Worlds of welfare in time: A historical reassessment of the three-world typology

Abstract

Gøsta Esping-Andersen’s ‘three worlds’ typology has become one of the principal heuristics for examining modern welfare states. Despite its widespread adoption, however, few efforts have been made to assess the typology’s validity across time. To gauge its historical relevance, this study examines welfare state data for each 5-year interval from 1950 to 2000 for evidence of tripartite clustering. Using two forms of cluster analysis – an agglomerative hierarchical approach and a model-based approach – on each of these cross-sections, this study finds that Esping-Andersen’s typological framework is largely incapable of identifying a lasting and meaningful three-world arrangement. To detect such a configuration, this study shows that the framework must be extended to include new dimensions relating to social services, gender, poverty, and activation. When this extended framework is employed, this study identifies three worlds of welfare that first began emerging by 1975, became salient by 1980, and had firmed by 1985.