Portal de Periódicos da CAPES

Innovation and Industrial Districts: A First Approach to the Measurement and Determinants of the I-District Effect

2008; Routledge; Volume: 43; Issue: 9 Linguagem: Inglês

10.1080/00343400801932342

1360-0591

Rafael Boix, Vittorio Galletto,

ICT Impact and Policies

Abstract Boix R. and Galletto V. Innovation and industrial districts: a first approach to the measurement and determinants of the I-district effect, Regional Studies. The paper analyses an exhaustive database of patents granted in Spain between 2001 and 2006 aggregated in a panel of 806 local labour markets classified by seven typologies of local production systems. The analysis shows that Marshallian industrial districts generate 30% of Spanish patents and an innovative output per capita that is 47% above the national average and 31% larger than the manufacturing production systems of large firms. The econometric estimates of a fixed-effects model confirm the existence of an innovation-district (I-district) effect and its size. The I-district effect is mainly related to the presence of Marshallian localization economies. Boix R. et Galletto V. Innovation et districts industriels: une première approche de la mesure et des déterminants des effets des districts d'innovation, Regional Studies. Nous analysons une base de données exhaustive de brevets délivrés en Espagne entre 2001 et 2006 dans un échantillon de 806 marchés locaux de l'emploi classés en sept typologies de systèmes locaux de production. Notre analyse montre que les districts industriels de type Marshall génèrent 30% des brevets espagnols et affichent une performance innovatrice par tête supérieure de 47% à la moyenne nationale et supérieure de 31% aux systèmes de production manufacturière des grandes entreprises. Les estimations économétriques d'un modèle d'effet fixe confirment l'existence d'un effet 'district d'innovation' et de son importance. L'effet 'district d'innovation' est lié, pour l'essentiel, à la présence d'économie de localisation de type Marshall. Districts industriels Innovation Économies extérieures Effet de district Boix R. und Galletto V. Innovation und Industriebezirke: ein erster Ansatz für die Messung und die Determinanten des I-Distrikt-Effekts, Regional Studies. Wir analysieren eine umfangreiche Datenbank mit zwischen 2001 und 2006 in Spanien erteilten Patenten, die in einem Panel von 806 lokalen Arbeitsmärkten zusammengefasst und nach sieben Typologien lokaler Produktionssysteme klassifiziert werden. Unsere Analyse zeigt, dass Marshallsche Industriedistrikte 30% der spanischen Patente und eine innovative Pro-Kopf-Leistung hervorbringen, die um 47% über dem Landesdurchschnitt liegt und 31% höher ausfällt aus die Produktionssysteme von Großfirmen. Die ökonometrischen Schätzungen eines Festeffekt-Modells bestätigen die Existenz eines Innovationsdistriktseffekts (I-Distrikt) sowie dessen Größe. Der I-Distrikt-Effekt bezieht sich in erster Linie auf die Präsenz von Marshallschen Lokalisationswirtschaften. Industriebezirke Innovation Externe Ökonomien Distrikteffekt Boix R. y Galletto V. Innovación y distritos industriales: una primera aproximación a la medición y determinantes del efecto I-distrito, Regional Studies. En la investigación se analiza una exhaustiva base de datos de patentes que entre 2001 y 2006 solicitaron protección en España, agregadas en un panel de 806 mercados locales de trabajo clasificados en siete tipologías de sistemas productivos locales. El análisis muestra que los distritos industriales marshallianos generan el 30% de las patentes españolas, así como un output innovador per capita 47% mayor que la media nacional y 31% mayor que los sistemas productivos manufactureros de gran empresa. Las estimaciones econométricas de un modelo de efectos fijos confirman la existencia de un efecto-distrito en innovación (efecto I-distrito) y su dimensión. El efecto I-distrito se asocia principalmente a la presencia de economías de localización marshallianas. Distritos industriales Innovación Economías externas Efecto distrito Keywords: Industrial districtsInnovationExternal economiesDistrict effectJEL classifications: O14O31R12 Acknowledgements The authors wish to thank Giacomo Becattini and two anonymous referees for helpful comments. They also thank the Spanish Patent and Trade Mark Office (Ministry of Industry, Tourism and Trade), the Centre for the Development of Industrial Technology (CDTI), and the Spanish Institute of Statistics (INE) for providing most of the data used in this research. Notes For a comparative review of the literature on territorial innovation models including industrial districts, clusters, milieux innovateurs, new industrial spaces, etc., see Moulaert and Sekia Citation(2003). The use of patents as indicators of innovation can be influenced by the industrial specialization of the LPS and firm size distributions. Griliches (Citation1990, Citation1992) and Khan and Dernis Citation(2006) provide further discussion on their advantages and limitations. The complete patent database includes 70 000 documents from 1991 to 2006. Patent counts include 'utility models', a figure granted by the OEPM which is similar to the patent, although legal requirements are less strict and protection covers only ten years. Similar figures exist in Austria, Denmark, Finland, Germany, Greece, Italy, Japan, Poland and Portugal. Employment data come from the 2001 Census of the Spanish Institute of Statistics (INE). Data treatment follows international standards: patents are located according to the first applicant with an address in Spain (the inventor's address is not available for national patents); reference date is the oldest application data in any register because it is the closest to the invention date and does not introduce biases due to legal or procedural delays. These include those LPS with the characteristics of an industrial district that Boix and Galletto Citation(2008) excluded because the number of employees in the main specialization was fewer than 250 employees (considered too small), and also some LPS where manufacturing as a whole is of the average size of a large firm but without any large firm in the main specialization. This framework facilitates to compare and discuss the results. The choice of the dependent variable (patents) was also related to comparability. This follows the line of other research that has used relative indicators in the measurement of the district effect, e.g. productivity (Signorini, Citation1994) or efficiency (Hernández and Soler, Citation2003). R&D and employment data are taken from the INE. It is also possible to use hierarchical multilevel models to avoid the assignation, although the hypothesis introduced for the data generates other restrictions. Results are controlled by using additional data on innovation grants and loans provided by the Ministry of Industry (CDTI and PROFIT databases). R&D per employee in the initial year is conceived as exogenous in the model. Additional test of exogeneity proves that this variable is empirically exogenous in all the regressions. Additional controls of the functional form of the model and the relation between the dependent and explanatory variables were introduced. The log–linear specification without non-linearities proved to be the most suitable specification. In fact, in industrial districts the average firm size is larger than in most of the other non-manufacturing systems. In Spain, more dynamic environments such as industrial districts provide numerous job opportunities for young people so that the necessity of higher levels of education to get work is not perceived. This result should not be interpreted as a direct indicator of the impact of contextual knowledge on innovation, although it suggests the importance of contextual knowledge mechanisms (learning by doing, on-the-job training, etc.) to make up for the lower levels of standard-educated people. When the data are pooled, the spatial lag (ρ = 0.14) is statistically significant, although it does not improve the fit. When fixed-effects and external economies are included, the lag decreases to ρ = 0.08 and again the most parsimonious model is preferred. This weak evidence and Fig. 2 suggest that the impacts of inter-LPS spillovers could be locally important in the East and North East of Spain where industrial districts and large firm LPS are concentrated.