A geographic concentration of interconnected companies, specialized suppliers, service providers, associated institutions, and firms in related industries (Porter (1998) Harvard Bus. Rev. 76, 6). Ketels in (2004) Innovative City and Business Regions, 3 identifies the critical characteristics of clusters as: proximity—firms need to be sufficiently close in space to allow spillovers and the sharing of common resources; linkages; some level of active interaction; and critical mass—there need to be enough participants for the interactions to have a meaningful impact. B. Asheim (2005) adds localization economies and specialization to this list, and Storper and Venables (2004) J. Econ. Geog. 4, 4 stress the importance of face-to-face contact. Florida (2003) City & Community 2, 1 sees ‘concentrations of talented people who power innovation and economic growth’ as vital. Glaeser, in G. Clark (2003) argues that places with higher levels of human capital are more innovative and grow more rapidly and robustly over time; see also Cushing (2001) W. Paper, U. Texas at Austin. Giuliani (2007) J. Econ. Geog. 7, 2 stresses the importance of embeddedness in local business networks. Bathelt (2005) Reg. Studs 39 argues that clusters can only create new knowledge if the cluster firms have linkages with external markets and employ a mix of local and non-local transactions. Clustering and globalization are closely intertwined; see M. Fujita et al. (2001)

Maskell and Malmberg (2007) J. Econ. Geog. 7, 5 conceive clusters in evolutionary terms: the growing cluster is made up of firms with similar competencies, reducing expenditure on locational search costs, and enjoying agglomeration economies. Further growth results from cumulative causation, the presence of support institutions, and the power of the name of the locality as an attraction to a worker: the benefits of being physically located within, for example, Silicon Valley are perceived to be worth the costs of relocation. See Delgado, Porter, and Stern (2012) NBER Working paper series, #18250 on clusters, convergence, and economic performance. Sadler (2004) Reg. Studs 38 argues that an effective regional cluster policy needs to identify clusters on a number of variables, including knowledge flows. The use of the cluster model as a paradigm of economic development has been criticized. Bathelt and Taylor (2002) Geografiska B 84, 2 argue that clusters ‘depend on specific circumstances in “time–space” and, because of their very transience and specificity, those conditions might be very difficult if not impossible to create through the blunt instruments of policy’; indeed, Huber (2012) J. Econ. Geog. 12, 1, 107 casts doubt on the benefits of clustering.