Analysts and policy makers frequently measure industrial localization and regional specialization. However, they rarely examine the nation's full array of industries or regions. So local indices, appropriate for specific industries or selected regions, are typically estimated. But in some instances global indices would be preferable in order to assess the wider features of the entire space-economy. This article constructs global indices from the local indices already used in assessing localization and specialization. Global localization and global specialization are shown to be identical when all local indices use the dissimilarity logic. Two-digit standard industry codes manufacturing data, taken from the U.S. during 1958-1995, are used to illustrate the results. The values of these global coefficients, like their local constituents, are shown to vary with geographic scale. The discussion addresses spatial distributions (evenness) but not geographic arrangements (clustering). Copyright 2005 Blackwell Publishing Ltd..
