Summary of the Basel Committee's: "The New Basel Capital Accord" 

On January 16, 2001, the Basel Committee on Banking Supervision (Committee) released the second consultative package on the new Basel Capital Accord (new Accord). The proposal modifies and substantially expands a proposal issued for comment by the Committee in June 1999 and describes the methods by which banks can determine their minimum regulatory capital requirements. Comments are due on the proposal by May 31, 2001 and the Committee intends to finalize the Accord by year-end 2001. The new Accord will apply to all "significant" banks, as well as to holding companies that are parents of banking groups.

The consultative package has three parts, each of which is increasingly more detailed. The package opens with an Executive Summary and Overview paper, which discusses the rationale behind the changes and highlights the primary elements of the new approach. The second section is known as the "Rules" document; it provides the details of the revisions and is intended to be the focal point of national rule-making processes. The final section comprises seven Supporting Technical Documents. Each of these documents focuses on a specific area of the proposal and provides the technical details on the different issues, along with focused questions. The seven documents address the standardized approach, the internal ratings-based (IRB) approach, supervisory review, asset securitization, interest rate risk, operational risk and disclosure.

The proposed new Accord contains a number of complex elements. Many of the new approaches include inputs and specific calculations by individual banks, which, in turn, will require supervisors to validate the models and methods used to develop the inputs. The Committee has proposed that implementation of the new framework begin in 2004. A great deal of work will need to be done in the interim to ensure that both banks and supervisors are ready to implement the new framework by that time.

It is important to recognize that discussion is still ongoing as to the population of banks to which the new framework will apply. While the 1988 Accord was applied to all banks in the U.S., it has not been determined how broadly the new approach will be applied, particularly given the many complex elements that may not be needed for smaller, less complex institutions. There are several factors that will determine the ultimate implementation in the U.S. One factor will be the results of the recent comment period on the advanced notice of proposed rule-making for non-complex institutions ¹ . Another factor, for the IRB approach, will be whether particular banks have the data, processes, and controls in place to implement the new framework.

Structure of the New Accord: The Three Pillars

The new Accord has three mutually reinforcing "pillars" that make up the framework for assessing capital adequacy in a bank. The first pillar of the new Accord is the minimum regulatory capital charge. The Pillar 1 capital requirement includes both the standardized approach, updated since the 1988 Accord, and the new IRB approaches (foundation and advanced). Since this first pillar is likely to be the focal point for industry comment, it is described in some detail below.

Pillar 2 is supervisory review. It is "intended to ensure not only that banks have adequate capital to support all the risks in their business, but also to encourage banks to develop and use better risk management techniques in monitoring and managing these risks." This pillar encourages supervisors to assess banks' internal approaches to capital allocation and internal assessments of capital adequacy, and, subject to national discretion, provides an opportunity for the supervisor to indicate where such approaches do not appear sufficient. Pillar 2 should also be seen as a way to focus supervisors on other means of addressing risks in a bank's portfolio, such as improving overall risk management techniques and internal controls.

The third pillar recognizes that market discipline has the potential to reinforce capital regulation and other supervisory efforts to ensure the safety and soundness of the banking system. Thus, the Committee is proposing a wide range of disclosure initiatives, which are designed to make the risk and capital positions of a bank more transparent. As a bank begins to use the more advanced methodologies, such as the IRB approach, the new Accord will require a significant increase in the level of disclosure. In essence, the tradeoff for greater reliance on a bank's own assessment of capital adequacy is greater transparency.

The Standardized Approach

The 1988 Accord introduced the standardized risk-bucket approach for setting the minimum regulatory capital requirement, which is still used in the U.S. today. The approach has been subject to increasing criticism that it lacks sufficient risk sensitivity. The revised standardized approach enhances the 1988 Accord by providing greater, though still limited, risk sensitivity.

Key changes to create a more risk-sensitive framework include the refinement and addition of risk buckets, the use of external credit ratings, and a wider recognition of credit risk mitigation techniques. The proposal removes the 50% risk weight cap on derivatives contracts and increases the credit conversion factor for business commitments under one year to 20%. The new Accord also provides for a lower risk weight on certain commercial real estate loans that have historically low default and loss rates. (The agencies note, however, that U.S. commercial real estate would not be eligible for the preferential treatment.) Risk weights will continue to be determined by the category of the borrower--sovereign, bank or corporate--but within each of those categories, changes have been made.

For sovereign exposures, membership in the Organization for Economic Cooperation and Development will no longer provide the benchmark for a preferential risk weight. Instead, the sovereign risk weight will depend on the assessments of "eligible external credit assessment institutions" (ECAIs). To be an ECAI, the entity must meet certain criteria. Under the proposal, a sovereign with a AAA rating would receive a 0% risk weight, while a sovereign rated below B- would be subject to a 150% weight. The Committee has also developed an alternative proposal to allow supervisors to use ratings developed by certain export credit agencies.

There are two options in the treatment of claims on banks. National supervisors must select one of the options to apply to all banks. The first option requires that banks be assigned a risk weight that is one category less favorable than that assigned to the sovereign of incorporation. The second option bases the risk weight on the external credit assessment of the bank. Under this option, the bank can obtain a more preferential risk weight vis-à-vis the sovereign, but the overall risk weight cannot be lower than 20%.

Several changes have been made to provide a more risk-sensitive framework for corporate claims. Moving away from the uniform 100% risk weight for all corporate credits, a corporate claim would receive a risk weight of 20%, 50%, 100%, or 150% depending on its external credit rating. Unrated corporate credits will be rated at 100%; that is considered a floor and supervisors may raise the risk weight where default rates or other conditions warrant a higher capital allocation.

The Internal Ratings-Based (IRB) Approach

The IRB approach represents a fundamental shift in the Committee's thinking on regulatory capital. It builds on internal credit risk rating practices of banks used by some institutions to estimate the amount of capital they believe necessary to support their economic risks. In recent years, as a result of technological and financial innovations and the growth of the securities markets, leading banking institutions throughout the world have improved their measurement and management of credit risks. These developments have encouraged the supervisory authorities to devote greater attention to developing more risk-sensitive regulatory capital requirements, particularly for large, complex banking organizations.

Banks must meet an extensive set of eligibility standards or "minimum requirements" in order to use the IRB approach. Because the requirements are qualitative measures, national supervisors will need to evaluate compliance with them to determine which banks may apply the new framework. The requirements vary by both the type of exposure and whether the bank intends to use the simpler "foundation" IRB framework or the more advanced IRB framework. A small sample of the minimum requirements includes:

• The bank has a risk rating system that can differentiate borrowers and facilities into groupings that are of similar levels of credit risk and across all levels of risk.
• There should be a meaningful distribution of exposure across grades with no excessive concentrations in any one grade.
• Borrower risk ratings must be assigned before there is a commitment to lend and must be reviewed periodically by an independent source.
• The board of directors and senior management have a responsibility to oversee all material aspects of the IRB framework, including rating and probability of default (PD) estimation processes, frequency and content of risk rating management reports, documentation of risk rating determinations, and evaluation of control functions.
• A one-year PD estimate for each grade must be provided as a minimum input.
• Banks must collect and store historical data on borrower defaults, rating decisions, rating histories, rating migration, information used to assign ratings, PD estimate histories, key borrower characteristics, and facility information.

As mentioned above, the requirements that a bank must meet are partially dependent upon which of the two IRB approaches a bank will use. The first methodology, called the "foundation" approach, requires few direct inputs by banks and provides several supervisory parameters that, in many cases, carry over from those proposed for the standardized approach. The second approach, the "advanced" approach, allows banks much greater use of their internal assessments in calculating their regulatory capital requirements. This flexibility is subject to the constraints of prudential regulation, current banking practices and capabilities, and the need for sufficiently compatible standards among countries to maintain competitive equality among banks worldwide.

There are four key inputs that are needed under IRB, for both the foundation and advanced approaches. The first element is the probability of default (PD) of a borrower; the bank is required to provide the PD in both the foundation and the advanced approaches. The second piece is the estimated loss severity, known as the loss given default (LGD). The final two elements are the amount at risk in the event of default or exposure at default (EAD) and the facility's remaining maturity (M). LGD, EAD and M are provided by supervisors in the foundation approach, but in the advanced approach banks are expected to provide them (subject to supervisory review and validation). For each exposure, the risk weight is a function of PD, LGD, and M.

The IRB approach envisions internal rating systems that are two-dimensional. One dimension focuses on the borrower's financial capacity and PD estimates that quantify the likelihood of default by the borrower, independent of the structure of the facility. The other dimension takes into account transaction-specific factors such as terms, structure, and collateral. These characteristics would determine the second dimension, i.e., the LGD. Implicit in this treatment is the assumption that when a borrower defaults on one obligation, it will generally default on all its obligations. (This assumption is relaxed with the IRB treatment of retail portfolios.)

Calculating the capital charge under the IRB approach involves several steps. The first of these steps is the breakdown of the bank's portfolio into six categories: corporate, retail, bank, sovereign, equity, and project finance. The IRB rules differ to varying degrees across these portfolios. As a result, the IRB charge is calculated by category, with the PD, LGD, and EAD inputs potentially differing across these categories. Supervisory approval is needed before banks can use the IRB approach for any of the six categories. The minimum requirements described above also differ somewhat across these six types of exposure. The IRB approaches are most developed for portfolios of exposures to banks, corporates, and sovereigns.

Another important step is the determination by the bank of the PDs for its loan grading categories in both the foundation and advanced IRB approaches. The PD of an exposure is the one-year PD associated with the borrower grade, subject to a floor of 0.03% (except for sovereign exposures, which are exempt from the 0.03% floor). The determination of PDs for borrowers supported by guarantees or credit derivatives is more complex. Banks under the advanced approach would use their internal assessments of the degree of risk transfer within supervisory defined parameters, while those under the foundation approach would use the framework set forth in the credit risk mitigation section. Overall, the PD must be "grounded in historical experience and empirical evidence," while being "forward looking" and "conservative." A reference definition of default has been developed for use in PD estimation and internal data collection of realized defaults.

Once the PD has been established, a second credit risk dimension -- loss severity or LGD -- must be determined. Under the foundation approach, the bank simply matches the collateral characteristics of the exposure to a specified list of LGDs, expressed as a proportion of the credit exposure. If the collateral type is not specified in the Accord, the exposure is considered unsecured and receives the corresponding LGD. If banks can meet the requirements for using their own LGD estimates, they can implement the advanced approach.

Then, for each facility the effective maturity, M, must be determined. To limit burden, under the foundation IRB approach each facility's M is assumed to equal three years. The Committee is considering several options for the advanced approach, of which the most developed would involve basing M on each facility's remaining contractual maturity.

After the bank determines the PDs and LGDs for all applicable exposures, those combinations can be mapped into regulatory risk weights. The risk weights are calibrated to include coverage for both expected and unexpected losses. Unexpected loss is a probability-based assessment of the losses that would occur under severe stress conditions. The risk weights are expressed as a continuous function, which provides maximum risk sensitivity and flexibility in accommodating diverse bank risk rating systems.

The capital charge is determined by multiplying the risk weight by the amount expected to be outstanding at the time of default, known as the EAD, and by 8%. If a bank has a high degree of single-borrower or single-group credit risk concentrations (a "non-granular" portfolio) within its non-retail credit portfolios, the bank would be required to increase the regulatory capital minimum by the granularity adjustment which is specified in the proposal. The adjustment would be a reduction in capital for a bank with a relatively low degree of single-borrower risk.

A final step in this process involves the ongoing review by the supervisors of the systems used to develop the IRB capital charge. Periodically, supervisors will need to validate these systems and review the internal controls that provide the foundation for the IRB approach. In addition, supervisors will also have to consider, under Pillar 2, whether the amount of capital generated by the IRB approach is commensurate with the bank's risk profile.

Operational Risk

One of the most significant changes in the new Accord is the proposal for an operational risk charge. It is expected to represent, on average, 20% of the minimum regulatory capital charge. The framework is based upon the following operational risk definition: "the risk of direct or indirect loss resulting from inadequate or failed internal processes, people and systems or from external events." Although the focus of operational risk is on the Pillar 1 capital charge, it also brings in elements of Pillar 2 (strong control environment) and Pillar 3 (disclosure).

The Committee is proposing a spectrum of approaches, which represent a continuum of increasing sophistication and risk sensitivity. The Basic Indicator Approach is the simplest of the three approaches to determine an operational risk charge. It allocates operational risk capital using a single indicator as a proxy for an institution's overall operational risk exposure. The current proposal would require banks to hold capital equal to a fixed percentage of its gross income. The Committee expects only the least sophisticated institutions to use this method.

To the extent that banks can demonstrate to supervisors an increased sophistication and precision in their measurement, management and control of operational risk, they would move along the spectrum to a more advanced approach, the Standardized Approach. Under this approach, supervisors establish standardized business lines (e.g., asset management), standardized broad indicators (e.g., total funds under management), and standardized loss factors (the beta) per business line. Within each business line, the capital charge will be calculated by multiplying the bank's relevant broad indicator measurement by the relevant beta factor. The total capital charge for operational risk would be the sum of the business line charges.

The most complex approach presented by the Committee as a current option is the Internal Measurement Approach. This approach, unlike the first two approaches, allows banks more direct input into calculating the operational risk capital charge. For standardized business lines banks would provide the following: an exposure indicator (EI), which is a proxy for the size or amount of risk of each business line; a parameter representing the probability of a loss event (PE); and a parameter representing the loss given that event (LGE). The product of EI*PE*LGE produces an expected loss (EL) for each business line/risk type combination. Regulators provide a standardized factor (gamma) per business line that translates the EL into a capital charge.

There are still a number of open issues related to operational risk and much work is needed to finalize the proposals, particularly the Internal Measurement Approach. The Committee has recognized this and is encouraging greater industry input in the development of an operational risk capital charge. In particular, there is currently only limited data to support the various operational risk charges. Additionally, more work is needed on defining loss types and loss events, risk categories, and business types. As part of this, the Committee will also have to look closely at the betas associated with the Standardized Approach to ensure that they appropriately reflect the risk associated with the individual business lines.

Asset Securitization

Given the rapid pace of innovation in the financial markets since the introduction of the 1988 Accord, the Committee believes that it is important to construct a more comprehensive framework to better reflect the risks inherent in the many forms of asset securitizations, including traditional and synthetic forms. The current Basel proposal is broadly similar to the proposals issued for public comment by the agencies in March 2000, although there are some differences. Similar to the U.S. proposal, the new Accord would use external ratings to assign asset-backed securities to the appropriate risk category. Also, the new Accord proposes methods to quantify the risks that are retained by banks after asset sales, e.g., the retention of residual or other subordinate tranches, servicing assets, etc. Another important feature is the proposal to incorporate an additional capital requirement in revolving securitizations that incorporate early amortization provisions.

Of particular interest to some U.S. banks is the new feature that would assess capital against short-term liquidity commitments, including servicer cash advances. Under the current Accord and U.S. regulations, no regulatory capital is required for these commitments. Finally, the Basel proposal outlines "clean break" criteria that must be met in order to remove securitized assets from the risk-based capital calculation. The proposed clean break criteria are analogous to the current sale criteria under GAAP. Unless all of the clean break criteria are met, a bank would receive no capital reduction when securitizing assets, because the bank would still be vulnerable to credit-related losses on those assets.

In addition to the ex ante criteria for a "true sale" the Committee also discusses ex post credit enhancements that might be provided by sellers of assets into securitizations. Specifically, there is concern that a bank might provide implicit recourse on the sold assets by supporting securitizations in instances where the pool of underlying assets experiences meaningful credit deterioration. This might be done through asset repurchases at prices that exceed the then-current market value, lending to the structure in ways that are not required by the contractual arrangement, or forgoing fees that the seller is otherwise entitled to receive. If a bank was to provide an ex post credit enhancement, the new Accord describes significant supervisory responses. These entail adding the credit risk amounts of the sold assets back to the bank's risk-weighted assets.

The papers highlight the Committee's intent to continue to work on a more risk-sensitive treatment for securitization under the IRB approach. In addition, treatments for synthetic securitizations under both the standardized and IRB approaches will be developed. The agencies expect this work to progress during the consultative period so that a more refined framework may be discussed with the industry in the latter part of this year.

Credit Risk Mitigation

The section on credit risk mitigation incorporates into the standardized and foundation IRB approaches rough approximations of the risk reduction attributable to various forms of collateralized credit exposures, guarantees, credit derivatives, and on-balance sheet netting arrangements. The Committee proposes a conceptual approach to these risk mitigation techniques that, while recognizing their risk reduction benefits, attempts to capture the additional risks posed by such transactions.

Collateralized credit exposures would include those arising from the lending of securities or the posting of securities as collateral that are secured by the cash or securities borrowed, such as occurs in repurchase agreements and securities lending transactions. Recognition is given only to financial collateral and includes listed corporate equities and investment-grade debt, in addition to cash and sovereign securities rated at least BB. Under the comprehensive approach, the risk that, in the event the counterparty defaults, the realized value of the collateral may be less--or that the value of the exposure may be more--than at last valuation is captured quantitatively in a haircut parameter, "H," that reduces the recognized collateral coverage. The values of H are collateral-specific and are set to approximate (in a very general way) the potential volatility in the value of the instrument over a ten-day holding period. Certain banks would be allowed to develop and use their own estimates of H based on specifications provided by the Committee.

A second parameter, "w", establishes a floor, below which the risk weight on the collateralized portion of the exposure will not fall. This formulation is based on the assumption that, regardless of the amount of collateral posted by the borrower, there remain unavoidable risks. For most transactions, "w" is set at 15%. For certain very low risk transactions, "w" is set at zero (and supervisors also may set H at zero). As an alternative to the comprehensive approach, a simple approach based on substitution with a floor is offered for collateral recognition.

A substitution approach that does not take into account double default effects is proposed for guarantees and credit derivatives extended by sovereigns and banks, as well as corporates rated A or better. Non-bank and non-sovereign guarantees would be discounted by a "w" factor of 15 percent, as would all credit derivatives. Only credit derivatives that meet certain criteria and take the form of a credit default or total rate of return swap are recognized.

The proposal would permit netting of on-balance sheet loans and deposits with a single counterparty where certain conditions are met. It would require portfolio netting arrangements to be decomposed and netted on an individual basis.

The proposal also describes a formulaic proportional adjustment to risk weights in instances where the collateral, guarantee, credit derivative, or netting arrangement will not be in place for the entire remaining maturity of the credit exposure. When the exposure enters the last year of its contractual maturity, no reduction in the risk weight would be permitted below that which would normally apply to the borrower (i.e., there would be no recognition given to the presence of the credit risk mitigant).

Future Work

The Committee will continue work on the revisions to the new Accord during the consultation period. There are a number of issues that are still to be resolved. These include:

• Equity Portfolios: The objective will be to develop a risk-sensitive approach to treating equity positions held in the banking book that is based on banks' internal approaches and takes appropriate account of the different types of equity holdings.
• Project Finance Portfolios: A working group will focus on mapping project finance exposures into a PD/LGD framework.
• Retail Portfolios: There are several key issues to resolve. The current risk weight formulas will need further refinement once additional data is available. The Committee will also explore the need for further risk weight parameters for different retail products.
• Securitization and Credit Risk Mitigation: Several elements remain to be completed, including the development of an approach to synthetic securitization under the standardized approach and treatments of securitization, guarantees, and credit derivatives under the IRB approach.
• Maturity: The Committee is examining various approaches to measuring M and calibrating the impact of effective maturity on risk weights.
• Overall capital: The Committee will be seeking to ensure that, on an overall basis, the proposals result in an appropriate amount of regulatory capital.

As part of these efforts, the Committee will also be working to collect data on the impact of the new Accord. It is possible that the Committee will release additional documents for industry consideration as some of the outstanding issues are resolved.

Following the conclusion of the comment deadline, the Committee will focus on finalizing the new Accord. Currently, the Committee plans to release a final version of the Accord by year-end 2001. Based on that release date, the implementation date has been set for 2004 to allow for domestic rule-making processes, and to allow banks and supervisors to adequately prepare for the use of this new Accord.

Discussion Issues for Industry Consideration

Respondents are encouraged to review and provide comments on the entire range of topics covered in the Committee proposal and to consider questions raised throughout. The agencies have highlighted below a number of specific issues of particular significance for the U.S. banking system and on which comment would be particularly welcome.

As commenters are analyzing the various approaches in the proposal, the agencies would be interested in any and all information and projections on the potential impact of the approaches to regulatory capital requirements. Most useful would be comparative analysis of the potential impact under each of the regulatory capital approaches set forth. This information may be provided directly to the agencies on a confidential basis. The agencies also will be participating with other members of the Committee in an exercise designed to provide estimates of the quantitative impact of various aspects of the proposal on a comparable basis.

Commenters on the Basel proposal who wish to submit comments directly to the agencies should address them as follows: Basel 2001 Capital Proposal, Office of the Comptroller of the Currency, mailstop 3-6, 250 E Street, SW, Washington, DC, 20219; Federal Reserve Board, Basel 2001 Capital Proposal, mailstop 179, 21 ^st and C Streets, NW, Washington DC 20551; Robert E. Feldman, Executive Secretary, Attention: Comments/OES, Federal Deposit Insurance Corporation, 550 17 ^th Street, NW, Washington, DC 20429.

Consistency with Broad Objectives

In all areas of the proposal, the agencies solicit comment on whether it achieves the following broad objectives:

1. Risk sensitivity: Would the proposal result in capital charges that are aligned with underlying risks? Would the proposal generate reasonably comparable levels of capital for equivalent risk taking? Are the proposed capital treatments, minimum requirements, and disclosure standards consistent with current and emerging sound banking practice?

2. Incentive compatibility : Would the proposal promote better risk management, rewarding better risk management processes while penalizing less effective processes? Do the different standards set forth in the proposal appropriately address and eliminate incentives for capital arbitrage (e.g., in comparing the standardized versus IRB treatments of credit risk, and in its treatments of securitization and credit derivatives)? Does the proposal provide adequate incentives for institutions to move beyond the standardized approach to the IRB approaches?

3. Competitive equity : Would the proposal place certain banks at competitive disadvantage to other banks or nonbanks? Among internationally active banks (using either the standardized or IRB approaches), would the proposal generate sufficient consistency in capital treatments across instrument types and national jurisdictions? Should some banks be required to operate under a particular approach? How could this be determined?

4. Safety and soundness : What are the industry's views on the overall amount of capital that would be generated by the proposal? Would the proposal generate prudential levels of bank capital while promoting economic efficiency and overall financial and macroeconomic stability? What effects, if any, are the changes in capital treatments under any of the approaches described likely to have on particular markets and/or the general availability of credit?

5. Market discipline : Would the proposal, which includes expanded risk disclosures by banks, enhance overall market discipline within the banking industry? Could the proposed disclosure requirements be improved? Has the Committee achieved the correct balance between burden and data required for effective market discipline? In particular, for disclosures related to the IRB approach, the Committee set forth a wide range of information from which its final disclosures will be crafted. Which of these disclosures are most relevant? For certain core disclosures that are not tied to explicit regulatory capital treatments, will market discipline effectively require banks to make such disclosures?

6. Implementation : Do the methodologies and standards required in the proposal achieve the right balance between rigor and burden? Can banks take the steps necessary to implement approaches under the proposal in a cost-effective manner within a reasonable amount of time? What are the industry's views on the proposed implementation date of 2004, given the current state of risk measurement practices?

Scope of Application

1. The proposal expands the application of the capital framework to holding companies that are parents of banking groups and requires the deduction of investments in insurance companies. What are the industry's views on these proposals?

The Standardized Approach to Credit Risk Capital

1. Are the "mappings" that assign ranges of external credit ratings to the various risk weight categories appropriate, for example with regard to corporate exposures? What overall percentage of the credit portfolio will likely be affected by the potential use of external credit ratings for corporate exposures? Are the risk weights that would be assigned to impaired and below-investment-grade exposures (which in many cases are less than called for under the foundation IRB approach) sufficient to ensure prudential amounts of capital and appropriate incentives to migrate to the IRB approach?

2. The proposed 20% and 50% credit conversion factors for short- and long-term lending commitments are substantially less than the 75% factor that would apply under the IRB foundation approach, although in the latter case capital requirements would also be affected by the PD and LGD associated with each exposure in question. Would such a disparity be important in influencing the decision to participate in the IRB approach or otherwise create a significant overall inconsistency between the two approaches?

3. Are the proposed criteria for "eligible" ECAIs sufficient to ensure the reliability of their rating processes? In addition to providing rating information to institutions with legitimate interests, should rating information also be made available to the general public to promote market discipline?

4. The proposal permits a preferential risk weight for certain commercial real estate loans in qualifying markets. Are the conditions for this preferential treatment set forth in the rules paper, along with the additional conditions that are available from the Basel Committee Secretariat, reasonable and appropriate given the risks traditionally associated with commercial real estate lending?

The IRB Approach to Credit Risk Capital

1. Are the proposed minimum requirements for the IRB approach, as well as the additional requirements for use of the advanced approach, consistent with sound banking practice and sufficient to provide confidence that minimum capital requirements for credit risk will be adequate and prudent?

2. Is the proposed structure of absolute and relative risk weights consistent with banks' internal credit risk measurement systems and, more generally, with ensuring prudential levels of capital?

3. The proposed IRB risk weights are calibrated to encompass both unexpected and expected credit losses because total capital is defined to incorporate general reserves. Therefore, not including expected losses could result in the same capital being assigned to cover both expected and unexpected losses. What are the industry's views on this issue? Are there alternative approaches that could achieve the same objectives in a different manner? 
 

4. Similarly, because the framework focuses on the amount of total capital required, calibration of the IRB risk weights entails assumptions about the average composition of capital between tier one and tier two elements. To what extent, if any, does the proposal create incentives for banks to modify their existing mix of such capital elements?

5. Under the foundation IRB proposal, risk weights for loans to banks, sovereigns, and corporates are calibrated under the assumption that all such loans have an average maturity of three years. What are the industry's views on the overall appropriateness of this treatment? Does this approach strike a reasonable balance between the additional burden of a separate maturity dimension and the desire to increase risk sensitivity? What impact, if any, would such an approach have on lending practices and, in particular, on the extension of short maturity credits?

6. Within the advanced IRB methodology--and possibly at national discretion, for the foundation methodology--the Committee is examining alternative approaches for linking risk weights and remaining maturity. One approach would permit use of banks' internal methods for estimating effective maturity. If banks were permitted to estimate effective maturity using their own internal methods, what minimum standards would be necessary?

7. The IRB Supporting Technical Document describes alternative approaches to calibrating the effects of maturity on risk weights. What criteria should guide the Committee in deciding how to incorporate maturity effects?

8. There is a tension between the prudent desire for banks to take a longer-term view of a borrower's credit quality and the need under the IRB approach to provide conservative estimates of average one-year default probabilities that can be used as a common metric for empirical validation and setting IRB risk weights. How do banks expect to resolve this tension? In particular, are credit assessments likely to be excessively focused on the next twelve months (i.e., in relation to well-established sound lending practice)? How could supervisors ensure that banks would conservatively and appropriately incorporate longer-term considerations in their rating and risk management processes?

9. Under the foundation IRB approach, senior and subordinated loans to corporates and banks would receive LGDs of 50% and 75%, respectively. In addition, real estate would be the only type of physical collateral recognized as a credit risk mitigant, and could potentially reduce the LGD to 40%. What are commenters' views on these proposed treatments? Is subordination sufficiently well defined for use in this context?

10. In regard to retail exposures, the proposal requires that PD, LGD, and/or EL be calculated separately for segments chosen by the bank but incorporating at least four risk factors - product type, borrower risk (e.g., credit score or equivalent), delinquency status, and time period of origination (i.e., vintage). Are these dimensions adequate to capture the risk characteristics of retail portfolios or should additional risk factors be required? Are the requirements consistent with sound retail lending and risk management practices?

11. Are the proposed risk weights for retail portfolios reasonable, especially for loans having high expected loss rates, such as credit cards? Specific recommendations as to alternative risk weights and estimation methodologies are encouraged.

12. Under the proposal, the same mapping from PDs and LGDs into risk weights would be applied to all retail exposures. Is this appropriate, or would separate risk weight calibrations for specific retail portfolios be worth the additional complexity?

13. A bank would not be required to estimate explicit credit conversion factors for uncommitted and undrawn retail lines of credit, such as credit cards. However, to ensure adequate coverage of potential credit losses, such a bank would be expected to reflect potential credit losses on such undrawn lines in its LGD estimates. Can such LGD estimates be developed in a robust and cost-effective manner? Should either approach be viewed as more risk-sensitive and more consistent with sound risk management practice?

14. Does the definition of retail exposures provide sufficient clarity as to whether and when small business lending should be considered a retail or a corporate exposure? How could this definition be enhanced to provide greater clarity and ensure that appropriate capital requirements are applied? 
 

15. Does the granularity adjustment strike an appropriate balance between complexity and additional risk sensitivity? Relative to the informational requirements associated with other aspects of the IRB proposal, what is the incremental burden associated with the proposed granularity adjustment?

16. As noted in the proposal, the Basel Committee's work in developing IRB treatments for equity and project finance exposures is at a relatively early stage. How might the proposed definitions of these two exposure types be improved? To what extent and in what respects do banks' internal risk and capital assessment methodologies for these exposures differ from those used for corporate loan exposures? What approaches should the Committee consider in addressing equities and project finance and how might the Committee ensure that, under the IRB approaches, banks maintain prudent levels of required capital against those portfolios?

Credit Risk Mitigation

1. Does the "H" and "w" framework for incorporating the risk mitigation effects of various forms of collateral, guarantees, and credit derivatives strike a reasonable balance between allowing additional elements of credit risk mitigation and providing safeguards against residual risks? Is the approach broadly consistent with current bank practices and risk measurement systems? Are the proposed values of H and w reasonable reflections of the credit risk inherent in different forms of collateral, as well as the different contract types? Are the required holding period assumptions appropriate for different types of transactions?

2. The agencies are preliminarily of the view that in lieu of the netting treatment set forth in the proposal, they would retain the netting criteria established under U.S. GAAP, which is generally more conservative. What are the potential competitive implications of this approach? Under the proposed treatment of on-balance-sheet netting, loans and deposits would have to be decomposed and netted individually. What effect would this treatment have on institutions' internal reporting systems, in particular for repurchase and reverse repurchase agreements netted under GAAP, and on incentives for appropriate risk management practices?

3. Is the availability of a simple approach to collateral based on substitution helpful and are there possible alternative simple approaches that result in at least as much capital as under the comprehensive approach?

4. Guarantees from banks and sovereigns are provided a more favorable treatment through a lower w factor than guarantees from other similarly rated entities. Is this treatment warranted?

5. What are the industry's views on the proposed approach for proportional recognition of mismatched hedges and the incentives that this approach might produce?

Asset Securitization

1. The risk weights for securitization tranches under the standardized approach are in several cases higher than the corresponding risk weight for a similarly rated corporate loan. For example, a BB-rated securitization tranche would receive a 150% risk weight while a BB-rated loan would receive a 100% risk weight. In part, this reflects that the ratings for securitization pools typically incorporate significant diversification benefits, while individual loans can benefit from additional diversification in the context of a large loan portfolio. What are the industry's views on the proposed risk weights for securitization tranches? In particular, are the risk weights for below-investment-grade tranches sufficient to address their increased risk?

2. The Committee proposes to apply a 20% conversion factor for liquidity facilities that enhance securitization. Is that conversion factor commensurate with the risk created by such facilities?

3. The Committee proposes to apply a 10% conversion factor to assets sold by sponsors into securitizations that contain an early amortization clause. Is it appropriate to characterize early amortization clauses within typical credit card securitization and CLO structures as a credit enhancement?

4. Under the standardized approach, what standards regarding the management of residual risks should banks have to meet in order to receive a capital benefit from a synthetic securitization transaction where the bank retains the most senior tranche of credit risk?

Operational Risk

1. Is the suggested definition of operational risk appropriate? How can loss types be better specified for inclusion in a robust definition of operational risk? To allow for the development of more advanced approaches to operational risk, what detailed guidance on loss categorization and allocation of losses by risk type will be needed?

2. Conceptually, a capital charge for operational risk should cover unexpected losses, while provisions and current revenues should cover those that are expected. Are accounting rules and practices among countries sufficiently compatible to accommodate a uniform, international regulatory capital standard based upon this concept?

3. How do the proposed approaches compare with a bank's own method for allocating capital for operational risk? What are the costs and benefits associated with collecting and maintaining operational loss information by business line?

Supervisory Review of Bank Capital Adequacy

1.Do the four proposed supervisory principles provide a sufficient basis for a consistent approach to supervisory review processes, including validation of minimum standards for participation in the IRB approaches, among international supervisors? 
 

Transition Arrangements

1. Do the proposed transition arrangements regarding the necessary data and other requirements for the IRB approach strike the proper balance between accommodating the need for banks to develop fully their systems and measurement capabilities while providing for prudent minimum capital requirements based on less than complete information?

2. During the first two years of implementation, the overall capital requirement of a bank following the advanced IRB approach could not fall below a floor equal to 90 percent of its requirement under a simplified calculation of the foundation IRB approach (see paragraph 162 of the "Rules" paper). What are the industry's views on the need for such a floor both in relation to competitive equity and to the reliability of bank internal estimates of LGD, EAD, and the effect of guarantees and credit derivatives? In particular, how significant are the differences between the foundation and advanced IRB approaches likely to be and what would be the source of such differences (e.g., more accurate estimates of risk)?

¹ "Simplified Capital Framework for Non-Complex Institutions," Federal Register, November 3, 2000.