Abstract
Net entry rates of manufacturing establishments across Greek regions are examined here by means of both an appropriately modified conventional shift-share method and a linear-model variant. All results accord with earlier findings for UK regions, signifying that the main source of variation in spatial net entry patterns stems from within-industry variation in locational preferences across space. However, analysis by establishment size indicates that the predominance of within-industry variation, expressed by the magnitude and the significance of the competition effect, is much less marked for larger establishments. The determinants of spatial variations in net entry rates across Greek regions and over manufacturing sectors emphasise local economic conditions, especially for smaller firms. In methodological terms, the one-way ANOVA-based variant of shift-share analysis is criticised on the grounds that the alleged numerical equivalence between the estimated industry-mix component and that calculated conventionally is an artefact of little value. It is demonstrated that this numerical equivalence is sustained even when fundamental estimation assumptions are violated and despite the use of only partial information.
Get full access to this article
View all access options for this article.