How the Battle for the House Is Shaping Up
If you had asked me six months ago who I thought would win control of the House of Representatives in 2018, I wouldn’t have hesitated before answering, “It’s early, but Democrats are heavily favored, although conventional wisdom has been very slow to catch up.” With a raft of GOP retirements in highly vulnerable open seats, a president with job approval ratings in the 30s, and a generic ballot lead for Democrats in the double digits, it was increasingly difficult to spell out a path to victory for Republicans. In fact, things were bad enough that it appeared their losses could grow into the 40 or even 50 seat range.
Things have changed. If the election were held today, it’s not clear who would hold the chamber. I might put a thumb on the scale for Republicans, but right now – and it is still early – the House is likely to be close. Once again, conventional wisdom seems slow to catch up, with analysts still discussing the toxic environment for Republicans. There are three things to consider:
1. Trump’s job approval has increased.
I’ve long taken the stance that elections are fundamentally referenda on the party in power, and specifically on the performance of the president. Other factors, such as exposure and the economy play a role, but the correlation between presidential job approval and electoral outcomes is undeniable.
Trump’s job approval hit a low of 37 percent in the RCP Average in mid-December, but since then has gradually improved. As of this writing, it stands at 43.7 percent, which is roughly where it was before Trump fired FBI Director James Comey.
There are a variety of potential reasons for this – the passage of the tax cut rallied Republicans, people are tiring of the Mueller investigation and increasingly of the opinion that it is politicized, and the good economic news is breaking through. The diplomacy with North Korea, even if it fails, contradicts the image of Trump as an irrational warmonger. Regardless of the reasons, his fortunes have clearly improved, and if you agree that elections are referenda, then we would expect them to raise all boats.
2. The generic ballot has moved.
Thus, it is no great surprise that GOP numbers on the so-called generic ballot, which asks some variant of whether people plan to vote for Republicans or Democrats in the next congressional election, have likewise improved. The generic ballot is an imperfect metric, but it is a decent gauge of where the mood of the country is. In mid-December, the Democrats’ lead was around 13 points. Today it stands at just four points.
This is a problem for Democrats, because to overcome GOP gerrymandering (which, in fairness, has been whittled away by court decisions to a large extent), the clustering of Democratic partisans and the Republican incumbency advantage, many expect that Democrats would have to win the popular vote by five to seven points. I think this overstates the case, but a three-point popular vote win would probably result in a very narrowly divided House in 2019.
3. Special elections are an unproven metric.
Finally, some point to the Democrats’ performance in special elections as a sign that their voters are energized. This is undoubtedly true. I pointed to a similar phenomenon in late 2013, so I do think special election outcomes are interesting.
But there are three problems here. I did not make specific projections from these in 2013, because we don’t have a lot of experience projecting a midterm election from special election results. These are informative datapoints, but until we have more experience seeing how models based upon them perform, we should prefer established metrics like job approval and the generic ballot.
Second, these special elections all involve open seats. The one set of elections that did not involve open seats exclusively – in Virginia and New Jersey – brought about a set of results that are a lot less encouraging for Democrats. They won in “Hillary districts” and lost “Trump districts.” If this is replicated in 2018, Democrats would fall just shy of winning the House.
Finally, many of those elections took place when the president’s job approval was at sub-40 levels. Perhaps unsurprisingly, the Democrats’ margin has narrowed as the Republicans’ position has improved. If the president’s job approval on Election Day (or, as some suggest, in early summer) is what drives election outcomes, using the average special election outcome will result in a lowball estimate of where we end up come November. Again, we should not dismiss these data. But until we have more experience using them, we are better off relying upon tried-and-true techniques.
Can we say with certainty how the Battle for the House will play out? Absolutely not. Six months is several lifetimes in politics. But there is little doubt that the Republicans’ chances have improved over the past five months, perhaps dramatically so. That is noteworthy, and the CW should adjust accordingly.