Whether urbanisation promotes or inhibits the risk of civil war is disputed: while case studies usually support the former, quantitative investigations have found either the latter or no significant correlation at all. I argue that this contradiction is due to a conceptual and operational over-aggregation of urbanisation, ignoring its intrastate variation. I claim that a high relative concentration of the urban population and political, economic and social institutions in the largest city -so-called metropolisation -can increase both the motivation for and the feasibility of rebellion in a country. Triangulating case study evidence with a quantitative cross-national time-series design, I show that metropolisation significantly and robustly increases the risk of governmental conflict in particular and hence civil war in general.