The 2018 midterms aren’t even finished yet — several House races still have not been called — but the conversation has already moved on to feverish speculation about what the results indicate for the 2020 presidential election.
Since Democrats picked up a Senate seat in Arizona while getting trounced in a governor’s race in Ohio, will President Donald Trump win the former and lose the latter? Does Andrew Gillum coming up short in Florida mean it’s now off the table for Democrats? Does Stacey Abrams coming close in Georgia mean it’s now on the table? And most of all, does Democrats’ resurgence in the former “blue wall” states of Wisconsin, Michigan, and Pennsylvania mean that Trump is doomed?
The boring answer is no.
It’s natural to look to the 2018 results as tea leaves that can help us predict the future. Trump fascinated Americans for decades long before he got into politics, and now that he’s president, everyone wants to know whether he will continue to be president.
But the truth is that nobody knows anything, life is unpredictable, and while politics is shaped by deep demographic trends whose influence is detectable in the midterms, things can also change very rapidly in unexpected ways.
Barack Obama swung from a landslide victory in 2008 to a horrific defeat in 2010 and then bounced back to a decisive win in 2012 — only to see the electoral map fundamentally shift by 2016. If the midterms remind us of anything, it’s that the future is hard to predict.
Florida’s growing Hispanic population has been on the verge of turning the state solidly blue for 20 years, yet it keeps not happening. Arizona’s growing Hispanic population has been on the verge of transforming it into a swing state for 20 years — and hey, it maybe is happening! Iowa swung way to the right in 2016 and then replaced half its House caucus with Democrats in 2018. Eight years ago, Scott Walker triumphed against the odds and was elected governor of the solidly blue state of Wisconsin. Last week, Tony Evers triumphed against the odds and beat Scott Walker in Trump country.
Life is weird. But here’s what we know.
You need to adjust for the national trend
The key thing to remember when looking at 2018 is that Democrats just did a lot better overall than they did two years ago. Hillary Clinton won the popular vote by about 2 points in 2016, while House Democratic candidates won by more than 7 in 2018.
That 5-point swing in Democrats’ direction should be the starting point for all analysis.
For starters, if Democrats win the popular vote by 7, they are going to beat Trump. Geography doesn’t matter in the face of a landslide that big. And winning by 7 is by no means unrealistic. Trump got 46 percent of the vote in 2016, Clinton got 48, and the remaining 6 percent splintered among third-party candidates. If Democrats can manage to fully consolidate the not-Trump vote, they’ll win big — and nobody will care whether Iowa swung back from red to blue.
This is important because some of the seemingly most dramatic 2018 results look very different when you adjust for the fact that 2018 Democrats performed about 5 points better than Clinton in 2016, on average:
- Beto O’Rourke’s Senate campaign in Texas attracted a ton of national attention, and he came very close in the end, losing to Ted Cruz by only 3 points. But Clinton only lost Texas by 8 points, so O’Rourke’s improvement on Clinton was exactly average and not necessarily indicative of any incredibly impressive political skills on his part.
- Similarly, Arizona “turned blue” in the sense that Clinton lost it by 4.5 points and then Democrats ran 5 points ahead of her nationally — which powered Democrat Kyrsten Sinema to a narrow Senate win based on an average performance.
- By contrast, the razor-close statewide elections in Florida feel familiar because Florida has had so many close elections over the years. But Trump only won Florida by 1 point. An average, or even below-average, performance relative to the national swing should have been enough for Democrats Bill Nelson and Andrew Gillum. The close races indicate a very strong underlying Republican trend in Florida.
- The anti-Florida of 2018 was Iowa, a traditional swing state where Clinton got creamed by 9 points. Democrat Fred Hubbell fell short in the governor’s race, but it was by closer than 4 points, and Democratic House candidates did very well in Iowa. Trump’s trade policies have been really bad for farmers, which may be hurting Republicans there.
All of which is to say that while Democrats did a lot better in 2018 than in 2016, they didn’t necessarily “change the map” in a fundamental way.
In 2012, not only did Obama win the popular vote by nearly 4 points, but he won each of Virginia, Colorado, Nevada, Pennsylvania, Michigan, and Wisconsin by more than 4 points. In other words, even if he’d lost the popular vote to Romney by a few tenths of a percentage point, he still would have won.
That changed in 2016, and it doesn’t seem to have changed back in 2018. Democrats are consolidating “wasted” votes in the suburbs of big cities in non-swing states like California, Texas, and New York.
All that said, the transformation between 2012 and 2016 was not necessarily foreseeable.
Midterms don’t tell us much
At the end of the day, before asking what the 2018 midterms tell us about 2020, you might want to ask what the 2010 midterms told us about 2012, or what the 1994 midterms told us about 1996.
In fact, George H.W. Bush’s GOP did pretty well in the 1990 midterms for an incumbent party — which told us nothing at all about 1992, when he lost to Bill Clinton.
In 2002, George W. Bush scored one of the strongest midterm elections for a president’s party. Then he nearly lost his 2004 reelection bid due to weakness in the pivotal Electoral College state of Ohio. That didn’t seem likely in 2002, when in the midterms, Ohio Republicans won a landslide 20-point win in the gubernatorial election and won the popular House vote by 12 points.
All of which is just to say that it’s hard to make predictions about the future based on midterm results. And that’s largely because actual events tend to intervene. The elder Bush was cruising to reelection until he was derailed by a recession. The younger Bush’s midterm wins were driven by a post-9/11 polling surge, and his Ohio weakness emerged later, when America’s manufacturing sector unexpectedly failed to bounce back from the early-aughts recession thanks to competition with China.
Bill Clinton reshaped his whole political persona in the wake of his early midterm losses, pivoted to the center by signing a welfare reform bill that sharply increased severe poverty, and was rewarded with an 8-point stomping of Bob Dole that saw him win all the swing states plus Louisiana, Arkansas, Tennessee, Kentucky, and Arizona (though not Virginia or Colorado, which were firmly Republican two decades ago).
One of the striking things about the Trump era is the extent to which, in contrast to the historical pattern, events do not seem to have moved his approval ratings.
That’s frustrating to Democrats, who feel in their bones that his numbers should be in the cellar. But it really ought to scare Republicans, because his holding pattern in the low 40s puts them on track to lose in 2020.
More than anything else, that is the lesson of the midterms. Trump’s numbers were never great during the 2016 election, but ever since April 2017 or so, his numbers have been in clear loser territory.
What happened last week is just what losing looks like, with Democrats beating Republicans badly enough that the geographical details didn’t matter very much. Two years is plenty of time for Trump to turn that around and become more popular. But if he wants to get reelected, he does need to actually do something to become more popular — and so far, at least, he’s shown no real capacity for growth and change.