On Election Night, the conventional wisdom seemed to be that the 2018 midterm was a split decision. Democrats won the House, Republicans expanded their margin in the Senate, and different pundits spun the results in different ways. The public had a similar intepretation: A post-election Huffington Post/YouGov found that 68 percent of people who voted for a Republican in the midterms were “ at least satisfied” with the national results, and 65 percent of Democratic voters said the same.

Since then, as more votes have been counted, the Democrats seem to be in a better position than they were a couple weeks ago.

Does the data show that such a shift in the conventional wisdom is warranted? I examined turnout numbers to see what it all means for 2020.

The Post-Election Vote-Counting Hasn’t Been Great for the GOP

Republicans ended Election Night with less-than-great results. By the time the famous New York Times Upshot needle powered down, Democrats were projected to win 229 seats in the House and hold a roughly seven-point edge in the House popular vote (with some room for error around both estimates). In the Senate, the needle thought the GOP would end up with 53 seats, with Arizona, Montana, Florida and Mississippi still undecided when most election watchers went to bed.

Now, Democrats look like they’ll end up with roughly 235 seats (rather than 229) as multiple close races have broken their way. A number of those flips came from California, where the final votes usually come in late and Democrats won well-educated, sometimes diverse, and traditionally GOP-held areas (see the Orange County seats or the retracted calls in the now-undecided race for Republican Rep. David Valadao’s seat). Democrats also got late wins in Utah’s 4 th District, Maine’s 2 nd District, New York’s 22 nd District and others.

Not every call went against the GOP (Rob Woodall barely held on in Georgia-7 and Will Hurd won in Texas-23), but on balance the last few weeks of vote-counting have gone poorly for House Republicans. Democrats will end up with more seats than they were projected to get on Election Night, and their overall margin in the House popular vote has gone up to eight points.

Republicans have some wins and some losses outside the House. Rick Scott and Ron DeSantis won their respective Florida recounts, and Brian Kemp became the governor of Georgia after lawsuits from his opponent, Stacey Abrams, and some serious clashes over voting rights. Republicans also held on in Mississippi's runoff election on Tuesday (a race that was a lot less competitive than the conventional wisdom indicated). But Krysten Sinema ended up winning in Arizona and Jon Tester won in Montana. So the post-Election Night vote counting on the Senate side seemed to come out in a wash for the GOP, while the most important undecided governor races went Republican.

Overall, these late calls don't warrant a huge reassessment of the 2018 midterms. We knew enough on Election Night to say that this was a really rough night for the House GOP and the post-election calls didn’t change too much on the Senate side (and governor elections are a little idiosyncratic). Democrats have caught more breaks than Republicans in the post-Election-Day counting, but if you look closely you’ll see that the original picture was always rough for the GOP and that things haven’t changed as much as some might think.

House Republicans Hit Their Turnout Numbers, but Democrats Blew The Doors Off

The most recently updated numbers have confirmed the conventional wisdom: Democrats had very strong turnout. As of now, House Democrats have almost 60.1 million votes, which is a pretty huge total considering that Hillary Clinton got 65.8 million votes in 2016. But the intense focus on Democratic turnout has obscured the fact that Republicans also turned out strongly.
Year POTUS Party Total House Votes for POTUS Party Raw Votes for POTUS -- Last Election Raw Votes for POTUS Party -- Last House Election Midterm as Percent of POTUS Vote Midterm as Percent of Last House Vote
2018 R 50,673,163 62,984,828 63,173,815 80.45297 80.21228
2014 D 35,624,357 65,915,795 59,645,531 54.04525 59.72678
2010 D 38,980,192 69,498,516 65,237,840 56.0878 59.75089
2006 R 35,857,334 62,040,610 55,958,144 57.79655 64.07885
2002 R 37,332,552 50,456,002 46,992,383 73.99031 79.44384
1998 D 31,490,298 47,401,185 43,507,586 66.43357 72.37887
1994 D 31,609,829 44,909,889 48,654,189 70.38501 64.96836

This table focuses on turnout among the president’s party in recent midterms. In midterm elections, the president’s party usually doesn’t turn out well – the opposition party (i.e. the party that doesn’t hold the White House) is often angry about whatever the President is doing and ends up more energized about the election. In midterms where the President’s party lost in a wave (e.g. 2010, 2006, 1994) their overall vote total usually came in way below what they got in the previous election.

But 2018 was different—Republicans turned out but still lost. As of now House Republicans are at about 80 percent of the GOP’s 2016 total. In fact, 50.7 million votes—the latest vote count for House GOP candidates—is larger than the 44.8 million or 40.1 million votes they got the Republican landslide of 2010 or their solid 2014 win. In other words, the overall vote count for the GOP wasn’t bad in comparison to recent past midterms. But it’s not so great when it’s compared to the other number that counts—the Democratic vote total.

So Where Does this Leave Trump and the Republicans?

The 2018 results weren’t good for the GOP, and Donald Trump bears a lot of responsibility for that. The 2018 House results were highly correlated with the 2016 presidential election results, and Democrats won important victories in GOP-held districts where Trump ran behind Mitt Romney’s 2012 margin. More generally, the results were consistent with an election where voters generally disliked Trump and took their frustration out on Republican House candidates.

But we should be careful before extrapolating this result out to 2020. As I’ve noted before, midterm results have a bad record of predicting the outcome of the next presidential election. Sometimes they’re right (2006 was a bad year for the GOP, 2008 was a terrible one) and sometimes they’re wrong (2010 was a great year for the GOP, 2012 was not) and sometimes it’s just vague (1974 was a terrible year for the GOP, and 1976 was close). The midterm elections simply don’t have a good track record of predicting what’ll happen next time around.

There is a sort of bizarre silver lining in these results for Trump. Republicans lost partially because they pushed unpopular policies. This is a pretty normal thing for parties to do—presidents often try to deliver for the base, and (regardless of what you believe about the merits of those policies) not everybody loves what the base loves. For example, Trump’s approval ratings took hits when major news about the various Affordable Care Act/Obamacare Replacements were rolled out. Those bills generally didn’t poll well, and they damaged Trump’s approval rating. But in 2019 and 2020, Republicans won’t have enough strength to push legislation like that, so Trump won’t have the weight of new unpopular-but-base-pleasing policies dragging his approval rating down.

That being said, Republicans also lost because Trump is Trump. His overall approval rating is less than it should be given the economy, and his personal style likely has something to do with that. If Trump tweeted less and handled controversy ( especially racial controversy) differently he’d probably be in a significantly better position. So far I haven’t seen much evidence that he’s rethinking his personal strategy, and that’s a potential problem for him.

I think this leaves Trump’s re-election odds at about where they were before Election Day—somewhere around 50/50. We don’t know who the Democrats are going to nominate, what national conditions are going to be like, what’s going to happen with the Mueller Investigation or almost any of the relevant details outside what we’ve already learned about Trump himself and the makeup of the next Congress.

These results also suggest that the GOP is almost totally tied to Trump for the time being. The 2018 election provided evidence that Trump’s relationship to the GOP is, electorally speaking, normal. Voters saw 2018 as a referendum on Trump and House Republicans took a beating. Right now there’s no reason to doubt that the down-ballot GOP will also be closely tied to him in 2020.