![Sample our Urban Studies journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110826/TOC-Subject-Sample-2021-Urban-Studies.jpg"})
Abstract
This paper examines existing disparities in technical efficiency levels across the European regions over the period 1986–2002. The results reveal that technical efficiency is not randomly distributed across space in the European setting. On the contrary, the different tests performed highlight the presence of positive spatial autocorrelation and spatial heterogeneity in the distribution under consideration. In fact, we have detected several regional clusters characterized by similar efficiency levels distinguishing them from the rest of the sample. Nevertheless, the estimates carried out show the existence of a process of regional convergence in terms of technical efficiency during the study period. Our findings also reveal that factors such as the regional stock of capital per worker or the patterns of productive specialization are relevant in explaining the changes in technical efficiency experienced by the European regions between 1986 and 2002.
Acknowledgments
The authors would like to thank two anonymous referees for their helpful comments and suggestions. The usual disclaimer applies. Financial support from Spanish MEC (Project ECO2008-05072-C02-02) is gratefully acknowledged.
Notes
Nevertheless, this line of research has been followed by numerous works in a cross-country context (e.g. Färe et al., Citation1994; Maudos et al., Citation2000a; Krüger et al., Citation2000; Kumar & Russell, Citation2002). Likewise, there are several studies based on single country data (e.g. Maudos et al., Citation2000b; Puig-Junoy, Citation2001; Karadağ et al., Citation2005).
NUTS is the French acronym for “Nomenclature of Territorial Units for Statistics”, a hierarchical classification of sub-national spatial units established by Eurostat. In this classification, NUTS-0 corresponds to country level and increasing numbers indicate increasing levels of sub-national disaggregation.
The full results of this analysis are available from the authors upon request.
See Cooper et al. Citation(2000) for a complete description of this methodology.
The data provided by Cambridge Econometrics are based mainly on information supplied by Eurostat. Eurostat nevertheless has some significant limitations, especially when it comes to data relating to the 1980s. Cambridge Econometrics has therefore opted to complete Eurostat data with national statistics from the various countries and interpolation methods.
The regional data provided by Cambridge Econometrics are only available from 1980.
As a result of lack of regional data, the countries incorporated into the EU in 2004 and 2007, the Länder of former East Germany and the French Overseas Departments and Territories had to be excluded from the analysis.
To check the robustness of our findings, we used various different spatial weight matrices. In particular, we constructed two additional matrices W based on the 15 and 20 nearest neighbours. Nevertheless, both matrices yielded in all cases similar results to those discussed later.
The implications of this issue are potentially relevant, especially if we are interested in obtaining information about the statistical significance of efficiency scores in individual regions. To this end, various studies have constructed confidence intervals around efficiency indices using bootstrap techniques (e.g. Simar & Wilson, Citation1998, Citation2000; Enflo & Hjertstrand, Citation2006). Nevertheless, as pointed out by Rey Citation(2001), the presence of spatial autocorrelation invalidates the use of standard bootstrap methods in the context of our paper.
The data on the set of control variables were drawn from Cambridge Econometrics.
In order to check the validity of the decision rule used to select this specification, we also estimated an alternative model including the spatial lag of the dependent variable in the list of regressors (spatial lag model). Nevertheless, the results show that the spatial error model achieves a better fit than the spatial lag model in all cases. Further details on this issue are available from the authors upon request.
In light of this result, we considered the possibility of analysing the contribution of public and private capital. However, lack of data for our sample regions prevented us from studying the effect of these variables on efficiency changes.