JP2002169958A - System and method for credit evaluation, and storage medium with credit evaluation program stored therein - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a credit evaluation system which is capable of evaluating debtors having mutually different kinds of businesses and scales with a unified reference. SOLUTION: The credit evaluation system executes credit evaluation of the debtors on the basis of a unified evaluation system. The system is provided with an evaluation segmentation estimating data registration part 11 which registers an individual evaluation model and calculate an incorrect decision rate in the individual evaluation modeling information registered therewith to set evaluation segmentation on the basis of the incorrect decision rate, a unified evaluation segmentation estimating expression determination part 12 which registers a model to become the unified evaluation system and calculates the set evaluation segmentation and a transformed variable to convert into a unified evaluation model by an evaluation segmentation estimating business data registration part 11, and a debtor evaluation processing part 13 which converts data on an individual evaluation model of the debtor that is input for use of the credit evaluation of a specified debtor on the basis of transformed variable calculated by a transformed variable calculating means to execute evaluation of the unified evaluation model.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a credit evaluation system, a credit evaluation method, and a storage medium storing a credit evaluation program, and more particularly, to a credit evaluation system, a credit evaluation method, and a credit evaluation system for a debtor based on a unified evaluation system. The present invention relates to a storage medium storing a credit evaluation program.

[0002]

2. Description of the Related Art A system for evaluating credit of a debtor has been proposed for credit management, investment management, and measurement of credit risk of asset-backed securities. At present, in a conventional system proposed or realized, a plurality of models divided by industry or scale exist alone. That is, a system and a method for evaluating the credit risk of large enterprises and small and medium enterprises based on a unified standard have not been established. In addition, there is no established system and method for evaluating credit risk based on a unified standard that integrates the attributes of borrowers such as the type of industry. Furthermore, there is no established system and method that can evaluate credit risk on a unified basis between indirect finance and direct finance. Therefore, evaluation results performed based on a plurality of models divided by industry and by size could not be compared with each other when the models were different.

[0003]

As described above, in the prior art, a plurality of models divided according to the type of business or scale exist independently, and are unified across the boundaries of the type of business and scale. However, there was no credit evaluation system, so that there was a problem that it was not possible to compare the credits of companies with different types of businesses and sizes.

[0004] Accordingly, an object of the present invention is to provide a credit evaluation system capable of evaluating obligors having different types of business, scales and the like based on a unified standard.

[0005]

A credit evaluation system according to a first aspect of the present invention is a credit evaluation system for executing a credit evaluation of a debtor based on a unified evaluation system, and registers information relating to individual evaluation models. An individual evaluation model registering means (for example, the evaluation section estimation data registering unit 11 in FIG. 1) and an individual evaluation model information registered by the individual evaluation model registering means calculate an erroneous judgment rate and evaluate based on the erroneous judgment rate. Evaluation section setting means (for example, evaluation section estimation data registration unit 11 in FIG. 1) for setting a section, and unified evaluation model registration means (for example, the unified evaluation section estimation formula in FIG. 1) for registering a model to be a unified evaluation system The determination unit 12) and the unified evaluation registered by the unified evaluation model registration unit by the evaluation division set by the evaluation division setting unit. A conversion variable calculating means (for example, the unified evaluation category estimation formula determination unit 12 in FIG. 1) for calculating a conversion variable to be converted into an evaluation category of the model; Obligor evaluation means for converting the data on the individual evaluation model based on the conversion variables calculated by the conversion variable calculation means and executing the evaluation on the unified evaluation model (for example, the obligor evaluation processing unit 13 in FIG. 1) ). This allows
Obligors with different industries and sizes can be evaluated based on uniform criteria.

[0006] A credit evaluation system according to a second invention comprises:
In the first invention, the evaluation category setting means calculates an erroneous judgment rate based on the default erroneous judgment rate and the erroneous survival judgment rate, and weights the default erroneous judgment rate and the erroneous survival judgment rate. Thereby, the accuracy of evaluation can be improved.

[0007] A credit evaluation system according to a third aspect of the present invention comprises:
In the first invention, the evaluation category setting means calculates an erroneous judgment rate based on the non-investment grade misjudgment rate and the investment grade misjudgment rate, and weights the non-investment grade misjudgment rate and the investment grade misjudgment rate. Things. Thereby, the accuracy of evaluation can be improved.

[0008] A credit evaluation system according to a fourth invention comprises:
In the first invention, the evaluation category setting means sets an evaluation category with a minimum point in the calculated erroneous determination rate as a section break. This allows
The accuracy of evaluation can be improved.

[0009] A credit evaluation system according to a fifth invention is characterized in that:
In the fourth invention, when the minimum point of the calculated erroneous determination rates is unclear, the evaluation division setting means uses the information of the division points other than the division to be set to perform the division setting. This is to derive the breakpoint to be tried. This makes it possible to derive a breakpoint even when the minimum false determination rate is unclear.

[0010] A credit evaluation system according to a sixth invention comprises:
In the fifth invention, the evaluation division setting means derives a division point to be set by using information on points on both sides of the division to be set. As a result, even when the minimum erroneous determination rate is unclear, the breakpoint can be accurately derived.

[0011] A credit evaluation system according to a seventh invention comprises:
In the fifth invention, the evaluation division setting means derives a division point to be set by using information on a plurality of points on one side of the division to be set. As a result, even when the minimum erroneous determination rate is unclear, the breakpoint can be accurately derived.

[0012] A credit evaluation system according to an eighth aspect of the present invention comprises:
In the fifth invention, information on a point other than the partition to be set as a partition is set as a parameter other than the erroneous determination rate. As a result, even when the minimum misjudgment rate is unclear, the breakpoint can be accurately derived according to the situation.

[0013] A credit evaluation system according to a ninth invention is characterized in that:
In the eighth invention, the capture rate is used as a parameter other than the erroneous determination rate. As a result, even when the minimum misjudgment rate is unclear, the breakpoint can be accurately derived according to the situation.

A credit evaluation system according to a tenth aspect of the present invention is the credit evaluation system according to the first aspect, further comprising a future outlook information input means for inputting information relating to the future outlook, and reflecting the future outlook information in the conversion variable calculating means. The conversion variable is calculated by using Thereby, the accuracy of the evaluation can be further improved.

[0015] The credit evaluation system according to an eleventh aspect of the present invention is the credit evaluation system according to the first aspect, further comprising a future outlook information input means for inputting information on a future outlook, and reflecting the future outlook information in the obligor evaluation means. The borrower's evaluation was carried out. Thereby, the accuracy of the evaluation can be further improved.

A credit evaluation method according to a twelfth invention is a credit evaluation method for executing a credit evaluation of a debtor based on a unified evaluation system, wherein an individual evaluation model registration step for registering information relating to individual evaluation models. Calculating an erroneous judgment rate in the registered individual evaluation model information and setting an evaluation classification based on the erroneous judgment rate; a unified evaluation model registration step of registering a model to be a unified evaluation system; A conversion variable calculating step of calculating a conversion variable for converting the evaluation category set in the category setting step into an evaluation category of the unified evaluation model registered in the unified evaluation model registration step; and The input data on the individual valuation model of the debtor is converted to the conversion variable calculated in the conversion variable calculation step. And convert Zui, those having a debtor evaluation step for performing evaluation on the uniform evaluation model. As a result, it is possible to evaluate obligors of different types and sizes based on a uniform standard.

The storage medium storing the credit evaluation evaluation program according to the thirteenth invention is a storage medium storing a credit evaluation program for executing credit evaluation of a debtor based on a unified evaluation system. The program includes, in a computer, an individual evaluation model registration step of registering information on an individual evaluation model, and an evaluation category setting for calculating an erroneous determination rate in the registered individual evaluation model information and setting an evaluation category based on the erroneous determination rate. Step, a unified evaluation model registration step of registering a model to be a unified evaluation system, and conversion for converting the evaluation division set by the evaluation division setting step into an evaluation division of the unified evaluation model registered by the unified evaluation model registration step. A conversion variable calculation step for calculating the variables and an input for the credit evaluation of a specific debtor Has been at the data on the debtor's individual evaluation model that executing the debtor evaluation step of performing an evaluation on the converted based on the calculated conversion variables the uniform evaluation model by the conversion variable calculation step. As a result, it is possible to evaluate obligors of different types and sizes based on a uniform standard.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention relates to, for example,
This is a credit evaluation system for debtors, characterized by having a unified evaluation system that can be used even when the types of finance such as indirect finance and direct finance are different. If the debtor's default rate and credit spread, etc., measured using the evaluation categories of this unified valuation system are used as credit risk benchmarks, the credit risk of various financial transactions of indirect financing and direct financing can be measured. The credit evaluation in the credit evaluation system according to the present invention is a concept including a rating, a score, ranking, and the like.

FIG. 1 is a system configuration diagram of a credit evaluation system according to the present invention. As shown in FIG. 1, this credit evaluation system basically has a server 1 of an evaluation service provider, a memory 2, and a terminal 3 of the evaluation service provider. Terminal 4
1 is connected.

The server 1 includes an evaluation section estimation data registration section 11, a unified evaluation section estimation formula determination section 12, a debtor evaluation processing section 13, and a data totaling analysis section 14. Memory 2
Is a debtor file 21, an evaluation method control file 22,
A unified evaluation division estimation formula file 23 and a debtor evaluation result file 24 are provided.

Evaluation section estimation data registration section 1 of server 1
The basic function of the unit 1 and the unified evaluation division estimation expression determination unit 12 is to create a unified evaluation model. In the present invention, creation of this unified evaluation model plays an important role.

The debtor evaluation processing unit 13 has a function of evaluating the debtor using the created unified evaluation model.

The data tabulation / analysis section 14 has a function of tabulating various data with the terminal 3 of the evaluation service provider.

The user terminal 4 is, for example, a personal computer (PC) and has at least a communication function for exchanging information with the server 1 via the communication network 5. The communication network 5 is a communication network such as the Internet or a public network.

FIG. 2 is a system configuration diagram showing a credit evaluation system according to the present invention, particularly showing a system in a mode different from that of the system shown in FIG. In FIG. 2, components denoted by the same reference numerals as those in FIG. 1 indicate the same or corresponding parts as the components illustrated in FIG. 1, and description thereof will be omitted. In the credit evaluation system of FIG. 2, the terminal 42 of the user 2 is directly connected to the server. However, the present invention is not limited to the case where the terminal 42 and the server are different computers, and may be the same computer or different computers. That is, a program for realizing the function of the server 1 may be installed in the terminal 42.

FIG. 3 shows an example of construction of an evaluation system of the credit evaluation system according to the present invention. As shown in FIG.
An external rating 301 is adopted for A to B-, and an internal rating 302 is adopted for CCC to D. Therefore, the evaluation categories 303 are 20 categories from the AAA to the bankruptcy destination. This is executed by the credit evaluation system in a process 304 to create a unified evaluation system 305. The evaluation categories in the unified evaluation system 305 are 20 categories from aaa to d. In this system, in terms of direct finance, aaa to bbb-
+ To d means non-investment grade. Also, in indirect financing, it is assumed that aaa to b-
This means that cc to d are abnormal destinations. This figure 3
As shown in the above, the credit rating of the debtor based on the unified valuation system enables measurement of the credit risk of all debtors for indirect or direct finance.

FIG. 4 is a flowchart showing a processing flow for registering data for estimating the evaluation category. First, registration of evaluation control data is executed (step S401). The registration is performed by the evaluation service provider using its own terminal 3 using input means such as a keyboard. The input evaluation control data is stored in the evaluation section estimation data registration unit 1 of the server 1.
1 is stored in the evaluation method control file 22. Specifically, the evaluation formulas of a plurality of models that have been divided according to the type of business or the size that exist conventionally are registered. For example, the evaluation formula itself can be used in an individual bank or incorporated in commercially available software. In addition, the correction method is registered. Here, the compensation method means, for example, that a model of three ranks is desired, but when only two ranks exist,
It is a method of how to rank three.
Then, classification categories of individual evaluation formulas such as a business category and a size category (for example, for large enterprises and small and medium enterprises) are executed.

Evaluation categories are set using internal ratings such as external ratings and debtor categories. For example, investment grade zones include AAA, AA +, AA, AA-, A +, A, A-, BBB +, BBB and B
Set 10 categories of BB-. In addition, ten categories of BB +, BB, BB-, B +, B, B-, CCC, CC, C and D will be set as non-investment grade zones. In the case of low-ranking categories such as CCC, CC, C, and D, where the number of external ratings is small, in addition to the external ratings,
You may set the category of internal evaluation, such as a substantially bankrupt borrower or a bankrupt borrower.

Registration of evaluation control data (step S40)
1) is executed on the screen shown in FIG. 13 and the screen shown in FIG. FIG. 13 shows an example of a data registration screen for estimating the evaluation category.
Shown in In the screen example shown in FIG. 13, an input column 1301 for the classification category of the individual evaluation formula, an input column 1302 for the individual evaluation formula, and an input column 1303 for the compensation method are provided. In the input field 1301 for the classification of the individual evaluation formula, the applicable range of the individual evaluation formula is input. For example, a business type, a scale, and the like are input. In an individual evaluation expression input field 1302, a model type, an explanatory variable, and its coefficient are input. The model type is, for example, multiple regression, logit, probit, or the like. For example, “ROA” · “0.05” and “debt ratio” · “−0.21” are input as explanatory variables and their coefficients. Then, an adjustment variable and an adjustment method are input in the input field 1303 of the adjustment method.
The compensation variable is input as, for example, “default capture rate”, “survival capture rate”, or the like. The correction method is, for example, "linear interpolation",
Enter "spline function" or the like. The input is executed at least for the screen shown in FIG. 13 for each individual evaluation expression.

FIG. 14 is an example of an input screen for future prospects. In the screen example shown in FIG. 14, the default number of cases for each type of industry is input as the expected number of cases. For example, "300 types" are input for "A type industry", and "200" numbers are input for "B type industry". The information input on this screen is used to bias the unified evaluation as described later.

After registering the evaluation method control data, the evaluation section estimation data registration section 11 executes a check for required input items (step S402). As a result of the check, when some of the required input items are not input or when there is an error in the input method, the evaluation section estimation data registration unit 11
A message prompting the evaluation service provider to execute registration of the control data of the evaluation method again (step S401) is displayed on the display of the terminal 3. The evaluation service provider recognizes this message and re-registers the control data of the evaluation method (step S401).

On the other hand, as a result of the check, when it is determined that the input for the required input item is appropriate, the evaluation section estimation data registration unit 11 further sets the sample data of the debtor data as sample data necessary for creating a model. A registration process is performed (step S403). Specifically, external ratings,
Information on evaluation categories such as debtor classification, accounting information, qualitative information, market information, and classification category information of individual evaluation formulas such as business category are registered. The classification information is needed to connect the model input in step S401 and the data input in S403.

FIG. 15 shows a registration screen of the debtor data.
The screen shown in FIG. 15 includes an input field 1501 for a classification category, an input field 1502 for an evaluation category of a debtor, and an input field 1503 for debtor data. In the input field 1501 for the classification, the applicable range of the individual evaluation formula is input. For example, "business type" and "scale" are input. In the input field 1502 of the debtor's evaluation category, the evaluation category is input. For example, "external rating", "obligor | AAA", and the like. In the input field 1503 of debtor data, accounting information, qualitative information, and market information are input. For example, assets, liabilities, and capital are input as accounting information. As the qualitative information, the series, parent company, unrealized profit and loss, etc. are input. As market information,
Enter market capitalization, bond spreads, etc. Note that the input is executed in debtor units, and, for example, sample data for 1000 companies is input. The assembling with the individual evaluation formula is executed based on the classification.

Then, the evaluation section estimation data registration section 11
Checks the registration of these debtor data (step S404). As a result of the check, when a part of the data to be registered is not input or when there is an error in the input method, the evaluation division estimation data registration unit 11 registers the debtor data again with the evaluation service provider. A message prompting execution of the process (step S403) is displayed on the display of the terminal 3. The evaluation service provider recognizes this message and registers the debtor data again (step S403). On the other hand, when it is determined that the registration of the debtor data is appropriate, the evaluation section estimation data registration unit 11 ends the evaluation section estimation data registration processing.

Next, the process of determining the unified evaluation section estimation formula will be described with reference to the flowchart of FIG. First,
Execute the process of setting the unified evaluation category (step S50)
1). This processing is executed by the unified evaluation division estimation formula determination unit 12 of the server 1. The process of setting the unified evaluation category is a characteristic part of the present invention, and will be described later in detail.

Then, similarly, the unified evaluation division estimation formula determination unit 12 determines whether or not it is necessary to perform a correction process on the set unified evaluation division (step S50).
2). If it is determined that the correction process needs to be performed, the unified evaluation section estimation formula determination unit 12 executes the unified evaluation section correction process (step S503). This correction processing will also be described later in detail. And, after this correction processing,
The unified evaluation section estimation formula determination unit 12 executes a process of storing the unified evaluation section estimation equation (step S504). When it is determined in step S502 that the correction processing is not necessary, the unified evaluation division estimation expression determination unit 12 executes the storage processing of the unified evaluation division estimation expression without performing the correction processing (step S504).

The setting of the unified evaluation section estimation formula (step S501) will be described in more detail. The determination of the unified evaluation section estimation formula includes an individual cutting point deriving process and a unified cutting point deriving process. Here, the cutting point refers to a point serving as a section break. In addition, individual cutting points are classified by industry,
This is a point that serves as a partition between sections in models individually set according to scale or the like. The unified cutting point is a point that serves as a section break in the unified model. The details will be described below.

(1) Processing for Deriving Individual Cutting Points FIGS. 7 and 8 are diagrams for explaining the processing for deriving individual cutting points. In the figure, the X-axis direction indicates the direction in which the business situation of the debtor deteriorates. The N-axis shows the number of debtors, and also shows the division of sections. Also, d1 shown by a solid line
Indicates the distribution of the total default number. This default number is the number of companies that went bankrupt within a predetermined period, for example, one year, in the sample. S1 shown by a dotted line indicates the distribution of the total surviving number. The surviving number is, for example, the number of companies in the sample that have not gone bankrupt for a predetermined period, for example, one year. D2 indicated by the diagonally shaded area on the left indicates the number that could not be determined to be the default out of the total default number. Also,
S2 indicated by a diagonally oblique line on the right indicates the number of the total surviving numbers that could not be determined as surviving. In general, as shown in the figure, the total number of survivors is larger than the total default number.

First, a search is made for a point where the erroneous determination rate is the minimum for each of the sections set as the objective. Specifically, X1, X2, X3, X4,... X10 shown in FIGS.
The value of the evaluation formula at each point of is calculated. FIG. 7 illustrates a case where X5 is a cutting point, and FIG. 8 illustrates a case where X3 is a cutting point. Then, a point at which the value of the obtained evaluation formula becomes smallest is obtained. The point of the minimum erroneous determination rate is defined as a cutting point. In the search method, for example, a point at which the value of the following evaluation equation of the erroneous determination rate is minimized, and a weight can be set by the utility α. When there is a zone where the misjudgment rate is the smallest, it is conceivable to adopt the average value or the median of the zones as the minimum misjudgment rate point. The misjudgment rate is, for example, 7 to 15%.

The following two evaluation expressions can be given as examples of the evaluation expression of the erroneous judgment rate. It is also possible to apply evaluation expressions other than these evaluation expressions.

Default evaluation formula = α (d2 / d1) +
(1-α) (s2 / s1)

Non-investment grade evaluation formula = α (n2 / n1) +
(1-α) (i2 / i1)

Here, as described above, d1 is the total default number, s1 is the total number of survivors, d2 is the number of the total default numbers that could not be determined as default, and s2 was the total number of survivors that could not be determined as alive. Is a number. Also, n1 is the total non-investment grade number, i1 is the total investment grade number, n2 is the total number of non-investment grade numbers that could not be judged as non-investment grade, and i2: the total investment grade number cannot be judged as investment grade. Is the weight.

D2 / d1 is a default erroneous judgment rate,
s2 / s1 is the survival misjudgment rate, n2 / n1 is the non-investment grade misjudgment rate, and i2 / i1 is the investment grade misjudgment rate. Here, α is a coefficient that determines which of the default misjudgment rate (d2 / d1) and the survival misjudgment rate (s2 / s1) is important. When the default misjudgment rate (d2 / d1) is important, , Α, and a relatively small value is input to α when emphasizing the survival error determination rate (s2 / s1).

The point at which the value of these evaluation expressions is minimized is determined in each model. Normally, the values of these evaluation expressions have the minimum value, but may be zones.

In this way, the division can be determined for each individual model. The classification for each model is unified in the following (2).

(2) Derivation process of unified cutting point A unified evaluation scale (master measure) is set which classifies various evaluation models for each type of business or scale by using a cutting point such as a default misjudgment rate. It is assumed that an objective variable representing a generally used debtor's creditworthiness and an evaluation formula for each type of industry and each size composed of n explanatory variables such as financial indicators and attributes are known. Next, a temporary evaluation scale serving as a master measure is provided. For example, the following simultaneous equations are used to integrate various types of evaluation models for each type of business and for each scale to set a unified evaluation scale.

For example, B- and D on the master measure
Are assumed to be 0.10 and 0.30. This master measure can use fictitious numerical values or simple model evaluation results that can evaluate all industries.

The model expression of credit evaluation by industry or scale is expressed as follows: the base e of the natural logarithm to the power of x / (1 + the base e of the natural logarithm x
Power). This expression is a general expression, and other expressions are also applicable. Note that x is an explanatory variable composed of n constants and n explanatory variables such as financial indicators and attributes. For example, it can be represented by x = ax1 + bx2. Here, a and b are coefficients, x1 is a capital ratio, and x2 is ROA.

Here, when the explanatory variables xB− and xD at the minimum erroneous judgment point, which are the intercepts of B− and D, are given, unknown variables are set as a constant μ and a coefficient β, and the following simultaneous equations are derived.

0.10 = μ + β · xB of base e of natural logarithm
−power / (1 + the power of natural logarithm base e to μ + β · xB−)

0.30 = μ + β · xD of the base e of the natural logarithm
Power / (1 + the base e of natural logarithm to the power of μ + β · xD)

By solving these simultaneous equations, μ and β can be obtained. If μ and β are known, a conversion formula from an individual model such as a specific business type to a master measure can be obtained. In this case, the value of μ and β is obtained between B− and D, but not limited thereto.

In other words, the above simultaneous equations when μ is set to 0 and β is set to 1 are model equations in each individual model. Therefore, μ and β are different in each model. Further, even in one model, the division is different for each division. For example, in the case of 20 divisions, the division is 19 divisions.
Therefore, there are 19 sets of μ and β in the model. These μ and β function as compensation coefficients for adjusting to a master measure from an individual model.

As described above, the variable conversion is performed by using the divided AA.
A, AA +, AA, AA-, A +, A, A-, BBB +, BBB, BBB-, BB +,
Classification of 19 sections of BB, BB-, B +, B, B-, CCC, CC, C, and D by applying to various evaluation models by industry and size and matching the position on the master measure The scale of unified evaluation to be done is completed.

It should be noted that various model formulas other than the above-described model formulas can be used. For example, y = ax1 + bx
An Altman model that can be represented by 2...

Next, compensation processing (FIG. 5: step S503)
Will be described. If there is no point or a plurality of points at which the erroneous judgment rate is the minimum, it is necessary to make correction. There are two methods: a method of obtaining an unknown point by using the values of μ and β for other cutting points, and a method of using an evaluation expression such as a capture rate without using μ and β.

First, a method of finding an unknown point by using the values of μ and β for other cutting points will be described. In this method, information on the unknown point is extracted from information on the cutting points on both sides of the unknown point (interpolation method), and information on the unknown point is extracted from information on one or more cutting points. (Extrapolation method). In addition, a method of complementing by adopting a point near an arbitrarily determined value can be considered.

FIG. 10 is a diagram for explaining a correction method using the interpolation method. The Y axis indicates the erroneous determination rate. In the figure, the minimum misjudgment rate point of y1 is a partition between A + and A, y
The minimum misjudgment rate point of 2 is a segment between A and A-, and the minimum misjudgment rate point of y3 is a segment between A- and BBB +. In this example, the minimum misjudgment rate points of y1 and y3 are clear, but y2 has a so-called zone shape, and the minimum misjudgment rate points are unknown. At this time, the minimum misjudgment rate point of y2 is determined so that the distances from the minimum misjudgment rate points of y1 and y3 are equal.

FIG. 11 is a diagram for explaining a compensation method using the extrapolation method. In this example, the minimum misjudgment rate points of y1 and y2 are clear, but y2 has a plurality of so-called minimum misjudgment rate points and is unclear. At this time, the minimum misjudgment rate point of y2 is obtained on an extension of the minimum misjudgment rate point of y1 and y2. Various methods can be considered for determining which point on the extension line is the minimum false determination rate point. For example, a point on the route along the line separated from y2 by the distance between y1 and y2 is set as the minimum false determination rate.
In the examples of FIGS. 10 and 11, the minimum false determination rate is clearly 2
Although the method of deriving one point that is unclear from the points has been described, the present invention is not limited to this, and one or more points that are unclear from three or more points where the minimum misjudgment rate is clear may be obtained.

Next, a method of using evaluation expressions of different measures such as a capture rate without using μ and β will be described. Using the evaluation formula of a different measure such as a capture rate, a break at which the erroneous determination rate is minimized is determined.

FIG. 12 is a diagram for explaining a method of using evaluation expressions of different measures such as a capture rate. In FIG. 12, the horizontal axis is the value of μ or β in the master measure, and the vertical axis is the value in a different measure such as the capture rate. In this example, the minimum misclassification rate of y1 and y3 is clear, but the minimum misclassification rate of y2 is not clear. In the example shown in FIG. 10, the value is obtained from the minimum misclassification rate of y1 and y3.
In this example, a value corresponding to y2 is obtained by interpolation of values in other measures corresponding to y1 and y3, and μ and β of y2 are obtained from these values. In this example, the value is obtained by the interpolation method. However, as shown in FIG.

For example, the following expressions are used as evaluation expressions for the capture ratio and the like. Default capture rate = 1- (d2 / d1) Survival capture rate = 1- (s2 / s1) Non-investment grade capture rate = 1- (n2 / n1) Investment grade capture rate = 1- (i2 / i1) ) Cumulative default rate = (d1-d2) / (s2 + d1-d)
2) Cumulative non-investment grade rate = (n1-n2) / (i2 + n1-n)
2) Default rate within category = kd1 / k1 Non-investment grade rate within category = kn1 / k1

Here, d1 is the total number of defaults, s1 is the total number of survivors, d2 is the number of the total defaults that could not be determined as default, s2 is the number of the total number of survivors that could not be determined as alive, and n1 is The total number of non-investment grades, i1 is the total number of investment grades, n2 is the total number of non-investment grades that could not be judged as non-investment grade, i2 is the number of total investment grades that could not be judged as investment grade, k1 Is the total number belonging to the category, kd1
Is the total number of defaults belonging to the category, and kn1 is the total number of non-investment grades belonging to the category.

Further, the following values can be used. Distance from the cutting point of the model before integration such as by industry =
X-Xmin

Here, x is a calculation result of n explanatory variables such as constants, financial indices and attributes in a model before integration by industry or the like, and Xmin is a cutting point of the model before integration by industry or the like.

Further, the unified evaluation can be biased based on the future default prospects of the whole or each industry. Compensation is performed using the following conversion formula. (Evaluation after reflecting future prospects) = γ (Evaluation before reflecting future prospects)

Here, γ is a weight, for example, an industry default contribution index, an industry default rate index,
An industry default ratio index can be used. Here, the industry-specific default contribution index is 1- (predicted value of the number of default cases by industry) / (predicted value of the total number of default cases)
Can be represented by The industry default rate index is 1-
(The number of defaults by industry) / (the number of companies by industry). The industry default ratio index is {1- (predicted number of defaults by industry) / (number of companies by industry)} / {1-
(Estimated value of the total number of default cases) / (Number of all companies)}. As the evaluation before reflecting the future prospect, for example, μ + βx can be used.

Next, a description will be given, with reference to the flowchart of FIG. 6, of a debtor evaluation process using a unified evaluation system.
First, the debtor evaluation processing unit 13 of the server 1 executes a registration process of debtor data used as an explanatory variable (step S601). The debtor data is stored in the terminal 41, 42 by the user.
Enter using. The input debtor data is transmitted to the communication network 5
Is transmitted to the server 1 via. The server 1 stores the debtor data in the debtor file 21 by the debtor evaluation processing unit 13. Note that the debtor data includes accounting information, qualitative information, market information, and the like, and further includes a classification category of an individual evaluation formula such as a business category.

FIGS. 14 and 16 show examples of registration screens for debtor data. FIG. 14 shows a future prospect input screen, which is the same screen as the screen described in the evaluation method data registration (FIG. 4: step S401). The information input on this screen is used to bias the unified evaluation.

FIG. 16 is an example of a screen for registering debtor data, in which an input column 1601 for classification and an input column 1602 for debtor data are provided. In the input section 1601 of the classification, the applicable range of the individual evaluation formula is input. For example, "business type" and "scale" are input. Input field 16 of debtor data
In 02, account information, qualitative information, and market information are input. For example, assets, liabilities, and capital are input as accounting information. As the qualitative information, the series, parent company, unrealized profit and loss, etc. are input. As market information, stock market capitalization, bond spread, and the like are input. It should be noted that the input is performed in debtor units. The assembling with the individual evaluation formula is executed based on the classification.

Then, the debtor evaluation processor 13 executes a registration check of the debtor data (step S60).
2). As a result of the check, if there is any omission in the debtor data or if there is a defect in the input method, the debtor evaluation processing unit 13 sets the user terminals 41 and 42
Via the communication network 5 for prompting re-registration. The message is displayed on the displays of the terminals 41 and 42. The user executes the re-registration of the debtor data in accordance with the message (step S601).
On the other hand, if there is no problem with the entered debtor data,
The debtor evaluation processing unit 13 executes a process of setting the evaluation category of the debtor according to the unified evaluation category estimation formula (step S60).
3). That is, a classification such as A or AA is assigned to a specific debtor input by the user.

The setting of the borrower's evaluation category using the unified evaluation category estimation formula will be described in further detail. First, the rating of the specific debtor in the individual model is calculated. Since the coefficients μ and β that convert the rating of each obligor in this individual model into the rating in the unified model have already been obtained, use these values to determine which rating this specific obligor belongs to in the unified model. be able to.

For example, the determination of AAA is made by AAA> (AAA and AA +
The unified cutting point). Also, the judgment of AA + is
It is executed based on the following formula. (Unified cutting point of AAA and AA +)
≧ AA +> (Unified cutting point of AA + and AA)

The determination of C is executed based on the following equation. (Unified cutting point of CC and C) ≧ C> (Unified cutting point of C and D) The determination of D is performed according to D ≦ (Unified cutting point of C and D).

With the credit evaluation system described above, it is possible to measure the credit risk of all the debtors related to indirect finance or direct finance by the credit assessment of the debtors based on the unified evaluation system. For example, it is possible to measure credit risk of asset-backed securities such as credit management (management of loans, accounts receivable and investment, a credit engine for B2B net transactions), investment management (bond rating, credit derivatives), and CLO.

Note that programs for implementing various functions provided in the server 1 for executing the above-described credit evaluation system may be stored in a storage medium. Here, for example, a floppy (registered trademark) disk, CD
-A storage medium or storage device such as a ROM, a magneto-optical disk, a semiconductor memory, and a hard disk is included.

[0078]

According to the credit evaluation system of the present invention, it is possible to evaluate obligors having different types of business, different scales, etc. based on a uniform standard.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of a credit evaluation system according to the present invention.

FIG. 2 is a system configuration diagram of a credit evaluation system according to the present invention.

FIG. 3 is a diagram showing an example of construction of an evaluation system of the credit evaluation system according to the present invention.

FIG. 4 is a flowchart showing a part of a processing flow of the credit evaluation system according to the present invention.

FIG. 5 is a flowchart showing a part of the processing flow of the credit evaluation system according to the present invention.

FIG. 6 is a flowchart showing a part of a processing flow of the credit evaluation system according to the present invention.

FIG. 7 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 8 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 9 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 10 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 11 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 12 is a diagram for explaining processing of the credit evaluation system according to the present invention.

FIG. 13 is a screen display example in the credit evaluation system according to the present invention.

FIG. 14 is a screen display example in the credit evaluation system according to the present invention.

FIG. 15 is a screen display example in the credit evaluation system according to the present invention.

FIG. 16 is a screen display example in the credit evaluation system according to the present invention.

[Explanation of symbols]

Reference Signs List 1 server 2 memory 3 terminal of evaluation service provider 5 communication network 11 evaluation section estimation data registration section 12 unified evaluation section estimation formula determination section 13 debtor evaluation processing section 21 debtor file 22 evaluation method control file 23 unified evaluation section estimation Formula file 24 Obligor evaluation result file 41, 42 User terminal

 ──────────────────────────────────────────────────続 き Continuation of the front page (71) Applicant 592259978 Daiichi Kangyo Bank, Ltd. 1-1-5 Uchisaiwaicho, Chiyoda-ku, Tokyo (72) Inventor Shin Miyoshi 3-8-1TT, Nihonbashicho, Nihonbashi, Chuo-ku, Tokyo − 2 Building Co., Ltd. Financial Engineering Research Institute (72) Inventor Takehiko Yasukawa 3-8-1TT, Nihonbashi Ningyocho, Chuo-ku, Tokyo 2 Building Rating Information Center (72) Inventor Tsutomu Tamura Nihonbashi Ningyo, Chuo-ku, Tokyo 3-8-1 TT-2 Machi Building Rating Investment Information Center (72) Inventor Shinka Aoki 1-1-5 Uchisaiwaicho, Chiyoda-ku, Tokyo Co., Ltd. Daiichi Kangyo Bank Co., Ltd. (72) Inventor Yoji Aoki Tokyo 1-5-5 Otemachi, Chiyoda-ku Inside Fuji Bank, Ltd. (72) Inventor: Akitsu Tominaga 1 Marunouchi, Chiyoda-ku, Tokyo Eye No. 3 No. 3 stock company Industrial Bank of Japan in the F-term (reference) 5B049 BB46 CC00 EE00 5B055 CC11

Claims (13)

[Claims] 1. A credit evaluation system for executing a credit evaluation of a debtor based on a unified evaluation system, comprising: an individual evaluation model registration means for registering information relating to an individual evaluation model; An evaluation category setting means for calculating an erroneous determination rate in the registered individual evaluation model information and setting an evaluation category based on the erroneous determination rate; a unified evaluation model registration means for registering a model serving as a unified evaluation system; A conversion variable calculating means for calculating a conversion variable for converting the evaluation classification set by the setting means into an evaluation classification of the unified evaluation model registered by the unified evaluation model registration means; and inputting for a credit evaluation of a specific obligor. The converted data on the individual valuation model of the obligor is converted based on the conversion variable calculated by the conversion variable calculating means, Credit rating system with a debtor evaluation means for performing an evaluation on a uniform evaluation model. 2. The evaluation classification setting means calculates an erroneous judgment rate based on a default erroneous judgment rate and a survival erroneous judgment rate, and weights the default erroneous judgment rate and the survival erroneous judgment rate. Credit rating system. 3. The evaluation classification setting means calculates an erroneous judgment rate based on the non-investment grade erroneous judgment rate and the investment grade erroneous judgment rate, and weights the non-investment grade erroneous judgment rate and the investment gradual misjudgment rate. The credit evaluation system according to claim 1, wherein 4. The credit evaluation system according to claim 1, wherein said evaluation section setting means sets an evaluation section by setting a minimum point in the calculated erroneous determination rate as a section break. 5. The method according to claim 1, wherein, when the minimum point among the calculated erroneous determination rates is unclear, the evaluation division setting means uses the information of the division points other than the division to be set. 5. The credit evaluation system according to claim 4, wherein a break point to be obtained is derived. 6. The method according to claim 5, wherein said evaluation division setting means derives a division point to be set by using information on points on both sides of the division to be set.
The credit rating system described. 7. The method according to claim 5, wherein said evaluation division setting means derives a division point to be set by using information on a plurality of points on one side of the division to be set. Credit rating system. 8. The credit evaluation system according to claim 5, wherein the information of the break points other than the break to be set as a parameter is a parameter other than the misjudgment rate. 9. The credit evaluation system according to claim 8, wherein a capture rate is used as a parameter other than the erroneous determination rate. 10. The credit evaluation system further comprises a future prospect information input means for inputting information about a future prospect, and a conversion variable is calculated by the conversion variable calculating means by reflecting the future prospect information. The credit evaluation system according to claim 1, wherein 11. The credit evaluation system further includes a future prospect information input means for inputting information regarding a future prospect, and the obligor evaluating means executes the evaluation of the debtor by reflecting the future prospect information. The credit evaluation system according to claim 1, wherein: 12. A credit evaluation method for performing credit evaluation of a debtor based on a unified evaluation system, comprising: an individual evaluation model registration step of registering information relating to individual evaluation models; and registered individual evaluation model information. Calculating an erroneous judgment rate and setting an evaluation categorization based on the erroneous judgment rate, a unified evaluation model registration step of registering a model to be a unified evaluation system, and an evaluation categorization set by the evaluation division setting step A conversion variable for calculating a conversion variable for converting the conversion into the evaluation category of the unified evaluation model registered in the unified evaluation model registration step, and an individual evaluation of the specific obligor input for credit evaluation of a specific obligor The data on the model is converted based on the conversion variable calculated in the conversion variable calculation step, and the unified evaluation model is converted. Credit evaluation method with a debtor evaluation step for performing evaluation on Le. 13. A storage medium storing a credit evaluation program for executing credit evaluation of a debtor based on a unified evaluation system, wherein the credit evaluation program registers information on an individual evaluation model in a computer. Registering an individual evaluation model, calculating an erroneous determination rate based on the registered individual evaluation model information, and setting an evaluation category based on the erroneous determination rate, and a unified evaluation for registering a model as a unified evaluation system. A model registration step; a conversion variable calculation step of calculating a conversion variable for converting the evaluation category set in the evaluation category setting step into an evaluation category of the unified evaluation model registered in the unified evaluation model registration step; The data on the individual valuation model of the debtor entered for the credit evaluation of Calculated conversion based on the calculated transition variable by step, the unified evaluation storage medium storing credit evaluation program for executing the debtor evaluation step of performing an evaluation on the model.