Пространственно – авторегрессионная модель для двух групп взаимосвязанных регионов (на примере восточной и западной части России)
The present study suggests a generalization of the spatial autoregressive model for the case when considered regions are split into two different groups, which have a mutual influence on each other. The weighted matrix in this model is split into four parts and four spatial coefficients are estimated. The proposel model is applied for the analysis of three macroeconomic indicators using the data for Russian regions, previously divided into western and eastern. Our analysis revealed, 1) a positive spatial correlation of the main macroeconomic indicators for the western regions, 2) both positive and negative externalities for the eastern regions and 3) the asymmetric influence of eastern and western regions on each other. Usually "impulses" from the western regions have a positive effect on the eastern regions, but the "impulses" from the eastern regions usually do not affect the western regions.