WO2012132355A1 - リスク管理装置 - Google Patents
リスク管理装置 Download PDFInfo
- Publication number
- WO2012132355A1 WO2012132355A1 PCT/JP2012/002006 JP2012002006W WO2012132355A1 WO 2012132355 A1 WO2012132355 A1 WO 2012132355A1 JP 2012002006 W JP2012002006 W JP 2012002006W WO 2012132355 A1 WO2012132355 A1 WO 2012132355A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- verification
- scenario data
- loss
- unit
- scenario
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Definitions
- the present invention relates to a risk management device, and more particularly to a risk management device having a function of statistically verifying the frequency of loss occurrence in scenario data that is input to a risk weighing device.
- risks such as earthquakes, system failures, clerical errors, and fraud. For this reason, it is required to measure the amount of risk using a risk weighing device and take measures against the risk.
- the risk weighing device inputs fragmentary information about an unknown risk profile in a company, and measures the characteristic value (for example, 99.9% value at risk (VaR)) of the risk profile of the company from this input data.
- the input data of the risk weighing device generally includes internal loss data and scenario data.
- Internal loss data is data related to loss events that actually occurred in the company.
- the internal loss data indicates how much loss has occurred for what kind of event.
- General risk weighing devices measure VaR using a technique called loss distribution technique (see, for example, Patent Document 1 and Non-Patent Document 1). Specifically, first, a loss frequency distribution is generated from the number of cases of internal loss data, and a loss scale distribution is generated from internal loss data and scenario data. Next, tens of thousands of processes to calculate the amount of loss per holding period by taking out the amount of loss for the number of losses generated using the above loss frequency distribution from the above loss size distribution by Monte Carlo simulation and adding them up. Generate a distribution of losses, repeated hundreds of thousands of times. Then, VaR of a predetermined confidence interval is calculated from the generated loss distribution.
- a loss frequency distribution is generated from the number of cases of internal loss data
- a loss scale distribution is generated from internal loss data and scenario data.
- tens of thousands of processes to calculate the amount of loss per holding period by taking out the amount of loss for the number of losses generated using the above loss frequency distribution from the above loss size distribution by Monte Carlo simulation and adding them up. Generate a distribution of
- the loss occurrence frequency in the scenario data described above is predicted using the following method (for example, see Non-Patent Document 1).
- a matrix consisting of a combination of risk assessment and internal control status assessment based on the number of occurrences per year of the actual loss-causing business and the scores for risk assessment and internal control status assessment conducted for that business.
- Create an average frequency evaluation table that describes the number of occurrences per year.
- the operational risk inherent in each business process is recognized as a scenario.
- the same risk and internal control status evaluation as described above is performed, and the average frequency evaluation table is subtracted from the risk evaluation score and the internal control status evaluation score. Number of occurrences of this event per year).
- the occurrence frequency can be estimated even in a scenario where there is no past loss record.
- An object of the present invention is to provide a risk management apparatus that solves the above-described problem, that is, the problem that there is no effective method for verifying the validity of the loss occurrence frequency of scenario data after the fact.
- the risk management device is: A plurality of verification units composed of one or more scenario data including a predicted value of loss occurrence frequency, a verification range that is a block of the plurality of verification units, and an actual number of occurrences of loss corresponding to the scenario data A memory for storing; and a processor connected to the memory; The processor Whether or not the total value of the number of loss occurrences corresponding to the scenario data included in the verification range follows a Poisson distribution that averages the total value of predicted loss occurrence frequencies in the scenario data included in the verification range. , Programmed to be determined using a goodness-of-fit test for Poisson distribution.
- the present invention Since the present invention has the above-described configuration, it is possible to verify the validity of the loss occurrence frequency of scenario data using actual loss cases.
- the risk management device 1 determines whether or not the frequency of loss of scenario data included in a plurality of verification units is appropriate for the entire verification unit, and shows an actual loss case. It has a function to use and verify.
- maintainability verification such verification is referred to as maintainability verification.
- the risk management apparatus 1 includes a communication interface unit (hereinafter referred to as a communication I / F unit) 11, an operation input unit 12, a screen display unit 13, a storage unit 14, and a processor 15 as main functional units.
- a communication interface unit hereinafter referred to as a communication I / F unit
- an operation input unit 12 a screen display unit 13
- a storage unit 14 a storage unit 14
- a processor 15 main functional units.
- the communication I / F unit 11 includes a dedicated data communication circuit and has a function of performing data communication with various devices (not shown) connected via a communication line (not shown).
- the operation input unit 12 includes an operation input device such as a keyboard and a mouse, and has a function of detecting an operator operation and outputting it to the processor 15.
- the screen display unit 13 includes a screen display device such as an LCD or a PDP, and has a function of displaying various information such as an operation menu and a verification result on the screen in response to an instruction from the processor 15.
- the storage unit 14 includes a storage device such as a hard disk or a semiconductor memory, and has a function of storing processing information and a program 14P necessary for various processes in the processor 15.
- the program 14P is a program that realizes various processing units by being read and executed by the processor 15, and is read by an external device (not shown) or a computer via a data input / output function such as the communication I / F unit 11. It is read in advance from a possible storage medium (not shown) and stored in the storage unit 14.
- main processing information stored in the storage unit 14 a plurality of verification units 14A1, a verification range 14A, that is, a loss occurrence number 14B, a first test condition 14C, and a first test result 14D, There is.
- the verification unit 14A1 is composed of one or more scenario data.
- One scenario data includes an identifier (ID) for uniquely identifying the scenario data and a predicted value of loss occurrence frequency. Since the predicted value of the loss occurrence amount is not used at the time of scenario data frequency verification, it may be removed from the scenario data.
- FIG. 2 shows a configuration example of the verification unit 14A1.
- the verification unit 14A1 in this example is composed of one scenario data.
- the predicted value ⁇ i of the loss occurrence frequency indicates the number of loss occurrences per year when the holding period is one year.
- the verification range 14A is a block of a plurality of verification units 14A1.
- FIG. 2 shows a configuration example of the verification range 14A.
- the verification range 14A in this example indicates that a set of scenario data having scenario IDs 1 to IDn is set as the verification range.
- the loss occurrence count 14B is data indicating the actual loss occurrence count corresponding to the scenario data.
- the loss occurrence number 14B is, for example, a set of a set of an identifier for identifying the corresponding scenario data and the number of loss occurrences per holding period.
- FIG. 2 shows a configuration example of the loss occurrence number 14B.
- the data in the first row of the loss occurrence number 14B in this example indicates that the number of loss occurrences per holding period of the scenario corresponding to the scenario ID 1 is one.
- the first test result 14D is data indicating the result of the first test process executed by the processor 15.
- the first test result 14D is one of three types of “conservative”, “valid”, and “non-conservative”.
- Conservative means that the frequency of loss of scenario data included in a plurality of verification units is larger than that estimated from the actual number of occurrences as a whole verification unit.
- Non-conservative means that, contrary to conservative, the frequency of loss of scenario data included in a plurality of verification units is less than that estimated from the actual number of occurrences as a whole verification unit.
- Valid means that it is neither conservative nor non-conservative, and the frequency of loss of scenario data included in a plurality of verification units is appropriate for the entire verification unit.
- the first test condition 14C indicates the condition of the first test process executed by the processor 15.
- FIG. 2 shows a configuration example of the first test condition 14C.
- the first test condition 14C in this example indicates that the first and second significance levels used in the first test process are ⁇ 11 and ⁇ 12.
- the first significance level ⁇ 11 is used for the maintainability determination
- the second significance level ⁇ 12 is used for the non-maintenance determination.
- the processor 15 has a microprocessor such as a CPU and its peripheral circuits, and reads and executes the program 14P from the storage unit 14, thereby causing the hardware and the program 14P to cooperate to implement various processing units. have.
- main processing units realized by the processor 15 there are an input storage unit 15A, a first test processing unit 15B, and an output unit 15C.
- the input storage unit 15A receives the verification unit 14A1, the verification range 14A, the loss occurrence number 14B, and the first test condition 14C from the communication I / F unit 11 or the operation input unit 12, and stores them in the storage unit 14. It has a function.
- the first verification processing unit 15B determines that the total value of the number of loss occurrences 14B corresponding to the scenario data included in each verification unit 14A1 in the verification range 14A is the loss in the scenario data included in each verification unit 14A1 in the verification range 14A. It has a function of determining whether or not to follow a Poisson distribution that averages the total value of the predicted values of occurrence frequency by using a test of the degree of fit with respect to the Poisson distribution.
- the first test processing unit 15B has a function of storing the test result in the storage unit 14 as the first test result 14D.
- the output unit 15C reads the verification result 14D of the first verification processing unit 15B from the storage unit 14, and outputs the result to the screen display unit 13 as the maintainability verification result of the entire verification unit, or externally through the communication I / F unit 11. Has a function to output.
- the input storage unit 15A communicates a plurality of verification units 14A1, a verification range 14A that is a block of the plurality of verification units 14A1, an actual loss occurrence number 14B corresponding to the scenario data, and a first test condition 14C.
- the data is input from the I / F unit 11 or the operation input unit 12 and stored in the storage unit 14 (step S1).
- the first verification processing unit 15B reads the plurality of verification units 14A1, the verification range 14A, the loss occurrence number 14B, and the first verification condition 14C from the storage unit 14, and converts them into scenario data included in the verification range 14A. Whether or not the total number of occurrences of the corresponding loss follows a Poisson distribution that averages the total of the predicted values of the loss occurrence frequency in the scenario data included in the verification range 14A is determined using a test of the degree of fitness for the Poisson distribution. The result is stored in the storage unit 14 (step S2).
- the output unit 15C reads the test result 14D of the first test processing unit 15B from the storage unit 14 and outputs it as a verification result to the screen display unit 13 or outputs it to the outside through the communication I / F unit 11 ( Step S3).
- FIG. 4 is a flowchart showing an example of the process of step S2 of FIG.
- an example of the process of the first test processing unit 15B will be described with reference to FIG.
- the first verification processing unit 15B calculates the total value ⁇ Ni of the number of loss occurrences corresponding to the scenario data included in the verification range 14A (step S11).
- the first verification processing unit 15B calculates the total value ⁇ i of the predicted values of the loss occurrence frequency in the scenario data included in the verification range 14A (step S12).
- the first test processing unit 15B sets the null hypothesis H0 and the alternative hypotheses H1 and H2 as follows (step S13).
- the null hypothesis H0 is “the total occurrence number ⁇ Ni follows a Poisson distribution with an average ⁇ i”.
- the alternative hypothesis H1 is “the average is smaller than ⁇ i (the scenario is conservative)”.
- the alternative hypothesis H2 is “average greater than ⁇ i (the scenario is non-conservative)”.
- the first test processing unit 15B assumes that the null hypothesis H0 is correct, and calculates thresholds n1 and n2 for comparison with the total value ⁇ Ni of the number of loss occurrences from the Poisson distribution of the mean ⁇ i (step S14).
- the threshold value n1 is a value at which the probability that the Poisson distribution of the mean ⁇ i takes a value less than or equal to n1 is greater than the significance level ⁇ 11, and the probability that the Poisson distribution takes a value less than (n1-1) is less than or equal to the significance level ⁇ 11. is there.
- the threshold n2 is a value at which the probability that the Poisson distribution of the mean ⁇ i takes a value greater than or equal to n2 is greater than the significance level ⁇ 12 and the probability that the Poisson distribution takes a value greater than or equal to (n2 + 1) is less than or equal to the significance level ⁇ 12.
- the first test processing unit 15B compares the total value ⁇ Ni of the number of loss occurrences with the thresholds n1 and n2 (steps S15 and S16), generates a test result according to the comparison result, and stores it in the storage unit 14.
- Store steps S17 to S19). That is, it is determined as “conservative” when ⁇ Ni ⁇ n1, “valid” when n1 ⁇ ⁇ Ni ⁇ n2, and “nonconservative” when n2 ⁇ Ni.
- the validity of the loss occurrence frequency is verified for the entire scenario group in which a plurality of scenario data is collected. Therefore, the frequency is low enough that the maintainability cannot be verified in a single scenario. However, accurate verification is possible. This point will be further described.
- the risk management device 2 according to the second embodiment of the present invention includes, in addition to the maintainability verification function of the risk management device 1 according to the first embodiment, between verification units in the verification range. It has a function to verify whether there is a bias in maintainability using actual loss cases.
- this latter verification is referred to as unbiased verification.
- the risk management device 2 includes a communication I / F unit 21, an operation input unit 22, a screen display unit 23, a storage unit 24, and a processor 25 as main functional units.
- the communication I / F unit 21, operation input unit 22, and screen display unit 23 have the same functions as the communication I / F unit 11, operation input unit 12, and screen display unit 13 of FIG. 1 in the first embodiment. is doing.
- the storage unit 24 includes a storage device such as a hard disk or a semiconductor memory, and has a function of storing processing information and programs 24P necessary for various processes in the processor 25.
- the program 24P is a program that realizes various processing units by being read and executed by the processor 25, and can be read by an external device (not shown) or a computer via a data input / output function such as the communication I / F unit 21. It is read in advance from a possible storage medium (not shown) and stored in the storage unit 24.
- the main processing information stored in the storage unit 24 includes a plurality of verification units 24A1, a verification range 24A, that is, the number of loss occurrences 24B, a first verification condition 24C, and a first verification result 24D. There are a second test condition 24E and a second test result 24E.
- the plurality of verification units 24A1, the verification range 24A, the number of loss occurrences 24B, the first test condition 24C, and the first test result 24D are the plurality of verification units 14A1, the verification range 14A, and the number of loss occurrences in the first embodiment. 14B, the first test condition 14C, and the first test result 14D.
- the second test result 24F is data indicating the result of the second test process executed by the processor 25.
- the second verification process 24F is either “no bias” or “with bias”.
- the second test condition 24E indicates the condition of the second test process executed by the processor 25.
- FIG. 6 shows a configuration example of the second test condition 24E.
- the second test condition 24E in this example indicates that the significance level used for the second test process is ⁇ 2.
- the processor 25 includes a microprocessor such as a CPU and its peripheral circuits, and reads and executes the program 24P from the storage unit 24, thereby realizing the various processing units in cooperation with the hardware and the program 24P. have.
- main processing units realized by the processor 25 there are an input storage unit 25A, a first test processing unit 25B, an output unit 25C, and a second test processing unit 25D.
- the input storage unit 25A inputs the verification unit 24A1, the verification range 24A, the loss occurrence number 24B, the first verification condition 24C, and the second verification condition 24E from the communication I / F unit 21 or the operation input unit 22. And has a function of storing in the storage unit 24.
- the first verification processing unit 25B has the same function as the first verification processing unit 15B of the risk management device 1 according to the first embodiment. That is, the total value of the number of loss occurrences 24B corresponding to the scenario data included in each verification unit 24A1 in the verification range 24A is the sum of the predicted values of the loss occurrence frequency in the scenario data included in each verification unit 24A1 in the verification range 24A. Whether or not to follow a Poisson distribution with an average value is determined by using a test for the degree of fitness for the Poisson distribution, and the result of the test is stored in the storage unit 24 as a first test result 24D.
- the number of occurrences of loss 24B corresponding to the scenario data for each verification unit 24A1 is the sum of the number of occurrences of loss corresponding to the scenario data included in the verification range 24A, and the ratio parameter.
- the multinomial distribution as a ratio of the total value of the predicted value of the loss occurrence frequency in the scenario data for each verification unit 24A1 to the total value of the predicted value of the loss occurrence frequency in the scenario data included in the verification range 24A. It has a function of determining using a fitness test for a multinomial distribution.
- the second test processing unit 25D has a function of storing the test result in the storage unit 24 as the second test result 24F.
- the output unit 25C reads the first test result 24D and the second test result 24F from the storage unit 24, and outputs them to the screen display unit 23 as the maintainability verification result of the entire verification unit and the unbiased verification result between the verification units. Alternatively, it has a function of outputting to the outside through the communication I / F unit 21.
- the input storage unit 25A includes a plurality of verification units 24A1, a verification range 24A that is a block of the plurality of verification units 24A1, an actual loss occurrence number 24B corresponding to the scenario data, a first test condition 24C, and a second Are input from the communication I / F unit 21 or the operation input unit 22 and stored in the storage unit 24 (step S21).
- the first verification processing unit 25B determines that the total number of loss occurrences corresponding to the scenario data included in the verification range 24A is the verification value. Whether or not to follow the Poisson distribution that averages the total value of the predicted values of the loss occurrence frequency in the scenario data included in the range 24A is determined using a test for the degree of fitness for the Poisson distribution, and the result is stored in the storage unit 24. (Step S22).
- the second verification processing unit 25D reads the plurality of verification units 24A1, the verification range 24A, the number of loss occurrences 24B, and the second verification condition 24E from the storage unit 24, and corresponds to the scenario data for each verification unit 24A1.
- the total number of loss occurrences corresponding to the scenario data whose total parameter is included in the verification range 24A is the total number of loss occurrences in the scenario data whose ratio parameter is included in the verification range 24A.
- Whether to follow the multinomial distribution which is the ratio of the total value of the predicted values of the loss occurrence frequency in the scenario data for each verification unit 24A1 to the value, is determined by using the fitness test for the multinomial distribution, and the result is stored in the storage unit 24. (Step S23).
- the output unit 25C reads the first test result 24C and the second test result 24F from the storage unit 24, and displays the screen display unit 23 as the maintainability verification result of the entire verification unit and the unbiased verification result between the verification units. Or output to the outside through the communication I / F unit 21 (step S24).
- FIG. 8 is a flowchart showing an example of the process of step S23 of FIG.
- an example of the process of the second test processing unit 25D will be described with reference to FIG.
- the second test processing unit 25D calculates the number k of verification units 24A1 (step S31). Next, for each verification unit, the second test processing unit 25D calculates the total values n1, n2,..., Nk of the number of loss occurrences corresponding to the scenario data included in the verification unit (step S32). Next, the second verification processing unit 25D calculates the total value ⁇ Ni of the number of loss occurrences corresponding to the scenario data included in the verification range 24A (step S33).
- the second test processing unit 25D calculates the ratio parameter predicted values p1, p2,..., Pk for each verification unit 24A1 (step S34).
- the ratio parameter pi of a certain verification unit 24A1 is the sum of the predicted values of the loss occurrence frequency of the scenario data included in the verification unit 24A1, and the total value of the predicted values of the loss occurrence frequency in the scenario data included in the verification range 24A. Calculated as a value divided by ⁇ i.
- the second test processing unit 25D sets the null hypothesis H0 and the alternative hypothesis H1 as follows (step S35).
- the null hypothesis H0 is “ratio parameters are p1, p2,..., Pk”.
- the alternative hypothesis H1 is “the ratio parameter is not p1, p2,..., Pk”.
- the second test processing unit 25D assumes that the null hypothesis H0 is correct, and in the multinomial distribution of the total parameter ⁇ Ni, the ratio parameters p1, p2,..., Pk, the realized values n1, n2,. , Nk realizes the probability px (step S36).
- the second test processing unit 25D determines that “k non-negative integers whose sum is ⁇ Ni” by combining all possible values in the multinomial distribution of the total parameter ⁇ Ni and the ratio parameters p1, p2,. (Step S37) Next, among these calculated probabilities, the sum of only the ones whose realized values n1, n2,. (Step S38).
- the second test processing unit 25D compares the calculated p value with the significance level ⁇ 2 (step S39). Then, the second test processing unit 25D generates a test result corresponding to the comparison result and stores it in the storage unit 24 (steps S40 and S41). That is, it is determined that “no bias” when p value ⁇ significant level ⁇ 2, and “no bias” when p value ⁇ significant level ⁇ 2.
- the present embodiment it is possible to verify the validity of the loss occurrence frequency of the scenario data with higher accuracy than the first embodiment by using actual loss cases.
- the reason for this is to verify the validity of the loss occurrence frequency for the entire scenario group that combines multiple scenario data, and to verify that there is no bias in maintainability among the verification units using actual loss cases. Because it is. This point will be further described.
- the scenario frequency is low enough that it is impossible to verify the maintainability by the verification unit alone.
- accurate verification is possible.
- scenario data for all verification units are combined into one, the scenario frequency for each verification unit is hidden. For this reason, if the total scenario frequency of the entire scenario group is the same, the result of the maintainability verification is the same.
- By performing unbiased verification it is possible to verify the maintainability bias between verification units, which cannot be verified by the maintainability verification.
- the risk management apparatus 3 according to the third embodiment of the present invention is verified in addition to the maintainability verification function and the unbiased verification function of the risk management apparatus 1 according to the second embodiment. It has a function of correcting the frequency of loss of scenario data based on the result.
- the risk management device 3 includes a communication I / F unit 31, an operation input unit 32, a screen display unit 33, a storage unit 34, and a processor 35 as main functional units.
- the communication I / F unit 31, the operation input unit 32, and the screen display unit 33 have the same functions as the communication I / F unit 21, the operation input unit 22, and the screen display unit 23 of FIG. 5 in the second embodiment. is doing.
- the storage unit 34 includes a storage device such as a hard disk or a semiconductor memory, and has a function of storing processing information and programs 34P necessary for various processes in the processor 35.
- the program 34P is a program that realizes various processing units by being read and executed by the processor 35.
- the program 34P can be read by an external device (not shown) or a computer via a data input / output function such as the communication I / F unit 31. It is read in advance from a possible storage medium (not shown) and stored in the storage unit 34.
- a plurality of verification units 34A that is, a verification range 34A, a loss occurrence number 34B, a first test condition 34C, a first test result 34D, a second test condition 34E, and a second test
- the result 34F includes a plurality of verification units 24A1, a verification range 24A, a loss occurrence number 24B, a first test condition 24C, a first test result 24D, a second test condition 24E, and a second test in the second embodiment.
- the result is the same as 24F.
- the processor 35 has a microprocessor such as a CPU and its peripheral circuits, and reads and executes the program 34P from the storage unit 34, thereby causing the hardware and the program 34P to cooperate to implement various processing units.
- the main processing units realized by the processor 35 include an input storage unit 35A, a first test processing unit 35B, an output unit 35C, a second test processing unit 35D, and a correction unit 35E.
- the input storage unit 35A, the first verification processing unit 35B, and the second verification processing unit 35D are the input storage unit 25A, the first verification processing unit 25B, and the second verification processing unit 25D in the second embodiment. Has the same function.
- the correction unit 35E reads the conservative verification test result 34D and the unbiased verification test result 34F from the storage unit 34, and based on these two test results, the correction unit 34A1 corrects the predicted value of the loss occurrence frequency. And has a function of correcting the predicted value of the loss occurrence frequency of the scenario data in the determined verification unit 34A1.
- the correction unit 35E has a function of storing the corrected scenario data in the storage unit 34.
- the correction unit 35E may overwrite the scenario data before correction with the scenario data after correction, or may store the scenario data after correction in the storage unit 34 separately from the scenario data before correction.
- the correction unit 35E has a function of resuming the processing from the first verification processing unit 35B when the correction is performed on at least one scenario data.
- the output unit 35C reads the first test result 34C, the second test result 34F, and the scenario data after correction from the storage unit 34, the maintainability verification result of the entire verification unit, the unbiased verification result between the verification units, In addition, it has a function of outputting the correction contents to the screen display unit 33 or outputting the correction contents to the outside through the communication I / F unit 31.
- the input storage unit 35A has a plurality of verification units 34A1, a verification range 34A that is a group of the plurality of verification units 34A1, and an actual loss corresponding to scenario data.
- the number of occurrences 34B, the first verification condition 34C, and the second verification condition 34E are input from the communication I / F unit 31 or the operation input unit 32 and stored in the storage unit 34 (step S51).
- the first verification processing unit 35B determines that the total number of loss occurrences corresponding to the scenario data included in the verification range 34A is the verification value. Whether or not to follow a Poisson distribution that averages the sum of predicted values of loss occurrence frequencies in the scenario data included in the range 34A is determined using a test of the degree of fit with respect to the Poisson distribution, and the result is stored in the storage unit 34. (Step S52).
- the second verification processing unit 35D verifies the total parameter by the number of occurrences of loss 24B corresponding to the scenario data for each verification unit 34A1.
- the scenario data for each verification unit 34A1 with respect to the total value of the occurrence number 34B of loss corresponding to the scenario data included in the range 34A and the total value of the predicted value of the loss occurrence frequency in the scenario data included in the verification range 34A Whether or not to follow the multinomial distribution as a ratio of the total value of the predicted values of the loss occurrence frequency is determined by using a fitness test for the multinomial distribution, and the result is stored in the storage unit 34 (step S53).
- the correction unit 35E determines a verification unit 34A1 for correcting the predicted value of the loss occurrence frequency based on the conservative verification test result 34D and the unbiased verification test result 34F, and the verification unit 34A1 thus determined.
- the predicted value of the loss occurrence frequency of scenario data is corrected, and the corrected scenario data is stored in the storage unit 34 (step S54).
- the correction unit 35E determines whether or not correction has been performed on at least one scenario data (step S55), and when the correction has been performed, returns control to the first verification processing unit 35B. As a result, the same maintainability verification and unbiased verification as described above are performed again using the corrected scenario data, and then the correction processing by the correction unit 35E is executed. Such processing is repeated until there is no scenario data to be corrected. On the other hand, when the scenario data is not corrected, the correction unit 35E passes control to the output unit 35C.
- the output unit 35C reads the first test result 34C, the second test result 34F, and the scenario data after correction from the storage unit 34, the maintainability verification result of the entire verification unit, the unbiased verification result between the verification units, And it outputs to the screen display part 33 as a correction content, or outputs outside through the communication I / F part 31 (step S56).
- FIG. 11 is a flowchart showing an example of the process of step S54 of FIG.
- an example of the process of step S54 executed by the correction unit 35E will be described with reference to FIG.
- the correction unit 35E determines whether the maintainability verification result 34D is “conservative”, “valid”, or “non-conservative”, and the unbiased verification result 34F indicates “no bias” or “bias”. It is determined whether it is “present” (steps S61 to S64). Then, it is classified into the following six cases according to the determination result, and correction processing corresponding to each case is executed (steps S65 to S70).
- Case 1 Conservative and non-biased In this case, the correction unit 35E performs correction to reduce the predicted value of loss occurrence frequency in the scenario data included in all the verification units 34A1 (step S65).
- Case 2 Conservative and biased In this case, the correction unit 35E performs correction to reduce the predicted value of the loss occurrence frequency in the scenario data included in the most conservative verification unit 34A1 among all the verification units. (Step S66).
- Case 3 Reasonable maintainability and no bias In this case, the correction unit 35E determines that there is no need for correction (step S67).
- Case 4 Reasonable maintainability and bias
- the correction unit 35E increases the predicted value of the loss occurrence frequency in the scenario data included in the most non-conservative verification unit 34A1 among all the verification units. Correction is performed (step S68).
- Case 5 Non-conservative and non-biased
- the correction unit 35E performs correction to increase the predicted value of the loss occurrence frequency in the scenario data included in all the verification units 34A1 (step S69).
- Case 6 Non-conservative and biased In this case, as in Case 4, the correcting unit 35E determines the frequency of loss occurrence in the scenario data included in the most non-conservative verification unit 34A1 among all the verification units. Correction for increasing the predicted value is performed (step S70).
- the correction unit 35E calculates an estimated value of maintainability for each verification unit and determines the relative maintainability and non-maintainability between the verification units based on the magnitude.
- the verification unit with the lowest probability is the most conservative verification unit, and the largest verification unit is the most non-conservative verification unit.
- the correction unit 35E follows a predetermined rule as to how much the predicted value is increased or decreased by the correction.
- a correction rule for example, a rule of increasing or decreasing by a predetermined ratio (for example, 30%) of the predicted value before correction may be used.
- a rule is used in which the frequency of loss occurrence that can be taken as a predicted value is corrected so that the frequency before correction in the frequency table arranged in descending order becomes a value that is higher or lower by 1 rank or 2 ranks. It's okay.
- the same effects as those of the second embodiment can be obtained, and the scenario data loss can be caused when an invalid verification result is obtained from the verification results of the maintainability verification and the unbiased verification.
- the frequency of occurrence can be automatically corrected.
- the risk management device 4 according to the fourth exemplary embodiment of the present invention includes the maintainability verification function, the unbiased verification function, and the correction function that the risk management device 3 according to the third exemplary embodiment has.
- scenario data to be verified is extracted from the scenario data group, and a verification unit and a verification range that is a set thereof are set.
- the risk management device 4 includes a communication I / F unit 41, an operation input unit 42, a screen display unit 43, a storage unit 44, and a processor 45 as main functional units.
- the communication I / F unit 41, the operation input unit 42, and the screen display unit 43 have the same functions as the communication I / F unit 31, the operation input unit 32, and the screen display unit 33 of FIG. 9 in the third embodiment. is doing.
- the storage unit 44 includes a storage device such as a hard disk or a semiconductor memory, and has a function of storing processing information and programs 44P necessary for various processes in the processor 45.
- the program 44P is a program that realizes various processing units by being read and executed by the processor 45, and can be read by an external device (not shown) or a computer via a data input / output function such as the communication I / F unit 41. It is read in advance from a possible storage medium (not shown) and stored in the storage unit 44.
- a scenario data group 44G As main processing information stored in the storage unit 44, a scenario data group 44G, a plurality of verification units 44A1, a verification range 44A, that is, a loss occurrence number 44B, a first test condition 44C, a first Test result 44D, second test condition 44E, and second test result 44F.
- the scenario data group 44G is composed of a plurality of scenario data.
- One scenario data includes an identifier (ID) for uniquely identifying the scenario data, a predicted value of loss occurrence frequency, the type of loss event, the department that created the scenario, and the department where the scenario is assumed. It consists of related departments that represent etc.
- FIG. 13 shows a configuration example of the scenario data group 44G.
- the scenario data group 44G in this example is composed of m scenario data.
- the predicted value ⁇ i of the loss occurrence frequency indicates the number of loss occurrences per year when the holding period is one year.
- the type of loss event is, for example, a system failure, fraud, or earthquake.
- a plurality of verification units 44A1 that is, a verification range 44A, a loss occurrence number 44B, a first test condition 44C, a first test result 44D, a second test condition 44E, and a second test
- the result 44F includes a plurality of verification units 34A1, a verification range 34A, a loss occurrence number 34B, a first test condition 34C, a first test result 34D, a second test condition 34E, and a second test in the third embodiment.
- the result is the same as 34F.
- the plurality of verification units 34A1 and the verification range 34A are data given as input information
- the plurality of verification units 44A1 and the verification range 44A of this embodiment are dynamically derived from the scenario data group 44G. It is different in that it is generated data.
- the processor 45 includes a microprocessor such as a CPU and peripheral circuits thereof, and reads and executes the program 44P from the storage unit 44, thereby realizing various processing units by cooperating the hardware and the program 44P. have.
- main processing units realized by the processor 45 there are an input storage unit 45A, a first verification processing unit 45B, an output unit 45C, a second verification processing unit 45D, a correction unit 45E, and a verification target setting unit 45F.
- the input storage unit 45A inputs the scenario data group 44G, the number of loss occurrences 44B, the first verification condition 44C, and the second verification condition 44E from the communication I / F unit 41 or the operation input unit 42, and the storage unit 44.
- the verification target setting unit 45F has a function of extracting a plurality of scenario data as a verification range from the scenario data group 44G, and further classifying the extracted plurality of scenario data into a plurality of verification units.
- FIG. 14 is an example of a verification range and a verification unit.
- the unit of the verification range is each part, and the verification unit is a scenario.
- the unit of the verification range is each business department, and the verification unit is a section. According to this setting 2, for example, if attention is paid to the first business department having the first sales department and the second sales department, the scenario data in which the related department in FIG. 13 is the first sales department or the second sales department.
- the set is a verification range, and among the sets, a set of scenario data whose related department is the first sales department and a set of scenario data whose related department is the second sales department are each one verification unit. Furthermore, in setting 3, the unit of the verification range is each business department, and the verification unit is the type of loss event. According to this setting 3, for example, when focusing on the above-mentioned first business department, a set of scenario data in which the related department in FIG. 13 is the first sales department or the second sales department is the verification range. A set of scenario data having the same type of loss event is one verification unit.
- the verification target setting unit 45F calculates a verification unit 44A1 and a verification range 44A that is a set thereof from the scenario data group 44G according to the defined settings, and stores them in the storage unit 44.
- processing is performed according to the defined order. For example, when the first ranking is set to each part as a verification range and a scenario as a verification unit, a verification range 44A and a verification unit 44A1 are created for each part such as an existing sales department, and verification and correction are sequentially performed. Process.
- the verification range 44A and the verification unit 44A1 are created according to the setting of the next order. Such processing is repeated for all the defined settings.
- the order of processing is as shown in the order of setting 1 and setting 2 in FIG. 14 in order to avoid that the part that is creating the correct scenario is subject to correction due to the influence of other parts that are not creating the scenario correctly.
- a bottom-up method that is preferentially implemented from a narrower verification range is preferable.
- the verification range is narrower than that of the verification range, if correction is performed during verification of the verification range in the middle, the correction results in conflict with the verification range of the narrower range that was previously performed. Therefore, when correction is performed in the middle, it is desirable to perform verification again from the narrowest verification range.
- the first verification processing unit 45B, the second verification processing unit 45D, the correction unit 45E, and the output unit 45C are the first verification processing unit 35B, the second verification processing unit 35D, and the correction unit in the third embodiment. 35E and the same function as the output unit 35C.
- the first and second test processing units 45B and 45D store the first and second test results 44D and 44F in a storage unit so as to clearly distinguish which setting is the test result for the verification range. 44.
- the output unit 45C outputs the first and second test results 44D and 44F to the storage unit 44 so as to clearly distinguish which setting is the test result for the verification range.
- the input storage unit 45A receives the scenario data group 44G, the actual loss occurrence number 44B corresponding to the scenario data, the first verification condition 44C, and the second verification condition 44E, from the communication I / F unit 41 or the operation input. Input from the unit 42 and stored in the storage unit 44 (step S81).
- the verification target setting unit 45F pays attention to the definition of the first order setting to be processed first (step S82).
- the verification target setting unit 45F extracts a plurality of scenario data as a verification range from the scenario data group 44G according to the definition of the setting under attention, and further classifies the extracted plurality of scenario data into a plurality of verification units.
- a verification unit 44A1 and a verification range 44A that is a group of the verification units 44A1 are generated and stored in the storage unit 44 (step S83).
- the first verification processing unit 45B stores a plurality of verification units 44A1 generated by the verification target setting unit 45F, a verification range 44A that is a collection thereof, a loss occurrence number 44B, and a first verification condition 44C.
- the total number of occurrences of loss corresponding to the scenario data included in the verification range 44A is the same as that in the scenario data included in the verification range 44A, as in the first verification processing unit 35B in the third embodiment.
- Whether or not to follow the Poisson distribution that averages the sum of the predicted values of the loss occurrence frequency is determined using a test of the degree of fit with respect to the Poisson distribution, and the result is stored in the storage unit 44 (step S84).
- the second verification processing unit 45D verifies the total parameter by the number of occurrences of loss 44B corresponding to the scenario data for each verification unit 44A1.
- the scenario data for each verification unit 44A1 with respect to the total value of the number of occurrences of loss 44B corresponding to the scenario data included in the range 44A and the total value of the predicted values of the loss occurrence frequency in the scenario data included in the verification range 44A Whether or not to follow the multinomial distribution as a ratio of the total value of the predicted values of the loss occurrence frequency is determined by using a fitness test for the multinomial distribution, and the result is stored in the storage unit 44 (step S85).
- the correction unit 45E corrects the predicted value of the loss occurrence frequency based on the test result 44D for maintainability verification and the test result 44F for unbiased verification.
- the verification unit 44A1 is determined, the predicted value of the scenario data loss occurrence frequency in the determined verification unit 44A1 is corrected, and the corrected scenario data is stored in the storage unit 44 (step S86).
- the correction unit 45E determines whether or not correction has been performed on at least one scenario data (step S87), and if correction has been performed, returns control to the first verification processing unit 45B. Thereby, using the scenario data after correction, the same maintainability verification and unbiased verification as those described above are performed again for the verification range 44A being processed, and then the correction processing by the correction unit 45E is performed. Is executed. Such processing is repeated until there is no scenario data to be corrected. On the other hand, when the scenario data is not corrected, the correction unit 45E returns control to the verification target setting unit 45F.
- the verification target setting unit 45F determines whether or not there remains a verification range that has not been processed regarding the definition of the setting under attention (step S88), and if it remains, returns to the processing of step S83.
- the verification unit 44A1 and the verification range 44A that is a group of the verification unit 44A1 are generated for the setting range that has not yet been processed in accordance with the definition of the setting under attention, and maintainability verification and unbiased verification are performed on the verification range 44A. And correction processing are executed.
- the verification target setting unit 45F determines whether the definition of the setting under attention is the first and only definition (step S89). If the definition of the setting under attention is the first and only definition, the verification target setting unit 45F passes control to the output unit 45C. In addition, if the definition of the setting under attention is not the first and only definition, the verification target setting unit 45F determines whether or not the scenario data has been corrected in the process of defining the setting under attention (step S90). ). If correction has been performed, the verification target setting unit 45F returns to the process of step S82. Thereby, the verification is repeated from the definition of the first setting again.
- the verification target setting unit 45F determines whether there is an unprocessed setting definition (step S91), and if so, pay attention to the definition of the setting to be processed next. Transfer (step S92), the process returns to step S83. Thereby, regarding the definition of the next setting, the same process as the process for the definition of the previous setting is repeated. If there is no definition of unprocessed settings, the verification target setting unit 45F passes control to the output unit 45C.
- the output unit 45C reads the first test result 44D, the second test result 44F, and the scenario data after correction from the storage unit 44, the maintainability verification result of the entire verification unit, the unbiased verification result between the verification units, And as a correction content, it outputs to the screen display part 43 according to each setting, or outputs outside through the communication I / F part 41 (step S93).
- the same effects as those of the third embodiment can be obtained, and the verification range 44A and the verification unit 44A1 can be automatically generated. be able to.
- the present invention has been described with reference to some embodiments, the present invention is not limited to the above embodiments, and various other additions and modifications can be made.
- the present invention can also be applied to risks other than operational risks, such as credit risks associated with credit transactions such as lending operations and market risks associated with foreign exchange and interest rate transactions.
- the present invention also includes the following embodiments.
- the significance level of maintainability verification and unbiased verification is set to a fixed value.
- those significance levels may be variable values.
- conservativeness verification, unbiasedness verification and correction processing are performed at the first significance level, and then conservativeness verification and unbiasedness verification are performed at a second significance level larger than the first significance level, Only the verification result based on the second significance level may be output.
- the verification results of the maintainability verification and the unbiased verification are classified into six cases, and correction is automatically performed for five cases other than cases with reasonable maintainability and no bias. I made it.
- only one of these five cases is non-conservative and non-biased, or only two cases are non-conservative and non-biased and non-conservative and non-biased, or non-conservative and non-biased.
- correction may be automatically performed only for the three cases of non-conservative and biased and appropriate maintainability and biased.
- the present invention can be used for the purpose of verifying the validity of the predicted value of the loss occurrence frequency in the scenario data used as input information of the risk weighing device, and correcting the predicted value according to the verification result.
- a part or all of the above embodiments can be described as in the following supplementary notes, but is not limited thereto.
- Appendix 1 A plurality of verification units composed of one or more scenario data including a predicted value of loss occurrence frequency, a verification range that is a block of the plurality of verification units, and the actual number of occurrences of loss corresponding to the scenario data Storage means for storing; Whether or not the total value of the number of occurrences of loss corresponding to the scenario data included in the verification range follows a Poisson distribution that averages the total value of predicted loss occurrence frequencies in the scenario data included in the verification range. And a first test processing means for determining using a degree of fit test for Poisson distribution.
- the number of loss occurrences corresponding to the scenario data for each verification unit is the total value of the previous period loss corresponding to the scenario data whose total parameter is included in the verification range, and the scenario whose ratio parameter is included in the verification range Test the fitness of the multinomial distribution to determine whether to follow the multinomial distribution, which is the ratio of the total predicted value of the loss occurrence frequency in the scenario data for each verification unit to the total predicted value of the loss occurrence frequency in the data.
- a correction unit is provided that determines a verification unit for correcting a predicted value of loss occurrence frequency in scenario data based on a test result of goodness of fit for the Poisson distribution and a test result of goodness of fit for the multinomial distribution.
- the risk management device according to attachment 2.
- the risk management apparatus according to appendix 3, wherein the correction unit corrects a predicted value of loss occurrence frequency in scenario data included in the determined verification unit.
- the storage means stores a scenario data group composed of a plurality of scenario data including a predicted value of loss occurrence frequency,
- the risk management apparatus according to any one of appendices 1 to 4, further comprising verification target setting means for extracting the verification range and the plurality of verification units from the scenario data group.
- a plurality of verification units composed of one or more scenario data including a predicted value of loss occurrence frequency, a verification range that is a block of the plurality of verification units, and the actual number of occurrences of loss corresponding to the scenario data
- a risk management method executed by a risk management device comprising a storage means for storing and a first verification processing means, The first test processing means averages the total value of the number of occurrences of loss corresponding to the scenario data included in the verification range, and the total value of the predicted values of loss occurrence frequency in the scenario data included in the verification range.
- a risk management method characterized by deciding whether or not to follow a Poisson distribution using a test of goodness of fit against the Poisson distribution.
- the risk management device comprises a second verification processing means,
- the second verification processing means is configured such that the number of loss occurrences corresponding to the scenario data for each verification unit is the sum of the number of loss occurrences in the previous period corresponding to the scenario data whose total parameter is included in the verification range.
- a parameter is a ratio of a total value of predicted values of loss occurrence frequency in scenario data for each verification unit to a total value of predicted values of loss occurrence frequency in scenario data included in the verification range.
- the risk management apparatus includes a correction unit, The correction means determines a verification unit for correcting a predicted value of loss occurrence frequency in scenario data based on a test result of goodness of fit for the Poisson distribution and a test result of goodness of fit for the multinomial distribution.
- a plurality of verification units composed of one or more scenario data including a predicted value of loss occurrence frequency, a verification range that is a block of the plurality of verification units, and the actual number of occurrences of loss corresponding to the scenario data
- a computer having storage means for storing; Whether or not the total value of the number of occurrences of loss corresponding to the scenario data included in the verification range follows a Poisson distribution that averages the total value of predicted loss occurrence frequencies in the scenario data included in the verification range.
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Technology Law (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Educational Administration (AREA)
- Debugging And Monitoring (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、上記複数の検証単位のあつまりである検証範囲と、上記シナリオデータに対応する実際の損失の発生件数とを記憶するメモリと、上記メモリに接続されたプロセッサとを備え、
上記プロセッサは、
上記検証範囲に含まれるシナリオデータに対応する上記損失の発生件数の合計値が、上記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する
ようにプログラムされている、といった構成を採る。
[第1の実施形態]
図1を参照すると、本発明の第1の実施形態にかかるリスク管理装置1は、複数の検証単位に含まれるシナリオデータの損失発生頻度が検証単位全体として妥当か否かを実際の損失事例を用いて検証する機能を有している。以下、このような検証を保守性検証と言う。
図5を参照すると、本発明の第2の実施形態にかかるリスク管理装置2は、第1の実施形態にかかるリスク管理装置1の有する保守性検証機能に加えて、検証範囲において検証単位間で保守性に偏りがないかを実際の損失事例を用いて検証する機能を有している。以下、この後者の検証を不偏性検証と言う。
図9を参照すると、本発明の第3の実施形態にかかるリスク管理装置3は、第2の実施形態にかかるリスク管理装置1の有する保守性検証機能と不偏性検証機能とに加えて、検証結果に基づいてシナリオデータの損失発生頻度を補正する機能を有している。
(ステップS53)。
この場合、補正部35Eは、全ての検証単位34A1に含まれるシナリオデータにおける損失発生頻度の予測値を減少させる補正を行う(ステップS65)。
(2)ケース2:保守的かつ偏り有り
この場合、補正部35Eは、全ての検証単位のうち最も保守的な検証単位34A1に含まれるシナリオデータにおける損失発生頻度の予測値を減少させる補正を行う(ステップS66)。
(3)ケース3:妥当な保守性かつ偏り無し
この場合、補正部35Eは、補正の必要性は無いと判断する(ステップS67)。
(4)ケース4:妥当な保守性かつ偏り有り
この場合、補正部35Eは、全ての検証単位のうち最も非保守的な検証単位34A1に含まれるシナリオデータにおける損失発生頻度の予測値を増加させる補正を行う(ステップS68)。
(5)ケース5:非保守的かつ偏り無し
この場合、補正部35Eは、全ての検証単位34A1に含まれるシナリオデータにおける損失発生頻度の予測値を増加させる補正を行う(ステップS69)。
(6)ケース6:非保守的かつ偏り有り
この場合、補正部35Eは、ケース4と同様に、全ての検証単位のうち最も非保守的な検証単位34A1に含まれるシナリオデータにおける損失発生頻度の予測値を増加させる補正を行う(ステップS70)。
図12を参照すると、本発明の第4の実施形態にかかるリスク管理装置4は、第3の実施形態にかかるリスク管理装置3の有する保守性検証機能と不偏性検証機能と補正機能とに加えて、シナリオデータ群から検証を行う対象とするシナリオデータを抽出し、検証単位とその集合である検証範囲を設定する機能を有している。
以上、本発明を幾つかの実施形態を挙げて説明したが、本発明は以上の実施形態にのみ限定されず、その他各種の付加変更が可能である。例えば、貸出業務などの信用取引にかかる信用リスクや、為替および金利取引にかかる市場リスクなど、オペレーショナルリスク以外のリスクに対しても本発明は適用可能である。また、本発明は以下のような実施形態も含まれる。
[付記1]
損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶する記憶手段と、
前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する第1の検定処理手段と
を備えることを特徴とするリスク管理装置。
[付記2]
前記検証単位毎のシナリオデータに対応する前記損失の発生件数が、合計パラメータを前記検証範囲に含まれるシナリオデータに対応する前期損失の発生件数の合計値、比率パラメータを前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値に対する、前記検証単位毎のシナリオデータにおける損失発生頻度の予測値の合計値の割合とする多項分布に従うか否かを、多項分布に対する適合度の検定を用いて決定する第2の検定処理手段
を備えることを特徴とする付記1に記載のリスク管理装置。
[付記3]
前記ポアソン分布に対する適合度の検定結果と前記多項分布に対する適合度の検定結果とに基づいて、シナリオデータにおける損失発生頻度の予測値を補正する検証単位を決定する補正手段
を備えることを特徴とする付記2に記載のリスク管理装置。
[付記4]
前記補正手段は、前記決定した検証単位に含まれるシナリオデータにおける損失発生頻度の予測値を補正する
ことを特徴とする付記3に記載のリスク管理装置。
[付記5]
前記記憶手段は、損失発生頻度の予測値を含む複数のシナリオデータから構成されるシナリオデータ群を記憶し、
前記シナリオデータ群から前記検証範囲と前記複数の検証単位とを抽出する検証対象設定手段
を備えることを特徴とする付記1乃至4の何れかに記載のリスク管理装置。
[付記6]
損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶する記憶手段と、第1の検定処理手段とを備えたリスク管理装置が実行するリスク管理方法であって、
前記第1の検定処理手段が、前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する
ことを特徴とするリスク管理方法。
[付記7]
前記リスク管理装置が、第2の検定処理手段を備え、
前記第2の検定処理手段が、前記検証単位毎のシナリオデータに対応する前記損失の発生件数が、合計パラメータを前記検証範囲に含まれるシナリオデータに対応する前期損失の発生件数の合計値、比率パラメータを前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値に対する、前記検証単位毎のシナリオデータにおける損失発生頻度の予測値の合計値の割合とする多項分布に従うか否かを、多項分布に対する適合度の検定を用いて決定する
ことを特徴とする付記6に記載のリスク管理方法。
[付記8]
前記リスク管理装置が、補正手段を備え、
前記補正手段が、前記ポアソン分布に対する適合度の検定結果と前記多項分布に対する適合度の検定結果とに基づいて、シナリオデータにおける損失発生頻度の予測値を補正する検証単位を決定する
ことを特徴とする付記7に記載のリスク管理方法。
[付記9]
前記補正手段が、前記決定した検証単位に含まれるシナリオデータにおける損失発生頻度の予測値を補正する
ことを特徴とする付記8に記載のリスク管理方法。
[付記10]
損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶する記憶手段を有するコンピュータを、
前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する第1の検定処理手段
として機能させるためのプログラム。
11、21、31、41…通信I/F部
12、22、32、42…操作入力部
13、23、33、43…画面表示部
14、24、34、44…記憶部
15、25、35、45…プロセッサ
Claims (10)
- 損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶するメモリと、前記メモリに接続されたプロセッサとを備え、
前記プロセッサは、
前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する
ようにプログラムされていることを特徴とするリスク管理装置。 - 前記プロセッサは、さらに、
前記検証単位毎のシナリオデータに対応する前記損失の発生件数が、合計パラメータを前記検証範囲に含まれるシナリオデータに対応する前期損失の発生件数の合計値、比率パラメータを前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値に対する、前記検証単位毎のシナリオデータにおける損失発生頻度の予測値の合計値の割合とする多項分布に従うか否かを、多項分布に対する適合度の検定を用いて決定するようにプログラムされていることを特徴とする請求項1に記載のリスク管理装置。 - 前記プロセッサは、さらに、
前記ポアソン分布に対する適合度の検定結果と前記多項分布に対する適合度の検定結果とに基づいて、シナリオデータにおける損失発生頻度の予測値を補正する検証単位を決定する
ようにプログラムされていることを特徴とする請求項2に記載のリスク管理装置。 - 前記プロセッサは、さらに、
前記決定した検証単位に含まれるシナリオデータにおける損失発生頻度の予測値を補正する
ようにプログラムされていることを特徴とする請求項3に記載のリスク管理装置。 - 前記メモリは、さらに、損失発生頻度の予測値を含む複数のシナリオデータから構成されるシナリオデータ群を記憶し、
前記プロセッサは、さらに、
前記シナリオデータ群から前記検証範囲と前記複数の検証単位とを抽出する
ようにプログラムされていることを特徴とする請求項1乃至4の何れかに記載のリスク管理装置。 - 損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶するメモリと、前記メモリに接続されたプロセッサとを備えたリスク管理装置が実行するリスク管理方法であって、
前記プロセッサが、
前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定する
ことを特徴とするリスク管理方法。 - 前記プロセッサが、さらに、
前記検証単位毎のシナリオデータに対応する前記損失の発生件数が、合計パラメータを前記検証範囲に含まれるシナリオデータに対応する前期損失の発生件数の合計値、比率パラメータを前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値に対する、前記検証単位毎のシナリオデータにおける損失発生頻度の予測値の合計値の割合とする多項分布に従うか否かを、多項分布に対する適合度の検定を用いて決定することを特徴とする請求項6に記載のリスク管理方法。 - 前記プロセッサが、さらに、
前記ポアソン分布に対する適合度の検定結果と前記多項分布に対する適合度の検定結果とに基づいて、シナリオデータにおける損失発生頻度の予測値を補正する検証単位を決定する
ことを特徴とする請求項7に記載のリスク管理方法。 - 前記プロセッサが、さらに、
前記決定した検証単位に含まれるシナリオデータにおける損失発生頻度の予測値を補正する
ことを特徴とする請求項8に記載のリスク管理方法。 - 損失発生頻度の予測値を含む1以上のシナリオデータから構成される複数の検証単位と、前記複数の検証単位のあつまりである検証範囲と、前記シナリオデータに対応する実際の損失の発生件数とを記憶するメモリに接続されたプロセッサに、
前記検証範囲に含まれるシナリオデータに対応する前記損失の発生件数の合計値が、前記検証範囲に含まれるシナリオデータにおける損失発生頻度の予測値の合計値を平均とするポアソン分布に従うか否かを、ポアソン分布に対する適合度の検定を用いて決定するステップ
を行わせるためのプログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/977,999 US20130311231A1 (en) | 2011-03-29 | 2012-03-23 | Risk management device |
EP12764199.1A EP2693379A4 (en) | 2011-03-29 | 2012-03-23 | RISK MANAGEMENT DEVICE |
KR1020137018339A KR101368103B1 (ko) | 2011-03-29 | 2012-03-23 | 리스크 관리 디바이스 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011072747A JP5697146B2 (ja) | 2011-03-29 | 2011-03-29 | リスク管理装置 |
JP2011-072747 | 2011-03-29 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012132355A1 true WO2012132355A1 (ja) | 2012-10-04 |
Family
ID=46930136
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/002006 WO2012132355A1 (ja) | 2011-03-29 | 2012-03-23 | リスク管理装置 |
Country Status (5)
Country | Link |
---|---|
US (1) | US20130311231A1 (ja) |
EP (1) | EP2693379A4 (ja) |
JP (1) | JP5697146B2 (ja) |
KR (1) | KR101368103B1 (ja) |
WO (1) | WO2012132355A1 (ja) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11373189B2 (en) | 2014-03-27 | 2022-06-28 | EMC IP Holding Company LLC | Self-learning online multi-layer method for unsupervised risk assessment |
CN106600424A (zh) * | 2016-12-15 | 2017-04-26 | 平安科技(深圳)有限公司 | 数据核实的提示方法及装置 |
US10388652B2 (en) | 2017-11-14 | 2019-08-20 | Globalfoundries Inc. | Intergrated circuit structure including single diffusion break abutting end isolation region, and methods of forming same |
US10403548B2 (en) | 2017-11-14 | 2019-09-03 | Globalfoundries Inc. | Forming single diffusion break and end isolation region after metal gate replacement, and related structure |
KR102248319B1 (ko) * | 2018-09-10 | 2021-05-03 | 김해동 | 주식 청약 대금 대출을 위한 금융 기술 서비스 방법 및 그 장치 |
KR101986954B1 (ko) * | 2018-09-10 | 2019-06-07 | 김해동 | P2p금융을 활용한 청약 대금 납입을 위한 금융 기술 서비스 방법 및 그 장치 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003036346A (ja) * | 2001-07-23 | 2003-02-07 | Mitsubishi Trust & Banking Corp | オペレーショナル・リスク評価方法及びそのシステム |
JP2004252893A (ja) * | 2003-02-21 | 2004-09-09 | Fujitsu Ltd | オペレーショナルリスク計量装置およびオペレーショナルリスク計量方法 |
JP2006011492A (ja) * | 2004-06-21 | 2006-01-12 | Fuji Xerox Co Ltd | 分布適合度検定装置、消耗品補給タイミング判定装置、画像形成装置、分布適合度検定方法及びプログラム |
JP2006309571A (ja) * | 2005-04-28 | 2006-11-09 | Sumitomo Mitsui Banking Corp | コンピュータ演算処理方法および残存リスク判定装置 |
JP2008542860A (ja) * | 2005-05-27 | 2008-11-27 | カム ルン レオン | リスクの評価及び提示のためのシステム及び方法 |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004127265A (ja) * | 2002-08-06 | 2004-04-22 | Mizuho Trust & Banking Co Ltd | 証券代行業務における要員管理システム |
JP3799326B2 (ja) * | 2002-12-02 | 2006-07-19 | Necインフロンティア株式会社 | パケット送信方式及びパケット受信方式 |
US8649277B2 (en) * | 2006-06-29 | 2014-02-11 | Nec Corporation | Communication apparatus and method |
US7627511B2 (en) * | 2007-06-28 | 2009-12-01 | Mizuho-Dl Financial Technology Co., Ltd. | Method and apparatus for calculating credit risk of portfolio |
US9400960B2 (en) * | 2009-01-22 | 2016-07-26 | Scott Clements | Methods for verifying satisfaction of prognostic algorithm requirements for a component having multiple failure modes |
US8892409B2 (en) * | 2009-02-11 | 2014-11-18 | Johnathan Mun | Project economics analysis tool |
US20120150570A1 (en) * | 2009-08-20 | 2012-06-14 | Ali Samad-Khan | Risk assessment/measurement system and risk-based decision analysis tool |
-
2011
- 2011-03-29 JP JP2011072747A patent/JP5697146B2/ja active Active
-
2012
- 2012-03-23 WO PCT/JP2012/002006 patent/WO2012132355A1/ja active Application Filing
- 2012-03-23 EP EP12764199.1A patent/EP2693379A4/en not_active Withdrawn
- 2012-03-23 US US13/977,999 patent/US20130311231A1/en not_active Abandoned
- 2012-03-23 KR KR1020137018339A patent/KR101368103B1/ko active IP Right Grant
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003036346A (ja) * | 2001-07-23 | 2003-02-07 | Mitsubishi Trust & Banking Corp | オペレーショナル・リスク評価方法及びそのシステム |
JP2004252893A (ja) * | 2003-02-21 | 2004-09-09 | Fujitsu Ltd | オペレーショナルリスク計量装置およびオペレーショナルリスク計量方法 |
JP4241083B2 (ja) | 2003-02-21 | 2009-03-18 | 富士通株式会社 | オペレーショナルリスク計量プログラム、オペレーショナルリスク計量方法およびオペレーショナルリスク計量装置 |
JP2006011492A (ja) * | 2004-06-21 | 2006-01-12 | Fuji Xerox Co Ltd | 分布適合度検定装置、消耗品補給タイミング判定装置、画像形成装置、分布適合度検定方法及びプログラム |
JP2006309571A (ja) * | 2005-04-28 | 2006-11-09 | Sumitomo Mitsui Banking Corp | コンピュータ演算処理方法および残存リスク判定装置 |
JP2008542860A (ja) * | 2005-05-27 | 2008-11-27 | カム ルン レオン | リスクの評価及び提示のためのシステム及び方法 |
Non-Patent Citations (3)
Title |
---|
KOBAYASHI; SHIMIZU; NISHIGUCHI; MORINAGA: "Operational Risk Management", 24 April 2009, KINZAI INSTITUTE FOR FINANCIAL AFFAIRS, INC., pages: 107 - 144,181- |
See also references of EP2693379A4 * |
TAKAAKI KOBAYASHI, OPERATIONAL RISK KANRI KODOKA ENO CHOSEN, 24 April 2009 (2009-04-24), pages 164 - 195, 214 TO 220, XP008171631 * |
Also Published As
Publication number | Publication date |
---|---|
JP2012208644A (ja) | 2012-10-25 |
US20130311231A1 (en) | 2013-11-21 |
JP5697146B2 (ja) | 2015-04-08 |
KR101368103B1 (ko) | 2014-02-27 |
EP2693379A4 (en) | 2014-09-03 |
KR20130086657A (ko) | 2013-08-02 |
EP2693379A1 (en) | 2014-02-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5697146B2 (ja) | リスク管理装置 | |
US7925638B2 (en) | Quality management in a data-processing environment | |
US7720822B1 (en) | Quality management in a data-processing environment | |
US20080262981A1 (en) | Method, System and Computer Program for Operational-Risk Modeling | |
WO2021032056A1 (zh) | 一种处理批量任务的方法、装置、计算设备及存储介质 | |
US20150088595A1 (en) | Systems and Methods for Evaluating Risks Associated with a Contractual Service Agreement | |
CN111612415B (zh) | 一种项目管理方法及装置 | |
KR101471797B1 (ko) | 리스크 관리 장치 | |
US20180330268A1 (en) | Method for adaptive tuning via automated simulation and optimization | |
US12039480B2 (en) | System, method, and apparatus for measuring, modeling, reducing, and addressing cyber risk | |
JP5348351B2 (ja) | リスクプロファイル生成装置 | |
CN113157583B (zh) | 一种测试方法、装置及设备 | |
Rotella et al. | Implementing quality metrics and goals at the corporate level | |
EP2693378A1 (en) | Risk-management device | |
JP5804492B2 (ja) | リスク管理装置 | |
Kojima et al. | Risk analysis of software process measurements | |
CN114510431B (zh) | 一种工作量感知智能合约缺陷预测方法、系统及设备 | |
Matsumura et al. | Analyzing factors of defect correction effort in a multi-vendor information system development | |
CN116777220B (zh) | 一种企业风控管理方法及系统 | |
Saif et al. | Robust design of service systems with immobile servers under demand uncertainty | |
JP2017157074A (ja) | 包括契約システム及びプログラム | |
JP5692800B2 (ja) | リスクプロファイル生成装置 | |
CN117828300A (zh) | 一种基于异常指标时序关系的银行业务根因指标分析方法、系统、设备及可读存储介质 | |
CN118014367A (zh) | 评估商户经营风险的方法和装置 | |
CN115905318A (zh) | 一种数据的处理方法、装置、电子设备及存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12764199 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012764199 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 20137018339 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13977999 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |