It’s my pleasure to introduce The Washington Post’s Election Lab, a new feature that provides both data and forecasts for the upcoming 2014 House and Senate elections. Election Lab allows you to drill down to individual districts and states and see key factors in each race. It provides an overall forecast for each race — in short, how likely it is, at this moment, that the Democratic or Republican candidate will win that race. It gives you a “daily digit” at the top — a noteworthy number about the elections that will be updated every day.
Today’s daily digit is the current forecast for the Senate. Almost every analyst sees a significant chance for GOP gains in the midterm. The forecasting model that underlies Election Lab is currently bullish about Republican chances in the Senate. Right now, the model estimates that Republicans have an 82 percent chance of re-taking the Senate.
How did we arrive at this conclusion? You can read more about that model here, but in brief: the model looks at Senate elections between 1980-2012 and estimates the effect of several key factors in the country and in individual states or races — the rate of economic growth, the popularity of the president, whether it’s a midterm or presidential year, the most recent presidential election outcome in that state, whether the incumbent is running, and each candidate’s qualification (measured as highest elective office to date).
Why do these factors add up to a significant GOP advantage? The main problem for Democrats is that it’s a midterm year — and the president’s party almost always loses seats in the midterm. Moreover, conditions make it difficult for Democrats to overcome this tendency: the economy is not growing that strongly and, partly as a consequence, President Obama is not that popular. Moreover, as many have noted, many seats that the Democrats must defend this year are in Republican-leaning states.
Given these conditions, the political science literature suggests that quality Republican candidates should emerge. This is because quality candidates are strategic: They tend to run when their chances of winning are higher. Thus, many Republican candidates have significant political experience in state legislatures, the U.S. House of Representatives, and in other offices. (In states where primaries haven’t taken place, we assume that the eventual party nominees will have an average experience level like that of nominees in similar races in the past.)
Put all of that together, and it’s a recipe for a likely Republican majority.
Now, some caveats. What’s not in this forecasting model? Most notably, fundraising and polls — although both will become part of the model. Indeed, recent polls in some key Senate races — Kentucky, for example — show the candidates nearly tied, while our model forecasts an eventual Republican win. If those polls don’t change, then our model’s prediction will move in that direction as it begins to incorporate and weight polls more heavily. Of course, it’s also possible that the polls, not the forecast, will move. Research has shown that polls often move toward the outcome that fundamental factors predicted at the outset of a campaign.
So stay tuned for more updates from the forecasting model and more features at Election Lab.
