Admittedly, my forecast for the 2018 House elections looks pretty pessimistic for the Democrats. The media has been quick to pick up on this fact, but they’re glossing over something important: Despite a grim forecast seat share, Democrats have an outside chance of winning the majority — and a landslide election.
I talked through this “skew” of expectations in a twitter thread the other day:
Let’s talk more about what happens if we’re wrong about the House generic ballot, according to my 2018 model. [THREAD] https://t.co/bFKXlGb8rV— G. Elliott Morris📈🤷♂️ (@gelliottmorris) September 3, 2017
The argument goes like this:
- Democrats have a disproportionately high number of safe seats
- Republicans, by comparison, have found their way to a majority with a healthy number of seats in competitive territory
- That territory will actually become competitive under a few scenarios (the purpose of this post)
- Because we can expect error in generic ballot polling, we have to imagine that those darker blue environments are possible
- Specifically, Democrats need to increase their vote share by 1.5% to be favored to win back the House. If the election were held today, the chance we are that wrong happens 30% of the time. If they're off by more, say 4 percentage points, Democrats win in a landslide. That happens 1 time out of ten. That's a pretty good chance
We can image that an error large enough to land Democrats the House majority exists by shifting the following graph to the right, where the seats in the highlighted area (a 5% margin of error) flip blue.
The only question is how Democrats are going to make these improvements to their position. There are a few paths they can take toward moving that needle the extra point and a half, aside from random chance.
The purpose of the above note is to point out that this seat disadvantage for Democrats, what I call their distributional skew, may actually be an opportunity. The benefits are two-fold. One, that we’ve already covered, is since Democrats have been gerrymandered and grouped into seats that are safe, it’s really hard for them to lose a lot of seats. But perhaps more importantly, this skew can be reversed.
One way for Democrats to move the needle toward them is to have Republican incumbents retire in some desirable districts. As Nate Cohn wrote in a piece for the New York Times Upshot,
If retirements were to stay ahead of 2008 pace, it would be enough to make Democrats favorites to win the House in a wave election like those in 2006 or 2010.
This is because each incumbent earns a “bonus” for being in office. Whether it is because their district knows their name, they entice more donors the second (and third, and fourth, etc.) time around, or what have you, members of the House who have already served a term overperform their colleagues who are running their first race. This is the famous “incumbency advantage,” and in the U.S. House I estimate it to be about 5% on average.
What we call the incumbency advantage is really just how much the incumbent overperforms our benchmark of the district — in our case, the benchmark is simply Clinton’s share of the vote in the 2016 U.S. presidential election. This is a suitable benchmark for several reasons.
A simple demonstration of the incumbency advantage is thus: In the 2016 election, the average Democratic House candidate won 50.1% of the vote. Democrats ended up winning 194 seats in the House of Representatives. In those same congressional districts, Hillary Clinton won an average 51.7% of the vote, a 1.6% disparity. If each district had given an equal share of the vote to the Democratic House candidate the Democrats would currently hold 209 seats. This represents a gain of 15 seats for the Democrats if no incumbency advantage exists (and, of course, if voters picked the same party for their House and Presidential candidate, which they don’t always do).
This 1.6% vote / 15 seat disparity actaully grows if we apply it to our 2018 prediction. Rather, if we assumed that the Democratic candidate will do 1.6% better in each Congressional district (IE: if their incumbent retires), the Democrats would win an additional 47 seats on top of their current projection of 208 seats — for a grand total of 255 seats in the House.
Of course, not every district would see that 1.6% swing; some incumbents have a stronger advantage than the others.
Our goal is to figure out how much better a given Democratic candidate would do in a district where they’re running against a generic, unelected Republican rather than the sitting incumbent. To do this, I use my 2018 forecast model’s current predicted share for the Democratic candidate and take away the Republican’s incumbency advantage for that district.. Using this process, we can answer the question of how large a share of the vote would we expect the Democrat to win in any given district if the incumbent retires? Is that better for Democrats, or worse?
The most valuable retirements for the Democrats are, by nature, the districts where the Republican had the largest incumbency advantages. Or is it? What Democrats really want are the retirements that will push their candidate over the edge in a district they are otherwise projected to lose. There are 47 of these districts:
|District||Dem. % 2014/2016||Clinton %||Forecast||Forecast if Retirement||Retirement Bonus|
There are some obvious cases included in this list, but some that are less so. Still, we have to consider that not all these members are the type of congressperson you would see in a retirement announcement; mostly, retirements come when a member is aging or they’re running for higher office. Some of these folks are just getting started!
Suffice to say, there are some key Republican lawmakers that would make great retirement pickups for the Democratic party. Naturally, there are also some Democratic incumbents that would make great retirement pickups for the GOP – they’re included in the table above, though aren’t the focus of this post.
Aren’t Democrats Already On a Path Toward Victory?
History is on the Democrats’ side.
Since 1972, the party in the White House has lost an average of 0.5% ground in generic ballot polls from now until election day. Even better for them: Since the 2000 election the average out-party has gained 1.5% in the generic ballot.
That’s the exact margin they need to gain to have majority odds at victory, per my model and the fifth bullet point at the beginning of this article.
What may encourage the left even more is that they have gained more than that much ground in my projection of the 2018 national vote since July. While it’s unlikely that they will continue to gain half a percentage point per month until the midterms — that would put them at an unreasonable 60% of the vote on election day — an increase is likely to continue.
The argument “history will save Democrats!” is a hard one to make in the context of the current House map, but the math adds up to say that the gains we can expect them to make will at least have them favored to win on election day even if our model only says that they’ll win 51% of the time. Notably, I’m not saying history will save them, I’m saying history could. But for every percentage point the Democrats gain in the generic ballot the likelihood they win the House increases about 20%, so they have a lot of room on which they could build a hefty probability of victory. Still, the chance they can build that foundation in the first place is lower than one might expect.
What About the South(western) Strategy
Democrats have been trying to forge a productive electoral coalition out of minority voters for decades, and it didn’t work out so well in 2016. Though midterms elections are white and more educated than presidential cycles, we have some reason to believe that congressional districts in the Southwest will be the backbone of a Democratic Wave.
For starters, 11 of the 47 “most valuable retirements” we detailed about above are located in California, Texas or Arizona. Seven more are in Florida or Virginia, areas of the country that slightly resemble the southwest in demographics and income — and, obviously, politics.
Moreover, 11 of the 23 Republican-held congressional seats won by Hillary Clinton are in the same three states. Only 4 of these 22 seats are repeats — which means there are 18 unique seats that Democrats have fair odds of flipping that are in the Southwest.
Take Arizona’s second congressional district as a prime example. Incumbent Representative Martha McSally ® is facing an onslaught of enthusiastic Democrats in her district. There are six members of the opposition, including former congresswomen Ann Kirkpatrick (D), running for their party’s nomination in the district. The district is a Whiter one with 82% of its residents being white, with at 30% college graduate rate (about average) and $44,000 median income (about 10,000 dollars below average). McSally’s district elected her with 57% of the vote, but 53 percent wanted Clinton in office.
Or look at TX-23’s Will Hurd ®. Hurd represents a district that he won by less than two percentage points and one where Clinton squeaked by with a four-point margin. The residents there, while less educated than those in McSally’s AZ-02, are white and somewhat urbaner, thanks to the seat’s adherence to a redistricting principle called “stove-piping” whereby a district is drawn to cover large swaths of urban land and “pipe” into an urban area. It’s projected to flip in 2018, though most observers see it as a truly Tossup contest.
Whatever seat you pick from this bucket, there is a good case to be made that it could be the backbone of a Democratic wave in 2018. And if there is a tsunami, however rare that would be, look to the Southwest to be the earthquake that sets it off.
Others have noted this trend in the data, and while this post offers a taste of what a fully-fledged Southwestern strategy could offer, there’s much more to be investigated here.
There are several ways that Democrats could go about picking off seats for a House majority. Their best hopes lie in some older Republican congresspeople getting tired of their jobs or looking towards more ambitious offices to vacate seats. Of course, the other options are good paths too. The trend of generic ballot polls from now until election day is in their favor, as is (more generally) their status as the opposition party in the Government. And how thankful Democratic activists and strategists in middle-California must feel right now. Finally, any one of the 23 GOP-held districts that Hillary Clinton won should be top targets for Democratic operatives — especially the four for which I calculated a favorable retirement bonus.
For now, the 2018 midterm cycle is more than 390 days away. We can only guess as to what Democrats might plan to do or what events could impact their chances at victory. But for now, the ball is almost entirely in their court. What play they decide to run could be the difference between a House majority in 2018, or falling just short.