Financial Risks, Stability, and Globalization


Omotunde Johnson
Published Date:
April 2002
  • ShareShare
Show Summary Details

This commentary looks at the subject of internal ratings and internal rating systems in three ways:

  • as a tool for banks in managing their credit risks;
  • as a tool for banks in setting, or helping to set, economic capital; and
  • as a tool for supervisors in setting regulatory capital.

Supervisors and all those interested in financial stability—and, of course, the banks themselves—have a keen interest in each of these three aspects of the subject. In particular,

  • Supervisors increasingly emphasize credit risk management; it is one of the key Basel Core Principles that IMF staff are monitoring in the context of assessments of countries’ observance of standards, in the course of carrying out work on financial risks and vulnerabilities, especially in the Financial Sector Assessment Programs being undertaken jointly with the World Bank.
  • As banks are developing their own estimates of economic capital, supervisors are increasingly interested in how this is done and the robustness of the methods. For many banks, for much of the time, actual capital exceeds minimum regulatory capital by a wide margin. Moreover, the Basel Committee’s new capital framework, through Pillar 2, supervisory review, will encourage supervisors to set minimum capital standards for an individual bank on the basis of the risks it runs—and the quality of a bank’s credit risk controls and environment is highly relevant.
  • One of the most important proposals in the Basel framework is the desire to use internal ratings for setting minimum capital requirements in Pillar 1.

Internal Ratings as a Tool for Banks in Managing Credit Risk

The paper by Stephen and Fischer is primarily concerned with this aspect of internal ratings. Their paper is a reminder—and the message is the same as that coming from surveys by the Federal Reserve Board and by the Basel Committee (Treacy and Carey, 2000;Basel Committee, 2000a)—that banks and especially the larger and more sophisticated banks have increasingly developed their own quantitative systems for managing credit risks; that these systems vary a good deal between banks; and that these systems are now being used by banks for making or helping to make major decisions on pricing loans, credit limits, provisioning, employee compensation, and more.

Stephen and Fischer’s paper emphasizes that credit risk needs to be seen as a stochastic process in which banks try not only to quantify likely outcomes but also take account of uncertainty by looking at, for instance, the potential for migration from one rating to another. The concluding remarks in this commentary focus on the regime shift implied by the new capital framework.

Stephen and Fischer emphasize the close links between external and internal ratings. There has been much criticism of the proposed use of external ratings in the new capital framework. Stephen and Fischer rightly say that the criteria for accreditation of external rating agencies need to be strengthened, a view shared by both the IMF and the World Bank staffs. An IMF Working Paper (Karacadag and Taylor, 2000) suggests more rigorous criteria based on the U.S. Securities and Exchange Commission (SEC) experience in designating “Nationally Recognized Statistical Rating Organizations.” One key element in these criteria is the use of market feedback: the rating agencies must be regarded as credible and reliable by major market makers. In addition, the SEC examines the agencies’ independence, resources, structure, procedures, etc. A proven track record extending over several years is also needed.

Stephen and Fischer’s paper brings out very clearly how, at least for CSFB, the use of internal ratings systems is benchmarked on external ratings wherever these are available. This is not surprising: risk management officers, when seeking to check internal estimates of riskiness in loans, will be bound to look closely at, and indeed be heavily influenced by, any comparable external ratings—easier in the United States than elsewhere.

This benchmarking means that there is likely to be a close connection between the strengths and weaknesses of external and internal ratings; they cannot be treated as two separate and entirely independent approaches to setting capital. The proposed use of external ratings has been criticized, including by IMF staff, because it may increase the procyclicality in the system. Suppose that a country or a corporate sector enters a cyclical downturn, followed or accompanied by a fall in the external rating of the country or corporate; this, in turn, leads to higher capital requirements for banks lending to this country or corporate, which causes the banks to cut back further on, or raise the cost of, credit. The result could be an internal credit crunch, a cut in capital inflows, or both.

Now ratings agencies claim to “rate through the cycle” (Treacy and Carey, 2000), so that a normal business cycle—to the extent that there is such a thing—should not of itself lead to shifts in ratings. Internal ratings, on the other hand, make no such claim: indeed the banks that do these ratings say they base ratings on the borrower’s “current condition,” although the evidence collected by the U.S. Federal Reserve suggests that banks vary a good deal in the time period they have in mind. Clearly, this is an area where more work is needed, but it may well be that internal ratings may be more cyclical than external ratings.

That is not the end of the story. It remains unclear how an explicit internal rating system compares, in terms of cyclicality, with the informal credit appraisal techniques still used in some banks and generally in use before internal rating systems became common in larger banks. This is a relevant comparison for those wishing to compare the proposed new system with the 1988 Accord.

Stephen and Fischer make the point that a complete perspective on credit risk requires estimates of both default probabilities and of losses in the event of default. They rightly go on to say that they expect regulators to want both, and to give banks the incentives to evolve much more robust approaches to measuring loss given defaults. I am sure this is very much in the minds of the Basel Committee as they strive to put flesh on the bones of their proposals.

Aspects of the New Capital Framework

I will now spell out in more detail the ways in which internal ratings are relevant to the new capital framework.

The first way is in Pillar 1, for setting minimum capital. Many supervisors believe that making more use of banks’ internal information systems, including internal ratings, is a valuable addition to the supervisors’ tool-kit, by tailoring capital more closely to the actual risks run by banks.2 Moreover, there is an increasing need for supervisors to develop the skills and resources to assess the quality of banks’ systems. IMF staff has indicated to the Basel Committee (IMF, 2000) that for internal ratings to be used successfully:

  • guidelines need to be developed for supervisory assessment of internal rating systems including through the development of a supporting methodology, which will help to make the resulting capital ratios comparable across countries;
  • more evidence is needed on the accuracy of internal rating systems; and
  • supervisors, too, will need guidance and training on how to assess banks’ risk management capabilities, not least to ensure that banks do not deliberately underestimate risks.

So internal rating systems are relevant to the new capital framework through the enhanced interest of supervisors in how the banks allocate their capital internally; this matters for the incentives that banks have to improve their decision making. But how do banks arrive at their estimates of the overall quantity of capital that banks should hold?

This is a much more difficult area for both banks and supervisors. Both are concerned not only about creating the right incentives for micro-decision making in banks but also about the total amount of capital in the system. Now, this is far from being a simple matter of what the regulators decree. For many banks, for most of the cycle, capital is well above the minimum: the Financial Services Authority (FSA) in London (Richardson and Stephenson, 2000) has published an interesting study showing that about half the banks in the United Kingdom hold capital, which is some 50 percent above the minimum—and this is in a regime where all banks are required by the FSA to hold capital to varying degrees above the 8 percent minimum. No doubt, some of this reflects the benign state of the banking cycle in the United Kingdom; some represents the banks’ desire to keep the supervisors in the FSA off their backs. But, even making allowance for these factors, it is clear that the constraint on banks’ capital is often not the supervisors’ minimum but the effect of market discipline.

This, however, is not to downplay the significance of regulatory capital. In many countries, regulators insist not on 8 percent, but on higher percentages, like 10 or 12 percent. And, of course, for most banks, credit risk is by some way the largest source of overall risk. Basel will not be prepared to accept whatever level of capital emerges from banks’ own internal rating systems. In its June 1999 paper, the Basel Committee was explicit: the overall level of capital in the system at least should be maintained. And the Committee is doing a lot of work on how to translate internal ratings into capital. No doubt, the results will be tested on a sample of internationally active banks. But what if the results show that such banks would hold not 8 percent, but 4 percent capital?

A partial answer may be found in the Committee’s readiness to impose a capital charge on “other risks”; this could compensate for some lower level of capital for credit risk. But in the event that a 4 percent capital charge resulted from the use of internal rating systems, the Committee might well have to consider some kind of multiplier, as it did with some success in the case of the Market Risk Amendment a few years ago, and which stood up well to the stresses of the financial crises in 1997-98.

Another way in which internal ratings are relevant to capital set by regulators is through Pillar 2 of the new capital framework—namely, supervisory review. Through this pillar, the Basel Committee encourages supervisors to set capital for individual banks above the minima set in Pillar 1 and to use as relevant factors the supervisors’ judgment of a bank’s risk management skills—including the quality of its internal ratings processes.

A fourth aspect of the capital framework that is relevant here is Pillar 3, market discipline. Here, Basel is encouraging banks to make more disclosures on credit risk (Basel Committee, 2000a, 2000b). A financial institution should summarize for the markets its policies for identifying, measuring, and managing credit risk on both an individual counterparty and portfolio basis. If applicable, the information should include a description of the internal credit rating classification system.

An important issue is whether, or how far, any of this work on internal ratings is relevant to banks and supervisors in developing countries. There are several reasons why it is likely to become relevant to countries outside the Group of Ten (G-10) and the Organization for Economic Cooperation and Development (OECD) groupings.

The first reason is that it is part of the wider argument about the role of the new Basel framework worldwide. In the view of IMF staff, this framework needs to pay attention, not only to the large internationally active banks that are at the heart of the international financial system, but also to the needs of banks worldwide. The framework can best do this not by diluting any key standards, but by providing for options. That seems to be a widely held view in the Basel Core Principles Liaison Group, which includes supervisors from G-10 and non-G-10 countries, as well as the IMF and the World Bank.3

The second point on the relevance of internal ratings is to watch developments in Europe. For the new framework to be adopted in the European Union, there will need to be new and revised directives, which will apply to all banks. Now supervisors in Europe are keen that capital standards using internal ratings should be available to all their banks that meet the conditions. So Europe is likely to see a simple standard version of the framework that is suitable for a wide variety of banks. Such a framework could be applied outside Europe.

Another development to look out for is the emergence of industry standards—ready-made systems that can be adopted to the needs of smaller, less sophisticated banks. Such standards would allow many banks to implement tested internal rating systems that could be tailored to these banks’ needs.


Much of the initial comment on the Basel capital framework focused on the questions of the appropriate weights on banks’ assets; the appropriate weighting of loans to countries, corporates, and banks; and trying to determine those who will gain and those who will lose. This is only a part—and not the most important part—of the story. The new framework needs to be seen much more broadly, as an incentive to banks to improve their credit risk management methods and to develop their own estimates of economic capital to be held against risks. This will represent a new regime, not just a change in a few parameters. Banks and supervisors will have to be aware, not just of appropriate weights at a point of time, but of the knowledge that, as the world and its risks change, so will the appropriate risk weights and capital. In turn, this will reinforce the pressures on supervisors to develop the skills needed to ensure that banks and their management know what they are doing.


    Basel Committee on Banking Supervision1999A New Capital Adequacy Framework,Basel Committee Publications No. 50 (Basel: Bank for International Settlements).

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision2000aA New Capital Adequacy Framework: Pillar 3—Market Discipline,Basel Committee Publications No. 65 (Basel: Bank for International Settlements).

    • Search Google Scholar
    • Export Citation

    Basel Committee on Banking Supervision2000bBest Practices for Credit Risk Disclosure” (Basel: Bank for International Settlements).

    • Search Google Scholar
    • Export Citation

    GreenspanAlan2000Evolving Challenges for Bankers and Supervisors,speech via videoconference to the National Association of Urban BankersUrban Financial Services CoalitionSan FranciscoMay.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund2000Comments on Proposals of the Basel Committee on Banking Supervision for a New Capital Framework.Available via the Internet:

    • Search Google Scholar
    • Export Citation

    KaracadagCem and MichaelTaylor2000The New Capital Adequacy Framework: Institutional Constraints and Incentive Structures,IMF Working Paper 00/93 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    MeyerL.2000The Challenges of Global Financial Institution Supervision,speech given to the Federal Financial Institutions Examination CouncilInternational Banking ConferenceArlington, VirginiaMay.

    • Search Google Scholar
    • Export Citation

    RichardsonJeremy and MichaelStephenson2000Some Aspects of Regulatory Capital FSA Occasional Paper No. 7 (London: Financial Services Authority).

    • Search Google Scholar
    • Export Citation

    TreacyWilliam and MarkCarey2000Credit Risk Rating Systems at Large U.S. Banks,Journal of Banking and Finance Vol. 24 (January) pp. 167201.

    • Search Google Scholar
    • Export Citation

It is with great regret that we inform the reader that Huw Evans passed away on February 15, 2002, shortly before this book went to press and just after he had cleared the final version of these comments.


For a statement by a governor of the U.S. Federal Reserve that, “capital requirements for individual banks will vary with their individual credit risk profiles,” see Meyer (2000).


For a view that smaller, simpler banks require an altogether different approach, see Greenspan (2000).

    Other Resources Citing This Publication