Forecasting a President-Centered Midterm: Leader-Centered Personalization in the 2026 US House Elections

Abstract

Forecasting models of US midterm elections rely on indicators such as presidential approval, economic conditions, incumbency, district partisanship, and the generic congressional ballot.This paper asks whether these indicators fully capture presidential influence when congressional competition is organized around a highly personalized leader.It distinguishes electoral nationalization, in which local contests respond increasingly to national parties and conditions, from leader-centered personalization, in which those forces are organized around an individual leader.The paper identifies three mechanisms through which personalization may shape district-level electoral choice: candidate alignment, asymmetric mobilization, and electoral translation through campaign framing.It then combines a prespecified district baseline with a structured qualitative assessment of leader centrality and these mechanisms to generate prospective expectations about whether personalization should reinforce, weaken, or leave the baseline unchanged.The diagnostic is applied prospectively to four competitive 2026 US House races.The paper also distinguishes explanatory from predictive value, that is, leader-centered personalization may clarify how congressional accountability operates even when separately measuring personalization does not improve forecast accuracy.