Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper analyses whether the propensity to secede by subnational regions responds mostly to differences in income per capita or to distinct ethnolinguistic identities. We explore this question in a quantitative political economy model where people’s willingness to finance a public good depends on their income and identity. Using high-resolution economic and linguistic data for the entire globe, we predict the propensity to secede of 3,153 subnational regions in 177 countries. We validate the model-based predictions with data on secessionist movements, state fragility, regional autonomy and conflict, as well as with an application to the dissolution of the Soviet Union. Counterfactual analysis shows that removing ethnolinguistic identity differences reduces the average support for secession from 10.1% to 0.2% of the population, while removing income differences has no major quantitative impact. Although both forces affect secessionism, identity trumps income in determining a region’s propensity to secede.