Categorising Executive-Legislative Relations

Characterising executive-legislative relations can frustrate due to subjectivity in classifying hybrid (or “semi-presidential”) forms that combine elements of presidentialism and parliamentarism. Much depends on how researchers determine the number of types of semi-presidential systems and which countries are classified to those types. The difficulty is often striking a conceptual balance between particular Constitutional arrangements and the exercise of executive power, along with standard issues of aggregation in any measurement exercise. Debate over the number of categories exists, with 3 (with semi-presidentialism) and 4 (premier-presidentialism, president-parliamentarism) categories the most common, but larger numbers of categories can be argued for as well. Furthermore, within these categories are wide variations in executive powers, leading to debates over classification concerning particular countries. To avoid such arbitrariness, we apply a measurement model for identifying typologies (latent class analysis) and apply Bayesian methods developed by one of the authors to simultaneously estimate the number of overall categories along with category membership. We apply this model both in an “exploratory” fashion, across all types in a general characterisation of executive-legislative relations, along with exploring variation of presidential power within traditional delineations (e.g., apply the same model on indicators of executive power exclusively within commonly defined boundaries for parliamentary, presidential and mixed systems). We apply these models to data from the Comparative Constitutions Project, Quality of Government Institute, and V-DEM, and compare to existing operationalisations.