The publication by the Basel Committee of its discussion paper on the future of the 1988 Capital Accord is designed to lead to a revision of the approach to the supervision of internationally active banks in the course of the year 2000. Since this initiative will establish the framework for the supervision of these firms for the foreseeable future, it is clearly essential that the approach which is finally decided upon should not only be soundly based but also sufficiently robust and adaptable to enable it to accommodate the further significant changes which are likely to affect global financial markets and the firms that operate in them in the future.

In view of these considerations, this paper seeks to identify the key characteristics of an effective international regulatory framework for modern financial markets. In particular, the paper argues that any effective international regulatory system has to be based on the central objective of minimising systemic risk in international markets, be applicable to all systemically significant firms, and operate in a way that is broadly consistent with the way in which firms themselves manage risk and capital.

The central proposition of the paper is that the changes which have occurred over the past ten or fifteen years both in the nature of international financial markets and in the structure and risk management processes within the firms are of such a substantive nature that they justify a fundamental reassessment of the basic premises on which the Basel Committee approach is currently based. In particular we highlight the following issues for consideration:

• Recent developments in risk management practice, and particularly in relation to the management of credit risk have led to a substantial reassessment of the nature of firm's risk exposure and the role which capital should play in protecting against this risk. In particular, firms now increasingly distinguish between the expected loss associated with credit risk which should be factored into the transaction pricing and accounted for as a cost of doing business and the unexpected and stress losses associated with credit risk which are measures of the potential volatility of the losses around the expected value and which have to be covered by the firm's other earnings and capital resources. This is a very different view of capital to that which underpins the Basel Accord and which assumes that capital is required to protect against all credit losses, both expected and unexpected.

• The firms which are involved in the international finance have also changed significantly since the mid 1980's. International finance is no longer dominated by cross border bank lending but includes a much wider range of activities in securities and derivative markets. Moreover, firms have increasingly sought to diversify their sources of revenue in order to provide a more secure earnings stream and to provide protection against any particular risk event. As a result, firms are now managing the link between risk and capital in a way that seeks to limit the total risk to their earnings from all material risk factors, not just the traditional risks of market and credit risk. This process of relating risk to the earnings capability of the firm provides a different perspective on risk, which is not currently captured by the current Basel Committee approach.

• International financial intermediation is increasingly characterised by the use of traded market techniques and instruments. As recent events in financial markets have shown, this means that the primary source of international systemic risk should be seen as the potential for a sudden drying up of liquidity and heightened price volatility in particular markets or a funding crisis at a particular firm, rather than the undercapitalisation of any particular group of banks as was the case in the mid 1980's. This issue has significant implications for both regulatory capital standards and for the supervision of financial firms more generally which is also not addressed in the Basel Committee proposals.

• There is currently no consensus about the circumstances which regulatory capital should seek to protect against and as a result there is no agreed basis for determining whether a particular level of regulatory capital is appropriate or not. We believe that the core purpose of the regulatory capital requirement should be to enable a firm not only to survive an initial stress event but also to remain solvent for long enough after such an event to allow an effective work out strategy to be put into effect. This requires a specific focus on the type of stress events which could give rise to a call on the firm's capital as well as an express margin of capital that should be required to ensure that the firm is able to remain solvent for a reasonable period after the event. At present there is no assurance that the existing capital requirements set by the Basel Committee would achieve this objective.

However, while we believe that there is a strong case for a change to the current methodology for calculating regulatory capital, we also believe that many of the concerns which we identify in the paper could be addressed by some relatively straightforward adjustments to the approach which is currently being considered by the Basel Committee. Moreover, we believe that an approach of the type we advocate which is based on a measure which we refer to as 'regulatory equity at risk' would have the great benefit of establishing common regulatory standards that could apply to all internationally active financial firms, not just those which are currently regulated as banks.

The key elements of our proposal are as follows

• At the outset, we recognise that a regulatory capital requirement should continue to play a central part in any international regulatory system. However, we believe that the Basel Committee should be clear that the regulatory capital requirement exists to protect against unexpected and stress loss events and not to protect against expected losses as is the case with the current system

• At the same time, we believe that it is important to recognise that any measure of the adequacy of a firm's capital to protect against the risk of loss can only be assessed effectively by comparing the risks that a firm faces with the likely earnings that are available to offset the loss.

• In this context we also argue that the class of capital to which the regulatory capital requirement applies should be limited to the firm's equity, own funds and equity-like instruments since these are the only class of liability which is available to the firm to absorb losses should they arise. Other forms of funding may help to enable a firm to survive a loss, and thus form an important part of a firm's liquidity risk management process, but if they cannot be used to absorb loss, we suggest that they should not constitute part of the regulatory capital of the firm.

• A further aspect of our proposal is that the regulatory capital system should expressly focus on the type of stress events which can give rise to systemic risk and thus should be based on measures of potential stress loss rather than on measures of 'normal' earnings volatility, which is what most portfolio risk measures such as VaR and credit portfolio models seek to measure.

• We also argue that the regulatory capital requirement should incorporate an explicit regulatory safety margin. While the determination of the appropriate measure for such a safety margin remains a decision for the regulators, we argue that this should be set at a level which provides a suitable level of assurance that the firm will be able to survive a stress event and still remain solvent for long enough to enable an effective work out plan to be put in place.

• Finally , we argue that any regulatory approach which seeks to minimise international systemic risk should place a much clearer emphasis on the importance of effective liquidity and funding management processes within firms. The objective of such a focus is to seek assurance that firms would be able to survive both firm specific funding problems as well as more general periods where market liquidity is limited without themselves causing additional systemic problems.

Whilst we recognise that any new approach of the type we advocate would require considerable further analysis and discussion, we are confident that it would have significant advantages over the current Basel Committee proposals. By comparing a firm's demonstrated earnings capability with the potential stress losses that it faces, the approach would not only align regulatory standards more directly to developing risk management practice within firms, but would also provide a common framework for setting capital standards for those firms which are active in international markets but whose primary business and risks are not related to a portfolio of trading and/or banking products. A further advantage of the approach is that it would not require the development of the sophisticated risk models for all classes of risk which are envisaged under the current Basel Committee approach, but could to a significant extent be implemented based on firms' existing risk management processes and management information.

In the rest of this paper, we develop the argumentation which leads to these conclusions in more detail. We start with a review of the current situation and the rationale for establishing a common internationally agreed standard for financial supervision. We then analyse the implications of recent changes in credit risk management practice and in the nature of international finance for the approach to setting regulatory capital standards and the understanding of systemic risk. In the light of these changes we examine the need for a common definition of risk, a clear basis for establishing the adequacy of the regulatory capital requirement as well as the need to address the linkage between liquidity and capital adequacy.

This analysis allows us to derive the key requirements which need to be met by an effective regulatory capital system. Using these requirements as our benchmark we then compare the proposed Basel Committee approach for setting regulatory capital standards with an approach based on an analysis of the sources of potential earnings volatility and conclude that the latter is a more robust approach. We then outline how our proposed 'equity at risk' measure might work in practice, based on a combination of developing good practice in stress testing market and credit risk portfolios and the information which is already available about the actual volatility of firms' costs and earnings. Finally we consider briefly the issues which would be involved in setting the 'regulatory multipler' that would be required to establish an adequate safety margin under such a system and we indicate how a transition could be effected between the current Basel Committee approach and the new approach.