Analysis Of The Influence Of Spatial Weight Matrixes On Estimates Of Regional Indicators[Анализ Влияния Пространственных Весовых Матриц На Оценки Региональных Показателей]

The article is devoted to modeling the spatial dependence between macroeconomic indicators of Russian regions, using spatial weight matrices. The authors perform a comparative analysis of five matrices taking into account only the values of the indicators themselves in neighboring regions, without the influence of other macroeconomic series of data. The article analyzes papers that describe ways to build weighting matrices, both based only on geographic data and on economic indicators; the five matrices are applied to three models built on Russian regional data. The study is based on regional inflation and GRP data. In addition to modeling on data about all regions, spatial models were also built separately for the Western and Eastern parts of Russia. Based on the results of the study, we can conclude that for a number of GRP figures, the choice of spatial matrix significantly affects the ratio estimates and errors of in-sample forecasts. The way in which these matrices are considered in the model is of secondary importance. For regional inflation, on the contrary, the model type is more important, and most matrices give nearly identical results in each model.