Testimony of D. Wilson Ervin

Managing Director and Head of Strategic Risk Management
Credit Suisse First Boston

Hearing: A Review of the New Basel Capital Accord
Wednesday, June 18, 2003

[view file]






Good morning Mr. Chairman.  I want to thank you for holding these hearings today and inviting me to appear before the Committee.  My name is Wilson Ervin and I am a Managing Director of Credit Suisse First Boston ("CSFB").1  I head our Strategic Risk Management (or "SRM") department and also chair its risk committee.  I am presenting testimony today on behalf of CSFB and on behalf of our trade group, the Financial Services Roundtable.2  CSFB employs approximately 20,000 people, primarily in the United States, and is a major participant in the capital markets.  It ranks among the top firms in raising money for companies around the world and is a leading underwriter of mortgage and credit card financing.  The firm is also among the largest managers of funds invested in private companies. 
My department is responsible for assessing the risk profile of CSFB on a global basis and for recommending corrective action where appropriate to protect our capital.  This objective is similar to many of the goals of bank supervisors, including the drafters of the proposed Basel Accord - to deter large losses and protect bank solvency. 
The Basel II Capital proposals have been the topic of intense discussion and debate in the financial and regulatory community for the past several years.  The industry supports the objectives of the Basel process:  to better align regulatory capital to underlying economic risks, promote better risk management and foster international consistency in regulatory standards.  The proposed Accord is not a minor refinement to the bank regulatory process, but is, instead, a wholesale reform of bank regulation - a regime that covers roughly $2 trillion of capital and is a key economic engine for most developed markets.  The impacts of these seemingly technical discussions will affect banks, the markets and the economy in a deep way, and we would be wise to consider the effects carefully before implementation. 
Before I start, I would like to note that I have personally developed tremendous respect for the diligence and stamina of the regulators who have worked on Basel II.  They have had to address a great many complex and challenging issues, and have been tenacious in trying to develop a "best practice" solution for each.  Balancing all of this and applying it to very different financial markets around the world - with political sensitivities in each - does not make this an easy job.  I wish to express appreciation for the efforts of Federal Reserve Board Vice Chairman Roger Ferguson, who has met with CSFB and Roundtable member companies several times in the past few weeks to listen to our concerns on the proposed Accord.  Comptroller Hawke and FDIC Chairman Powell have also had open doors for discussion throughout the long process of developing the new Accord.  We look forward to continuing this dialogue as Basel II moves closer toward formal adoption and throughout the implementation period.
CSFB and the Roundtable have worked hard to be constructive commentators on the new rules, particularly in respect to practical implementation issues.  The recent revision of the proposals - called CP3 - included significant improvements, and demonstrated a willingness by regulators to address specific issues raised by industry and academic critics.  Just last week, the Federal Reserve announced that, in implementing Basel II in the U.S., the regulators propose to reduce the capital charges on many types of commercial real estate loans, in response to comments and new data from the banking industry.  We support the direction in which the Accord has been moving recently, and appreciate the regulators' willingness to reexamine earlier conclusions and consider further changes.
However, in spite of the hard work of the Basel Committee and industry, we believe substantial areas for improvement still remain.  Basel II has considerable momentum, and most people in the industry believe it will likely be implemented in the relatively near future.  On balance, we believe that the advantages of the reform now outweigh the drawbacks, although that balance remains close, and in several areas, open issues remain.  This is a frustrating outcome for an initiative with so much potential.  We hope these hearings will help illuminate some of the important remaining issues that need to be addressed, so that the Basel II reforms can live up to their original, very worthy goals.
Today, without getting too involved in the technical details of the Accord, I would like to highlight four "macro" issues which we believe are particularly important: 
1. The current Basel proposal is unnecessarily complex and costly, and suffers from an excessive reliance on detailed, prescriptive rules.  Under the rubric of comparability, these international rules could bring a more formulaic, inflexible style of regulation to the U.S., which currently enjoys a much better balance between black-letter rules and supervisory consultations.

2. The new Accord and its sensitivity to credit ratings could reduce liquidity in the credit markets during economic downturns, potentially extending or deepening economic recessions ("pro-cyclicality").

3. The operational risk capital charge proposed by the Basel Committee remains highly controversial.  Some Roundtable members support the proposed Pillar I operational risk charge; others believe operational risk should be addressed through Pillar II supervisory reviews instead.

4. The disclosures required under Pillar III of the new Accord are likely to add perhaps 20 pages of highly technical data to bank reporting requirements, raising costs and adding little information of value to the reader.  While we appreciate that the Pillar III disclosure requirements have been reduced, they continue to be burdensome and potentially confusing.

Prescriptiveness, Cost, and Adaptability
The first topic I would like to address is the overall cost and prescriptive tone of the new capital rules, and the effect this will have on whether the rules remain relevant over time.  The new rules shift the regulatory regime toward a highly complex, formula-based system, and will diminish the important role that is currently played by human judgment.  Implementation of these rules will be high cost, but not highly cost effective.  Moreover, we believe the very complexity of the new rules and the delicate political balance represented in them will make it challenging to update the rules over time.
Most of this prescriptiveness is to be found in Pillar I, which describes the "recipe" for calculating capital requirements.  The most recent draft of the Pillar I calculations ran to nearly 200 pages, roughly 5 times the length of the original Basel Accord (not including technical papers and additional guidance that is expected to be issued).  This is a common result from this kind of process.  Once you start developing a system that attempts to capture the complexity of the real world in a series of mathematical rules, it is very hard to stop halfway.  One issue or another will always be of major concern for some institution or country.  Many of the Pillar I rules reflect a political compromise as much as the results of a scientific approach to risk management.  The result is a very elaborate system that tries to address all circumstances by being ever more complex, and currently staggers under its own weight.  The Basel Committee has done a commendable job in streamlining the earlier drafts in CP3 - the earlier drafts of Pillar I rules were even longer - but this remains a fundamental issue.
Perhaps the underlying issue in this respect is the prescriptive nature of the new Accord.  Conceptually, the Committee has attempted to capture current industry best practices and boil them down into  fixed formulae, adding burdensome qualification, testing, and reporting requirements.3  These new regulatory requirements, while well-intentioned, will be unduly burdensome and inconsistent with changing market reality and evolving best practice.4  It is our recommendation that the Committee establish some basic requirements largely around the key input parameters and exposure calculations and publish best practices that provide guidance to banks and supervisors rather than a rigid rulebook.
CSFB and other Roundtable members are also concerned about the cumulative effect of numerous conservative choices and assumptions that are built into this complex fabric.  Each of these can be debated separately, and many are extremely technical.  But the combined effect of each of these individual items adds ups to regulatory capital requirements that can depart significantly from the true economic capital needs that Basel II was aiming to emulate.5 
Home/Host Country Issues:  The complexity of the new rules poses particular challenges for an international bank that is regulated by supervisors in multiple countries.  CSFB, for example, will be required to implement Basel II as both a Swiss bank and a U.S. financial holding company.  Our implementation will be governed primarily by the Swiss Federal Banking Commission, in conjunction with the Federal Reserve in the U.S. and the Financial Services Authority in the U.K., and also by other regulators around the world.
Most international banks face a similar set of interlocking regulation in which both home and host countries interpret and enforce rules.  This can give rise to conflicts, even under an international standard like the Basel Accord.  At times, we have been given conflicting requirements by home and host regulators under Basel I, making compliance an impossible "Catch-22".  While we have been able to resolve these issues to date, the potential tension between "home and host" regulators will become a bigger issue given the much wider and more detailed Basel II regime.  If each country decides to require its own local rules and local data for each of the many calculations required under Basel II, the compliance burden will go from bad to worse.  The Basel Committee has formed an Accord Implementation Group to deal with cross-border implementation issues, but experience shows that some differences between multiple supervisors are inevitable.
We are pleased to note that, in a speech last week, Vice Chairman Ferguson indicated that the U.S. banking regulators expect to accept the Basel II approaches and calculations followed by a bank's home country supervisors, when evaluating an international bank with U.S. branches and for purposes of eligibility of Gramm-Leach-Bliley Act financial holding company status.  This is reassuring to hear.  We hope that other host countries adopt similar policies that defer to home country regulators, and that similar issues related to subsidiary banks also are addressed.  We believe that stronger proposals should be developed to resolve home/host country conflicts in a timely and more predictable manner.
Securitization:  A germane example of Basel II's complexity and excess prescriptiveness is its proposal for asset securitization.  Asset securitizations are a cornerstone of how the U.S. markets finance residential mortgages, consumer credit card balances, automobile loans and other receivables.  The draft rules here are daunting, potentially quite burdensome, and often difficult to interpret.  The result is that only a few experts in each area are likely to understand this and other specialized rules of the Accord.  Yet, the interpretation of these experts on some technical points can have enormous impact on the capital calculation.
These rules are written to deter possible arbitrages in the new rules, but risk throwing the "baby out with the bath water."  The industry and regulatory communities generally agree on the objective that capital should be similar before and after securitization, since the total economic risk is unchanged.  However, apportioning the risks properly among the different securities poses a difficult challenge for any set of static rules.  The Basel Committee's current proposal under CP3 takes a conservative approach to this problem, focusing on avoiding improper capital arbitrage by building a technically complex system with a "belts and suspenders" philosophy.  Unfortunately, this approach can also interfere with legitimate transactions and could undermine a widely accepted risk management tool used by many U.S. institutions.
Several problems remain that should be reviewed by the regulators.  First, the mere act of securitization and distribution will tend to increase the capital charge assigned to the same pool of assets.6  This increased capital is an important issue for the U.S. markets in particular, as foreign markets are much less reliant on securitization technology.  This could raise costs for funding U.S. consumer loans and other asset classes where securitization techniques are important.  Foreign regulators have much less at stake in their local markets.
Second, the calculations are subject to difficult interpretations, which can give rise to "cliff edge" uncertainties, where capital charges can change by a factor of ten or more depending on whether a particular instrument can be fit into a specific regulatory box.  For example, a credit line provided to support a credit card or receivables facility might attract a risk weighting of 100% if the bank can satisfy a number of technical tests about the structure of the credit facility.7  However, this charge can skyrocket to 1250% (i.e., an outright deduction from capital) if a bank cannot meet one of these compliance requirements.  This is a conservative approach8 that will certainly help deter arbitrage, but it may also deter good finance.  It also will tend to restrict the evolution of new markets and new securities, since these future instruments might not fit easily into today's compartments.  As with other areas of the Accord, we believe that moving to a more principles-based system that leaves more discretion to banks - subject to thorough supervisory oversight - will provide a more durable and flexible solution for the long term.  It will be important to incorporate these changes into both the final text and in the practical implementation of the rules.9
Cost:  The monetary cost of complying with the Basel II rules will be significant.  For Credit Suisse Group, our holding company, we estimate that our initial costs will be $70mm to $100mm just to implement the system, plus substantial ongoing costs.  Multiply that by thousands of banks globally and this will amount to many billions of dollars of additional costs.  Some of these costs will be passed on to consumers and corporations, and some of these costs may force banks to exit certain activities leaving these markets to unregulated entities.
A major driver of the cost / benefit ratio of the new rules will depend on how they are applied.  For example, there are more than 50 specific requirements that must each be met to use the so-called IRB advanced credit system.  If each of them is interpreted and tested to rigorous audit standards, there will be enormous costs in compliance though the relevance to better risk management will be small.  I would note that implementation costs will be substantial for regulators as well as for the banking community.
Even more important, perhaps, than the direct monetary costs, are the indirect costs.  These will depend on whether the new rules support the real risk management needs of the business, or whether they become an extra bureaucratic burden or even a diversion.  CSFB's internal assessment is that most of the additional resources required will not be in the risk control departments.  Instead, most of these new resources will be needed in the areas of financial reporting and IT support systems, in order to generate the volume of data and reports that Basel II requires to a reliable, audit quality standard.  While further systems development provide some important benefits, this result suggests that the gains in risk management quality from the new proposal are likely to be relatively modest 
Adaptability:  The proposed Basel rules are based on the financial markets as they work today, but are so complex and heavily negotiated that they will be difficult to update over time.  Indeed, some commentators have suggested that the Accord will be outdated by the time of implementation.10
The draft Accord also requires banks to use the Basel II processes in their internal management in many areas, regardless of whether they remain relevant for business practices.  If bank management is required to compute and manage by the Basel II rules anyway, further improvements in internal practice could be seen as both costly and irrelevant.  As a result, the Basel Accord could actually slow the progress of better private sector risk management techniques.
Proposal:  Our suggested response to the problems of prescriptiveness and high cost is for the Basel Committee to place a much greater emphasis on the principles-based approach that underlies the "Pillar II" section of the proposed Accord.11  Whereas Pillar I sets out regulatory capital calculations in a detailed, prescriptive way, the approach of Pillar II is to force firms to develop their own internal models, based on evolving best-practice, and then to scrutinize the results through the examination process and regulatory guidance.  This "principles-based" approach, subject to some reasonable benchmarks and guidelines to maintain consistency, has some important natural advantages compared to the complex "black-letter" style rules currently prescribed by regulators under Pillar I.  Pillar II encourages banks and regulators to work together over time to improve risk management practice, rather than forcing compliance with a potentially dated rulebook.  That approach permits steady, evolutionary improvement and should therefore be more durable and relevant than Pillar I rules that are designed with today's markets in mind. 
Addressing this issue will not be simple in the short time left before the rules are finalized.  If these rules are all applied as black letter law and interpreted strictly, the new rules will be both costly and - since the risk management advances that lead in part to Basel II will not end in 2003 - potentially irrelevant to ongoing best practice.  We encourage an approach that emphasizes principles and simplicity as the rules are finalized, and a less onerous "trust but verify" approach to compliance.  Specifically, we would support adding statements to the Accord to emphasize that compliance with the rules will be based not on "box checking" but with the spirit of the rules, based on economic content. 
Impact on Competition:  We believe that the cost and complex rules of Pillar I will have significant impacts on competition, and could tilt the current playing field significantly in various markets.  This will be particularly important in the U.S., where non-bank competitors like investment banks, finance companies and insurance companies represent a large part of the financial system.  The Basel rules do not apply to them.  If the costs of Basel II are high, banks will earn a lower return on capital, will grow more slowly and may lose market share.  There may even be some incentives to exit businesses or to de-bank altogether.  We believe that the Basle Committee needs to do significantly more work in assessing the competitive impact of the rules across the financial marketplace.

The new rules will change how banks calculate and manage their capital and the amount of business they choose to do.  If banks all act in concert - as they will tend to do under a common regulatory regime - this can significantly increase or decrease liquidity in the credit markets and ultimately affect the real economy.  We have analyzed this effect over the last 20 years of credit cycles.  Our calculations suggest that the impact on required bank capital will be substantial.  In particular, the new Basel II calculations could require much more bank capital during economic recessions than the current system.  The process by which these rules could widen economic swings is called "pro-cyclicality." 12  This is, in effect, an implicit change in macro-economic policy and it would be wise to consider that carefully. 
As a practical example, consider the credit environment of the last two years.  We have seen a huge number of credit rating downgrades, which have increased the real risk of bank portfolios.  The current system is relatively indifferent to this change in terms of required regulatory capital, but the proposed system will require significantly more capital when companies are downgraded.  Banks will have to choose between raising more capital during recessions or reducing the amount of lending that they do.
Some regulators have suggested that the fear of a capital shortfall will change banks' risk assessment and lending behavior so that this issue will disappear.  Implicitly, they suggest that bank's risk assessment will improve so much that mistakes will be a thing of the past.  While it would be wonderful if banks could always foresee the future, I don't think that's realistic.  Bank management already makes risk assessment a top priority - it is perhaps the core judgment that determines whether a bank thrives or fails.  Unfortunately, economies are likely to remain cyclical and predictions about the future will inevitably turn out to include their share of mistakes.
Cutting lending during a downturn is probably smart, if your perspective is focused solely on bank solvency.  However, it raises significant issues for the wider economy.  My personal estimate is that my bank would have cut back its lending by perhaps an additional 20% to 30% if the Basel II rules were in place during 2002.  If all banks cut back at the same time, the potential adverse impact on the real economy could lengthen and deepen the recession.  We are currently working through an economic slowdown; it is difficult to think that adding pressure on bank capital during this period would be helpful to economic recovery.  In fact, it defeats part of the reason for regulating banks in the first place - in order to have a stable supply of capital to support the underlying economy.  We need to be particularly careful here because the new system is imposed across the whole banking system and everyone will have to operate at the same time on the same rules.  Herd behavior can make smaller problems into bigger ones.
The regulatory community has acknowledged this as a potentially serious issue, but we believe that further attention is warranted, because the consequences of getting this wrong are potentially quite important to the broader economy.  When the first quantitative proposals in January 2001 revealed a significant potential problem, the regulators did react with a revised and somewhat "flatter" risk-weight curve.13  However, while this reduces the scale of the issue somewhat, it does not grasp the nettle.
The current Pillar II proposals include a credit risk "stress test" which is directly linked to possible additional capital requirements.14  The exact design of this test remains unclear but the language suggests it amounts to an extra layer of buffer capital so that banks will not need to dig into their core capital in tough times.  In effect, this is like creating a second fire department, because you want to always keep the first fire department in reserve.  Creating two fire departments or requiring two pools of capital is unnecessarily expensive and doesn't seem to address the fundamental issue.  That issue is that a risk sensitive system will inevitably lead to varying capital requirements through time, and that is a result that will require explicit management and thoughtful preparation.  As with other areas of the Basel Accord, adding some flexibility to the rules is the simplest and most practical way of preventing these inevitable stresses from building up into major crises.
At a minimum, we suggest that the Basel Committee add to the proposed Accord an explicit acknowledgment that capital levels may fluctuate, and that Pillar II reviews and stress tests not become one-way ratchets that only increase regulatory capital requirements.  If a stress test is to work properly, then when tough times arrive banks should be permitted to live within their plans, and regulators should resist the temptation to continue to require the same untouched capital cushion.  Otherwise, Basel II's stress test will not in fact reduce pro-cyclicality, but will simply amount to an unpublished higher minimum capital standard.

Operational Risk
In addition to reforming capital charges for credit risk, Basel II establishes a new capital charge for operational risk - the risk of breakdowns in systems and people.  This is the most controversial element of the proposed Accord. 
Financial Services Roundtable Comment:  It is important to distinguish between the concepts of managing operational risk and imposing a separate, quantitative capital requirement for it.  All of the Roundtable's member companies agree that evaluating and controlling operational risk is important and should be required as a supervisory and business matter.  Roundtable members do not agree on whether or how operational risk should be reflected in regulatory capital calculations.  Many companies believe operational risk can best be addressed through case-by-case supervisory reviews under Pillar II; others favor a quantitative and a publicly disclosed capital charge under Pillar I.
In several forums, the Roundtable itself has opposed a separate capital charge for operational risk and has argued for handling the issue through supervisory reviews under Pillar II, much as interest rate and liquidity risk are handled.  The Roundtable's senior management has expressed its concerns directly with Federal Reserve Board Chairman Greenspan and Vice Chairman Ferguson.  Many Roundtable member companies strongly oppose any Pillar I operational risk capital charge.  However, several Roundtable member companies just as firmly support Basel II's proposed Pillar I approach, following the development of the Accord's "Advanced Measurement Approach" (AMA), which gives banks flexibility to use their own internal methods for determining the regulatory capital needed for operational risk.  Institutions that support a Pillar I operational risk charge believe it would improve transparency and comparability and bring regulatory capital requirements into closer alignment with the "economic capital" determinations used in these banks' internal management decisions.  These institutions contend that any approach other than an explicit Pillar I charge for operational risk would impede progress toward a level playing field, by affecting the process of calibrating regulatory capital minimums.  That is, these members believe that if an operational risk charge were not included in Pillar I, the resulting capital charges on credit risk and market risk would remain higher to compensate, making it more difficult for international banks to compete with institutions that are not covered by the new Accord.
The Roundtable continues to have concerns about the proposed operational risk capital charge, as well as several technical questions about its implementation.  One problem that all of our members agree upon is that the proposed Accord fails to give enough recognition to the benefits of insurance in mitigating operational risk.
CSFB Comment:  CSFB is concerned about the attempt to base an operational risk capital charge on new, unproven models, and believes this approach is problematic and possibly even counter-productive.  We agree that operational risk is a critical risk to manage, and we set aside significant capital to cover potential surprises in our internal capital allocation process.  However, we do not believe that operational risk can be modeled in the quantitative way proposed under the Basel II rules.  Many efforts to measure operational risk have been proposed, often focusing on limited areas (e.g., operations processing losses) that happen to be susceptible to statistical techniques.  But these methods are not generally relevant to major risks, such as fraud, a changing legal environment or a major disaster, which are the risks that require capital.  Operational risk capital is primarily to insure against the risk of being fundamentally surprised by a major event, but it is difficult to predict and measure what you don't expect.
Basel II and other regulatory initiatives will push banks to devote significant resources toward operational risk systems and loss databases, but I personally feel that these resources could be better utilized elsewhere.  Basel II's Advanced Measurement Approach to operational risk requires banks to attempt to verify their models statistically.  Many are working hard on this, but we have yet to see any model that has actually been verified in a robust way.  In fact, by emphasizing quantitative numbers for operational risk, we may be creating a real danger - creating a false sense of security that we have measured operational risk and hence controlled it.  I am a model-oriented, technical person by training, but I do not want to rely on a model that is built on speculative assumptions. 
It is encouraging that the Basel Committee has sent signals suggesting an increased degree of flexibility in operational risk calculations.  I am hopeful that this will bear out through the implementation.  But we will still have a long way to go, and I am concerned that there will be a tendency to revert to prescriptive and unscientific requirements as regulators develop specific rules for approving operational risk models.

Pillar III - Disclosure Rules
One of the strengths of the Basel II proposals is that they look beyond just calculating and maintaining capital levels.  In designing Basel II, regulators realized that capital requirements - the so-called "Pillar I" - could never ensure the safety and soundness of the banking system alone.  They understood that ultimately it is more important to encourage constructive relationships between financial institutions, their supervisors and the market to produce good risk management.  This reasoning, which has the strong support of the banking industry, has lead to the creation of the two qualitative Pillars of the Basel Accord.  Pillar II deals with the supervisory review process and, in particular, regulatory oversight of banks' internal economic risk assessments.  Pillar III seeks to enhance market discipline through increased public disclosure requirements.
The concepts behind the proposed rules for Pillar II and III are well accepted by the industry and regulators alike.  However, many of the detailed proposals in the Pillar III market disclosures section are cause for concern in the industry.  Unfortunately, the development of Pillar III is an area where consultation between the industry and the regulators came late in the process.  Although CP3 has improved the situation somewhat, we believe the proposals still are overly prescriptive, burdensome and subject to misinterpretation.  The Pillar III requirements also reflect a somewhat narrow view of risk, focusing exclusively on a specific regulatory view of risk capital.
We currently publish approximately 20 pages of risk information in our annual report, and we support transparency and disclosure as very worthwhile goals.  The Pillar III proposals would add a large mass of additional disclosure which is highly technical in nature and which we believe will be of little benefit to the reader.  Indeed, few people are able to digest all of the information that is already presented on risks, but now this information could be lost in a deeper, more technical pile of data.  The additional requirements proposed under Pillar III are more likely to confuse than illuminate. 
As Chairman Greenspan has recently remarked, transparency is not the same as disclosure:  "Transparency challenges market participants not only to provide information, but also to place that information in a context that makes it meaningful."15  In this, we believe the prescriptive, volume oriented focus of Pillar III falls short.
Of particular concern are the numerous required disclosures that relate directly to the capital calculations performed within Pillar I.  Instead of disclosing measures of risk used in internal risk management systems, these disclosures mandate an explicit regulatory capital view of risk.  In the most complex areas, such as asset securitization, these disclosures will surely be mystifying to all but the most expert audiences.
Moreover, given the likely longevity of the Basel II Accord (the current Accord is in its 14th year), there is a need to ensure risk management practice is able to mature beyond the concepts now embedded in the Basel II proposals.  Just as the market has moved beyond the current accord, there will inevitably come a time when some Pillar I calculations are no longer regarded as good measures of risk for all products.  In that case, it must be possible for banks to alter disclosures to represent emerging best practices.  Under Pillar III as currently proposed, banks will likely find themselves constrained to disclosing risks under a system that is no longer wholly relevant.
In designing the details of Pillar III, the Basel Committee has placed too much emphasis on quantity, rather than quality, of disclosure.  It is emphasizing consistency by prescription instead of consensus.  In contrast, the demands of the market have produced broadly comparable and largely voluntary disclosures of market risk by banks.  This is an example of how Pillar III should work.  It would be more effective if Pillar III established a general set of principles, and then allowed the discipline of the market to produce continuous improvement in risk disclosure.  This would produce information that the market actually desires, rather than seeking to impose today's ideas on future market participants by fiat.
We are at an important crossroads in the reform effort.  A lot of good hard work on designing the framework and gaining political consensus has been accomplished.  We have a high regard for the efforts of the Basel Committee and the regulators who have worked so hard to capture the best current practices in risk assessment.  CSFB and the Roundtable have tried to contribute to the specifics of those discussions in a constructive manner.  We believe that the current proposal should be streamlined significantly, reducing the level of prescriptiveness and cost, so that the advantages of this project are not tarnished by its current shortcomings.  . 
Simplifying the massive weight of detailed rules in Pillar I will require continued discipline in the final round of drafting  It will also require a new emphasis on the "spirit" of the rules, both as the rules are finalized and when they move to the implementation phase with national regulators.  If, instead,  these rules are written and interpreted as black-letter regulations, set at a highly technical audit standard, the cost of overall implementation will be high.  Such an approach would mean the calculations could also become increasingly outdated and less relevant to risk management best practice over time.  We can hope that all national regulators will avoid this pitfall, but international banks will tend to be driven by the standards set by the strictest and most literal of their major regulators. 
Much hard work has been put into Basel II, but much also remains ahead.  The timetable for implementation is challenging, particularly since the Accord's requires a minimum of three years of data for the advanced calculations - meaning that banks will need to revise systems to begin collecting the new information by early next year.  In the pressure to finalize and implement the Accord, we hope that enough time will be provided for everyone - banks and supervisors alike - to digest and think about the implications of the new regime, and to develop appropriate transition rules.
As a final comment, I believe that much more can be accomplished by increasing the emphasis on the concepts of Pillar II and Pillar III, and a focus on the principles of evolving best practice rather than fixed formulae.  This approach would not only help address "prescriptiveness, cost and adaptability", but could also help address the issues of operational risk and pro-cyclicality.  Pillars II and III have real people on the other side - regulators and the market.  Human judgment can adapt to changes and new markets more easily than a rulebook can.  This approach, properly applied, also puts the burden back where it should be - on the shoulders of bank management to demonstrate to the regulators and the public that they are doing a good job.  That is in the spirit of the Sarbanes-Oxley reforms, and I think it is a smart, durable way to improve discipline and maintain best practice standards. 
Lastly, it should also make the new system more responsive to change and therefore more relevant over time.  Without adjustments to make Basel II more flexible and to allow it to evolve over time, I am afraid we might have to start work on a Basel III before the ink is dry on the current effort.
Thank you.

1   Credit Suisse First Boston (CSFB) is a U.S. financial holding company and leading global investment bank serving institutional, corporate, government and high net worth clients.   CSFB's businesses include securities underwriting, sales and trading, investment banking, private equity, financial advisory services, investment research, venture capital and asset management.  CSFB operates in more than 89 locations across more than 37 countries on six continents.  The Firm is a business unit of Zurich-based Credit Suisse Group, a leading global financial services company.
2   The Financial Services Roundtable is a national association representing 100 of the largest integrated financial services companies in the U.S. providing banking, insurance, securities, and investment products and services to American consumers.
3   One editorial recently described this approach as "prescribing and proscribing in equal measure . . .a monster that can't clear the first hurdle:  flexibility."  Risk Magazine, editorial page, June 2003 edition.
4   For example, the eligibility requirements for institutions to qualify to use the Accord's advanced IRB methods for credit risk capital charges are too detailed and burdensome.  In general, we believe that these eligibility requirements should be scaled back and replaced with more general guidance. 

A specific example is the testing requirements for credit exposure in repurchase agreements, an area with historically very low losses.  To its credit, the Basel Committee permits the use of internal market risk models to estimate potential collateral shortfalls under stress, which is in line with modern practice.  However, the Committee requires substantial additional testing to use this technique, even though  this calculation is based on the same model that governs overall market risk, which is a much bigger risk and already subject to comprehensive regulatory oversight.
5   To mention a few examples:

(i)  The Accord significantly overstates the credit risk capital charges for exposures hedged by guarantees and credit derivatives, by failing to recognize the much lower risk of joint defaults by debtors and guarantors and by applying overly conservative rules on maturity mismatches.

(ii)  The proposed Accord requires capital against Expected Losses, even though these losses are already covered by loan loss reserves, and Future Margin Income is generally recognized only for credit card exposures.
6   For example, the originating bank is charged the full risk of the pool if it retains a sufficiently large position in the junior securities.  A second bank that purchases the senior securities also will be charged significant capital, meaning that the capital required of the banking system will be higher than if the assets had simply been held on an institution's balance sheet directly.
7   In particular, questions remain regarding the proposed treatment of liquidity facilities for asset-backed commercial paper programs, which would face capital charges that seem disproportionately high relative to the level of risk.
8   Some also have argued that the risk weights on such securitized assets are too high as a more general matter.  Similarly rated corporate loans often attract a much lower capital charge.
9   For example, banks that qualify for the Advanced IRB approach should be allowed to use internal ratings to determine risk weights, which is not allowed for securitizations under CP3.  Ratings based on rating agency methodologies or reasonably equivalent approaches, for example, should provide supervisors sufficient comfort that a market test has been met.  Liquidity facilities and credit enhancements for asset-backed commercial paper conduits are prime examples where this approach could be easily adopted.
10   See Risk Magazine, footnote 3 above.
11   Our Pillar II comments here are strictly focused on the credit risk capital charges.  As noted later in this testimony, Roundtable members have differing views on whether any operational risk charge should be addressed under Pillar I or Pillar II.
12   CSFB has taken particular interest in this issue among Roundtable members.  See American Banker, "Basel Capital Accord Must Leave Some Room for Human Judgment" by Wilson Ervin and Joseph Seidel, August 30, 2002, and Risk Magazine, "Pro-cyclicality in the New Basel Accord" by Wilson Ervin and Tom Wilde, October 2001.
13   Basel Committee on Banking Supervision, Working Paper "Potential Modifications to the Committee's Proposals", Bank for International Settlements, November 2001.
14   Basel Committee on Banking Supervision, CP3,  April 2003, Paragraph 397-399 and 724.
15   Remarks by Federal Reserve Board Chairman Alan Greenspan, Corporate Governance, at the 2003 Conference on Bank Structure and Competition, May 8, 2003.