Forthcoming in the American Journal of Political Science

The presence of large cities increases the probability of authoritarian breakdown, but the literature has offered little empirical insight as to why cities are dangerous beyond noting the concentration of protest in urban areas. I develop a theory of cities as complex socio-political spaces that are difficult to govern, particularly in the absence of democratic institutions. This complexity makes both cooptation and coercion difficult, meaning the very tactics that authoritarian cities use to control discontent can become its proximate cause. Using a large, city-financed housing project in Moscow targeted at rewarding regime supporters, I utilize a Bayesian semi-parametric model to demonstrate that even a seemingly well-targeted clientelistic exchange contributed to a surprising defeat for the regime in a subsequent municipal election. My results suggest that treating cities as simple agglomerations distinct from rural areas obscures important dynamics in the development of opposition to authoritarian rule.