Posted on March 19, 2014
Last Tuesday was the first congressional election of 2014, and pundits had a lot to say about it. The race, a special election in Florida's 13th district, saw Republican David Jolly narrowly defeating Democrat Alex Sink.
National Republican Congressional Committee Chairman Greg Walden trumpeted that "Tonight, one of Nancy Pelosi’s most prized candidates was ultimately brought down because of her unwavering support for ObamaCare, and that should be a loud warning for other Democrats running coast to coast."
The Hill wrote that "It’s hard to see Jolly’s win as less than evidence that Democrats still face a turnout problem and difficulties with ObamaCare this cycle."
The National Journal credited Jolly's "upset" win to a fancy new Republican voter database.
Even the New York Times confidently stated that "Tuesday’s special election showed that [Sink's] campaign could not outrun the tsunami of advertisements tying her to President Obama’s health care law."
These conclusions were largely baseless. Everyone wanted to find a story in the candidates, or the campaigns, or a supposed referendum on Obamacare. But what if you ignored all those factors, and had predicted the outcome of the race based solely on the partisanship of the district? You would have been accurate to within 0.2%.
Florida 13's partisanship (how it voted for president in the 2012 election relative to the candidates' national averages) is 51.2% Republican. In an open seat race with no incumbency advantages, our Monopoly Politics model therefore projected that David Jolly would receive 51.2% of the two-party vote. In reality, David Jolly received 51.0% of the two-party vote.
Even in an highly competitive district, a slight edge in partisanship proved to be decisive.
Special elections are somewhat unique electoral environments, meaning that the outcome of this race cannot necessarily be extrapolated to predict what will happen in November. Still, the results of Florida 13 suggest that neither party has a significant advantage in terms of nationwide support at this point in the election cycle.
A "balanced" 2014 election is bad news for Democrats in the House. Due to the way congressional districts are drawn and the incumbency advantage of Republican representatives, our model suggests that if voters nationwide are split evenly in their support for Democrats and Republicans in November, Republicans will win 235 districts and Democrats just 200.