Hearing: A Review of the New Basel Capital Accord
Wednesday, June 18, 2003
It is an honor to appear today before this Committee to discuss the potential ramifications of the international risk-based capital rules under consideration in Basel for U.S. financial institutions and - even more important - for the economy that depends upon them. I am the managing partner of Federal Financial Analytics, a consulting firm that advises on U.S. legislative, regulatory and policy issues affecting strategic planning. In this capacity, we advise a variety of companies on the implications of specific sections of the Basel proposal. We also advise the Financial Guardian Group, which represents those U.S. banks most concerned with the proposed operational risk-based capital charge.
In my testimony I will focus on the most recent version of the Basel rules - the third consultative paper or CP3, as well as on the advance notice of proposed rulemaking (ANPR) on which the U.S. regulators are now working. Although the effective date for the new version of the international capital rules - Basel II - is December 31, 2006, its actual impact will be felt far more quickly. Indeed some markets have already begun to change in anticipation of the Basel standards. As the final shape of the rules becomes more clear, financial markets - and the larger economy - will change more noticeably. Congress' review of the rules thus comes in a timely fashion that ensures any policy concerns posed by the rules can be addressed well in advance.
Much in CP3 is very worthwhile. Overall, Basel II is a worthy and overdue effort to fix the problems in Basel I that have created all too many opportunities for banks to "arbitrage" the capital rules. When regulatory capital diverges from the "economic" capital dictated by the markets, banks change their portfolio, pricing and risk decisions. This has profound impact on overall franchise value and on key lines of business, as well as affecting the cost and availability of credit to consumers and companies across the country and around the world.
When regulatory capital is too low, banks can take undue risk - a problem observed since Basel I went into effect that needs a quick remedy. When regulatory capital is too high, banks cannot compete against non-banks, and assets flee the banking system with possible adverse consequences for overall market stability.
However, Basel II now has gone far from its initial clear goal of ending regulatory arbitrage. In fact, the most recent statement of the purposes of Basel II - posed two weeks ago by U.S. regulators - no longer even mentions this. Now, the goals of Basel II are said to be: improvements to internal risk management and capital allocation, enhanced market discipline and - of all things - a new capital charge for operational risk. I shall have more to say about the operational risk capital charge later, but suffice it to say that this proposal worsens the relationship between regulatory capital and risk - absolutely the reverse of where Basel II initially intended to go. If Basel II can't be brought back to its initial and important purpose, then the U.S. capital rules alone should change to do so.
Based on our review of the third consultative paper and recent statements from U.S. regulators about our own implementing rules:
* There is a "first things first" solution that fixes the Basel II's complexity problem. Much in the proposal can be quickly implemented at reasonable cost for all banks and savings associations here and abroad. U.S. regulators should act now on those sections of Basel II on which they can agree unanimously, and defer other sections until they can do so.
* A "one size fits all" approach won't work in the U.S. Capital should go up or down with risk, not be squeezed into the current requirements that were originally set with scant regard for actual credit losses. Unique factors in the U.S. market make it especially important that bank capital appropriately reflect risk.
* The operational risk-based capital section of Basel II remains deeply flawed and should be dropped. Regulatory capital for operational risk will increase risk, not reduce it, and strong supervision with enforced standards is the right way to address operational risk.
* Simple capital rules are essential for effective supervision. Agencies here and elsewhere cannot administer over-sophisticated rules. Further, laboring to do so will divert resources from emerging risks that often prove the undoing of individual institutions or serious risks to the financial system as a whole. Capital is not the only driver of safety and soundness. Banks have collapsed in the past and will fail in the future even as they hold more than the minimum amount of regulatory capital.
Economists and financial analysts have spent literally thousands of hours working to revise the risk-based capital standards that govern internationally-active banks around the world and all insured depositories in the United States. This effort is an important one, as problems in Basel I have led to undue risk-taking and other concerns that warrant immediate attention. However, in the five years in which Basel II has been crafted, more and more attention has been devoted to the increasingly complex models that attempt to anticipate expected and unexpected losses in every line of business, under every scenario in each country for all time. The defense of this effort is that financial markets are now complex, so capital must be too. However, the universe is complex, yet Einstein found a very simple formula that helped to explain it. Complexity is a weakness, not a strength, and Basel II should be difficult only when absolutely necessary to capture subtle risks with potentially severe consequences.
Basel II rightly rests on three pillars: improved regulatory capital standards, better supervision and more disclosure. If Pillars 2 and 3 work well, then Pillar 1 - the capital standards - need not be as formulaic and far-reaching as currently proposed because supervisors will have ample tools to tailor regulatory capital to individual circumstances and markets will know when this isn't being done.
One reason regulators rely so much on regulatory capital is the lack of effective supervision in many major financial markets. Here, though, supervisors have ample authority to discipline banks for problems that have nothing to do with capital standards. Companies must, for example, be "well managed," as well as "well capitalized" to be financial holding companies and enjoy the privileges provided in Gramm-Leach-Bliley. Further, supervisors measure banks on a "CAMELS" scale in which capital - the C - is just one of a range of factors - all weighted equally - on which critical enforcement actions hinge. The other factors are asset quality, management, earnings, liquidity and sensitivity (to various risk factors). If non-U.S. regulators adopted a similarly wide-reaching supervisory regime - and backed it up with meaningful sanctions such as those deployed here - then much of the complexity in Basel II could fade away and the rules could focus on ending major sources of regulatory arbitrage that, on the one hand, threaten safety and soundness and, on the other, unnecessarily undermine bank profitability.
First Things First
Despite the intention of having a balanced international regulatory framework that emphasizes more than just regulatory capital, the vast majority of staff time has been spent on the regulatory capital charges. U.S. regulators, I think, could have done much for the global financial system and avoided many of the pitfalls in Basel II if more attention had been paid to exporting our strict supervisory standards and their effective enforcement. Japan, in particular, would benefit greatly from this - it's a clear case in which nominal adherence to regulatory capital has done nothing to prevent a grave banking crisis with serous macroeconomic impact.
As Basel II advanced and the capital models grew ever more complex, U.S. regulators rightly became increasingly concerned about how this would work in our unique banking system. In sharp contrast to Japan and the European Union, we here have thousands of banks and savings associations; foreign banking systems are far more concentrated into a few nationwide banks. Regulators also - rightly - became concerned about several pieces of the simpler sections of Basel II that resulted from complex multilateral negotiations in which the end goal was often obscured. This is particularly true with the more simple versions of the operational risk-based capital standards, which are not only flawed, but could also actually increase - not reduce - banking risk. Further, even the relatively simple sections of Basel II grew ever more complex as negotiators sought to solve each problem and individual national political objectives as the rules worked their way along over the years.
Based on these fears - some of them quite right - U.S. regulators have come up with a solution - mostly wrong. They now plan to impose only the most complex versions of Basel II and then to do so only for the nation's largest banks. This may limit the pain, but it also undermines the gains close at hand in Basel II. Where Basel II drops regulatory capital - which it does dramatically in traditional lines of business like mortgages and small-business lending - banks left out of Basel II, which will still be required to comply with Basel I, will be at a serious competitive disadvantage to big ones in it. Where the models are overly-complex or - worse - wrong, the fact that only big banks must comply with them does nothing to redress the adverse impact they might have.
Further, writing off the most flawed sections of Basel II in the U.S. does nothing to address potential serious consequences in the global economy. U.S. banks - especially large ones - compete head-on with non-U.S. banks here and abroad. If differences in the simpler parts of Basel II - called the standardized approaches - give non-U.S. banks an excuse to rely on over-lax rules and inadequate enforcement, then the major strength U.S. banks now have in the international financial services market will be undermined. Worse still, major financial services firms could operate under capital rules that don't actually address real risk.
For all its flaws, much in the standardized proposal for credit risk reflects broad agreement on improvements to Basel I. U.S. regulators should turn this into clear language and propose it for smaller banks, while refining the advanced models and offering them to large ones. Where no agreement is in sight - on asset securitization, for example - regulators should act now on those areas where broad consensus exists and defer the others until it emerges or regulators are sure - absolutely sure - they're right and the industry is wrong. U.S. regulators now diverge on many key areas of Basel II, and they should act in unison on issues where they intend to contradict the best evidence and advocacy the industry can muster.
One Size Won't Fit All
As U.S. regulators have turned their attention from the international negotiations to implementation of Basel II at home, a major dispute has arisen over whether to follow the new rules where they lead. Under Basel II, capital could go down below current levels, especially for very large banks with major retail or mortgage operations. It is for that reason that the operational risk-based capital proposal has been super-imposed on the credit risk reforms Basel II initially sought. It is also the reason why some U.S. regulators are now reasserting the importance of the most primitive of all capital charges - a simple leverage one - on banks and their parent holding companies. "Topping off" the right amount of credit risk capital with the operational charge and a surcharge for "leverage" will so undermine Basel II - especially in light of its high implementation cost - as to raise serious questions about whether the entire exercise is worthwhile.
I shall have more to say about the operational risk charge below. With it, Basel II should not be implemented at all. Without it, a sound regulatory capital scheme is in sight.
The leverage rule is a unique U.S. capital standard, and it's one that should be dropped as Basel II comes into force. Indeed, it's one that should have been dropped years ago. The leverage standard is a simple ratio of capital to on-balance sheet assets calculated without regard to risk. Under the leverage standard, a bank holds the same amount of capital if its book of business is solid gold or unsecured credit-card loans to dubious borrowers. It's a capital standard that couldn't be more crude, but U.S. regulators clung to it in 1988 because they weren't sure they trusted Basel I. They wanted some form of insurance because they knew then - as now - that credit risk rules didn't capture interest-rate risk. You will recall that this latter risk was the predicate cause of the collapse of our savings-and-loans - which cost taxpayers more than $250 billion and kept this Committee extremely busy for over a decade.
Ironically, Basel II still doesn't address interest-rate risk (IRR). Although the regulators think they know enough about operational risk to put it in the Pillar 1 mandated capital standards, they have decided to leave IRR in Pillar 2. In 1988, regulators were right about the problems measuring IRR; now, they're not. Markets price trillions of dollars of IRR each year in a fashion that Fed Chairman Greenspan has rightly praised.
Why then keep the leverage rule? U.S. agencies appear to be clinging to it because they are afraid to follow Basel II's models where they lead. In some cases, the advanced models propose massive drops in regulatory capital. This is particularly true in mortgages and small-business loans - key lines of business for smaller banks that will face major competitive problems if big banks get to drop regulatory capital under Basel II while they are kept in the cold of Basel I. Of course, in other cases, Basel II will dramatically raise capital - for high-risk loans and certain equity holdings, for example. To adopt Basel II when it goes up and block it when it goes down is to create a regulatory capital regime that leaves arbitrage largely in place - again profoundly undermining why all this started in the first place.
The best way to protect the deposit insurance funds from risk and small banks from competitive harm is to introduce Basel II's most advanced model-driven sections in an incremental way that - essentially - hedges the model-builders' bets. How to do this? Despite the complexity of the advanced internal ratings-based approach to credit risk, it can be introduced in a remarkably easy way. Upon conclusion of Basel II's comment period and a review of all the analyses of the sophisticated models, regulators should make up their minds about the "right" amount of credit risk-based capital for specific assets. Where they can't agree, as noted, they should defer action. Where they can, they should implement Basel II - but only in a phased-in fashion. If, for example, the "right" amount of capital is a dramatic drop, then set a schedule in which capital slides down year after year across the board for all banks that qualify to use the advanced models. Where it goes up a lot, capital should similarly be phased in.
This incremental approach has two advantages. First, as noted, it hedges the regulators' bet on the skills of their model builders and the ability of supervisors to handle the complex new rules (on which more below). Second, it addresses concern that Basel II will exacerbate booms and worsen busts - "procyclicality" in Basel speak. To be sure, phasing in Basel over time doesn't eliminate procyclicality, but it ensures that regulators are certain of their capital models when these come into full force, while giving them time also to assess the value of stress testing and other measures now under consideration.
Eliminate the Pillar 1 Operational Risk Capital Charge
Basel II's pending proposal and, we are told, the draft U.S. implementing rules will include a new regulatory capital charge for operational risk. Operational risk is that resulting from human or systems failures, natural disasters and even terrorist attack. There is, though, no accepted definition of operational risk for supervisory purposes - for example, does it include reputation risk? Basel II says no - for now - but this risk has frequently proven the most serious of all in a business fundamentally founded on investor and depositor confidence. What about events like September 11 - catastrophic operational risk? Basel II now has them in - although they were out at the end of last year - but who knows how to measure the likelihood of another attack and then to decide just how much capital is enough and whether capital the right antidote?
It's particularly hard to understand why Basel II has a specific capital charge for this risk when one notes that many of its own documents agree that it cannot be well defined. The Basel Risk Management Group, for example, said its own data need to be used with "caution" and that a specific capital charge cannot now be based on them. A major Basel Committee on global financial safety also concluded earlier this year that there is now no way to determine a quantitative regulatory capital charge.
Another major unanswered question: is any amount of capital enough against catastrophic risk? I don't think so, and indeed imposing an operational risk-based capital (ORBC) requirement will create a serious and perverse incentive for banks to skimp on the forms of operational risk management and mitigation that proved their worth in the most recent and terrible case of catastrophic operational risk, the attack on the World Trade Center. What worked after the terrorist attack - apart from undaunted heroism - were the back-up systems and contingency plans that well-prepared financial firms had put in place. What worked in the terrible days thereafter - again, other than sheer courage and determination - was insurance. In Basel-speak, these are operational risk management and mitigation. Both are costly - indeed, the back-up systems, which U.S. regulators have mandated since 9/11, are very much so. Imposing a simple, arbitrary charge against operational risk will lead many banks to rely on this, not proven ways to protect themselves, their customers and the financial system more generally.
Basel II now includes three variations on a regulatory ORBC requirement. Two of these - the "basic indicator" and "standardized" ones - rely on a simple percentage of gross income to calculate ORBC. This is, quite simply, nonsensical. There is no correlation between income and risk. In fact, operational risk also runs counter to gross income because banks that spend more on risk management and mitigation have less profits. Banks that generally have trouble making money also tend to be riskier - again, an inverse correlation between gross income and operational risk, not the positive, linear one on which Basel II relies.
U.S. regulators have, apparently, realized that the two simple approaches to operational risk in Basel II don't work. As a result, they are planning only to impose the "advanced measurement approach"(AMA) here. This will, though, leave the other two methods in place in the EU and Japan, creating a perverse incentive for big banks there to run undue amounts of operational risk. We can't wall ourselves off from the problems this will create, and U.S. regulators should thus push hard for meaningful supervisory standards for operational risk that bind all financial services firms, not compromise on a deeply flawed regulatory capital model.
Further, the AMA does not solve the fundamental problems with an ORBC charge in Pillar 1. Some of these are unique to U.S. banks - which must compete with major non-banks in lines of business like asset management and payments processing. Basel II in the U.S. will not cover non-banks. Specialized banks will thus face major competitive pressures that may force them to review whether continuing to remain a bank is worthwhile. ORBC not only creates incentives for increased operational risk, but it also may create one for non-bank charters. This would drive assets outside our sound, proven system of bank supervision.
The pending ORBC charge will also put U.S. banks at a competitive disadvantage against EU and Japanese ones because "legal risk" results in a regulatory capital charge. Our legal system is unique -- no other nation has our plaintiffs' bar or our extensive array of laws designed to protect consumers, prevent discrimination and promote workplace safety. There is no evidence that any of this legal risk has ever caused any U.S. bank to fail, and current law already requires reserves for material legal risk (and these must also be disclosed).
The AMA epitomizes the problems in Basel II where reliance is placed on unproven models over which U.S. regulators rightly do not agree. Acceptance of these models now puts banks at undue and unnecessary risk - risk far better addressed through effective supervision with meaningful enforcement.
Can Supervisors Supervise Under Basel II?
Finally, I would like to turn to the question of whether the complexities in Basel II's advanced models are so daunting that supervisors at home and abroad will not be able to ensure that banks actually comply with the new capital rules. This is a major concern, and one the regulators are already trying to address through a major Basel committee focused on supervisory implementation. In the U.S., the agencies now think the best way to handle the complexity problem is to make Basel II apply only to the biggest banks, whose examiners tend to be those most familiar with complex financial arrangements. However, as noted, applying Basel II here only to the biggest banks will create a range of competitive and safety problems, while leaving the supervisory capability question largely unresolved.
A recent survey of the cost of Basel implementation for the larger banks expected to use the advanced models indicates that it will reach $200 million per bank. One has to ask how it can cost so much for banks and not pose a comparable burden on supervisors who must assess these elaborate models. In point of fact, the rules must be as costly for the supervisors as for the supervised or undue reliance will be placed on untested models. If supervisors instead rely on "benchmarks" they will in effect superimpose standardized credit and operational standards that obviate the flexibility hoped for from the advanced approaches.
These problems are not addressed by the proposed qualifying conditions for use of the advanced models - more board and senior management involvement, for example -- because none of the proposed standards addresses the fundamental problem posed by complexity, let alone how top management can divert resources from their many other pressing investor protection and safety-and-soundness responsibilities.
The right solution to the supervisory resource problem is the same as the right solution to the other challenges posed by Basel II: impose a uniform system of improved rules across the board and then change them gradually over time as the rules are tested and we all learn how to work under them. Back up the more sophisticated models with meaningful supervision that binds banks in the EU and Japan, not just U.S. ones and give investors simple, clear disclosures to help them understand just how much capital banks have and whether the supervisors are concerned about it.