The week before Christmas 2017, the likelihood that the new year would bring Democrats a hefty electoral bounty seemed very high. The overall record in 2017 off-year and special elections for the Donkey Party was stellar. Polls were showing a strong and growing advantage for Democrats on the generic congressional ballot, a measure strongly correlated with actual House results. House Republican veterans were beginning to head into retirement in big numbers. Democrats had won a shocking Senate victory in Alabama, and suddenly beginning to look like possible net winners in Senate seats in the midterms instead of roadkill. And throughout 2017, the president’s job-approval ratings ran far below the 50 percent that has historically been necessary to avoid big midterm losses for his party.
Yes, there seemed to be a mild positive trend for Trump and Republicans as 2018 began, some of it probably attributable to companies crediting the GOP tax bill for bonuses and wage boosts and some of it likely reflecting generally good economic news. But toward the end of January, something a little deeper seemed to be happening. The big eye-opener was a Monmouth poll conducted from January 28–30, and released the day after the president’s generally well-received State of the Union address. It showed Trump’s job-approval rating leaping 10 points since the December poll from the same outlet. More surprisingly, it showed the Democrats’ generic ballot margin dropping from 15 points to 2.
It’s hard to generalize short-term polling trends, if only because the leading aggregators (FiveThirtyEight and RealClearPolitics are the two I regularly consult) don’t necessarily use the same polls. But it’s now quite clear that the numbers have moved pretty strongly in the GOP’s direction since Christmas.
FiveThirtyEight’s polling averages gave Democrats a 12.8 percent advantage (50.1/37.3) on December 24. Today it’s down to 6.5 percent (46.7/40.2). RealClearPolitics had Democrats up by 13 points on December 24 (49.1/36.1). Now it’s a 6 point advantage (44.4/38.4). So we are talking generally about that “wave” that looked so apparent at the end of 2016 being roughly halved.
And that’s not because of a bad mix of polls. The latest to hit the news is from Quinnipiac. At first blush it represents good news for Democrats: a nine-point lead on the generic ballot (49/40). But that’s the smallest Democratic advantage on a Q-Pac poll since last October; in the last December survey, Democrats were up by 15 points (52/37). And in most polls, Trump’s job approval ratings have been drifting upward, too: In the RealClearPolitics averages it’s gone up from 38.8 on December 24 to 42.4 today. At FiveThirtyEight, it was 37.4 on Christmas Eve and 40.7 today. Meanwhile, just about everyone is reporting that public opinion toward the GOP tax-cut bill has gone from toxic to lukewarm. And that trend is very likely to continue when the IRS begins withholding less wage income and then again when taxes are filed.
So are Democratic dreams of a big midterm sure to be dashed? Of course not. Aside from the great distance at which we stand from the 2018 general election, there are counter-indicators of any pro-GOP trend, most recently in yesterday’s special state legislative elections in Missouri. As New York’s Eric Levitz argues, polls may be missing an especially intense Democratic determination to vote that make existing turnout models obsolete.
Typically polls get more focused on measuring likelihood to vote as an election grows nearer, and that’s particularly true with midterms, where turnout is lower and more variable than in presidential years. We’re probably going to see some inaccurate polls down the stretch based on likely voter “screens” that heavily rely on past midterm voting — i.e., in the heavily Republican election years of 2010 and 2014 — rather than stated interest in voting.
These are all things to watch closely this summer and fall, along with developments in the many individual races that will determine the actual outcome. For the time being, polls are probably accurately reflecting a better fundamental position for Republicans than they had a few weeks ago, but hardly any guarantee they will get through this cycle without a bloodbath. Yes, Democrats will have to earn a 2018 victory with the kind of “ground” effort they have displayed in many special and off-year elections. But at this point you’d much rather be in their shoes than in those of the party going into the midterms with Donald Trump at the helm and battling a vast history of anti–White House midterm voting.