Revisiting Clarke and Stewart’s (1995) investigation of PM approval and vote intentions in the UK

Abstract

Economic Evaluations, Prime Ministerial Approval and Governing Party Support: Rival Models Reconsidered by Harold D. Clarke and Marianne Stewart (1995) makes major contributions both substantive and methodological. Substantively, the article contributes to literatures on economic voting and the presidentialization of British politics. Methodologically, the paper is a model of how to conduct cointegration and error correction analyses between closely related political time series. In all, the authors established a long-run equilibrium between prime ministerial approval and vote intentions for the governing party. The two variables may diverge for short periods, but vote intentions will always revert back to a level that is in line with the popularity of the PM. We extend Clarke and Stewart’s data to the present day and examine how the PM-vote intention link has endured while the effects of the economy on vote intentions have varied. We also show that, while time series analysis in political science has undergone several revolutions, Clarke and Stewart’s process of conducting two-step cointegration analysis (Engle and Granger, 1987) stands up 30 years later as a blueprint for how to investigate long-run relationships with political data.