School spending and extension of the youth voting franchise: Evidence from an experiment in Norway

Changes in population age composition is challenging in modern welfare states. Intergenerational conflicts may have important consequences for provision of services directed towards specific age groups as schooling and care for elderly. A relevant question is to what extent the supply side responds to changes in the age composition of the electorate in terms of actual spending policies. This paper exploits a novel experiment that took place in Norway in the 2011 local elections to estimate the causal relationship between local government school spending and the age composition of the electorate. We exploit that the voting age was reduced from 18 to 16 years in local elections in selected local governments (experimental governments), while voting age was kept at 18 in the rest (control governments). Using a difference in differences strategy, we find that compulsory school spending decreased by approximately 2% in the experimental governments. The results are robust across a number of econometric specifications and robustness checks. Since all the newly enfranchised voters had just finished compulsory school and receive no direct benefits from local government school spending, the result is consistent with selfish voter behavior.