New Public Management (NPM) has made bureaucracy a topic for inquiry in public sector economics and rational choice theory, but assessing the contribution of the state and its employees to economic growth remains a challenging task. This paper examines micro data on Russia’s federal executive bodies from 2004 to 2023, comprising a total of 1,168 observations which were then processed by a classification method, time series analysis, and regression modeling. Mapping economic activities to their nationally standardized categories provided a way to link each federal executive body to the sector it regulates. The impact of bureaucracy on economic growth was ascertained by observing changes in the structure of value added by the various economic sectors. Models were constructed to include both bureaucracy performance indicators (the number and salaries of civil servants) and control variables influencing the independent variable (the number and average salary of employees in economic sectors), as well as time-fixed effects, agency type, and agency hierarchy. The hypothesis that civil servant salaries are positively correlated with the average salary of employees in economic sectors was confirmed. The hypotheses that the number of civil servants has a positive effect on economic growth and a negative effect on employment were empirically refuted. The hypothesis that the level of civil servant salaries has a negative effect on growth was also not confirmed. On the contrary, each 1% increase in the salaries of civil servants in federal executive bodies is accompanied, all other things being equal, by a 0.64% increase in added value for an economic sector. These results are correlated with those of earlier studies and clarify previously known data on the economic role of bureaucracy.
One distinctive feature of Russian households is that they hold a large proportion of their financial assets in nominal ruble contracts (or assets), that is, contracts whose nominal payment is fixed in rubles: bank deposits and bonds as well as cash. The purchasing power of savings held in such assets is subject to inflation risk. The author arrives at several key conclusions and policy recommendations aimed at maintaining the purchasing power of household savings. First, a nominal interest rate of 8.5% may be regarded as the minimum long-term threshold that provides sufficient incentives for long-term saving given large and historically frequent asymmetric (upward biased) inflationary shocks. Second, inflation targeting constitutes an important part of the incentive for keeping savings in nominal contracts. The introduction of inflation targeting in Russia in 2015 made it possible to systematically counteract the long-standing decline in the purchasing power of accumulated savings. Third, insofar as long-term macroeconomic stability depends on preserving the purchasing power of savings, the proposed inflation targeting regime entails that the policy rate should be increased in response to one-off inflationary shocks whether or not inflation expectations are well anchored. An inflation targeting approach which allows monetary policy to accommodate temporary price-level changes may cause permanent (cumulative) erosion of the purchasing power of savings. The paper concludes that the key factor that would support household savings is not inflation targeting per se, but rather the preservation of the inflation targeting approach that has been established in Russia over the past decade.
This article provides a comparative analysis of the performance of neural network and econometric models in generating pseudo-out-of-sample multistep forecasts. Multistep forecasts with horizons of up to 12 months were generated for two periods: January 2024 to December 2025 and January to December 2025. The robustness of those results was additionally verified over a 48-month evaluation period from January 2022 to December 2025. The hyperparameters of the LSTM model were selected using random search and the Tree-Structured Parzen Estimator (TPE) Bayesian optimization algorithm. The result indicated that, without properly tuned hyperparameters, even a complex nonlinear model provides no advantage over the simplest naive forecast. The stochastic nature of neural network training was taken into account to ensure a valid comparison and reproducibility of the results by testing the selected configurations for stability across a series of runs with different random seeds for which mean forecast errors and their standard deviations were calculated. Forecast accuracy was evaluated using MAPE, RMSE, and the aggregated GMFQI measure; the run time of the models was also compared. The results were compared with forecasts generated by classical econometric models: random walk (RW), ARIMA, and the linear trend (TS) model. The econometric forecasts were constructed using recursive and moving window forecasting. The conclusion was that, with appropriate hyperparameter tuning and control for reproducibility of results, the LSTM model can provide more accurate forecasts, primarily over longer evaluation periods. Nevertheless, simple econometric models provide comparable forecasting performance at substantially lower computational cost.
This study compares the impact of industrial policy and the energy transition on the life cycles of coal mining enterprises in 31 countries from 2017 to 2022 and takes into account differences in national coal industry development strategies by using panel data on enterprise entry and exit, climate policy, energy mix, and government support. Econometric techniques, such as fixed and random effects models, clarify nonlinear relationships between policy measures and coal enterprise dynamics. The findings show that these dynamics are shaped by three opposing forces: the global climate agenda, national energy security strategies, and institutional inertia. Three key paradoxes are identified: the expansion of nuclear power correlates with fewer coal mine closures; the Climate Change Performance Index is significant only in countries where coal is not a strategic resource; and green subsidies do not prevent the entry of new coal enterprises. Four country groups are identified: major exporters, major importers, domestically oriented coal producers, and green transition economies. Environmental goods exports accelerate mine closures, whereas coal rents and the share of coal in the energy mix slow this process. For Russia, the authors propose a differentiated climate policy, cooperation with countries facing similar enterprise dynamics, and investment in clean coal technologies as ways to ensure that climate objectives are integrated into industrial policy without compromising energy security or export competitiveness.
Measuring spatial inequality and assessing its impact on economic development and growth is an urgent challenge in strategic planning and management. Traditional approaches have focused on the scale of inequality and differences between territories, whereas the key issue is whether inequality stimulates growth or hinders it. This paper tests a comprehensive methodological approach for diagnosing the transition of an economy into a disincentivizing developmental phase, which the authors term the “imbalance trap,” that results from the degree of inequality. The toolkit employs a three-component system of indicators modified by the authors: the Spatial Concentration Index (SPCI) adjusted for spatial discounting; the Socioeconomic Polarization Index (SEPI), based on a multidimensional assessment of regional development (SED); and the Spatial Mobility Index (SPMI) computed using transition probability matrices across tercile groups. Panel data for 82 constituent entities of the Russian Federation from 2000 to 2023 in a nonlinear Panel Smooth Transition Regression (PSTR) model empirically verified the threshold values of indices beyond which spatial inequality ceases to stimulate growth and becomes destructive. Analysis of the Russian economy indicates that by 2023 it was not in an imbalance trap. However, the fact that the mobility was approaching critical values and polarization readings exhibited ambiguous dynamics warrants consideration in policy decisions. Additional testing of the methodology on the Siberian Federal District confirmed the universality and scalability of the toolkit for sub-federal diagnostics. This index system can provide governmental authorities with an early warning about destructive trends which require prompt adjustments in regional policy.
