Regulation of Religion: Identification of Cross-Country Similarities and Differences with Help of Novel Cluster Analysis Techniques

Quantitative research on regulation of religion has so far focused on explaining the cross-country variation in various types of regulation by employing regression analysis as the method. This can be contrasted to other fields of public policy research, (e.g., social policy, law and order), where cluster analysis has been increasingly and successfully used to classify countries based on policy outputs. Inspired by the recent methodological advances in the field of public policy, this paper employs data from the third round of the Religion and the State- project (RAS3) and uses various cluster analysis techniques (hierarchical clustering as well as the new technique of clustering ensembles) to identify country clusters based on the way in which countries regulate religion. Clustering ensembles (or consensus clustering) allow for the application of a wide variety of clustering algorithms to verify and compare results in detail between different algorithms (K-means, GMM, Spectral Clustering and Affinity Propagation). The research focus will be on four regulatory categories, namely government regulation of the majority religion or all religions; government discrimination against minority religions; societal discrimination against minority religions and specific types of religious support. Further, we separate between the qualitatively different types of regulation following the grouping of the indicators in the RAS3-dataset. Concretely, the paper will answer the following questions: (1) How do countries cluster when it comes to regulation of religion, and how many clusters are there? (2) How does the clustering of countries change depending on the employed indicators and cluster analysis techniques? (3) What are the defining characteristics of the different regulatory clusters, that is, what are the main similarities and differences across the countries when it comes to regulation? Given the argument that democratic and authoritarian countries should be studied separately when it comes to regulation of religion, the paper gives particular emphasis on understanding how the countries clusters if all countries are studied jointly, or if democracies and autocracies are studied separately. The findings contribute to the discussion about cross-country similarities and differences in regulation of religion. The clustering of countries should also be of assistance for the case selection for qualitative comparisons between the countries.