Unemployment Clubs in Russian Regions: spatial econometric comparative analysis

Transcription

1 Unemployment Clubs in Russian Regions: spatial econometric comparative analysis Danilenko Tatiana (NRU HSE, Moscow, Russia), Demidova Olga (NRU HSE, Moscow, Russia), Signorelli Marcello (University of Perugia, Italy) X World Conference SEA216, Università Cattolica del Sacro Cuore, Rome, June 13-15, 216

2 Motivation The concept of socio-economic development of the Russian Federation till 22 states that the priorities of the state regional policy are (i) balanced socio-economic regional development and (ii) the reduction of interregional disparities. 2

3 Literature review (1) Regional real convergence and regional labour market dynamics Barro and Sala-i-Martin, 1991; Blanchard and Katz, 1992; Armstrong, 1995; Obstfeld et al., 1998; Boldrin and Canova, 21; Magrini, 24; Albu et al., 21; Fischer, 215 Elhorst, 23; Huber, 27; Ferragina and Pastore, 28; Decressin and Fatás, 1995; Francis, 29; Caroleo and Pastore, 21; Falk and Leoni, 21; Marelli and Signorelli, 21; Beyer and Smets, 214; Marelli et al., 214; Beyer and Stemmer, 215; Mussida and Pastore, 215a and 215b); Gray, 24; Marelli et al., 212; Bruno et al., 214; Scarpetta and Worgotter, 1995; Scarpetta and Huber, 1995; Lopez-Bazo et al., 25; Tyrowicz and Wojcik, 21 3

4 Literature review (2) Russian regional development and unemployment Solanko (28); Ledyaeva et al. (28); Kholodilin et al. (212); Akhmedjonov et al., (213); Lehmann and Silvagni (213); Mikheeva (1999); Dolinskaya (22); Galbraith et al., (214); Popov (1999); Ahrend (25); Desai et al. (25); Oshchepkov (215); Demidova and Signorelli (212); Demidova, Marelli and Signorelli (213); Demidova, Marelli and Signorelli (215). 4

5 Focus on some studies for Russia Lugovoy O., Dashkeyev, I. Mazayev, D. Fomchenko, Е. Polyakov. (27). Analysis of Economic Growth in Regions: Geographical and Institutional Aspect. Consortium for Economic Policy Research and Advice. Moscow: IET.: Even during a relatively short period under consideration ( ) one can talk about significant spatial heterogeneity in economic development of Russian regions, which obviously should be taken into account in empirical studies of regional growth. Kolomak, E. (211). Spatial Externalities as a Source of Economic Growth. Regional Research of Russia, 1, 2, pp Moscow: Springer. For the western regions of Russia spatial effects for economic growth positive and statistically significant for both the neighborhood matrix and the distance matrix. For the eastern regions of Russia spatial externalities on their economic growth are limited to the neighboring territories and negative. Kholodilin, K. A., Oshchepkov, A., & Siliverstovs, B. (212). The Russian regional convergence process: Where is it leading?. Eastern European Economics, 5(3), 5-26: Our results show that the overall speed of regional convergence in Russia, being low by international standards, becomes even lower after controlling for spatial effects. However, when accounting for the spatial regimes, we find a strong regional convergence among high-income regions located near other high-income regions. 5

6 Focus on some studies adopting a club/cluster approach Many studies on real regional growth follow a club approach (Quah, 1997; Baumont et al., 23; Canova, 24, Alexiadis (213), Fischer & LeSage, 214). As for a survey, see Alexiadis (213). However, in the literature on unemployment a cluster approach (Overman and Puga, 22) is more widespread. More recently, some chapters in Mussida C. and F. Pastore (Eds.), (215) follow a cluster approach for investigating geographical labour market imbalances. 6

7 Club vs cluster The main technical difference between a club and cluster approach: Using a cluster approach, researchers try to unify regions with close values of independent variables (this requires the using of special multi-dimensional distance between objects, for example, the Euclidean or Mahalonobis). Under the club approach researchers unify regions with close values of the dependent variable. 7

8 Data and weights matrix 8 Russian regions; period ; The dependent variable is regional unemployment rate. W = w w1 w21 K w2 M M O M wn1 wn2 K 12 K n n w ij = lengthinkmof joint boundariesbetweenregions iand totallengthinkmof allboundariesof region i j 8

9 The Moran plots Moran scatterplot, 24 Moran scatterplot, 25 Moran scatterplot, 26 Wz Wz Wz z z z Wz Moran scatterplot, Wz Moran scatterplot, Wz Moran scatterplot, z z z Moran scatterplot, 21 Moran scatterplot, 211 Moran scatterplot, 212 Wz Wz Wz z z z 9

10 Clubs 1

11 Map of Russia High-high club Low-Low club High-low club Low-high club 11

12 The dynamic of average unemployment rate in Russia for , % 12

13 Main Hypotheses H1: spatial effects for the High-High and Low-Low clubs differ from spatial effects for other regions; H2: the determinants of unemployment for the High- High and Low-Low clubs differ from other regions. 13

14 Independent Variables Three groups of variables: 1) variables about the attractiveness of the region 2) socio-demographic variables 3) variables of the industrial structure of the employed population. 14

15 Independent Variables 1) variables about the attractiveness of the region: GRP per capita (variable grp, million rubles/ thousand people) the share of city population (variable city, %) population density (variable dens, %) 2) socio-demographic variables: the age structure of the population (variables below and above, %) the proportion of people with higher education in labour force (variable highed, %). 15

16 Independent Variables 3) variables of the industrial structure of the employed population: the share of employment in agriculture (variable agro, %), the share of employment in industry (variable indust, %), the share of employment in construction (variable build, %), the share of employment in wholesale and retail trade (variable trade, %), the share of employment in the public sector (variable public, %). 16

17 Several minor hypotheses 3) The higher the GRP per capita, the lower the unemployment rate; 4) The higher the share of city population, the lower the unemployment rate; 5) The higher the share of the young, the higher the level of unemployment; 6) The higher the share of the elderly, the higher the level of unemployment; 7) The higher the share of educated population in labour force, the lower the unemployment rate. 17

18 Econometric model ihl hl il l ih h ilh il ih ilh il ih WY WY WY Y Y Y Y Y Y = ρ ρ ρ θ 18 t ihl il ih t ihl il ih t hl ihl t l il t h ih t ihl t t t ilh t ilh c X X X ε ε ε α α α β β β 1

19 Formalization of two main research hypotheses Hypothesis 1. There are no differences of spatial effects in regional clubs. Alternative hypothesis 1. There are differences of spatial effects in regional clubs. Formal main and alternative hypotheses 1: H : ρ = ρ = h l ρ hl H 1 : ρ h ρ l or ρ h ρ hl 19

20 Formalization of two main research hypotheses Hypothesis 2. There are no differences in the influence of the factors on unemployment rates in the regions belonging to different regional clubs. Alternative hypothesis 2. There are differences in the influence of the factors on unemployment rates in the regions belonging to different regional clubs. Formal main and alternative hypotheses 2: H H 1 : β = β = : β h h l l β β orβ h hl β hl 2

21 The results of estimation 21

22 Conclusions - 1 The first hypothesis about differences in spatial effects for regions from different unemployment clubs gained partial empirical confirmation. A negative spatial effect for the High- High club was found. The second hypothesis also gained partial empirical confirmation. Therefore, the same factor may decrease unemployment rate in regions from one club and have the opposite effect or no effect at all in an other club. Two minor hypotheses about the direction of influence of the GRP per capita (hypothesis 3) and the share of city population (hypothesis 4) did not get the empirical confirmation, as the obtained coefficients were insignificant. The increasing the share of young people raises unemployment in the regions (hypothesis 5), as expected, but this factor did not demonstrate a club effect. 22

23 Conclusions 2 The increase in the share of the elderly unexpectedly decreases unemployment in the regions (hypothesis 6) of the High-High club. Hypothesis 7 about the negative influence of higher education on the unemployment rate was not empirically confirmed. The higher the share of educated people in regions of the Low-Low club, the higher their unemployment. The increasing share of employed people in the construction industry and trade sector raise unemployment in the High-High club, while in the Low-Low club the increase of the share of employed people in industry lowers unemployment. The increasing share of employed people in the public sector, which is presented with education and health, increases unemployment rates both in the Low-Low and High-High clubs. 23

24 Thank you!