Winning the Presidency Negatively Impacts a Party’s Voter Turnout

In the midterm election following a party change in the White House, the President’s party generally incurs a net turnout decrease.

In a conversation about Andrew Gillum’s surprise upset in Florida’s Democratic gubernatorial primary, an acquaintance declared wantonly that one of the biggest challenges Gillum will face is the confrontation with a surge in voters precipitating from Trump’s energizing force among his Republican base.

It has been a piece of conventional wisdom for some time that Republicans register better turnout in midterm elections than Democrats (some data below draws this into question), but 2018 is a special kind of mid-term — one that immediately follows a party change in the White House.

It is easy to think of the way the Democratic base abandoned Obama in the 2010 midterm amid the rise of the Tea Party and assume the much-hyped “Blue Wave” of 2018 will have a similar impact, but one data point does not a data set make.

The Data Set

In order to analyze party-line turnout, I pulled the available historical results for House of Representatives elections. Election data is preferable to registration data because party registration and party identification (that is, the way someone actually votes) are statistically distinct.

I chose House elections because, unlike Senate and Governor, these elections happen every two years, providing some uniformity in the voter pool. And, since most districts are smaller than the states they are in, the average House election is isolated a bit more from noisy data effects such as an atypical race or a controversial candidate.

The statistic I calculated was the proportion of the eligible voting population that voted for each party. This normalizes for population shifts, and also accounts for overall turnout changes that are washed away in analyses of strict voting margin.

This is a good start, but of course, there is a ton of noise from Presidential elections (the spikes every other election). Luckily, we are only investigating midterms. These are the elections between the spikes. Here’s what they look like.

Now we’re seeing some signal beneath the noise!

Results

We are not necessarily interested in every midterm (although the graph above does seem to challenge the conventional wisdom that Republicans show up more consistently to midterms). We want to look at the midterm elections immediately after a party change in the White House. Those party-change Presidential elections are represented by the vertical dashed lines below.

Next, I calculated the change in the voting population percentage from one midterm to the next. Note that this means we are not comparing the midterm to the preceding Presidential election year, but to the previous midterm election. (For example, 2014 is compared to 2010.)

The results are below, with “hangover” elections (those following a party change in the White House) highlighted with the color of the new President’s party. “Net hangover” is the change for the President’s party minus that of the opposition.

As a word of caution: eight data points is a small sample size. Still, it bears consideration that with the exception of 2002 — post-9/11, pre-Iraq War, when George W. Bush still had an approval rating well above 60% — the new President’s party has incurred a net loss in the hangover midterm.

Although the impact of the hangover hits Democrats harder than Republicans (the last two Democrat hangovers — 1992 and 2008 — were historic for the numbers of seats lost) there is a non-negligible hangover for Republicans as well.

Trump is, of course, quite strange. None of this data addresses his perpetually increasing oddity as a politician and a President. However, as a correlary for a Republican President with an energized base, consider Reagan’s hangover in 1982. Despite an increase in Republican voter turnout compared to 1978, the Democratic Party turnout increased even more.

That year, the Democrats gained 26 seats in the house. A 26-seat gain in 2018 would give the Dems a 219–211 majority.