Volume 30.4 December 2006 894–917 International Journal of Urban and Regional Research DOI:10.1111/j.1468-2427.2006.00695.x © 2006 The Authors. Journal Compilation © 2006 Joint Editors and Blackwell Publishing Ltd. Published by Blackwell Publishing. 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA Cities as the Industrial Districts of Housebuilding MICHAEL BUZZELLI and RICHARD HARRIS Abstract In North America the housebuilding industry is ubiquitous and locally autonomous. In Ontario during the 1990s, 81% of urban single-family homes were erected by locally based builders, a proportion that varied with urban isolation. Urban areas may be regarded as the industrial districts of home builders: numerous small, specialized firms interact frequently within a rich, embedded market network; subcontracting is the norm; networks and firm boundaries are fluid. The theory of industrial districts offers a useful vocabulary for analysing the neglected building industry. Analytically, the building industry offers unequalled opportunities to explore the dynamics of industrial districts, and how economic globalization meets local limits. Examples of regional conglomerations are the specialized industrial districts of northern and central Italy, the New York City garment district, and the construction industry in any number of U.S. cities (Piore and Sabel, 1984: 265). The geography of housebuilding is unlike that of all other important forms of economic activity: it combines substantial features of manufacturing with many of the locational aspects of services. Some building — about a fifth of all new units in the United States, and less in Canada, Australia, Britain and Western Europe — occurs in factories, in much the same way that automobiles and many other goods are produced. 1 Most, however, still takes place on site. This means that, like personal services, it is ubiquitous, occurring wherever there are buyers. The assembly of dwellings on site can take advantage of the efficient, factory fabrication of materials and subassemblies, but dispersed sites limit the builder’s capacity to realize economies of scale. As a result, for many decades observers of the industry have debated how such limits might be overcome. At the same time, they have puzzled over how it is that, faced with factory competition from larger and apparently more efficient producers, the traditional builder has persisted. This article proposes a new way to answer to that question. We argue that on-site housebuilding is efficient because it operates in much the same way as those manufacturing industries that cluster in manufacturing districts. The classic manufacturing district consists of a fluid network of small, interdependent firms in a particular industry that operate within a limited area. This, we suggest, describes the housebuilding industry in any North American city: most builders are small, relying on 1 Manufactured housing accounts for about a quarter of all single family units produced in the United States each year (Wallis, 1998: 347). We would like to thank the Social Science and Humanities Research Council of Canada for its financial support, and Henry Friesen of the Ontario New Home Warranty Program for his cooperation. Robert Lewis, Donna Rilling, and two anonymous referees offered valuable comments on an earlier draft.