Linguistic Distance and Market Integration in India

We collect data on grain and salt prices, as well as language, for more than 200 South Asian markets in the 19th and early 20th centuries. Conditional on a rich set of controls and fixed effects, we find that linguistically distant markets are less integrated as measured by the degree of price correlation. While linguistically distant markets exhibit greater genetic distance, greater differences in literacy, and fewer railway connections, these factors are not sufficient statistics for the negative correlation between linguistic distance and market integration. Our results indicate that a one standard deviation increase in linguistic distance predicts a reduction in the price correlation between two markets of 0.121 standard deviations for wheat, 0.167 standard deviations for salt, and 0.088 standard deviations for rice. These differences are substantial relative to other factors such as physical distance that hinder market integration.