Predicting bank defaults is an important task for the entire economy. Early identification of troubled banks helps to prevent impending bank failures or minimize the losses associated with them. The paper discusses the state of the art of instrumental methods and data used for this purpose. The theoretical background, the evolution of methodological approaches used to predict bank defaults, the specifics of data handling, and the lists of predictors that are included in early warning models are successively reviewed. We conclude that there is still considerable controversy in the literature regarding both the methods and the variables to be used in predictive models. Machine learning methods show a better ability than traditional statistical models to detect non-linear dependencies and to handle large samples. Their advantages are often offset by out-of-sample estimation. Other limitations of such methods are the risk of overfitting and the difficulty in interpreting the results. The lists of potential predictors of bank defaults also vary from country to country. Most commonly, predictive models use bank balance sheet data and financial ratios. However, there are studies that show that forecast accuracy improves when market, macroeconomic and non-financial indicators are included for special countries. Prospects for further research in this area include finding an optimal combination of parametric and non-parametric approaches, investigating the potential of non-financial indicators as factors in bank failures, and research on large samples including both developed and developing countries.
This article discusses ways to obtain convergence of calculation results obtained using before-tax and after-tax cash flow discounting models. Various ways of achieving this goal are proposed. Two concepts are considered - the concept of equivalence of post-tax and pre-tax bases of calculation, within which the theoretical equality of the results of calculations of the value of an asset, calculated based on post-tax or pre-tax cash flows, is implied, and the concept of non-equivalence of post-tax and pre-tax bases of calculation, within which the value of an asset generating cash flows is determined based on the required basis of the value of the asset itself. To align the results of estimates calculated in pre-tax and post-tax bases, the use of cash flow duration is considered. In the third part of the article, a recommendation is provided for the analysis and adjustment of the observed risk premium concerning post-tax and pre-tax cash flows of shareholders. In the fourth part of the article, a conclusion is drawn about the incorrect application of the results of calculations based on the CAPM model when assessing companies and investment projects, and recommendations are given on how to achieve more correct discounting of cash flows at the rates of alternative income calculated via the CAPM and WACC models. Suggestions are provided on the analysis and adjustment of the observed risk premium in relation to after-tax and before-tax capital cash flows. The appendices consider the theoretical impact of inflation on the amount of the observed risk premium and examples explaining recommendations for the correct accounting of the tax factor.
The article was prepared on the base of results of research carried out at the expense of budgetary funds under a state assignment from the Financial University under the Government of the Russian Federation. Abstract. This article examines the interaction between the world of economic systems and the world of their computer-mathematical models. The significance of this issue is growing with the expansion of digitalization and intellectualization of the domestic and global economies. The article consists of two parts. The proposed solution of the problem is based on the consistent typology of economic systems and their models that was developed in the first part of the article. The world of real economic systems was presented in accordance with the systems economic theory as a set of object (organizational), process (logistics), project (innovation) and environmental (infrastructure) systems operating in economic space-time. The world of the most well-known computer-mathematical models is currently represented by agent-based (agent-oriented), econometric, cognitive and equilibrium (optimization) models. In this situation, the relevant task is to develop methods for forming pairs "economic system - its mathematical model" that form the basis of computer-mathematical modeling of economic systems. The relationship between the subject and instrumental spheres of modeling is reflected in the concept of the systems modeling paradigm developed in this article. An in-depth study of the interactions between economic subsystems leads to an understanding of the role of subsystems not only in the processes of intra-system exchange of spatiotemporal and cognitive resources, but also in the integration and separation of these resources. Since each model type reflects one of the aspects of economic system functioning, there comes the task to aggregate models for their comprehensive and adequate reflection of economic system's activities. The application of the proposed framework creates the prerequisites for improving the adequacy of modeling, expanding the possibilities for the effective use of models for forecasting and regulating the behavior of economic systems, and ensuring a higher level of user confidence in the modeling results. The methodology of the research is based on the principles of a systems paradigm, spatiotemporal, and cognitive-competence analysis.
The paper proposes a model for optimizing government support for research and development (R&D) prior to the implementation of an investment project in a real sector enterprise. Budget subsidies an enterprise for compensating part of R&D expenses, and deductions from the income tax base (with the certain coefficient) are considered as such support measures. After the R&D stage, an innovative project starts implementation only with some probability. The enterprise operates under uncertainty, its profits' flow is modeled by a stochastic process, and after the project implementation it changes to another stochastic process. The study of optimizing the support provided by the government to an enterprise for R&D is based on the principle of maximizing the expected integral budgetary effect from the operation of this enterprise. The control parameter is the index of government support for R&D, which characterizes the total direct and indirect budget expenditures (subsidies and lost tax revenue, respectively) per unit of R&D costs. The formula for the optimal government support index is explicitly derived. We analyze the dependence of this optimal index on the tax burden, the volatility of the enterprise profits, the probability of the project implementation, and the effectiveness of innovation project. The situations are described when the optimal budget effect will be achieved without any support from the state, as well as when the maximum possible state support for R&D is not optimal. The analysis considers situations where the optimal budgetary effect can be achieved without government support, as well as situations where maximum acceptable support for R&D may not be optimal.
The development of urban agglomerations is impossible without long-term scientifically based forecasting of traffic flows, which should take into account not only the existing transport infrastructure and established travel routes, but also changes in both the technical and technological properties of vehicles and patterns of behavior of the population. At the same time, the overwhelming majority of models currently used to forecast traffic flows are based on the analysis of a relatively small number of variables that influence the choice of route. The purpose of this article is to demonstrate the capabilities of the agent-based approach to modeling traffic flows in comparison with currently used models based on the use of physical analogies, in particular, the gravity model. To implement this task, the authors developed a multifactor model of job selection by the population, taking into account the multiplicity of criteria for an individual to select a place of employment, as well as the different significance of these criteria for different groups of the population. The model assumes that the choice is made not only by potential employees, but also by employers, and the formation of paired combinations occurs on the basis of the Gale-Shapley algorithm. Comparative calculations carried out using multifactor and gravity models on the same conditional data showed that the roughening assumptions of the gravity model lead to significant deviations in the data obtained. At the same time, the multifactor model can be used for more accurate forecasting of traffic flows, provided that the results obtained are calibrated and adjusted based on statistical data.
The growing uncertainty due to large-scale changes happening at the current stage of socioeconomic development increases the relevance of the problem of analyzing and shaping the future quality of life. The article suggests that in order to prevent possible negative consequences of the impact of major challenges, it is necessary to develop methodological foundations for predicting the quality of life indicators in the Russian regions. The innovative essence and complexity of the research problem led to the use of quality economics methodology and advanced computational methods, which involves the use of a supercomputer. The goal of the work is determined by the need to develop a theoretical and methodological basis for assessing quality of life as an instrument for reconciling strategic priorities and forecast indicators of socioeconomic development of the region. The foundations established at the state and regional levels to ensure the economy's ability to withstand major challenges were considered. Prospects for further use of quality-of-life assessment indicators in quality-of-life modelling to account for significant changes taking place in the economy. The step was taken to create a prognostic model of quality of life. The quality assessment methodology developed in the IRES RAS was verified, with emphasis on its applied nature. In development of this methodology, a conceptual model is presented that describes three levels of contribution of exogenous and endogenous factors to the processes of formation of quality of life. In the future, it is proposed to direct research on specific quality of life issues towards a unified approach that combines the methodology of quality economics and the latest supercomputer technologies.
The article presents a sensitivity analysis of credit risk parameters - including Value-at-Risk (VaR), Expected Shortfall (ES), the range of potential losses, skewness, and excess kurtosis coefficients - to the level of default correlation. The results indicate that VaR may decrease as default correlation rises, confirming core findings of H. Penikas that challenge established assumptions about monotonic relationships in these metrics. The scientific novelty of the present paper lies in evaluating default correlation's impact on typical credit portfolios of varying quality, derived through scaling of factual data from PJSC "Sovcombank". A bottom-up approach is employed, accounting for borrower-specific characteristics including Probability of Default (PD), Loss Given Default (LGD), and Exposure at Default (EAD). Two primary factors driving the non-monotonic relationship between VaR and default correlation are identified: the range of potential losses and concentration in the statistical distribution's shape. Their interaction varies with default correlation levels and average default frequencies: exhibiting opposing effects in high-quality portfolios and predominantly aligned effects in low-quality portfolios. High exposure concentration is also noted as a potential contributor to non-monotonic dynamics, wherein the default of a major borrower significantly impacts portfolio metrics. All, these findings underscore the relevance of refining portfolio analysis methodologies to ensure comprehensive credit risk assessment.
The decision to lend to individuals is associated with many factors, including expert opinions. Some factors are of uncertain nature, which cannot always be described mathematically, but affect the decision-making process. Among the various methods for assessing the risk of default, one can single out the scenario approach using the methods of fuzzy set theory to calculate the values of membership functions. The problem of ranking a set of default risk scenarios taking into account the mental properties of experts and borrowers was not sufficiently studied. The purpose of the study is to develop the new methods for assessing the risks of lending to individuals taking into account the temperament of the borrower and the preferences of experts based on logical and linguistic classification of images. To achieve this goal, an analysis is made of the influence of the experts' opinions with different temperaments on the decision on the lending risk by assigning borrowers scoring points selected randomly from a given set of membership functions. Based on the proposed method, an algorithm was developed for calculating the risk of default on the analyzed individual's loan with determination of his rating taking into account the experts' opinions. The results of computer modeling showed that with a spread of expert opinions establishing the coefficients of significance of borrowers' indicators at about 30%, individual experts can give a significantly different forecast of the lending risk from the average assessment of all experts, which affects the determination of the borrower's rating. Therefore, when assessing the risk of default on loan funds, it is advisable to use the opinion of at least five experts and rank the borrower by the average assessment. The research results can be used to shape recommendations for experts assessing the risks of loan default, as well as to design an expert system that can speed up the risk analysis of lending to individuals.
The issue of strategic factors for providing national healthcare with human resources is a strategic task for all countries of the world without exception, including Russia. A significant shortage of medical personnel in our country is a strategic challenge that requires long-term measures based on established patterns of various levels. According to methodology of strategizing of V. L. Kvint, RAS Foreign member the authors identified patterns in the impact of healthcare workforce potential on its performance at various levels - global (particularly, intercountry) and national (particularly, interregional). Existing global trends indicate the importance of a high level of medical personnel for health care outcomes. It is shown that factors such as the healthcare system's educational function, aimed at improving the population's medical literacy, play an increasingly important role in healthcare workforce management. A study of the dynamics of the age distribution of physicians in Russia revealed that challenges to human resource potential primarily lie in creating conditions for retaining physicians in the profession, which is especially important after their starting years of work. A study of the dynamics of the age distribution of medical personnel in Russia revealed that the challenges to deficit in human resources primarily in creating conditions for the retention of doctors in the profession, which is especially important after the beginning years of work. This issue is also of strategic importance in terms of using the positive opportunities of state policy to increase the number of state-funded places in medical universities and expand the training of medical personnel. In general, on the basis of V.L. Kvint's strategic methodology, a space of opportunities for increasing and improving the efficiency of the human resources of the national health care system was identified.
The paper aims to develop a model that enables to construct density forecasts of inflation in Russia. The reason is that domestic researches are mainly concentrated on providing point forecasts for inflation with little attention paid to constructing confidence intervals. However, predicting the entire conditional distribution of inflation may provide insights into uncertainty and risks associated with future price level movements implying a point projection as well. Therefore ordinary quantile regression and quantile regression neural network are used as forecasting tools, with a large number of potentially informative indicators being considered. Several ways of imposing the L1-regularization term are employed to implement variable selection. Among them are standard quantile Lasso-regression, Bayesian quantile Lasso-regression and linear Lasso-regression. The performance of the first one turns out to be the most successful compared to benchmarks such as linear and quantile autoregressions. Ordinary quantile regression with selected predictors provides qualitative results when constructing both interval and point forecasts. In turn, the use of the neural network approach allows for improved inflation forecasting over longer time horizons. Additionally, we found that exchange rate volatility, housing starts, government debt and natural gas prices are variables that significantly enhance the predictive properties of the model when incorporated into equations for some quantiles. Taking into account the error correction mechanism has also proved its importance. The models proposed in this paper can be used for constructing point forecasts of inflation as well as evaluating inflation risks.
At present development of advanced civil aircraft is one of the key priorities in Russia. To become commercially successful advanced civil aircraft should provide competitive lifecycle costs. Accordingly, it requires balancing operational expenses and consumer-acceptable pricing during an aircraft design phase. However, price evaluation of a prospective aircraft with the analog method or the income method can be challenging due to the lack of a direct analog or the impossibility of obtaining a reliable long-term forecast of air transportation tariffs for a horizon comparable to the development cycle of the object of assessment. To address this problem, this study proposes a methodology for estimating the price of future commercial transport aircraft, combining cost-based and analogy approaches. The price is calculated as the sum of a modern analog aircraft price and the discounted savings future aircraft operation, adjusted for projected amount of passenger traffic. New regression models of price dependence on technical parameters such as passenger capacity, range, thrust-to-weight ratio, maximum takeoff weights are used to estimate the analog aircraft price and an advanced aircraft operational costs. Additionally, the article describes the results of testing the proposed approach for estimating the price of a next-generation passenger aircraft.
The article provides a detailed overview of research related to the problems of economic dynamics. It is shown that well-known models of economic growth are based primarily on assessing the impact of internal factors on the Russian economy. An econometric assessment of economic growth factors and priority areas of its development was carried out. The following hidden growth factors and their importance were identified: focusing investments in fixed and human capital on the technological development of the country (68.99%); the growth of investments in fixed and human capital mainly in export-oriented industries (13.70%) and others. The implementation of the approach proposed in this paper, based on a comprehensive assessment of the impact of hidden factors on economic dynamics, GDP, and GDP per capita, will consistently increase investment in priority projects and achieve the best result from a variety of potential economic opportunities. At the same time, the fact of the increasing importance of openness of national economies should be taken into account. The latter, being more or less integrated into the global economic system, can both benefit from this participation and incur losses. The main promising areas of development of the Russian economy were identified.
We consider machines used in the production process of an enterprise, the failure of which leads to a stoppage of this process and losses. The technical condition and operating characteristics of machines deteriorate with an increase in their operating time. We construct a model for determining the optimal service life of such machines, which assumes that a failed machine is disposed and immediately replaced with a new machine of the same brand. However, the replacement process usually requires random, sometimes quite a long time. Then it is advisable to use a backup machine, and the subject of consideration is a complex of interchangeable main (working) and backup machines of the same brand. In this case, losses from stopping the production of the enterprise arise only after the failure of both machines of the complex, when orders for the delivery of replacement new machines is not yet been fulfilled. Methods of reliability theory and general principles of valuation theory allow us to construct a model for solving the problem of optimal control of the specified complex in discrete time. It turns out that decisions on decommissioning of one machine of the complex, on its designation as the main or backup machine should be made depending on the existence and condition of the second machine. The calculations carried out allow us to assume that of the two machines in the complex, the machine with the lower operating time should be designated as the main one, however, it is not possible to prove the validity of this hypothesis.
This article analyzes the dependence of stock index dynamics on two types of factors - objective and subjective - which are understood, on the one hand, the dynamics of economic development, and, on the other hand, the activities of various stock market participants, affecting both the dynamics of the entire stock market and the values of stock indices. At the beginning of the article, the structure of the ratios of these factors is considered, and a summary of the content of the article is given. It is assumed that the main trend in the dynamics of the stock index, represented by the regression model, characterizes the results of objective factors, and deviations from it - subjective factors. After reviewing the works devoted to the construction of regression models for various stock indexes, a brief description of the history of the creation of the most famous stock index, the Dow-Jones index, is given. Based on the dynamics of this index for 1897-2024, regression models describing its main trend are built and the deviations of the index from this trend are analyzed. It is assumed that this trend depends on the objective characteristics of economic development, and deviations from it depend on the influence of subjective factors. It is shown that the main reasons for these deviations are incorrect time and intensity changes in the list of companies for which it is calculated, made by the managers of the index. This is interpreted as the influence of such a subjective factor as the "human factor" on the dynamics of this index. The level and significance of these deviations are further assessed against the background of the dynamics of US GDP, and it is shown that although in some periods of time the "human factor" deviates the dynamics of the index from the dynamics of US GDP, in general, the dynamics of the US economy has some "attractive force" for the dynamics of the Dow-Jones index. In conclusion, it is noted that when constructing models of stock indices, it is necessary to take into account the specific features of their structure, including the principles of changing their lists, the role of coefficients in the formulas for calculating these indices, and their other characteristics.
The topological model presented in this article is a tool for the economic and mathematical analysis of diversification processes in enterprises of the military-industrial complex carried out by "Rosatom" State Corporation. Based on an analytical review of scientific literature, the most pressing issues of diversification of defense industry enterprises were identified. The lack of a theoretical and normative-methodological framework for studying transformation processes is noted. Critical analysis of the methodology governing the activities of defense enterprises revealed shortcomings such as a standardized approach to diversification, a predominance of quantitative indicators, and the absence of adequate indicators to reflect economic feasibility. The existing problems can be eliminated mainly by using the topological model described in the article. Its development was based on conditional data that was structurally close to real information on a representative sample of enterprises, including scientific, scientific-production, and production organizations. Based on this model, seven categories of enterprises were identified, characterized by similar patterns in the transformation processes under consideration. Within the model framework, the joint changes in performance indicators and the levels of diversification of these groups of enterprises were investigated, and their dynamic stability was analyzed by comparing the identified patterns for two time periods (2018-2020 and 2021-2023). This work demonstrates graphs showing the topological areas obtained for the identified categories of companies. Numerical boundaries for these topological objects are also provided. The results obtained substantiate the conclusion that a differentiated approach to the problems of transforming the industry in question is necessary and can also be recommended for improving methodological provisions governing the planning processes of diversification at "Rosatom" State Corporation enterprises.
The current stage of development of the Russian economy is accompanied by fundamental changes in the innovation sphere. The processes of obtaining new knowledge and technologies and their effective use to improve the level of socio-economic development of the country largely determine its significance and place on the world stage and play an important role in ensuring national security. However, at the stage of technical preparation for the implementation of innovative projects, not all entities and not always can correctly determine the potential prospects and dynamics of transformations. This indicates the relevance of the topic of scientific research. The current practice of organizing technical preparation of innovative projects is often accompanied by various kinds of errors. The most significant of them are associated with incorrect estimates of the timing of individual stages of the project, a low level of communication between its potential participants and insufficient elaboration of various options for its implementation. The consequence of this state of affairs is obviously incorrect estimates of the dynamics of external and internal environmental factors, insufficient consideration of the physical and cost volumes of material and technical support, a low level of preparation of pre-project and project justifications for the implementation of innovative projects. These and a number of other circumstances entail a decrease in the effectiveness of investments in the innovative sector of the real economy. The authors believe that the use of probabilistic planning methods will increase the level of interaction between all investors and performers, and will also help to better assess innovation risks taking into account the project implementation timeframes. It is proposed to make a decision on choosing a project option for its further implementation using probabilistic planning and game theory methods.
This research aims to develop a model and methodology for forecasting logarithmic returns (logreturns) for a new asset class of AI (Artificial Intelligence) tokens. To obtain one day ahead forecasts for the 0.9, 0.5, and 0.1 quantiles, the use of XGBoost quantile models is proposed, which represent an ensemble, based on gradient boosted regression trees. Quantile models have an advantage in terms of forecasting over traditionally used regression models because they allow for the estimation of not only point forecasts but also of their confidence intervals, while remaining robust to outliers. This is especially important when forming the forecasts for various cryptocurrencies' market characteristics, which are known to be highly volatile. In addition to forecasting, the study conducts a post-forecast analysis using the SHAP (Shapley Additive explanations) method, which allows to interpret the XGBoost model, revealing key factors that are important for forecasting AI tokens' logreturns. Based on the results of feature importance analysis with SHAP, a significant influence of AI stocks' market characteristics, cryptocurrency market investor sentiment, seasonal fluctuations, as well as features related to the Blockchain ecosystem was identified. The paper also discusses and addresses the shortcomings of modern forecasting and post-forecasting analysis approaches for time series in general. The obtained results, in addition to academic interest, are relevant for private investors, risk managers, firms and regulators.
In this work, hedonic indices and hedonic regressions are estimated for smartphone prices. The indices are built using a special collection of big data from online retailers that comprises a large number of representative observations and a wide range of product attributes (1.9 million distinct smartphone price observations since July 2022). We demonstrate that, for smartphones, the dynamics of hedonic indices built with the techniques suggested in the literature are substantially greater than Rosstat price index. Furthermore, we indirectly validate that the skimming pricing strategy of smartphones, in which a high price is initially set for a variety and then lowered. In the case of using simple methods of accounting for quality in the consumer price index, that is common among statistical agencies, the use of pricing strategies by manufacturers and retailers, along with frequent changes in product ranges, ultimately result in the depletion of the observation sample and a bias in the consumer price index. This supports the use of hedonic regression-based techniques to enhance the ways in which price dynamics are adjusted for changes in quality. Hedonic regressions should use all quality attributes that are pertinent to price fluctuations as regressors rather than fixed effects of varieties, according to another finding. A bias in the consumer price index happens when simplified regression specifications are used. For price adjustments, Rosstat will be encouraged to collect information on all qualitative traits of the varieties. Increased confidence in the published data will result from the addition of hedonic methods to Rosstat's toolkit, which will help make the published CPI more representative.
The development of distributed ledger technologies (blockchain) led to the alternative ways to finance projects based on the issuance of tokens - digital accounting units in blockchain platforms. Unlike traditional financing mechanisms, the use of tokens allows for greater flexibility in fundraising and income and management rights allocation among stakeholders. An urgent problem that arises when using this mechanism is to determine the proportions in which tokens are distributed among the stakeholders of the project. Currently, this problem is solved by each project individually, based on its organizers' subjective vision. In the article, the problem of token allocation of a decentralized project is investigated from the point of view of cooperative game theory. A token allocation model is considered in the form of a three-party cooperative game and the properties of its solutions are investigated, based on the core, Shapley value and nucleolus. The model allows to obtain estimates of rational token allocations corresponding to the participants' contributions in the project, as well as stable to individual parties and their coalitions deviations. It also makes possible to solve the inverse problem of assessing the contribution of participants and their coalitions to the value of the project based on information about the past token allocation.
The paper examines the impact of an increase in the export customs duty rate on the functioning of the Russian sunflower seed market. Based on theoretical considerations, a logical scheme of the impact of export duties on market processes is proposed. The revealed interrelations between the main indicators characterizing the state and development of the market became the foundation for the construction of economic and mathematical model of the sunflower market in the Russian Federation. The proposed model is a system of power-law equations describing the dynamics of domestic prices, the volume of demand from domestic processors, the amount of exports, and the amount of gross receipts of the crop in question. The calibration of the model parameters is based on the estimation of linear-logarithmic regression coefficients obtained after linearization of the equations included in the model. Based on the proposed model, a scenario calculation of the main indicators of the sunflower market in 2021-2023 was made for the case of maintaining the export customs duty rate on this crop at 6.5%. With the current trends in the development of the sunflower market and unchanged parameters of foreign trade regulation, average annual prices could be higher by 3.1-4.8 thousand rubles per ton, acreage - by 300-350 thousand hectares, gross yields - by 5-6 million centners, exports - by 1.5-1.8 million tons. Domestic demand under the influence of duties would increase by 664.6-1047.8 thousand tons that is insignificant in the scale of processors' purchases. The export restriction policy had a negative impact on the income of sunflower producers. The revenue of the agricultural sector decreased by 6-7% from sales in the domestic market and 8-10 times from export sales compared to hypothetical values. The increase in the export duty rate, contrary to expectations, did not lead to an increase in budget revenues from the duties paid.
