From Variables to Mechanisms in Federalism Research

Federalism merits attention because of how it affects things that matter—outcomes as fundamental as democracy and the welfare state. This article critiques the common, variable-based approach to understanding federalism’s impact, arguing that it violates multiple rules of good inference such as the number and independence of cases or consistent variable coding. It then proposes an alternative, case-based approach that tries to understand cases as constituted by the interactions of variables (mechanisms) rather than as opportunities to identify the impact of single variables. It shows the method in the abstract, using a hierarchy of variables to understand the development of the welfare state in federations with illustrations from the United States. The impact of federalism and interesting research questions, become clearer by examining the mechanisms at work in its interaction with other variables in a concrete case.