CC BY
62
Методы прогнозирования банкротства предприятия Methods for predicting bankruptcy of an enterprise
Яценко Александра Андреевна
Студент 2 курса Финансовый факультет Российский экономический университет им. Г.В. Плеханова
Москва, Россия
Yatsenko Alexandra Andreevna
Student 2 term Faculty of finance Plekhanov Russian University of Economics
Moscow, Russia
Аннотация.
Статья посвящена методам прогнозирования банкротства в строительной отрасли, рассмотрены как зарубежные, так и отечественные методы. Автор подробно рассматривает 3 logit-модели: шестиступенчатый logit - модель Д. Чессера, девятифакторный logit - модель Дж. Олсона и пятифакторный logit - модель Е. А. Федоровой, Я. В. Тимофеева. Подробное рассмотрение трех методов представлено автором на основе финансовой отчетности крупной строительной компании ООО «ПСО Казань». В результате исследования автором разработаны меры для компаний строительной отрасли по повышению эффективности применения методов оценки риска банкротства в российских условиях.
Annotation.
The article is devoted to the methods of forecasting bankruptcy in the construction industry, both foreign and domestic methods are considered. The author examines in detail 3 Logit - models: Six-factor logit - D. Chesser's model, Nine-factor logit - model of J. Olson and Five-factor logit - model of E. A. Fedorova, Ya. V. Timofeeva. A detailed consideration of the three methods is presented by the author on the basis of financial statements of a large construction company PSO Kazan LLC. As a result of the study, the author has developed measures for companies in the construction industry to improve the effectiveness of the application of methods for assessing the risk of bankruptcy in Russian conditions.
Ключевые слова: методы прогнозирования банкротства, logit - модели, модель Д. Чессера, модель Дж. Олсона, модель Э.А. Федорова, Я. В. Тимофеева.
Key words: methods of forecasting bankruptcy, Logit models, D. Chesser's model, J. Ohlson's model, model of E.A. Fedorova, Ya. V. Timofeeva.
Introduction
The problem of forecasting and assessing the risk of bankruptcy occupies a special place in the existing theories for assessing business efficiency and the financial condition of a company. According to the data of the unified federal register of bankruptcy information, the number of Russian companies declared bankrupt in 2019 decreased by 5.5% compared to 2018 - to 12,401 (see Fig. 1). The number of rehabilitation procedures (external management and financial recovery) imposed by courts decreased to 1.01% in 2019, compared to 1.26% in 2018. [9]
16000 14000 12000 10000 8000 6000 4000 2000
12923 13044
13541
12549
13117
12401
10762
10040
2012
2013 2014 2015 2016 2017 2018 2019 Figure 1. Dynamics of company bankruptcies in the period from 2012 to 2019
Table 1. Number of court decisions from 2015 to 2019.
Number of court decisions 2015 2016 2017 2018 2019
on declaring the debtor bankrupt and starting bankruptcy proceedings 13044 12549 13541 13117 12401
on the introduction of surveillance 10198 10487 11495 10547 10134
on the introduction of external management 434 372 363 278 209
on the introduction of financial recovery 38 52 32 19 19
to terminate the proceedings 1943 2342 2495 2802 3817
share of rehabilitation procedures,% 2,03% 1,84% 1,58% 1,26% 1,01%
Based on the above data, we can conclude that in Russia the problem of forecasting bankruptcy of companies is still acute. On average, only 3.9% of companies filing for bankruptcy experience minor losses (through external management, financial recovery or settlement). The majority is liquidated, which leads to losses (not only material ones) for a wide range of legal entities and individuals. Currently, this issue is more relevant than ever for Russian companies in view of the deterioration of economic and political ties between countries, which cause the instability of the processes taking place in the market and directly affecting their activities.
Currently, many studies in the field of bankruptcy risk assessment are aimed at finding and developing alternative approaches and methods that include not only an analysis of the company's financial condition, but also qualitative detection and forecasting of bankruptcy. These methods include binary choice models in which the dependent variable takes on two different values. In practice, they are used to study the influence of certain factors on the presence or absence of some sign, in our case, the onset of bankruptcy. Logit models are of particular interest, since they have a fairly accurate predictive power.
Logit models.
Logit - models are a statistical predictive model for assessing the likelihood of bankruptcy. As with other models, two samples of companies are used for development: a group of companies that are bankrupt by an arbitration court and a group of financially stable companies. Various financial ratios are calculated for them, and then a regression model is constructed that most accurately describes these two samples. The main difference between the logit model and others is that the probability distribution function is described by a logistic curve.
0
In addition, the logit analysis model makes it possible to build nonlinear dependence models, which can be considered a significant advantage. Thus, a report by scientists from the Turku Center for Computer Technology in Finland showed that often, especially for bankrupt companies, the condition of subordination of discriminant variables to the normal distribution law is not met, and the use of logit models removes this restriction. [1]
The probability of bankruptcy in each logit - model is calculated by the following general formula of the logistic function (formula 1):
i
where:
P is the probability of bankruptcy (takes a value from 0 to 1);
e is the base of the natural logarithm;
Y is an integral indicator calculated depending on the developed model.
Next, let's take a closer look at the classic logit models. Despite the fact that J. Olson (1980) is considered the founder of logit analysis, who was the first to use the logistic regression apparatus, D. Chesser (1974) was the first to use this formula, who developed a special model for the banking sector for assessing the probability of a borrower's default on the terms of a loan agreement. [3] So, in his model, the calculation of a certain integral indicator Y is presented on the basis of 6 weighted variables, which are financial ratios that characterize the efficiency of the company and its financial stability. The formula for calculating the model is as follows (formula 2):
Y = -2,0434 - 5,24x1 + 0,0053x2 - 6,6507x3 + 4,4009x4 - 0,0791x5 - 0,1220x6 (2)
where:
X1 = (Cash + Liquid Securities) / (Total Assets);
X2 = (Net Sales) / (Cash + Liquid Securities);
X3 = (Gross - income) / (Total assets);
X4 = (Total debt) / (Total assets);
X5 = (Fixed Capital) / (Net Assets);
X6 = (Working Capital) / (Net Sales).
If the final value of the probability is less than 0.5, then the probability of bankruptcy is considered low, and if it is more than 0.5, then, accordingly, high.
Another model used in this bankruptcy forecasting study is the model described by J. Olson (1980). Based on a sample of 105 bankrupt and 2,058 financially healthy public industrial companies between 1970 and 1976, the scientist calculated the probability of bankruptcy, selecting the most significant coefficients for assessing the financial stability of companies. [7] The probability of bankruptcy is determined by the formula (formula 3):
y = -1,32 - 0,407 * X1 - 0,603 * X2 * -1,43 * X3 + 0,076 * X4 - 1,72 * X5 - 2,37 * X6 - 1,83 * X7 + 0,285 *X8 - 0,521 *X9 (3)
where:
X1 = log (Total assets) / GDP deflator index;
X2 = Total Liabilities / Total Assets;
X3 = Working Capital / Total Assets;
X4 = Current liabilities / Current assets;
X5 = If total liabilities > total assets, the value is 1; if total liabilities < total assets, the value is 0;
X6 = Net profit (NP) / Total Asset;
X7 = Revenue / Total Liabilities;
X8 = If the NP has been negative for the last 2 years, the value is 1; if on the contrary, the value is 0;
= NPt - NPt-i 9 |NPt| - |NPt-i|-
The main advantage of J. Olson's model, as in all other models based on logit-analysis, is in the unambiguous interpretation: the resulting indicator can take values from 0 to 1, that is, from a low probability of bankruptcy to a high one. However, when constructing the model, the scientist used data for a 7-year period, which is a significant drawback, since macroeconomic factors are not taken into account. [11]
Based on the idea of adapting logit - regression to predict the likelihood of bankruptcy in Russia, E.A. Fedorova and Ya.V. Timofeev in 2015 developed a model designed to assess the viability of Russian companies in the construction industry. The formula for the probability of bankruptcy for construction companies is presented as follows (formula 4): Z = -1,75 - 0,28X1 - 2,33X2 - 15X3 + 1,38X4 - 0,34X12 (4)
where:
X1 = Quick ratio;
X2 = Cost Effectiveness;
X3 = Return on assets;
X4 = Short-term debt / Total liabilities;
X5 = Total Capital / Total Liabilities.
Thus, we can conclude that the approach based on the use of logit - regression is universal, since it does not have strict restrictions, and therefore it has a wider field of application. Some modern studies prove its effectiveness: in practice, nonlinear models allow you to get significantly more effective assessment results than models related to discriminant analysis can provide. [6]
Diagnostics of the bankruptcy probability of OOO PCA "Kazan" based on Logit models
For the purpose of assessing the risk of bankruptcy, the author chose the company OOO "Production and Construction Association "Kazan" (hereinafter referred to as PCA "Kazan") - this is a large construction company of the Republic of Tatarstan, the main activities of which are the construction of industrial and civil facilities, design of buildings and structures, reconstruction of historical and architectural monuments.
The calculation of indicators was made on the basis of the "Balance sheet as of December 31, 2019." and "Statement of financial results for 2015-2019".
The above 3 bankruptcy forecasting models will be used for the calculation:
1. Six-factor logit - D. Chesser's model;
2. Nine-factor logit - J. Ohlson's model;
3. Five-factor logit - model of EA Fedorova, Ya. V. Timofeeva.
1. D. Chesser's model.
Table 2. The results of calculating the bankruptcy risk assessment of the Kazan PSO according to the six-factor model
of D. Chesser for 2017 - 2019.
Coefficient Calculation algorithm 2017 2018 2019
X1 (Cash + Fast Selling Securities) / Assets 0,1028 0,0660 0,0690
X2 Net Sales (Revenue) / (Cash + Fast Selling Securities) 3,4394 2,5815 1,6264
X3 Income (gross, working capital) / Assets 0,0018 -0,0372 -0,0401
X4 Total debt / Assets 0,9895 1,0305 1,0321
X5 Fixed Capital / Net Assets 1 1 1
X6 Working capital / Net - sales (revenue) 0,0052 -0,2186 -0,3579
Y 1,6992 2,3505 2,3706
Z 0,8454 0,9130 0,9146
If Z > 0,5 - bankruptcy High High High
Table 3. The results of calculating the bankruptcy risk assessment of the PCA Kazan according to the nine-factor model
of J. Ohlson 2017-2019
Coefficient Calculation algorithm 2017 2018 2019
X1 Ln (Assets / deflator-GDP index) 17,52949 17,72136 17,73099
X2 Total debt / Assets 0,98950 1,03051 1,03215
X3 Working Capital / Assets 0,00184 -0,03724 -0,04014
X4 Short-term liabilities / Current assets 0,99814 1,03749 1,04046
X5 If total liabilities > total assets, 1 If total liabilities < total assets, 0 - - -
X6 Net Income / Total Assets 0,00003 0,00000 0,00000
X7 Revenue / Total Liabilities 0,35719 0,16531 0,10867
X8 If the NP has been negative for the last two years, 1 If NP is positive, 0 - - -
X9 NPt - NPt-1 0,00305 0,04634 0,01404
|NPt| - |NPt_1|
Y -9,63325 -9,34856 -9,22860
P 0,00007 0,00009 0,00010
If P > 0.5 - bankruptcy Low Low Low
According to the results calculated by J. Ohlson's model, the risk of bankruptcy for the company is unlikely. This contradictory meaning may be due to some peculiarities: the model is foreign, and it is also based on a sample of industrial companies. That is, the factors that reflect the specifics of companies in the industrial sector were taken into account. Since in the analyzed period the net profit indicator was always positive, the value equal to 0 is taken as the X8 coefficient, which has the greatest specific weight in this model.
Table 4. The results of calculating the bankruptcy risk assessment of the Kazan PSO according to the five-factor model _of E. A. Fedorova, Ya. V. Timofeeva for 2017 - 2019.
Coefficient Calculation algorithm 2017 2018 2019
X1 Quick ratio 0,86059 0,77618 0,74640
X2 Return on costs 0,00025 0,00023 0,00035
X3 Return on assets 0,00003 0,00000 0,00000
X4 Short-term debt / Total liabilities 1 1 1
X5 Equity / Total Liabilities 0,01061 -0,02961 -0,03115
Y -0,61559 -0,57782 -0,56922
P 0,35079 0,35943 0,36142
If P > 0 - bankruptcy High High High
As can be seen from the table, the estimates obtained are very contradictory. According to the results of the study, the models of D. Chesser and E.A. Fedorova and Ya.V. Timofeev show a high probability of bankruptcy risk. At the same time, the results of calculations based on the J. Olson model show the opposite.
In addition, it can be established that the model of E. A. Fedorova and Ya. V. Timofeev gives the most accurate result for the object of research, since it includes indicators of quick liquidity and financial independence (autonomy) in the calculation. It is fair to say that this model takes into account the industry specifics, as it was developed on a sample of construction companies. This confirms the logic of this study. The result according to the model under consideration indicates the risk of bankruptcy of the PCA "Kazan" company.
The need to assess the effectiveness of the application of methods for predicting bankruptcy risk is due to a wide variety of models. The study of the use of foreign models in Russian practice (see paragraph 3.1.) Showed that the use of
a number of models for assessing the probability of bankruptcy is inappropriate in Russian conditions. The following reasons. First, the use of various models leads to contradictory results, since they do not take into account the specifics of the economic situation and factors of the business environment in Russia. The analysis shows that 2 of the models under consideration show a high probability of bankruptcy: these are the models of D. Chesser, E. A. Fedorova and Ya.V. Timofeeva. As of 2019, J. Ohlson's forecasting model shows a fairly high score. This can be explained by the fact that the value of the largest ratio in terms of specific weight X4, which is characterized as the ratio of total debt to total assets, is almost 1. However, the model was developed specifically for the banking sector in order to assess the likelihood of a borrower's default on the terms of the loan agreement, therefore, the results calculated for the research object are not indicative.
The most accurate result, taking into account the branch affiliation of the research object, is provided by the model of E. A. Fedorova and Ya. V. Timofeeva. According to a study by the authors of the model, its overall predictive power for the construction industry is 81.33%. This confirms the logic of this study. The result according to the model under consideration indicates the risk of bankruptcy of the PCA "Kazan" company.
Based on the studies studied, it can be argued that the industry affiliation of a company has a significant impact on its financial performance, and therefore on the assessment of the likelihood of bankruptcy, obtained using the considered models.
Existing models include a limited range of indicators that determine liquidity, solvency, profitability and, as a rule, are extended or modified Western models. One more conclusion follows from this: to assess the bankruptcy of a construction company, one should take into account certain indicators characterizing the specifics of the industry.
Today the legislation offers an approved system of 10 different indicators required to assess the financial condition of a company. [8] Nevertheless, the question arises about the effectiveness of the standards of these indicators in relation to the construction industry. In the article by E.A. Fedorova, M.A. Chukhlantseva. and Chekrizova D.V. the main financial ratios for the construction industry are highlighted: liquidity ratios, financial dependence, provision with own circulating assets (hereinafter - OCA), maneuverability of OCA. [5] Using IBM SPSS Statistics software and box and mustache methods, the authors of the article refined the guideline values for various industries, including the construction industry:
1. Current liquidity ratio > 1;
2. Coefficient of supply SOS > 0;
3. Ratio of financial independence > 0.012;
4. Absolute liquidity ratio > 0.06;
5. Quick ratio > 0.7;
6. Coefficient of maneuverability OCA < 0.9.
The results of the author's analysis of the use of foreign methods for assessing bankruptcy risk in Russian practice allow us to draw the following conclusions:
1. Inconsistency of the initial data used to build a particular model;
2. Foreign models take into account the specifics of the countries for which they were developed (the macroeconomic situation in the country, features of its tax system, information and legal framework, etc.).
3. When adapting the considered models into Russian practice, it is necessary to remember about their sectoral characteristics.
Based on the foregoing, the author has developed measures for companies in the construction industry to improve the efficiency of methods for assessing the risk of bankruptcy in the Russian context (Fig. 2).
Figure 2. Proposals to improve the efficiency of bankruptcy risk assessment methods in Russian conditions
conclusions
Conclusions
Summarizing all of the above, we can draw some of the most significant conclusions. Bankruptcy is the most unfavorable consequence of the impact on the company's activities of a whole set of entrepreneurial risks. In Russia, the problem of predicting bankruptcy of companies is still acute. Only a small percentage of companies filing for bankruptcy experience minor losses. The majority is liquidated, which leads to losses for a wide range of legal entities and individuals.
In order to analyze the accuracy of calculating the probability of bankruptcy according to the considered models, modern research in the field of practical application of various forecasting methods in Russian conditions was considered. Based on the work of modern researchers, many methods are overly generalized, not taking into account the industry and regional characteristics of companies' activities, which can lead to incorrect forecasts and irrational actions of external and internal stakeholders of the company. In particular, it was revealed that foreign models (J. Olson and D. Chesser) are not advisable to use in assessing the likelihood of bankruptcy of a Russian company. In order to clarify the results obtained, a point method for assessing the financial condition was calculated, with the help of which it was possible to identify the most suitable models for the object of research: the five-factor logit - the model of E. A. Fedorova and Ya. V. Timofeev. Moreover, the author proposed possible ways to improve the efficiency of the use of insolvency forecasting methods in the Russian context.
Список используемой литературы:
1. Back В., Laitinen Т., Sere К., van Wezel M. Choosing Bankruptcy Predictors Using Discriminant Analysis, Logit Analysis, and Genetic Algorithms // Turku Centre of Computer Science, September, 1996. Technical Report No 40. P. 20.
2. Bogdanova T. K., Alekseeva Yu. A. Predicting the probability of bankruptcy of enterprises taking into account changes in financial indicators in dynamics // Business Informatics 2011. No 1 (15) P. 50-60.
3. Chesser D.L. Predicting Loan Noncompliance // The Journal of Commercial Bank Lending, 56(12), 1974.
4. Federal Law of 26.10.2002 N 127 - FL (as amended on 27.12.2018) "On insolvency (bankruptcy)" (last
edition).
5. Fedorova E.A., Chukhlantseva M.A., Chekrizov D.V. Regulatory values of the coefficients of financial stability: features of types of economic activity // Administrative sciences. 2017; 7(2).
6. Lennox C. Identifying Failing Companies: A Re-evaluation of the Logit-, Probit- and DA Approaches // Elsevier Science Inc, 1999.
7. Ohlson J. Financial Ratios and the Probabilistic Prediction of Bankruptcy // Journal of Accounting Research. 1980. No 18. pp. 109 - 131.
8. Resolution of the Government of the Russian Federation of June 25, 2003 No. 367 "On Approval of the Rules for Conducting Financial Analysis by the Arbitration Manager"
9. Statistical Bulletin of the UFRBI as of December 31, 2019 [Electronic resource]. URL: http://download.fedresurs.ru/news/CTaTHCTHqecKHH%206roflfleTeHb%20EQPCE%202019.pdf (date of the application: 14.09.2020)
10. The system of professional analysis of markets and companies SPARK [Electronic resource]. URL: http://www.spark - interfax.ru (date of the application: 21.09.2020)
11. Yaw M. Mensah Journal of Accounting Research Vol. 22, No. 1 (Spring, 1984), pp. 380-395.
12. Zemskov V.V., Soloviev A.I., Soloviev S.A. Insolvency (bankruptcy) risk assessment models: history and modernity // Economics. Taxes. Law. 10(6): 91 - 100. - 2017.
13. Zhdanov V.Yu., Afanasyeva O.A. A model for diagnosing the risk of bankruptcy for airlines taking into account the economic factors of the external environment. Vestnik UrFU, no. 6, 2011, p. 126-138.
