Towards Indicators of Social Capital for Regional Development Issues: The Case of French Rural Areas

Abstract

Callois J.-M. and Aubert F. (2007) Towards indicators of social capital for regional development issues: the case of French rural areas, Regional Studies 41, 809–821. This paper aims at giving empirical content to the notion of social capital for regional development issues. It first provides a typology of both positive and negative influences of sociological factors on economic performance, and a set of available social capital indicators. Second, using the results of four case studies, the relevance of these indicators is assessed. Third, it presents an econometric study on the role of social capital on economic development. The results show that both local cohesion (‘bonding’) and external social links (‘bridging’) are important in order to define social capital and explain performance.

Callois J.-M. et Aubert F. (2007) Vers des indicateurs de capital social pour les problématiques de développment régional: le cas des zones rurales Françaises, Regional Studies 41, 809–821. Le but de cet article est de construire une méthode de mesure empirique de la notion de capital social pour les questions de développement régional. Nous commençons par proposer une typologie des mécanismes positifs et négatifs par lesquels les facteurs sociologiques influent sur la performance économique, ainsi qu'une liste d'indicateurs reliés à ces mécanismes. Puis, à l'aide des résultats de quatre études de cas, nous évaluons la pertinence de ces indicateurs. Enfin, nous présentons une application économétrique. Les résultats montrent que la cohésion locale (‘bonding’) et les liens sociaux extérieurs (‘bridging’) sont tous deux importants pour définir le capital social et expliquer la performance économique.

Développement régional Capital social Etudes de cas Indicateurs Zones rurales

Callois J.-M. und Aubert F. (2007) Entwicklung von Indikatoren des Sozialkapitals für Themen der Regionalentwicklung: der Fall der ländlichen Gebiete in Frankreich, Regional Studies 41, 809–821. Mit diesem Aufsatz soll dem Begriff des Sozialkapitals für Themen der Regionalentwicklung empirischer Gehalt verliehen werden. Zuerst wird eine Typologie der positiven und negativen Einflüsse soziologischer Faktoren auf die Wirtschaftsleistung aufgestellt, begleitet von mehreren erhältlichen Sozialkapital-Indikatoren. Anschließend wird anhand der Ergebnisse von vier Fallstudien die Relevanz dieser Indikatoren untersucht. Schließlich wird eine ökonometrische Studie der Rolle des Sozialkapitals für die Wirtschaftsentwicklung vorgestellt. Die Ergebnisse zeigen, dass sowohl die lokale Kohäsion (‘Bonding’) als auch die externen sozialen Verknüpfungen (‘Bridging’) zur Definition des Sozialkapitals und zur Erklärung von Leistung eine wichtige Rolle spielen.

Regionalentwicklung Sozialkapital Fallstudien Indikatoren Ländliche Gebiete

Callois J.-M. y Aubert F. (2007) Indicadores de capital social para cuestiones de desarrollo regional: ejemplo de zonas rurales en Francia, Regional Studies 41, 809–821. El objetivo de este ensayo es aportar un análisis empírico a la noción del capital social en lo que afecta a cuestiones sobre el desarrollo regional. Primero damos una tipología de las influencias positivas y negativas de los factores sociológicos sobre el rendimiento económico y una serie de indicadores disponibles de capital social. Segundo, valoramos la importancia de estos indicadores con ayuda de los resultados de cuatro estudios de casos. Tercero, presentamos un estudio econométrico sobre el rol del capital social en el desarrollo económico. Los resultados indican que tanto la cohesión local (‘bonding’) como los vínculos sociales externos (‘bridging’) son importantes para definir el capital social y explicar el rendimiento.

Desarrollo regional Capital social Estudios de casos Indicadores Zonas rurales

Keywords:

• Regional development
• Social capital
• Case studies
• Indicators
• Rural areas

Notes

Note: French translation is provided by author.

1. Other works on social capital emphasize the importance of a third, intermediary, category, which is termed ‘linking’ social capital. ‘Linking’ involves social links that connect individuals with complementary resources or hierarchical relationships. Although this category is conceptually relevant, it was not introduced in this work for both theoretical and practical reasons. On the theoretical side, the bonding/bridging distinction is already rich enough to understand most mechanisms involved in regional development issues. On the practical side, linking is very difficult to distinguish empirically both from bonding and from non-social (i.e. ‘anonymous’) relationships.

2. A well known example in development studies is the case when cultural norms forces you to give some part of your income to your family. Clearly, nobody will be incited to work efficiently, knowing that he can himself rely on others.

3. Other measures of social heterogeneity are possible, such as measures of the evenness of social categories. As they give similar results to the Gini indicator, they were not discussed here.

4. For constructing the phonebook indicator, it was assumed that every household had a phone at home. The market for fixed phones was liberalized only very recently in France, so the measure of the rate of phone users on the phonebook should not be too much distorted.

5. The validity of this indicator is probably contextually specific. It would be interesting to see how indicators of farm structure perform in other countries. It is likely that for other European countries, a link should be found between cooperation and farm structure indicators.

6. This is of course restrictive, as bridging links may be internal to regions. However, restricting bridging to external links is a reasonable approximation for rural areas, which are usually considered as being very homogenous. Note that the bonding/bridging distinction varies a lot among authors. For example, Beugelsdijk and Van Schaik (2005) define bridging in a way that would probably be interpreted as bonding by many authors (including us). The advantage of using a very clear-cut representation of bonding and bridging is to highlight the fundamental trade-off between cohesion and diversity.

7. The number of phone calls/mail exchange between regions would measure the actual communications that take place between a region an the rest of the world. Unfortunately, such an indicator is not available at local levels.

8. In May 2005, there were 316 pays, covering about 70% of the French territory.

9. However, using figures from the 2000 census does not modify the qualitative results on these study areas.

10. Because the pays do not cover the whole French territory, and their definition lacks statistical objectivity, bridging indicators were computed at the bassin de vie level (see the next section) and then averaged over the pays, weighting each unit by its 1999 population. Probit analysis allows the testing of the significance of various determinants in a binary variable. Here, a test was carried out to see whether there were significant differences between the four regions for binary variables such as ‘Trust indicator’ or ‘Lend question’. The results of probit analysis is available upon request from the authors.

11. Employment growth may seem a crude indicator of economic performance. However, employment is almost the only reliable (and robust) economic measure that can be obtained at such a local level. Unemployment is also available, but does not seem to be a very relevant indicator for rural areas. Some authors use professional tax base as a proxy for local GDP. Unfortunately, the rules for calculating tax bases change frequently and display high biases and spatial variations. Note that using as dependent variables unemployment rate and tax base per inhabitant gives results coherent with employment growth concerning social capital variables.

12. In this case, principal component analysis allows the extraction of a ‘common tendency’ from a set of quantitative variables. All the variables, though very indirect, share something in common (intensity of sociability in the case of bonding, openness to external opportunities in the case of bonding). PCA helps to remove the unwanted features of the indicators. A very satisfactory property of the bonding index constructed with the PCA is that it is very robust to changes in the list of variables: it changes very little if a few variables are added or suppressed.

13. All results are available upon request from the authors.

14. Other potential control variables such as initial employment density, initial per capita income or age structure were tested, but were not significant in any regression.