Interregional Effects of Innovations in Russia: Analysis from the Bayesian Perspective

Abstract

This study analyzes the interregional effects of innovation in Russia. The hypothesis of the presence of interregional effects is tested by combining the methods of spatial econometrics and Bayesian approach. Using panel data on Russian regions for the period from 2000 to 2021, the author calculates posterior probabilities for a set of spatial regression models that model interregional effects of innovation in different ways. Within the framework of Bayesian approach 6 models were selected for comparison: model without spatial effects (OLS), model with spatial lag of the dependent variable (SAR), model with spatial lag of the error (SEM), model with spatial lags of the explanatory variables (SLX), spatial Durbin model with lags of dependent and explanatory variables (SDM), as well as spatial Durbin model with lags of the explanatory variables and error (SDEM). Based on our calculations, we can conclude that the spatial correlation of innovation in Russian regions is not as strong as it has been assumed in previous studies. This can be considered as evidence in favor of the fact that the concept of interregional spillovers of innovations is poorly consistent with the historical, institutional and territorial peculiarities of Russia, and the methods generally accepted in other countries for such analysis are unsuitable in the Russian context. The results obtained can be taken into account in further research involving spatial modeling of regional innovations. More attention should be paid to the spatial effects of explanatory variables, in particular, interregional spillovers of R&D expenditures, as well as dynamics in the innovation process

Keywords

• innovation
• spatial econometrics
• bayesian methods
• regions
• russia