A Stochastic Semi-Non-Parametric Analysis Of Regional Efficiency In The European Union

This paper uses a novel semi-non-parametric approach, the so-called stochastic non-parametric envelopment on Z-variables data (StoNEZD), to measure the performance of 279 European regions (in 28 EU member states and Norway) in the years 2000, 2007, and 2014. The StoNEZD approach combines the main virtues of the parametric and non-parametric methods in a unifying framework. The proposed model accounts explicitly for the presence of contextual/exogenous factors that might affect the regional performance and allows for the use of statistical inference methods to explore the effects of such variables (agriculture-to- gross value added (GVA) share, employment rate, and euro area membership). According to our results, a larger agriculture-to-GVA ratio has a negative impact on regional growth rates, whereas a higher employment rate has a positive influence on regional economic levels. In overall terms, the euro area membership appears to reduce regional average growth rates but seems to enhance regional average efficiency scores.