DP15677 Urban specialisation; from sectoral to functional

Author(s): Antoine Gervais, James R. Markusen, Anthony Venables
Publication Date: January 2021
Keyword(s): agglomeration economies, firm organization, fragmentation, Geographic concentration, multiple equilibria, Regional specialization
JEL(s): F12, F23, R11, R12, R13
Programme Areas: International Trade and Regional Economics
Link to this Page: cepr.org/active/publications/discussion_papers/dp.php?dpno=15677

The comparative advantage of many cities is based on their efficiency in the production of 'functions', e.g., business services such as finance, law, engineering, or similar functions that are used by firms in a wide range of sectors. Firms that use these functions may choose to source them locally, or to purchase them from other cities. The former case gives rise to cities developing a pattern of sectoral specialization, and the latter a pattern of functional specialization. This paper develops a model to investigate circumstances under which either of these outcomes is more likely, and finds that predictions of the model are consistent with changes in the pattern of specialization in the US over recent decades. The model combines elements of the literatures on economic geography, multinational firms, urban economics, and trade theory. A two-city country trades with the larger world, and workers within the country are mobile between the two cities. Productivity in a given function varies across cities, giving rise to urban comparative advantage. This may be due to exogenous technological differences (Ricardian) or to city- and function-specific scale economies. Sectors differ in the intensity with which they use different functions, giving rise to a pattern of sectoral and functional specialisation. We generate a number of economic insights, including that, as costs of remote sourcing fall, cities' functional specialization tends to increase and their sectoral specialization falls. We examine the model's predictions empirically over a 20-30-year period for US states. In line with the predictions of the model, we find that functional concentration rises and sectoral concentration falls over this time span. Similarly, we find that regional specialization in functions rises and regional specialization in sectors falls over the period.