Analyzing polycentricity

Abstract

Spatial restructuring processes occurred in many economically advanced countries with opposing forces of concentration and deconcentration leading to the emergence of polycentric urban configurations. However, diverging empirical approaches are applied in several studies – a matter that contributes to conceptual confusion and a ‘fuzzy’ use of the term polycentricity. Against that background, a discussion is held to what extent this conceptual fuzziness can be narrowed down to derive a working definition for a quantitative study of polycentricity. Likewise, it is evaluated how to translate this concept into an empirical approach. A spatial scale of 1 km² grid cells ensures a high analytical depth and a joint consideration of employees and built-up volume reflects the multidimensional nature of urban spatial structure. Thus, prospects and limitations of jointly considering employees and built-up volume are evaluated by a literature review and own empirical analyses. Additionally, several conceptually different methods for subcenter identification are applied to the same data and the same study regions. These methods are threshold based ones, exploratory spatial data analyses and semi-parametric regressions. A revisit of these methods’ rationales as well as their outcomes for both employees and built-up volumes finally permits assessing how far a fuzzy concept can be narrowed down to make it analytically feasible. Regarding the polycentricity of four German city regions, the finding is that spatial disparities exist in all study regions. Yet, comparing the empirical results reveals that an agreement on the number, size and precise location of the identified centers and subcenters cannot be reached. This result holds for both employees and built-up volumes although these variables’ spatial patterns differ in details. The empirical conclusion is to apply a ‘double-funnel’ concept to cope with conceptual and analytical fuzziness of polycentricity. This concept’s base is to successively increase the level of detail of both the spatial scale and the applied set of methods. From a conceptual point of view, it can be concluded that it is necessary to consider not only one urban or regional economic model but to incorporate several conceptually different models to understand the variegated nature of urban spatial structure – and thus that of polycentricity.