Federal Deposit
Insurance Corporation

November 3, 2003

Dear Sir or Madam:

State Street Corporation is pleased to have the opportunity to comment on the Advance Notice of Proposed Rulemaking (ANPR) and draft supervisory guidance issued by the Office of the Comptroller of the Currency, the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation, and the Office of Thrift Supervision (the Agencies) in relation to implementation of the New Basel Capital Accord (New Accord). As an internationally active bank, State Street appears to meet the proposed criteria for "core bank," and, under the ANPR, would be required to comply with both the Advanced Measurement Approaches (AMA) for operational risk and the Advanced Internal Ratings Based approach (A-IRB) for credit risk.

State Street remains very concerned by the proposal to create new regulatory capital requirements for operational risk. While management of operational risk is an important goal of both banks and their regulators, the proposed Pillar I treatment of operational risk will create unnecessary competitive disadvantages for U.S. banks compelled to operate under the Accord, compared to non-banks, overseas competitors, and U.S. "general" banks. We continue to urge the Agencies to eliminate the proposed new Pillar 1 capital requirement for operational risk, and create instead a rigorous Pillar 2 supervisory approach.

In addition, State Street is concerned that the mandatory application of the A-IRB to all U.S. banks subject to the New Accord will be unnecessarily burdensome and costly for some banks. We believe the goal of creating a more risk sensitive credit risk regime could be better achieved through a more flexible approach. The A-IRB may be suitable for very large banks or banks focusing on traditional bank lending activities. State Street, however, is a specialized bank, focused on fee-based investment servicing and management, with relatively modest exposure to traditional bank lending. For such specialized institutions, the high cost and complexity of the A-IRB will create significant compliance burdens, with little or no risk management benefit. We urge the Agencies to provide for an alternative treatment for credit risk for such institutions. Suggested parameters for such a proposal are detailed below.

We appreciate the Agencies' interest in industry comments, and we would be pleased to further discuss potential changes with the Agencies. Specific comments on these and other issues raised by the ANPR and draft supervisory guidance follow below.

Application of the Advanced Approaches in the United States

"Core Bank" Definition Lacks Risk Sensitivity

It is unclear how the definition of "core bank" based on the level of foreign exposures relates to the mandatory imposition of the AMA and A-IRB. While "core banks" meeting the ANPR's asset size criteria may be assumed to have suitable exposures --- and resources --- for these approaches, the existence of over $10 billion in foreign exposures does little to distinguish the risk profile of banks such as State Street from their peers and competitors that would not be subject to the New Accord. In fact, for State Street, these foreign exposures are generally in the form of low risk, short-term exposures to highly rated international banks. Such exposures provide little rationale for triggering costly and inappropriate compliance with only the most advanced approaches to capital offered by the New Accord.

The Agencies should identify and clarify the risks raised by foreign exposures, and offer an alternative approach to implementation of the New Accord which more closely relates to these risks.

U.S. Implementation of New Accord Remains Overly Prescriptive

The Agencies should, to the greatest extent possible, adopt a "principles-based" approach to implementing the New Accord. Unlike the highly prescriptive requirements of the ANPR, a principles-based approach would allow regulators the flexibility necessary to address the wide variation in market position, management structure, and risk profile of banks expected to be subject to the New Accord. In addition, a principle-based approach would provide more suitable flexibility for regulators and banks to adapt to changing circumstances over time, including the introduction of new products, evolving client needs, and changing economic conditions.

The U.S. Proposed Implementation Lacks Regulatory Incentives

We are concerned that the Agencies appear to have decided not to adopt a key component of the Basel Committee's concept for the New Accord --- the creation of regulatory incentives for the adoption of more advanced approaches to regulatory capital.

As proposed by the Basel Committee, banks will be provided the option of choosing the capital measurement approach most suited to their market position, size, and corporate structure. The potential benefits of more risk sensitive capital calculations are expected to function as incentives for adoption of the more advanced approaches.

The Agencies have proposed a different approach. Instead of the "incentives-based" approach offered by the Basel Committee, the Agencies have chosen to impose a "sanctions-based" approach, which requires adoption of the most advanced --- and expensive --- capital measurement approaches by all banks subject to the new Accord, regardless of their risk profile, corporate structure, or size. As a result, we are concerned that "core" and "opt-in "banks will be forced to adopt expensive systems which may not match their risk management needs, while remaining "general" banks will face considerable disincentives to invest in improved, more risk-sensitive, risk management techniques.

Timing and Transition Issues

The requirement to have in place both the AMA and the A-IRB by year-end 2006 may be overly optimistic, and will create undue compliance burdens for institutions designated as core banks. As noted in the ANPR, compliance with the New Accord will require substantial and time-consuming effort, and result in considerable costs. In fact, to meet the compliance targets proposed by the ANPR, certain types of data would need to be collected starting nearly immediately, long before the New Accord is finalized, and well before the Agencies have defined regulatory requirements and issued final implementation rules.

While it may be appropriate to offer the option of movement to the New Accord for banks desiring to make immediate, expensive investments in new systems and procedures, the Agencies should avoid compelling banks to make such investments prior to the finalization of the new capital regime. The Agencies should delay proposing a mandatory compliance date for the New Accord until regulations are finalized. At that point, regulators should ensure that any mandatory compliance date allows ample time for the development of systems and collection of appropriate data by core banks.

Leverage Ratio Requirements Unnecessary

While the goal of the New Accord is to create a more risk sensitive regulatory capital regime, the ANPR retains numerous non-risk sensitive features. The most restrictive of these non-risk sensitive elements is the proposed retention in the U.S. of the leverage ratio requirements.

Even under Basel 1, the leverage ratio functioned as a non-risk sensitive constraint above and beyond the risk-based capital requirements. In many cases, despite the stated goal of creating risk-based capital requirements, banks have been required to manage balance sheets primarily with regard to the leverage ratio. The Agencies, in the ANPR, concede that under the New Accord, in some cases, the leverage ratio will remain "as the most binding regulatory capital restraint."

This situation is particularly inappropriate under the system proposed by the ANPR, which requires "core banks" to make extensive investments in new systems for both credit and operational risk. In addition to meeting difficult quantitative, data, and system requirements, the capital requirements derived from these systems will be subject to strict qualitative supervisory standards. Once a regulator determines that all of these standards are met, a bank will then be required to operate a "parallel-run year," where it calculates regulatory capital under both the New Accord and existing general risk-based capital rules. Finally, once a bank is permitted to use the New Accord on a stand-alone basis, the Agencies will impose capital floors based on the existing system for at least the first two years of the New Accord --- and retain the ability to maintain these floors indefinitely, on a bank-by-bank basis.

With all of these precautions in place, regulators will have more than adequate resources at their disposal to address any perceived capital deficiencies. Under such a system, it is difficult to envision the value added by defaulting to the long outdated leverage ratio as a binding regulatory capital restraint, and it should be abandoned.

Materiality of Exposures

We appreciate the Agencies' proposal to exempt from the advanced approaches exposures in non-significant business units and immaterial asset classes. Providing an exemption for non-significant business units is appropriate, and any criteria for "nonsignificant" will necessarily be related to the size of the business. In defining "immaterial asset classes," however, we urge the Agencies to focus criteria primarily on the level of risk associated with a given portfolio or type of exposure, within the context of an institution's overall risk profile. The quality of an asset is far more material to an institution's risk profile than either its relative or absolute size. Banks should be provided the option of deeming very high quality types of exposures "immaterial," and thus subject to the general risk-based capital rules.

Home/Host Issues

We agree with the Agencies' comments identifying a "level playing field" as an important consideration in implementing the New Accord. In addition to requesting general comments on the competitive impacts of the proposal, the Agencies specifically invite comment on the treatment of U.S. banking subsidiaries of foreign banking organizations. This is an important issue, but, for U.S. banks, the treatment of U.S. banks operating overseas is also a critical aspect in the implementation of the New Accord.

While not strictly under the domestic rulemaking authority of the Agencies, it is essential that this issue be definitively resolved between jurisdictions prior to implementation of the New Accord.

While the Basel Committee's publication in August 2003 of "High-level principles for the cross-border implementation of the New Accord" provided some insights into the nature of issues raised by cross-border implementation, it is far from definitive on the subject. The decision by the Agencies to limit U.S. implementation of the New Accord to only the AMA and A-IRB raises additional issues related to the treatment of U.S. core banks overseas, due to both the presumed need for overseas regulators' acceptance of the supervisory requirements of these approaches, and the competitive and other factors raised by U.S. banks operating in jurisdictions where the full range of capital methodologies offered by the New Accord are in use by domestic banks.

We urge the Agencies to aggressively seek resolution of the home/host issue in Basel, and to defer any decisions on regulatory treatment of U.S. subsidiaries of foreign banking organizations until the issue is resolved on an international basis in a manner that provides reasonable and fair treatment of U.S. banks operating abroad.

Boundary Issues

We appreciate the Agencies' explicit acknowledgement of the importance of boundary issues. It is essential that regulators provide simple, explicit guidance regarding the treatment of losses which may be attributable to multiple risk factors, both to facilitate the development of comprehensive risk management systems, and to avoid potential "double-counting" of risk factors.

Cost and Complexity

The Agencies have requested comment on the potential implementation cost of the New Accord for core and opt-in banks. Public estimates suggest that implementation of the New Accord will be very costly. A recent study by Oliver Wyman & Company, for example, estimated that implementation of the New Accord will require investments approximating 5 basis points of assets, resulting in a global implementation cost of $25 billion. The study estimates an individual cost of from $50 million to $200 million for the largest banks. The Financial Service Roundtable has reported that one of its member companies has estimated the implementation cost at between $70 million to $100 million.

The complexity and lack of clarity regarding certain aspects of the proposal, the numerous changes to the proposal in recent months, and the decision by U.S. regulators to require use of only the most advanced approaches for core banks have made it difficult to predict a specific anticipated cost for State Street. However, it is clear that implementing the New Accord will require substantial investment over the next four years.

The Agencies also have requested comment on the balance achieved between the objectives of simplicity and consistency across banks on one hand, and risk sensitivity on the other. We appreciate the Agencies' explicit acknowledgment of the importance of all three of these factors, and recognize that an appropriate balance will be difficult to achieve. As mentioned in other areas of these comments, for a specialized bank such as State Street, we believe the proposed mandatory use of the A-IRB is overly complex and expensive relative to our risk profile. In addition, we believe that the AMA creates an overly complex system of regulations, aimed at arriving at a capital requirement for risks that could be better addressed through a Pillar 2 supervisory approach.

Advanced Internal Ratings-Based Approach (A-IRB)

Proposal for Hybrid Approach to Credit Risk

As mentioned above, State Street is concerned by the Agencies' proposal to require all banks operating under the New Accord to use the A-IRB for credit risk. While the AIRB may be appropriate for very large banks, or banks focused on traditional bank lending activities, it provides little risk management benefit to specialized banks focused on fee-based business lines, such as State Street, and would require unnecessary investment in highly complex risk management systems and processes.

We believe that an alternative to the A-IRB could be made available for both core and opt-in banks meeting certain criteria for asset size and credit quality. For example, banks with assets below the proposed $250 billion core bank asset size criteria, and with credit risk profiles consisting of predominantly investment grade or equivalent exposures, could be provided an alternative to the A-IRB In general, an alternative credit risk treatment could follow the general risk-based capital rules, with use of more advanced approaches reserved for selected exposure types where the resultant increased risk sensitivity is most relevant. Criteria for use of the advanced approaches could include a variety of factors, including such elements as the significance of a bank's business activity within an industry segment.

Such an approach would provide a more suitable regulatory capital regime for banks with specialized credit portfolios, providing the flexibility needed to develop a risk management approach with a level of system sophistication, oversight, and governance commensurate with the risk profile of an institution's underlying portfolios. Under such a regime, unnecessary costs for core banks would be reduced, and general banks specializing in non-credit business lines would be provided greater incentive to opt-in to the New Accord.

Asset Securitizations

Asset securitizations are an important component of U.S. capital markets, providing an important source of liquidity. While it is important for regulators to address the risks associated with securitizations, the proposal included in the ANPR is overly complex and unworkable. As a result, the ANPR proposal would risk significant disruptions to the financial markets.

As indicated in our comment letter on CP3, State Street is concerned that the Basel Committee's proposal does not properly recognize all types of asset securitization activity by banks. CP3, and the ANPR, recognize only two types of asset securitization participants --- investors and originators. As a sponsor of asset-backed commercial paper conduits, State Street would be deemed an originator under the ANPR. However, State Street does not directly or indirectly originate any of the underlying exposures held by the conduits it sponsors, and does not have access to the proprietary PD (probability of default) data required under the Supervisory Formula Approach (SFA). Therefore, for State Street, the Agencies' assumption that originators, as defined by ANPR, "are presumed to have much greater access to information about the credit quality of the underlying exposures" is incorrect, raising significant challenges to meeting the requirements of the Agencies' proposal.

We are encouraged by the recent announcement by the Basel Committee that the securitization proposal will be simplified, and that the "supervisory formula" will be replaced by a less complex approach. We look forward to working with the Agencies on the development of a simplified approach that recognizes the full range of banks' asset securitization activity.

AMA Framework for Operational Risk

As we have commented in the past, State Street strongly opposes the proposed new capital requirement for operational risk. We continue to believe that the proposed treatment of operational risk will have negative competitive effects for U.S. banks compelled to operate under the Accord. The ANPR and associated draft supervisory guidance fail to address these broad policy implications.

State Street is highly attentive to the issue of operational risk, and effectively managing such risk is a key element of confidence in our relationship with our clients. Our long record of extremely low operational losses validates our ability to manage such risks. However, we continue to believe that operational risk is first and foremost an "earnings-at-risk," not a "capital-at-risk," issue. In our experience, operational losses have been covered many times over by earnings.

Regardless of the outcome of the deliberations of the Basel Committee, the Agencies should address operational risk under a rigorous Pillar 2 supervisory system. Placing operational risk management under Pillar 2 will allow the Agencies to impose strict operational risk regulatory requirements, including requirements related to capital adequacy, without creating the negative competitive and technical impacts of a Pillar 1 regulatory requirement.

U.S. banks operate in a highly competitive environment, competing directly with investment management firms, broker/dealers, insurance companies, investment banks, mutual funds, leasing companies, and business services and software companies. None of these non-bank competitors are subjected to the proposed capital requirements. In addition, under the ANPR, the great majority of U.S. banks will remain under Basel 1, unless they choose to opt-in to the New Accord.

While the credit risk benefits of the New Accord may offset the negative impact of the new operational risk requirement for banks engaged in traditional lending activities, precisely the opposite is true for banks providing fee-based services, such as investment servicing and management. For banks engaged in these businesses, the New Accord and the ANPR impose an additional capital requirement, largely through the operational risk requirement, while leaving their competitors with no capital requirement for operational risk at all. The result is a marketplace distortion, creating a regulatory incentive for banks to move activities outside of the reach of the New Accord.

We are also concerned that the operational risk components of the New Accord may be applied inconsistently across national jurisdictions. As has often been noted by the Comptroller of the Currency, U.S. banks operate in a far different, in many ways stricter, regulatory environment than our non-U.S. competitors. For example, the common use by U.S. regulators of on-site examination teams is simply not the practice in most other jurisdictions. In addition, the U.S. proposal to retain both the Leverage Ratio and Prompt Corrective Action regime for U.S. banks operating under the New Accord creates additional capital demands on U.S. banks. As a result, any additional capital requirement for U.S. banks, including the operational risk requirement, will have a greater impact on U.S. banks than banks operating in other national jurisdictions.

The ANPR, and CP3, attempt to treat operational risk under a model more suited to credit risk. The proposal, by focusing primarily on regulatory capital, shifts attention and resources away from critical risk management efforts, and creates perverse incentives to avoid investment in important operational risk infrastructure, processes, and people.

We urge the Agencies to address these concerns by adopting a strong Pillar 2 approach to operational risk.

We also remain concerned by more technical aspects of the operational risk proposal. While these concerns are particularly acute under the ANPR's proposed Pillar 1 treatment of operational risk, they would also apply to a Pillar 2 supervisory system based on the AMA proposal as well. These concerns include:

• Use of External Data: Under the ANPR, the AMA continues to rely too heavily on the potential use of external data. External data provides valuable information for qualitative reviews of an institution's risk management systems and internal controls. It is not, however, suitable as a basis for a quantitative capital calculation.

Scaling external data to make it relevant to a bank's risk profile, system of internal controls, and other risk management factors is a difficult and uncertain process. Moreover, the integrity, completeness, and general data quality of external database are often questionable, and difficult to ascertain and control. Much publicly available operational loss data is based on relatively extreme risk management failures. Mandating the use of such data risks imposing capital requirements based on the "lowest common denominator" of risk management practices --- an approach that would penalize banks with well-developed control systems, and low losses. Sharing such data between institutions, or other third parties, will also raise serious privacy, confidentiality, legal, and competitive issues.

• Operational Risk Measurement: We remain concerned that the nascent state of operational risk measurement methodologies creates a serious obstacle to identifying a precise, risk-sensitive regulatory capital requirement for such risks. As a result, any capital adequacy evaluation should take place in the more flexible Pillar 2 supervisory element on the New Accord.

• Indirect Losses: The Agencies specifically solicit comment regarding the possible inclusion of the risk of indirect losses, such as opportunity cost, in the definition of operational risk. We urge the Agencies not to add such risks to the operational risk definition. First, quantifying the risk of such losses, and converting that risk to a capital requirement, will be next to impossible. Second, such indirect losses are far more related to a bank's business plan than its risk management function, and should continue to be appropriately accounted for in a bank's income statement, not its regulatory capital requirements.

• Legal Risks: While legal risk is certainly a factor in an institution's risk profile, we are concerned that such risks are among the most difficult to quantify for purposes of a Pillar 1 capital requirement. Litigation loss history provides limited insights into future losses, creating significant challenges to modeling. Since legal losses are typically closely linked to individual events and circumstances, the use of external data is particularly inappropriate for legal risk. Finally, U.S. securities law already addresses litigation losses, requiring reserving for material legal risks.

• Insurance: Finally, the Agencies should eliminate the proposed 20% cap on mitigants, such as insurance. This 20% cap does not appear to have any analytical or statistical basis. In fact, insurance can provide far more leverage than capital in addressing the low frequency, high severity loss events which may exceed an institution's ability to cover with earnings. It is appropriate for regulators to retain the ability to impose conditions on the use of insurance as a mitigant, and, perhaps, to limit credits against capital on a case-by-case basis. However, imposing a specific regulatory cap of 20%, or any other percentage, will create a disincentive for banks to hold insurance, will stifle innovation in new insurance-related (and other) risk management products, and will greatly reduce the risk sensitivity of the proposed New Accord.

Disclosure

In general, we are concerned that the Agencies' Pillar 3 disclosure proposals create significant comparability issues between banks within the U.S., between U.S. and nonU.S. banks, and between banks and non-banks.

First, core and opt-in banks will be required to make significantly greater and different disclosures than general banks, greatly reducing comparability between banks, and reducing the transparency for the overall banking system.

Second, the use of internal modeling and other highly subjective methodologies will result in significant inconsistency of disclosures among core and opt-in banks. For example, for credit risk, the Agencies propose to allow supervisors to exempt exposures in non-significant business units and immaterial asset classes from the A-IRB. While this may be a desirable option from a risk management perspective, the subjective nature of determining factors such as materiality create additional challenges to comparability.

Third, inconsistent application of the Pillar 3 requirements in other regulatory jurisdictions will create further competitive disadvantage for U.S. banks.

In addition:

• We note that the Agencies' proposal will require a summary table on banks' public websites indicating where all required disclosures may be found. This will create the need for users of financial data to access several sources (even though directed by a single website) to assemble a complete discussion of a bank's risk environment. An alternative, and more desirable, solution would be to require all required disclosures in Forms 10-Q and 10-K filed with the Securities and Exchange Commission.

• Quarterly disclosure requirements, as proposed, will add a significant burden to our reporting requirements, with little or no added benefit. We believe annual disclosure for most of this additional information is adequate.

• While there is no specific requirement to have this additional information audited by external auditors, it is likely that these disclosures will be subject to audit as part of our quarterly and annual reports, increasing the scope and cost of our audits.

• In response to the specific request for comment on the Agencies' description of the required formal disclosure policy, we believe additional guidance in the form of a sample policy is necessary. We recommend that the Agencies' guidance take into consideration the requirements related to Sections 302 and 404 of the Sarbanes-Oxley Act, and not create duplicative or conflicting requirements.

• For the operational risk-related disclosures, the requirement to discuss relevant internal and external factors considered in the Bank's measurement approach would require disclosure of operational loss history. Such disclosures raise serious privacy and competitiveness concerns. We believe this information should be available to supervisors only and not to outside parties.

All of these factors contribute to a regulatory regime under which the extensive and highly technical disclosures required by Pillar 3 will provide little meaningful transparency or market discipline improvements.

We urge the Agencies and the Basel Committee to work closely with the Securities and Exchange Commission, the Financial Accounting Standards Board, and the International Accounting Standards Board to develop disclosure requirements consistent with existing standards and practices.

Corporate Governance and Management Structure

The Agencies, in both the A-IRB and AMA, propose extensive new requirements for banks' corporate structure and management. While we appreciate the Agencies' assurances that they do not propose to dictate management structure of banks, we are concerned that the highly prescriptive structure of the Agencies' proposal provides little flexibility in establishing an internal risk management structure best suited to a bank's particular needs.

In particular, we are concerned that the Agencies' proposal places inappropriate and unduly burdensome responsibilities on a bank's board of directors, and does not clearly delineate the respective responsibilities of the board and senior management.

As proposed by the Agencies, the responsibilities imposed on the board of directors are excessively detailed, and go well beyond the board's appropriate supervisory and strategic role. For example, the ANPR requires the board to: maintain "effective internal controls over the banking organization's information systems and processes for assessing adequacy of regulatory capital and determining regulatory capital charges," approve all "significant aspects of the rating and estimation processes," and "ensure that appropriate resources have been allocated to support the operational risk framework." The ANPR also requires the board or a committee of the board to "oversee the development of the firm wide operational risk framework."

The board of directors' role in the risk management process should be supervisory, focusing on the oversight of management's activities and broad supervision of the implementation of regulatory requirements. The board should not be charged with the responsibility for the day-to-day risk management function of a bank. The board of any banking organization is unlikely to be comprised of directors with the time or skill set required to carry out the highly technical and extensive requirements proposed by the Agencies. The unintended consequence of placing excessively detailed demands on the board will be less board resources, time, and attention devoted to broader supervisory, strategic, and risk management responsibilities.

In addition to these broader concerns related to the proper role of the board of directors, the proposed rules do not sufficiently delineate the separate responsibilities of the board of directors and senior management. Many of the requirements proposed by the Agencies impose duties on the board and management combined, but do not specifically allow for division of responsibility between these two groups. These requirements do not provide sufficient guidance for the board to determine the extent to which it needs to be involved. For example, it is unclear if the mere approval by the board of funding to support an operational risk framework is sufficient to fulfill its responsibilities, or if the board's responsibilities can only be met through extensive development of detailed plans, policies, and budgets.

The Agencies should revise the ANPR to more closely align any new board responsibilities with the board's strategic and oversight roles, to avoid placing management functions on the board, to provide for clearer delineation of board and management responsibilities, and, when appropriate, to allow board delegation of its responsibilities to either board committees or senior management.

Conclusion

Once again, State Street appreciates having the opportunity to comment on this important issue. Despite our concerns, we are supportive of the Agencies' efforts to reduce risk in the global financial system, and remain willing to work with regulators on appropriate implementation of a potential New Basel Capital Accord.

Sincerely,

David A. Spina
Chairman and Chief Executive Officer
State Street Corporation
Boston, MA

cc:

Catherine Minehan, President and CEO, Federal Reserve Bank of Boston