XI-th INTERNATIONAL CONFERENCE

«MULTIVARIATE STATISTICAL ANALYSIS, ECONOMETRICS AND MODELING OF REAL PROCESSES» NAMED AFTER S.A. AYVAZYAN

On June 24-25, 2024, the XI-th International Conference "Multivariate Statistical Analysis, Econometrics and Modeling of Real Processes" named after S.A. Ayvazyan, organized by the Central Economics and Mathematics Institute of the Russian Academy of Sciences in cooperation with the Moscow School of Economics of Lomonosov Moscow State University, Yerevan State University and the Armenian Mathematical Society, was held in Moscow.

The conference «Multivariate Statistical Analysis, Econometrics and Modeling of Real Processes» has been held every four years since 1979. It has become a traditional and integral part of the scientific life of Russia and Armenia. In previous years, the conference participants were skillfully and cordially received by hospitable Armenia in one of its most beautiful corners - the village of Tsaghkadzor (translated into Russian as «Valley of Flowers»). The XI-th conference dedicated to the 90^th anniversary of the birth of its founder S.A. Ayvazyan was held in Moscow at the Central Economics and Mathematics Institute of the Russian Academy of Sciences.

The conference was held in a face-to-face and online format with the possibility of remote participation.

Head of the Department of Econometrics and Applied Statistics of the Central Economics and Mathematics Institute of the Russian Academy of Sciences Afanasiev M.Yu. opened the conference: "We are starting the work of the I International Conference "Multivariate Statistical Analysis, Econometrics and Modeling of Real Processes" named after S.A. Ayvazyan. Today marks the 90^th anniversary of the birth of S.A. Ayvazyan, Honored Scientist of the Russian Federation, Head of the Scientific Department of the Central Economics and Mathematics Institute of the Russian Academy of Sciences, Doctor of Physical and Mathematical Sciences, Academician of the National Academy of Sciences of the Republic of Armenia – we dedicate our conference to this event.

In 1979 S.A. Ayvazyan became one of the founders and organizer of the school-seminar which was held in the village of Tsaghkadzor in the Republic of Armenia approximately once every four years. S.A. Ayvazyan held his last school in Tsaghkadzor in 2016. The Xth Anniversary School named after S.A. Ayvazyan was held in Moscow on the basis of the Central Economics and Mathematics Institute of the Russian Academy of Sciences in 2021. For 45 years a lot has changed in this event: the name, the form of the event, the composition of the participants has changed, the topics have expanded. But today in this hall there are friends of Sergey Artemyevich, his employees, students, who help us preserve the best traditions of the school-seminar. For example, the active participation of young people. More than 30 young scientists take an active part in our conference with reports. We wish all the participants of the conference successful work!"

At the opening of the conference, welcoming speeches were made by:

Bakhtizin A.R., Director of CEMI RAS, Corresponding Member of the Russian Academy of Sciences. "On behalf of the leadership of CEMI RAS and on behalf of our directorate, let me greet you all and congratulate you on the start of the next school-seminar! The school-seminar was actively developing, initially it was planned as a seminar on statistical methods and econometrics, but starting from 2012, S.A. Ayvazyan began to actively consider other areas related to computer modeling within the seminar, these are agent-based models, equilibrium models, etc. S.A. Ayvazyan took a broad look into the future and involved new directions in every possible way within the framework of the seminar. The seminar is a unique event, traditionally it was held during the week, like a real intensive school. For example, in 2012 the school-seminar was held from June 26 to July 6, that is, two weeks of intensive sessions devoted to considering various aspects of multivariate statistical analysis. Another uniqueness was that this is a meeting of professionals in the field of statistical methods and econometrics, there is no other such in our country.

What is important? Now we are in a difficult phase not only in the development of our country, but also in global development, so there is a special demand for applied methods of information processing and the creation of tools to help government agencies and decision-makers. Many things that pass in national projects are not sufficiently worked out, these are the tax scale, the calculation of the key rate, monetization, all of them require additional calculations. In this regard, special attention is focused on people who are able to offer instrumental methods for working out these solutions. The main call from us, as the organizers of the event, is that we actively share our knowledge, our vision in these areas and contribute to the development of more elaborate decisions in the field of economic policy. I would like to wish all the participants good health and productive work within the framework of the two-day intensive course, which we are starting today."

Makarov V.L., Scientific Director of the Central Economics and Mathematics Institute of the Russian Academy of Sciences, Academician of the Russian Academy of Sciences: "I sincerely welcome everyone! Within the framework of the conference, two sections will be held within two days, I think that everything will be very interesting and fun, and most importantly, productive! It is very important for us that people receive and share new information at the sessions of our conference. I want to emphasize that I am especially interested in the issue of big data, the discussion of this issue is included in one of the sections of the conference. They can be analyzed in various ways, including with the help of artificial intelligence, which is in high demand among young people these days. From the analysis of numerical data, we move on to speech analysis.

S.A. Ayvazyan created a club of like-minded people and friends, and when you are a member of this club, you already enjoy the fact that you just communicate with your friends! Good luck to everyone!"

Movsisyan Yu.M., President of the Armenian Mathematical Society, Doctor of Physical and Mathematical Sciences: "Dear colleagues, I express my deep gratitude to you for holding the XIth International Conference "Multivariate Statistical Analysis, Econometrics and Modeling of Real Processes" after S.A. Ayvazyan, Academician of the National Academy of Sciences of the Republic of Armenia. In his life and fruitful work, Sergey Artemyevich left a deep unique mark of a scientist, whose heirs are many generations of Russian, Armenian, as well as scientists of the former Soviet Union and other countries. Ayvazyan's multiple talent as a scientist and organizer has been embodied in many excellent scientific works, results and undertakings. His legendary seminars in Tsaghkadzor, Armenia, are still unsurpassed examples of the fruitful work of a scientist and organizer. In Armenia, they love and will remember Sergey Ayvazyan! Therefore, I invite all of you to spend the 100^th anniversary of Sergey Ayvazyan in Armenia. I wish you all success!"

Eliseeva I. I., Scientific Director of the Department of Statistics and Econometrics of St. Petersburg State University of Economics, SI RAS, Corresponding Member of the Russian Academy of Sciences. "Dear friends! Thanks to such a bright person, Sergey Artemyevich, we gather here and maintain our close professional contacts, we are always happy to see each other and help each other. Sergey Artemyevich loved St. Petersburg very much and loved our conferences. Our conferences are held every two years and are more devoted to the study of statistical methods. But Sergey Artemyevich always took advantage of the opportunity, because he was a great lover of painting, and always visited the Hermitage, the Russian Museum and especially the French impressionists and the Russian avant-garde. These were wonderful meetings.

I hope that statistics will take its rightful place in our society, because all these words about "machine learning", "artificial intelligence", "big data" – everything ultimately comes down to statistics. Statistics are the root of everything. It is a tree that has yielded so much fruit that we are reaping and our descendants will reap.

Life becomes more complicated and shows that it is impossible to live without stochastics, without statistical thinking, without understanding statistical patterns, it is impossible to make decisions. In my plenary report, I will also talk about the difficulties that have to be overcome.

Thanks to CEMI, thank you for having been, is and will be! That this is our platform that consolidates us all! Many thanks to the management of CEMI and all employees, especially those who are engaged in statistics!"

Kharin Yu.S., Director of the Scientific Research Institute of Applied Problems of Mathematics and Informatics of the Belarusian State University, Academician of the National Academy of Sciences of the Republic of Belarus: "Good afternoon, dear colleagues! First of all, allow me to greet all of you on behalf of the statisticians of the Republic of Belarus on the opening of the conference. And, of course, I would like to thank CEMI RAS for organizing this conference. I must say that it is very great that the RAS has the opportunity to have along with the mathematical institute, along with the Institute of Economics, to have CEMI RAS, which not only applies mathematical methods in economics very actively and productively, but also changes and refines problems for mathematicians, that is, mutual cooperation.

This conference is dedicated to the 90^th anniversary of the birth of the deeply respected and dear to all of us Sergey Artemyevich Ayvazyan. S.A. Ayvazyan made a huge contribution to the formation and development of applied statistics, econometrics, the theory of stable statistical inferences, computer statistics in the USSR, Russia, Armenia, Belarus, Lithuania, Estonia and other countries. In Belarus, since 1985, Sergey Artemyevich has been a permanent co-chairman of the program committee of the International Conference "Computer Data Analysis and Modeling", which was previously held annually, and now is held once every three years.

In the 90s, Sergey Artemyevich involved Belarusian scientists in international projects and grants, consulted Belarusian statisticians, employees of the National Bank of the Republic of Belarus in the field of econometric modeling and forecasting, and participated in the opposition of dissertations. Belarusian statisticians are deeply grateful to Sergey Artemyevich for his help, contribution to development and are ready to continue cooperation with the scientific school created by Sergey Artemyevich at the CEMI RAS, at the Higher School of Economics, and at Moscow State University. All the best!"

In total, seven plenary reports and 84 section reports were announced at the XI-th International Conference from:

- Russia, Armenia, the Republic of Belarus;

- 13 cities: Blagoveshchensk, Voronezh, Grodno, Yerevan, Minsk, Moscow, Novosibirsk, Orenburg, Petrozavodsk, Rostov-on-Don, St. Petersburg, Saratov, Ufa;

- 40 universities, scientific and financial organizations.

Over the past 45 years, there have been great changes in the theme of the conference. The emphasis is shifting to the maximum use of modern computer technologies and the creation of intellectualized decision support systems. Econometrics has won its rightful place among the effective research tools and in the curricula of higher education, the mathematical means of which are almost entirely related to multivariate statistical analysis. The scope of econometrics and applied statistics in the modeling of real processes is expanding.

Seven reports were heard at the plenary session.

Bakhtizin A.R. made a report on the topic "Development of methods for computer modeling of socio-economic processes". The report identified important areas for further work. It is noted that the scope of agent-based modeling has expanded significantly over the past quarter of a century, incorporating many areas at a variety of scales – from molecular to global. Three most priority areas for further development have been identified:

1.  building models within the framework of economic, environmental, epidemiological systems, regardless of their compliance with existing political boundaries;

2.  creation of agent-based models that link the above systems and a huge amount of data that is increasing in progression;

3.  filling models with realistic, cognitive agents.

Separately, the author noted that the development of AI will become the main trend in the near future. The author draws attention to the fact that the potential of large language models has not yet been revealed, but time will tell whether they will be able to replace traditional methods of modeling and forecasting socio-economic systems[1].

Kharin Yu.S. presented a joint report with Shibalko S.A. on the topic "Statistical analysis of multivariate binary time series". The author noted that multidimensional time series are one of the most common forms of presenting economic and statistical data in the analysis of the dynamics of real processes. Most statistical analysis methods are designed for continuous time series. However, the digitalization of the economy and the entire world around us leads to an increase in statistical data registered in the discrete space of states. For the mathematical description of such data in dynamics, discrete, including binary, time series are used. In the report for multidimensional binary time series, low-parameter Markov models of order s are constructed on the basis of basic functions for two cases (conditionally independent components and fully connected dependence) and a low-parameter neural network Markov model of order s. Consistent statistical estimates of the parameters of these models are constructed, an algorithm for predicting a multivariate binary time series is developed, computer experiments on model and real data illustrating the applicability of the theoretical results obtained in solving applied problems[2].

Kovalenko A.P. performed jointly with Perminov A.I. report on the topic "Construction of high-density clusters by a fully connected neural network with a piecewise linear activation function". The author noted that the practical application of the hierarchical method of the statistical package of SAS applications called the "probabilistic method" is limited only to small values of the dimensionality of the initial feature space and the sample size due to the need to perform a large number of operations to calculate the values of the metric and sorting or a large volume memory to store intermediate values. The report proposed an approach to the construction of high-density clusters based on the use of a fully connected neural network (other names: Rosenblatt's multilayer perceptron, ramjet neural network) with a piecewise linear activation function[3].

Eliseeva I.I. presented jointly with Dekina M.P. report on the topic "On the Issue of Instrumental Variables and Models". At the beginning of the report, the author drew attention to the fact that the last component of the passport of an important scientific specialty for today's event, introduced on February 24, 2021 under the code 5.2.2., called "Mathematical, Statistical and Instrumental Methods in Economics" (Order of the Ministry of Education and Science of Russia No118) is not sufficiently developed in terms of content. On the topic of the report, the author drew attention to the fact that in social and economic sciences, instrumental methods include those variables that allow revealing the essence of the analyzed variables and the relationships between them. The use of instrumental variables involves taking into account a variable that correlates with the dependent variable, but not directly, but through "explanatory" variables. One of the methods that can be considered instrumental can be multilevel modeling, since estimates of the parameters of the dependence of the result on factors depend on which variable is the basis of the zero level[4].

Movsisyan Yu.M. made a presentation on the topic «Stochastic mappings form an algebra with hyperidentities»[5].

Ohanyan V.K. made a presentation on the topic of Reconstruction of Convex bodies by probabilistic methods. Complex geometric patterns are found in many fields of science. Their analysis requires the creation of mathematical models and the development of special mathematical tools. The relevant field of mathematical research is called stochastic geometry. The most popular application of stochastic geometry is tomography. Reconstruction of a body by its cross-sections is one of the main tasks of geometric tomography, the term was introduced by R. Gardner. If D ∈ Rn (Rn - n-dimensional Euclidean space) intersects k-shaped area, then there is a k-shaped section, containing some information about D. The report estimated the possibility of restoring D if we have a subclass of k-dimensional sections[6].

Varshavsky A.E. made a presentation on the topic "A model based on a finite functional sequence and its application to the study of the problem of inequality". The report discusses the possibilities of using the model developed by the author, described by the final functional sequence, to study the problem of income inequality. The model provides a fairly high accuracy of approximation of the distribution of income by equal groups of the population, which is confirmed by the results of empirical studies and justified theoretically. At the same time, a new indicator of income inequality is proposed, interrelated with the quintile and decile coefficients of funds, as well as with the Gini coefficient, with the help of which it is possible to calculate the values of income shares of 20% and 10% of groups (quintiles and deciles) for different levels of inequality[7].

The reports of the plenary session were devoted to the development of methods for computer modeling of socio-economic processes, research on the problem of income inequality, issues of instrumental methods, issues of multivariate statistical analysis and modeling of real processes. All plenary speakers noted the significant contribution of S.A. Ayvazyan to the development and implementation of methods of multivariate statistical analysis and econometrics in the practice of socio-economic research.

In total, more than 100 specialists were registered and took part in the work of the XI International Conference named after S.A. Ayvazyan, including external listeners representing the State Duma Apparatus, the Skolkovo Innovation Center, representatives of the "RAMO-M" Group of Companies and others.

At the XI International Conference named after S.A. Ayvazyan, there were two sections: "Multivariate Statistical Analysis and Econometrics" and "Modeling of Real Processes". Each section held six meetings.

Abstracts of plenary reports and reports of section 1 are published in in the proceedings of the Conference:

Multivariate statistical analysis, econometrics and simulation of real processes. Proceedings of XI-th International Conference / Part 1. Plenary reports. Section 1. Multidimensional statistical analysis and econometrics / By ed. V.L. Makarov. – M.: CEMI RAS, 2024. – 143 p.

Reports of section 2 are published in the proceedings of the Conference:

Multivariate statistical analysis, econometrics and simulation of real processes. Proceedings of XI-th International Conference / Part 2. Section 2. Modeling of real processes / By ed. V.L. Makarov. – M.: CEMI RAS, 2024. – 112p.

The main results presented in the sectional reports in the two sections are described below. Footnotes are given indicating the names of the authors, the titles of reports and the numbers of sections. The full text of the abstracts of the reports can be found in the published collections. In the description of the results, several thematic sections are highlighted.

Methods of Multivariate Statistical Analysis

•  In modern conditions, with large dimensions of observations in economics, medicine, insurance, and other applications, it often turns out that observations have a block structure, that is, they consist of blocks that can be considered stochastic independent, which makes it possible to consistently use not only these large-dimensional observations as a whole, but also the blocks themselves. As a result, the constructed sequential decision rule can be used for statistical testing of hypotheses under the conditions of missing part of the components of observations, and an additional reduction in the mathematical expectation of the sample size is provided, which is especially important when the assumed number of large-dimensional observations is small.

For the problem under consideration, a consistent statistical decision rule is constructed. For this rule, asymptotic decompositions of efficiency characteristics were obtained: probabilities of erroneous decisions and mathematical expectations of sample size. The effect of skipping on performance characteristics has been investigated. Robust (resistant to deviations from model assumptions) decision rules in the conditions of distortions of the probabilistic observation model, when distortions are represented by "clogging"[8], are constructed.

•  Singular spectrum analysis (SSA) is an increasingly popular method of analyzing and forecasting time series. The method is based on the singular decomposition of a trajectory matrix built on the basis of a time series. Singular expansion is the mathematical core of principal component analysis used in multivariate statistical analysis. Therefore, the SSA method (in Russia it was given the name "Caterpillar") is also called the analysis of the main components of time series. An algorithm is proposed, as a result of which it is possible to automate the SSA method so that the construction of time series decompositions, in particular, series of sales volumes with seasonality, is reasonable. The ability to isolate a time series component using SSA is called the separability of that component with the remainder. Mixing of components due to insufficient quality of separability or non-uniqueness of singular decomposition with coinciding eigenvalues leads to the impossibility of their identification. Therefore, it is important to improve separability. One of the possible approaches to solving the problem of automatic decomposition of a series into interpreted components is to sequentially isolate the trend and then the periodicities from the remainder[9].

•  On the basis of thematic modeling, including such methods of multivariate statistical analysis as dimensionality reduction, clustering, classification of multivariate observations, an approach is proposed to reflect the transformation of the perception of the image of the USSR in Russian and foreign media over the past three decades. An analysis of the similarity of the evolution of topics over time for different publications, including the dynamics and thematic structure of publications, as well as the composition of keywords and bigrams by periods, is carried out[10].

•  Estimates of the performance of individual enterprises in the stochastic boundary model may be untenable, since no increase in the sample size can eliminate the uncertainty associated with stochastic shocks. In the report, the authors showed what accuracy in estimates can be achieved on the basis of real data, while the accuracy of estimates means a rank correlation between the assessments of the ineffectiveness of subjects and true, not observed in practice, indicators. Thus, we are talking about the ability of the stochastic boundary model to distinguish more efficient subjects from less efficient ones.

The results of estimating models of the stochastic boundary from a number of studies were collected and the Harrell and Kendall coefficients were calculated based on the estimates of the parameters of the distributions of random components. The values calculated from the data of the fundamental articles are given. The obtained estimates indicate that the lack of ranking ability of the stochastic boundary model is not uncommon, so researchers and regulators should be very careful when relying on statistical methods when measuring the efficiency of enterprises[11].

•  A class of low-parameter high-order Markov models for discrete time series was constructed and a probabilistic-statistical analysis was performed on their basis. The proposed models are determined by special basic functions. The type of sufficient statistics makes it possible to interpret basic functions as some meaningful informative features. The search for such informative features can be considered as one of the ways to build small-parameter models adequate to the problem to be solved. All kinds of special cases of models from the constructed class and their connection with the models known in the literature are described. An alternative to the assessment of maximum likelihood is an asymptotically consistent and effective statistical assessment of the model parameters based on frequencies. This estimate is applicable to some special subclass of the constructed model class. An algorithm for statistical prediction of discrete time series based on the constructed models is described. The computational complexity of algorithms for statistical parameter estimation and forecasting is evaluated[12].

•  Estimating the parameters of the final mixtures of normal distributions is a complex statistical problem that arises when predicting volatility in the energy sector; development of features that increase the accuracy of time series forecasting; creation of intelligent trading systems; choosing the optimal investment portfolio, etc. To solve it, it is traditionally used to maximize plausibility using the EM-algorithm. However, this algorithm converges to the local maximum of the likelihood function, and the resulting estimates may depend on the initial approximation. Moreover, the likelihood of a mixture of normal distributions does not have a global maximum and in certain cases tends to infinity when the dispersion of at least one component tends to zero. The authors consider the problem of zeroing out the dispersions of the components of the mixture from the point of view of statistical learning. It is proposed to interpret it as the result of retraining the mixed model. It is also suggested to use the reparameterization of the mixture. On this basis, an optimization problem is formulated for estimating the parameters by the method of maximum likelihood. Using simulation, it is shown that there can be a non-zero hyperparameter that provides a minimum model error on validation sampling[13].

Structural Complexity of Regional Economies and Economic Growth

•  The complexity of professional groups and structures of professional employment in the regions of the Russian Federation is estimated on the basis of the concept of economic complexity. A comparative analysis of the estimates of the complexity of professional groups is carried out, taking into account their prevalence. The information base of the study is the data of a sample survey of organizations in the constituent entities of the Russian Federation, as well as data on the number of doctors and candidates of sciences in the regions of the Russian Federation. Within the framework of the study, a formal description of the structures of professional employment in the region is carried out on the basis of the concept of identified comparative advantages based on data on the number of eleven professional groups. Estimates of the complexity of employment structures and occupational groups were obtained from data for 2018, 2020 and 2022[14]. It is substantiated that the assessment of the complexity of the structure of professional employment in the region can be considered as a relative characteristic of the level of development of its human capital. Assessment of the complexity of a professional group is a relative assessment of the human capital of its typical representative. The ranks of estimates of the complexity of occupational groups generally correspond to the existing ideas about the average level of human capital development of representatives of these groups of the employed population.

•  The results of the study aimed at identifying the main drivers of growth of the gross regional product are presented. Key among these factors are sound tax policies, interest rates, and labor demand, which have a significant impact on economic growth. Given the interlinkages between these factors, targeted strategies can be implemented to improve efficiency, stimulate investment and increase human capital, thereby creating an environment conducive to sustainable economic growth[15]. The solution of the problem of rational distribution of production factors (capital, labor) in the regions of the Russian Federation depends on the indices of sectoral specialization. In regions characterized by low values of the extractive and manufacturing industry indices, the optimal ratio of costs depending on the value of fixed assets to costs depending on the number of employees reaches its maximum value. Thus, in these regions, investment in fixed assets requires the main focus to stimulate growth. A fairly high share of investments in fixed assets is also required in regions with a developed mining industry. In regions with a developed manufacturing industry, the optimal ratio of costs depending on the cost of fixed assets to costs depending on the number of employees reaches a minimum value. This suggests that in these regions there is a high demand for highly qualified personnel who meet the specialization and level of manufacturing enterprises in the region.

•  An approach to assessing the economic complexity of Russian regions for 24 types of economic activity (TEA) is proposed. A comparative analysis of previously obtained estimates of economic complexity for 82 sectors and 24 TEA for 79 regions according to data for 2019 is carried out. Estimates of the economic complexity of the regions remain highly stable in the transition from data on tax revenues when assessed by sectors to data on production volumes when assessed by TEA. Assessment of the economic complexity of regions according to 24 TEAs can be useful in solving management problems aimed at increasing the economic complexity of the region.

Using the developed program in Python, the trajectories of regional development were modeled. Various trajectories of regional development were built with a focus on maximizing economic complexity based on data on tax revenues for 82 sectors and on the basis of data on shipped products for 24 TEAs for 2019. It is shown that the orientation towards maximizing economic complexity leads to changes in the structures of regional economies. Economic complexity increases in regions that purposefully diversify their economies, in contrast to regions that do not change the structure of the economy[16].

•  Based on regression analysis, the impact of the level of public debt on economic growth is estimated[17]. The issue of the optimal level of public debt of the constituent entities is important for improving the effectiveness of budget policy not only at the regional, but also at the federal level. It is shown that for Russian regions with a high level of socio-economic development, the public debt of the subject can be increased to cover the budget deficit if necessary. For regions with low and medium levels of socio-economic development, exceeding the public debt of the region to 2.36% of GRP is undesirable.

•  Due to the high heterogeneity of the regions of the Russian Federation and the presence of spatial effects between them, it is suggested that the impact of sanctions on micro, small and medium-sized enterprises will be different in different regions. The purpose of the study is to empirically analyze the impact of exogenous shocks on micro, small and medium-sized enterprises in Russian regions using the 2022 sanctions as an example. The authors collected primary data on sanctions imposed on the constituent entities of the Russian Federation and created a database on indicators of sanctions pressure on households and SMEs in the regions of Russia from 2014 to 2022. On this basis, the authors formed a database of the main indicators of the effectiveness of SMEs and sanctions in Russian regions. The need to study the impact of new restrictions on Russian economic agents and compare their effects with the sanctions of 2022 is noted[18].

Modeling the behavior of complex socio-economic systems

•  An agent-based approach has been developed to find optimal strategies for individual behavior of sellers and buyers in a stochastic model of trade interactions using the proposed hybrid genetic algorithm[19]. A new stochastic agent-based model of goods exchange (trade interactions) has been created. The proposed model makes it possible to form optimal conditions for sellers and buyers when choosing the moments of concluding barter and monetary transactions at the individual level of each agent, maximizing the value of the utility function of future consumption. A software implementation of such a model has been performed using the supercomputer agent modeling framework FLAME GPU and the AnyLogic.

•  A novel parallel hybrid real-coded genetic algorithm (RCGA-PSO) has been developed, combining evolutionary search methods based on well-known heuristic operators with particle swarm optimization and machine learning algorithms. The proposed algorithm (RCGA-PSO) provides better time efficiency in comparison with other known genetic algorithms of real coding (RCGA), while maintaining the necessary level of accuracy of the solutions obtained. The possibility of using the RCGA-PSO algorithm to optimize the characteristics of the environment and strategies for making individual decisions by agents involved in barter and monetary transactions is demonstrated.

•  A novel simulation model of the multi-agent socio-economic system (MSES) is proposed that implements individual inter-product interactions between agent-producers and agent- consumers and provides the opportunity to form optimal strategies for individual behavior, taking into account several objective functions, such as average profit and number of buyers, the average utility of future consumption and monetary savings, which should be maximized.

•  A novel algorithmized procedure for the synthesis and training of an artificial neural network (ANN) has been developed to generate surrogate ANN models replacing previously developed agent-type simulation models in order to approximate the values of objective functions and constraints in solving large-scale multi-objective optimization problems. As a result of the conducted numerical experiments, it was shown that even in conditions of the predominant frequency of using ANN-based surrogate  models (instead of the initial agent-based one), it is possible to ensure a sufficiently high level of accuracy of the solutions obtained.

•  A medium-scale agent-based model of the evolutionary development of intelligent transport systems (ITS) has been developed within the controlled configuration of intelligent transport systems[20], in particular, the variated geometry of digital road networks (DRNs), belonging to the "Manhattan Lattice" type. The model allows us to explore the possibilities of using "Smart Traffic Lights" (STL), taking into account the behavior of interacting agents of the ITS through adaptive control of STLs’ states and cycle lengths of control signals.

•  An important bi-objective optimization problem has been formulated and solved using the BORCGA-BOPSO heuristic algorithm to maximize traffic and pedestrian flows according to a set of control parameters, including the states of "smart" traffic lights, the cycle lengths of control signals of the the STLs, etc. The Pareto fronts have been computed for various configurations of the DRNs. The high efficiency of adaptive control of ITS using the BORCGA-BOPSO genetic algorithm and the proposed adaptive traffic light control algorithm (FCA-DBSCAN) has been demonstrated.

•    A model of the organization of railway freight transportation between two nodal stations has been built, which allows synchronizing input and output flows at stations. The ranges of parameter changes under which the freight transportation system can function smoothly have been determined. In addition, for a given value of the demand characteristics for cargo transportation, the most acceptable achievable levels of the degree of inconsistency between the reception and dispatch of goods at all stations have been established by controlling the characteristics of the degree of use of the technical potential of the stations and the mode of distribution of goods from the terminal node station[21].

•  The technologies of broadband Internet access (fixed and mobile), their distribution in time and space across groups of countries with different income levels are considered[22]. High-income countries are accepted as "growth poles" for low-income countries. It is in these countries that new ICTs appear first of all, which are introduced a few years later in countries with lower incomes of the population.

•  As "spatial characteristics", the following are considered: the share of urban residents in the total population of the country; the logarithm of population density; the logarithm of foreign direct investment; the ratio of the logarithm of GDP per capita of a group of countries to the corresponding indicator of high-income countries and a number of others. The actual and model-calculated indicators of the spread of fixed broadband Internet access across groups of countries with different income levels, depending on the proportion of urban residents in the total population, are presented as a "spatial" characteristic for the period 2006-2021.

•  The analysis of the robotization of production and the main factors influencing this process in the countries of Central and Eastern Europe (CEE) has been carried out [23]. The dependence of the density of robotics on the indicators of economic development of these countries (GDP per capita at purchasing power parity, the share of R&D costs in GDP, the volume of foreign direct investment (FDI), the share of employed in industry in the total number of employed, the share of value added of manufacturing in GDP, etc.) is studied. The indicator of the density of robotics is calculated as the number of robots per 10 thousand. It is used to assess and compare the level of robotics in different countries. It is concluded that in Romania and Slovakia, an increase in the density of robotization is influenced by an increase in the share of FDI in GDP, and in Slovenia by an increase in the share of manufacturing in GDP.

Labor markets, employment and migration of the population

•  The qualification composition of the employed population in the regions of Russia has been analyzed, and the index characterizing the ratio of the number of specialists with the highest level of qualification to the number of unskilled workers has been calculated. Additionally, considering the importance of secondary vocational education, the index showing the ratio of the number of specialists with higher qualifications to the number of specialists with intermediate qualifications has been calculated. The calculations were based on data from the Federal State Statistics Service regarding the number of employees in organizations by professional groups in the subjects of the Russian Federation for the years 2018, 2020, and 2022. It is shown that in most regions of the Russian Federation, the number of highly qualified specialists significantly exceeds the number of unskilled workers, and the trend of growth in the value of this index continues. The differentiation in terms of the ratio of specialists with the highest level of qualification to those with an average level of qualification is less significant.[24].

•  One of the most widely studied economic laws is Okun's Law, which states that there is an inverse relationship between changes in real gross domestic product (GDP) and fluctuations in the unemployment rate. Using a hidden class model, an analysis of the heterogeneity of unemployment sensitivity to GDP was conducted, during which regions that are most responsive to crises were identified. The results of the study can be used in the development of economic policies and state support measures during crises, which is particularly relevant at this time.[25].

•  A forecasting algorithm for migration with minimal time lag has been developed, using data from Google Trends Index search query statistics. The analysis shows that the use of exogenous data in migration forecasting improves the predictive power of models. The findings of this study indicate a high potential for utilizing the digital footprint data of migrants on the Internet. The proposed approaches allow for quicker estimates of migration compared to those published by official statistical agencies. Additionally, the use of such data increases the predictive power of models and reduces prediction error compared to SARIMA models, which plays a particularly important role during periods of external shocks, such as the Covid-19 pandemic.[26].

•  In the framework of the study[27] a classical approach to the production function was applied with some modifications. The modified production function was used as a basis for forecasting the dynamics of product output and the labor market, taking into account the influence of scientific and technological progress on improving the efficiency of the use of key production factors. The hypothesis of the study is that a common labor market in the Union State will allow for an increase in product output and smooth out labor market imbalances. In this context, the technological parameter of the function combines wages, migration, and labor productivity. When this parameter is high, these factors are significant for the economies of the Union State, and with the presence of a common labor market, there will be an improvement in the quality of the workforce, an increase in labor productivity, and conditions will be created for leveling the wage and living standards in the Russian Federation and the Republic of Belarus. The analysis showed that the modified production function does not adequately reflect migration flows. There is a substitution of production assets with labor from abroad, resulting in unrealistic data in certain sectors. As part of the continuation of the research, it is planned to assess the model for labor using a similar method and identify the substitution effect for sectors where it exists.

•  The model presented in the work [28] largely relies on a general scheme of macroeconomic models for forecasting the labor market, allowing for scenario analysis and short-term forecasting of employment numbers and the demand for specialists in specific types of economic activities (TEA) by broad professional groups, including specialists of higher and medium qualifications in the fields of science and technology and information and communication technologies (ICT), as well as skilled workers in industry.

In developing the scenarios, data from short-term forecasts by the Ministry of Economic Development, as well as results from the authors' research on foreign experience, were used. In particular, based on the analysis of production structure by sectors, dynamics of investments in fixed assets, and professional-qualification structures of employed individuals, as well as projected values of demand for specialists in countries such as Germany, Poland, Canada, Turkey, and several economies in Eastern Europe, additional scenarios for changes in the demand for specialists and labor force were proposed depending on the trajectory of economic development.

The developed model also allows for refining previously obtained estimates of the imbalance between supply and demand for engineering and technical specialists by presenting alternative methods for calculating the additional need for engineering and technical specialists, adjusted for changes in the sectoral structure of output and taking into account the demand assessment for filling vacant positions. However, this model still cannot analyze the problem of structural imbalance, including that related to wage differentiation among workers in various sectors and professional groups, which requires further research.

•  In the study[29] a meta-analysis of existing estimates of the motherhood penalty was conducted. The sample included over 2,000 estimates obtained from data for 38 countries.

The analysis confirmed the significance of mechanisms leading to the motherhood penalty, such as losses caused by employment interruptions and underinvestment in human capital, the exchange of part of earnings for more convenient working conditions, and reduced labor efforts, including due to high engagement in unpaid domestic work. The hypothesis of lower productivity among mothers was not confirmed.

Control for regional variables revealed a relatively higher motherhood penalty in Western European countries and the United States, and a relatively lower penalty in Latin American countries. This provides certain empirical evidence in support of the hypothesis proposed during the study regarding a possible relationship between the size of the motherhood penalty and fertility heterogeneity. In countries where there is significant heterogeneity among female populations in terms of birth rates (high prevalence of childlessness and/or large families), employers may lean towards smaller gender pay gaps while simultaneously imposing a higher motherhood penalty.

•  The objects of the study[30] are the cities of the Russian Federation. The research sample includes 187 cities, including 25 administrative centers of the subjects of the Russian Federation and 35 mono-industrial towns, which determined its qualitative heterogeneity. A system of statistical indicators has been developed, including performance indicators that assess the labor market in the cities, as well as factor indicators that cover a range of thematic sections: "demography," "economy," and "social sphere." The combination of these methods allowed for the division of the researched sample of cities into homogeneous clusters, which serves as a basis for further investigation of socio-economic factors affecting employment in the cities of the Russian Federation.

Quality of life of the population

•  The assessment of the impact of public transport fare on the demand for transportation services in megacities is an important task addressed by municipal authorities within the framework of fare regulation. Dynamic models of changes in passenger flow on public transport in Moscow have been developed based on the cost of travel, types of tickets, and frequency of trips. The research was based on the results of an online survey of Moscow residents – representatives of a certified consumer panel for conducting marketing research by OMI. Based on the obtained models, it becomes possible to calculate the marginal prices for a specific share of passengers willing to use public transport.[31].

•  An analysis of the social capital of the region was conducted, taking into account the spatial interaction of regions in Russia. Factors that can serve as indicators of social capital were identified. It was found that the results obtained using the SAR model have higher statistical significance than the results of the cross-sectional regression model. The positive estimation of the spatial lag coefficients for the western regions indicates a positive influence of social capital factors in these areas. The positive estimation of the spatial lag coefficients characterizing the influence of western regions on eastern ones suggests the spread of regional social capital from west to east. The assessment of the spatial lag coefficients characterizing the influence of eastern regions on other eastern regions showed instability in trends and requires further investigation.[32].

•  The conducted research demonstrates that peers play a crucial role in the formation of children's health, with the effect working both ways: they can either improve or worsen health. An increase in the average body mass index (BMI) of peers raises the child's BMI, and the presence of friends with weight issues increases the likelihood of the child developing similar problems. Additionally, in all specifications, it was shown that the magnitude of the peer effect for the sample with normal and overweight is several times higher than the effect on the sample with underweight and normal weight. A likely explanation for this result lies in the differences in mechanisms that lead to these issues.[33].

•  • The results of the statistical analysis confirm that the salaries of teachers are currently almost entirely determined by the socio-economic development of the region. However, to reduce the differentiation in teachers' salaries across regions, the level of teacher qualification should be taken into account first and foremost when determining wages. This will help retain the most qualified personnel in the regions and, consequently, contribute to improving the quality of education for the younger generation.[34].

•  The hypothesis that government spending on the budget items under consideration affects individual life satisfaction has been tested[35]. To verify the hypothesis, the authors used an ordered probit model of panel data with random effects, which was estimated using the STATA statistical package. According to the obtained estimation results, partial confirmation of the proposed hypothesis was found. Spending on education and healthcare increases the life satisfaction of Russians, while spending on social policy does not have an effect.

•  The factors affecting housing affordability in 401 regions of Germany from 2004 to 2020 have been studied[36]. The main premise for a detailed examination of housing affordability in Germany is the rise in housing prices that outpaces the growth of population incomes. This trend has been observed in many European Union countries for an extended period and potentially poses a significant risk to the well-being of the population. At the same time, Germany represents an example of a country with pronounced regional differences in social, economic, political, and other aspects. Studying housing affordability without considering these peculiarities may lead to inaccurate results. In this regard, the primary objective of the conducted research is to identify the factors influencing changes in housing affordability in Germany, taking into account the spatial interconnections of territories.

•  To study changes in consumer spending during the atypical crisis of 2022, data from a sample survey of households conducted by the Federal State Statistics Service was used. To achieve the research objective[37], the Kolmogorov-Smirnov test was employed to verify the hypothesis of equality of distributions, and the t-test was used to check the hypothesis of equality of means. Within the framework of the analysis, distributions and mean values of dependent variables (shares of spending on food at home, dining out, alcohol, non-food goods, and services in total consumer spending) were compared at different points in time.

The results indicate a significant increase in the share of spending on food at home in 2022 compared to the years 2019-2021. Interestingly, the share of spending on food at home even exceeds the level during the pandemic, when households had limited opportunities for dining out. In contrast, the share of spending on dining out continues to recover after its decline in 2020, but this share in 2022 is still significantly lower than in 2019. The share of spending on alcohol is also decreasing regardless of the comparison period. The share of spending on non-food goods significantly decreased in 2022 compared to 2021 and 2020, when there was an increase in the share of non-food goods due to the development of marketplaces and the availability of imported non-food products. Conversely, 2022 is characterized by the exit of many companies from the Russian market, which led to a reduction in assortment as well as the share of non-food goods. The share of spending on services has been declining regardless of the comparison period, which may be related both to the exit of companies from the market and the economic shock during which many small businesses providing a wide range of services closed down.

•  In the study and analysis of the standard of living of the population, subjective aspects of well-being play an important role. The research presents the results of analyzing their characteristics among different groups of Russians[38].

As a key indicator for assessing the level of subjective well-being, the question of overall life satisfaction and satisfaction with its individual aspects is presented. These aspects include the following areas of a person's life: personal relationships with loved ones, leisure activities, the environmental situation in the place of residence, access to qualified medical care, and housing conditions.

In addition, other components of subjective well-being assessment are also considered: opportunities for self-development, expectations regarding improvement or deterioration in the future, and emotions that reflect, on one hand, a general sense of safety and stability in the current situation, and on the other hand, opportunities and resources for further development.

According to the results of regression analysis, life satisfaction is largely determined by income level. In second place is the type of settlement in which a person lives (megacity, large cities, medium-sized cities by population, rural areas).

•  An analysis of the impact of child benefits on the level of absolute income poverty among children in Russia has been conducted[39]. The first part of the study is dedicated to analyzing the dynamics of the contribution of child benefits to reducing the level of child poverty in Russia from 2013 to 2021. In the second part, microsimulation modeling is performed to assess the impact of universal, categorical (criteria for payment – child's age), and targeted (criteria for payment – income) benefits on child poverty. The study is conducted for all children as a whole, as well as for specific age groups. For each type of benefit considered, the expenses for its payment are evaluated.

The results of the analysis show that the most effective form of budget expenditure for reducing child poverty is targeted assistance provided based on the criterion of need. However, considering the particular vulnerability of young children, it is useful to explore the possibility of making assistance for young children fully universal. This would eliminate coverage errors (inclusion/exclusion errors) for this group of children under the targeted assistance program. Additionally, the universality of the benefit would ensure support not only for young children living below the poverty line but also for others, requiring only a moderate increase in expenditures. This measure will help ensure that children receive comprehensive development at the earliest stage, making a significant contribution to their future growth.

Demography

•  Based on event occurrence analysis methods, new results have been obtained regarding the impact of the coronavirus pandemic on birth rates. The empirical basis is a nationwide representative survey of the population titled "Person. Family. Society," conducted in the spring of 2023 with a sample of 9,500 individuals aged 18 to 72. According to the results of the regression analysis, the variable associated with the crisis showed statistical significance. The crisis negatively affects the chances of having a first child, reducing them by 0.6 times, all else being equal. The linear variable of the generalized value index is also significant and has a negative impact—respondents who share emancipatory values have lower chances of having a first child, all else being equal. Religiosity is significant and has a positive effect; however, the strength of this factor's influence is small. Respondents from rural areas and cities with populations under one million have higher chances of having a first child compared to residents of million-plus cities (by 2.16 and 1.56 times, respectively). The timing of the crisis was not significant in assessing the chances of having a second child. The value index reversed its influence—if a respondent shares emancipatory values, it does not change the chances of having a second child compared to choosing mid-range responses; however, if they share traditional values, it increases the chances by 1.3 times.[40].

•  To assess the dynamics of population changes in the regions of the Far Eastern Federal District (FEFD), this study[41] proposes not only to evaluate relative indicators but also to identify factors influencing the established trends through regression analysis. This will also allow for a more precise determination of differences between the regions of the federal district in question. Thus, it is possible to highlight regions with more complex demographic situations, not only at this stage of their development (i.e., those regions that require urgent attention) but also in the future, since the factors identified in the study affect not only the current situation but also lay the groundwork for future changes (or their absence). Regions have been identified that not only exhibit the most complex demographic situation (the greatest decline in population) at this stage of development but are also expected to experience further deterioration in the future. The results of the conducted study can be used in the development of strategic socio-economic development programs for individual regions of the FEFD, for the federal district as a whole, as well as for medium-term forecasts of population changes (taking into account the trends in the identified factors).

•  The factors influencing changes in the reproductive plans of Russian citizens during the socio-economic shocks of 2022–2023 have been studied[42]. As part of the strategic project of the National Research University Higher School of Economics (HSE) titled "Life Choices and Decision-Making in Conditions of Instability," a survey on "Reproductive Behavior of the Population in Conditions of Socio-Economic Shocks of 2020–2022" was conducted in May 2023, involving 7,967 respondents. The sample included Russian citizens aged 18 to 44, representing the population of Russia in terms of gender, age, and place of residence.

The novelty of this work lies in the study of individual heterogeneity and the role of psychological factors in making various reproductive decisions during periods of uncertainty. The results show that despite socio-economic upheavals, people make very different decisions: some do not change their reproductive plans, others refuse to have children in the near future, while there are those who wish to have children right now. This decision is not solely explained by material well-being; rather, it is largely justified by psychological factors and attitudes toward the events occurring in the country.

To analyze the factors influencing changes in reproductive plans, binary and multiple choice models were assessed. The analysis of binary choice models revealed that reproductive plans are significantly influenced by both regressors related to the objective living conditions of respondents and variables describing their subjective attitudes toward the situation in the country and in their lives. For respondents over 30 years old, the restraining factors include fear, anxiety, having children, and physiological deadlines, while for younger respondents, career prospects are the main concern.

•  The results of the statistical analysis of socio-economic factors determining family decisions regarding the choice of the month of a child's birth are presented[43]. In Russia, as well as globally, there is a seasonality in births, with the summer months—July and August—being the most popular during the period from 2000 to 2022.

The study utilized data from the Rosstat household sample survey "Comprehensive Monitoring of Living Conditions of the Population in 2020" and data from Rosstat on registered births by month across regions for the years 2019 to 2022. The conclusions were drawn using estimates based on an unordered multiple choice logit model, as well as cluster and correlation analysis.

It is shown that the choice of summer months for a child's birth is influenced by household income (as income increases, families' preferences for July births also rise); in regions where the quality of life improves, the share of the urban population increases, and the level of contraceptive use rises, there is a growing demand for summer months for childbirth.

•  Taking into account the trends in the variability of marital structure, the relationship between family status and an expanded set of factors is examined, which includes not only traditional characteristics (gender, age, education) but also some socio-psychological indicators (trust in people, level of optimism, attitude towards religion) that characterize individual preferences in the marriage market. [44] Considering the fact of the modern expansion of the age structure of family status, we analyze this relationship separately in each age group. A comparative analysis of statistically significant relationships across these groups and categories of marital structure has been conducted.

Modeling the marital structure of the population by age revealed a variety of existing relationships between family status and the factors that shape it. The features of the multinomial logistic regression model used in this work allowed for a comparative analysis of the dependencies of the indicators used with specific categories of marital structure. An expanded set of indicators was considered, including both ordinary socio-demographic factors and individual personal characteristics.

•  The dynamics of morbidity from diseases of the digestive organs are analyzed, along with the main methods for assessing the value of an average statistical life[45]. Additionally, estimates of the economic damage resulting from premature death due to diseases of the digestive organs are provided. To assess the economic damage from premature death, the value of an average human life and the cost of one year of average life are calculated.

Reducing mortality, including from diseases of the digestive organs, will improve the demographic situation and, consequently, reduce economic damage.

Analysis of industries and enterprises, environmental

•  Identified factors determining the development of volume of manufacturing production; periods characterized by different behavior of the initial data, which is characteristic for the economy of the Russian Federation (RF) are investigated. It is intended to use the data obtained in the future to assess the risks of economic security of the country: dependence of domestic commodity markets on imports, deterioration of production structure. [46]

•  The possibility of econometric evaluation of parameters of the Cobb-Douglas production functions of for individual sectors of the Russian economy in the period 2005-2022 is being considered. Analysis is conducted for 15 industries: 2 mining industries (fuel and energy minerals and non-fuel) and 13 manufacturing industries.[47]

This method of estimating production functions allows the identification of sectors where output increases are due, among other things, to scientific and technological progress. This group included five industries: food production; mining of mineral resources, other than fuel and energy; textile and clothing production; Leather, leather and footwear production; manufacture of vehicles and equipment.

•  The problem of constructing short-term forecasts and current forecasts in “nowcasting mode” for key macroeconomic indicators is of particular importance for public authorities.  The solution of task of operational forecasting of annual growth rate of total industrial production in economy of Belarus is proposed on the basis of data of monthly surveys of enterprises.  Using monthly balance time series of answers, the indices are calculated: entrepreneurial optimism, current average wealth and forecast average wealth.  Based on the established co-integration dependence between time series of variables under consideration, a model for error correction using the Engel-Granger ECM-EG approach and a vector model for error correction using the Johansen approach are constructed.  The following indicators are used as characteristics of accuracy of forecasts: RMSE is the standard deviation of the residuals (prediction errors), MAE is the medium absolute error, MAPE is the medium absolute prediction error in %. At the time interval under consideration, the average absolute error of one-step forecasts for a month averaged about 2% for both models.  However, the problem of structural changes or anomalous observations in the forecast period remains and can sometimes be solved by special methods to remove the mixing of forecasts.[48]

•  The preferential business regime in the Russian Arctic is a new regulatory mechanism requiring implementation monitoring and evaluation of effectiveness. The study aims to develop and test a tool for assessing the effectiveness of preferential treatment at the level of motor companies, as well as to identify general trends in the implementation of the regulatory framework.

The authors note the pronounced dual impact of preferential treatment on resident companies depending on their time of establishment.  The group of long-established companies differs from the group of first-time resident companies in the predominant factor in generating revenue: for «old» is characterized by a predominance of capital contribution, whereas for new - labor.  In this regard, the importance of «new residents» projects for the development of the regions of presence may be more related to the implementation of the function of social stabilization of the labor market, Creation of new high-productivity jobs and conditions for population retention in the territories through professional implementation and employment of broader segments of the population.[49]

•  The impact of GDP growth, agricultural and aquaculture production, energy consumption and transport infrastructure development on carbon emissions is examined. The aim of this study is to fill the gap in ecosystem research in the context of economic and agricultural growth. The studies carried out showed the presence of characteristics of influence of the mentioned factors for different countries and economies, which was the reason for the research task of this work.

The intercountry analysis, implemented by developing multiple regression models, estimates carbon dioxide emissions in different countries.  There is a big difference in how much the change in the underlying factors affects the level of carbon emissions.  The values of the regression factors based on CO2 emission degrees are ranked for each country. [50]

•  The influence of environmental factors on the morbidity of respiratory organs of the population is evaluated using econometric modelling.  An analysis of the dynamics of the number of registered diseases with first diagnosis of respiratory disease for 1997-2022 is presented (per 100,000 people).  Multiple linear models of dependence of the number of patients with diseases of respiratory organs on environmental factors according to statistical data 1997-2022. The model results that increasing individual sulphur dioxide and nitrogen oxides in air by 1,000 tons would increase respiratory disease by 55.74 and 421.75, respectively; a 1% increase in domestic public health spending will reduce the number of cases per 100,000 population by 96.98. [51]

•  The regularities of formation and functioning of holistic social and economic-ecological (“econological”) systems at global level and methodological aspects of the issue of the sustainable development index (SDI) have been examined, in order to better reflect its essence.  The proposed modified index aims to provide a balanced approach to the SD Goal Index (SDGI), complementing the commonly accepted analytical framework of the system with large aggregated indicators of national programmes.  For particular groups of countries, such as those in conflict zones, national security is the indicator, because it dominates welfare and other metrics of the SDI. [52]

Based on the developed concept of sustainable development, a classical model of optimal control with corresponding phase coordinates, motion equations and control parameters is embedded in the common system.

Analysis of finance, banks, interest rates

•  The authors have studied more than 35 time series.  Different segments were considered: stock, debt and money markets.  In the proposed study, the problem of volatility heterogeneity is solved by using the synergy between GARCH model and the fuzzy inference system.  The volatility of financial instruments was modeled using a fuzzy inference system, in which the GARCH model is used.  The indices and instruments of three sectors of the Russian financial market were taken into consideration: the stock market, the bond market and the money market.  More than 35 tools were reviewed.[53]

•  The comparison of described approaches was carried out on the example of reproduction of the dynamics of the index of the Moscow Exchange (MOEX) for one quarter term (frequency of rebalancing) by different measures of similarity.  In the main part of the analysis data on quotes from 17.06.2014 to 17.12.2021 were used.  Estimates were made for the investment of 50 thousand rubles in three, ten or twenty shares according to each approach.  The authors concluded that the methods considered do not allow for an improvement of the quality of stock index repetition.  Therefore, the study of alternative approaches such as neural network use, reinforcement training and existing techniques should be continued.[54]

•  The important role of fiscal expenditures, including those directed to investment activities, where there is a strong need for turbulent periods in the economy, has been identified.  The total government expenditure and the budget expenditures for investment can be used as indicators of economic growth both in the Russian Federation as a whole and by region.  In addition, to monitor the sustainable economic development of territories, as well as to build forecast estimates of economic growth achieved through implementation of budget investments, as thresholds, Indicators of growth achievement/failure should be guided by the development targets set in strategic planning documents, on the basis of which it is necessary to develop a flexible system of indicators of economic development, taking into account the sectoral, natural-climatic and resource specificities of the regional composition of the territories of the Russian Federation.[55]

•  The financial technologies based on the concept of distributed registry, namely blockchain, are gaining in popularity and weight in the economy.  The objective of static liquidity provision on a decentralized exchange with concentrated liquidity is formulated, the mathematical model of position of liquidity provider is described and the optimal properties of the liquidity provision process are investigated.  Using the stochastic model of geometric Brownian motion for exchange rate process, the author has obtained a number of statements and theorems illustrating the optimal properties of the formulated problem.  The results obtained by the author can be applied to create derivative financial instruments on liquidity provided by the decentralized exchange provider, and also to create strategies for liquidity provision in real decentralized markets by institutional investors.[56]

•  The use of digital technology increases the accessibility of financial services to the population and has a number of advantages, including convenience and speed. However, the digital transformation process is accompanied by an increase in the risks associated with it.  The main assumption of the author is that increasing use of digital technologies in the financial market will determine the increase of risks associated with them, so the assessment of online-technology provides an insight into the increasing risks.  The following characteristics of respondents were used as factors influencing the probability of population activity in the financial market using digital technologies: gender; age; educational level; economic activity; type of settlement (urban and rural); federal district.  All the factors presented in the logistic regression have a significant influence on the use of the internet for banking transactions and the purchase of banking products and services.  The most active users of digital banking technologies are women with higher education, living in the city, employed, aged 20-24 years, living in the Southern Federal District (for banking) and the Volga Federal District (for the purchase of banking products and services).[57]

•  The impact of «surprise» component of monetary policy shocks (MPS) of the Federal Reserve System (FRS) on the ability of fund managers to generate excess yield was analyzed.  The work analyzed the monetary policy of the Fed at the current stage, revealed the «surprise component» of the MPS for the period from 2007 to 2022. The high-frequency identification procedure was used, and its impact on the profitability of US mutual funds in the period under review was verified.  Funds and their characteristics were selected from the Bloomberg terminal, based on the obtained panel data models with surplus fund yield as a dependent variable were built.[58]

The main hypothesis on the importance of «surprise» components of the MPS in the evaluation of the effectiveness of activities of mutually actively managed funds of the USA has been confirmed in the periods 2007-2009 and 2020, when the USA was in recession. The result allows to consider the «surprise component» as a macro-factor in further analysis of the activity of investment funds. The authors also carried out a check of the significance of bond yield spreads, which are an indicator of economic activity of financial market participants.

•  Using econometric methods, the factors influencing both the probability of bank failure and some continuous indicators of its activity - risk (Z-score and volatility of asset profitability) and profitability (ROA and ROE) were evaluated.  The bank’s probability of default was estimated by means of logistic regression, in which the explanatory variables were at a 1-4 block lag.  The risk and profitability assessment were performed on panel data using fixed effect (FE) models, which considered time effects in addition to individual effects.[59]

The results showed that the balance sheets are significantly correlated with both the probability of a bank default and its risk of insolvency and profitability. There is a diminishing return on scale in banks: the larger the bank, the lower its default probability, risk of insolvency, and profitability. It is estimated that high levels of liquidity compared to the banking sector put more pressure on the risk of bank default, increase the probability of default, but do not correlate with profitability.

•  The methodology of construction of metrics of client experience and integrated index of satisfaction of users in implementation of automated system of operational risk management (AS-ORM) in bank intended for collection was considered analysis, evaluation and monitoring of data on operational risk incidents, reporting, development and control of risk mitigation activities.[60]

The presented approach to building a system of metrics of user satisfaction and construction of an integral indicator allows not only to form adequate evaluations of feedback from users of systems, but also to form a register of improvements, optimal for improving the system, which includes analysis of all comments received during the CSI survey; highlighting key issues based on comments received; development of an action plan to eliminate the highlighted key comments.

Bilder: depositphotos

Die Meinung des Autors/Ansprechpartners kann von der Meinung der Redaktion abweichen. Grundgesetz Artikel 5 Absatz 1 und 3 (1) „Jeder hat das Recht, seine Meinung in Wort, Schrift und Bild frei zu äußern und zu verbreiten und sich aus allgemein zugänglichen Quellen ungehindert zu unterrichten. Die Pressefreiheit und die Freiheit der Berichterstattung durch Rundfunk und Film werden gewährleistet. Eine Zensur findet nicht statt.“

[1] Bakhtizin A.R. Development of Methods of Computer Modeling of Socio-Economic Processes // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[2] Kharin Yu.S., Shibalko S.A. Statistical Analysis of Multidimensional Binary Time Series of Processes // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[3] Kovalenko A.P., Perminov A.I. Construction of high-density clusters by a fully connected neural network with a piecewise linear activation function // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[4] Eliseeva I.I., Dekina M.P. On the issue of instrumental variables and models // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[5] Movsisyan Yu. M. Stochastic mappings form an algebra with hyperidentities моделях // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[6] Ohanyan V.K. Reconstruction of Convex bodies by probabilistic methods // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[7] Varshavsky A.E. Model based on finite functional sequence and its application for the study of the problem of inequality // Multidimensional statistical analysis, econometrics and simulation of real processes: abstracts of reports of the XI International Conference. Part 1: Plenary reports. Section 1. Multivariate Statistical Analysis and Econometrics, edited by V.L. Makarov – Moscow: CEMI RAS, 2024. Pp. 16-19. DOI: 10.33276/978-5-8211-0827-2

[8] Kharin A.Yu., Pashuk P.A. Sequential statistical testing of hypotheses for multidimensional observations with a block structure. Section 1.

[9] Golyandina N.E. Possibilities of Automation of the Singular Spectrum Analysis Method as a Principal Component Analysis Method for Time Series. Section 1.

[10] Balash V.A. Application of Multidimensional Statistical Analysis Methods for Modeling the Perception of the USSR Image in Russian and Foreign Mass Media. Section 1.

[11] Nikolsky I.M., Furmanov K.K. Is It Possible to Recognize Inefficient Enterprises? On the ranking ability of the model of the stochastic boundary. Section 1. 

[12] Voloshko V.A., Kharin Y.S. On an Approach to Probabilistic-Statistical Analysis of Discrete Time Series Based on Sufficient Statistics. Section 1.

[13] Ivanov M.A., Roshchina Ya.A., Korolev V.Y. Retraining in finite mixtures of normal distributions. Section 1.

[14] Afanasiev M.Yu., Gusev A.A., Nanavyan A.M. Estimates of Professional Groups and Structures of Professional Employment Based on the Concept of Economic Complexity. Section 1.

[15] Kudrov A.V. Optimization of Production Factors: Strategies of Regional Economic Growth. Section 1.

[16] Gusev A.A. Comparative Analysis of Approaches to the Assessment of the Economic Complexity of the Regions of Russia. Section 1.

[17] Matevosova A.M. Assessment of the Impact of the Public Debt Level on the Growth Rate of Regional Output in the Context of Clustering of Russian Regions by the Level of Socio-Economic Development. Section 1.

[18] Urazbaeva A.R. Modeling the Reaction of Entrepreneurship to the Severity of Sanctions Using the Dose-Effect Curve for Russian Regions. Section 1.

[19] Akopov A.S. Optimization of behavior strategies in the agent model of trade interactions using genetic optimization algorithms and clustering methods. Section 2.

[20] Beklaryan L.A., Akopov A.S. Improving traffic in Manhattan road networks using genetic optimization algorithms and clustering methods. Section 2.

[21] Khachatryan N.K. Synchronization of input and output flows in the model of organization of railway freight transportation. Section 2.

[22] Dubinina M.G. Spatial and temporal models of diffusion of broadband Internet access technologies. Section 2.

[23] Dubinina V.V. The main factors of the growth of robotization density in the countries of Central and Eastern Europe. Section 2.

[24] Nanavyan A.M. Highly qualified specialists and unskilled workers: an analysis of the qualification composition of the employed population in the regions of Russia. Section 1.

[25] Ryabukhina D.S. Analysis of the Heterogeneity of Unemployment Sensitivity to GRP in the Regions of Russia Based on a Hidden Class Model. Section 1.

[26] Bronitsky G.T., Vakulenko E.S. Sciencecasting migration using Google Trends. Section 1.

[27] Turko V.A., Arzhaev F.I. Modeling of production output and employment using the Cobb-Douglas production function: analysis in modern realities. Section 2.

[28] Kochetkova E.V. Analysis of Changes in Demand for Engineering and Technical Specialists and Skilled Workers Considering Scenarios of Macroeconomic Development. Section 2.

[29] Kuznetsova P.O., Kalabikhina I.E., Zhuravleva S.A. Meta-analysis of the motherhood penalty in the labor market. Section 2.

[30] Zalmanov I.A. Application of Cluster Analysis for Multidimensional Grouping of Cities by Socio-Economic Factors. Section 2.

[31] Mkhitaryan S.V. Analysis and modeling of perceived demand elasticity for public transport services in a metropolis. Section 1.

[32] Trofimova N.A. Analysis of social capital in the regions of Russia considering spatial interactions. Section 1.

[33] Osavolyuk A.A. Application of instrumental variables for modeling the relationship between peer health. Section 1.

[34] Bogdanova T.K., Zhukova L.V. Modeling of teachers' wages under conditions of heterogeneity of the socio-economic status of regions. Section 1.

[35] Demidova O.A., Dem'yanenko A.V., Gromov D.V. The Impact of Government Spending on the Life Satisfaction of Russians. Section 2.

[36] Pilyuk A.S., Semerikova E.V. Spatial Regression Analysis of Housing Accessibility in Germany. Section 2.

[37] Vojtenkov V.A. The Impact of the Atypical Crisis of 2022 on the Structure of Household Consumption: An Analysis of Russian Regions. Section 2.

[38] Salmina A.A. Features of the Subjective Well-Being of Russians and Factors of Their Differentiation. Section 2.

[39] Kartseva M.A., Kuznetsova P.O., Serechkina E.A. Child Benefits as a Tool for Reducing Child Poverty in Russia: Microsimulation Analysis. Section 2.

[40] Makarenkova A.O. Assessment of the Impact of Crisis Time on Actual Reproductive Behavior: Event Occurrence Analysis. Section 1.

[41] Nikonova M.A. Features of the Demographic Situation in the Regions of the Far Eastern Federal District. Section 2.

[42] Vakulenko E.S., Gorskiy D.I., Kondratieva V.P., Trofimenko I.A. Objective vs. subjective factors of reproductive intentions of Russians in 2022-2023. Section 2.

[43] Rodionova L.A., Kopnova E.D. Why do people in Russia prefer summer births for children? Section 2.

[44] Kopnova E.D., Rodionova L.A., Ivanova M.A. Statistical analysis of marital status factors in Russia considering age and gender. Section 2.

[45] Kuznetsova M.S. Assessment of economic damage from premature death of individuals due to diseases of the digestive system. Section 2.

[46] Komkina T.A. Analysis of dynamics of development of manufacturing industries to assess risks of achievement of technological sovereignty of the Russian Federation. Section 2.

[47] Borisova S.V.. Evaluation of production functions of branches of the Russian economy through the reduction of production factors. Section 2.

[48]Malyugin V.I. Short-term forecasting of macroeconomic indicators using survey data under conditions of shock effects.  Section 1.

[49] Roslyakova N.A., Volkov A.D. Tools for evaluating the effectiveness of preferential treatment for enterprises in the Russian Arctic. Section 1.

[50] Gazaryan Z.M., Ordyan N. P. Cross country analysis on carbon dioxide emissions. Section 1

[51] Buniatian M. M., Hukasyan G.M Econometric modelling of the effects of emissions into air on respiratory diseases (according to data from the Republic of Armenia). Section 2.

[52]Sargsyan G.L., Gevorgyan R.A., Kochinyan N. S., Sargsyan S.G., New approach to the assessment of the Sustainable Development Index.  Section 2.

[53] Swedes A.S., Sviazov V.A. The unclear model ARMA-GARCH-TS and its application to financial time series. Section 1.

[54] Gavrish B.G., Tregumova G.I.. Approaches to the replication of stock exchange indices. Section 1.

[55] Kovalenko N.N. Assessment of the impact of fiscal investments on economic growth. Section 1.

[56] Savin A.L. The objective of optimal liquidity provision in decentralized finance markets with concentrated liquidity. Section 1.

[57] Tregumova A.A., Sinyavskaya T.G. Determinants of the use of digital technologies by the population in the financial market. Section 1.

[58]Artamonov N.V. Kurbatsky A.N. Surprise component in the analysis of excess return of mutual funds of the USA. Section 2.

[59] Bekirov O.A. Analysis of the factors of stability of Russian banks. Section 2.

[60]Voronovskaya O.E. Metrics of client experience and construction of the integrated index of user satisfaction in the implementation of automated systems for operational risk management.  Section 2.