Political Economy Forecasts: The 2026 Midterms

Abstract

We employ a political economy model to forecast the outcome of the 2026 midterm elections in the House and Senate.Earlier versions of our model have been utilized, to good effect, since 2010.The model draws on sound theory, which posits the importance of political and economic fundamentals.Midterm election contests are considered referenda on the president's party.If the White House party appears to be delivering a good performance, it is rewarded with more votes; otherwise, it is punished at the ballot box.We begin with a baseline political economy model containing three independent variables: presidential approval, income growth, and the term sequence.We adjust that model, by incorporating local conditions, and then by revising the specification to include variables that tap candidate quality and time in office.These revisions allow for a range of possible incumbent seat outcomes for the House in 2026, ranging from a lesser loss of 17 seats to a greater loss of 33 seats.We follow the same strategy for the Senate, where the forecast numerical range is quite narrow, -6 to -7.This suggests that the pickup by Democrats needed for Senate control (i.e., 4 seats) is more likely than not.