Modeling of differentiation of consumer prices in neighborhoods, remoteness and interregional trade

The paper presents the results of econometric modeling of interregional differentiation of consumer prices using three different weight matrices, which allow to take into account different forms of spatial interaction of regions. Estimates obtained when using the inverse distance matrix were biased and insignificant on the basis of dummy variables. The use of normalized neighborhood and interregional trade matrices allowed for comparable parameter estimates, with the neighborhood matrix being preferred in the case of the wage effect and the interregional trade matrix in the case of the share of workers in industrial production. The estimation of parameters using information on interregional trade volumes was carried out on the basis of a two-step procedure, where at the first step the non-zero elements of the matrix are restored with the help of auxiliary regression, and at the second step the parameters of the main model are estimated.