Christophe Ernaelsteen (p), Marcus Dejardin

Traditionally, neoclassical literature generally use Cobb-Douglas functions to model production of firms or, at a more aggregate level, regional or national production (Pendharkar and al., 2008, Felipe and Adams, 2005, Douglas, 1976, Fraser, 2002). These production functions have been generalized by the CES (Constant Elasticity of Substitution) functions and the VES functions (Variable Elasticity of Substitution). Less known, the WDI functions (Weak Disposability of Inputs) consist of another generalization of production functions. It is this last generalization that is used in our contribution.
According to Cobb-Douglas functions, an increase of inputs leads to an increase of production. However, it is possible that by congestion/shortage effect, production will decrease due to an increase in factors. The WDI functions take this kind of phenomena into account. By syllogism with a motorway, increasing the number of cars will initially result in improving the flow of cars but from a moment, by congestion effect, increasing the number of cars can lead to decrease the flow of cars (Derycke, 1997).
The aim of our paper is to test and measure congestion effect in regional production functions in Europe. To do this, we use European data in particular from the EU KLEMS database. The paper brings also some recommendations that can lead to a better orientation of public policies, in particular to the bottlenecks that, once loosened, would improve economic performance.

Boris A. Portnov (p), Natalya Rybnikova

Data on geographic concentrations of economic activities (such as manufacturing, construction, wholesale and retail trade, financial services, etc.) are important for identifying clusters of economic activities (EAs) and concentration forces behind them. However, such data are essentially sparse due to limited reporting by individual countries and administrative entities. For example, at present, Eurostat provides EA data for less than 50% of all regional subdivisions of the third tier of the Nomenclature of Territorial Units for Statistics (NUTS3). Light at Night (LAN), captured by satellite sensors, is likely to differ by intensity, depending on its source. As a result, LAN levels can become a marker for EAs; the present study attempts to verify this possibility. As the present analysis indicates, the inclusion of LAN intensities into multivariate models (in addition to standard economic and locational variables) helps to explain up to 88.8% of the economic activity variation, performing especially well for manufacturing, construction, and agriculture (the coefficient of determination, R2-adjusted=0.754-0.888). The study thus confirms the feasibility of using LAN satellite measurements for reconstructing geographic patterns of economic activities, information on which may be restricted or unavailable due to sparse or incomplete reporting.

Anna Bufetova (p)

The introduction of market mechanisms regulating the economy in the Russian Federation has led to change of spatial distribution of economic activity and triggered the process of its concentration and further increase of regional economic development disparities. Concentration of economic activity is usually considered to be positively related to productivity. So we expect that spatial concentration of economic activity in the RF has led to growing inequality in regional productivity, including labour productivity.
The aims of this work were to study the general trends in regional labour productivity, to test the hypothesis of its positive relation to concentration of economic activity, to analyze the evolution of distribution of regions by labour productivity and the role of spatial externalities in the dynamics of regional labour productivity.
We study the evolution of regions distribution by labour productivity and convergence process among regions by analyzing the change of the region relative position inside the cross-sectional distribution of regions by labour productivity. The growth process is modeled as a first-order Markov chain. Spatial effects in our study are introduced within the Markov chain framework using regional conditioning and spatial Markov chain. By applying this technique we achieve a more detailed picture of the evolution of regions distribution by labour productivity and spatial disparities. These tools allow studying how the economic performance of a region and its position inside the labour productivity distribution can be explained by its geographical environment.
The results of the analysis, based on a data set for 79 Russian regions over 2001-2014 period suggest that the process of economic growth has been characterized by slight divergence of regional labour productivity and discover some specific features of its dynamics in different group of regions. Regional conditioning and spatial Markov chain clearly indicate that location and physical geography matter to explain inequality and unevenness of growth of regional labour productivity. The changes of the relative position of a region in the cross-sectional distribution are highly constrained by its geographical environment. The results of our analysis confirm the presence of positive spatial externalities but their development is conditioned by the level of regional inequality in labour productivity. Too high level of inequality prevents their manifestation.