Анализ региональной безработицы в России и Германии: пространственно-эконометрический подход
This paper analyzes the regional unemployment in Russia and Germany in 2005–2010 and addresses issues of choosing the right specification of spatial-econometric models. The analysis based on data of 75 Russian and 370 German regions showed that for Germany the choice of the spatial weighting matrix has a more significant influence on the parameter estimates than for Russia. Presumably this is due to stronger linkages between regional labor markets in Germany compared to Russia. The authors also proposed an algorithm for choosing between spatial matrices and demonstrated the application of this algorithm on simulated Russian data. The authors found that 1) the deviation of the results from the true ones increases when the spatial dependence between regions is higher and 2) the matrix of inverse distances is more preferable than the boundary one for the analysis of regional unemployment in Russia (because of the lower value of the mean squared error). The authors are also planning to apply the proposed algorithm for simulated data of Germany. These results allow accounting the spatial dependence more correctly when modeling regional unemployment which is very important for making proper regional policy.