Date added: September 2017

Rene Bernard

FiFo/ December 2017/ Discussion Paper, FiFo-Köln

Abstract
The government fragmentation hypothesis (GFH) states that coalition governments spend more than single -party govern-ments due to an underlying common pool problem. Using a large panel data set on 604 local governments in the German state of Baden-Württemberg for the 1994-2014 period, I test the GFH for tax rates as well as expenditures and its sub-cate-gories. Studies using standard regression methods fail to iden-tify causal effect as the type of government is generally not random. I apply a RDD, that exploits quasi-random variation generated by close elections. I add external validity to the re-cent quasi-experimental literature by investigating the GFH for a mayor-council system. I find that contrary to the theoreticalprediction, coalition governments do not increase taxes. There is a non-robust, negative effect on total expenditures, which is mainly driven by administrative expenditures and material ex-penditures.

Download directly from our server :

Download