Problems and countermeasures of implementing internal rating law in BOJ industry

Challenges and tackling strategies of Internal Rating System Implementation, a Japanese perspective wen/Wang Shengbon one, retail loan risk parameter estimation　　The main problems include: (1) Lack of the number of asset pool types of industry standards, (2) the lack of a standard definition of default, (3) insufficient data accumulation. Number of types of asset pools The internal rating method requires commercial banks to carry out the management of retail loans, and to estimate the credit risk parameters, such as default probability, default loss rate and default risk exposure, on the basis of risk-pooling. The number of retail loan asset pools in Japan has not yet formed an industry consensus.　　Therefore, even for similar retail loans or similar credit risk models, the estimated risk parameters and the measured capital requirements may vary.　　In order to ensure the reliability of the results of capital measurement, commercial banks should establish a reasonable pool type of retail credit portfolio according to the attributes of different products, such as housing mortgage loan, credit card loan, and considering the accuracy and sample size of the estimated risk factors. The definition of default is the precondition of estimating credit risk parameters and measuring capital requirements. The new capital Accord provides a reference definition of default based on economic implication, allowing the regulatory authorities and commercial banks to adjust the definition of default to some extent according to practice.　　Japan's major banks have a different number of days to default on retail loans, and there are also differences in conditions for the return of non-default status from default, which will result in differences in the risk parameters and capital requirements of similar retail loans.　　Japan's regulatory authorities believe that a default definition should be at least consistent with a group of retail credit products with similar risk characteristics, and that commercial banks should be able to explain the rationale for defining choices if different definitions of defaults are used in a retail credit portfolio. If data accumulation is insufficient, it will be unable to support the division of asset Pool and the determination of risk factors, so it is difficult to make an objective judgment on risk. As the data accumulates, banks will find that the asset pool partitioning method based on limited data is no longer appropriate.　　Lack of time-series data makes it difficult to understand the sensitivity and maturity effects of retail credit portfolios on economic conditions (seasoning effect). The commercial bank thinks that if we find that the risk factors based on the current asset pool division are unstable, we should check the method of asset pool division; If the dividing method is difficult to change, considering the instability of risk factor valuation, the risk factor should be estimated conservatively. When estimating risk factors, it is very important to identify the dynamic characteristics of risk factors through longer time series data.　　If a long time series data is not available, the commercial banks should consider the influence of economic fluctuation and the maturity effect when using model and scenario analysis. Second, the internal rating system to use commercial banks to deal with the internal rating system and risk parameters of the valuation of full confidence. Therefore, the risk parameters of default probability and default loss rate areand the internal rating results in the business decision of the practical application is very important, the new capital Accord on the internal rating system testing has a clear stipulation. However, the estimated value of risk parameters for the measurement of regulatory capital and business decision can be different. If the two are inconsistent, it is important for commercial banks to prove to third parties the quality of risk parameters used to measure regulatory capital, the relationship between these risk parameters and the risk factors used for business decisions.　　It is also critical to determine which departments in the commercial banks use risk parameter estimates and which departments are responsible for checking the use of risk parameter estimates. First, commercial banks must be able to explain how the internal rating system is used for business decisions and strategies. In order to guarantee the reliability of risk parameters such as default probability and default loss rate, the key is to determine the extent of use in the following areas: (1) credit approval and set the limit, (2) Determine the loan interest rate or loan rate of the determination of the Rules, (3) credit portfolio management; (4) Economic capital allocation; (5) The ability to demonstrate the difference between risk parameters for internal risk management and regulatory capital objectives.　　In addition, given the direct impact of default probability, default loss rate and other risk parameters on the calculation of regulatory capital requirements, it is particularly important to use these risk factors in ways that third parties consider to be able to guarantee the quality of these parameters. The assessment of these key elements should be based on two criteria: first, the degree of importance, can be determined by the impact of risk parameters or internal rating systems on business decisions; second, consistency, sustainability, through risk parameters or internal rating system in the day-to-day risk management of the use of　　and determine whether consistency is maintained throughout the organization. Secondly, if the risk parameters of internal risk management and capital regulation are different, commercial banks should be able to explain clearly to the third party the relationship between the two sets of risk parameters and the reasons for the difference. For example, if: (1) two sets of risk parameter estimates use the same data source, (2) in the long run, the two sets of risk parameter deviations are not large (from the point of view of the return test, there is no deviation).　　This indicates consistency between the two sets of risk parameters. Third, commercial banks have the responsibility to verify that the results of regulatory capital measurement are used for business and internal risk management.　　Within commercial banks, while the front line and risk control departments are responsible for the use and validation, senior management and internal audit departments should establish the necessary processes for validation and use approval. Third, the credit risk management organization system internal rating system effective implementation must be based on strong corporate governance mechanism. The corresponding corporate governance mechanism includes: (1) The formation of risk management rules; (2) The document of the risk management process; (3) The senior management participates in the credit risk management through the internal rating system, (4) establishes the risk Control department which is independent of the foreground line of business, (5) The external third party appropriately intervenes, provides the support for the internal audit.　　The present report focuses on the latter two issues. The main issues of the role and independence of the risk Control department include: (1) How to operate at the front deskEstablishment of an independent department responsible for data collection and risk management (called the Middle and middle section of the front office) or a credit review department independent of other operations in the front office? (2) How to establish an appropriate division of business between the middle-and middle-office departments of the front office and the Independent Risk control departments (real middle and middle departments)? (3) The role of risk control departments in the review of individual loan interest rates?　　(4) Is there a risk of an incentive mechanism that could undermine the independence of the sector and the credit review sector? What are the minimum requirements to avoid this risk?　　Japan's banking industry believes that the establishment of independent risk-control departments and credit-review departments is essential to building effective and accountable risk management systems that can challenge the front-office business. In order to ensure the independence of the Credit Review department, it is very important to ensure the independence of the Credit Review department in the command chain of the front office business. From the perspective of scale and degree of risk, the independence of the Middle-Office department in the front office is different depending on the importance of the relevant business. In addition, if the level of risk, for less important lines of business, the middle and middle departments are mainly responsible for the inspection of the middle and middle departments established in the Front Office Department to implement the risk management process; With increasing importance, the Office will also be responsible for other risk control functions consistent with the importance, possibly including related risk data collection,　　Monitor risk levels and monitor risk-related compliance. The front office and credit review departments mainly refer to the earnings of other products, business growth prospects and corporate credit ratings for the single loan pricing. In this case, when the whole credit portfolio is priced, the Office should check whether the credit risk/cost is taken into account. In assessing the performance of the middle-and middle-office sector, short-term profits should not be the minimum requirement to avoid the risks posed by incentives that could harm the independence of the sector.　　In addition, the middle and middle sector performance evaluation personnel should be independent of the front Office business unit. The main issues of participation of internal audit and external audit include: (1) To what extent can internal audit departments conduct detailed audits of internal rating and risk quantification models? Is it feasible?　　(2) If the internal audit department lacks sufficient quantitative analysis personnel, how to establish the necessary risk management system? The industry believes that the verification of internal ratings and credit risk quantification models should be carried out continuously by the middle-and-middle sector. Obviously, the internal audit department does not need to carry out the same level of detailed certification with the middle-office, or the implementation of the process of verification activities. However, the internal audit department should review the front office and other departments to ensure the effective operation of the internal rating system and prevent conflicts of interest. The internal audit may have to cooperate with external agencies in the implementation of audit activities. In this case, the internal audit department should also bear the final responsibility for the internal evaluation system Operation inspection.　　If the same person in the same department is responsible for risk factor estimation, as well as model and risk factor verification, internal audit should have beyond the normal supervision authority and have the corresponding expertise. If you are familiar with the new goldThe limited number of professionals involved in the product, or the inability of the audit Department to retain competent personnel, commercial banks should consider other methods of overcoming internal auditor shortages, such as strengthening the verification of the internal rating process, rather than focusing too much on quantitative inspection, cooperating with external audit bodies or advisory bodies. The estimate of default loss rate is different because of the definition of default and loss. For example, whether to use the definition of default defined by the regulatory department or the definition of accounting breach of contract to provide special preparation. How to clarify and deal with the conceptual differences between the two definitions, and focus on the impact of differences on accounting and information disclosure is critical. For the purpose of service risk management, it is estimated that the loss of default loan should be based on the actual recovery amount of the final loan, and since the disposal of any single default loan is unique, it is inappropriate to adopt a unified estimation method in some cases. The alternative approach is to use specific preparations based on accounting objectives to replace loan default losses that serve the risk management objectives.　　In this case, reasonable validation should be based on the data to see if the assumption leads to an estimated deviation in the rate of default loss. In order to explore the best method of estimating the loan default loss rate, it is necessary for commercial banks and related departments to collect the loan default data and check the empirical results of the loss estimate carefully. When estimating the loss rate of default loans, consideration should be given to how to estimate the impact of the economic downturn.　　It is very difficult to use actual data and empirical analysis to solve these problems because of the limited sample of default loan loss data and the effective method of estimating default loss rate of default loan. In contrast to the estimated default loss rate of non-default loans, the default loss rate of default loans should be considered as follows: (1) The default loss rate of default loans is based on the assumption that the default state of loans is established.　　For non-default loans, the state of default is random, and the estimation of default loss rate has nothing to do with the state information of default.　　(2) As a partial default loan can be recovered over time, the estimate of the default loss rate should be adjusted on the basis of the residual debt (defined as ead[default risk exposure] minus the recovered amount), whereas for non-default loans, the rate of default loss is estimated to be based on exposure to default risk. The above analysis shows that cash flow discount method (CCF) can be used to estimate the default loss rate of default loan. This method can be used as an alternative to estimating the rate of default loss if the cash flow discount method is applied to the accounting preparation.　　However, there is still the question of how to consider the impact of the economic downturn. The estimation of default probability and loss rate of professional loan due to the lack of sufficient default data in professional loan (special lending), it is difficult to use actual data to validate risk parameter estimates.　　Theoretically speaking, the framework of the rating system/model can be divided into the following three kinds of methods. (1) using the same rating system as the company to estimate the default probability internal credit rating is based solely on the probability of default and is not related to the credit type and the mode of transaction. The same default probability is used for corporate and professional loans at the same risk level.　　(2) The use of external data related to professional loans is estimated to map the professional loan rating to an external rating, using the default probability issued by a credit rating agency. (3) using the model to estimate the default probability and other risk parameters of a single professional loan in a risk level by using the model directly. For example, based on Merton's model or scoring model, a debtor or loan is deemed to be in default if its value is below a threshold.　　If the rating is based on the result of the model (equal to the assumed default probability), the rating process is the process shown in the right figure.　　For a single professional loan, the default probability is calculated by model. The three methods mentioned above can be used in the estimation/verification process.　　For example, the probability of default is estimated by a third method, the first method is mapped, or the probability of default is estimated by the first method, which is confirmed by a third method. The industry believes that the three types of methods used in the process, should pay attention to the following issues: The first method: if the professional loan rating system and the company risk exposure rating system, should be based on empirical data validation, Examine whether the actual default probability of professional loans and corporate risk exposures at the same level is different from the trend of transfer probabilities.　　Due to the small number of professional loan defaults and the imposition of similar validation restrictions, it is still important to ensure that corporate risk exposures and professional credit rating systems are of the same nature, such as comparing the rate of migration between the two ratings. The second method: if the external rating of the asset securitization transaction is matched with the professional loan rating, and the asset securitization transaction is not separated from the professional loan, the various problems in the mapping process should be reasonably explained.　　For example, the use of commercial housing mortgage-backed securities external ratings to estimate banks without recourse to real estate loans. The third method: if the establishment of independent professional loan rating system, not with the corporate credit rating mapping, it is best to use the actual default data to verify.　　With the scarcity of professional loan data, it is difficult to establish a rating system that covers only professional loans, and the use of a common database (data Federation) is a future solution to the problem. (The author is the China Banking Regulatory Commission International Department)