Hearing: A Review of the New Basel Capital Accord
Wednesday, June 18, 2003
Thank you, Mr. Chairman. I'm here today on behalf of KeyCorp, the 11th largest banking company in the United States. KeyCorp has total assets of approximately $85 billion, and spans the northern half of the U.S. from Maine to Alaska. While the vast majority of our business is domestically based, we do have a modest level of international business activity.
KeyCorp is not one of the institutions included in the definition of "top ten most internationally active institutions". Accordingly, under the present regulatory guidance, we will not be required to comply with Basel II when it becomes effective in 2006. Nonetheless, it is our intent to qualify as an advanced practice institution. We simply believe that it is good banking practice to develop the risk management tools that are the foundation of Basel II: if that qualifies us as an advanced practice company under the new accord, so much the better.
I believe my testimony today provides a rather unique perspective on the issue of whether or not Basel II is good for the banking industry. For the first 17 years of my professional career I was a bank regulator with the Office of the Comptroller of the Currency (OCC). Much of my time with the OCC was spent dealing with problem and failing institutions. During my last several years with the OCC, I was Deputy Comptroller for Special Supervision. That's a nice way of saying I was responsible for the department that dealt with severely troubled and failing financial institutions.
My tenure in the Special Supervision department ran from 1986 through 1990, a time when a significant number of banks failed in the U.S. I was able to see first hand the myriad of reasons that caused banks to get into trouble. Not the least of these was the inability to appropriately identify and manage their risks.
I left the OCC in 1990 to join the deeply troubled Ameritrust Corp. in Cleveland, OH. Ameritrust was a $12 billion company that had encountered difficulties arising from its loan portfolio. I was part of the new management team focused on turning the company around. Over an 18-month period, Ameritrust lurched from one crisis to another, but we eventually were able to stabilize the company. During the interim period I lived, first hand, through the effects of a firm that had little in the way of risk management practices and tools.
My experience with OCC's failing banks division and the Ameritrust debacle convinced me that there had to be a better way of managing risk in the banking industry.
In 1992, Ameritrust was acquired by Society Corp., the precursor of today's KeyCorp. I was placed in the position Executive Vice President of Credit Policy & Risk Management. In this capacity, I was given the opportunity to explore and experiment with new risk management tools that were beginning to bud in the industry. I was encouraged to do so by our CEO who expressed a desire to have a system whereby he could understand the totality of risk that our company faced on a daily basis.
Our CEO envisioned a process that could tell him how much aggregate risk the company was taking, including the risks that emanated from our credit, market and operational activities. He wanted a system that could allow us to increase, decrease or maintain our risk position as circumstances warranted. Neither of us realized it at the time, but he was describing a process that is today commonly called "enterprise-wide risk management".
In 1993 I commenced the first step of his vision by installing a Value-at-Risk (VAR) system in our company's trading floor. VAR was a highly complex model designed to measure risk in the bond, equity and foreign exchange trading we undertook on a daily basis. Due to the complexity of a VAR model, I had to engage several PhDs to help us implement it. During the course of their engagement, I happened to mention my frustration in finding an enterprise wide system that could aggregate the risk of each of our banking activities. One of the PhDs suggested that I look into the concept of economic capital allocation, now commonly known as "risk based capital".
Once I investigated the premise of risk-based capital allocation, I concluded I had discovered a powerful risk management tool. Implementing such a model at KeyCorp would enable us to allocate capital to our lines of business based on the amount of risk they took. Each line of business would be charged for the amount of credit risk, market risk and operational risk they encountered. Using the aggregate of that capital charge as the denominator, and the revenue they generated as the numerator, we could determine which lines of business were getting appropriately paid for the risk they took. For the first time, we would be able to put all our lines of business on an apples-to-apples comparison basis. Hence: the ability to know our level of risk and whether or not we would be paid for the risk being taken. Further, we would be able to aggregate the total amount of capital being allocated to all our lines of business to understand the totality of risk our company was taking. It was the enterprise-wide solution we had been looking for.
KeyCorp commenced building an economic capital allocation program in the mid 1990s because we firmly believed it was the right thing to do. It has taken us a nearly a decade to build it, and we are still not finished with it. Nonetheless, even after nearly 10 years we remain convinced that it is the best way to run our company. No regulator has told us that we must do this.
We are pleased to note that this powerful risk management tool, economic capital allocation, is now the underlying driver of Basel II. Our company was highly critical of the initial version of Basel II and publicly stated as much. We felt that it failed to address the sophistication and complexity that our industry routinely operated in. We felt it was inadequate and little better than the original Basel I. Put simply, it did not adequately address risk sensitivity. However, over the next several years we were pleasantly surprised to see how Basel II became a much better document. The regulators working on the new accord have been genuinely receptive to hearing the concerns that KeyCorp and others have raised. We haven't always gotten our way, but at least we have been heard.
We believe that Basel II is now on the right track. Financial institutions will need to develop more sophisticated risk management tools to support the risk based capital premise upon which it is built. This is a good thing. In today's world of complex financial markets, tools such as value-at-risk, two-dimensional loan grading systems, enterprise data warehouses and operational loss databases are not a luxury; they are a necessity. In order to understand their risk positions, banks should be calculating risk based capital and using these tools to do so. While models are no substitute for human judgment, they certainly create a more informed human with whom to make the decision.
One of the benefits we see in the Basel II proposal is that we will finally be free to price our products and services commensurate with the risk they entail. As previously mentioned, Basel I provides very little in the way of risk sensitivity. One of the perversities of this shortcoming is that it has driven high quality borrowers away from the banking industry. These clients can access providers of credit not subject to the costly level of capital that banks are currently required to hold. In essence, banks are forced to overprice for this business, and they lose it to other cheaper, non-regulated providers. Conversely, Basel I's simplistic 8% capital requirement has allowed banks to hold less capital than they should against borrowers that are high risk. This has resulted in banks underpricing such credit. It should be no surprise, then, that Basel I has chased high quality credits away from banks, while attracting low quality credits to them.
If banks are allowed to calculate the proper level of capital to be held based on a realistic stratification of credit risk, this serious problem will largely disappear. This is one of the tenets that Basel II is based upon: you hold the level of capital necessary to support the risk, and price for it accordingly.
I would now like to address some of the criticisms that have been leveled against Basel II. These would include its cost, complexity, inflexibility and propensity to foster pro-cyclicality. I would also like to provide a few comments on the merits of Basel II's Pillar 1 versus Pillar 2.
Much has been said about the cost of building the models necessary to comply with Basel II. At KeyCorp, we wonder how anyone can afford not to build them. We, ourselves, have painfully learned the cost of not having them. In 1996, our risk based capital process was still in its embryonic stage: in truth it didn't begin to take hold until 2000. In '96 we were still calculating profitability measures utilizing the primitive 8% capital standard stipulated by Basel I. On this basis, one of our loan portfolios, leveraged lending, was producing an eye-popping return on equity close to 30%. As a consequence, we unfortunately pursued expansion of leveraged lending over the next several years. At the end of 1998, the quality of this portfolio began to collapse and we have written-off many millions of dollars since.
We have looked retrospectively on our experience with this portfolio. We believe if we had had our risk based capital model in place (the kind proposed by Basel II) our anticipated return would have been in the single digit range. Such knowledge would have caused us to avoid this particular lending activity and to seek other opportunities that offered better risk/reward ratios.
Through this experience, we have learned an important lesson from which others can benefit. The entire cost of the nearly 10-year effort to implement our economic capital model (the same kind proposed by Basel II) pales in comparison to the cost of not having it in place.
We have read that others estimate the cost of compliance with Basel II to be staggeringly high. We are not convinced this is the case, and it certainly has not been so at KeyCorp. Yes, we have spent multiple millions of dollars over the years investing in risk management tools and models, but we've done so because we believe those tools are necessary to conduct our business in a safe and sound manner. Frankly, they will also make us a better competitor. The more we understand our risk, the better we will be at managing and pricing for it.
Some have criticized the cost of auditing and back-testing the accuracy of the models that Basel II is based upon. We view such activities as nothing more than good common sense. Auditing and back-testing of outputs is critical to ensuring that the model is producing reasonable numbers. Auditing/back-testing serve as the tuning devices necessary to modify the models' calculations. For example, auditing and back-testing of VAR models is an accepted practice in the industry now: everyone knows their benefit. We view auditing/back-testing as necessary investments needed to create a better model. Better models create better understanding of risk and the ability to better manage it. Better management of risk results in lower losses to banks. We believe the cost of auditing/back-testing is inconsequential compared to the losses that can occur due to inferior risk management processes.
Before one accepts the large figures attributed to Basel II compliance, one must subtract the costs of building the risk management systems that a good financial institution would invest in, regardless. We do not believe the gap between the two is significant.
We cannot deny that Basel II is a complex document. It is. Yet, it needs to be. Banking is a complex business that needs complex solutions to the issues it faces. We should not run from complexity but instead be willing to face it and manage our way through it.
I have previously mentioned that KeyCorp installed a VAR system for its trading floors in the early 1990s. At that time, many were saying VAR systems were exceedingly complex, expensive and too mathematically driven. Yet, today VAR systems are widely recognized as the standard by which to manage risk in their trading books. VAR is a superior risk management tool that never would have come to be had the financial services industry been intimidated by its complexity. I reiterate: when VAR first surfaced, it was accused of being too complex, costly and mathematically driven, the same crimes Basel II stands accused of today. Yet, VAR has become the industry standard.
Some fear Basel II will trap the industry with year 2000 era risk management tools and stifle creation of new ones. We believe this concern is overstated. The 1988 Basel Accord was a woefully inadequate document from the start. Its simplistic approach mandated a specific capital level and made no provisions to the contrary. Yet, over the past 15 years, the financial services industry has continued to develop new risk management tools never envisioned by the '88 accord. These would include: VAR models, two-dimensional loan grading systems, economic capital models and enterprise-wide data warehouses. The fact that such tools were not contemplated by Basel I did not interfere with the industry's pursuit of them. We anticipate a similar situation with Basel II - banks will continue to pursue a better risk management mousetrap. We will acknowledge, however, that regulators must be willing to consider the new tools as they are developed, and work with the industry to accommodate them as their effectiveness is demonstrated.
We have frequently heard that regulators are concerned that Basel II might allow substantial capital to escape from the banking system. We believe the whole premise of pro-cyclicality is evidence that such concerns may be overstated. Basel II capital levels represent the minimum level of capital that an institution is to hold. The premise of pro-cyclicality assumes that banks operate at or near the minimum capital level. We believe it is highly unlikely that any banking company worth its salt will allow their capital to sink to the lowest acceptable level.
Some argue that under Basel II, economic downturns will cause financial institutions to become more reluctant to lend when liquidity is most needed. Banks would be placed in a position of making a difficult choice: immediately raise new capital or stop lending. In truth, there is a third choice that most banks will probably follow: retain a buffer level of capital to accommodate cyclical changes in risk that everyone knows will inevitably occur.
We also believe that even in times of economic stress, banks genuinely desire to make new loans to drive their own revenue streams. Our current economic situation is a prime example: banks are anxious to lend money. The demand simply isn't there.
Pillar 1 versus Pillar 2
One of the basic principles of Basel II is to make risk transparent so that it is comparable from one institution to another. Pillar 1 encourages a formulaic based system that will enable this to occur. Consistency of methodology is critical to empower investors, regulators and depositors with the information they need to gauge the risk of the institution with whom they are dealing. Without Pillar 1's consistency of approach, a Tower of Babel syndrome can occur.
Pillar 2 relies more on flexible judgment as to how much capital is warranted at an institution. We acknowledge and accept that regulators must have the flexibility to invoke their authority to ignore the results of Pillar 1 when circumstances so dictate. However, completely abandoning Pillar 1 in favor of Pillar 2 yanks any comparability benefit away from investors and depositors. The invisible hand of the market will be impeded in its ability to quickly discipline a wayward institution.
For example, much has been said about the need to place operational risk under a Pillar 2 approach. In this regard, the individual regulator that happened to be examining a particular bank would largely determine the adequacy of capital held for its operational risk. This lends itself to varying assessments, interpretations, methodologies and enforcements. An investor attempting to compare the level of capital held for operational risk at multiple banks must assume that different examiners will utilize the exact same thinking in their operational risk assessments. That simply doesn't happen. A more formulaic approach, where all banks are using the same scorecard, lends itself much more to consistent comparability.
The mere presence of a Basel II draft has caused many in the industry to start contemplating new ways of tracking operational risk. This would include KeyCorp. We have commenced building an operational risk database that will give us better information regarding the source, size and amount of operational losses. This database will ultimately serve as the system that feeds our operational risk model. We believe it can be supplemented by exchanging information on operational risk losses with other financial institutions. This will help us build the critical mass necessary to create reliable, predictive loss forecasting models. I will readily admit that we have a way to go in this particular area, but the presence of Basel II over our heads encouraged KeyCorp and others in the industry to get moving on building the databases sooner.
In conclusion, KeyCorp believes that Basel I is hopelessly broken and that a new accord needs to be implemented. Basel II is a major step forward and we applaud its approach. It is not perfect now, nor will it be perfect when implemented, nor perfect 10 years after implementation. Regardless, it is light years ahead of Basel I as well as any other proposal we have seen to date. It should be supported.
We acknowledge it is complex, but banking is a complex business. A simple solution to complex issues is probably not the right medicine. As an industry, we should not shy away from the remedy simply because it is complex. Instead, we should work collectively with the regulators to find the right solution, not the easy one.
We have our doubts as to the high cost figures attributed to Basel II. Our own experience to date has proven to the contrary. Further, we believe many Basel II costs are simply expenditures we should otherwise be making as a matter of sound banking practice. Good risk management costs money, but it is intended to help avoid even bigger costs that arise from bad risk management.
We do not believe adoption of Basel II will trap the financial services industry in a time warp. Banks will continue to develop better methods of managing risk regardless of what Basel II requires. However, regulators must be open and responsive as these new tools are developed.
We believe there is substantial merit to including as much as we can in Pillar 1 versus Pillar 2. One of the greatest benefits that Basel II promises is that it will utilize the invisible hand of the market to discipline wayward institutions. In order to do that, investors must have adequate information to compare the risk of one institution against another on an apples-to-apples basis. Pillar 1 is the best vehicle for ensuring that banks report on a consistent basis.
Mr. Chairman, KeyCorp appreciates the opportunity to share our views on Basel II. We want to make sure our industry operates within a safe and sound environment. We know this is the goal of the committee as well as our friends in the regulatory world. While Basel II is far from perfect, it certainly moves us further down the path.