JP2006099492A - Loan decision index calculation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To calculate a loan decision index reflecting qualitative data about a company. <P>SOLUTION: A loan decision index management server 1000 accesses a loan customer company management server 2000 to acquire trade attribute data of the loan customer company and retrieves an instruction operation command given by a system user and received from a terminal device 4000 to digitize the trade attribute data. A program performs grouping processing of the loan customer companies based on the digitized data to create a grouping definition table 1160. After accessing the loan decision data management server 3000 and calculating a group index for each group, data about a company considered as a loan customer company are acquired by accessing the loan customer company management server 2000. Based on the data, a company matching a trade form of the company considered as a loan customer company is selected from existing loan customer companies. As to the selected company, a loan decision index is calculated from the group index of the assigned group and the trade attribute data. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a technique for calculating an index for determining whether or not a loan provided by a financial institution is possible using a computer. Among them, in particular, an index for lending to a company is calculated.

  Conventionally, credit judgment and rating data based on financial statements of each borrower company are used as loan judgment indexes. These disclose a technique for calculating credits and ratings according to individual rules established by each bank using a credit risk management system or the like. For example, in Patent Document 1, an index called a customer asset value growth rate is calculated. Moreover, in patent document 2, the parameter | index which shows the creditworthiness of the company used as loan judgment is calculated.

JP 2002-329055 A JP2003-16261

Here, the data handled in the bank loan business includes qualitative evaluation items such as the relationship with the companies directly dealing with each company and the qualities of the manager, which are not shown in the financial statements.
However, collecting the data of these qualitative evaluation items has the following problems.
-It takes time because data collection of qualitative items is carried out personally.
-Judgment criteria are not standardized by customer representatives.
-Evaluation is low compared to quantitative items such as financial statements, and especially small and medium-sized enterprises with poor financial statements are often evaluated as low-cost loans.

In the above-mentioned Patent Documents 1 and 2, the current situation is that there is a problem in quantifying qualitative items that are used for loan determination.
For example, in Patent Document 1, it is a content to calculate an index called a customer asset value growth rate, but the customer asset value growth rate is not a value that takes into account the transaction (customer or trading partner) of the company. Therefore, it is difficult to reflect the actual situation of growth.

  In Patent Document 2, an index indicating the creditworthiness of a company that is a loan decision is calculated. However, the basis for calculating the index is limited to financial data. Nothing is considered. For example, regarding the transaction of the company, the strength of ties between companies in the transaction relationship is calculated as an index taking into account the transaction continuation time as information other than financial data.

  The present invention is a system that quantifies a business relationship, which is a qualitative item that does not appear in financial data, and calculates a loan judgment index using grouping means. The aim is to calculate the loan judgment index that leads to bottom-up evaluation of SMEs by quantifying the business relationships of companies that are usually considered as qualitative items.

  Therefore, in the present invention, regarding the business relationship that is a qualitative item of the loan destination company, the business relationship with each company directly dealing with the loan destination company is quantified. Further, the present invention includes a configuration for quantifying the business relationships of a plurality of loan destination companies in the form of a business partner of the business partner company. With the above quantification, the loan companies are grouped as shown in FIG. 2 by handling precise and realistic attribute data such as the flow of funds among a plurality of companies. It is a feature of the present invention to calculate a loan judgment index as a material for future loan judgment using the grouping means.

  According to the present invention, it is possible to calculate an index that is a more detailed loan determination material than in the past.

  Embodiments of the present invention will be described with reference to the drawings. In the present embodiment, the loan judgment index of the loan destination company that is considering the loan in the future is calculated by the grouping means based on the data of the loan destination company. However, the present embodiment is applicable to other cases. For example, it may be used for investing or measuring the ability of a company. First, FIG. 1 shows a system configuration diagram in the present embodiment. Each computer is connected to each other via a network. Each computer has a processing device such as a memory, a storage device including a hard disk, and a CPU, and the processing device executes data processing in accordance with a program stored in the storage device. However, the loanee company management server 2000 and the loan judgment data management server 3000 in FIG. 1 are not limited to the method of collecting data via the network, but are used by storing data in the index calculation server 1000 in advance. It may be in the form of

Hereinafter, the processing content of this Embodiment is demonstrated.
First, the loan judgment index calculation program 1110 of the loan judgment index calculation server 1000 accesses the loan destination company DB 2100 of the loan destination company management server 2000, and acquires basic data relating to the transactions of the loan destination companies that have previously provided loans. This basic data includes accounts receivable, accounts payable, and business partner data. A transaction relationship definition table 1120 between the loan companies is created based on the basic data regarding the loan companies.

Thereafter, the loan judgment data management server 2000 obtains rating data of the loan destination company, and performs a leveling process of the loan destination company based on the rating data. That is, information having the same rating data (within a predetermined range) is set to the same level, and information for identifying a loan destination company is stored for each level. Furthermore, in order to quantify the business relationship, based on the transaction attribute data, an index indicating the strength of the business relationship (hereinafter referred to as the transaction strength index) and the revenue ratio of each transaction held by the borrower company (hereinafter referred to as the transaction strength index) , Referred to as transaction return). First, in order to calculate the transaction strength index, the system user selects the items to be considered in the transaction attribute data indicated by reference numerals 1132 to 1135 in FIG. 7, and for each of the selected items, FIG. An input of weight setting shown by reference numeral 1137 is accepted. The transaction attribute data items include “transaction duration” and “transaction amount” items as minimum necessary data. In addition to this, for example, “the degree of love of each transaction's profit rate” can be cited. The so-called love degree means that the rate of return arising from the business relationship with Company B as viewed from Company A accounts for 70% of the transaction revenue of Company A, and conversely from the business relationship with Company B as viewed from Company A. Profit rate accounts for 10% of Company A transaction revenue, Company D's revenue from Company C accounts for 40%, Company D's revenue from Company C accounts for 50% This is a way of thinking compared to Case 2. For example, this is expressed as follows:
Case 1: 70 + 10 = 80
Case 2: 40 + 50 = 90
By comparing the above formulas, it is an idea that it is judged that the relationship with which the business relationship is held is stronger, and the business relationship in Case 2 is emphasized.

  In addition, the intensity level 1138 in FIG. 7 indicates which level the numerical value of each item deserves from a relative comparison. In this strength level calculation method, levels are classified into specific stages based on the maximum value, minimum value, and average value based on relative comparison, and the level value is determined by creating the transaction attribute item leveling table 1140 shown in FIG. In the example shown in the drawing, in the case of 0001 → 0002 in the transaction relationship in FIG. 7, the transaction duration is 10 years. Compared with this, when the relationship between the transaction duration and the strength level in FIG. 8 is examined, the transaction duration of 10 to 14 years is positioned as level 8. Therefore, it is determined that the strength level of the transaction duration in FIG. Intensity levels are similarly calculated for reference numerals 1133, 1134, and 1135 in FIG.

The transaction strength index is calculated according to the following (Equation 1) for each transaction relationship by comparing the items considered in the transaction relationship with the weights and level values for the items.
Each transaction strength = W1 x "transaction duration level" + W2 x "transaction amount level" + W3 x "transaction increase rate level" + W4 x "transaction company revenue increase rate" + ... (Equation 1)
As a specific example of calculating the transaction strength index, (Equation 2) is calculated by taking 0001 → 0002 of the transaction relationship in FIG. 7 as an example.
5 x 8 + 4 x 9 + 3 x 9 + 2 x 7 = 117 ... (Equation 2)
From (Equation 2), the transaction strength index from 0001 to 0002 is calculated as 117.

  Next, the transaction revenue ratio is calculated. Based on the revenue ratio data of each transaction, the supplier revenue ratio relationship table 1150 in FIG. 9 is created. That is, what shows how each company has a business relationship with each trading company is created. This is realized by taking the sum of the profit ratio data of each transaction for each business partner.

Further, for the purpose of grouping loan companies, the system user accepts input of setting conditions relating to business relationships as shown in step A12 of FIG.
First, as condition 1, a minimum value is accepted from the system user for the transaction amount and the calculated value of the transaction intensity in each transaction, and the transaction relationship below the minimum value is not considered. In addition to Condition 1, the transaction revenue ratio of the loan company is the highest among the entire loan companies until the arbitrary value received from the system user becomes the minimum value. Select a business partner. Based on the two conditions, the loan destination companies are grouped by selecting a more closely related business relationship. By the grouping means, the business relationship, rating level, and affiliation group of the borrower are defined, and a grouping definition table 1160 is created.

  Based on the grouping definition table 1160 in FIG. 10, a diagram as shown in FIG. 22 grouping result output screen A5000 is displayed on the output screen 5000 for the system user. If the system user determines that there is a problem with the grouping definition with respect to the screen display result, the process returns from step A17 to step A12 shown in FIG. 4, and the processing from step A12 to step A17 is repeatedly executed. The grouping definition can be examined based on the condition set by the system user.

Next, an index for each group is calculated on the basis of the grouping definition in which the grouping is performed. A data item for calculating the group index is received from a user of the system. The items to be selected by users of the system are not limited to “transaction amount within the group” and “revenue amount brought to the bank by the lending company belonging to the group”, but also “increased rate of revenue from the transaction” Examples include “gold”, “equity ratio” and “credit balance”. According to the item selected by the user, data is acquired for each item from the loan determination attribute DB 3100 shown in FIG. Based on the data acquired for each item, the numerical value for each group to which the loan belongs is calculated by accumulating the numerical value of the loan company. For example, it is assumed that a company belonging to group A is a CIF denoted by reference numeral 3110 in FIG. The numerical values of group A at this time are calculated as (Equation 3) and (Equation 4).
Group A transaction amount: 300 + 150 + 80 = 530 (3)
Earnings to Group A banks: 10 + 5 + 1 = 16 (4)
Further, comparison with past data is performed according to the selection item. During a period dating back as past data to be referenced, input information from the system user is accepted. For each item, an increase / decrease rate is calculated for each item by comparing the past data back to the set period and the current data. For example, an example of calculating the increase / decrease rate will be shown taking the group A as an example. The period of the past data to be compared received from the user of the system is the previous period. The rate of increase is calculated as in (Equation 5) from the past transaction amount of Group A and the transaction amount of Group A in the current term calculated from (Equation 3). A table of these results is shown in the group index table 1170, and the transaction amount (increase rate) of the group A is shown in 1172.

Furthermore, after calculating the rate of increase in the amount of revenue for the bank as well as the transaction amount increase / decrease rate, the group index 1174 is calculated from the relative comparison with the economic growth rate during the same period, and the group index table shown in FIG. create. For example, in the case of Group A, the transaction amount increase rate is + 10%, and the revenue increase rate for the bank is + 30%. Calculate the average value of Group A growth rates for various items, compare it with + 5% of economic growth during the same period, and calculate the group index as shown in (Equation 6). (10 + 30) ÷ 2 ÷ 5 = 4 (Equation 6)
As shown in the above (Equation 6), since it shows an increase rate of 4 times compared with the economic growth rate, the index result of × 4 and group A is shown.
Next, a loan judgment index of a company that is considering financing in the future (hereinafter referred to as a loan considering company) is calculated. First, data stored in the loan consideration company DB 2300 shown in FIG. 17 is acquired from the loan company management server 2000 of the system. For example, in the case of CIF 0000 as the loan consideration destination company shown in FIG. 17, the partner company in this company is shown in the partner company name 2307, and the data of transaction amount actual 2308 and transaction duration 2309 of each partner company is shown. get. Calculation of the transaction strength index and the transaction rate of return of the loan company using the actual transaction amount with XX Corporation, which is one of the client companies in CIF 0000, and the numerical value of the transaction duration According to the same procedure, the transaction increase rate 1187 and the revenue increase rate 1188 of the counterparty company in the loan consideration destination company transaction data table 1180 shown in FIG. 18 are calculated, and the transaction strength index 1189 is calculated. Specifically, the level value is determined by comparison with the FIG. 8 level table 1140 for each transaction attribute item. When comparing a CIF 0000 loan-considering company with a business partner XX Co., Ltd., the transaction duration is 5 years. Compared with this, if the relationship between the transaction duration and the strength level in FIG. 8 is examined, the transaction duration of 5 to 9 years is positioned as level 7. Therefore, it is determined that the strength level of the transaction duration in FIG. Intensity levels are similarly calculated for reference numerals 1133, 1134, and 1135 in FIG. By comparing the items to be considered in the business relationship with the weights and level values for the items, the transaction strength index is expressed by (Equation 7) for the business relationship between CIF 0000 and OO Co., Ltd. ).
5 x 7 + 4 x 9 + 3 x 5 + 2 x 5 = 96 (Equation 7)
From the above (Equation 7), the transaction strength index of CIF 0000 and OO Co., Ltd. is calculated as 96 and is shown in the transaction strength index 1189. Based on the data table in FIG. 18, the business relationship of the loan considering company to be considered for grouping is determined according to the business relationship condition setting set in the system user in A12 to A15 in FIG.

  The transaction-related conditions set by the system user shown here are not limited to “transaction amount within the group” and “revenue amount brought to the bank by the lending company belonging to the group”, but “revenue from transactions” ”Increase rate”, “Capital”, “Equity ratio” and “Credit balance”. Each item is determined by a relative comparison between the transaction data of the loan-considering company and the transaction data of each loan company shown in FIG. 7 with respect to the transaction-related condition setting items set by the system user from the data table shown in FIG. The borrower companies that have approximate values are extracted. The corresponding company is selected from the borrower companies in order from the smallest error. At that time, it is also possible to set a condition such that the rating of the loan considering company and the corresponding company is absolutely referenced. For example, when a loan company having a business relationship similar to CIF 0000 is extracted, the rating is absolutely referenced, and as shown in a rating 2303 in FIG. The data on the number of trading companies, the trading strength index, the rate of return of each trading company and the data on the number of trading companies of CIF 0000, trading strength index, and the rate of return of each trading company are standard values. Set and calculate the standard error for each item. Further, an average value of the standard errors is obtained to calculate an error rate. This is indicated by the error rate 1198 in FIG. The companies with the lowest error rate are selected in order, and loan companies such as CIF 0055 and CIF 0012 are extracted as shown in FIG. Based on the result of the selection, a corresponding loan company DB 1190 shown in FIG. 19 is created. The group index of the group to which each selected loan company belongs is obtained from the group index 1174 in FIG. 14, and is shown in the group index 1203 in FIG. Further, the contribution degree for each item to the group is calculated as shown in (Equation 8) using CIF 0055 as an example, and an average value is calculated based on this, as shown in (Equation 9), as shown in FIG. The group contribution index 1204 is calculated.

800 (total transaction amount for CIF 0055) ÷ 2000 (total transaction amount for Group A) × 100 = 40% (Equation 8)
40% (contribution of transaction amount) + 20% (contribution of income to banks) ÷ 2 (number of items) = 30% (Equation 9)
Further, the group ranking 1205 in FIG. 20 is determined based on a relative comparison with the group contribution index of other companies in the group to which each borrower company belongs.

  From the above, the loan judgment index table 1200 shown in FIG. 20 is created, and the output screen 5000 displays the FIG. 23 loan judgment index calculation result output screen B5000 to the system user.

System outline diagram in an embodiment of the present invention The image figure after the grouping process in the embodiment of the present invention Overall outline processing flow in an embodiment of the present invention Processing flow until grouping by index calculation program 1110 and loan judgment data management program 2200 in the embodiment of the present invention Data table handled in step A2 of FIG. 4 in the embodiment of the present invention In the embodiment of the present invention, the data table created in step A3 of FIG. In the embodiment of the present invention, the data table created in steps A9 to A10 of FIG. In the embodiment of the present invention, the data table created in step A10 of FIG. Data table created in step A11 of FIG. 4 in the embodiment of the present invention Data table created in step A16 of FIG. 4 in the embodiment of the present invention Processing flow until calculation of loan judgment index for each group by the index calculation program 1110 and the loan judgment data management program 3200 in the embodiment of the present invention Detailed processing flowchart of step B8 of FIG. 11 in the embodiment of the present invention. Data table created in step B7 of FIG. 11 in the embodiment of the present invention Data table created in step B8 of FIG. 11 in the embodiment of the present invention Processing flow until calculation of a loan judgment index of a company corresponding to a loan consideration destination company by the index calculation program 1110 and the loan destination company management program 2200 in the embodiment of the present invention Detailed processing flowchart of step B18 of FIG. 15 in the embodiment of the present invention. In the embodiment of the present invention, the data table created in step B14 of FIG. Data table created in step B15 of FIG. 16 in the embodiment of the present invention Data table created in step B16 of FIG. 16 in the embodiment of the present invention In the embodiment of the present invention, the data table created in step B18 of FIG. An example of a grouping process screen for a loan company An example of a loan decision index calculation result screen at a loan consideration partner company

Explanation of symbols

1000 ... Loan decision index management server, 1110 ... Loan decision index calculation program, 1120 ... Data table defining transaction relationships of borrower companies, 1130 ... Data table defining strength index of each transaction, 1140 ... By attribute data of transaction Leveled data table, 1150 ... Client's revenue ratio data table, 1160 ... Data table defining group of borrower companies, 1170 ... Data table defining growth index for each group, 1180 ... Transactions of loan considering companies Relational data table, 1190: Database of loan destination company corresponding to loan considering company, 1200 ... Data table of loan judgment index calculation result, 2000 ... Loan company management server, 2100 ... Database on loan company, 2200 ... Loan company Data management program 2300 ... Database on loan considering company, 3000 ... Loan judgment data management server, 3100 ... Loan judgment attribute database, 3200 ... Loan judgment data management program, 4000 ... Terminal device for inputting user instruction operation, 5000 ... Processing Display device for displaying results, 6000... Network connecting various servers

Claims (5)

In a loan decision index calculation system that calculates an index for determining whether or not a loan is considered for a loan considering company,
A storage means for storing loan data relating to the loan of the loaned company in which the loan is implemented;
Using the loan data, quantification means for quantifying the business relationship between the loan companies,
A transaction strength index calculating means for calculating an index indicating the strength of the transaction from the quantified business relationship; a grouping means for grouping the borrower companies into groups according to the index;
For each group created by the grouping means, group index calculating means for calculating a group index;
Comparing the loan data of the loan destination company and the loan data of the loan consideration destination company stored and grouped in the storage means, and based on the comparison result, the company corresponding to the loan consideration destination company is Selection means to select from the borrower company,
A loan determination index calculation system comprising means for calculating a loan determination index using the group index of a group to which a selected loan destination company belongs. A loan determination index calculation system according to claim 1,
The said quantification means quantifies a transaction relationship from the transaction data containing the transaction amount and transaction continuation period for every said partner company from the said memory | storage means, The loan judgment parameter | index calculation system characterized by the above-mentioned. A loan determination index calculation system according to claim 1,
The transaction strength index calculating means uses the transaction continuation period and the transaction amount in each transaction, based on the transaction attribute data stored in the storage means, and the transaction increase rate of the loan-destination company and the customer of the loan-destination company A means of calculating an increase rate indicator, including the rate of increase in the company ’s revenue,
Means for generating a transaction level indicating a level of a transaction with the counterpart company by using a relative comparison with a transaction with another counterpart company using the calculated numerical value;
Calculating a transaction strength index indicating the strength of connection for each transaction using the transaction level and a weighting value indicating a weight for the transaction level determined according to an input of a user of the loan determination index calculation system. A characteristic loan decision index calculation system. In the loan judgment transaction index calculation system according to claim 3,
The transaction strength index calculating means is:
A loan judgment transaction index characterized in that, from the transaction attribute data, a mutual love degree of a business relationship between companies indicating how much the loan company depends on the transaction to the client company is calculated as the transaction strength index. Calculation system. In the loan judgment index calculation system according to claim 3,
The grouping means includes
For each transaction relationship, the index calculated by the transaction strength index calculation means and the profit ratio data obtained for each of the client companies of the loan company are acquired, and the conditions input by the user of the loan determination index calculation system are obtained. A loan judgment index calculation system characterized by grouping by selecting a business relationship to be considered in response.

Images