Panel data analysis and partisan variables: how periodization does influence partisan effects

Abstract

One central result of macro-quantitative studies in comparative public policy is that the importance of partisan politics on policy outputs has strongly decreased in recent decades. This finding may well be a methodological artifact. I argue that ad hoc standards in panel data analysis, especially using country-years as periodization, create estimation problems which potentially influence results against partisan variables. Therefore, I propose a simple and straightforward, as well as theoretically suitable, alternative to test the influence of partisan politics on policies and use cabinets instead of country-years. Using comparative welfare state research as an example, I show that partisan effects are strong and stable when using a cabinet-based periodization and fragile and weak within the standard procedure based on annual data. This paper aims at suggesting that annual periods do not need to be the best simplification of time in empirical analyses.