Political economy models and UK election forecasting: End game?

Abstract

Political economy models have been applied to election forecasting for some time. However, in the United Kingdom, as well as elsewhere, other methodologies have come to the fore to take their place alongside the forecasting methodology of vote intention polling. Returning to a classic Political Economy model first successfully tested on the 2001 General Election, we ask whether it still has relevance today. After various time series analyses of UK general elections (1955 to the present), we find that it does. The model manages to forecast the vote share of the incumbent party rather accurately, via three predictor variables: economic performance, executive/prime ministerial approval, and the number of terms in office. For the 2024 contest, it forecasted, before-the-fact, a Conservative defeat of historic proportions.