JP2002288421A - Method and system for simulating loan portfolio optimizing action - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system for simulating loan optimizing action which constructs a loan portfolio which has the credit risk minimized and which is realistic and implementable. SOLUTION: The system for simulating loan optimizing action which optimizes a loan portfolio is provided with a first means which calculates the cash flow, the credit risk, and the earning rate related to a first loan which has been already contracted for, a second means which calculates the cash flow, the credit risk, and the earning rate related to a second loan concerned with a new contract, and a third means which executes optimizing calculation on the basis of calculation results of the first means and the second means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a loan portfolio optimal behavior simulation system and method.

[0002]

2. Description of the Related Art Hitherto, systems for optimizing a portfolio of financial assets have been proposed. For example,
JP-A-6-35918 and Patent Publication 2000-29
No. 3569. Financial assets subject to these optimizations are stocks and bonds.

On the other hand, financing is widely performed in the financial industry. In this financing business, the business condition and the like of each individual company to be financed are analyzed to determine whether the financing is possible, and the conditions such as the amount and interest rate. However, even if decisions were made on an individual company basis, the criteria from the perspective of the loan portfolio as a whole were not clear. At this time, even if the system for optimizing the portfolio of financial assets described above is applied to the loan portfolio as it is, the financing business has a unique property different from financial assets as securities such as stocks and bonds. , It is not possible to build an optimal loan portfolio.

[0004]

As described above, in the prior art, there is a problem that a loan portfolio optimum behavior simulation system for optimizing a loan portfolio cannot be constructed.

It is therefore an object of the present invention to provide a loan portfolio optimal behavior simulation system that takes into account characteristics of a loan portfolio.

[0006]

According to the present invention, there is provided a loan portfolio optimizing behavior simulation system for optimizing a loan portfolio, comprising: a first means for calculating a cash flow and a credit risk amount relating to a first loan as a contract. A second means for calculating the cash flow and the amount of credit risk for the second loan on the new business;
The apparatus further comprises third means for executing an optimization calculation based on the calculation results of the first means and the second means. With such a configuration, optimization suitable for the loan portfolio can be executed.

Preferably, the first means and the second means
The bankruptcy probability variability is used as a calculation parameter when calculating the credit risk amount in the means of (1). Since the present invention targets a loan contract with no marketability, it is not possible to directly obtain a risk from the volatility of profit. Therefore, means for calculating the amount of credit risk from the bankruptcy probability variability is provided.

[0008] Further, when calculating the credit risk amount in the first means and the second means, the credit risk amount may be calculated using an actual allocation amount. As a result,
It is possible to control the corporate group loan balance for each contract and the total loan balance of the entire portfolio.

Further, the optimization calculation by the first means, the second means and the third means may be executed over a plurality of periods. As a result, it is possible to confirm the effects of portfolio status changes and optimization results over a long period of time.

It is desirable that the first means, the second means, and the third means execute the calculation in units of a company group including a company to be a borrower. As a result, the amount of calculation can be reduced, and the credit risk can be controlled by adjusting the loan amount on a corporate group basis.

Further, a credit risk amount decomposing means for decomposing a credit risk amount using correlation information for each type of business may be provided. Thereby, it becomes possible to grasp the details of the credit risk factors in controlling the credit risk amount.

It is desired that the industry-specific correlation information be an industry-specific price-earnings rate correlation.

Further, the group correlation may be calculated using the industry-specific correlation information. Thereby, the credit risk amount can be accurately calculated.

[0014] Further, means may be provided for setting a constraint condition for the first and second loans.

Preferably, the constraint includes a constraint on the balance of the entire loan portfolio, a constraint on the upper limit and / or lower limit of the balance by company group, by industry or by credit rating, by company group, by industry or by credit. Includes constraints on the upper and / or lower limit on new loan disbursement for balances by rating, and constraints on upper and / or lower limit on new loan disbursement for repayments by company group, industry or credit rating. As a result, it is possible to obtain a result suitable for practical operation.

On the other hand, the loan portfolio optimizing behavior simulation method for optimizing a loan portfolio according to the present invention calculates a cash flow and a credit risk amount relating to a first loan which is a loan already held and held. A first step, a second step of calculating a cash flow and a credit risk amount relating to a second loan, which is a loan for which a new contract is to be concluded in accordance with assumed unique loan conditions, and the first step; A third step of executing an optimization calculation based on the calculation result of the second step;
And the following steps. With such a configuration, optimization suitable for the loan portfolio can be executed.

Here, the first step and the second step
When calculating the credit risk amount in the step, the bankruptcy probability variability may be used as a calculation parameter. The present invention
Since it is aimed at loan assets with no marketability, it is impossible to determine the volatility of profits. Thus, by using the volatility of bankruptcy probability, the credit risk amount can be accurately calculated.

It is desirable that the credit risk amount in the first step and the second step is calculated using an actual allocation amount. Thereby, the credit risk amount can be accurately calculated.

The optimization calculation in the first step, the second step and the third step may be performed over a plurality of years. Thereby, optimization can be performed over a long period of time, so that more appropriate optimization calculation can be realized.

It is desirable that the first step, the second step, and the third step be performed in units of a company group including a company to be a borrower. As a result, credit risk control becomes possible by adjusting the amount of lending in units of other companies that can reduce the amount of calculation.

Further, a credit risk amount decomposing step of decomposing a credit risk amount using industry-specific correlation information assumed for a loan contract for each individual company may be provided. Thereby, the credit risk amount can be accurately calculated.

The industry-specific correlation information may be an industry-specific price-earnings rate correlation.

Preferably, the credit risk amount decomposition step calculates a group correlation using the correlation information for each industry. Thereby, the credit risk amount can be accurately calculated.

[0024] The method may further comprise the step of setting a constraint condition for the second loan.

Preferably, the constraint includes a constraint on the balance of the entire loan portfolio, a constraint on the upper limit and / or lower limit of the balance by company group, by industry or by credit rating, by company group, by industry or by credit. Includes constraints on the upper and / or lower limit on new loan disbursement for balances by rating, and upper and / or lower limit on new loan disbursement for repayments by business group, industry, or credit rating. As a result, it is possible to obtain a result suitable for practical operation.

On the other hand, a loan portfolio optimum behavior simulation program for optimizing a loan portfolio according to the present invention, wherein the program is a first loan which is a loan which has already been concluded and held with a computer. A first step of calculating a cash flow and a credit risk amount relating to a loan; and a second step of calculating a cash flow and a credit risk amount of a second loan, which is a loan for which a new contract is concluded according to assumed unique financing conditions. And a third step of executing an optimization calculation based on the calculation results of the first step and the second step.

Further, it is possible to store the loan portfolio optimizing program in a storage medium. When the above-described processing is executed by a computer, basically, the data stored in the storage unit is read out and calculated by the arithmetic unit. That is, on the computer, each of the above-described data is stored in the storage unit, read out by the arithmetic unit, and the calculation is executed. For example, the first means reads out data related to the first loan, calculates a cash flow and a credit risk amount by the calculation means, and stores the calculated cash flow and credit risk amount in a predetermined storage means. Further, the second means reads out the data related to the second loan, calculates the cash flow and the credit risk amount by the calculation means, and stores the calculated cash flow and credit risk amount in a predetermined storage means. further,
The third means reads out the data relating to the calculation result from the storage means for storing the calculation results of the first means and the second means, executes an optimization calculation, and stores the result in a predetermined storage means. . In addition, the above-described constraint condition is also temporarily stored in a predetermined storage unit.

[0028]

DETAILED DESCRIPTION OF THE INVENTION Loan Business (FIG. 1) First, an example of a loan business that is an object of the present invention will be briefly described with reference to FIG.

1. Project Candidate (S101) First, a financing officer of a financial institution that performs financing visits various companies and obtains information on financing from a financing officer of the company to find a financing contract candidate.

2. Examination (S102) When a loan contract candidate is found, it is examined whether the loan can be executed based on the contents of each case. In conducting the examination, the credit rating (S1021), loan amount (S1022), loan period (S1023), interest rate (S1023)
1024) are considered. Here, the financing is usually a continuous aspect, except for the first financing of the company, the additional financing at any time before the end of the financing agreement already concluded, Additional financing may be provided at the end of an existing contract.

As a result of the examination, if it is determined that the financing contract candidate cannot be worked on, the financing is not executed, and the operation related to the judgment as to whether or not the case can be executed ends (S103).

On the other hand, as a result of the examination, if it is determined that the loan can be executed, a loan contract is concluded (S104).

II. Outline of Flow of Financing Optimal Behavior Simulation System (FIG. 2) Next, an overview of the flow of the financing optimal behavior simulation system will be described with reference to FIG.

(1) Basic Data Acquisition and Generation of Early Contract Data (S201) First, in order to execute calculation and analysis at the beginning of the term,
The process starts by acquiring the attribute / balance data of the early contract and generating the early contract data including the cash flow (CF) and the like.

Here, the term "beginning" means, for example, when the calculation is performed with the year as the "unit calculation period", the beginning of the year, that is, the beginning of the year. The point in time is the “end of the term”. However, in the present invention, the unit calculation period is not limited to the fiscal year, but may be a half-year unit or a monthly unit. Further, a plurality of unit periods can be continuously analyzed.

The term "existing contract" refers to a loan contract already concluded at the beginning of the term. On the other hand, "new contract"
Is a loan agreement (candidate loan agreement) that has not been concluded at the beginning of the term, but may be contracted during the unit calculation period (current term), or will be executed based on the optimization calculation result. Says a new loan agreement.

In the CF, “principal CF” corresponding to the loan principal is used.
And "interest CF" corresponding to interest. The loan contract to be analyzed usually includes a plurality of principal CFs and interest CFs, but both the loan amount (hereinafter sometimes referred to as “balance”), the repayment period, and the interest rate level as loan conditions And attribute data including the repayment method. Alternatively, the CF data itself may be input to the external DB in advance, and may be acquired.

As required, the CF data may be stored as an actual amount as in the case of a balance, or may be stored at a CF rate per unit of the balance. Both can be mutually converted via the balance.

The CF data also includes information on a period up to the occurrence thereof (hereinafter, referred to as a “CF period”), as well as the information on the above-mentioned quantity (actual amount or balance rate).

(2) Beginning Analysis of Beginning Contract Data (S20)
2) Based on the acquired or generated early contract data, calculate the credit risk amount and the rate of return at the beginning of the early contract portfolio.

(3) End-of-Term Processing of Existing Contract Data (S203) Next, in order to execute calculation and analysis at the end of the term, end-of-term processing of existing contract data is executed.

Specifically, for the contract data at the beginning of the term, the CF and balance are adjusted in consideration of the influence of the passage of time,
Generate term-end contract data.

(4) End-of-Term Analysis of Existing Contract Data (S204) Based on the end-of-term contract data, a credit risk amount calculation and a return rate calculation of the existing contract portfolio at the end of the term are executed.

(5) Generation of New Contract Data for Current Term (S205) Next, new contract data for the current term is generated in preparation for the optimization calculation. Here, a new contract CF is reasonably assumed and generated from the existing contract attributes, market interest rate data, assumed input data, and the like.

(6) Creation of input data for optimization calculation (S2
06) Next, input data for optimization calculation is specifically created using the results of S203, S204, and S205. The main input data for optimization prepared here are the data obtained by re-arranging the credit risk amount already incurred calculated in S204 and the credit risk amount in new business calculated from S205, and data relating to corporate group correlation. It is.

(7) Creation of constraint data for optimization calculation (S207) Constraints are created based on various data generated from the processes up to this point and data given from the input unit, and are used in the optimization calculation process. Sent in.

(8) Optimization Calculation (S208) The end-of-period portfolio is optimized using the aforementioned input data for optimization and constraint data. The optimization calculation executed here can be repeatedly performed in accordance with the constraint condition input from the input unit, and each optimization calculation result can be held at the same time. In addition, the processes subsequent to the next process can be performed on each optimization calculation result, and a plurality of calculation patterns can be held until final form output.

(9) Term-End Default Processing (S 209) Based on the optimization result, term-end existing contract data and term-end new contract data default processing is performed. Here, using the average bankruptcy probability or the like, the occurrence of defaults for the optimized balance and CF is estimated, and the existing contract and new contract data are respectively corrected.

(10) Post-Default Analysis of Existing / New Business Data (S210) Thereafter, based on default-processed existing business data and new business data at the end of the term, credit risk amount calculation and return rate calculation for the portfolio are performed again. And so on.

(11) Form Calculation and Form Output (S211) The output results from S202, S204 and S210 are tabulated and calculated by various methods, and after outputting some additional indices, output to the form is executed. I do. In this process, the calculation result regarding the loan portfolio in each state at each time point is confirmed, and the processing for one unit calculation period (for example, the first year) ends.

(12) Calculation period and calculation pattern discrimination (S
212) It is determined whether the calculation for the predetermined period has been completed. Here, the “predetermined period” is, for example, three years (three unit calculation periods). If the calculation for the predetermined period is not completed, the analysis result (S210) of the plurality of calculation patterns executed in S207 is selected. S208 corresponding to the selected calculation pattern
And the data from S209 is used as the beginning contract data of the next unit calculation period. Similarly, the analysis result of S210 is carried over to the calculation of the next unit calculation period, and is reflected on the output result of S210 again.

(13) Calculation relating to the next unit calculation period (S21
3) In response to the determination processing of S212, the above-mentioned S202 to S211
The same processing as described above is repeatedly executed.

III. Details of the portfolio lending optimal behavior simulation system flow (Figure 3)

Next, the process of optimizing the loan portfolio according to the present invention will be described in detail with reference to the flowchart shown in FIG.

In the process of the present invention, an optimization process is performed over a plurality of unit calculation periods (for example, for three years). In this figure, one unit calculation period (for one year) is displayed. . In the figure, S11 to S15 are processes executed on the basis of the beginning of the term (the beginning of the fiscal year), and
56 is a process executed based on the end of the term (end of the fiscal year). Hereinafter, each step will be described in order.

1. Basic Data Acquisition and Analysis of Early Contract Data (S11 to 15) 1-1 Basic Data Acquisition from External DB 1 and Input Unit 2 and Storage in Main Program DB As shown in FIG. In executing the optimization process, data is obtained from an external database (external DB 1) and the input unit 2. The external DB 1 is constituted by storage means such as a memory and a hard disk, and stores basic data mainly related to a loan contract and a market environment. The basic data includes data related to a loan contract (eg, company name, balance, CF, industry, credit rating), data related to bankruptcy probability (eg, average bankruptcy probability, bankruptcy probability change), and data related to market environment (eg, bankruptcy probability). For example, market interest rates).

The input unit 2 is composed of a keyboard, a mouse, and the like, and inputs information relating to an assumed rate of return and constraints. The constraints may include constraints on the balance of the loan portfolio as a whole, constraints on the upper and / or lower limits of balances by business group, industry or credit rating,
Restrictions on the upper and / or lower limit on the amount of new lending for balances by business group, industry or credit rating, and on the upper and / or lower limit on the amount of new lending for repaid amount by business group, industry or credit rating Constraints are included.

The acquired data is stored in a DB area (hereinafter, “MDB”) on the main program. This MD
B has an area for temporarily storing data obtained from the external DB 1 or processing data calculated in the present system, and a storage area for a final output result.

Most of the data acquisition from the external DB 1 and the input unit 2 is executed in this step, but may be executed in the subsequent steps as necessary.

1-2 Generation of Beginning Contract Data (S1
1) Each data acquired in 1-1 is arranged and arranged according to the unit of data processing, and CF data is generated as necessary. Basic data processed in the system according to the present invention basically holds data for each company to be financed, that is, for each individual company, but for each individual contract. Is also good. Therefore, in order to handle data consistently in the subsequent processing, it is necessary to convert and arrange the data holding unit as necessary and store it again in the MDB.

Even after being stored again in the MDB, the basic data and various processed data (intermediate calculation data) are
"Company group unit" to which the company belongs according to the purpose of use
Or "portfolio unit" or "industry unit", or disassembled again and passed to the next step,
It will be output and stored.

In the following description, it is basically assumed that the basic data and the intermediate calculation data are stored in individual companies. However, depending on the calculation items, it is necessary to pay attention to the mutual conversion. In some cases, it is necessary to explicitly specify a specific data holding unit. In this case, an annotation is appropriately added.

For the contracted CF data at the beginning of the term,
The CF data as obtained from the external DB 1 is used. If the calculation extends over a plurality of unit calculation periods, the CF data finally generated in the processing of the previous period will be used. However, it may be regenerated from the basic data according to the method described in S31.

1-3 Gross Risk Amount (VCF) at Beginning of Contract
Calculation (S12) The initial contract gross risk amount (hereinafter sometimes referred to as gross risk amount or VCF) as input data (intermediate calculation data) for calculating the credit risk amount is calculated.

VCF refers to a gross risk amount caused by bankruptcy probability fluctuation, and can be specifically expressed by the following equation.

[0066]

(Equation 1) <Example of symbol> Cash flow to be evaluated: CF Assumed recovery rate: θ Discount coefficient: d Bankruptcy probability variability: σ

In this equation, each parameter is represented by θ is a recoverable ratio at the time of default, d is a coefficient (discount coefficient) for converting a cash flow into a present value,
σ is the bankruptcy probability variability. Therefore, in other words, VCF
Indicates how much the present value of the cash flow excluding the recoverable ratio fluctuates due to the bankruptcy probability change.

The present system simulates the measurement of the credit risk amount and the optimization behavior of the portfolio composition based on the measurement.
The CF value is treated as a credit risk amount inherent to the contract, assuming that there is a single loan contract to be analyzed and evaluated.

More precisely, each parameter is a numerical value depending on an attribute of an individual loan contract or cash flow. For example, the assumed recovery rate is set according to the loan contract, the discount coefficient is set according to the CF, and the bankruptcy probability change is set according to the credit rating to which the loan contract belongs.

Note that, at this stage, the calculation results are held for each company. Assuming a perfect correlation between CFs at the same company at different points in time, rewriting based on the contents of the above parameters gives the following.

[0071]

(Equation 2) <Example of symbol> Company name (ID): Subscript i CF date: t

1-4 Calculation of Disassembled Credit Risk Amount at Beginning of Contract (S13) (1) Meaning of Calculation of Disassembled Credit Risk Amount of disassembled credit risk is calculated from correlation data on individual company VCF and bankruptcy probability fluctuation between companies Is calculated.

The amount of decomposed credit risk is different from VCF,
The credit risk amount of each individual loan contract given to the calculation of the credit risk amount of the entire loan portfolio, which is a collection of loan contracts to be analyzed and evaluated. Calculated assuming there are multiple contracts to be considered. By performing such processing, the credit risk amount (gross risk amount) calculated assuming that the company exists alone
Can be calculated to calculate more accurate portfolio and credit risk amount, and the credit risk amount decomposition calculation results for each credit risk element useful for controlling credit risk amount (Group factor credit risk amount and industry factor credit risk amount) ) Can be obtained.

Before describing the specific calculation of the amount of disassembled credit risk in this system, a basic theory as a premise will be explained. The method of credit risk decomposition is basically based on the risk calculation method defined in the so-called “portfolio theory”. Here, the risk amount of the entire “portfolio”, which is an aggregate of individual securities, is measured by using the “variability” of the return rate of the individual securities and the “correlation” between them.

[0075]

(Equation 3) <Example of symbol> Return rate variability of individual securities: σ Correlation of above rate of return variability: ρ Fund allocation: w Individual security name (ID): Subscript i, j Portfolio: Subscript p

(2) Parameter Calculation in Credit Risk Calculation In the present system, the concept of the general method described above is used as a base, while taking into account the specificity of analysis for loan contracts and the availability of objective data. , Parameters, calculation methods, and how to give data.

<Credit Risk Amount of Assets Alone> First, the volatility of individual assets alone, that is, the amount of risk, is generally an asset that is marketable and for which objective price data can be obtained (for example, In the analysis of stocks, etc., the variability can be calculated directly from the actually acquired price data in the form of a standard deviation or the like.

On the other hand, with respect to a loan contract, which is an asset based on bilateral transactions, which is generally not marketable and for which price data is difficult to obtain, as handled by the present system, the loan contract is directly obtained from objective price data. It is not possible to obtain the risk amount. Therefore, as shown in 2-1 in the present system in which the value fluctuation due to the corporate credit factor is considered as the risk amount to be analyzed, it is indirectly derived from the bankruptcy probability variability. That is, the VCF value is the credit risk amount (gross risk amount) of an individual loan contract alone, and
(Accurately, wσ).

<Correlation> Regarding the correlation, for the same reason as the credit risk amount, it is difficult to obtain direct observation data on the loan contract. Therefore, in this system, the stock price is used as a proxy observation variable that reflects the corporate credit factor of the company to some extent and provides objective observation data required for correlation measurement. For the correlation used as a parameter for the above calculation, information of the asset value itself is not used, and thus such a method can be selected.

However, even in this case, this method cannot be simply applied when dealing with a loan contract that includes an unlisted company or a non-stock company. Therefore, this system uses the stock prices of companies listed on the First Section of the Tokyo Stock Exchange to temporarily classify each company into an industry, calculate the correlation between the industries ("industry correlation"), and analyze In this method, the industry correlation is assigned as a proxy variable based on the business nature of the loan contract.

<Fund Allocation> In general, w is often calculated as the ratio of individual securities held to the total portfolio held. However, in this system, VC
As is clear from the definition of F, the calculation using the actual allocation amount is performed. It is desired to select the actual amount calculation from the relationship with the constraint condition setting method described later and the relationship with the final output method.

(3) Credit Risk Decomposition Calculation In this system, since w is not the investment ratio but the actual amount, the equation (Equation 3) relating to the credit risk is rewritten as follows.

[0083]

(Equation 4) Since the first two terms on the right-hand side are numerical values calculated only from the credit risk amount peculiar to the individual company, they are defined as “the credit risk amount attributable to the individual company”. Furthermore, in this system, since the corporate credit factors of individual companies within the same corporate group are assumed to be fully correlated, the simple sum of the individual company-related credit risk amounts of different individual companies is referred to as the “group-factor credit risk amount”. It can be defined as

On the other hand, in the latter two cases, in addition to the credit risk amount peculiar to the individual company, it is calculated from the credit risk amount and the correlation value of the target company in which the correlation is considered. Moreover, since the correlation is given by the correlation of the industry to which each individual company belongs, these terms are defined as "industry-related credit risk amount".

Therefore, from the formula 4, the credit risk amount of the entire portfolio can be accurately calculated in consideration of not only the gross risk amount of the individual company loan contract, which is a component of the portfolio, but also its correlation. In addition, we were able to break down the credit risk of the entire portfolio into “group factor credit risk” and “industry factor credit risk” as the contributory credit risk of individual company loan agreements .

(4) Actual Data Processing The calculations of Equations 1 and 2 are executed based on the contract data at the beginning of the term, and the elements of Equations 3 and 4 are converted into a combination of individual financing contracts included in the portfolio. Calculate for As a result, the portfolio / credit risk amount and the individual company decomposition credit risk amount can be obtained, but they may be aggregated into groups or business units as needed. Further, the credit risk rate may be calculated and held by dividing by the balance corresponding to each credit risk amount.

1-5 Calculation of Beginning Contract Return Rate (S14) The return rate is obtained or calculated by the following method.

A) The corresponding rate of return assumed for each loan contract is obtained from the external DB 1 or the input unit 2. The input numerical value is not limited to each individual contract, but may be for each specific contract group (for example, each credit rating, each company group, each business type, each repayment term, etc.) that can be determined by attributes.

B) Calculate the ratio of interest CF per unit calculation period to the balance (principal, book value) of each contract. If the CF amount differs at each point in time, the average amount per one unit calculation period may be used.

Regarding b), it is considered that the “gross profit rate” means “including various costs”, but the average bankruptcy probability, the cumulative bankruptcy probability acquired in S11,
Using the recoverable ratio or the like, an average expected loss related to bankruptcy may be deducted from interest CF as a cost related to corporate credit factors, and calculated as “net rate of return”. Similarly, for a), a numerical value corresponding to the net profit rate may be stored in the external DB 1 in advance.
Also, multiply each "rate" by the corresponding balance,
The profit amount may be calculated and held.

Further, these numerical values can be calculated for each company, but when they are consolidated for each group or portfolio, they may be weighted by the balance acquired in S11.

1-6 Calculation for Beginning Contract Form (S15) Basically, the balance, profit rate, credit risk rate (credit risk amount), etc., as the calculation result in each step calculated for each company are calculated for each company. The information is organized together with the attributes, and is totaled for each item in a group unit / industry unit / credit rating unit and stored in the MDB. At this stage, the form can be output, but may be output together with the calculation result from the subsequent calculation process.

2. End-of-term analysis of existing contract data (S21-2
5) 2-1 End-of-Term Processing of Existing Contract Data (S21) Perform processing to remove the CF repaid in the current term (the unit calculation period) from the existing contract CF at the beginning of the term, and generate the existing contract CF corresponding to the term-end analysis. .

The CF corresponding to the current term is specified from the principal CF term data of the contracted data at the beginning of the term, and the CF is deleted from the original contracted principal CF data. Also, the remaining principal CF
, The respective CF period data is shortened by the elapsed period corresponding to the unit calculation period and re-recorded. Interest CF
Also performs the same processing as the principal CF.

The specific input data includes the beginning contracted principal CF / interest CF, the corresponding CF period, and the beginning contracted balance (book value). Book CF
/ Generation of interest CF and various tabulations are executed. The output date includes the term-end contracted principal CF / interest CF, the CF period, the term-end contract book balance, and the term-end contracted repayment amount (total of deleted principal CF).

Basically, data is processed in units of individual companies. However, if necessary, the data may be added and totaled in units of company groups, portfolios and business units.

2-2 Calculation of gross risk amount of contracts already in
22) (1) Calculation of forward rate First, calculation of discount coefficient and interest CF used for VCF calculation
For generation, the forward rate is calculated by reasonably estimating the interest rate term structure at the end of the period.

The forward rate interest after the elapse of the unit calculation period is calculated based on the spot rate interest (for example, 10 years for every 0.5 years) corresponding to the analysis start time already stored in the MDB. At this time, in consideration of the analysis period over a plurality of periods, the forward rate finally required at all the analysis points may be obtained. Furthermore,
Forward rate and discounted CF, CF
A discount coefficient can be obtained from the period.

(2) Calculation of the Gross Risk Amount at the End of the Term The gross risk amount at the end of the term of the existing contract is recalculated from the contracted CF data processed at the end of the term and the forward rate. By this recalculation, it is possible to obtain a gross risk amount that reflects a change in CF of an existing contract due to repayment in the current period, a change in a discount coefficient due to a different time point, a corresponding change in bankruptcy probability variability, and the like. The input data includes the term-end contracted principal CF / interest CF, the forward rate (corresponding to the term-end), the cumulative probability of bankruptcy / recoverable ratio.
The VCF is calculated by a method according to the method of -3. Output data includes VCF and individual company ID at the end of the first year of contract.
/ Company group ID / industry.

2-3 Calculating the amount of disaggregated credit risk at the end of the term (S23) Using the existing contract data and the gross risk amount obtained and calculated in the above 2-1 and 2-2 after the term end processing, The same method is used to calculate the amount of credit disintegration at the end of the period.

Specifically, the same processing as the processing executed in step S13 is performed. Here, the input data includes the VCF at the end of the term, industry / group data, and industry correlation. The output data includes the end-of-period portfolio and credit risk, and group / industry credit risk.

The calculation result in this step is not only sent to the form output calculation as the loan portfolio analysis result at this time, but also used as an input value for the later-described group correlation calculation and optimization calculation. You.

2-4 Calculation of Returned Contracted Rate at End of Term (S24) By the same method as in the above 1-5, the rate of contracted return at the end of the term is calculated.

2-5 Calculation for Existing Contract Form at End of Term (S25) As in 1-6, the balance, profit rate, credit risk rate (credit risk And the like are organized together with the attributes of individual companies, and the items are totaled for each item in a group unit / industry unit / credit rating unit and stored in the MDB. At this stage, the form can be output, but may be output together with the calculation result from the subsequent calculation process.

3. Year-end analysis of new business data for the current term (S31
-34) 3-1 Generation of new contract data for the current term (S31) An execution candidate loan contract that may be newly contracted in the current term or a loan contract that is deemed to have been contracted in the current year based on the optimization calculation result Generate data for new contracts for the current term.

The final balance is calculated by optimization (S41, S4
2. Since the determination is made after performing S43), a CF is generated for the temporarily set balance, and thereafter, the risk rate and the return rate are calculated.

For example, as for the principal CF, the actual amount of the contract original principal CF itself is used. In addition, the principal C
F is multiplied by the “interest rate” obtained by adding the expected rate of return on new business acquired from the input unit or the external DB and the forward rate corresponding to the calculation time point (here, the end of the term) already calculated in S21, Generate interest CF. Furthermore, when the assumed rate of return used here is the net rate of return, contrary to 1-5, the average bankruptcy probability, the cumulative bankruptcy probability,
The average amount of losses related to bankruptcy that is estimated using the recoverable ratio, etc. is calculated as the cost related to corporate credit factors.
By returning to F, a gross standard interest CF can be obtained.

3-2 Calculation of Gross Risk Amount of New Business for Current Term (S3
2) Based on the new contract data generated in 3-1 above, VCF calculation is performed in the same manner as in 1-3. Note that the risk amount is basically calculated on a company basis, but here, a total calculation is performed on a group basis for a group correlation calculation in the next step.

3-3 Group Correlation Calculation (S33) (1) Purpose of Group Correlation Calculation Variable control in optimization calculation for credit risk control in the present system, which will be described later, is performed on a company group basis. This is because, when performing a portfolio analysis involving thousands of individual company loan contracts, the system selects variable control on a group basis as a result of taking into account the computational load and the utility of the analysis results.

However, as described above, due to the problem of obtaining objective observation data, the credit risk amount has been measured for individual company loan contracts or individual CFs using the industry correlation as a proxy variable. .

Therefore, in order to execute group variable control (specifically, loan balance control) by optimization calculation based on the so-called least squares method, the correlation (ρ) corresponding to the “corporate group” which is the control unit. And variability (σ) data are required.

(2) Calculation of Group Correlation Group correlation is calculated from known individual company unit data.

In Equations 3 and 4, the credit risk amount was calculated from the credit risk amount per company and its correlation. In the following, Equation 4 before deformation is re-presented.

[0114]

(Equation 5) This is a self-evident result because the correlation between the two companies is shown. However, if the above equation is expanded on the assumption that the correlation between corporate credit factors within the same group is complete, the correlation between two corporate groups including multiple companies each with a known individual company Can be defined as

[0115]

(Equation 6) <Example of use of symbols> Corporate group (ID): Subscript k, l Individual company financing contract (ID): Subscript i, j However, the company (plural) indicated by subscript i is the corporate group k.
And the company (plurality) indicated by the subscript j belongs to the company group l.

The numerator item is the portfolio / credit risk amount calculated using the VCF of each individual brand, and the denominator item is the gross risk amount of the corporate group. After all, using the intermediate calculation data calculated in the above-described steps, the group correlation can be calculated as follows.

[0117]

(Equation 7) (3) Actual data processing Interim calculation data and industry correlation data, such as the amount of new business gross risk from 3-2, the amount of existing contracted credit risk at 2-2, and the amount of existing portfolio and credit risk at 2-3 at the end of period Calculate the group correlation from

Here, the corporate group correlation relating to the new business data for the current period is calculated by the method of the above formula 7, and the correlation between the portfolio at the end of the period and the new business for each group unit is also calculated. This is because, in the optimization calculation described later, it is necessary to execute the control calculation of the entire loan portfolio including the existing contracts / new contracts in a form suitable for the asset characteristics of the loan contract.

When calculating the correlation between the portfolio of existing contracts and the new business in group unit, in equation (7), one of the group VCFs for k or l included in the denominator is used as the portfolio / credit risk amount. Need to be replaced. As for the numerator, as described above, either k or l indicates the contracted portfolio.

The output result is a correlation coefficient between the group-based new contract data and the existing contract portfolio.

3-4 Calculation of expected return on new business for the current term (S
34) This is the calculation of the expected return on new business acquired from the input unit for each group. Since the result of calculation of the profit rate here is input data for optimization described later, the net profit rate is basically assumed, and the data is processed as group unit data.

4. Loan portfolio optimization calculation (S4
4-1) Creation of input data for loan portfolio optimization calculation (S41) Input data for optimization calculation is basically calculated in each of the above steps, but is aggregated in this step. And sent to the optimization calculation program. In the optimization related to minimizing the amount of credit risk, the core input items are “portable portfolio / credit risk amount / credit risk ratio”, “balance of existing group units”, “VCF of new business group”, “new business unit”. Group and contract portfolio correlation. " However, it is also possible to set the constraint conditions described in the following sections, and in that case, additional input data is prepared.

4-2 Creation of Constraint Condition Data for Financing Portfolio Optimization Calculation (S42) In addition to the base optimization input values described in 4-1 above, the following constraint conditions are added to further improve the effectiveness. Calculation can be performed. Since the variable in the optimization calculation process is the new contract group balance on a group basis, the constraint on the calculation needs to correspond to the new contract group unit balance.

However, since the actual constraint condition input numerical value may be for the entire portfolio including the existing contracts, the input support calculation at the time of input and correction by internal processing based on a predetermined rule are required. Then, the constraint condition data is passed to the next optimization calculation process. Hereinafter, the logic will be mainly described.

(1) Constraint Condition on Return Rate of Entire Loan Portfolio Credit rate amount is controlled by assigning a profit rate or a target value of the profit amount to the entire loan portfolio as a constraint condition, thereby controlling the credit target amount and controlling the profit target amount. An optimization calculation for minimization can be performed.

Constraint: Return rate of entire loan portfolio = Input target return rate

The input values included in the right side of the constraint expression may be stored in the external DB 1 in advance, but are basically input from the input unit 2 according to each calculation pattern. (The same applies to (4) below.)

(2) Constraints on the Balance of the Loan Portfolio By assigning the target value of the balance to the entire loan portfolio as a constraint, it is possible to preliminarily define the balance of the loan portfolio after term-end calculation. Become. The financial institution that conducts the lending business finances the funds to be lent out to the company by financing separately, so in the process of optimizing the portfolio, it also applies constraints to the overall balance to optimize it. Will increase the meaning.

Restriction: balance of entire loan portfolio
= Enter target balance value

(3) Restrictions on the upper limit of the balance by company group, industry, etc.

In addition, a financial institution that conducts a loan operation generally performs various types of risk management by setting the upper limit of the balance at various points such as individual companies, corporate groups, and industries. ing.

By providing the upper limit of the balance as a constraint in this way, optimization of the loan portfolio taking credit risk management into consideration is achieved. In this system, since the system load is reduced by aggregating data for individual companies for each corporate group, the upper limit of the balance for each corporate group is given as a constraint. Further, it is also possible to incorporate the upper limit of the balance such as credit rating as a constraint.

Restriction: Group unit balance ≦ balance upper limit input value (for each group) Industry unit balance ≦ balance upper limit input value (for each industry)

(4) Restrictions on Upper and Lower Limits of New Loan Execution Amount to Repayment Amount by Company Group Generally, a loan is a balance due to being repaid according to an already concluded contract for an existing contract portion. Except for the decrease, there is no asset decrease. This is in principle different from stocks and bonds that have marketability that can be exchanged for stock at any time.

In the optimization formula, the existing contract is treated as if no increase or decrease in the balance other than the contract is performed, while the new contract is financed with an arbitrary amount under the constraint. Treated as performing.

Here, in consideration of the fact that a loan contract is often renewed by refinancing with an existing client company other than a transaction with a new company, a new contract for the current period repayment amount from the corporate group is taken into account. By setting the range of the ratio of the loan amount, a constraint condition is set that a new loan can be executed within the range of the upper limit (upper limit) and the lower limit (lower limit) of the range.

Restriction: Upper limit of loan amount for new contract ≤ current period repayment amount x input value of repayment upper limit ratio Constraint: Lower limit amount of loan for new contract ≥ current period repayment amount x input value of lower limit of repayment amount

(5) Constraint on New Loan Execution Limit for Balance by Company Group The constraint in (4) above is the ratio of new loan execution to the current period repayment amount. In effect, there is no new loan disbursement amount. Similarly, if the current period repayment amount is excessively small, the newly executed amount does not become so large.

In view of the fact that the amount of new financing is affected by the existing contract balance, the upper limit of new financing is set differently from the above (4) by setting the upper limit ratio to the balance at the end of the previous term for each corporate group. Constraint condition as the upper limit ratio of

Restriction: Upper limit of loan amount for new contract ≦ balance at the end of the previous term × end value of the balance at the end of the term

(6) Setting of Priority Relationship of Constraints Since the constraints of (2) to (5) are constraints on some “balance”, consistent constraint setting with a reasonable solution Therefore, it is necessary to set a priority relationship between the conditions. In this system, it is also possible to input and set this priority relationship from the input unit 2.

The input data of the constraint conditions can be arbitrarily set for each company group, each business type, etc., and can be arbitrarily set for the same company group or over time. . That is, the constraints include the constraints on the balance of the entire loan portfolio, the constraints on the upper limit and / or lower limit of the balance by company group, industry or credit rating, and the balance on the balance by company group, industry or credit rating. Includes constraints on the upper and / or lower limit on new loan disbursement, and constraints on upper and / or lower limit on new loan disbursement for repayments by business group, industry or credit rating. As a result, it is possible to obtain a result suitable for practical operation. Similarly, the priority relationship of the constraint conditions can be arbitrarily set with the passage of time.

4-3 Executing Optimization Calculation (S43) After setting the input values and the constraint conditions, the optimization calculation for the entire loan portfolio is executed. Specifically, the credit risk minimization calculation is executed by the quadratic programming method. Here, the balance corresponding to the existing contract is fixed,
The objective function is set so that the new loan balance becomes the optimization variable. However, as described above, since the correlation data including the existing contract balance and the existing contract is obtained as the input value, it is possible to execute the target setting of minimizing the credit risk amount of the entire portfolio. However, if necessary, the portfolio and credit risk amount of only the new business may be minimized.

[0144] 5. Period-end analysis of optimization portfolio (S
51-55) 5-1 Default Processing of Existing Contract Data (S51) The existing contract data is corrected on the assumption that a default occurs for the existing contract. As for the default occurrence probability, average bankruptcy probability data corresponding to the credit rating acquired from the external DB will be used. The “average” referred to here is, for example, an objective probability estimated “average” for loan contracts belonging to the same credit rating, and “average” for the elapsed period after execution of the contract. "Probability". However, the “unique” probability corresponding to the elapsed time after execution may be derived from the accumulated bankruptcy probability data that can be obtained in the same manner as the bankruptcy probability for the current term.

From the obtained bankruptcy probability data, a “default amount” is obtained. The default amount is the balance multiplied by the probability of bankruptcy, and is the expected value of the principal loss that would be incurred by bankruptcy. However, in order to more reasonably estimate the default amount corresponding to the current term, the balance used here may be the average value of the beginning balance and the ending balance of the existing contract.

The obtained default amount is subtracted from the existing contract term end balance to calculate the contract term end balance after the default processing (hereinafter, “contract term end balance (Def)”). Although the balance may be a negative number, holding the negative balance has a problem in terms of analytical concept. In such a case, a correction process is performed. Basically, the existing contract end balance (Def) is set to 0, but an amount corresponding to a negative number may be deducted from a new contract end balance (Def) described later of the individual company. Even in such a case, a negative number is not allowed for the new business term end balance (Def).

In addition, the CF data is corrected. The ratio obtained by dividing the obtained default amount by the contracted balance at the end of the period is multiplied by each of the principal CF and the interest CF, and the defaulted contracted end CF after the default processing (hereinafter referred to as “the contracted end CF (De
f) ").

The “default collection amount” is calculated by multiplying the obtained default amount by the recoverable ratio.

5-2 Default Processing of New Contract Data (S52) Similarly, the default processing is performed on the new contract data after the optimization at the end of the term. (Hereinafter, the data after the default processing including the existing contract and the new contract may be simply referred to as “data after the default processing”.)

5-3 Calculation of Gross Risk Amount of Data after Default Processing (S53) For new / contracted data (data after default processing) generated in 5-1 and 5-2, 1-3 and The gross risk amount calculation is executed in the same manner as in 2-2.

5-4 Calculation of Decomposed Credit Risk Amount After Default Processing (S54) 1-4 and 2 are applied to new / contracted data (data after default processing) generated in 5-1 and 5-2. -3, the calculation of the disaggregated credit risk amount is performed.

5-5 Calculation of Return Rate after Default Processing (S
55) A return rate calculation is performed on the new contract / existing contract data (data after the default processing) generated in 5-1 and 5-2 by the same method as in 1-5 and 2-4.

6. Form calculation and form output (S61-6
2) 6-1 Calculation for Form (S61) Here, a calculation including a part of analytical contents is executed for data to be included in a form output at the end of one unit calculation period. Basically, "processing" on data display, such as totaling / distribution (for example, individual company unit-group unit-industry unit-portfolio unit) or display unit change of data calculated in each of the above steps. Processing is in progress. However, some items include the analysis items newly calculated in this step. Broadly speaking, the former is data related to optimization calculations (hereinafter referred to as “optimization-related data”), and the latter is data on other loan agreements (individual, group, portfolio, etc.) Data related to the status indicator (hereafter,
"Index data").

In the calculation over a plurality of periods, if the unit calculation period corresponds to the second or later unit calculation period, the same optimization-related data and index data in the previous unit calculation period, and the form output result May be included.

The optimization-related data includes loan contract data such as cash flows and balances calculated in the above-described steps, and credit risk amounts, profit rates or constraint data calculated based on the loan contract data. It is. In addition,
Some data has already been processed in S15 and S25. Since the index data is not used in the steps related to the above-described optimization calculation, the calculation and the output can be arbitrarily selected. For example, as a result of the optimization calculation, it is displayed whether or not each group-based financing contract or each industry-based financing contract violated a preset constraint condition, or a credit rating distribution within the group.

6-2 Form Output (S62) The contents prepared in 6-1 are output as a form from a screen or a printer.

[0157] 7. Calculation period and calculation pattern discrimination (S
71) If it is determined that the calculation period input from the input unit has come in advance, the processing ends at that point. If it is determined that the calculation period has not arrived, a plurality of possible calculation patterns are selected. A unique “trial number” is assigned to a plurality of calculation patterns, and by specifying the selected trial number from the input unit, the calculation pattern of the unit calculation period sent to the calculation process of the next unit calculation period is designated. It is determined. In addition, in specifying the calculation pattern, the calculation pattern may be automatically executed by a method such as a permutation based on quantitative form output item contents or a numerical range specification.

8. Sending data to the next unit calculation period and calculating the next unit calculation period (S81) The state of the loan portfolio associated with the processing in the unit calculation period is referred to as the next unit calculation period process (hereinafter sometimes referred to as the "next calculation process"). The data corresponding to the selected calculation pattern is sent to the next calculation process in order to reflect the above. The data to be sent are basically two types: loan contract data reflecting the optimization result and the default processing result, and form output data. First, 5-1 and 5-2
Send the post-default data related to the new contract and the existing contract processed in. These data will be treated as an early contract in the next unit calculation period. However, regarding data corresponding to a loan contract that was an existing contract or a new contract during the period, data processing depending on the number of years elapsed is included in the calculation of the next unit calculation period. Make a clear distinction. Therefore, when the multi-period calculation is performed, in the processing in the second and subsequent unit calculation periods, 3
The data after the default processing, which is managed more than once, is sent. It should be noted that such a distinct management state is continuously maintained throughout the execution of the next process.

Further, regarding the form output contents of 6-2,
Send the data. Since this data is output simultaneously when the form is output for the next unit calculation period, it is possible to display the portfolio attributes and the transition of each calculation result at each analysis time.

In the above example, VCF is used as an example of measuring credit risk. However, the present invention is not limited to this, and credit risk may be used according to other definitions. In addition, it is applicable to other risks, not just credit risks. Also,
Regarding data other than the above, since the acquired / input data and the intermediate calculation data of each step are stored in the MDB, they may be sent to the next process as needed.
In the above-described example, the lending optimal behavior simulation system can be constituted by a general-purpose computer and a predetermined program in addition to the dedicated computer. This program can be stored on a storage medium,
It can also be transmitted via a communication medium.
Here, examples of the storage medium include a floppy (registered trademark) disk, a hard disk, a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD, a ROM cartridge, a RAM memory cartridge with battery backup, a flash memory cartridge, and a nonvolatile RAM cartridge. And so on. Further, the communication medium includes a wired communication medium such as a telephone line, a wireless communication medium such as a microwave line, and the like, and also includes the Internet. Further, the computer constituting the present system includes at least arithmetic means such as a CPU and storage means such as a memory, and the various calculations described above are executed by the arithmetic means, and the calculation results are appropriately stored in the storage means. Stored and read as needed.

[0161]

According to the present invention, it is possible to provide a loan portfolio optimal behavior simulation system that takes into account characteristics of a loan portfolio.

[Brief description of the drawings]

FIG. 1 is a flowchart for explaining a loan operation.

FIG. 2 is a flowchart showing an outline of a processing flow in the loan portfolio optimal behavior simulation system according to the embodiment of the present invention.

FIG. 3 is a flowchart showing details of a processing flow in the loan portfolio optimal behavior simulation system according to the embodiment of the present invention.

[Explanation of symbols]

 1 External DB 2 Input section

Continued on the front page (72) Inventor Sugio Suzuki 1-1-1 Yurakucho, Chiyoda-ku, Tokyo Dai-ichi Mutual Insurance Company (72) Inventor Kazuyuki Shigemoto 1-1-1 Yurakucho, Chiyoda-ku, Tokyo Dai-ichi Life Within Mutual Insurance Company (72) Inventor Takashi Tsuda 1-1-1 Yurakucho, Chiyoda-ku, Tokyo Dai-ichi Life Within Mutual Insurance Company (72) Inventor Tomohiko Asano 1-5-1, Otemachi, Chiyoda-ku, Tokyo Within Technology Co., Ltd. (72) Inventor Shinya Takamura 1-5-1 Otemachi, Chiyoda-ku, Tokyo, Japan First Financial Technology Co., Ltd. (72) Inventor Ryuji Fukaya 1-5-1, Otemachi, Chiyoda-ku, Tokyo, Japan I Financial Technology Co., Ltd.

Claims (28)

[Claims] Claims 1. A lending portfolio optimization behavior simulation system for optimizing a lending portfolio, comprising: calculating a cash flow and a credit risk amount regarding a first lending which is a lending already held and held. A first means, a second means for calculating a cash flow and a credit risk amount relating to a second loan which is a loan for which a new contract is concluded in accordance with an assumed unique loan condition, the first means and the And a third means for executing an optimization calculation based on the calculation result of the second means. 2. The loan portfolio optimal behavior simulation according to claim 1, wherein the bankruptcy probability variability is used as a calculation parameter when calculating the credit risk amount in the first means and the second means. system. 3. The loan portfolio optimal behavior simulation system according to claim 1, wherein the credit risk amount in the first means and the second means is calculated using an actual allocation amount. . 4. The loan portfolio optimal behavior simulation system according to claim 1, wherein the optimization calculation by the first means, the second means, and the third means is performed over a plurality of periods. 5. The loan according to claim 1, wherein said first means, said second means, and said third means execute the calculation for each company group including a company to which the loan is made. Portfolio optimal behavior simulation system. 6. The loan portfolio optimization according to claim 1, further comprising credit risk amount decomposing means for decomposing a credit risk amount using correlation information assumed for a loan contract for each individual company. Behavior simulation system. 7. The loan portfolio optimal behavior simulation system according to claim 6, wherein the correlation information is a price-to-earnings ratio correlation by industry. 8. The loan portfolio optimal behavior simulation system according to claim 1, wherein the correlation relating to the loan contract for each company group is calculated using the industry-specific correlation information. 9. The loan portfolio optimal behavior simulation system according to claim 1, further comprising means for setting a constraint condition for said first and second loans. 10. The loan portfolio optimal behavior simulation system according to claim 9, wherein said constraint condition includes a constraint condition relating to the balance of the entire loan portfolio. 11. The loan portfolio optimal behavior simulation system according to claim 9, wherein said constraint includes a constraint relating to an upper limit and / or a lower limit of a balance by company group, by industry or by credit rating. . 12. The loan portfolio according to claim 9, wherein the constraint includes a constraint on an upper limit and / or a lower limit of a new loan disbursement amount for a balance by company group, industry, or credit rating. Optimal behavior simulation system. 13. The loan according to claim 9, wherein the constraint includes a constraint on an upper limit and / or a lower limit of a new loan disbursement amount with respect to a repayment amount by company group, industry, or credit rating. Portfolio optimal behavior simulation system. 14. A loan portfolio optimization behavior simulation method for optimizing a loan portfolio, the method comprising calculating a cash flow and a credit risk amount relating to a first loan which is a contract already held and held. And a second step of calculating a cash flow and a credit risk amount relating to a second loan which is a loan for which a new contract is to be concluded in accordance with an assumed unique loan condition; and the first step and the second step. And a third step of performing an optimization calculation based on the calculation result of the second step. 15. The bankruptcy probability variability is used as a calculation parameter when calculating the credit risk amount in the first step and the second step.
4. The loan portfolio optimal behavior simulation method described in 4. 16. The method according to claim 14, wherein the credit risk amount in the first step and the second step is calculated using an actual allocation amount. 17. The loan portfolio optimal behavior simulation method according to claim 14, wherein the optimization calculation in the first step, the second step, and the third step is performed over a plurality of periods. 18. The loan according to claim 14, wherein the first step, the second step, and the third step execute the calculation for each company group including a company to be a borrower. Portfolio optimal behavior simulation method. 19. A credit risk amount decomposing step for decomposing a credit risk amount using correlation information assumed for a loan contract of each individual company.
4. The loan portfolio optimal behavior simulation method described in 4. 20. The method according to claim 19, wherein the correlation information is a price-to-earnings ratio correlation by industry. 21. The method for simulating a loan portfolio optimal behavior according to claim 14, wherein the correlation regarding the loan contract for each company group is calculated using the correlation information for each industry. 22. The loan portfolio optimal behavior simulation method according to claim 14, further comprising means for setting a constraint condition for said first and second loans. 23. The loan portfolio optimal behavior simulation method according to claim 22, wherein said constraint condition includes a constraint condition relating to the balance of the entire loan portfolio. 24. The loan portfolio optimal behavior simulation method according to claim 22, wherein said constraint condition includes a constraint condition on an upper limit and / or a lower limit of a balance by company group, by industry or by credit rating. 25. The loan portfolio according to claim 22, wherein the constraint includes a constraint on an upper limit and / or a lower limit of a new loan disbursed amount for a balance by company group, industry, or credit rating. Optimal behavior simulation method. 26. The loan according to claim 22, wherein the constraint includes a constraint on an upper limit and / or a lower limit of a new loan disbursement amount with respect to a repayment amount by company group, industry, or credit rating. Portfolio optimal behavior simulation method. 27. A loan portfolio optimizing behavior simulation program for optimizing a loan portfolio, the program comprising a cash flow relating to a first loan which is a loan already held and held by a computer. And a first step of calculating a credit risk amount, and a second step of calculating a cash flow and a credit risk amount relating to a second loan, which is a loan for which a new contract is made according to assumed specific financing conditions. And a loan portfolio optimal behavior simulation program for executing a third step of executing an optimization calculation based on a calculation result of the first step and the second step. 28. A storage medium storing the program according to claim 27.