Why did Democrats lose House seats?

MONDAY, NOVEMBER 9, 2020

Statistical thinking is hard: Democrats won the White House last week. They also lost some seats in the House.

Why did Democrats lose House seats? Dems are at each others' throats, and journalists seem to be puzzled by this. Statistical thinking is hard!

Why did Democrats lose House seats? There are always a million ways you can answer a question like that. But we'd suggest you start by pondering voter turnout in recent House elections, starting in 2008:

Voter turnout, House elections. Off-year elections in bold:
2008: 117.5 million
2010: 83.8 million 

2012: 117.9 million
2014: 75.7 million 

2016: 124.9 million
2018: 111.4 million

2020: 147 million and counting; maybe 155 million when done.

Just consider the last two elections. In 2018, 111.4 million people voted. They elected Democrats to 235 seats.

This year, as many as 155 million people may have voted. In short, this was a vastly different electorate from the one which showed up in 2018.

On its face, there's no reason to assume that this year's (much larger) electorate would produce the same results as the (much smaller) electorate from two years ago. Beyond that, consider this:

Off-year elections tend to favor the party which doesn't hold the White House. So it went in 2018, with angry Democrats turning out to push back against Trump.

This year, with Trump himself on the ballot, Republicans had a major incentive to turn out too. The  smaller electorate in 2018 was largely driven by angry Dems. This year's vastly larger electorate featured a very large wave of Republican voters too.

With these patterns in mind, why wouldn't Democrats have lost some House seats this year? With massive turnout by both tribes, wouldn't that seem like a likely outcome?

Last time, the electorate was Democrat-heavy. This year, both sides came pouring out.

Opponents of the sitting president tend to turn out: 2010 was Obama's first off-year election. More Republican voters turned out. (Roughly, 45 million R to 39 million D.)

2018 was Trump's first off-year election, Many more Democratic voters turned out. (Roughly, 61 million D to 51 million R.)

This year, both parties turned out, bigly. As compared to the turnout last time,  wouldn't this make it fairly likely that Dems might lose some seats?