A theoretical framework for evolutionary economic geography: industrial dynamics and urban growth as a branching process

We propose a framework that specifies the process of economicdevelopment as an evolutionary branching process of productinnovations. Each product innovation provides a growth opportunityfor an existing firm or a new firm, and for an existing cityor a new city. One can then obtain both firm size and city sizedistributions as two aggregates resulting from a single evolutionaryprocess. Gains from variety at the firm level (economies ofscope) and the urban level (Jacobs externalities) provide thecentral feedback mechanism in economic development generatingstrong path dependencies in the spatial concentration of industriesand the specialization of cities. Gains from size are also expected,yet these are ultimately bounded by increasing wages. The contributionof our framework lies in providing a micro-foundation of economicgeography in terms of the interplay between industrial dynamicsand urban growth. The framework is sufficiently general to investigatesystematically a number of stylized facts in economic geography,while at the same time it is sufficiently flexible to be extendedsuch as to become applicable in more specific micro-contexts.A number of extensions related to the concepts of knowledgespillover and lock-in, are also discussed.     

An information-theoretic approach to the analysis of location and colocation patterns

The study of location and colocation of economic activities lies at the heart of economic geography and related disciplines, but the indices used to quantify these patterns are often defined ad hoc and lack a clear statistical foundation. We propose a statistical framework to quantify location and colocation associations of economic activities using information-theoretic measures. We relate the resulting measures to existing measures of revealed comparative advantage, localization, specialization, and coagglomeration and show how different measures derive from the same general framework. To support the use of these measures in hypothesis testing and statistical inference, we develop a Bayesian estimation approach to provide measures of uncertainty and statistical significance of the estimated quantities. We illustrate this framework in an application to an analysis of location and colocation patterns of occupations in US cities.