Sunday, November 22, 2020

Who Does High Turnout Help?

For as long as I can remember, it has been accepted wisdom that Democrats benefit from higher turnout. This is the view that motivates "if we can just get more people off the sidelines, Democrats will win every election", as well as more pessimistic declarations of how Democrats fare in midterms, off-cycle races, and run-off elections in, oh, let's say, Georgia.

But is it true today? The 2020 election is giving me a bit of pause.

2020 was a big turnout year. We had record turnout -- the highest percentage in at least 100 years, in all likelihood -- and that's with COVID throwing a wrench in things. But while Joe Biden won, and won clearly in the national popular vote, it's not the case that the additional turnout was all a tidal wave of new blue voters. Trump, too, has shown himself to be a turnout machine for the red column. Texas is a good example, where Joe Biden added 1.4 million votes to Hillary Clinton's 2016 total, only to see Donald Trump roughly keep pace by stacking an additional 1.2 million votes on top of his performance in the prior election. That's a lot more people voting, but not a huge net gain for Democrats -- especially given the general "blue-ing" of the state that had been observed over the past four years.

So what's going on? One thing to consider is who the marginal non-voter is, and who they're likely to support if they do come out to the polls. Non-voters are likely less politically engaged and aware -- the classic "independent" voter (which is to say, low-information and ideologically incoherent), and probably exhibit less trust in and affinity towards American political institutions generally. In our current climate, it's far from clear that these aren't easier for a Trumpist style populist politician to win.

More than that, though, is the issue of the broader realignment we're seeing in partisan identity. Historically, the case for Democrats being aided by high turnout has I think relied on the notion that Democratic voters skew poorer, and poorer voters are less likely to turn out, so the marginal vote gained by heightened turnout is more likely to be a Democratic one. But while it is not the case, contra some lazy takes, that Democrats are now the party of wealthy coastal elites, it is the case that the biggest divide between the parties right now does not track class but rather education. Democrats are overperforming among college-educated voters (of all economic backgrounds), Republicans do much better among those lacking a college degree (again, regardless of economic background). And highly-educated voters are a high turnout group -- they're likely to hit the polls even when other actors do not.

So it's quite possible that reductions in turnout could end up, counterintuitively, aiding Democratic candidates. You can imagine dividing voters into different turnout "tranches", where the highest tranche turns out in every election (that is, even in ones where nobody else votes), the ones below that in slightly more active races, the ones below that in moderately high turnout affairs, and so on down the line until the final tranche which never votes at all. If Democrats are disproportionately represented among the highest tranches, they'd be better served if elections remain low-turnout affairs, since they'd be the only ones showing up to the polls.

Again, this is just a hypothesis and an oversimplification at that. But I do think the education realignment may require adjusting some of our assumptions regarding who benefits from high turnout.


Benjamin Lewis said...

There's another facet which you touched on but didn't analyze thoroughly - why are voters turning out in a given race/election who aren't reliable?
For example, my pet hypothesis is that membership in the crazy-colony that fundamentally distrusts media institutions (and therefore declines to participate in polls) is correlated to support for Trumpism independent of other weighted factors - this creates a hard-to-avoid self-selection bias in polling, which explains the repeat of 2016 polling error, even given methodological adjustments like accounting for the education realignment. So those people will turn out based on... not negative partisanship, but personality based activation with a key candidate?

Whereas, there almost certainly continue to be lots of progressive-sympathetic voters who aren't participating because the system is unfriendly. (Total voter participation is still ... < 70% ?)

The implication wolud be that there's not tradeoff-in-expectation between doing the right thing about elections and winning them: progressive voting *reforms* still systemically favor progressive policy because the procedural obstacles they remove are the primary reason that eligible progressive are inactive, but they are not the primary reason that eligible crazy-campers are inactive.

Related, the whole "too much" vs "too little" debate going on about the US House elections etc. AOC claims that progressive wish-list items like "relied on AOC as a surrogate" and "endorsed Medicare for All" were correlated with success. I want that to be true, and also it seems obviously difficult to disentangle that data from any self-selection bias that's present - candidates who think AOC / progressive policies could hurt them in their particular districts aren't necessarily wrong just because they lost.
However, the Florida minimum wage initiative seems to be at least somewhat indepedent evidence in support of that hypothesis - more than 60% of the vote makes it vastly more popular than either presidential candidate, tying onself to it probably would have been a good choice
- based on some nth-hand analysis I read, essentially no FL democrats endorsed that initiative, and some who were running this year lost close-ish races in districts where the initiative was very popular. Was it somehow going to hurt a candidate to endorse something that did in fact get more votes than them in their own district?

Erl said...

"not tradeoff-in-expectation between doing the right thing about elections and winning them"

Gosh, I hope not. If expanded voting rights are electorally neutral, then they will be much easier to pass; if they are seen by Republicans as a stalking horse for electoral domination they can't make it through. I want a settlement on the subject of substantive universal suffrage, not all this endless bullshit. (of course, it is hard to convince the Republicans that anything is a matter of principle.)