According to the theory of industrial districts, a new wave of economic growth is being led in a number of regions in Europe, North America and East Asia by spatially concentrated networks of mostly small and medium sized enterprises, often using flexible production technology and characterized by extensive local interfirm linkages. Does this amount to a re-emergence of the dominance of what urban and regional economists call 'agglomeration economies' over the well-known pressures on business to spatially disperse its operations? Neoclassical economic theorizing from Marshall to Perroux provides one perspective on the contemporary industrial district phenomenon. Another is afforded by Granovetter's more recent elaboration of the ideas of 'embedding', 'under-' and 'over-socialization'. Confronting each of these theoretical approaches with the other leads me to conclude that the industrial district prototypes involve more than simply a reassertion of agglomeration economies. Nor can the industrial districts be satisfactorily explained by Williamsonian concepts about the internalization (within the firm or within the region) of transaction costs. While emphasizing the mutual/shared benefits to individual firms/plants/production units of co-location (such as access to a larger and more specialized local labour pool, and the realization of scale economies in infrastructure provision), both agglomeration theory and transaction cost economics nevertheless follow standard neoclassical logic in conceptualizing local economies as collections of atomistic competitors, formally aware of one another solely through the intermediation of price/cost signals, embodied in contracts of varying completeness. By contrast, contemporary industrial district theory emphasizes the contextual significance of communal non-economic institutions and the importance of relations of 'trust' in reproducing sustained collaboration among economic actors within the districts. Whatever their empirical importance to economic growth and development, their ethical/distributive implications, and to whatever extent they prove over time to be stable institutions, industrial districts are definitely more than 'old wine in new bottles'.
