Convergence of Labour Costs in the European Union : Spatial Econometrics Approach

The last five decades saw several waves in the process of European integration and it is not surprising that the issue of convergence within the EU has received considerable attention in regional economic analysis. Most studies still regard regions as isolated in space, and spatial dependencies have also long been overlooked in the standard statistical and econometric work. The valid argument for their explicit incorporation in the analysis of convergence is that regional data cannot be regarded as independently generated, due of the presence of spatial similarities among the neighbouring regions. As a consequence, the standard estimation procedures, employed in many empirical studies, can be invalid and lead to serious biases and inefficiencies in the estimates of the convergence rate. This paper investigates the spatial dispersion and the process of convergence of labour costs across NUTS2 European regions in the period 1996-2004. The methods we utilize are based on the concept of spatial dependence, which we use to augment the standard β-convergence measure in the estimation of a spatial error model. Our results confirm absolute convergence process of labour costs, and also reveal a process of narrowing of the nominal labour cost gap between high- and low-wage regions even after controlling for their different productivity growth. This is an important signal to firms, seeking location for production, and at the same time bad news for regions, relying on labour-costs as a main source of their competitiveness to attract new firms