The recent growth in local school subsidies in Korea offers a chance to test the effect ofpopulation aging on public education expenditure. In the existing literature, theintergenerational conflict model suggests a negative effect, although a growing number ofstudies are finding theoretical and empirical evidence of a positive effect. After controllingfor the district fixed effect, the analysis of the Korean data shows that the elderly populationshare has a significant positive effect on the local school subsidy per pupil. This result isrobust after instrumenting the real proportion of the elderly population by the predictedproportion, assuming no population mobility, to handle the possible bias caused by Tieboutsorting.