How reliable is pooled analysis in political economy? The globalization-welfare state nexus revisited

Panel data analysis has become very popular in comparative political economy. However, in order to draw meaningful inferences from such data, one has to address specification and estimation issues carefully. This paper aims to demonstrate various pitfalls that typically occur in applied empirical work. To illustrate this, we refer to the debate on the globalization-welfare state nexus. We reexamine a model by Garrett and Mitchell (2001), a leading study in this regard. Utilizing a data set of 17 OECD countries and the time period 1961 to 1993, they find evidence that globalization and partisan composition have a significant impact on the extent of public activity. However, because they apply a dynamic specification in levels, they do not adequately take into account both the dynamic and spherical nature of the data. In contrast, we propose an autoregressive model in first differences that is shown to perform well in statistical terms. Further, we explicitly pay attention to the time pattern of the globalization-welfare state nexus. Substantively, we find evidence that government spending is primarily driven by the state of the domestic economy. Neither partisan effects nor the international economic environment have affected public expenditure considerably.