Advances in Spatial Economic Theory
: New Techniques for the Analysis of Agglomeration

  • Andrew Crawley

    Student thesis: Doctoral Thesis


    Economic clusters have been one of the key research areas of the new economic geographers over the last 25 years. The topic has received much attention from policy makers and academics alike, however, there has been a great deal of confusion both in terms of defining and identifying a cluster. This work seeks to reincorporate traditional economic thinking into the study of clusters by bringing back the focus of this phenomenon through a spatial diagnostic framework, allowing the construction of a new theoretical model. After reviewing a large number of studies, it has become clear that methodological approaches need to incorporate more of an intuitive element, thereby reflecting the fluidic nature of a cluster. The research draws on the method constructed by De Propris (2005) and applies it to manufacturing data from South Wales. The results give a contemporary description of the distribution of industry in the region, but fail to show the existence of clusters. The method was found to be inconsistent when applied to disaggregated data, which prompted an investigation into what a cluster is thought to be and how they link into the traditional agglomeratory notions of Marshall (1890).

    Firstly, the traditional tool of cluster investigation, (i.e. the location quotient) was amended to allow for variance. This was then complimented with the construction of a new decomposition model that enables a more detailed position of analysis to be achieved rather than making gross generalisations about individual sectors. The results for both methods when applied in South Wales have been positive, permitting more detailed information about sectoral specialisations to be uncovered. Finally, a new measure of agglomeration was introduced that is able to capture the main attributes of the force. This work has devised new methods of analysis and prompted a rethink regarding economic clusters as well as advocating the continued development of more detailed spatial frameworks in the future.
    Date of AwardJul 2008
    Original languageEnglish

    Cite this