This is the second week of DW's Learning Curve coverage of two consultative papers issued by The Basel Committee on Banking Supervision. After looking at market discipline last week, we will focus this week on issues surrounding internal ratings systems.
The Basel Committee on Banking Supervision has produced two further consultative papers on individual aspects of the new capital adequacy framework. They build on the proposals contained in the original proposals for updating the Capital Accord ('the original paper'). The first of the papers, which was examined last week, is designed to support and strengthen the original proposals for the nature and extent of disclosures that banks should make in order to promote the role of market discipline in promoting bank capital adequacy. The second paper, to be discussed in this article, surveys the current applications of, and methodologies behind, internal ratings systems used by banks. The original proposal emphasized the Committee's commitment to basing new approaches to regulatory capital standards on internal ratings--an emphasis that has attracted considerable support in subsequent coverage of the regulatory debate.
INTERNAL RATINGS SYSTEMS
The second supplementary document sets out in more detail the Committee's approach to the increased use of internal ratings systems. In early 1999, a body known as the Models Task Force was asked by the Committee to produce a study of internal ratings systems with a view to evaluating the options for linking such systems to capital regulation. The Framework Paper explained that internally generated ratings should generally be preferable to externally generated on the basis that a bank's understanding and experience of its counterparties' businesses should lead to a more sophisticated analysis of creditworthiness. The paper contrasted this with the assessment criteria applied by the rating agencies which, in order to be of value for comparative purposes, need to be applied completely consistently. The Task Force spent much of last year conducting a survey of 30 financial institutions, chosen as having particularly well-developed modelling systems. Other banks and regulators also made presentations to the Task Force, which will continue to consider the issues surrounding the introduction of ratings into capital regulation.
This paper is therefore interim in nature--its purpose is not to provide detailed coverage and analysis of individual proposals, nor to judge the merits of individual approaches. Rather, it should be approached as a survey of current practice in the context of which policy issues relating to the 'architecture' of capital regulation using internally generated ratings can start to be publicly addressed. The key function of the paper is to generate feedback on the extent to which the range of practices identified does represent 'best' or 'sound' practice; whether the Task Force has failed to identify key aspects of banks' rating processes; and, generally, whether its conclusions are fair and reasonable.
Having set out its findings on current industry practice, the paper then identifies key issues that need to be effectively addressed as part of the process by which internal ratings-based regulation could be introduced. These are briefly set out below.
RANGE OF INTERNAL RATINGS PRACTICE
In its examination of the structures of ratings systems, the Task Force looked a number of issues, starting with numbers of grades for performing and non-performing/impaired assets. The conclusion was that banks have widely differing ideas as to what constitutes potential weaknesses, and that direct comparisons would be difficult. In terms of the risk evaluation process, the survey identified three main categories depending on the extent to which reliance is placed on quantitative techniques on the one hand and on the personal experience and expertise of individuals on the other. Despite the underlying methodological differences, however, the Task Force concludes that the distinctions between these categories may be less precise in practice. In terms of risk factors considered in assigning grades, the Task Force found that the traditional considerations in assessing borrowers--such as balance sheet, income statements and cash flow performance--were in use very widely.
The Task Force identifies various attributes of 'judgmentally oriented' and 'statistically oriented' banks. Almost all of the former make use of historical and trend data to some extent, as well as formal industry and peer group analysis. The latter tend to rely heavily on specific types of financial data, for example, on leverage or debt service coverage. In terms of considerations in assessing facilities, all of the banks surveyed explicitly considered such facility specific considerations as third-party guarantees, collateral and the seniority or otherwise of the facility. Almost all banks surveyed use statistical modelling to some extent, but the Task Force suggests that models rely on much the same inputs as the more 'judgmentally oriented' institutions.
LOSS CHARACTERISTICS
Once a bank has assigned a rating to an individual exposure, the next step is to assess the 'loss characteristics' that may be expected from that category. This information has a number of applications in the contexts of portfolio management and credit risk modelling, among other things, but the most obvious is to evaluate the accuracy and consistency of existing rating criteria. Since every bank will strive to improve the performance of its rating system, risk managers will benefit from an understanding of regulatory perceptions of loss characteristics.
The first point to note is that most banks combine a general sense of a counterparty's creditworthiness with one of the main 'quantifiable loss concepts.' These are probability of default; loss given default; and exposure after default. Detailed consideration of patterns in the methodology of these three is outside the scope of this article. However, it is worth noting, in brief, that for the first, default probability, the key questions concern the use of external data. The second, which at present is used only by a handful of banks, tends to rely on very similar inputs (such as borrower and facility-specific issues) to the others. Similar considerations applied to, and a similar conclusion was drawn for, the third form of loss characteristic, exposure at default.
While this will be of most immediate relevance to risk management professionals within the top tier of internationally active financial institutions, the principles that the Task Force draws from its survey are nonetheless an important milestone in the development of best or sound practice in the use of internal ratings. Particularly for financial conglomerates, where investment banking divisions will typically have better developed modelling and rating practices in place, the speed at which these practices can be rolled out into other business areas may have critical implications for the group's competitive position. This article therefore closes with a summary of the Task Force's elements of a possible architecture for the Internal Ratings Based approach:
This week's Learning Curve was written by Simon Gleeson, attorney at Allen & Overy'sfinancial services group in London.
