There is broad consensus among economists that regions‚Äô competitiveness heavily relies on their ability to produce innovative goods and services (Baumol 1967, Romer 1990, Grossman and Helpman 1991, Barro and Sala-i-Martin 1997, Los and Verspagen 2006). Main drivers of innovation include, but are not limited to, human and cognitive capital (Quelle), R&D expenditures (Quelle), industrial clusters and structure (Quelle) and foreign direct investments (Quelle). Most empirical studies confirm the presumed positive correlation of these inputs and regional innovativeness, measured for example by patent applications. At the same time, regions operating at similar input level show significant differences in the degree of innovativeness. These differences can, to some extent, be explained by the regions efficiency in using their available input factors (Quelle). The presented paper aims, in a first step, to identify this efficiency by using an outlier robust enhancement of the data envelopment analysis (DEA), the so-called order-α-frontier analysis (Daouia and Simar 2005, Daraio and Simar 2006), for a sample of more than 200 EU regions (NUTS 2). The findings of this model suggest that the regions‚Äô efficiency is partly affected by a spatial factor. Therefore, the study foresees to decompose regional efficiency into a spatial and non-spatial part by introducing a geoadditive regression analysis based on markov fields. The spatial part reveals differences of the efficiency for greater areas. Regions located in efficient areas, for example, are likely to be efficient as well, since they benefit by the efficiency of neighboring regions. In contrast, the non-spatial effect gives an idea on a region‚Äôs efficiency compared to the neighboring and nearby regions.
