Corruption in space: A closer look at the world's subnations

The level of corruption differs not only between countries, but also between subnations within countries. In this paper, we analyze spatial interdependencies in corruption levels for a large sample of 1,232 subnations in 81 countries. Based on a spatial autoregressive model, which additionally corrects for spatial autocorrelation in the error term, we find that a subnation’s corruption level is positively affected by neighboring subnations’ corruption levels. This suggests that subnational corruption levels are strategic complements. Extending the core model and allowing for heterogeneous spatial interdependencies our results indicate that in particular high income subnations and subnations with a relative low corruption level tend to spill in space. This is due to their high degree of connectivity in terms of economic, sociocultural and political exchange with other subnations. Our findings underline the importance to consider not only a subnation’s own characteristics but also their degree of connectivity with other subnations when implementing effective anti-corruption policies on a local level.