The aim of this paper is to analyse co-location patterns of manufactures and service industries at a microgeographic level using Spanish data from the Mercantile Register. Our approach allows us to analyse joint-location and co-location patterns of firms in different industries, and to overcome previous technical constraints in this type of analyses, partially thanks to using homogeneous cells instead of administrative units. This paper contributes to the empirical literature on industry location by developing a multisectorial co-location index computed by comparing differences between observed data about firms’ location and randomly generated data. Multisectorial relationships are analyzed by transposing bilateral relations onto an n-dimensional space. Our results show that dispersed industries tend to locate jointly and that industries with lower joint-location patterns have spatial structures similar to those obtained through input–output relationships, suggesting weak role of co-location patterns as interindustry linkages are not the main location determinants.
This is a preview of subscription content, access via your institution.
Buy single article
Instant access to the full article PDF.
Tax calculation will be finalised during checkout.
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
Tax calculation will be finalised during checkout.
Nevertheless, apart from these methodologies there is as well an extensive list of alternative approaches (see for instance Van Egeraat et al. 2016, for a review). The reason for the distinction between Clusters and Industrial Districts was to use widely known concepts (Industrial Districts and Clusters) in order to group contributions into those that consider some subjective determinants as trust among economic agents (Industrial Districts), and those that only consider some objective characteristics (Clusters).
There are also other approaches such as those that use the stochastic Point Pattern methodology and those that use Neuronal Networks for pattern recognition. However, these approaches are not able to perform the multisectorial analyses that are the goal of this paper.
In this paper we carry out an introductory analysis from a national (aggregated) point of view, and although an alternative urban approach is also reasonable, we have left urban issues for further contributions.
In this paper, “spatial proximity” means to be located in the same cell.
Although Duranton and Overman (2008) use the terms joint-localization and co-location, for the sake of simplicity, we will refer to them simply as joint-location and co-location.
It is worth noting that the data set refers to firms (not establishments) and that each firm could have more than one establishment. Nevertheless, not taking into account multi-plant firms is not a major problem, as only around 1% of firms in Spain are multi-plant (see Jofre-Monseny et al. 2015, for details).
Other alternative statistical sources such as the Censo de Locales (INE) are not currently updated, although having firms as observation units instead of establishments also provides useful information since it highlights the role of municipalities when firms are choosing where to locate their headquarters.
There are alternative datasets such as DIRCE (INE) but the geographical location of the firms is highly spatially aggregated.
The Spanish Input Output matrix is from 2000 and covers all economic industries at 2 digits of NACE classification (Spanish National Statistics Institute).
We tested several alternative cell sizes (5 km * 5 km, 20 km * 20 km, 25 km * 25 km, and 50 km * 50 km) but the results were essentially unchanged.
The cells’ size is obviously related to the total area analysed. A similar approach is used by Larsson and Öner (2014) in their analysis of Swedish retail markets in metropolitan areas. Nevertheless, as they focus on smaller areas the cells also become smaller, at 250 m * 250 m.
By way of an example, municipality sizes in Spain range from 5 to 1000 km2.
This is a (simple) starting point that could be easily improved by taking into account the intensity of land use, considering certain indicators such as the number of jobs, the production value and levels of sales, among others. This should allow the expected results to be compared with real results to determine the number of jobs, for example. However, there are also some (potential) limitations regarding data accuracy.
Nevertheless, the main problems concern the heterogeneity of firm size, so it seems that a better solution would be to use the size of firms (e.g. the number of employees) rather than just the number (or the existence) of firms.
This latter requirement implies that firms localise randomly inside “occupied” cells (i.e. areas where real firms are located) as suggested by Duranton and Overman (2008). This approach means that firms are expected to be located only in those places that are available for economic activity (as the real data shows). Unfortunately, a major shortcoming of this approach is that it assumes that firms could be located elsewhere with other firms, regardless of the industry they are involved in, which is not as realistic (especially at a 2/3-digit level). An extension of this study (and a possible solution to this shortcoming) would be to regard manufacturing, services and agricultural firms as being located with other firms from the fields of manufacturing, services and agriculture, respectively.
The idea about “equivalence” was suggested by a reviewer of this paper.
These regular cells have an area of 100 km2 (10 km * 10 km).
As an example, indices of high-tech industries such as Office machinery, computers and medical equipment, Precision and optical instruments (0.644) and Electrical machinery and apparatus (0.664) are clearly lower than those of some low-tech industries such as Food, beverages and tobacco (1.452) and Agriculture and fishing (1.424).
Specifically, Textiles, leather clothes and shoes, Paper and publishing, Rubber and plastic products, Machinery and equipment, Business services and Health and veterinary activities.
If an industry is joint-located with the same industry the value of the index is 1. In terms of the novelty of this proposal, it is similar to Correspondence Analysis, but the approach proposed by this paper is more appropriate because there are no information loses linked to dimensionality reduction. Among other advantages, we may highlight that there are no random biases in measures derived from criteria used for histogram construction given that spatial position is unique.
One reviewer suggested that our method could be similar to that of correspondence analysis, but we consider that though there are some similarities our proposal is more appropriate because there are no information loses linked to dimensionality reduction. Among other advantages of our proposal, we may highlight that there are no random biases in measures derived from criteria used for histogram construction given that spatial position is unique. Consequently, although if this is not a correspondence analysis in that case we would prefer Hellinger distance rather than Chi2 (Rao 1995).
Using a geometric perspective, they are close to the bisectriz.
Acosta M, Coronado D, Flores E (2011) University spillovers and new business location in high-technology sectors: Spanish evidence. Small Bus Econ 36:365–376
Albert JM, Casanova MR, Orts V (2012) Spatial location patterns of Spanish manufacturing firms. Pap Reg Sci 91(1):107–136
Arauzo-Carod JM (2005) Determinants of industrial location: an application for Catalan municipalities. Pap Reg Sci 84:105–120
Arauzo-Carod JM (2008) Industrial location at a local level: comments on the territorial level of the analysis. Tijdschr Econ Soc 99:193–208
Arauzo-Carod JM (2013) Location determinants of new firms: does skill level of human capital really matter? Growth Change 44(1):118–148
Arauzo-Carod JM, Manjón-Antolín M (2004) Firm size and geographical aggregation: an empirical appraisal in industrial location. Small Bus Econ 22:299–312
Arbia G (2001) Modeling the geography of economic activities on a continuous space. Pap Reg Sci 80:411–424
Barlet M, Briant A, Crusson L (2009) Location patterns of services in France: a distance-based approach. Working paper, Paris School of Economics
Bickenbach F, Bode E, Krieger-Boden C (2013) Closing the gap between absolute and relative measures of localization, concentration or specialization. Pap Reg Sci 92(3):465–479
Blöch F, Theis FJ, Vega-Redondo F, Fisher EO (2011) Vertex centralities in input-output networks reveal the structure of modern economies. Phys Rev E 83:1–9
Boix R, Galletto V (2008) Marshallian industrial districts in Spain. Sci Reg Ital J Reg Sci 7(3):29–52
Brenner T (2004) Local industrial clusters: existence, emergence and evolution. Routledge, London
Brenner T (2006) Identification of local industrial clusters in Germany. Reg Stud 40(9):991–1004
Devereaux M, Griffith R, Simpson H (2004) The geographic distribution of production activity in the UK. Reg Sci Urban Econ 34(5):533–564
Duranton G, Overman HG (2005) Testing for localization using microgeographic data. Rev Econ Stud 72:1077–1106
Duranton G, Overman HG (2008) Exploring the detailed location patterns of U.K. manufacturing Industries using microgeographic data. J Reg Sci 48(1):213–243
Ellison G, Glaeser EL (1997) Geographic concentration in US manufacturing industries: a dartboard approach. J Polit Econ 195:889–927
Graham RL, Hell P (1985) On the history of the minimum spanning tree problem. Ann Hist Comput 7(1):43–57
Hoffmann J, Hirsch S, Simons J (2017) Identification of spatial agglomerations in the German food processing industry. Pap Reg Sci 96(1):139–162
Isard W (1960) Methods of regional analysis. MIT Press, Cambridge
ISTAT (1996) Rapporto annuale. La situazione del paese nel 1995. ISTAT, Rome
Jofre-Monseny J, Sanchez-Vidal M, Viladecans-Marsal E (2015) Big plant closures and agglomeration economies. SERC discussion paper no. SERCDP0179
Larsson JP, Öner Ö (2014) Location and co-location in retail: a probabilistic approach using geo-coded data for metropolitan retail markets. Ann Reg Sci 52:385–408
López-Bazo E (2006) Definición de la metodología de detección e identificación de clusters industriales en España. Dirección General de la Pequeña y Mediana Empresa (DGPYME), Madrid
Marshall A (1890) Principles of economics. MacMillan, New York
Maurel F, Sédillot B (1999) A measure of the geographic concentration in French manufacturing industries. Reg Sci Urban Econ 29:575–604
Mulligan GF, Schmidt C (2005) A note on localization and specialization. Growth Change 36(4):565–576
Olsen J (2002) On the units of geographical economics. Geoforum 33:153–164
Openshaw S, Taylor PJ (1979) A million or so correlation coefficients: three experiments on the modifiable areal unit problem. In: Wrigley N (ed) Statistical applications in the spatial sciences. Pion, London, pp 127–144
Paluzie E, Pons J, Tirado D (2004) The geographical concentration of industry across Spanish regions, 1856–1995. Rev Reg Res 24(2):143–160
Porter M (1998) Clusters and the new economics of competition. Harvard Bus Rev 76(6):77–90
Rao CR (1995) A review of canonical coordinates and an alternative to correspondence analysis using Hellinger distance. Qüestiió 19(1–3):23–63
Sforzi F, Lorenzini F (2002) I distretti industrali. In: IPI-Ministero delle Attività Produttive, L’esperienza italiana dei distretti industriali, Rome, pp 20–33
Silverman B (1986) Density estimation for statistics and data analysis. Chapman & Hall, London
Titze M, Brachert M, Kubis A (2011) The identification of regional industrial clusters using qualitative input–output analysis (QIOA). Reg Stud 45:89–102
Van Egeraat C, Morgenroth E, Kroes R, Curran D, Gleeson J (2018) A measure for identifying substantial geographic concentrations. Pap Reg Sci 97(2):281–300
Viladecans E (2004) Agglomeration economies and industrial location: city-level evidence. J Econ Geogr 4(5):565–582
Ward JH Jr (1963) Hierarchical grouping to optimize an objective function. J Am Stat Assoc 58:236–244
Wrigley N (1995) Revisiting the modifiable areal unit problem and the ecological fallacy. In: Cliff AD, Gould PR, Hoare AG, Thrift NJ (eds) Diffusing geography. Wiley, Oxford, pp 49–71
We are grateful to the seminar participants at the 12th EUNIP Conference (Universitat Rovira i Virgili) and at the International Meeting on Regional Science (Universidad de Badajoz), and especially to Anxo Sánchez for his valuable comments. Any errors are, of course, our own.
This paper was partially funded by ECO2014-55553-P, ECO2013-42310-R, the “Xarxa de Referència d’R+D+I en Economia i Polítiques Públiques”, the SGR programme (2014 SGR 299) of the Government of Catalonia and the R&D Spanish project ‘CITiTALENT’, CSO2016-74888-C4-4-R (AEI/FEDER, UE).
See Table 6.
About this article
Cite this article
Pablo-Martí, F., Arauzo-Carod, JM. Spatial distribution of economic activities: a network approach. J Econ Interact Coord 15, 441–470 (2020). https://doi.org/10.1007/s11403-018-0225-8
- Microgeographic data
- Network analysis
- Firm location