Economic theories of systems of cities explain why production and consumption activities are concentrated in a number of urban areas of different sizes and industrial composition rather than uniformly distributed in space. These theories have been successively influenced by four paradigms: (i) conventional urban economics emphasizing the tension between economies due to the spatial concentration of activity and diseconomies arising from that concentration; (ii) the theory of industrial organization as it relates to inter-industry linkages and to product differentiation; (iii) the New Economic Geography which ignores land markets but emphasizes trade among cities, fixed agricultural hinterlands and the endogenous emergence of geography; (iv) the theory of endogenous economic growth. Among the issues examined are specialization versus diversification of cities in systems of cities, how city systems contribute to increasing returns in national and the global economies, the factors that determine skill distribution and income disparity between cities, the impacts of income disparity on welfare, whether population growth should cause economic activity to become more or less concentrated in urban areas, and how resources should be allocated efficiently in a system of cities. Related to the last issue, we consider models where cities are organized by local planners or developers as well as cities that self-organize by atomistic actions. A conclusion of the theoretical study of city systems is that markets fail in efficiently allocating resources across cities when certain intercity interactions are present and that a role for central planning may be necessary.