Do industrial districts influence export performance and export intensity? Evidence for Spanish SMEs' internationalization process

Abstract

This paper seeks to enhance the understanding of the role of the industrial districts in the internationalization process of small and medium-sized enterprises (SMEs). The study focuses essentially upon the following issue: Can the location inside industrial districts influence the export performance and export intensity of the Spanish SMEs? To address this question, this study draws upon a sample of 285 manufacturing firms located in the Valencian community (a Spanish region) surveyed during the period January 2000 to March 2000. It is shown that industrial district location, marketing differentiation, institutional networks, clients' networks, competitors' networks and global orientation of sector and company have a clear influence on firms export performance and export intensity. The results are basically consistent with the limited previous research.

Acknowledgements

The author wishes to acknowledge the solicitude of the anonymous referees who read several drafts of the paper and provided helpful comments.

Notes

1. The difficulties in “operationalizing” Becattini's (1992) socio-territorial entity have lead most of the empirical research to a quantitative identification of the industrial district, basing on the presence of small and medium-sized firms and the density of specialized firm in certain activities (Sforzi, 1992).

2. Most of the empirical studies have adopted a quantitative criterion for identifying the district depending on how various factors come together to demonstrate a high density of small and medium firms specializing in certain activities. The study developed by Costa (1988) began with the creation of an industrial census. This census is complemented using information provided by a wide group of enterprises regarding degree of innovation, firm's presence in the international markets, integration within the social context, industrial capacity and decentralization of the production process. Ybarra (1991) applied a method whose starting point was the concept of industrial district. The author identified the industrial districts using three variables: the territory with all its resources and peculiarities, the main economic activity including all complementary industrial activities, and the existence of a productive system in which SMEs predominate. This classification, based on a quantitative criterion, has been also used by Camisón (2003), Molina and Martínez (2003), Tomás and Such (1997) or Soler (2000). Recently, Giner and Santa María (2002) determined the existence of industrial districts in the Valencian community using the methodology specified by the Italian Ministry of Industry. First, the authors identified the different labour systems in the region. These labour systems were identified through analysis of the spatial configuration of daily movements between place of residence and the workplace.

3. ARDAN is an informational service supported and sponsored by the main industrial policy agencies of the different Spanish regional governments. In the case of the Valencian region this database is one of the services offered by Valencian Institute of Medium and Small Enterprises (IMPIVA), a public industrial policy agent under the authority of the regional government. The directory provides financial information, based on official sources, of manufacturing firms with annual revenue of more than 240,000 euros.

4. DIRNOVA is another database provided by Valencian Insitute of Medium and Small Enterprises (IMPIVA). The directory offers information about international strategy of 2055 companies located in the Valencian community.

5. Previous literature about economic structure of Alicante and international firms located in this area reflect a similar profile (Pedreño, 1993; Martínez, 1997; ITEM, 1999).

6. White's test results for the export intensity regression model: χ² = 94.3260 with probability = 0.075718; White's test results for the export intensity regression model: χ² = 88.65141 with probability = 0.152074.