via e-mail
October 31, 2003
Office of the Comptroller of the Currency
250 E Street, SW
Public Information Room
Mailstop 1-5
Washington, DC 20219
Attention: Docket No. 03-14
Ms. Jennifer J. Johnson, Secretary
Board of Governors, Federal Reserve System
20th Street and Constitution Avenue, NW
Washington, DC 20551
Re: Docket No. R-1154
Robert E. Feldman
Executive Secretary,
Federal Deposit Insurance Corporation
550 17th Street, NW
Washington, DC 20429
Attention: Comments
Regulation Comments
Chief Counsel’s Office
Office of Thrift Supervision
1700 G Street, NW
Washington, DC 20552
Attention: No. 2003-27
RE:
Advance Notice of Proposed Rulemaking Risk-Based Capital Guidelines;
Implementation of New Basel Capital Accord
Dear Sir or
Madam:
Mellon Financial Corporation, the parent of Mellon
Bank, N.A., Pittsburgh, Pennsylvania, appreciates the opportunity to
comment to 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 (collectively, the
“Agencies”) on the Advance Notice of Proposed Rulemaking
(the “ANPR”).
Mellon Financial Corporation (Mellon) has a number
of concerns with the Basel II Accord, and with the proposed U.S. rules
and standards laid out in the ANPR. The most significant concerns for
Mellon continue to focus on an explicit Pillar I capital charge for
operational risk, its inapplicability to many of our competitors and the
limited recognition of mitigants other than capital for operational
risk.
The provisions of Basel II and the rules and
guidelines considered by the ANPR, are unnecessarily complex. This
degree of complexity will lead to a number of problems; inconsistency in
the definition and enforcement of international regulatory standards,
difficulties for banks in interpretation of regulations, and the risk of
non-comparable (and potentially misleading) information being provided
to third parties under Basel disclosures. The Accord and the ANPR
attempt to imply a level of precision in determining capital that, in
reality, does not exist.
The provisions of Basel II, and the rules and
guidelines considered by the ANPR, are too prescriptive in nature.
Institutions and regulators have limited opportunities for the exercise
of reasoned business judgment in such a rule-based approach. The ANPR
should allow institutions to make distinctions relating to the degree of
risk and materiality in credit portfolios and operational processes.
With such an approach, an institution and its regulators would have the
flexibility to engage in cost effective risk management. The ANPR rules
do not fully allow differentiation of high quality or low dollar size
credit risk portfolios. Similarly, a Pillar II approach to operational
risk should allow the needed flexibility.
We support the intent of the Agencies to ensure
that boards of directors and senior management take responsibility for
appropriate risk management. However, the involvement and
responsibility of the board must be balanced with the fact that bank
directors have numerous responsibilities, including those increasingly
related to ensuring appropriate corporate governance safeguards are in
place and working. The ability of directors to set policy and to ensure
that management adheres to it is undermined if directors must at each
meeting review lengthy and detailed mandated reports. Buried in detail
that is best delegated to management, boards can become unable to spot
key emerging risks and address them. The board or a designated
committee should approve and periodically review the bank’s operational
risk management framework; the design, implementation, operation and
monitoring of a risk management system should be within management’s
duties. The board of directors should be kept informed of material
issues as they arise, with periodic reports as appropriate.
The Agencies refer to the capital requirements
currently in place in the United States under Prompt Corrective Action
legislation, specifically the leverage requirement. We believe the U.S.
banking regulators should consider the elimination of the leverage
capital ratio in conjunction with the adoption of Basel II. The
leverage ratio is fundamentally incompatible with an advanced,
risk-based capital regime. The primary purpose of adopting Basel II is
to introduce a broader menu of risk weightings for different asset
categories. This is in response to the single largest criticism of
Basel I, that there were too few risk categories and that loans to
triple-A rated corporations carried the same risk weighting as sub-prime
consumer loans.
The leverage approach, where all assets are risk
weighted identically, is an additional step backward even from Basel I.
Further, there is no provision in the leverage approach for capitalizing
off-balance sheet risks. It appears that the leverage ratio, like the
Pillar 1 ORBC requirement, exists solely to impose a “floor” on the
amount of equity capital that Banks and FSHC’s are required to
maintain. If that is indeed the case, it makes no sense to require
institutions to spend tens of millions of dollars for advanced
measurement systems and then to have the results of those measurements
essentially thrown out by having the leverage ratio become the minimum
capital standard.
The Agencies have indicated that U.S. banks
adopting Basel II capital standards, may only adopt the Advanced
Internal Ratings Based (A-IRB) approach to credit risk, and the Advanced
Measurement Approach (AMA) to operational risk. This limitation has
implications for overall bank capital, the soundness of the banking
system, and the ability of regulators in the field to appropriately
assess minimum capital standards. Limiting U.S. banking institutions to
these two approaches will lead to international inequality, as non-U.S.
institutions may pick from three credit approaches, and three
operational risk approaches.
Operational Risk Capital
Due to the numerous problems inherent in the
Accord, which we outline below, a Pillar II approach (which contemplates
regulators working with institutions to best understand and dimension
operational risks) provides a much more workable solution to the
determination of required capital.
·
Basel II will result in an incremental capital charge
for operational risk, which in the case of specialized trust and
processing banks, will not be offset by a reduction in required
capital for credit risk.
·
Since many of the
competitors of specialized trust and processing banks will not be
subject to the Accord, such a capital charge imposes an unfair
competitive burden.
·
The
Accord introduces arbitrary constraints on risk mitigants.
Insurance, which is an appropriate mitigant to unexpected losses in all
areas of commerce, is inappropriately limited to 20%. Further, the
one-year time limits imposed on insurance recoveries is problematic as
well.
o
The
one-year time limitation fails to consider the timetable of commercial
litigation. Frequently the determination of loss amount, and insurance
recovery, is not determinable until a date well into the future.
o
Institutions should be allowed to model loss data taking into account
their insurance coverages and recovery histories, bound by neither an
arbitrary percentage limit, or time of recovery limitation. Taking
these factors into account provides a realistic approach to the degree
of risk that exists in any loss situation. Modeling of losses and
recoveries should of course be subject to review and signoff by the
institution’s banking regulators.
Stable,
recurring fee based earnings for businesses that do not also contain
credit or market risk should be considered as a mitigant for unexpected
losses in other businesses.
·
Most
U.S. institutions can benefit from significant tax savings associated
with operational losses, via charges in the current year, as well as
loss carry back and carry forward. Failure to consider loss data on
an after tax basis overstates the impact of modeled losses.
·
Operational risk capital
is required for expected and unexpected losses; it should only be
required for unexpected events. At Mellon, expected operational
losses are incorporated into the business planning cycle.
·
There
is limited evidence that operational risk can be modeled accurately
and with any predictive power. This is further exacerbated by the
requirement of modeling operational risk capital at the 99.9%
confidence level. A confidence level of 99% is more appropriate.
This confidence level is typically used in Value at Risk calculations
for modeling market and interest rate risk.
Operational risk data for external events is not a
reliable or even relevant indicator of the future and also does not
contain critical root cause or scalar information. Most institutions
lack significant internal data – due in large part to their success in
running effective operational units. External loss data only reflects
the largest losses. This overstates the severity of loss
distributions. When only large losses are modeled, this calls for a
higher level of capital due to the severity of the presumed
distribution.
Balance Sheet Issues
The ANPR seeks to impose risk based capital rules
on institutions whose balance sheets vary greatly in terms of size,
quality and liquidity. Where the asset type considered is either small
in comparison to the overall capital structure of the institution, is of
high quality, and/or is extremely liquid, such assets should not be
subject to the burden of new control and system requirements.
- For instance, very liquid assets such as
investment grade securities, intrabank deposits, money market
investments, etc., with low credit risk, should require controls
commensurate with their risk.
- Such assets, including bank investment
securities portfolios should not fall subject to requirements for
additional systems and controls, which are not sensitive to the degree
of risk posed by an asset type.
- Within the credit portfolio, where loan assets
or commitments are publicly rated, the systems and controls should be
commensurate with the level of risk in the portfolio.
Credit Risk Capital
Mellon has examined the credit risk
components of Basel II and the ANPR. The size and quality of our credit
portfolio do not merit exhaustive modeling and review of the proposed
rules. Nonetheless, there are numerous elements of the ANPR that require
our response.
• The A-IRB approach to credit risk is
not an appropriate solution for institutions where credit risk
exposure is not a large risk. Investment in such models should not be
necessary for institutions whose primary focus is in the trust and
processing businesses. On the other hand, merely reverting to a Basel
I approach is hardly an enhancement in risk management. Mellon thus
proposes below several ways to address this problem.
• The A-IRB approach also has a number
of shortcomings that will make the jobs of regulators and field
examiners more difficult.
- As the determination of the
appropriate capital amount is left to each institution and its
regulator, it is possible for two institutions holding exactly the
same asset to hold differing capital amounts against that asset.
With many institutions holding the same assets, in a shared national
credit environment, this result is not appropriate.
- This result reveals the likelihood
that many banks will be motivated to understate their capital needs,
through the use of complex models.
- Due to the difficulties in adopting
an A-IRB approach (at the individual bank, and system wide level) we
believe that over time, the U.S. will (as examiners in charge
compare credit allocations for the same loans at their various
institutions) move to an approach similar to the Foundation Internal
Ratings Based Approach (F-IRB). (Here regulators will establish the
inputs for loss given default, exposure at default, and remaining
maturity, with banks determining probability of default for their
individual loans and portfolios.) If definition of the reasonable
range for these variables is inevitable, why shouldn’t the F-IRB be
an option from the onset?
- In the case of Shared National
Credits, the probability of default should be determined at the
agent bank for all institutions in the bank group.
• In light of these issues, we believe
U.S. institutions should be able to choose the Standardized, F-IRB or
A-IRB approach to credit risk.
• Were the Agencies to proceed with the
adoption of the Basel Capital Accord, we believe that the mandated use
of the A-IRB approach in conjunction with the AMA for operational risk
poses a significant cost burden and disincentive for trust and
processing banks. With this in mind, we believe the Agencies should
consider a more appropriate structure for non-credit intensive banks,
as shown in the two modified approaches to credit risk capital below:
Option 1: Permit Selection of
Either Basel I, Standardized or Foundation IRB for Credit with AMA for
Operational Risk.
Benefits
o Provides a simplified approach for
banks that have a constant to improving credit risk profile with
portfolios or that have stable to declining exposure levels.
o More cost effective for Banks.
o Maintains a level playing field for
asset management focused banks.
o Permits institutions to properly
allocate resources to the areas of greatest risk.
Option 2: For Credit IRB, at the
Portfolio Level and Applied to Exposure, Use a Materiality Threshold of
100% of Total Capital (subject to rolling five year average credit
losses in those portfolios being less than 1% of net operating income
before tax).
Benefits
o Balances the Accord principle of
requiring increasing sophistication in risk management tools and
technology for larger credit portfolios with material exposure
levels at risk.
o More cost effective for Banks.
o Maintains a level playing field for
asset management focused banks.
o Permits a phase in period for
growth portfolios.
o Permits institutions to properly
allocate scarce resources to the areas of greatest risk.
Within the ANPR, the Agencies posed
numerous questions regarding the A-IRB approach to credit risk and
credit capital allocation. Notwithstanding our strong objections
described above, we have examined a number of these points, and our
responses are contained in Attachment 1*. We have not commented on a
number of issues for which we do not have material exposure, such as
securitizations.
Market Discipline
The Agencies provide a discussion of
Pillar III disclosure issues in the ANPR. This appears to be cursory in
nature, and we would anticipate more detail in the proposed rules at a
later time. We remain concerned that such mandatory disclosure is
dangerous to the banking industry.
• Although we feel it is appropriate to
openly share risk information with our regulatory agencies, and have
and will continue to do so, the requirement for mandatory disclosure
of detailed risk capital elements is not appropriate.
- Although providing this information
might foster a greater level of transparency, it is questionable how
individuals and other entities would comprehend or use that
information. Banking institutions in the United States already
provide substantial disclosures of financial information, and it is
our perception that additional mandatory disclosure is not
warranted.
- Wide scale disclosure as
contemplated will lead to confusion among users of that information.
Although banks would disclose their loan portfolio composition in
gross terms, the underlying portfolios themselves may be radically
different – especially in the higher risk and unrated categories.
- This problem is further compounded
by the high likelihood of an uneven playing field for many banks.
Non-bank competitors, not subject to this level of disclosure may
well be advantaged in terms of the public’s perception. At a
minimum, their cost structure for reporting compliance would be
significantly less.
• Public access to risk/loss
information can have a number of consequences, including inappropriate
use of the information for competitive purposes and used against banks
by class action lawyers. Raw data is prone to misinterpretation. Some
losses, which may have reasonable explanations or which resulted from
problems that have been remedied, might require the organization to
defend its data in numerous forums, including responding to RFPs and
securities analysts. Such open dialogues might jeopardize confidence
in the banking system in general, if not in specific institutions, by
artificially heightening concern and focusing the debate on matters
that might otherwise not be of concern to experienced regulators.
Also, disclosure of such information might provide a roadmap for
litigators, particularly the class action bar, thus exposing the
banking industry to unwarranted litigation with its attendant expense
and reputation risks. This information would establish a floor for
negotiations and always result in increased cost for the bank.
• As a result of the options presented
to institutions under Basel II, data will rarely be comparable from
institution to institution. This results because of a diversity of
models that will be utilized among different institutions. Diverse
models, using varying assumptions, will yield a broad distribution of
results. Data from those models is not comparable, and will be
misleading to those who try to compare it.
• Mandatory disclosures such as those
set forth in the ANPR should be eliminated. Principles for disclosure
in lieu of prescriptive rules would be less burdensome, and more
appropriate to banks and third party users of that information.
We thank you for the opportunity to
comment on ANPR. If you should have any questions about our comments or
would like to discuss them further, please call Michael Bleier, General
Counsel, at 412-234-1537.
Sincerely,
Steven G. Elliott
Senior Vice Chairman
Mellon Financial Corporation
*
The Attachment document can be viewed in the FDIC Public Information
Center, 801 17th St NW, Washington, DC, during business days from
8:00 am to 5:00 pm.
|
