As of last Tuesday, Democrats are averaging a 27% swing in special elections since the 2016 election. This has translated into a net gain of 40 total legislative seats, including the Alabama Senate seat. If Democrats had a 27% lean in the national vote, not only would they flip the House, but they’d get over 300 seats and, assuming Trump’s approval stays around 55-40, would flip 5 Senate seats, giving them a majority in both Congressional houses. This is, of course, highly unlikely, but it shows the extent of the Democratic wave and how much Republicans should be worrying. Just for fun, here’s what the House and Senate races would look like under the current environment.
Notice how Democrats make gains into unlikely areas in the House race, such as Utah and many North Carolina seats. That is because of the downside to gerrymandering: vote spreading. When Republicans gerrymander seats to give themselves a majority, they simultaneously spread Republican voters across multiple districts. Instead of having a lot of Republicans in few districts, they draw borders to give each district just enough Republicans to secure the seat. This results in a lot of 55-45 or closer districts, meaning that a large enough wave can begin to flip seats exponentially.
To be clear, the scenario above is highly improbable. It could technically happen, but Democrats probably won’t have a 27 point lean come November. However, the average special election swing can show us some noteworthy things. First of all, Democrats have a backed momentum advantage. When I say backed, I mean that their momentum is proven by actual votes. These aren’t polls or predictions, these are raw votes cast by real people in real elections. More Democrats are turning out, proving that voters are excited and active going into the midterms. This is extremely important for Democrats, as they most likely lost the 2016 elections due to overconfidence and therefore lower turnout. Liberal voters were so sure of a Clinton victory that many of them didn’t feel the need to vote, and many Progressive voters didn’t turn out because of Clinton’s center appeal. So far, special election results show that the party as a whole is turning out in large numbers, no matter how likely a victory. A wave cannot be a wave unless there is water, and for Democrats, it’s raining. The second thing the special election results can tell us is the campaign effectiveness of the Democratic Party. While the DNC largely focuses on resisting Trump and Republicans, many Democratic candidates choose to focus on local issues specific to their voters. Each campaign method has proven to be successful, and a mix of both can help turn out Progressives and Moderates for a single candidate. With Trump’s approval ratings being so low, campaigning against him on a national level will be instrumental in turning out the Democratic base. However, independents and moderates need more than just a unity against something, they need a unity for something. Local candidates are providing that unity and, with the help of the national sentiment, winning elections because of it. The third and final thing the special election swing can tell us is the expected 2018 national vote margin. Since 1994, special election swings have predicted the national House margin in midterms with just a 3 point difference on average. On top of that, the 3 point difference doesn’t always hurt the party gaining on the swing, it is actually about 50/50 help/hurt. It should be noted that in 2006, when Democrats had a 15 point average swing, they only won the House vote by 8 points, recording a 7 point difference and the largest since 1994. If historical statistics is a predictor of future results, then Democrats could have anywhere from a 20% to a 34% lead in the national House vote. Those numbers seem extremely high, but keep in mind this is using data from actual elections throughout the past year. An outlandish 34% may not be as outlandish as one would expect. The only data point keeping pundits from expecting an all-out blowout is the generic ballot poll. While Democrats have averaged a 27% swing in special elections, the average generic ballot poll only gives them about a 9% lead, meaning the House and Senate would be much closer. It’s hard to say which is a better predictor, as both have their flaws. Special elections can be swayed by odd turnout demographics, while polls can obviously be swayed based on who is polled. However, as Nate Silver points out,
But important to remember that generic ballot polls are, at this early stage, mostly of registered voters rather than likely voters. Not crazy to think the Dem advantage could grow once pollsters switch over to likely voter models in the fall.
— Nate Silver (@NateSilver538) February 28, 2018
Basically, pollsters are polling all registered voters, which could be less favorable to Democrats than likely voters. Special elections, being an election, are a “poll” of only actual voters, not registered voters. At this point in the election cycle, a 60-40 weighted average of the two metrics would be the best predictor, in my opinion, weighing the special elections more than the generic ballot polls. Using the 60-40 weighting, Democrats have a 19.8% advantage, which, while still huge, is not quite as large as the 27% lead they currently hold. They’d still be favored to win both the Senate and the House, but everything comes down to actual turnout and who gets out to vote. Luckily, we have another special election coming up soon, so we can better predict favorability with an ever larger election sample. You can expect an article after the Pennsylvania 18th Congressional District Special Election on March 13th.