credit to $5000 and bring back SALT deductions and abolish taxes on tips, and abolish taxes on overtime pay and abolish taxes on Social Security income were enough to sway 1% of the electorate.  That doesn’t seem all that implausible.

Another objection is that the Democrats also made expensive promises, and perhaps the various promises balanced out.  That’s a reasonable counterargument.  For instance, the Democrats have been trying to forgive student loans, although the initiative has been tied up in the courts.  Harris also promised to exempt tips, but only after Trump had done so.  Thus her promise achieved less attention.  

But Trump’s tax cut promises were much larger than Harris’s and only partly offset by higher tariffs.  In addition, some voters wrongly believe that tariffs are paid by foreigners.   Thus I suspect that Trump’s tax program was more popular than the one proposed by Harris, even among lower paid workers.  This is a source of extreme frustration to progressives, who see the Democrats as the party of the working class.  

I don’t have any firm conclusion here.  Rather I’d encourage people to be open minded about election explanations in a close race.  Thus the statement that 98% of voters would not reject a candidate because she was a black woman does not in any way refute the claim that Harris lost the election because she’s a black woman.  (To be clear, I believe the main reason the Democrats lost related to other factors, such as those I listed at the top of this post.  But in a very close race, almost any single factor could be decisive.)

The analysis above applies even more strongly to complex historical events.  Thus there might be a dozen factors that led to something like the Great Depression or World War II, where a different outcome for any single factor could have led to a radically different outcome.  This is of course related to the famous “butterfly effect” in chaos theory.

PS.  In an ultra-close race like 2000, almost any single factor could plausibly be cited as decisive, even if it merely moved a few hundred votes.