Urban life cycle and long-run violence: Colombia 1938–2018

Abstract

ABSTRACT This paper revisits and improves upon the traditional urban life cycle theory, using it as a long-term conceptual baseline model to assess the effect of the Colombian internal armed conflict on urbanization during 1938–2018. The paper makes three innovations: (1) It uses a third-degree autoregressive panel estimation to detect the underlying Data Generating Process of the urban life cycle, a feature that eluded original scholarship in the field; (2) It uses the baseline urban life cycle model to assess the impact of long-term violence in Colombia; and (3) It produces an inductive conceptual approach to the relationship between urbanization and economic development. Our third-degree autoregressive panel models adequately explain the urban concentration cycles experienced by Colombia’s 20 largest metropolitan areas, regardless of using different specification structures. It also correctly controls the long-term trends of the demographic transition that the country experienced during that period: its rate of urbanization increased from 31 to 68% between 1938 and 2018; the largest 20 metro areas increased their participation in the total population from 17 to 54%; and yearly total population growth increased from 2.12% in the 1940s to its peak 3.19% in the 1970s, decreasing to 1.18% in the 2010s. The homicide rate had a controlling effect on the increasing parts of the urban life cycle, acting as a deterrent of urban concentration per metropolitan area.