Portal de Periódicos da CAPES

Measuring the Decline of a Marshallian Industrial District: The Birmingham Jewellery Quarter

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

10.1080/00343400802070894

1360-0591

Lisa De Propris, Luciana Lazzeretti,

Regional Economics and Spatial Analysis

Abstract De Propris L. and Lazzeretti L. Measuring the decline of a Marshallian industrial district: the Birmingham Jewellery Quarter, Regional Studies. This paper presents the findings of a study on the decline of a typical Marshallian industrial district: the Birmingham Jewellery Quarter in the UK. The paper contributes to the current debate on clusters' life cycle by presenting a multidisciplinary methodology that combines historical and economic analyses with demography and organizational ecology models. The paper seeks to explore the patterns of firms' birth and mortality rates, as well as firm density across branches of production activities to measure and understand the decline of the Jewellery Quarter over the last decades. De Propris L. et Lazzeretti L. Mesurer le déclin d'un district industriel du type Marshall: le quartier de la bijouterie à Birmingham, Regional Studies. Cet article cherche à présenter les résultats provenant d'une étude au sujet du déclin d'un district industriel du type Marshall: à savoir, le quartier de la bijouterie à Birmigham (R-U). L'article contribue au débat actuel sur le cycle de vie des regroupements d'entreprises, en présentant une méthodologie pluridisciplinaire qui associe des analyses chronologiques et économiques aux modèles de la démographie et de l'écologie organisationnelle. On cherche à étudier la structure des taux de naissance et de décès des entreprises, ainsi que la densité du parc d'entreprises à travers des branches d'activités de production afin de mesurer et comprendre le déclin du quartier de la bijouterie pendant les décennies récentes. Districts industriels du type Marshall Cycle de vie des regroupements Théorie écologique De Propris L. und Lazzeretti L. Messung des Niedergangs eines Marshallschen Industriedistrikts: das Juwelierviertel von Birmingham, Regional Studies. In diesem Artikel werden die Ergebnisse einer Studie über den Niedergang eines typischen Marshallschen Industriedistrikts vorgestellt, nämlich des Juwelierviertels von Birmingham (GB). Der Artikel versteht sich als Beitrag zur aktuellen Debatte über den Lebenszyklus von Clustern; hierfür präsentieren wir eine multidisziplinäre Methodologie, in der historische und ökonomische Analysen mit demografischen und organisationellen Ökologiemodellen kombiniert werden. Wir untersuchen die Muster der Raten von Firmengründungen und -schließungen sowie die Firmendichte in verschiedenen Branchen der produzierenden Industrie, um den Niedergang des Juwelierviertels in den letzten Jahrzehnten zu messen und zu verstehen. Marshallsche Industriedistrikte Lebenszyklus von Clustern Ökologische Theorie De Propris L. y Lazzeretti L. Medición del declive de un distrito industrial marshalliano: el Jewellery Quarter de Birmingham, Regional Studies. En este artículo presentamos los resultados de un estudio sobre el declive de un distrito industrial marshalliano típico: el Jewellery Quarter (barrio de las joyas) en Birmingham (RU). Aquí contribuimos al debate actual sobre el ciclo de vida de las aglomeraciones presentando una metodología multidisciplinaria que combina los análisis históricos y económicos con los modelos demográficos y la ecología organizativa. Lo que pretendemos es analizar los patrones de tasas de nacimientos y mortalidad de las empresas, así como su densidad en todos los sectores de las actividades de producción para medir y entender el declive del Jewellery Quarter en las últimas décadas. Distritos industriales marshallianos Ciclo de vida de aglomeración y teoría ecológica Keywords: Marshallian industrial districtsCluster life cycleEcological theoryJEL classifications: L23N80O14 Acknowledgements The authors gratefully acknowledge the financial support of the British Academy for the funding of this research (Grant No. SG-33827). The authors would also like to thank David Bailey, Gabi Dei Ottati, Marco Bellandi, Davide Parrilli, and two anonymous referees for useful comments and suggestions. The usual disclaimers apply. Notes In 1947 taxes on luxury items like jewellery increased to 145%, remained as high as 75% in the 1950s, to be abolished only in 1973 when value added tax (VAT) was introduced (Mason, 1998 Mason, S. 1998. Jewellery Making in Birmingham, 1750–1995, London: Phillimore. [Google Scholar]). Organizational populations could correspond, for instance, to sectors as defined by the standard SIC classification. For the Birmingham Jewellery Quarter, one could not rely on such a structured classification of economic activities; service and production activities, therefore, had to be grouped according to their role along the value chain. If the density curve is not non-monotonic, some descriptive statistics (univariate, bivariate, multivariate, spatial, etc.) can be carried out in order to support the qualitative study of the phenomenon. For more details on the Birmingham Jewellery Quarter location, see Cherry (1994) Cherry, G. E. 1994. Birmingham. A Study in Geography, History and Planning, Chichester: Wiley. [Google Scholar], Hopkins (1998) Hopkins, E. 1998. The Rise of the Manufacturing Town. Birmingham and the Industrial Revolution, Stroud: Alan Sutton. [Google Scholar] and De Propris and Lazzeretti (2007) De Propris, L. and Lazzeretti, L. 2007. The Birmingham Jewellery Quarter: a Marshallian industrial district. European Planning Studies, 15: 1–31. [Taylor & Francis Online] , [Google Scholar]. Data were not available for the following years: 1885, 1887, 1889, 1891, 1893, 1909, 1920, 1928, 1949, 1951, 1955, 1959, 1961 and 1962. When computing the density, these gaps have been filled by estimating proxies; it is standard procedure to calculate proxy values for each firm as the average between the density of the year before and the year after the gap. The following formula is applied: where z is a missing year. There is little information about how data were collected, but it is well accepted that the most likely way was for publishers to have had people who physically walked around the streets of Birmingham and stopped at every shop or household asking for names and types of economic activity/trade or looking at name plates. In this way, data and infomation were gathered every year and then published in a Directory. Each Directory classified economic activities by street, name and trade. For the history of the Birmingham Assay Office, see Ryland (1866) Ryland, A. 1866. “The Birmingham Assay Office”. In Resources, Products and Industrial History of Birmingham and Midland Hardware District, Edited by: Timmins, S. London: Robert Hardwicke. [Google Scholar] and Gledhill (1988) Gledhill, A. 1988. Birmingham Jewellery Quarter, Reddich: Brewin. [Google Scholar]. For more information of the evolution of the Birmingham Jewellery Quarter, see De Propris and Lazzeretti (2007) De Propris, L. and Lazzeretti, L. 2007. The Birmingham Jewellery Quarter: a Marshallian industrial district. European Planning Studies, 15: 1–31. [Taylor & Francis Online] , [Google Scholar]. For more information, see Horton (1908) Horton, J. 1908. Business training of jewellers. Interview with Professor Ashley. Modern Business, 1 [Google Scholar] and Smith (2003) Smith, B. M. D. 2003. A Hundred Years of Business Studies at the University of Birmingham, Birmingham: University of Birmingham. [Google Scholar]. Previous contributions that have described the organization of production in the Birmingham Jewellery Quarter include Roche (1927) Roche, J. C. 1927. The History, Development and Organisation of the Birmingham Jewellery and Allied Trades, Birmingham: Dennison Watch Case Co. Ltd.. [Google Scholar], Sargant Florence (1948) Sargant Florence, P. 1948. Investment, Location and Size of Plant, Cambridge: Cambridge University Press. [Google Scholar] and Gilbert (1972) Gilbert, C. 1972. “The evolution of an urban craft: the gold, silver and allied trades of the West Midlands”. Birmingham: University of Birmingham. PhD thesis [Google Scholar]. This number is the sum of all firms in 1880 and the number of firms that were born between 1880 and 1973. Firms' births are considered as discrete events since they are computed at t 1, t 2, t 3, etc. usually corresponding to years, rather than as continuous events, because this would require the exact day, month and year of birth. The most frequently used model to represent these relationships is the log-quadratic approximation (Hannan and Freeman, 1989 Hannan, M. T. and Freeman, J. 1989. Organisational Ecology, Cambridge, MA: Harvard University Press. [Crossref] , [Google Scholar]; Hannan and Carroll, 1992 Hannan, M. T. and Carroll, G. R. 1992. Dynamics of Organizational Populations: Density, Legitimation and Competition, New York, NY: Oxford University Press. [Google Scholar]). The log-quadratic approximation applied to the founding rates has the form: with hypotheses: H 1: θ1 > 0; and H 2: θ2 < 0. L t captures the legitimation process; and C t captures the competition process. This means that the effect of density due to N t , namely the first-order derivative of density, is positive but less than 1; whereas the second-order derivative due to N t 2 is negative. The relationship is non-monotonic and has the form of an inverse ‘U’-shape with a maximum at: Other models to represent these relationships are the generalized Yule models (GY) and the log-quadratic approximation model. However, some authors have demonstrated how these latter models do not always offer convergent estimates (Hannan and Freeman, 1989 Hannan, M. T. and Freeman, J. 1989. Organisational Ecology, Cambridge, MA: Harvard University Press. [Crossref] , [Google Scholar]; Hannan and Carroll, 1992 Hannan, M. T. and Carroll, G. R. 1992. Dynamics of Organizational Populations: Density, Legitimation and Competition, New York, NY: Oxford University Press. [Google Scholar]). In case the Poisson distribution model does not adequately fit the model (because it does not pick up the tendency of the founding rate to vary faster than its average – over-dispersion), the negative binomial model might fit the relation better. Nevertheless, a valuation of the variance of the estimated of the Poisson regression does not reveal high levels of over-dispersion. Period variables have been computed as dummy variables each covering 15 years: 1881–95, 1896–1910, 1911–25, 1926–40, 1941–55 and 1956–70. In density dependence models it is often the case that populations' characteristics are considered as independent variables. In particular, when analysing firm birth in a population, it is reasonable to verify whether the firms' founding rate at time t is correlated with the lagged variable at t – 1. A large wave of firms' births at time t can increase the founding rate due to imitation processes, but at the same time, it can hinder the birth of new firms due to the depletion of resources that competition processes have put in motion. Again, the firm founding rate can be drawn as an inverse ‘U’-shape (Delacroix and Carroll, 1983 Delacroix, J. and Carroll, G. R. 1983. Organizational foundings: an ecological study of the newspaper industries of Argentina and Ireland. Administrative Science Quarterly, 28: 274–291. [Crossref], [Web of Science ®] , [Google Scholar]). To calculate the multiplier of the founding rate, the two parameters of the density setting equal to zero the covariates are considered. This is given by the ratio of the value of density function over the value the density function takes at N min.