The effects of cultural diversity on innovation and public provision of schooling: Evidence from nineteenth century Prussia
Final Report Abstract
Aim of the project was to estimate the effect of religious diversity on innovation examining historical data of nineteenth-century Prussia. We tested the hypothesis that an economic environment characterized by religious pluralism is more open to new ideas and is conducive to technological progress. To study this question we digitized county-level data from population, education, and occupational censuses which allow reconstructing the religious composition of the population. Religious diversity is measured through the fractionalization index. These data have been matched with individual-level data on valuable patents granted in the period 1877-1890. Our main results indicate a strong and robust positive relationship between religious diversity and innovative activity. One of the main findings is that the relationship between religious diversity and innovative activity is non-linear with decreasing marginal returns. We address the issue of causality using a shift-share approach. By using the religious composition of the county population in 1816 augmented by denomination-specific population growth rates at the national level for the period 1816-1871 we impute the religious composition of counties in 1871. The imputed religious composition in 1871 is then used as an instrument for the observed religious diversity in 1871. The logic behind this approach is that the identified variation in 1871 depends only on the initial religious composition of the population and it is not affected by denomination-specific migration flows between 1816 and 1871. Instrumental variable estimates support the result that religious diversity was conducive to innovation. The instrumental variable estimates support also the finding of a non-linear effect of religious diversity on innovation. Regarding the mechanism, we find that religious diversity is positively associated to the probability of finding two people employed in different industrial sectors. This is consistent with the notion that religious diversity fostered production diversification. This result is confirmed by the analysis of town-level data for 1816. Exploiting variation across towns within counties we find that religious diversity is positively related to diversity in craftsmanship trades in 1816 and to the number of industrial establishments. Finally, we unveil important effect heterogeneity. We find that the positive effect of religious diversity on innovative activity is mostly relevant for large-firms innovation and in counties with a relatively higher level of human capital. These findings are consistent with the notion that religious diversity and pluralism are important drivers of innovation in environments characterized by social interactions between highly educated people as in R&D departments.
