Credit Registries, Efforts to Strengthen Credit Risk Measurement, and New Basel Capital Accord (Basel II)
Credit registries possess enormous potential as a key tool in the hands of supervisory authorities that would enable those authorities to face the challenges of implementation of Basel II.13 Moreover, effective use of the information contained in credit registries, whether public or private, will enable credit institutions to improve the identification and control of their banking risks, thereby helping to pave the way for more advanced risk and capital measurement approaches envisaged in Basel II.
As already explained, credit registries facilitate the sharing of information among lenders and with supervisors—subject to adequate safeguards—on credit history loan characteristics and specified characteristics of borrowers (households and firms separately), which enables each bank to assess the quality of its credit assets and enables the supervisors to monitor credit risk in the entire system. The access to credit information by banks helps to impose discipline on borrowers and fosters greater transparency, as well as more competition.
Information from credit registries can be used to support both onsite and offsite supervision, as well as to facilitate macroprudential surveillance. Because supervisors have access to the entire population of loans granted by each credit institution, they can use this information to construct a range of financial soundness indicators for individual banks, peer groups, and the system as a whole. The information can be used to select samples for more detailed examination in onsite inspection. Also, comparison of the information reported by different credit institutions can help those conducting offsite surveillance to detect the potential of any one bank’s systematic overvaluation of credit worthiness of its borrowers or a deterioration in the credit quality of a bank’s loan portfolio relative to the rest of the system. The information from credit registries can help when analyzing the dynamics of aggregate credit risk—and bank-specific risks—and its macroeconomic and institutional determinants. Finally, information in credit registries—together with other information outside the registries—can help when estimating (or validating bank estimates for) probability of default of different borrowers, when providing input into estimating loss given default (LGD), and when verifying the bank’s estimate of exposure at default.
Credit registries can be a useful tool to validate the bank’s own internal ratings and internal assumptions about credit risk modeling. The statistical techniques to verify borrower rating systems are well developed, and it is relatively easy to discriminate among the relative positions of obligors. However, the validation of probabilities of default associated with each rating is more difficult because data are scarce, particularly on defaulted obligors and on the correlation among defaults, which is hard to quantify. In this context, a rating system for borrowers—developed by supervisors and based on data on the entire population of all credit institutions—could provide a yardstick with which to compare and validate ratings and probabilities used by individual institutions. This approach would require credit registries to be managed by supervisors and to contain a certain minimum quantity of information so an overall rating system could be developed.
The estimation of LGD is typically based on market prices of defaulted loans and bonds or on a credit institution’s own data on discounted cash flows—revenues and expenses—following default so best estimates of loan losses can be obtained (using both internal and external data). Little progress has been made on the techniques to validate LGD. Information from credit registries can be used to estimate the key determinants of LGD (by means of a regression model), and the possibility of using credit registries to document loan losses offers a realistic option to develop estimation and validation procedures for LGD. Similar observations apply to exposure at Default (EAD). In addition, the transition matrix for the entire credit system, as well as the sectoral and geographic differences in credit quality, can all be monitored using the credit registries. Finally, the broad recognition of credit risk mitigation techniques in Basel II calls for the credit registries to carry precise information on loan characteristics so they can be used to estimate the value of guarantees, collateral, and other risk mitigants accurately.
If they are to harness the potential of credit registries, their information structure should have adequate information to estimate the value of Probability of Default (PD), EAD, LGD, maturity, risk mitigation factors, and loan loss provisions so various parameters of credit risk models can be estimated by banks and validated by supervisors. For this purpose, required minimum information that should be included in the data structure of
credit registries should be evaluated so credit registries can contribute to effective implementation of Basel II.