Please use this identifier to cite or link to this item:
Beneyto, Maria Jesus Santa Maria
Perez, Jose Miguel Giner
Olivares, Antonio Fuster
Year of Publication: 
Series/Report no.: 
45th Congress of the European Regional Science Association: "Land Use and Water Management in a Sustainable Network Society", 23-27 August 2005, Amsterdam, The Netherlands
A range of quantitative techniques have been employed by researchers in economic geography and other social science disciplines to measure and, spatially, define agglomerations of industrial activity. However, the application of these techniques in the literature results in a low consistency level. Because of this, new quantitative techniques have introduced solutions to solve the problems founded in the location's analysis. One of these problems is the discrimination between geographic concentration arising from individual plants locating near to each other and that due to the concentration in an industrial structure. A relevant limitation of traditional location indexes is the absence of data about the differences in the size distribution of firms between geographic units. Recent papers by Ellison and Glaeser (1997) and Maurel and S├ędillot (1999) have proposed indexes designed to measure agglomerations or geographic concentrations in excess of that which would be expected given industrial concentrations. These measures are all based on the distribution of activity over discrete geographic units. Another problem is the use of arbitrary cut-off values for determining what level of industrial specialization defines an agglomeration. O'Donoghue and Gleave (2004) have proposed a new measure, the 'standardized location quotient (SLQ)', which recognizes agglomerations as being comprised of locations with statistically significant location quotient values for the industry/activity under analysis. Other questions that appear when constructing these measures are the specification of the regional division's level and the suitable use of administrative territorial units. New quantitative techniques of spatial econometrics solve this question. The use of a spatial autocorrelation indexes will allow us to know if the location of a concrete economic activity in a municipality is influenced by the location of the same activity in other neighbouring municipalities. We use global spatial autocorrelation statistics as I Moran Index (Moran, 1948) and Local Measures of Spatial Autocorrelation (LISA). The cluster map (LISA map) shows the significant locations by type of association. With LISA map, we measure geographic concentration of employment in industry clusters by detecting spatial association patterns in administrative areas (in this case, municipalities). In the empirical analysis the municipality, the micro level of administrative regions (NUTS5) in Spain, will be used as territorial unit. The data will be provided by the Industrial Register (Ministry of Industry, 2000) that contains information about the population of production plants in Spain at two and/or three-digit industry level. This includes the location of the plant (given by municipality), the plant's three-digit industrial classification and the number of employees. So, the objective of this work will be to identify spatial agglomerations within the Spanish industrial sectors using all these new contributions to the spatial analysis and, as a secondary objective, to compare the difference of the results obtained with each quantitative technique. The results will offer a wide view of the geographic concentration and agglomeration of industrial activity in Spain.
Document Type: 
Conference Paper

Files in This Item:

Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.