Concentration, agglomeration and the size of plants

Abstract

This paper investigates whether the geographic distribution of manufacturing activities in Italy is likely to differ according to the scale of plants. We find strong evidence of a significant positive relationship between size and concentration, as in Kim (1995) or Stevens (2002, 2004). However, we go one step further in examining how sensitive is this feature to the consideration of spatial dependence between geographic units. We show that, while large plants exhibit a clear tendency to cluster within narrow geographical units such as local labor systems, small establishments, by contrast, rather co-locate within wider areas in which a distance-based pattern emerges. These findings are consistent with plants of heterogeneous size engaging in different transport intensive activities.JEL classification: C21, L11, R12, R30, R34.

Spatial Localization in Manufacturing: A Cross-Country Analysis

“…Beside these theoretical contributions, a good deal of empirical research has investigated localization in manufacturing industries (see among others, Ellison and Glaeser, 1997;Maurel and Sédillot, 1999;Barrios et al, 2005;Lafourcade and Mion, 2007;Duranton and Overman, 2005). All these works, mainly focusing on single countries, confirm the expectation that firms are generally clustered in space.…”

“…Previous attempts in the same direction can be found in Barrios et al (2005) and Lafourcade and Mion (2007). However, differently from Barrios et al (2005Barrios et al ( , 2008, we consider a larger number of (size-heterogenous) countries.…”

“…2 In addition to US-focused research (Ellison and Glaeser, 1997;Rosenthal and Strange, 2001; Holmes and Stevens, 2002a;Kim, 1995), existing contributions have been studying industry localization patterns in UK (Devereux et al, 2004;Duranton and Overman, 2005); Belgium (Bertinelli and Decrop, 2005); France (Maurel and Sédillot, 1999); Italy (Lafourcade and Mion, 2007); Germany (Brenner, 2006); Ireland and Portugal (Barrios et al, 2005). See Combes and Overman (2004a) for a review.…”

“…the degree of spatial correlation in firms' location decisions (cf. Arbia, 2001; Lafourcade and Mion, 2007;Duranton and Overman, 2005). Two indices that account for the spatial features of industrial data are those proposed by Moran (1950) and by Duranton and Overman (2005).…”

“…Finally, clusters of firms located at the borders of neighboring regions and/or spanning over the area covered by a single region are treated in the same way as clusters in two very distant regions. More precisely, as pointed out in Arbia (2001) and Lafourcade and Mion (2007), indices requiring the division of the space into smaller units are only able to capture "spatial concentration" of industrial activity into some areas. They are not able to measure instead "true agglomeration", that is the degree of spatial correlation between firms' location choices.…”