Synthesis of Socio-Economic Maps and Visualization of Deviant Activity Measures of Financial Monitoring of Entities

In the field of financial monitoring, analysts have to deal with a considerable amount of information. Statistics provides more details on these volumes.

According to the provisions of the Federal Law No. 115-FZ, the Federal Financial Monitoring Service receives information of about 100,000 financial transactions daily. It contains details of payers and recipients of funds, their accounts, and credit organizations.

As of January 2020, 3.7 million legal entities were registered in the Unified State Register of Legal EntitiesThe Unified State Register of Legal Entities. URL: https:// www.nalog.ru/rn77/related_activities/statistics_and_analyt- ics/forms/8376083 (accessed on 24.01.2020).. 286 organizations have a license to carry out brokerage activities, 314 have dealer licensesThe register of brokers. Electronic resource. URL: https:// www.cbr.ru/registries/ (accessed on 24.01.2020)..

According to official information of the Central Bank of Russia, as of 01.01.2020, 442 credit organizations and their 618 branches carry out activities in the Russian FederationDirectory of credit organizations. URL: https://www.cbr.ru/. Bank statements contain hundreds of data fields.

In addition, the federal database of Rosfinmonitor- ing maintains its own accounting information for each credit institution with up to 50 data fields.

Also, the Federal Financial Monitoring Service compiles statistics for each type of object, and additional identification information, such as addresses, data of identity documents, etc.

The departmental information of Rosfinmonitor- ing is enriched with data from various state registers, data on the foreign economic activity of entities, and tax information [6].

The visualization of assessments of identified entities solves the problem of primary identification of financial transactions.

Currently, methods of socio-economic mapping are widely used to visualize and analyze data in various fields.

The definition of mapping is given in GOST R 52438-2005banking_sector/credit (accessed on 24.01.2020). as a method of modeling, graphic, digital display of geoinformation, geospatial information "with an indication of its identifier, coordinate and attribute data". Another, broader interpretation of the term "mapping" is found in spatial studies. Mapping is defined as a method of modeling, visualization, graphic representation of any spatially localized information in accordance with the specified parameters in order to cognize the depicted phenomena [7-10].

Mapping is widely used in sociology, economics, philosophy, history, psychology to study pressing problems of territorial development and systematization, the visual presentation of information about the objects of study.

Socio-economic mapping as a scientific method dates back to the 1889 study of Charles Booth, who charted London's poverty8. This was followed by the work of researchers of the Chicago school E. Burgess and R. Park [11] on the territorial zoning of cities in relation to migration, urbanization, inequality, crime. The compilation of social maps was based on factual information collected through interviews, statistical observation, document analysis, which was then generalized and systematized.

The works of U. Teichler, I. Ferencz, B. Wдchter serve as an example of socio-economic mapping put into practice. The works present a map of international academic mobility in Europe [12]. E. S. Kuzmina [13] provides maps of the exports and imports of the education services with quantitative indicators of the social process under study.

Another example of the social mapping method put into practice is found in the analysis of urban space and presented in the works of O. I. Vendina, N. N. Veselkova, K. P. Glazkova, N. D. Vavilina, I. A. Scalaban, K. Lynch, S. Milgram [14].

The study [15] focuses on the socio-territorial structure of Moscow with statistical data put on the city map, and the authors used a socio-statistical approach.

Thus, an analysis of the sources showed that socioeconomic maps generally have one or two characteristics, quantitative one-dimensional data expressed in physical units - quantity, amount, etc. For example, the number of foreign students, or data expressed in a ratio - the ratio of resident to non-resident students.

At the same time, various industries, financial monitoring in particular, have tasks that require to display objects defined by a large number of characteristics. In other words, vector objects, which have no defined order relations.

In practice, there are situations in financial monitoring when ordinal relations between regions and constituent entities are important, and specific values of rating estimates are not.

To establish ordinal relations between regions, it is necessary to rank them. However, the objects of research are of vector nature and have sets of characteristics. As known, there is no mathematical definition of ordinal relations for vectors.

Thus, it is necessary to find a scalar function of the vector argument to solve the problem of data mapping in the field of financial monitoring.

Usually, the problem of finding a scalar function of a vector argument in various fields of human activity (politics, economics, sociology, sport, etc.) is solved by experts assigning weighting factors to vector elements. In other words, a weight vector in which the scalar product is the initial vector of the object's characteristics is selected, and the desired scalar estimate is generated. However, a subjective or politically driven judgment may impact this approach.

Method of principal components

A comparative analysis of methods for reducing the dimension, convolution, and scalarization of vector indicators makes it possible to classify the method of principal components as promising [16-20].

The principal components method enabled us to synthesize integral estimates of objects of financial monitoring, regions, and federal districts. The obtained scalar values made it possible to rank them and served as a basis for visualization and mapping.

The most interesting indicators demonstrate maximum variability in passing from one object to another when dealing with practical and data processing issues.

At the same time, it is not obligatory to use characteristics directly measured for a particular object. Two artificial parameters can serve as an example - size and height when choosing clothes, instead of a tailor taking specific body measurements of a person. Although some of the information gets lost and coarsened, daily practice shows that such an approach gives quite acceptable results.

These prerequisites are fundamental in finding a linear transformation of the original set of indicators, which allows us to obtain principal components.

The method of principal components analysis is based on a linear model. Taking the number of analyzed objects for N, the number of parameters of objects for n, the mathematical model will be written as:

where r = 1, 2,..., n; j = 1, 2,..., n; f-- r-th main component; ajr - weight of the r-th component in the j-th variable; yj - normalized value of the j-th attribute, known from observations or obtained as a result of the experiment.

Finding an integrm indicator of the deviant activity of business entities for visualization and mapping

The Supreme Arbitration Court of the Russian Federation analyzes the financial and economic activities of a liquidated legal entity during bankruptcy procedure, with an absent debtor. The results of such analyses are given in court decisions.

Shell companies are usually established for certain operations. When a shell company fulfills its assigned role, it is simply "abandoned" without any liquidation procedures to the extent required by the applicable law. Thus, shell companies often fall into the category of absent debtors, and court decisions on their forced liquidation have indicators of shell companies.

The selection, analysis, and systematization of the decisions of the Arbitration Court of the Russian Federation made it possible to form a sample of legal entities that have features of shell companies.

The value of this sample is that it includes organizations classified as shell companies, based not on intuitive guesses of experts, but on confirmed evidence supported by court decisions.

According to the obtained sample, the variances of the main components and the correlation coefficients of the characteristics of economic entities with internal factors (principal components) are calculated on the basis of this matrix.

The interpretation of the projections of the source variables on the principal components (Table 1) proves that the second principal component is adequate from the point of view of assessing the situation in the industry.

Ten principal components (Fig. 1) describe the overall data variation.

Consider the second principal component. Part of the indicators correlates positively with it, another part - negatively. Wherein, features have a negative connotation.

Table 1 and Fig. 2 show that the most important indicators that positively correlate with the second principal component are "lack of staff", "organization activity period" and "absence of non-current assets".

These are inherent characteristics of shell companies established to conduct suspicious financial transactions and cover illegal activities.

The second internal factor is negatively related to the indicators such as "lack of movement of funds in accounts", "absence of settlement accounts" and "absence at address". These are typical signs of organizations being on the verge of bankruptcy, facing financial difficulties under adverse circumstances.

The higher the value of the second principal component, the higher the money laundering risks of the business entity.

Integral indicator of the deviant activity of professional securities market participants used for visualization

The securities market plays an important role in a state's financial system and has its specifics. It is an attractive platform for money laundering, due to favorable trading conditions (for example, electronic bidding) and simplified process of international transactions. The securities market is a place to generate legitimate profit from money laundering.

Based on the federal financial monitoring database for professional securities market participants, the principal components variances and the characteristics' correlation coefficients of business entities with internal factors (main components) have been calculated.

The projections of the source variables on the principal components (Table 2) describe that the tenth principal component is adequate for the assessment process in the industry.

Table 1

Source: compiled by the author.

Projections of the source variables on the new subspace

Table 2

Source: compiled by the author.

Ten principal components describe the overall data variation (Fig. 3). Table 2 presents calculations of factor loads - projections of the initial variables on each of the principal components.

The tenth internal factor is the most interesting. This factor corresponds to the deviant component, i.e. the securities market is vulnerable to the activities of money laundering.

The obtained results reflected the subject area under consideration, and the securities market experts confirmed it.

Fig. 3. The total contribution of principal components to the total variance

Source: compiled by the author.

Fig. 4. Financial transactions diagram

Source: compiled by the author.

Thus, it is the first scientific solution to the problem of visualization of the measure of the deviant activity of entities: business entities, credit organizations, and professional securities market participants. The article explores and solves the problem of synthesizing measures of the deviant activity of entities' financial monitoring.

A socio-economic map, based on the deviant activity measures of business entities, has been created. It reflects the money laundering situation in Russian regions (Fig. 5).

These are important results for strategic decisionmaking, with the main focus on problematic regions.

We study if this approach applies to other official data analysis.

Russian legal entities data. Business activity mapping

Fig. 5. Mapping the propensity to legalize cash

Source: compiled by the author.

The Unified State Register of Legal Entities (hereinafter referred to as USRLE) provides information about Russian legal entities. The USRLE is a part of the Federal Tax Service of Russia, which contains information about registration, reorganization, and liquidation of Russian companies.

We study the geographical component of this information by considering the legal entities' registration statistics.

We apply the method of principal components to the initial data.

We calculate the variances of the main components and the correlation coefficients of the indicators with internal factors (main components) based on this matrix. Six main components describe the overall variation of the data. The first two main components contribute 75% of the total variance. Figure 6 presents a graphical illustration of the contribution of the internal factors to the total variance.

Consider the first major component. There is a negative correlation between the signs of "legal entities that have ceased their activity due to bankruptcy", "legal entities that have ceased their activities, total", "legal entities that ceased their activities by the registration authority decision" and "legal entities that ceased their activities as a result of liquidation".

In addition, a positive correlation is evident with the signs of "legal entities under reorganization", "legal entities that ceased their activities as a result of reorganization", "total operating legal entities" and "total legal entities created". On this basis, we can talk about the bipolarity of the first principal component.

Fig. 6. Total contribution of the main components to the total dispersion

Source: compiled by the author.

The tendency to register new legal entities and reorganize previously registered ones is a positive sign of the factor. The tendency to liquidate organizations for various reasons is negative. The first major component reflects the business activity of the region. Figure 7 shows the mapping of business activity in federal districts. A more intense colour represents higher values of the first principal component.

The inverse factor solution allowed to rank the federal districts according to the situation assessment and use the results for visualization purposes.Correlation coefficients of indicators and principal components

Table 3

Source: compiled by the author.

Fig. 7. Mapping business activity of federal districts

Source: compiled by the author.

Table 4

The value of the principal components for the federal districts

Source: compiled by the author.

Fig. 8. Total contribution of the principal component to the total variance

Source: compiled by the author.

Crime data mapping

1. Beketnova Yu.M., Prikazchikova G. S., Prikazchikova A. S. The modification of the T. Saaty's analytic hierarchy process in order to improve the risk management system of the Federal Customs Service. Vestnik Rossiiskoi tamozhennoi akademii = The Russian Customs Academy Messenger. 2016;(3):128-136. (In Russ.).

2. Beketnova Yu. M. Expert assessments of financial activity subjects. Diskussiya = Discussion. 2013;(8):52-54. (In Russ.).

3. Huang M. L., Liang J., Nguyen O. V. A visualization approach for frauds detection in financial market. In: 13th Int. conf. information visualisation (Barcelona, 15-17 July, 2009). New York: IEEE; 2009:197-202. DOI: 10.1109/IV.2009.23

4. Cardenas A. A., Manadhata P. K., Rajan S. P. Big data analytics for security. IEEE Security & Privacy. 2013;11(6):74-76. URL: https://mlsec.info/pdf/sp13.pdf

5. Khine M. S. Spatial cognition: Key to STEM success. In: Khine M. S., ed. Visual-spatial ability in STEM education. Cham: Springer-Verlag; 2017:3-8. DOI: 10.1007/978-3-319-44385-0_1

6. Beketnova Yu.M., Krylov G. O., Larionova S. L. Models and methods for solving analytical problems of financial monitoring. Moscow: Prometei; 2018. 274 p. (In Russ.).

7. Evteev O. A. Design and compilation of socio-economic maps. Moscow: MSU; 1999. 219 p. (In Russ.).

8. Volodchenko A. The e_glossary "Cartosemiotics". Dresden. 2009. URL: https://docplayer.ru/77039023-A-vo- lodchenko-kartosemiotika.html (In Russ.).

9. Vajjhala Sh.P. "Ground truthing" policy: Using participatory map-making to connect citizens and decision makers. 2006. URL: https://media.rff.org/archive/files/sharepoint/WorkImages/Download/RFF-Resources-162_ GroundTruthing.pdf

10. Sutcliffe A. G. A design framework for mapping social relationships. PsychNology Journal. 2008;6(3):225-246.

11. Davies G., Burgess J., Eames M., Mayer S., Staley K., Stirling A., Williamson S. Deliberative mapping: Appraising options for addressing `the kidney gap'. Final report. 2003. URL: https://www.researchgate.net/publica- tion/228530224_Deliberative_mapping_appraising_options_for_addressing_'the_kidney_gap'

12. Teichler U., Ferencz I., Wдchter B., eds. Mapping mobility in European higher education (in 2 vols.). Vol. 1: Overview and trends. Brussels: the European Union; 2011. 271 p. URL: http://www.acup.cat/sites/default/ files/teichlerferenczwaechtermappingmobilityineuropeanhighereducation_0.pdf

13. Kuzmina E. S. Social mapping used to study international academic mobility. Obshchestvo: sotsiologiya, psik- hologiya, pedagogika = Society: Sociology, Psychology, Pedagogics. 2018;(12):115--119. (In Russ.). DOI: 10.24158/ spp.2018.12.19

14. Skalaban I. A. Social mapping as a method for analyzing socio-territorial space. Zhurnal issledovanii sotsial'noi politiki = The Journal of Social Policy Studies. 2012;10(1):61-78. (In Russ.).

15. Goloukhova D. V. The research methodology of the socio-territorial structure of a Russian city (on the example of Moscow). Cand. sociol. sci. diss. Synopsis. Moscow: MGIMO University; 2017. 25 p. (In Russ.).

16. Geiger B., Kubin G. Relative information loss in the PCA. In: Proc. IEEE information theory workshop (Lausanne, 3-7 Sept. 2012). New York: IEEE; 2012:562-566. DOI: 10.1109/ITW.2012.6404738

17. Howard M. C. A review of exploratory factor analysis decisions and overview of current practices: What we are doing and how can we improve? International Journal of Human-Computer Interaction. 2016;32(1):51-62. DOI: 10.1080/10447318.2015.1087664

18. Amerioun A., Alidadi A., Zaboli R., Sepandi M. The data on exploratory factor analysis of factors influencing employees effectiveness for responding to crisis in Iran military hospitals. Data in Brief. 2018;19:1522-1529. DOI: 10.1016/j.dib.2018.05.117

19. Eskindarov M. A., Solov'ev V.I., eds. Paradigms of the digital economy: Artificial intelligence technologies in finance and fintech. Moscow: Kogito-Center; 2019. 325 p. URL: http://www.fa.ru/org/div/uoonir/Documents /%D 1%82%D 0%BE%D 0%BC_4_print2.pdf (In Russ.).

20. Krylov G. O., Beketnova Yu.M., Prikazchikova A. S. Factor analysis and clustering theory application in financial monitoring tasks. Vestnik Irkutskogo gosudarstvennogo tekhnicheskogo universiteta = Proceedings of Irkutsk State Technical University. 2016;20(10):102-110. (In Russ.). DOI: 10.21285/1814-3520-2016-10-102-110

Размещено на Allbest.ru