Market discipline should be used to lower the risk of arbitrary supervisory discretion in the implementation of pillar 2 of the Basel II Accord. Professor Harald Benink addresses the latest global developments on Basel from a European perspective.
The Basel II Accord of June 2004 seems to mirror a subtle but important emphasis with respect to the potential use of pillar 2 (supervisory review or qualitative supervision) as a counterbalancing mechanism to the quantitative assessment of banks’ internal ratings-based (IRB) systems (pillar 1 of the accord).
In particular, the accord seems to express (much more so than the 1999, 2001 and 2003 consultative papers) supervisory concern about the reliability of the internal ratings systems that banks design to determine the regulatory capital requirement. For instance, in point 9 of the accord the Basel Committee explicitly states that “national authorities may use a supplementary capital measure as a way to address, for example, the potential uncertainties in the accuracy of the measure of risk exposures inherent in any capital rule or to constrain the extent to which an organisation may fund itself with debt”.
In point 10, the committee says: “Even in the case of the internal ratings-based approach, the risk of major loss events may be higher than allowed for in this framework.” These major loss events – which usually have a low probability but are hard to capture in terms of probability distributions and IRB systems – are key in terms of stability of the banking and financial system. Apparently, the Basel Committee is concerned that the IRB approaches are not able to capture low-risk, high-severity events and, therefore, it raises the question of requiring additional capital above the minimum based on IRB models.
Impact study
In point 14 of the accord, the Basel Committee reiterates that it wishes to keep the aggregate level of capital in the banking system stable. In the US, the preliminary results of Quantitative Impact Study 4 (QIS-4), which was undertaken in late 2004 and early 2005, revealed that regulatory capital to be held by the large US banks participating in Basel II could fall significantly.
It also showed a considerable dispersion of outcomes, even for banks with asset portfolios of a similar profile. According to various sources, there would be an aggregate fall of about 20% in the regulatory capital of participating banks, compared with the Basel I regime. Moreover, the range of outcomes extends from a 50% capital reduction to a 60% capital increase.
However, in Germany, one of the few EU countries that undertook a QIS-4 exercise, the preliminary results reveal nothing comparable to those obtained for US banks. These sharp differences between the US and Germany add importance to the QIS-5 studies to be conducted in most Basel Committee member countries late this year.
On April 29, the four federal US supervisory agencies – the Federal Reserve Board, the Office of the Comptroller of the Currency, the Federal Deposit Insurance Corporation, and the Office of Thrift Supervision – issued a joint statement postponing the notice of proposed rule-making with respect to Basel II. The agencies said that additional work was necessary to determine whether the QIS-4 results reflected differences in risks between banks, revealed limitations of QIS-4, identified variations in the stages of bank implementation efforts (particularly related to data availability), and/or suggested the need for adjustment to the Basel II framework. For the supervisors, such an analysis is essential because it determines whether they can have sufficient confidence in the reliability of the banks’ IRB models and the extent to which they are likely to require supplementary (bank-by-bank) capital on a discretionary basis.
Implementation postponed
On September 30, the four federal US supervisory agencies announced that the Basel II rules would come into effect in 2009 instead of 2008, which is a year later than the target date set by the Basel Committee for banks using the advanced IRB approach. The motivation for this delay is to allow the agencies to introduce additional prudential safeguards to address concerns identified in the analysis of the QIS-4 results.
What seems to be clear, and a matter of concern, is that the US supervisors have, until now, not been able to reach final conclusions on the main reasons that are behind the QIS-4 outcomes. Although the supervisors are not yet in agreement, there is a lot of discussion in regulatory and banking circles.
One main reason driving the spectacular QIS-4 results seems to be that US banks have been using significantly different methodologies in assessing a borrower’s probability of default. The issue is that these assessments can be based on current economic conditions (where the rating will be conditioned on the current point in the economic cycle) or can take into account the effect on the borrower of a possible adverse change in the economic climate (‘through-the-cycle’ estimates).
Capital requirements
In a working paper published by the Bank of England in 2003, Eva Catarineu-Rabell, Patricia Jackson and Dimitrios Tsomocos demonstrated that under an approach based on current economic conditions, regulatory capital requirements could increase by as much as 50% for high-quality banks in a recession, while an approach using through-the-cycle estimates would lead to little, if any, increase in capital requirements. Unless US supervisors are able to create more consistency across banks in the choice of estimation methodology, the huge diversity of outcomes will persist. It will be interesting to see how prescriptive the US supervisors are willing to become on this matter.
Another issue related to the implementation of Basel II in the US is that the regulations will only be applicable to about 20 of the largest banks and can only be implemented in its most sophisticated form using advanced IRB models. All other US banks, more than 8000 of them, will have to stay under the Basel I regime. This has caused a lot of concern among these banks: they argue that they would face competitive disadvantages in certain loan markets (such as residential mortgages), whereas the banks operating under Basel II would be able to lower substantially the regulatory capital requirement and, hence, the interest rate that they would charge on these loans.
To address these concerns, the four federal US supervisory agencies announced on October 6 that they would introduce a modified version of Basel I (referred to as “Basel 1-A”) for these banks. In particular, Basel 1-A will increase the number of risk weights, permit greater use of external credit risk ratings, and will expand the types of guarantees and collateral that may be recognised. However, because the risk weights of Basel 1-A will remain less refined than under Basel II, it seems unlikely that Basel 1-A will be able to exclude entirely competitive distortions. This can be expected to be a matter of concern for the Basel 1-A banks.
The EU is heading for a full implementation of Basel II as of 2007/2008. On September 28, the European Parliament approved the proposal for a Capital Requirements Directive (CRD), which is the translation of Basel II into EU law. Contrary to the situation in the US, the CRD will be applicable to all banks in Europe, giving them a choice between either of the three risk measurement approaches identified in the Basel II Accord (standardised, foundation IRB and advanced IRB approaches). The fact that European banks will be allowed to implement the advanced IRB approach one year earlier than US banks is raising interesting competitive issues.
Market discipline
Although the Basel II Accord provides sound incentives for banks to professionalise their risk measurement and risk management, it creates all types of problems with respect to risk measurement from the supervisory point of view. These may have to be addressed by the application of a potentially substantial degree of supervisory discretion based on pillar 2 of the accord. Market discipline (pillar 3 of the accord) could play a counterbalancing role because the risk assessment by professional investors on financial markets (such as pension funds and insurance companies) would prevent banks from lowering their capital too much and, hence, would reduce the need for supervisory discretion.
Pillar 3 of the Basel II Accord contains many information disclosure requirements. At the same time, however, it fails to create incentives for professional investors to use this information in an optimal way. The reason for this is, as long as professional investors holding bank liabilities have the perception that large banks are too big to fail, they will have the idea that their money is not really at stake, mitigating their incentives to use the disclosed information. A mandatory requirement for large banks to issue credibly uninsured subordinated debt as part of the regulatory capital requirement could improve the strength of pillar 3.
International benefits
The IRB approach should be complemented by a more credible and effective form of market discipline. To the international banking community, the big advantage of this new approach would be that there will be less risk of arbitrary supervisory discretion, which is evident given the current discussions on QIS-4/5, because the supervisory assessment of the reliability of a bank’s internal ratings system will be assisted and objectified by the assessment of professional investors on financial markets.
Harald Benink is professor of finance at RSM Erasmus University, Rotterdam, the Netherlands, and a senior research associate to the Financial Markets Group of the London School of Economics. He is also chairman of the European Shadow Financial Regulatory Committee.
