This paper test the hypothesis that cultural distance between groups impedes local public goods provision. Using a unique historical dataset of surnames in the Lower Yangzi area of China, we are able to measure cultural distance between migrants and natives (MNCD) at a county-level. We exploit the quasi-exogenous mass migration following the Taiping Rebellion in the mid-19th century for variation in MNCD. Using a fully flexible difference-in-differences model, we find a one-standard-deviation increase in MNCD is associated with 29.7% decrease in primary school enrollment rates in the early 20th century given a structural change in the financing of primary schools. This finding is robust to the inclusion of a variety of controls and to various subsamples. In addition, our results survive the consideration of competing hypotheses including population diversity, migrant human capital, war destruction, age structure and missionary influence. We propose weakened decision-making power as a possible mechanism through which MNCD impedes public goods provision. This suggested mechanism is supported by the fact that only schools relying on the decision making of self-governed local communities are negatively affected in their enrollment. MNCD has no effect on schools established and financed by upper-level governments