Determinants and Effects of Local Public Education Expenditures - an Empirical Approach

There are many (international) debates on education and education expenditures from a political economic as well as from a scientific perspective. However, the role of German counties and local agents in Germany in general has not been analyzed from an economic perspective yet. This is the reason why my doctoral thesis investigates the role of German counties as the “forgotten” agent for the determinants and effects of local public school expenditures. This cumulative dissertation consists of three econometric papers – two of these papers deal with the determinants of the expenditures of German counties for vocational schools. The first paper addresses the question whether there is any interregional competition between counties concerning expenditures for vocational schools. Using a dataset on 193 West-German counties between 2001 and 2006 (including municipalities with county rights) several different spatial econometric specifications are applied. No spatial correlation is found so the authors conclude that there is no interregional competition between German counties in the field of vocational schools. Additionally, no significant differences in the teacher per student ratio between counties at the state border and interior counties can be shown. Secondly, a closely related research paper tries to answer the question if the political composition of German county councils – in terms of seat shares of different parties- exerts any influence on the level of county expenditures for vocational schools. A panel dataset of 301 West-German counties from 2001 to 2013 is analyzed using panel fixed effects models as well as mixed effects models. Our results indicate that expenditures increase in the Banzhaf-Index of the share of Christian Conservatives and decrease in the Banzhaf-Index of the share of Social Democrats in the county council. The latter index is a common measure of the voting power of a party. Furthermore, they appear to be higher in election years. Moreover, no significant impact of the decentralization measure applied here can be found. The third paper analyzes the short-run effects of public expenditures for grammar schools from different federal levels on the rate of successful grammar school finishers. As endogeneity is a problem usually found in related empirical research, standard IV and also Lewbel instrumental variable estimations are applied. The data set is similar to the one used in the second paper. Significantly positive effects are found for the teacher density as the state input (in this case only weekly significant) as well as for expenditures per pupil by counties for grammar schools as a local input. There does not seem to be any statistically significant impact of the decentralization measure applied here. In all three research studies, a self-collected regional panel dataset on the German counties is used which – among others- contains variables on public expenditures, socioeconomic characteristics, the structure of the labor force as well as political variables. All in all, an important role of German counties for public school expenditures can be shown by the analyses of two school forms.