John Sides and I recently released a forecast for this year’s U.S. House elections. This forecast shows that it will be hard, though not impossible, for the Democrats to retake the House, despite the problems facing Republican presidential nominee Donald Trump and the potential for him to hurt candidates down the ballot. Even if Clinton wins by 12 points and has strong coattails for House candidates, the Democrats have no better than a 50/50 chance of gaining the majority.
So does that mean that voters just prefer having the House in Republican hands?
Not exactly.
While our forecasting model gives Democrats a slim chance of winning a majority of seats, it gives them a very good chance (better than 80 percent) of winning at least a plurality of the votes in House races. This is true even in scenarios where Clinton wins by a landslide and has significant coattails.
This isn’t the first time: In 2012 Democrats received just over 50 percent of the two-party vote but won only 46 percent of the seats. If our model is right, it’s about to happen again.
What explains this discrepancy? One factor is incumbency: There are 49 more Republican than Democratic incumbents running this year, and each of them has a leg up over the competition. In some of our past simulations, this single factor accounted for essentially all the difference between Republican votes and seats.
Another possibility is gerrymandering. Republicans ended up drawing many of the congressional plans in 2011, and a number of those plans are heavily tilted in a Republican direction. Although this is a popular explanation, it is important to consider these facts:
- Republicans also had an advantage before the last round of redistricting.
- At least part of that earlier advantage was due to fundamental political geography — Democrats are concentrated in urban centers and so “waste” more votes on large victories.
- Even the largest and most indisputable Republican gerrymanders (looking at you, North Carolina) amount to only a few seats.
- A modest incumbency advantage can account for almost all of the supposed effect of the 2011 redistricting — that is, lots of Republicans ran as incumbents for the first time while also running in new districts.
To be clear, these are claims about the overall effect of the last redistricting. It is still true that the playing field is clearly tilted toward Republicans and has been for a while. Some of that advantage may also be due to earlier gerrymanders.
And even if the overall effect is smaller than often portrayed, it does not rule out the existence of lopsided gerrymanders in individual states. If Democrats come close to winning a majority in 2016, gerrymandering could make the difference.
In fact, with assumptions that favor a redistricting effect, Republicans may have as many as seven extra seats compared to what we would expect given the boundaries before the 2011 redistricting.
If we assume Clinton’s current polling margin and strong coattails, it predicts 212 Democratic seats. In such a world, an extra seven seats would make a big difference.
Eric McGhee is a political scientist and occasional contributor. He studies elections and political reform.