Risks of lending to small and medium-sized businesses. Foreign experience of government support for SMEs

But, as both theory and experience in assessing and managing credit risk show, credit risk has a complex internal structure. In addition to the risks caused by the characteristics of each individual borrower, there are other components. The SME lending market in our country is the last segment of the credit market, which is still insufficiently covered by banking services. How should systems be designed to assess and manage the risks of lending to enterprises in this segment?

What does credit risk consist of?

Simultaneously with the rapid growth of the credit market in Russia in the first decade of the 21st century. The risks associated with the lending business have also increased. Because of this, the importance of methods for assessing the creditworthiness of borrowers for Russian banks also increases. But to create and use methods for measuring and managing credit risks, you must first understand what types of credit risks exist and how they relate to each other. In Fig. Figure 1 shows a hierarchy of types of credit risk, which is formed into a three-level structure.

Picture 1

At a basic level, credit risk is represented by transaction risk. This risk is associated with the variability of the creditworthiness of individual borrowers, which arises in response to changes in economic, industry, socio-demographic and other factors affecting it. This risk manifests itself in the variability of the cash flow of enterprises or the income of individual borrowers. Because of this, the likelihood of repayment or, conversely, non-repayment of borrowed funds may also undergo changes. At the next level of the hierarchy there are risks associated with the “behavior” of large groups of loans, united by the principle of similarity into a “single large loan,” called a portfolio. Combining loans into a portfolio is dictated by the need to reduce management costs: it is assumed that such a portfolio can be managed as one large loan. But then such a meta-credit must be characterized by some parameters that allow one to assess its inherent risk - the so-called portfolio risk. The portfolio combines loans that are exposed to the same risk factors, which include both economic (for example, the state of demand in the industry) and social (for example, the level of income of the population) factors. Let's take an example of a possible situation with a portfolio of SME loans: in the event of a fall in demand, the weakest enterprises with the least diversified product portfolios or with the least diversified distribution systems will have their income fall first and, therefore, the likelihood of default will increase. The next, third, level of the hierarchy is represented by allocative credit risk - a risk caused by the distribution of bank assets across industries, regions of its presence and bank products. Different dynamics of development and different states of regional economies, industries and, for example, demand for different types of bank loans determine the variability in the quality of loan portfolios formed by the bank. Thus, investing in different proportions of funds in lending to the same composition of industries, which are offered the same credit products, and the industries are localized in the same regions, will lead to the fact that each of these different possible allocations of credit resources will generate its profitability and will be characterized by its level of credit, allocation, and risk. In this article, we will focus on the lowest—basic—transaction level of credit risk.

The foundation of transactional credit risk management systems is scoring

Let us begin the description of the methodological support for assessing and managing transactional credit risk with the definition of the term “scoring”. Credit scoring is a fast, accurate, objective and sustainable procedure for assessing credit risk, which has a scientific basis. Scoring always represents one or another mathematical model that correlates the level of credit risk (the probability of borrower default) with many different parameters that characterize the borrower - an individual or legal entity. Let us immediately note that there can be many scoring models for solving the same problem, for example, assessing the credit risk of SMEs. Moreover, each of these models is built according to an individual algorithm, uses its own set of factors characterizing the risk associated with lending to the borrower, and as a result receives a threshold assessment, which allows us to divide borrowers into “bad” and “good”. The point of credit scoring is that each loan applicant is assigned an individual assessment of credit risk - the probability of default. Comparing the value of the probability of default obtained for a particular borrower with a threshold assessment specific (we emphasize this) for each scoring model helps to solve the most difficult problem of choosing when issuing a loan: to give funds to a given borrower or not. Thus, scoring is essentially an automatic or automated procedure that classifies borrowers into the required number of classes. In the simplest case, there are two such classes - those to whom a loan can be issued, and those to whom it is strictly “contraindicated”.

Thanks to the use of scoring, the bank is able to reduce the number of “bad” loans by filtering the flow of client loan applications. As evidence, we present data on lending to individuals using the Fair Isaac scoring system. After “passing” the factors characterizing the borrower through the scoring model, we obtain a number (score) that determines the level of credit risk inherent in this borrower. This number takes one of the values ​​in the range from 500 to 800. Each of the values ​​in this interval characterizes a different probability of repayment of loan obligations by the borrower. That is, different credit scoring values ​​imply different ratios of “good” and “bad” borrowers (Fig. 2).

Figure 2

In Fig. 2, the horizontal axis shows the value of the scoring score calculated by the model, and the ordinate axis shows the probability of default of the borrower corresponding to this scoring score. As can be seen from the figure, an increase in the borrower's scoring score is accompanied by a decrease in the probability of his default: the higher the scoring score, the more resistant a particular borrower is to credit risk. The figure illustrates this dependence: if 100 people with a scoring score exceeding 800 contact the bank, then only one of them will not return the funds taken. And vice versa, if 100 people with a scoring score of 499 or less contact the company, then 87 of them will not return the funds borrowed. Thus, by lending to borrowers with a high scoring value, the bank reduces the likelihood of loan default. This reduces losses and increases profits from lending activities without lowering lending standards.

How does scoring work? What is inside?

Several ingredients are needed to create scoring systems. We will begin our consideration with an analysis of scoring models used to assess the creditworthiness of enterprises, since scoring models have already been developed for enterprises, the structure of which is described in scientific periodicals. The most famous of these models is the E. Altman model, the first version of which was developed in 1968 based on statistical data from less than 70 American companies, half of which went bankrupt. This model is designed to assess the creditworthiness of large public companies in basic sectors of the American economy. The Altman model cannot be used to assess the creditworthiness of, for example, small businesses. Therefore, in 1984, researcher D. Fulmer created a special model for assessing the creditworthiness of small businesses with an annual turnover of about $0.5-1 million. The third of the models we are considering was created by the world famous company Fair Isaac - a recognized leader in the development of scoring models for lending to individuals . This is one of the least public models, with little known about its internal structure. Can anything unify scoring models for such different entities: large enterprises, small businesses and individuals? It turns out - yes, it can! This unifying point for all three types of models is equality:

where Z is the scoring value (scoring point);
a k are weight coefficients characterizing the significance of risk factors;
X k are risk factors that determine the borrower’s creditworthiness.

This formula is designed to calculate the credit scoring value, or a numerical value characterizing the quality of the borrower's creditworthiness. It is this (or similar) formula that is the “core” of almost any scoring system. In particular, in the Altman model it takes the form:

where the model coefficients take values ​​1.2; 1.4; 3.3; 0.6; 0.999 and are weights that determine the significance of risk factors; characters A, B, C, etc. - risk factors. For example, A is the ratio of working capital to total assets; B is the ratio of retained earnings from previous years to total assets; C is the ratio of earnings before interest and taxes to total assets; D is the ratio of market capitalization to the total book value of debt obligations; E is the ratio of sales volume to total assets.

In the Fulmer model, a similar formula for assessing creditworthiness takes the following form:

Z = 6.075 + 5.528V 1 + 0.212V 2 + 0.073V 3 + 1.270V 4 + 0.120V 5 +
+ 2.335V 6 + 0.575V 7 + 1.083V 8 + 0.849V 9,

where V 1 is the ratio of retained earnings from previous years to total assets;
V 2 - the ratio of sales volume to total assets;
V 3 - the ratio of profit before taxes to total assets;
V 4 is the ratio of cash flow to total debt;
V 5 is the ratio of debt to total assets;
V 6 - the ratio of current liabilities to total assets;
V 7 - logarithm of tangible assets;
V 8 is the ratio of working capital to total debt;
V 9 is the logarithm of the ratio of earnings before interest and taxes to interest paid.

The two scoring models described, like many other models, share a common property - their multidimensionality, which can be illustrated in the simplest case for two risk factors by a geometric “interpretation” (Fig. 3), where credit risk factors are certain variables X1 and X2 (their specific meaning is not important in this case).

Figure 3

Borrowers of two classes are represented in the figure by ovals of different colors: some, for example, “bad” ones, are represented by a gray oval, while others (“good”) are represented by a black one. It is not possible to distinguish “bad” borrowers from “good” ones by any single risk factor (due to the significant overlap of risk factor distribution functions - bell-shaped curves). In Fig. 3 bell-shaped curves along the axes of risk factors are formed by projecting groups of “good” and “bad” borrowers onto these risk factors. These projections—probability density functions—describe the frequency of occurrence of borrower properties used for scoring in classified groups. The larger area of ​​intersection of these curves for any of the risk factors indicates the impossibility of distinguishing “bad” borrowers from “good” ones. Borrowers of different classes are very similar to each other when assessed by the first and second risk factors. The scoring model “searches,” using the statistics of previously processed loans, for such a “viewing angle” on the data in the space of risk factors (in our case it is two-dimensional, but in the general case it is multidimensional) so that objects of different classes viewed from this “angle” are as are not similar to each other. In Fig. 3, this “angle of view” is indicated by a dotted line passing between the gray and black ovals and separating them. The perpendicular to this straight line is the scoring axis, projection onto which the images of “bad” and “good” borrowers makes it possible to distinguish them from each other. The intersection point of these straight lines gives the threshold scoring value (cut-off level) - Z*. The density functions of borrowers of different classes when projected onto the Z scoring axis become different from each other. Where do the numerical values ​​of the coefficients that weigh the risk factors included in the model come from? These coefficients are the result of a training procedure, when, to configure the model, it is presented with available statistical data on loans issued and the effectiveness of this process (“bad” and “good” borrowers) and it iteratively “selects” the coefficients in such a way that the accuracy of recognizing “bad” and “good” borrowers “good” borrowers was maximum. In Fig. 3 is the selection of the angle of inclination of the straight line cutting the gray and black ovals, and the point of intersection of this straight line with the ordinate axis.

To determine the model coefficients, it is necessary that the statistical sample be divided into those groups of borrowers (in the simplest case, there are two of them - “bad” and “good”) that the scoring model should recognize. This problem is referred to as “credit graveyard.” Moreover, the data used to select ratios is subject to rather strict requirements: in order for these ratios to “feel” “bad” borrowers, there must be quite a lot of them (and many of our banks have a small number of “bad” borrowers, since they are just trying to learn lend to SMEs). The figures characterizing the ratio of “bad” borrowers to “good” ones are typical for many banks: from 1 to 100 to 10-15 to 100 (in our experience, the order of magnitude does not vary much). Of course, as a result of the 2008 crisis, the number of “bad” borrowers grew to such an extent that many banks were close to collapse, but... problems with statistical databases still exist even for such banks, since their fall was due to high concentration enterprises of specific industries in their loan portfolios. Such industries that prevailed in bank portfolios included construction and trade. It is not yet possible to talk about the possibility of implementing the statistical learning procedure, even with such “credit cemeteries”. In addition to the quantitative ratio in the training statistics of “bad” and “good” loans, an important factor is the total number of examples for each industry. And the formation of portfolios from such economically different sectors as trade and construction leads to the fact that we are trying to describe various objects with one model. Construction is characterized by a significant volume of fixed assets and a fairly slow capital turnover, while trade is characterized by a small volume of fixed assets and high capital turnover. Note that the more detailed the description of the borrower (which, of course, any lender strives for, using a larger number of characteristics), the greater the number of both “good” and “bad” examples the “credit graveyard” used should contain. Thus, to create a scoring system that uses a supervised learning procedure (both of the models we are discussing are included here), it is necessary to have a sufficient number of borrowers who have caused damage to the bank. There is a workaround that requires the use of expert knowledge. However, when choosing it, you should understand how you can assess the composition of the characteristics required for scoring, the significance of a particular sign of creditworthiness, and how to combine the opinions of many experts on this matter, since relying on the opinion of one person when issuing loans is dangerous. That is why, when formalizing expert knowledge, we still find ourselves on the “road” that leads us to statistical scoring.

To conclude this section, let us note that the algorithms we chose to train the scoring model belong to the class of statistical models: to build models, a training sample and a statistical training procedure are required. This means that the scoring module must have at least two modes of operation. The first, if data is available, is the mode for training the model: finding such model coefficients that will best classify a sample of statistical data. The second mode is the actual operation of the constructed model; in this mode, the model ensures the implementation of the classification of the input flow of borrowers into classes pre-established in the training mode. To implement the first mode—training a scoring model—several preconditions must be met. Firstly, statistical data must be pre-prepared in a special way: the data sample must be divided into two parts - training and testing. In the training sample, it is necessary to collect excess data on potential borrowers. It should include variables that could potentially be useful in deciding the creditworthiness of borrowers, and the selection of specific variables for inclusion in the scoring model is carried out during the learning process and without human intervention. Secondly, the user (banking specialist) should be able to choose from several types of scoring models (we are talking about the two most common algorithms: logistic regression and decision trees). All of the above is illustrated in Fig. 4.

Figure 4

The left side of the figure shows a block diagram of the functioning of the scoring module in the training mode, and the right side shows the operation mode. The first of the described operating modes of this module ensures the selection from a redundant set of features of that subset that provides the required level of classification, that is, it is in this mode that the mathematical scoring model is built (for example, logistic regression coefficients are determined). But to build a scoring model, it is necessary to somehow ensure that the training procedure is stopped, for which a special model is used that calculates the so-called ROC curve and the quality indicator of the scoring model - AUC. After the required level of model quality is achieved during the statistical training process, the statistical training procedure is completed and the model goes into operation mode. In Fig. Figure 5 shows the ROC curve obtained when implementing a scoring system in one of the Russian banks included in the top 100. The vertical axis on the graph shows the percentage of “bad” borrowers that the model catches from the total number of borrowers with insufficient credit quality. The horizontal axis shows the share of borrowers from the total flow of borrowers - potential clients who will be refused to receive credit funds. The right angle bisector going from left to right in the figure shows the scoring model, which “flips a coin” (random classifier) ​​to make a decision. It is clear that the better the scoring model, the steeper the ROC curve should be.

In an ideal model, it should coincide with a right angle (top left). This means that the model recognizes all “bad” borrowers in the training sample, but does not unreasonably deny anyone a loan, and this cannot happen. As can be seen from the illustrative example characterizing scoring in an almost ideal situation, there is always some overlap between the images of “good” and “bad” borrowers. Therefore, in reality, the curve should lie in an intermediate position (between the bisector and the upper left corner). The quality of the model, the ROC curve for which is shown in Fig. 5 was quite high, with an AUC of 0.85.

Figure 5

Problems of obtaining a scoring model for SME lending

As mentioned above, the introduction of scoring in banking management is becoming very relevant due to the growth of both consumer and commercial lending. Let us outline the problems that the banking community will have to face along this path.

An attempt to apply the Altman model to Gazprom, Rosneft or LUKOIL, at least from a formal point of view, will not encounter any difficulties. There are official reporting data, there are weighting coefficients, which means it is possible to calculate an assessment of the borrower’s creditworthiness. But what to do if you need to evaluate not the largest companies mentioned, but “father Fedor’s candle factory,” whose shares are not quoted not only on the NYSE, but even on the MICEX (remember that we are considering an assessment of the creditworthiness of SMEs). Even a quick glance at the corresponding formula is enough to see that out of five explanatory variables in the case of a “candle factory,” only four variables remain in the formula. The scoring value (the Z value in the formula) will decrease (although, strictly speaking, this will only happen if 0 is put in place of D), which in fact does not correspond to the case under consideration). As noted, the scoring value for a specific borrower is compared with a threshold value:

Z > Z* - “good” borrowers;
Z< Z* — «плохие» заемщики.

However, if it is impossible to take into account some variables (if the company’s shares are not quoted on the stock exchange, then variable D is absent in the description of credit quality), the “measuring instrument” itself, represented by the model, breaks down (a numerical assessment without taking into account factor D, for example, will always be shifted to the region worst scoring scores). In reality, the situation is even more complicated: without taking into account the variable D, we unwittingly change the geometry of the space of risk factors and, as a consequence, the weighting coefficients for other risk factors of the borrower. The model itself changes: the critical value of the scoring assessment (cut-off threshold), with which the assessment of each borrower is compared, becomes different. Consequently, in our conditions, the very choice of explanatory variables for assessing the scoring of Russian companies is a very non-trivial task. E. Altman built his model on data from only 60 companies, it reflects the very specific industry specifics of business (basic sectors of the American economy), it does not take into account the risks associated with business cycles in Russia and the risks inherent in companies with other industry affiliations . Therefore, we can state the following: the use of this kind of model by mechanically transferring it to our conditions becomes a powerful risk factor for credit assessment - what is called model risk in risk management. A very striking example of the risks associated with the use of statistical scoring models is given in one of the works devoted to studying the effectiveness of scoring models1. It states that the set of variables that form the score may change over time and that the “border” between the analyzed groups may not be linear (as shown in Fig. 3), but have a significantly more complex shape that cannot be described the simplest formula such as the Altman model. The authors of this article examined several mathematical approaches for constructing scoring, where 31 financial ratios were used as risk factors, characterizing various aspects of the company’s financial condition. They analyzed 11 scoring models developed in the period from 1931 to 1996, for the construction of which three mathematical approaches were used: discriminant analysis, logit model and genetic algorithms. The authors of the article showed two main points. The first is related to the fact that the composition of risk factors in the scoring model changes: it changes depending on time - the sooner before the future bankruptcy the scoring system needs to “see” it, the more variables must be taken into account in the model. The second is due to the fact that the boundary between classes of borrowers is nonlinear: the accuracy of estimates obtained using scoring based on genetic algorithms (they generate a nonlinear boundary) is significantly higher than that of models based on discriminant analysis (it generates a linear boundary). True, the first approach requires on average three times more variables than the second.

Consideration of the problem of credit risk management will not be complete if we do not touch upon the problems and possible ways to solve them when designing a system for managing credit risks not only for individuals, but also for legal entities. We will do this using the example of assessing the creditworthiness of small and medium-sized businesses. The purpose of this section of the article is to show that the proposed composition and architecture are preserved during the transition from lending to individuals to lending to small and medium-sized businesses. Proof of this fact will allow us to say that the composition of modules of the credit risk management system we propose and its functional architecture are universal.

Obviously, the first problem that we will encounter when moving to measuring the credit risk of small and medium-sized enterprises is the fact that sufficient factual material has not been accumulated for this type of borrower. To maintain the universality of our proposed functional architecture and the composition of the modules that form the system for managing the transactional part of credit risk, we propose to slightly expand the logical scoring model for SMEs. The proposed extension is represented by the block diagram in Fig. 6.

Figure 6

The essence of the changes in the logical structure of the scoring model is that the composition of algorithms traditional for solving this problem (decision trees, logistic regression, etc.) is expanded, and we propose to include SD-modeling algorithms in the model. The reasons why we offer this solution are as follows:
— in the absence of statistical data, a traditional set of algorithms turns out to be simply useless due to the inability to implement model training;
— for making lending decisions in the case of small and medium-sized businesses, the borrower’s creditworthiness is influenced by a significantly larger range of variables than for individuals;
— it is a widely known fact that the financial statements of enterprises in the Russian economy often very poorly reflect the real state of affairs in business due to the fact that they are very distorted due to tax “optimization”;
— the scoring model should reflect the industry specifics of SMEs.

In particular, in the context of the second statement, we can talk about the need to take into account not only financial information in the scoring model for SMEs. It is very important, due to scale effects, to assess the economic environment of the enterprises being scored. It is important to take into account environmental variables such as supply and demand in scoring models. With their variations, the cash flow of an enterprise and all its financial indicators, on which scoring models for legal entities are usually based, can undergo sharp changes. In addition, the state of enterprises, again due to their small scale, may also be affected by management features. Therefore, it is extremely important to be able to take into account the quality of enterprise management. And if it is not difficult to evaluate in numerical form the state of supply and demand in the industry to which the enterprise belongs, then assessing the quality of enterprise management is a non-trivial task. This requires that the model be able to consume expert information. The inclusion of an SD modeling block in the logical structure of the scoring model for SMEs allows us to adequately take into account all of the above requirements. To implement any SD model, as is known, as a first step it is necessary to draw up a cognitive map. A cognitive map is a diagram of causal influences in the form of a directed graph. The nodes of this graph represent variables that are included by the analyst in describing the creditworthiness of the enterprise, and the edges represent the causal influences of the variables on each other. In Fig. Figure 7 shows a cognitive map of a hypothetical forest industry enterprise that is engaged in timber harvesting and processing.

Such a change in the logical structure of the scoring model provides not only a solution to the four problems we listed earlier, but also the ability to dynamically assess the creditworthiness of SMEs. The use of the SD model in the scoring model allows us to generate the data that is missing for constructing statistical scoring. By directly taking into account the volume, timing and type of loans (Fig. 7), the model allows you to generate a “credit graveyard”, defining a state of default as the inability to repay the current debt within, say, three months. By varying the input indicators of the model, such as demand, supply, quality of management, loan parameters, in the process of SD modeling we obtain different trajectories for the enterprise’s cash flow, and therefore different conditions for the borrower to default (with different combinations of input parameters of the model, we have and various financial indicators). Having thus generated an artificial “credit cemetery”, we can apply traditional scoring algorithms in the form of the same logistic regression on the specified statistics in a standard way. But, in addition, we get a significant benefit from changing the logical structure of the model due to the fact that such a model allows us not only to solve the problem of lack of statistical data, which is important for the training mode, but also significantly expands the scoring functionality in the operating mode of the scoring model.

Let us explain what is new from the expansion of the logical structure for the last mode. Firstly, we can generate arbitrarily large data samples, which ensures the accuracy of statistical learning models that will take into account the industry specificity of the business of the SMEs being financed. Due to the absence of restrictions on the volume of generated artificial data on defaults of SME enterprises (which in real life never exists in the required quantity), we take into account industry specificity in the structure of the cognitive map: trading enterprises are characterized by rapid capital turnover and low fixed assets, while manufacturing enterprises are characterized by large values ​​of fixed assets and a slow turnover rate. By using expert information in the cognitive map, we can model the impact of management quality on cash flows, and when using macroeconomic statistics, the impact of variations in supply and demand on the sales volumes of the company under evaluation. And finally, we can model the company’s cash flows over time using SD models in the scoring structure, which will allow us to more rationally formulate a payment schedule. In addition, the dynamism of the resulting assessments will allow us to naturally, within the same model, implement not only applicative, but also behavioral scoring, if during the loan servicing period the assessed enterprise suddenly has problems repaying its debt. This structure of the scoring model will allow us, when problems arise, to effectively assess the prospects for repaying the debt that has arisen and make more informed management decisions regarding such a borrower.

Figure 7

Instead of a conclusion: what else is needed for a “quiet” life in the credit business?

Due to the fact that credit risk is structured hierarchically (has three levels - from transactional to allocation), to manage credit risks we will need two more sets of models. The first will serve the purposes of managing portfolio credit risk, and the purpose of the second is to support management decisions regarding the allocation of credit capital by region of the bank’s presence and by products sold. Moreover, even at the transactional level, to manage the credit risks of borrowers, it is not enough to simply classify the input flow of borrowers into “good” and “bad”. A number of other functions are needed, without which the use of scoring will not give satisfactory results. These two topics will be discussed later.

Estimate:

1 0

Loans provided by the MB are a special type of loan provided to its recipients in the form of deferred payment for purchased raw materials, materials, semi-finished products and goods. In terms of economic content, MB loans for replenishing working capital differ from other loans in that they are not of an investment nature, but are aimed at meeting the current needs of the population for goods, medical, household and educational services, therefore the risks for these types of loans are of a special nature.

Analysis of the data presented in table. 4.1 shows that the existing mechanism for issuing and repaying MB loans to replenish working capital in many banks is of a similar nature.

A common disadvantage of the loan process is that the application review period is too short, when the bank’s security services do not have time to check the correctness and authenticity of the documents submitted by the borrowers. Forged documents are a significant factor in the future emergence of credit risk, so the Bank of Russia needs to take steps to formalize the process of lending to small banks to replenish working capital in Russian banks. It is necessary to develop for banks an approximate methodology for issuing and providing loans to small banks to replenish working capital. This proposal is supported by large differences in the interest rates applied by banks on loans, additional commissions and the presence of constantly increasing overdue debt on MB loans.

To a large extent, the listed shortcomings are explained by the incorrect approach to the classification of MB loans for replenishment of working capital. Thus, many banks understand by them only overdraft loans for the purchase of goods for resale and provision of services. Other banks, on the contrary, classify all loans provided to the MB and MB owners as individuals as loans for current needs. While the development of approaches to organizing lending and managing emerging risks depends on the correct classification of loans.

Analysis of the mechanism for providing MB loans to replenish working capital based on studying the practice of individual banks

Table 4.1

Naturally, all MB loans are divided into loans of an investment nature, i.e. creating new investment value (loans for construction, loans for the purchase of cars, equipment), and loans for replenishing working capital that do not create investment value.

Among MB loans for replenishing working capital, it is legitimate to highlight loans for the purchase of goods and loans for the provision of services. It is possible to classify the mechanism of their lending by type of borrower, terms of provision and consideration, conditions of issuance and repayment, interest rates, commissions, risks, documents, types of collateral, terms and methods of repayment, lending objects, volume. For example, loans for the purchase of goods are sometimes, in violation of the law, provided in cash using the bank cards of the owner of the MB enterprise, using an overdraft. Loans for services can be received through a non-cash transfer of funds to the borrower’s current account, and from the borrower’s current account to the account of the organization providing the services. This leads to different terms for consideration and provision of a loan, different approaches to ensuring its repayment and calculating risks, participation of various organizations in the lending process, and excellent interest rates.

Let's analyze the features of MB loans for the purchase of goods. They are provided for the purchase of durable goods - refrigerators, televisions, computers, furniture, etc. for the purpose of their further resale. This type of lending is extremely common in foreign countries and in Russia. At the same time, calculations show that a credit purchase turns out to be more expensive than for funds in a current account. The level of the current interest rate in the Russian market does not provide benefits from credit conditions.

To prevent risks and improve the efficiency of lending, it is necessary to specify in the lending agreement when ownership of the purchased goods passes to the borrower - immediately or after full repayment of the loan. These conditions must be agreed upon in advance between the bank and the client. The transfer of ownership of the purchased goods to the borrower after full repayment of the loan reduces the risks of collateral for this type of lending.

Based on the development trend of this type of lending, it is possible to predict further growth in MB loans for the purchase of goods in 2011-2013, when all previously taken loans will not yet be repaid and new ones will be issued. Consequently, this type of loans will continue to expand its boundaries.

Competition between banks in the lending market for the purchase of durable goods by small business trading enterprises is extremely aggressive. The result is the presence of a large number of banks and a fairly dense distribution of the market between them. However, an opposite trend can also be observed, when individual banks are reducing these operations due to high risks and the proportion of late payments. In leading banks, commodity lending under MB loans is also problematic.

Let's consider the results of the work of one of the Moscow banks (let's call it bank N) on MB lending for six months of 2010 (Table 4.2).

As can be seen from the table, for six months of 2010 the bank issued loans to replenish working capital in the amount of 31.8 billion rubles, while the effective interest rate was 37.44%, and the first three months of the period under review it was about 55% and decreased due to a large-scale advertising campaign in which MB was offered a product with an average term of 24 months and a low rate of 18%.

Noteworthy is the high default rate for homogeneous loan portfolios, associated with the terms of lending and the duration of overdue payments.

In order to reduce emerging risks when calculating actual costs and reserves for homogeneous bank portfolios in accordance with Regulation No. 254-P, as well as reducing the effective interest rate for a loan, when calculating actual costs and reserves for homogeneous MB portfolios, it is proposed to include the costs of an advertising campaign.

The full advertising cycle from approval of an advertising campaign plan to its implementation takes about a month. According to the requirements of the Bank of Russia, banks are required to adjust their interest rate policy at least once every two weeks. If the advertising material does not comply with the interest rate policy, the bank will be subject to fines for unfair advertising in the amount of 500 thousand rubles. Taking into account the above, we recommend that the Bank of Russia, in order to assess the adequacy of the bank’s interest rates to market conditions, set the frequency of reviewing interest rates on MB loans once a month.

One of the mandatory requirements of the Bank of Russia for banks should be the indication of effective interest rates not only when including a loan in a portfolio of homogeneous loans, but also in all lending agreements concluded with the Bank. It is necessary to oblige banks to include information on the size of the effective interest rate in all interconnected document registers: application form

Dynamics of key indicators and quality of MB lending for working capital replenishment

for six months of 2010

Table 4.2

client; to the tariff table; on monthly credit statements, especially when using credit cards, where the payment amount, fees, and loan balance may vary.

The main processes for assessing risks on loans for small businesses, including consideration of the application and decision-making by the bank, monitoring risks, accrual of reserves for possible losses, are completely similar to those used in lending to all legal entities. To increase the efficiency of lending to small businesses and reduce risks, it is necessary to recommend that the Bank of Russia require banks to regularly prepare forecast risk calculations based on the first facts of non-repayment of the next loan payment, using forecast mathematical models.

Consideration of the features of the influence of the type of loan provided by the MB on credit risks allows us to draw conclusions that MB loans have both general and special risks (Table 4.3).

Table 4.3

Classification of credit risk of MB loans

In addition to the structural elements of risk of MB loans listed in the table, it is necessary to distinguish between aggregate (general) and individual types of credit risk, as well as take into account the features of credit and other risks that arise when lending to MB.

Lending to small and medium-sized businesses is traditionally classified as a high-risk category. However, statistics on non-returns and growth rates of portfolios in this market segment indicate the opposite. Without a borrower assessment model that has been proven over the years, banks are forced to constantly balance between the quality and cost of risk management techniques.

Shine light on dark places

According to the Central Bank, in 2007 the volume of loans provided to enterprises and organizations increased by 50.5% and reached 8.7 trillion rubles. The small business lending market, according to RBC estimates, grew somewhat faster - the increase over the year was 55%. But with such rapid growth, there are also many pitfalls.

The main problems that arise when lending to small and medium-sized businesses are the low transparency of this segment and the lack of reliable collateral. Anna Malysheva, head of the lending department for medium and small businesses at Rus Bank, stated: “In approximately 50% of cases, financial statements do not reflect the real financial and economic state of the enterprise’s activities, so we have to appeal to management accounting data, which is difficult to confirm and control in the future.”

As Andrey Kuznetsov, head of the small business development department at MDM Bank, said, his bank periodically conducts research on small and medium-sized businesses. And so far the results indicate that this market is at the “maturing” stage. Entrepreneurs, as a rule, think about lending only at the moment when they need to close the current “holes” of the business. Another inhibitory factor is the closeness of entrepreneurs to the bank and some of their cunning when working with an expert. “The majority of entrepreneurs,” the speaker noted, “receive a refusal precisely because during negotiations with the bank they do not know how to properly demonstrate their business.”

Although many banks have only a few years of experience in this segment (if we take it in bulk), credit institutions have accumulated enough experience to predict which skeleton might fall out of the closet.

Irina Bychkova, head of the lending and guarantees department at Investtorgbank, shared her experience: “Sometimes problems arising due to conflicts between owners lead to serious consequences. There have been cases of business being taken away from the owner by company management, and attempts at raider takeover have been observed. Problematic enterprises are those that depend on a narrow circle of suppliers or buyers. There is also a risk of unbalanced growth of a company that has difficulties, including in repaying borrowed funds. Often, due to a lack of collateral, it is necessary to attract the personal property of managers, although, in essence, this is not entirely correct. Often you have to act as an auditor, sometimes problems are solved even by replacing employees in the company.”

These factors are rather subjective. With a serious and thoughtful approach to business and the loan application procedure, most issues can be removed from the agenda. But the difficulties that arise when lending to small businesses can also be of an objective nature, and even “experienced” employees can be baffled. In particular, Natalya Golovanova, head of the small and medium-sized business department of the Russian Development Bank, noted that when assessing enterprises that apply special tax regimes, it is more difficult to verify some of the borrower’s financial indicators.

Bankers are faced with the question of how to check enterprises operating under the so-called “simplified system”. Some experts believe that it can be extremely difficult to assess the real condition of the borrower using a minimum set of documents. Thus, even complete financial statements may not contain the necessary information and may not fully reflect current business processes, therefore, experts believe, unambiguous conclusions cannot be drawn only on the basis of official documentation.

Therefore, banks, based on global experience and their own data, are developing various methods for assessing the creditworthiness of a small enterprise.

From particular to general

There is a pattern in the banking industry: if a bank switches to a scoring model for assessing the borrower, it means that the product is becoming mass-produced. For example, after the introduction of scoring in mortgage lending, this product became available to the general public.

When lending to small businesses, the task becomes somewhat more complicated. Firstly, we are often talking about larger amounts. Secondly, it is much easier to check information and evaluate an individual than a business. If when lending to an individual, most questions can be answered unambiguously (age, marital status, number of children, etc.), then when assessing an enterprise, important factors are the personal qualities of the manager, relationships with partners, reporting form, relationships in the team, prospects for the development of the industry, etc.

Nevertheless, today, when the interest of banks in lending to small and medium-sized businesses is constantly growing (fueled, among other things, by government programs), and those wishing to receive a loan are practically lining up, automated systems for assessing borrowers may come in handy.

Formally, there are two main methods for assessing a borrower: individual analysis and scoring. It is no secret that the scoring model has two main advantages: high processing speed and low cost. But at the same time, according to the law of “genre”, the quality of verification with this approach decreases. Ivan Khomenko, head of the small and medium-sized business lending department at Moscow-Credit Bank, explained: “The cost of the scoring assessment is certainly lower, since the application is reviewed in less time and employees with a lower level of qualifications and wages can be used for its analysis. As loan amounts increase, the borrower’s analysis becomes more individual, as the loan’s profitability in absolute terms increases.”

At the initial stage, it is important for the bank to correlate possible risks and cost items. It is for this reason that most credit institutions use a mixed system for assessing borrowers.

Thus, VTB 24 Bank, Russian Development Bank, UniCredit Bank, Moscow-Credit Bank, Soyuz Bank use a scoring system for assessing borrowers when implementing “microcredit”, that is, under programs in which the loan amount is the minimum for a given segment at this credit institution.

At Investtorgbank, the choice of borrower assessment methodology depends not only on the loan amount, but also on the region in which and how long the client has been working in this business. Chairman of the Board of Investtorgbank Vladimir Gudkov explained this approach: “Regions differ very much from each other in terms of wages, living wages, infrastructure and service sector development... This imposes certain features on the principles of doing business. The Moscow region has switched almost 90% to a “transparent” business scheme, so scoring can be used in the capital district. In the regions, on the contrary, “gray” schemes predominate, which means that it is necessary to carry out an individual assessment of the borrower.”

However, the methodology used by banks cannot always be called scoring. As the head of the department for work with SMEs, Alexandra Bugaeva (Svedbank), noted, the SME lending market in Russia is young: “The necessary credit history has not been accumulated, the work of the credit history bureau is still ineffective. In addition, scoring is applicable only for standard applications and, when making a decision, it uses a finite number of factors provided for when developing the program, which means it does not take into account the specifics of enterprises, possible additional factors, which is very important for SMEs, and which a credit analyst would pay attention to " Rus Bank has created a special simplified procedure for assessing a borrower, close to scoring. Anna Malysheva (Rus Bank) explained that this procedure is based on an analysis of certain stop factors and their compliance with the parameters of the lending program.

Natalya Golovanova (Russian Development Bank) drew attention to the other side of the problem of assessing the borrower: “Scoring assessment systems are not able to fully objectively evaluate the borrower’s business, and it often happens that the scoring system refuses “good” borrowers, while “bad” or even fraudsters , on the contrary, gives a positive answer. This is undesirable both for banks and for borrowers themselves.”

The line between a “reliable” and an “unreliable” borrower can be very thin. Two grocery stores in the same city may turn out to be completely different, since they have different suppliers, a different range of products, different managers, different points of sale, different service, etc. It is extremely difficult to take into account all these nuances mechanically. Therefore, bank employees prefer to make field trips and get acquainted with the production and management of the enterprise. Sometimes one visit can provide more information than a pile of documents.

Many banks are developing their own methodology for interacting with clients, which does not require a potential borrower to provide numerous certificates from various departments. A personal manager goes to the site and independently analyzes management reporting. In this way, it is possible, firstly, to minimize temporary losses on the part of the borrower, secondly, to analyze and provide for all possible risks and, thirdly, to optimize the process. The data is compiled into a standard form, on the basis of which it will no longer be difficult to make a fateful decision.

Of course, with this approach, the costs are immeasurably higher. But the choice of methodology can be influenced by many factors: from the size of the branch network and the volume of business to the minimum loan amount determined by the bank. As Eduard Issopov, a member of the board of UniCredit Bank, noted, lending to small businesses can be attractive for a bank if it has a clearly set up risk management system: either the required number of people have been recruited for underwriting, or a scoring system has been developed.

How to balance costs and acceptable risks - today credit institutions decide this issue on their own.

Reserves wholesale and retail

Improving the quality of the loan portfolio for banks is associated not only with the prevention of defaults, but also with a decrease in the amount of reserve funds.

Regulation of the Central Bank No. 254-P “On the procedure for the formation by credit institutions of reserves for possible losses on loans, on loan and similar debt” allows for the purpose of reserving to combine minor (no more than 0.5% of the bank’s capital) loans with similar conditions lending, into portfolios of homogeneous loans.

Alexandra Bugaeva (Svedbank) noted the general trend in the lending market: “Banks are striving to increase the number of standard loans by developing a wide product line. This allows loans to be grouped into portfolios in order to form reserves for the entire portfolio, which saves the bank’s time and effort, and also reduces the cost of the loan product for both the bank and the borrower.” Therefore, the bulk of loans belong to the category of so-called standard loans.

Anna Malysheva (Rus Bank) stated: “In accordance with the opportunity provided by the Central Bank, Rus Bank combines all loans issued under the lending program for medium and small businesses into portfolios of homogeneous loans, and also creates a reserve for these portfolios with the amount of reserve rates starting from 1%. This mechanism significantly simplifies the bank’s work with this business segment, reduces labor intensity and allows for a more flexible approach to the process of issuing and maintaining a large number of standard loans.”

It is important for banks to accurately determine which category of borrower a client belongs to at the assessment stage and attach it to the appropriate portfolio. If the system fails and there are delays in servicing the debt, this loan is removed from the portfolio of homogeneous loans, the borrower is assigned a lower quality category, and additional reserves are placed on the loan. Of course, such a procedure is undesirable for banks.

If a loan does not correspond to any portfolio of homogeneous loans in a number of ways (for example, the loan size exceeds the permissible volume or there is a shortage of collateral), the bank uses individual reserves. As a rule, the share of such non-standard loans in the bank is insignificant.

It is more convenient for banks to adhere to a standardized approach and not issue high-risk loans. However, as Alexandra Bugaeva (Svedbank) notes, credit institutions often accommodate enterprises halfway, taking into account their management statements when analyzing them: “If the bank is willing to take on a higher risk, this will make the loan more expensive for both the borrower and the bank. The bank is forced to increase the amount of reserves, which entails additional costs, and provide compensation for the risk. For the borrower, this is reflected in an increase in the interest rate on the loan and more stringent lending conditions. Banks encounter similar situations quite often.”

As practice shows, behind humanity in banks there is always a logical justification and financial guarantees. Perhaps it is precisely thanks to this approach that the risks in lending to small and medium-sized businesses remain quite low today.

How are the risks associated with small business lending assessed? What requirements have been established by the Bank of Russia for creating reserves for possible loan losses? The Department of Banking Regulation and Supervision of the Bank of Russia answered these questions specifically at BO’s request.

Risk assessment for loans provided to small businesses is carried out in the manner established by Bank of Russia Regulation No. 254-P dated March 26, 2004 “On the procedure for credit institutions to form reserves for possible losses on loans, on loan and similar debt” (hereinafter - Regulation No. 254-P).

In accordance with clause 3.1.1 and clause 3.1.2 of Regulation No. 254-P, an assessment of the credit risk of an issued loan (professional judgment) must be made by a credit institution based on the results of a comprehensive and objective analysis of the borrower’s activities, taking into account its financial situation and quality of debt servicing on the loan, as well as all information available to the credit institution about any risks of the borrower, about the functioning of the risk on which the borrower operates.

Regulation No. 254-P does not provide for any special requirements for the creation of minimum reserves for possible losses on loans provided to small businesses.

At the same time, the Bank of Russia pays due attention to the issues of creating conditions for the bank to implement more effective risk assessment procedures and the formation of reserves for possible losses on loans provided to these entities. The implementation of these approaches helps to save banks' labor costs on lending to small and medium-sized businesses, while at the same time the use of modern risk assessment techniques makes it possible to form adequate reserves for loan losses.

For example, as part of clarifying approaches to evaluating loans and forming reserves for possible loan losses in accordance with Bank of Russia Directive No. 1759-U dated December 12, 2006 “On amending Bank of Russia Regulation No. 254-P dated March 26, 2004” On the procedure for the formation by credit institutions of reserves for possible losses on loans, on loan and equivalent debt”, which entered into force on July 1, 2007, provides for the exclusion from the requirements of clause 3.14.1 of Regulation No. 254-P, which establishes a list of loans for which a reserve is formed in the amount of at least 21% of loans provided to pawnshops, cooperatives, small business support funds and used by them to provide loans to small businesses and individuals. This clarification makes it possible not to classify loans provided to these entities and used to provide loans to small businesses as quality category III with the formation of a reserve in relation to such loans.

Directive of the Bank of Russia dated December 28, 2007 No. 1960-U “On amending clause 6.3 of Bank of Russia Regulations dated March 26, 2004 No. 254-P “On the procedure for credit institutions to form reserves for possible losses on loans, on loan and similar debt » it is possible to recognize as collateral for the purposes of Regulation No. 254-P the guarantees of business support funds established by constituent entities of the Russian Federation and funds to promote lending to small and medium-sized businesses, which allows the formation of a reserve for possible loan losses taking into account such collateral (that is, reducing the amount of the reserve by security amount).

In addition, in order to simplify the assessment of loans that are not significant in size, which, as a rule, include loans provided to small businesses, Regulation No. 254-P (clause 1.5 Chapter 5) provides for the possibility of combining similar loans into portfolios. This approach involves assessing the risk in relation to the entire loan portfolio based on data on the amount of losses for a group of similar loans for the previous period, while ensuring the comparability of all material circumstances relating to the nature, volume of loans, operating conditions of borrowers and other circumstances.

Currently, the issue of extending to loans granted to legal entities - small businesses, an approach that provides for the possibility of forming a portfolio of loans with a general level of impairment, determined by the presence and certain duration of overdue payments, established by paragraph 5.1 of Regulation No. 254-P, is being considered.

Currently, this approach is implemented in relation to loans provided to individuals.

Expert opinion

The prospects for the development of the small business lending system and the existing system of reserving funds for possible loan losses were assessed by the President of the Association of Regional Banks "Russia" Anatoly Aksakov:

Currently, banks independently determine the amount of reserves for loans issued to small and medium-sized enterprises, based on a thorough analysis of the financial and economic activities of enterprises. Of course, this slows down the process. In an effort to simplify the procedure, some banks issue loans to individual entrepreneurs, registering them as loans to individuals. After all, a loan to an individual can be issued within 24 hours, whereas upon receiving an application from an entrepreneur, it is necessary to conduct a financial analysis of the borrower. The Rossiya Association, in cooperation with bankers, has prepared proposals to amend Bank of Russia Regulation 254-P “On the procedure for forming reserves for possible loan losses.” In our opinion, the amount of reserves for possible loan losses to calculate the optimal interest rate for loans to small and medium-sized businesses should be at least 1-1.5% (for a portfolio of loans without overdue payments) and increase depending on the duration of the delay.

Loans to small and medium-sized businesses are classified as high risk. Most likely, they will become more expensive, and borrowers will face tougher conditions and refusals. However, in general, the implementation of the Central Bank’s initiative will contribute to the development of lending to small and medium-sized businesses and stimulate the activity of banks in this sector, because individual reservation of a mass borrower is too expensive for banks, and the introduction of a single algorithm will automate the process.

The problem of reporting and transparency of small and medium-sized businesses remains relevant. Many entrepreneurs still prefer to work according to “gray” and “black” schemes without disclosing their turnover. Of course, this affects the possibility of obtaining a loan, significantly reducing it. An increase in reserves for possible loan losses will entail a tightening of lending conditions - because bankers will begin to check the information declared by a potential borrower in even more detail. Perhaps this measure will encourage many market participants who are still working under “gray” and “black” schemes to legalize.

Credit risk is the risk of default or late payment on a bank loan.

The main reasons for the risk of loan non-repayment:

§ reduction (loss) of the borrower’s creditworthiness;

§ deterioration of the borrower’s business reputation.

Credit risk can arise for each individual loan issued by a bank or for the bank's entire loan portfolio (total credit risk). Therefore, it is important for the bank to develop a credit policy - a documented organizational scheme and a system of control over credit activities.

The main requirement for the formation of a loan portfolio is the balance of the latter, which should compensate for the increased risk of some loans with the reliability and profitability of other loans. The structure of the loan portfolio is formed under the influence of the following factors:

§ profitability and risk of individual loans;

§ borrower demand for certain types of loans;

§ credit risk standards established by the Central Bank;

§ structure of the bank’s credit resources by loan repayment terms.

Bank credit operations themselves are risky, so credit risk management should be aimed at reducing them, the main methods of which are:

§ assessing the borrower’s creditworthiness and establishing his credit rating;

§ pursuing a policy of diversification of loans (by size, type, groups of borrowers);

§ insurance of loans and deposits;

§ formation of reserves to cover possible losses on loans provided;

§ formation of an effective organizational structure of the bank in order to minimize credit risk.

In modern conditions of functioning of Russian banks, it is necessary to take into account the development of external sources of information about the creditworthiness of borrowers, foreign experience in corporate risk management and assessment of the solvency of potential bank clients.

When assessing credit risk at the preliminary stage, you need to use certain criteria. Stand out five main risk assessment criteria:

§ reputation, i.e. clarification of the relationship between the borrower - banking client and creditors (suppliers). The assessment of this condition can be made both on the basis of written information provided by the borrower, and oral conversation and based on recommendations provided by the borrower, especially when it comes to a personal loan or a loan to a group of individuals (for example, a partnership);

§ possibilities, i.e. determining the borrower’s solvency for the last period (or several years) depending on the volume of the upcoming credit transaction;

§ Capital. Availability of own (share) capital and the consent of the borrower to use it in some part, if necessary, to repay the loan;

§ Conditions. Finding out the current state of the economy (regional, national), but especially the sectoral one, which includes the borrower;

§ Pledge- This is one of the reliable loan collaterals. Sometimes it makes it possible to overcome the weakness of other credit risk assessment criteria.

Credit risk is directly dependent on the quality of the loan portfolio. The loan portfolio is the result of the bank's activities in providing loans, which includes the entire set of all loans issued by the bank for a certain period of time.

Any commercial bank is interested in high profitability of its loan portfolio. Since credit risk has a direct impact on this profitability, it is important to assess the impact of credit risk on the profitability of the loan portfolio. This work should be carried out systematically in order to be able to take prompt measures to prevent negative processes accompanied by credit risk.

To quantify the impact of credit risk on the profitability of a loan portfolio, you can use a system of coefficients.

The most important integral coefficient that determines the profitability of the loan portfolio is the net interest margin (NIM). Taking into account credit risk, it is determined as follows:

§ NIM - net interest margin;

§ D p - interest income;

§ R p - interest expenses;

§ Ук - loan losses;

§ KB - credit investments.

The peculiarity of this indicator is that it assesses the effectiveness of the credit risk management system in the bank as a whole. It takes into account both losses due to credit risk and income received due to the bank's acceptance of credit risk.

Speaking about the problems of business lending to small enterprises, it is necessary to understand that all the problems that exist today are closely related to each other and flow into one another.

The opinion that there is a high degree of risk when working with small borrowers is quite widespread. Firstly, it is difficult for banks to objectively assess the degree of stability of a potential borrower. Secondly, the activities of small businesses are often not transparent, which means that the bank is not always provided with reliable information about the income and expenses of the enterprise and the state of affairs in the business.

Here it is immediately worth making the following reservation: the risk of lending to small businesses associated with failure to provide reliable information is not as high as the risk of consumer lending, where loans are issued to individuals upon presentation of a passport and very limited information about place of work, income, etc.

Emerging businesses also occupy a special place here. This category of small business is associated with very high risks. The emerging business has no credit history or collateral, and it is unclear whether it will be able to develop enough to repay the loan.

The problem of collateral is one of the top priorities for all banks working in the field of small business lending. It is quite rare that small companies can provide the bank with liquid property as collateral, so banks have to take risks and issue unsecured or partially secured loans.

If the company is interested in amounts of 10-20 million rubles. and the loan terms are 5-10 years, then in this case banks will require collateral, the estimated value of which is at least 200% of the loan amount. Only a few can provide such collateral; therefore, it is not possible for most small businesses to obtain such a loan.

However, a number of banks sometimes, with the support of government agencies, issue loans without collateral. But often such loans are issued on rather unattractive conditions: the amounts are small - up to 50 thousand dollars (microloans), the effective rate is quite high (28-30% per annum in rubles), the terms are limited - as a rule, they do not exceed 1.5 years. However, this loan still solves some problems for entrepreneurs. For example, it can be used to eliminate cash gaps when there is not enough money in the cash register to conduct current financial activities.

In addition, small businesses can be supported by the Small Business Lending Assistance Fund. Conditions for providing a guarantee from the Fund: registration in the Register of Small and Medium Enterprises of Moscow (for small and medium enterprises located in the city of Moscow), in the corresponding Register of Small and Medium Enterprises in the territory of a constituent entity of the Russian Federation (except for the territory of the North Caucasus) or otherwise confirming its status as a small and medium-sized enterprise (in the event that a register of small and medium enterprises is not maintained in a constituent entity of the Russian Federation);

The period of economic activity is at least 6 months;

The field of activity is not the gambling business, the production and sale of excisable goods, as well as the extraction and sale of minerals, enterprises should not be parties to production sharing agreements; absence of overdue debts to budgets of all levels, as well as violations of the terms of previously concluded credit agreements, loan agreements, leasing agreements, etc.

The amount of the Fund's guarantee is no more than 50% of the amount of obligations regarding the return of the actually received loan amount and the payment of interest on it under the loan agreement and in any case no more than 30 million rubles. The amount of the Fund's remuneration for the provided guarantee is 1.75% per annum of the guarantee amount. But in any case, the issue of granting a loan and the sufficiency of collateral remains at the discretion of the bank. In addition, both the Small Business Lending Fund and most banks do not operate under start-up programs. An enterprise must exist and conduct economic activity for at least six months in order to receive a loan from a bank and, accordingly, a guarantee from the Fund (more on the conditions for granting loans below).

State support for small businesses gives banks the opportunity to increase the volume of lending to small businesses that has fallen recently. If earlier this direction was developed with the goal of gaining market share, now it is becoming profitable in terms of profits and risks - the state has made it clear that it will not abandon small businesses and will help refinance loans.

Small and medium-sized businesses in modern conditions are one of the most profitable target segments for Russian banks. There are several reasons for this increased interest:

lower level of competition in this sector compared to servicing large businesses;

more “market” nature of the relationship with the client, less risk of using “administrative resources” when resolving disputes;

the absence of a “buyer’s” market, respectively, the possibility of receiving higher interest rates and commissions (with increased risk, of course);

the presence of both state and international (EBRD, International Finance Corporation) programs for refinancing loans to small and medium-sized businesses - an opportunity for banks to attract relatively cheap financial resources;

the ability to sell mass banking products, which reduces costs and leads to increased efficiency of the banking business.

But potentially high returns are also associated with high real risks. Of course, lending to small and medium-sized businesses is much riskier compared to large businesses. And the reasons are not only the incomparable financial power of enterprises in these sectors. Many risks are associated with the specifics of the activities of small and medium-sized businesses in Russia. One of the most obvious risks is associated with the lack of transparency (unreliability) of financial reporting, the use of tax optimization schemes, etc., in other words, the inability to assess the true financial position of an enterprise according to official reports, while standard credit analysis techniques used to assess potential borrowers , either do not work or give a distorted picture. The reason for this provision is the desire of enterprises to avoid taxation. Tax avoidance or a significant reduction in the tax burden has been seen as a competitive advantage since the 90s, so the use of “special accounting” and “tax-saving technologies” remains widespread among small and medium-sized businesses. Of course, the scale of evasion and the degree of aggressiveness in tax optimization are currently much lower compared to the 90s. Moreover, the demand for loans and the spread of bank lending played a significant role. What are the most common problems with financial reporting of small Russian enterprises:

insignificant amount of the enterprise's own capital;

non-transparent ownership structure of the company;

the presence of fictitious borrowed capital or debt not reflected in the balance sheet;

concealment of “losses” in the balance sheet asset or lack of reflection of real assets in the balance sheet;

fictitious expenses of an enterprise designed to reduce the tax base;

the presence of “off-balance sheet” obligations of the enterprise that are not reflected in the financial statements;

lack of standard reporting forms (balance sheet, profit and loss account, etc.) in the case of using the simplified taxation system (STS).

The critical role of banks in the development of small businesses is explained by the difficulty for small and medium-sized enterprises to obtain financial resources from other sources. At the present stage in Russia there are more than 5 million small and medium-sized enterprises, the bulk of which operate in such industries as wholesale and retail trade, the service sector, construction, and manufacturing.

This segment has a huge growth potential, and lending volumes to small and medium-sized businesses will grow. The growth of the SME lending market is a direct consequence of the general stabilization of the situation in the economy, increasing demand for loans, as well as increasing access of SME representatives to borrowed funds.