S. E. Demidova
In conditions of economic uncertainty and the need to attain technological sovereignty estimation of innovation policy efficiency in Russian regions becomes more and more important. However, traditional econometric methods often do not take into account endogen nature of factors and dynamic character of innovation processes, which can cause mixed estimations. The present research proposes to use dynamic panel model using systemic generalized method of moments (System GMM) to analyze determinants of innovation development. On the basis of data of 82 regions of Russia for 2014–2023 the model including lagged values of dependent variable and key factors of innovation policy was estimated. Validity of the model was confirmed by complex of diagnostic tests: the absence of autocorrelation of the second order (p-value = 0.1586) and validity of tools (Hanson test, p-value = 0.6186). Results showed complicated time dynamics. Positive effect of R&D expenses is clear with two-year lag, special economic zones demonstrate positive impact in the short-term period with subsequent inverse effect while human capital of high qualification showed stable negative impact on innovation output. Methodological contribution of the research lies in demonstration of System GMM application for correct estimation of dynamic effects in panel data by Russian regions and elaboration of practical recommendations on taking into account of time lags in developing scientific and technological policy.