Election Forecasts Favor Republican Gains In Midterm
In the weeks leading up to the 2010 midterm elections, five forecasters or teams of forecasters offer models and predictions for the House in the most recent issue (October 2010) of PS: Political Science and Politics, a journal of the American Political Science Association. The models offer a broad consensus that the Republicans will make substantial gains in the House, although there is not a consensus over how large those gains will be. A 30-seat spread between the low and high end of the seat change forecast range exists, with two forecasters giving an edge to Democrats maintaining their control of the House and three forecasters anticipating Republicans retaking the chamber.
James Campbell of the University at Buffalo, SUNY, presents a “seats in trouble” model for his forecast. The model is based on the Cook Political Report’s district-by-district assessment of the status of House races. He finds that the more seats a party has that are rated as toss-ups or worse before Labor Day, the more seats they actually lose in November. His model combines Cook’s expert handicapping of districts with national variables such as presidential approval and the number of seats a party holds going into the election. Campbell’s model combines the in-depth analysis of the Cook Report with the rigorous historical analysis of national statistical models. Based on his model, he forecasts that Democrats will lose 51 or 52 seats, allowing the Republicans to regain a majority.
Models created by Alan Abramowitz (Emory) and Joseph Bafumi (Dartmouth), Robert S. Erikson (Columbia), and Christopher Wlezien (Temple) corroborate the prediction of steep Democratic losses. Abramowitz uses the generic ballot and presidential approval to measure the national political climate, estimating that Republicans will gain 43 seats in the House. He further offers a forecast for the Senate, predicting a Republican gain of four seats. Bafumi, Erikson, and Wlezien give the Republicans a 79% chance of retaking the House, forecasting a 229 Republican to 206 Democratic split of seats, but they note that there is a wide dispersion of possible outcomes, and that seemingly minor variation in the national vote can have major consequences for the distribution of seats.
In contrast, the forecast by Michael Lewis-Beck (University of Iowa) and Charles Tien (Hunter College and CUNY) and the forecast by Alfred Cuzán (West Florida University) both predict that the Democrats are likely to maintain their control of the House. Lewis-Beck and Tien rely on established voting behavior theory in the formulation of their model and estimate a Democratic loss of only 22 seats, allowing the party to retain their majority. Cuzán uses a structural model relying on national-level variables and election outcomes as far back as 1914 to forecast Democratic losses of 27 to 30 seats, staving off Republican attempts to retake control. However, this model leaves a one in three or four chance that the Democrats will lose at least 40 seats, reducing them to a minority status.
Stepping beyond the collection of congressional predictions is Carl Klarner’s (Indiana State University) forecast of state legislative elections. He calls attention to the 43 state legislative elections that will be held this November, noting that many of the winners will have the responsibility for drawing new district lines based on the 2010 census. According to his model, Republicans will gain control of 11 legislative chambers, leaving them with a total of 46 of 98 partisan chambers, and the Midwest is shaping up to be a battleground of note.
Primaries held in a number of states in mid-September offer a taste of how hotly contested the midterm elections in November are likely to be. Whether Democrats or Republicans ultimately assume control of the House, Campbell notes that this majority is likely to be much narrower than the current Democratic majority, which may present a roadblock to the Obama administration’s legislative agenda and will quite probably make control of the House a real question again in 2012.
On the Net: