ПОСТРОЕНИЕ ЭКОНОМЕТРИЧЕСКОЙ МОДЕЛИ ПРОСТРАНСТВЕННОЙ АВТОКОРРЕЛЯЦИИ

Abstract

The article analyzes panel data using statistical indicators for the period 2014-2022 for 85 subjects of the Russian Federation, and a spatial autocorrelation table is constructed. The authors aim their research to analyze and build a spatial model of endogenous growth, characterized by an assessment of the impact of research and development costs and the number of researchers under the age of 39, their flows, on GRP per capita. To build a spatial model, linearization was carried out by logarithm of the studied indicators, as well as statistical software packages (MS Excel, Gretl, STATA) were used to analyze and plot scattering graphs. When evaluating the adequacy of the model using the Broich-Pagan, Hausman tests and the Akaike and Schwartz information criteria, a spatial model with fixed effects, rather than random ones, was chosen. The author identified spatial heterogeneity across the entire data set using global and local Moran indices. The calculated Moran Index for the period from 2014 to 2022 shows that positive spatial autocorrelation (Z>0), respectively, rejects the null hypothesis, that is, there are spatial patterns. Direct, indirect and general spatial effects are also interpreted, as well as geographical maps based on endogenous and exogenous parameters are constructed. However, it should be noted that direct and indirect side effects have a positive impact, emphasizing the complexity of the role and impact of total costs on the implementation of scientific and innovation policy of the state, their redirection, and in increasing the effectiveness of the current economic situation in the subjects of the Russian Federation.