This statement sets out the principles on which a system of bank capital regulation should be based, shows how the proposed New Basel Capital Accord fails to meet these standards, and provides recommendations which would significantly improve capital regulations. These recommendations are based on our strong belief that regulators are not generally in a better position to assess risk than financial market participants.

I. Principles

All proposals to reform or improve capital regulation should, in our view, adhere to the principles below. While we recognize that making these principles operational and implementing them will have to take account of specific conditions in individual countries, they should nevertheless guide the development and implementation of capital regulation in all countries.

1. Banks should maintain a level of capital that is sufficient to:
(a) reduce the likelihood of bank insolvencies to a level consistent with a stable banking system;
(b) immunize taxpayers from losses incurred by government-guaranteed bank claimants in the event of bank insolvencies; and
(c) align the incentives of bank owners and managers with those of uninsured bank claimants with respect to the risk assumed by banks.

2. Capital should be measured so as to maximize the use of market information.

3. Capital levels and risk exposures should be disclosed publicly and frequently.

4. Bank supervision should be administered by competent regulators who are independent of political and bank industry pressures and are publicly accountable.

5. Rules and supervision should be designed to enhance market discipline.

6. Rules should be designed to identify and disclose connected lending and provide appropriate safeguards.

7. Rules should include an effective mechanism for enforcement of capital regulations.

II. Inadequacies of the Basel proposal with respect to the principles of capital regulation

1. The Basel proposal will not cause banks to maintain a sufficient level of capital relative to their risk exposures.

The use of a bank's internal ratings provides many opportunities to game the rules. The proposed accord allows banks to use internal ratings under certain conditions, subject to evaluation and acceptance by the bank’s supervisory authority. This evaluation is complicated by the fact that banks’ practices in credit-risk assessment vary substantially, from the highly intuitive placement of credits into risk categories to the use of fairly sophisticated risk-assessment models. The ability of supervisors to prevent gaming is likely to be limited despite the more intensive supervision recommended in the proposal. This is important because banks have an incentive to minimize required regulatory capital for a given level of economic risk.

External ratings cannot be counted on to measure adequately the risks associated with bank loans. With respect to external ratings, there are a number of difficulties. Rating agencies have little experience in risk-rating bank borrowers; rating each and every bank borrower would be very costly; and the agencies’ track record suggests that they have been better at risk confirmation than risk diagnosis. Furthermore, were ratings to be used for the purpose of determining required capital, it is likely that rating firms would be established to provide biased ratings.

A separate concern is that firms and banks in emerging market economies may be adversely affected by any sudden downgrade of their home country's sovereign rating.

2. The proposal does not require that capital be measured on a market value basis.

3. Although the recommendations and requirements for additional disclosures are desirable and will provide more information, the proposal does not go far enough to enhance the role of market discipline in determining capital adequacy. The Basel Committee relies on information disclosure to create market discipline. However, effective market discipline requires not only that information be available to the market, but also that market participants have the incentive to act on it. As long as depositors and other creditors of banks are explicitly or implicitly protected against loss, they will be less inclined to demand relevant disclosure concerning risks or losses and to act on the information that is disclosed.

4. The number and complexity of the proposed rules will make it harder to hold regulators and supervisors accountable for their judgements about bank risk, and may result in increased regulatory forbearance.

5. Regulators need to address more urgently significant conflict of interest problems that have often caused bank failures in many countries, particularly the problem of "connected lending" — loans to insiders, major shareholders, and their affiliates.

6. Although the Basel Committee apparently did not consider enforcement to be within their mandate, it should be clear that no capital regulation regime can be effective without a strong enforcement mechanism.

III. Recommendations

The following recommendations, which could be adapted to the circumstances of individual countries, reflect our concerns on two key issues. The first is the need for an effective enforcement regime. The second is the fact that regulatory reliance on internal ratings lacks transparency, strengthening the need for supplementary measures aimed at increasing market discipline.

Capital requirements need effective enforcement. In the United States, for example, a system known as prompt corrective action attempts to mimic the sanctions the market would impose in the absence of a government-sponsored safety net. An effective enforcement mechanism based on this model should include the following features:

1. multiple capital zones with progressively stricter regulatory sanctions,

2. a specified equity closure rule when the equity is still above zero,

3. resolution of insolvencies at least cost to tax payers.

The appropriate amount of capital should resemble what banks would hold in the absence of a government-sponsored safety net. Thus, in addition to the Basel risk-weighted ratios, the relevant capital standard would include a leverage ratio based on total assets (both on- and off-balance sheet assets).

In order to strengthen market discipline, we recommend that banks in industrialized countries issue a minimum amount of credibly uninsured subordinated debt. Holders of such debt are sensitive to and uniquely positioned to monitor default risk. Therefore, the Basel capital rules should be supplemented with signals from these uninsured creditors. To encourage this, we recommend that the distinction between Tier I and Tier II capital be removed.

Although banks face a degree of market discipline from the stock market in most industrialized countries, the information provided by stock prices becomes progressively less reliable as banks approach insolvency.

This subordinated debt proposal can be implemented experimentally in phases so as to minimize costs and permit review of its effectiveness. Initially, the requirement could be limited to large institutions whose debt is actively traded. In addition, at the outset, the market signals from subordinated debt need not mandate any required supervisory action; however, the yields at which this debt trades will provide helpful information to both the market and the supervisors.

In the case of emerging markets, effective use of subordinated debt as capital and as a signal of bank strength may be limited by the absence of a liquid and developed capital market. However, this should not prevent emerging market countries from identifying and developing signals of bank strength--by for example encouraging banks to offer credibly uninsured certificates of deposit.