History suggests the 2018 election will almost surely be a bad one for Republicans. Midterms are generally considered a referendum on the president, and the results are almost always bad for said president. Layer on the fact that President Trump is the most unpopular new president in the modern era, and it would seem to be — at least at this early juncture — a clear recipe for a Democratic wave.
The key word there being “seem.”
A new Washington Post-ABC News poll, as Mike DeBonis and Emily Guskin report, presents a pretty mixed bag for Democrats. It shows that registered voters say they want Democrats to control Congress to be a check on Trump by a 52-38 percent margin, but it also shows Democrats are — rather remarkably — less enthusiastic about voting than Republicans are. While 65 percent of Republicans and GOP-leaning adults say they are “almost certain to vote,” just 57 percent of Democrats and Democratic-leaning adults say the same.
The question from there is which is more predictive of what lies ahead. And the answer won't necessarily make Democrats feel better.
That first number would seem to be a big one: People prefer a Democratic Congress reining in Trump by a 14-point margin! That is a big margin. It's actually similar in size to what it was for Republicans before their big wins in the 2010 and 2014 midterms. When the Post-ABC poll asked this question in April 2014, the GOP led on it by 14 points. When it was asked twice toward the end of the 2010 election, the GOP led by between eight and 16 points.
But then there's 2002. That midterm election was close to a stalemate, but just over a month beforehand, a similar question rendered a 19-point advantage to Democrats — quite similar to today's 14-point edge. Despite this, Democrats would actually go on to lose some seats in the House and the Senate.
In addition, the choice in the poll question is between being a check on the president and voting for the party that supports the president's agenda, and poll respondents are often drawn toward more middle-ground, moderate positions. Hence, Post-ABC polls show people have consistently leaned on the side of checking a president's power. But if you look at the so-called generic ballot — a simple question about which party you prefer — it's almost always closer than this. And sure enough, that's the case today, too.
So Democrats have an advantage there, though it's not clear how predictive it is. Which brings us to the flip side of the coin: enthusiasm. How bad is it that Democrats are somehow less enthusiastic about this election than are Republicans?
For past polling on that question, we have to look at registered voters rather than all adults, as we did at the top of this post. There, Republicans have a smaller, four-point advantage — with 70 percent of Republicans and Republican-leaning registered voters “almost certain to vote” vs. 66 percent for Democrats.
That's actually just shy of where the GOP's advantage was for much of the 2014 election, when they went on to a big win. On the eve of that midterm, the GOP had a seven-point edge on enthusiasm in Post-ABC polling. The GOP had a bigger, double-digit edge on this heading into the 2010 election in which they won big.
And finally, many of the voters who give Democrats an advantage on that first question above say they are "almost certain to vote" but didn't actually vote in the last midterm in 2014. Democrats-as-a-check-on-Trump leads by 34 points with this group, but by just 8 points with "almost certain" voters who did vote last time. This, as much as anything, suggests enthusiasm is hugely important to Democrats.
These are indicators that fluctuate quite a bit, and we're still more than 15 months away from the 2018 election, so we'll have to stay tuned. But there seems to be almost an expectation that Trump as president will spur big Democratic turnout, and this poll calls that into question. Watch these "almost certain to vote" numbers going forward, because they are a pretty solid leading indicator.
The Post-ABC poll was conducted July 10-13 among a random national sample of 1,001 adults reached on cellular and landline phones. The margin of sampling error for overall results is plus or minus 3.5 percentage points and four points among the sample of 859 registered voters.