Olga Demidova (p), Francesco Pastore , Alena Kolyagina

This paper is devoted to the study of diversification and specialization influence on one of the main indicators of Russian labour market, the unemployment growth. The purpose of the work is to find out which effects dominate in the Russian regions, Marshallian or Jacobs, and whether this predominance is stable for different time intervals. The following hypotheses were empirically tested: 1) the dependence of the unemployment rate on the degree of concentration or diversification is non-monotonic due to possible overlapping effects of urbanization and localization; 2) the influence of the degree of concentration or diversification on the level of unemployment depends on the time period. To test these hypotheses nonparametric additive models with spatial effects were used. Both hypotheses found empirical confirmation. It was shown that in Russia, depending on the period, various effects dominated: in 2008-2010, and 2013-2016 Marshallian effects predominated, while in 2010-2013, Jacobs effects dominated. Understanding the key differences between the regions of the Russian Federation will allow the state to conduct a competent structured socio-economic policy that will help to eliminate the negative social and economic consequences from the high concentration in some regions. So, in the crisis period the state should support enterprises whose specialization does not coincide with the main specialization of the region through tax benefits and special subsidies, and in the period of growth - to develop the most promising sectors in each region.

Boriss Siliverstovs (p)

We suggest a novel approach to monitor price level convergence based on the spatial filtering of the data. We capitalise on the observation that the inclusion of the Eastern European Countries (EECs) into the EU had the following consequences:
1) created a large group of homogeneous EEC countries with standards of living (initially much) lower than that of the Western European countries (WECs);
2) distorted the prevailing pattern of spatial correlation among the incumbent EU countries compared to the sample of WECs.
These two facts resulted in overestimation of the degree of spatial correlation for the EU as a whole compared to that when estimated only for Western European countries before the EU Eastern Enlargement.
The convergence (catch-up) process is expected to manifest itself in bringing the pattern of the spatial correlation back to normal through the following channels:
1) by making EECs more heterogeneous among themselves;
2) by making EECs more homogeneous with respect to WECs.
We test our hypothesis by using price indices in Purchasing Power Parity for 37 separate consumption analytical categories for EU 28 countries collected by Eurostat. The data set covers the period 1999-2017. We use the Bayesian Spatial Error Model of De Oliveira and Song (2008) to assess the evolution of spatial correlation over the time period in question. Our simulation evidence as well as empirical results support our hypothesis.
Our main finding is that except for a few consumption good sub-categories the convergence process of the Eastern European countries to the level of the incumbent member states is not over yet.

Camille Regnier (p), Catherine Baumont

In this study, we estimate the role of cities’ connectivity in their economic performance (i.e. their economic growth). To do so, we contruct a network of european cities trough the location of multinational firms’ offices in these cities. The objective of this study is twofold : firsly we contribute to the literature on the role of cities’ place in a global network on their economic performance. A whole body of literature has developed aroud the idea that connectivity is a necessary element to economic growth but there is a lack of proper empirical studies. Secondly, we pursue a methodological goal by linking spatial econometrics and network analysis : we construct a W matrix based on economic relation, and we derive several index of connectivity. Specifically, we derive Bonacich centrality index and we show that this measure of connectivity is directly linked to spatial econometrics. We demonstrate that the Bonacich centrality plays a significant role in cities economic performance, a more important role than other measures of centrality such as degree or closeness centrality. We thus demonstrate that the position held by cities in the world cities network is important for explaining their economic growth, especially when we take into account not only network topology but also the prevailing spillover effect strength inside the network.