Marshall's versus Jacobs' Externalities in Firm Innovation Performance: The Case of French Industry

Abstract

Galliano D., Magrini M.-B. and Triboulet P. Marshall's versus Jacobs’ externalities in firm innovation performance: the case of French industry, Regional Studies. This paper analyses the influence of different types of spatial externalities related to the location of firms on their innovation performance and how those externalities combine in the territories with regard to the Marshall–Jacobs dichotomy. The originality of this study also lies in the consideration of a larger definition of the firm, one that takes into account the location of all its units. Based on a dataset of French industrial firms and specific indicators to evaluate the specialization and/or diversification of the employment zones, the impacts of the spatial profile of the firm on its innovation performance are tested.

Galliano D., Magrini M.-B. and Triboulet P. 企业创新表现中的马歇尔外部性对雅各布斯外部性：法国工业的案例，区域研究。本文分析与企业区位有关的不同类型的空间外部性，对企业的创新表现产生的影响，以及这些外部性在马歇尔—雅各布斯的二元对立中，如何在领域中相互结合。本文的创新性，亦在于将企业的所有单位之区位皆纳入考量，以更广泛地考量企业的定义。本研究根据一个法国工业企业的数据集和评估就业区的专殊化以及／或多样化特定指标，测试企业的空间形态对创新表现的影响。

Galliano D., Magrini M.-B. et Triboulet P. Les externalités Marshall versus Jacobs dans la performance innovatrice de la firme: le cas de l'industrie française, Regional Studies. Ce papier analyse l'influence des externalités spatiales en lien avec la localisation de la firme sur sa performance à l'innovation et comment ces externalités spatiales se combinent dans les territoires au regard de la dichotomie Marshall–Jacobs. L'originalité de notre étude réside également dans la prise en compte d'une définition élargie de la firme qui intègre la localisation de tous ses établissements. Basé sur un ensemble de données d'entreprises industrielles françaises et deux indicateurs pour évaluer la spécialisation et/ou la diversification des zones d'emploi en France, nous testons l'impact du profil spatial de la firme sur sa performance innovatrice.

Galliano D., Magrini M.-B. und Triboulet P. Marshall- und Jacobs-Externalitäten bei der Innovationsleistung von Firmen: der Fall der französischen Industrie, Regional Studies. In diesem Beitrag wird der Einfluss verschiedener Arten von räumlichen Externalitäten im Zusammenhang mit dem Standort von Firmen auf ihre Innovationsleistung analysiert, und es wird untersucht, wie sich diese Externalitäten in den Gebieten hinsichtlich der Marshall-Jacobs-Dichotomie miteinander kombinieren. Die Originalität dieser Studie liegt auch in der Verwendung einer breiteren Definition des Begriffs der Firma begründet, bei der der Standort von allen ihren Niederlassungen berücksichtigt wird. Ausgehend von einem Datensatz französischer Industriefirmen und von spezifischen Indikatoren zur Bewertung der Spezialisierung und/oder Diversifizierung der Beschäftigungszonen werden die Auswirkungen des räumlichen Profils der Firma auf ihre Innovationsleistung überprüft.

Galliano D., Magrini M.-B. y Triboulet P. Externalidades de Marshall y Jacobs en el desempeño de innovación de las empresas: el caso de la industria francesa, Regional Studies. En este artículo analizamos la influencia de los diferentes tipos de externalidades espaciales con relación a la ubicación de las empresas en cuanto a su desempeño de innovación y cómo se combinan estas externalidades en los territorios con respecto a la dicotomía Marshall–Jacobs. La originalidad de nuestro estudio radica en considerar una definición más amplia del concepto de empresa que tiene en cuenta la ubicación de todas sus filiales. A partir de un grupo de datos de empresas industriales francesas y de indicadores específicos para evaluar la especialización y/o diversificación de las zonas de empleo, comprobamos los efectos del perfil espacial de la empresa con respecto a su desempeño de innovación.

• Innovation
• Spatial externalities
• Agglomeration economies
• Multiunit firms
• French industry
• Location quotient

• 创新
• 空间外部性
• 聚集经济
• 多单位企业
• 法国工业
• 区位商数

• Innovation
• Externalités spatiales
• Économies d'agglomération
• Firmes multi-établissements
• Industrie française
• Quotient de localisation

• Innovation
• Räumliche Externalitäten
• Agglomerationsökonomien
• Firmen mit mehreren Niederlassungen
• Französische Industrie
• Standortquotient

• Innovación
• Externalidades espaciales
• Economías de aglomeración
• Empresas con varias filiales
• Industria francesa
• Cociente de localización

JEL classifications:

• L2
• L6
• O30
• R3

Acknowledgments

This work was carried out as part of the LABEX SMS [reference number ANR-11-LABX-0066] and also received financial support from the Institut National de la Recherche Agronomique (INRA) and the Regional Council of Midi-Pyrénées as part of PSDR project ‘Compter’.

Funding

The authors would like to thank the editors and two anonymous reviewers for helpful comments.

Notes

1. As noted by Beaudry and Schiffeauerova (2009), certain indicators of the Jacobs externalities do not measure diversity per se, but the size of the urbanization externality. These indicators capture instead the ‘global urbanization externalities’ related to market size but not the diversity implied by Jacobs externalities per se (measured by the Hirschman-Herfindahl index).

2. An employment zone is a geographical area within which most of the labour force lives and works and in which establishments can find the majority of the labour force necessary to occupy the offered jobs (INSEE definition). This zoning therefore has an economic meaning, unlike the divisions into administrative areas such as ‘départements’. The 2010 zoning is used here conducted on the basis of the 2006 home-to-work flows.

3. More precisely, level A38 of the 2008 NA is used, which is defined by INSEE as an international intermediate level between the sections and divisions, the two ‘standard’ levels that fit into the International (CITI rev. 4) and the European (NACE rev. 2) nomenclatures of activities. The ‘division’ level A88 will give 23 subsectors compared with the 12 subsectors with level A38, but the disaggregated sectors are, for example, part of the Food and beverages industry (divisions 10–12) or Textile, wearing apparel and leather products sector (divisions 13–15), which are not particularly suitable for the purpose of characterizing a breakdown into industrial sectors. See the description of the categories in the Table A2 in Appendix A.

4. As 304 employment zones and 12 sectors of activities are being worked with, i.e. 3648 possible combinations, it can be concluded that one sector is not represented in one zone for the 335 combinations.

5. The size of area is measured by the working population.

6. In this approach, all the units of the firm have for reference the main activity of the firm. This approach seems the simplest and most consistent to apply to an innovative firm because it allows taking into account all the units of the firm, including those with non-industrial activity. In addition, due to the organization into broad subsectors (two digits), the probability of having units in very different industrial sectors is relatively low.

7. New-to-firm or new-to-market products. The variable I includes ‘radical’ and ‘incremental’ product innovations.

8. Because the share of innovative sales is bounded by 0 and 1 (or by 0% and 100%), it is preferable to use the ‘logit share’ variable which can vary from (Mohnen et al., 2006).

9. To interpret the results more simply, the choice here is to use the logistic distribution of the error term, which corresponds to a logit model.

10. As suggested by a referee, one may question the robustness of the choice of retaining the 50% threshold to affect multi-unit firms in a specialized environment or not (or in a diversified environment or not). To deal with this issue, 40% and 60% thresholds were tested. Approximately 3% of multi-unit firms see their environment to change with these new thresholds, compared with a 50% threshold. The results of the models (see the next section) were then compared with these different thresholds. The results appear robust, as shown by the significance of the explanatory variables which is not changed. This is supported by the small differences (< 0.01%) for the values of Akaike criterion in the estimates with different thresholds. Those estimations are available from the authors upon request.

11. Eight dummy variables are necessary to test all combinations of the diversity specialization and the type of firm. It was preferable to use an interpretation angle with respect to the specialization externalities, not seeking to detail the externalities of diversification (or not) in the non-specialized areas. As the effect of diversification for single-unit firms in the non-specialized areas was not significant, the choice was not to distinguish diversified versus undiversified environment for this type of firm: the single-unit firms in non-specialized areas are taken as a reference for Model 3. And so on, the same was done for the multi-establishment firms. The estimates with eight dummies do not change the results (the detailed tables are available from the authors upon request).