Replication data for: \"Does Partisanship Really Matter for the Provision of Welfare Policies?\"

This paper examines and applies of more advanced modeling methods for the time-series-cross-sectional data by replicating the article by Huo et al. (2008). We make a claim that the original article’s causal inference is problematic without taking consideration of country-level heterogeneity, even though we agree their defense of not using country dummies: They absorbs the most of variances in the time-invariant or slowly changing factors which are common in institutional analysis, and thus analysts are not able to estimate cross-sectional variances which are great interests of comparative political studies. Therefore, we propose and apply more advanced method actively measuring cross-country variances. Followed by Beck and Katz (2007) we estimated random coefficient model (RCM) with most typical TSCS data in comparative political studies. By examining a variety of aspects of RCM, we find evidences that the main theoretical arguments in Huo et al. (2008) are not supported.