How does government size affect economic growth? : new results from a historical dataset

This paper contributes to the still unresolved issue of the growth impact of government size by analysing a historical panel data set of 17 developed countries that ranges from 1880 to 2016. The unique feature of the long-time dimension allows for conducting a kind of natural experiment. Government size is closely related to economic-policy paradigms. The time span covers different economic policy paradigms, in particular, 'laissez-faire' before World War II and Keynesian economic policy after World War II. Before WW II government size is small, after WW II it is (has grown) big. Furthermore, this paper contributes to filling a gap in the literature by testing the non-linear hypothesis (Armey curve). We take particular attention to a key shortcoming of panel-data analysis - parameter or individual heterogeneity. Overall, this analysis suggests a systematic positive, albeit quite small, linear relationship of government size with economic growth. As a consequence, rather than concentrating their attention to the sheer size of government, policy makers are advised caring for an efficiently run and highquality government sector as a prerequisite for a steady growth path.