Detection of geographical clustering: cultural and creative industries in Barcelona

ABSTRACT

Creative clusters are increasingly being recognized as vital tools in the promotion of the competitiveness, innovation, urban development, and growth of cities in developed countries. This paper studies the geography of Cultural and Creative Industries (CCIs) in Barcelona (Spain) for the years 2009 and 2017. We investigate the spatial distribution of firms using the Scan methodology, which identifies the localization of clusters and assigns them statistical significance. Our findings indicate that CCIs are not located haphazardly— they tend to cluster in and around Barcelona’s prime districts. The evolution of the clusters over these nine years reveals distinct patterns of clustering among the twelve CCI sub-sectors. The mature clusters in Barcelona’s core tend to have greater growth and enhanced transformation capabilities. Our results can guide CCI cluster policy, taking into account the specificity of each sub-sector. In addition, they can direct place-based development strategies, creative urban and rural planning, and restructuring in a polycentric context.

KEYWORDS:

• Creative industries
• clusters
• scan statistics
• Barcelona

JEL CODES:

• C38
• Z10

Acknowledgements

This research was partly funded by Martí i Franquès COFUND (European Comission, Horizon 2020), FEDER/Ministerio de Ciencia, Innovación y Universidades (ECO2017-88888-P), the ‘Xarxa de Referència d’R + D+I en Economia i Polítiques Públiques’, the SGR Program (2017 SGR 159) of the Catalan Government and Programa de Ayudas a Grupos de Excelencia de la Región de Murcia (Fundación Séneca, Grant #19884-GERM-15) and PID2019-107800GB-I00/AEI/10.13039/501100011033.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Census tracks (CT), ‘Secciones Censales’, represent the smallest territorial unit for which population data is available in Spain. The number of inhabitants of each CT ranges between 1,000 and 2,500 inhabitants. We consider this spatial unit as a reference to identify spatial clusters. Latitudinal and longitudinal coordinates (centroids) were assigned to each CT and the distance between two CTs was defined as the distance between centroids.

2 For a detailed comparison with alternative methodologies see Appendix 2.

3 The significance level is essential to test whether this local excess of events (in this case the existence of CCIs firms) is the product of mere chance or not (Kulldorff 1997).

4 We decided to use a smaller threshold (10% of the population) in our research. This tuning parameter must be chosen before launching the test in order to avoid problems of multiple comparisons. The selection of this parameter is not relevant to the testing of the null hypothesis of independence but is relevant to geographically identifying clusters. High values of this parameter could identify one cluster formed of several small clusters. Low values allow the identification of more complex forms (e.g., a cluster with an ‘S’ form). The value of 10% is usually selected in the literature.

5 In any case, as shown by Boix, Hervas-Oliver, and De-Miguel-Molina (2015) for a European analysis, CCIs clusters differ at the industry level.

Additional information

Funding

This work was supported by Generalitat de Catalunya: [Grant Number SGR Program (2017 SGR 159), Xarxa de Referència XREPP]; European Comission, Horizon 2020: [Grant Number Martí i Franquès COFUND]; FEDER/Ministerio de Ciencia, Innovación y Universidades: [Grant Number ECO2017-88888-P]; Región de Murcia: [Grant Number Fundación Séneca, Grant #19884-GERM-15,PID2019-107800GB-I00/AEI/10.13039/501100011033].