A Political-History Forecast Model of Congressional Elections: Lessons Learned from Campaign 2022

Abstract

InDiscourses, Machiavelli opined, “it is easy, by diligent study of the past, to foresee what is likely to happen in the future in any republic.” The message: there is value in exploring history to predict. A strong pedigree of political science research acknowledges the importance of path dependence, “lock-in,” and “Laws of Politics.” Such recurrences at face bode well for forecasting. Election forecasting models traditionally use political-economic variables to predict results. In 2022, we formulated a model forecast of U.S. Congressional elections with a twist—spurning any public opinion or macroeconomy measure. Instead, we tested whether historical junctures, state-level party strength, and federalism dynamics offered solid guides to the performance of the Democratic Party, historically dominant in Congress since 1946. Our analysis demonstrated that this Political History model offered credible estimates of Democrats’ performance in thirty-eight Congressional elections from 1946–2020, with out-of-sample predictions