November 3, 2003
US BANCORP
Ms. Jennifer J. Johnson,
Secretary
Board of Governors of the Federal Reserve System
Twentieth Street and Constitution Avenue, NW
Washington, DC 20551
Attention: Docket No. R-1154
Office of the Comptroller of the Currency
250 E Street, SW
Public Information Room
Mail Stop 1-5
Washington D.C. 20219
Attention: Docket No. 03-14
Robert E. Feldman
Executive Secretary
Federal Deposit Insurance Corporation
550 17th Street, NW
Washington, DC 20429
Attention: Comments
Ladies and Gentlemen:
U.S. Bancorp welcomes the opportunity to respond to the
Agencies' Advance Notice of Proposed Rulemaking ("ANPR") for
advanced approaches for measuring regulatory capital. U.S.
Bancorp appreciates the effort the Agencies are making to work
with the banking industry on developing a more risk sensitive
capital accord.
This letter is composed of the following sections.
1. General Comments
2. Capital Framework
3. Credit Risk Capital
4. Operational Risk Framework
5. Securitization Data Collection Requirements
6. Disclosure Requirements
7. Cost of Implementation
1. General Comments
U.S. Bancorp supports the idea of updating the regulatory capital
framework to include a more risk sensitive measure of capital. We
believe that this is consistent with a safe and sound banking system.
U.S. Bancorp broadly supports the Advanced Internal Ratings Based
Approach ("AIRB") conceptual framework of probability of default, loss
given default and exposure at default. U.S. Bancorp also supports the
announcement of the Basel Committee to simplify the Basel II Accord. We
believe that the current capital rules should be updated to a more risk
sensitive measure. However, the complexity of the current proposal is
greater than necessary.
Given the recent announcement that the Basel Committee will be
redefining the regulatory capital requirements to include only
unexpected losses and effectively excluding expected loss from
regulatory capital, we urge the agencies and the Basel Committee to open
the issue of the definition of regulatory capital. U.S. Bancorp believes
that you cannot separate these two issues. We believe that the
regulatory definitions of capital and the accounting definitions of
capital need to be aligned.
U.S. Bancorp supports the three pillar concept used in the ANPR. We
believe that shifting the operational capital requirements to pillar two
should further strengthen this concept. This belief is founded on the
idea that the best method of managing and reducing operational risk is
the prudential use of controls. We are very concerned that a focus on
operational capital requirements will shift focus from managing
operational risk to managing operational capital requirements. We don't
believe that the capital attribution process for operational risk has
evolved sufficiently to warrant a pillar one treatment. We believe that
the adoption of sound practices for operational risk management is the
prudent course of action.
U.S. Bancorp urges the Agencies to delay the implementation date of
the capital rules. The proposed implementation timeframe specifies that
all data collection be in place by January 2004, however, the final
version of the rules have not been specified and are not expected until
the middle of 2004. The changes required to existing data systems and
creation of new data systems to collect the detailed data required by
the ANPR will take at least a year to create. These data systems will
require U.S. Bancorp and our software venders to expend substantial time
and effort to complete the required changes. In order to reduce the
implementation costs we believe that the requirement for the start of
data collection should be specified as one year after the finalization
of the rule. The final implementation of the rule should remain three
years after the inception of data collection.
2. Capital Framework
Competitive Impact
The proposed scope of application of the ANPR to only the largest
banks raises the issue of the impact on the banking industry. It is
clear that these rules are having an effect already on the behavior of
the banking industry. The proposed rules will create a systemic capital
arbitrage within the industry. The higher quality credit portfolios will
be incented to migrate to the larger banks and the lower quality
portfolios will be incented to migrate towards the smaller banks not
required to implement the ANPR. This shift of credit risk from the
larger more sophisticated banks to the smaller banks and non-banks will
have a profound impact on the ability of the banking sector to provide
credit to the economy.
Capital Definition
U.S. Bancorp recommends that the Agencies and the Basel Committee
reconsider the definition of capital. We believe that current definition
of regulatory capital excludes some elements of shareholder equity that
serve as capital. U.S. Bancorp strongly believes that the definition of
capital must be revisited contemporaneously with the risk based capital
requirements. The current definition of regulatory capital excludes
intangible assets that we believe should be included.
The treatment of the capital requirements for asset management and
merchant processing intangibles should be revised. The capital
requirement for these businesses is greater than the capital
requirements of non-bank entities. We believe that the treatment of
these businesses gives an unfair competitive advantage to non-bank asset
managers and merchant processors. Asset management and merchant
processing are important businesses for many banks. The customers of
banks need these services, however, if the capital requirements for
these businesses remains artificially high, banks will over time lose
this business to non-bank competitors due to the higher capital
requirements.
U.S. Bancorp believes that substantial improvements over the past
several years in GAAP standards have changed the nature of these
intangible assets. The accounting standards hold the valuation of these
assets to a high standard that provides a high degree of assurance that
the fair value can be realized. These accounting standards have created
a level of assurance in the valuations such that the automatic exclusion
of these assets is no longer warranted. U.S. Bancorp urges the Agencies
to consider the definition of regulatory capital as a critical component
to the revision of the capital rules.
Revised Capital Framework
On October 11, the Basel Committee announced a number of changes in
the proposed framework of the Basel II Capital Accord. One of these
changes was the exclusion of expected losses from the regulatory capital
requirements. In principle U.S. Bancorp supports the exclusion of
expected losses from regulatory capital requirements. I encourage the
regulators to work in consultation with the industry on determining the
appropriate methods for framing the capital rules. U.S. Bancorp is
prepared to work closely with the Agencies as part of an industry group
in formulating a capital framework and developing the calibration of the
capital rules.
U.S. Bancorp supports the principle that the overall capital in the
system remains the same, provided the treatment of innovative capital
instruments remains unchanged. Many banks use innovative tier one
capital instruments as a means to provide necessary regulatory capital.
Potentially, FIN 46 may eliminate these instruments on a go-forward
basis. Absent the development of new replacement instruments, we would
encourage the Agencies to revisit the tier one capital requirements. We
believe that in the long-run replacing innovative tier one capital with
more expensive forms of capital will have a negative impact on the risk
appetite, profitability and competitiveness of the banking industry.
Excessive Prescriptiveness
The Draft Supervisory Guidance (DSG) is very detailed and
prescriptive in the specifications of the infrastructure and management
processes surrounding the proposed capital rules. The level of
specificity contains some inherent contradictions and overly burdensome
stipulations. The data retention requirements include stipulations for
five years of historical data, and also "cradle-to-grave" data
retention. The ANPR references maintaining the key data supporting a
rating, while the DSG suggests that all possible rating factors be
retained whether or not they are currently used in establishing the
rating.
We would suggest a more principle-based approach. Clearly it is
important to retain key data for a relevant period of time. We believe
that this determination of key data elements should be subject to
management discretion and supervisory oversight. Establishing data
retention requirements for all data for an indefinite period of time is
not efficient, practical or necessary. Ratings migration data should be
maintained for a minimum. of 10 years. Loan rating factors by loan
should be maintained for at least five years. The use of stress testing
should be retained as a means of ensuring that prior credit cycle events
be evaluated. The use of stress testing is a best practice risk
management technique for evaluating the prior credit cycle events that
fall outside of the 10 year data retention time period.
3. Credit Risk Capital
Definition of Default
The definition of default is too prescriptive for wholesale and
retail exposures. Few banks recognize all types of defaults specified by
the regulators. As an example "silent defaults" are not generally
recognized for wholesale exposures. For retail exposures, inclusion of
distressed restructuring or workout involving forbearance is not
currently recognized as a default. The definition imposed by the ANPR
guidelines will impose additional costs on banks, as they will be
required to maintain and report on multiple default types to external
sources.
SME Definition
U.S. Bancorp recommends that the definition of SME be revised. The
current proposal of $50mm of assets or annual sales, in our opinion,
creates an unnecessary data collection burden. The purpose of the SME
category is to reflect a lower correlation factor. We believe that this
is due to the size of the loan rather than the size of the borrower.
Take the example of a leasing portfolio. An office equipment leasing
portfolio is comprised of small ticket equipment. Companies of all sizes
use this type of equipment. Using the size of the loan or lease as a
proxy for diversification is a reasonable simplification of the capital
rules. Mandating the collection of revenue data for the company as part
of the capital calculation is a complication that is not necessary in
our view.
We believe that concern that banks will arbitrage the capital rules
by making many small loans as a means of circumventing the higher
capital requirements of the corporate classification can be mitigated by
the Agencies through pillar two. Banks as a normal part of their risk
management practices track concentration risk. This reporting can be
effectively utilized by the regulatory agencies as part of a pillar two
review of capital adequacy.
We believe that pillar two can be used as an effective means of
addressing concentration risk. U.S. Bancorp has been steadily working
over the past several years to reduce the risk profile of the bank.
Reducing concentration risk has been an important element of this risk
reduction. We recommend that the Agencies consider eliminating the SME
curve and recalibrate the corporate curve using assumptions of greater
diversification. Calibrating the capital curve for a diversified
portfolio will serve as an incentive towards better risk management when
coupled with a pillar two review of diversification against the pillar
one standard.
LGD Floor for Residential Mortgages
The proposed 10% LGD floor for residential mortgages is too high. We
believe that establishing an arbitrary floor is not necessary. A 10%
floor is not appropriate for certain segments of the residential
portfolio, particularly low loan-to-value loans or loans with PMI. The
reliance on loss experience through-the-cycle and the conservative
requirements built into the LGD specifications will provide an
appropriate level of conservatism. A 10% LGD floor is an excessively
conservative assumption that we believe should be eliminated from the
ANPR.
Home Equity Loans
The inclusion of home equity loans with first lien residential
mortgages causes home equity loans to have a capital requirement that we
believe is too high. The correlation factor assigned to residential
mortgages is too high for home equity loans. We encourage the Agencies
to either establish a new category for home equity loans or
alternatively change the correlation factor to correspond to observed
industry data. We have analyzed our own experience in home equity loans
and have measured a significantly lower correlation factor than that
assumed in the ANPR. We believe that the use of too high of a
correlation will significantly discourage the offering of home equity
products by banks using the AIRB approach.
We believe that the example of the home equity product illustrates
the importance of further calibration. The variation in capital
requirements for different products to the same customer due to the
selection of the correlation factors in each curve should be further
reviewed. We believe that it is important that the capital requirements
for various bank products be commensurate with their relative risk. When
we compare the capital requirements of retail products by customer, the
capital requirements can vary substantially. We recommend that as the
curves are recalibrated for unexpected loss, that more scrutiny be
applied to the relative capital requirements of the various bank
products for comparable risk levels.
Short Maturity Loans
The primary purpose of the new capital accord is to establish a more
risk sensitive minimum capital requirement. The assumption of a one-year
maturity term for all exposures creates an increased capital requirement
for loans with maturity less than one year, compared to longer maturity
loans. These short dated loans should have a lower capital requirement
than longer dated loans of similar risk. We believe that this
enhancement of the rules can be accomplished relatively simply and
assign the capital requirements in a more equitable manner.
Unused Revolving Lines
U.S. Bancorp supports the Basel Committee's recent announcement to
revisit the treatment of unused revolving lines of credit. U.S. Bancorp
believes that these commitments are a low source of risk because of the
management practices that monitor unused revolver commitments. U.S.
Bancorp actively manages unused commitments as a means of limiting
credit risk. We believe that these practices should be factored into the
capital requirement calculations.
4. Operational Risk Framework
Operational Capital Requirements
U.S. Bancorp does not see evidence that the industry has evolved
measurement methodologies sufficient to warrant an addition to minimum
regulatory capital requirements under pillar one. The measurement of
operational risk is clearly in the art phase and has not yet progressed
to a science. U.S. Bancorp believes that the management of operational
risk is very important. The proper focus of operational risk management
should be on the continuous improvement of controls and risk
assessments. U.S. Bancorp believes that the focus on capital for
operational risk will detract from management's time devoted to
operational risk management. U.S. Bancorp is firmly convinced that
operational capital, like other risks such as interest rate risk, should
be in pillar two. At a minimum, the scope of the AMA should be
restricted to those banks with operational risk as their primary risk
and credit risk as a secondary risk.
Operational Risks
Given the recent announcement by the Basel Committee to exclude
expected losses from the risk based capital requirements, we believe
that the Agencies should also revisit the topic of operational risks
covered by the ANPR. In our view, the so-called
high-frequency-low-severity operational losses are associated with
expected loss not unexpected losses. These types of losses are
characterized by credit card fraud and demand deposit fraud. These types
of losses are actively managed by banks with dedicated groups of risk
managers and priced into the product and services.
We believe that the proper events to be covered by a capital
framework should be high severity events. Given the difficulty of
modeling these types of events and the paucity of data we believe that a
Pillar two approach is most appropriate. A pillar two approach can take
into consideration the overall risk profile of the bank, the earnings
strength of the bank, the strength of the controls and management
judgement coupled with supervisory oversight.
We believe that scant attention has been paid to the importance of
strong pre-provision income in any economic capital framework for
operational risk. We strongly believe that banks with strong
pre-provision income require less capital for operational risk, all
other things equal. Banks with a strong risk management culture and
strong earnings have a significant ability to absorb unexpected
operational losses. We believe that this is a compelling reason for a
pillar two operational capital treatment.
Opportunity Costs
Opportunity costs and other indirect costs are currently not included
in the definition of operational risk. U.S. Bancorp believes that this
is appropriate for the purpose of establishing minimum capital
requirements. U.S. Bancorp does consider opportunity costs an
operational risk, but we would characterize many of these opportunity
costs within the bounds of expected losses. We believe that managing
indirect costs is a leading practice for sound operational risk
management. The definition of indirect costs should be limited to those
costs that can be quantified.
For example, an ATM will have a targeted cash level. If the ATM
carries too much cash, or runs out of cash the cost can be quantified.
When too much cash is carried, the balance sheet is inflated and the
excess cash must be funded with interest bearing liabilities. When the
ATM runs out of cash, the number of customer transactions that were not
processed can be used to quantify the processing fee opportunity cost.
Another example is a security fail to deliver. This is an opportunity
cost that can be quantified by the payment made to the counterparty for
the delivery failure or the cost to fund the failed security delivery.
An example of an event that is difficult to assign costs would be a
server outage that resulted in no lost business or overtime hours. If a
server goes down and is replaced by a backup server, there may not be
any financial impact to be quantified. There is an opportunity cost from
the perspective that a system support person spent time returning the
server to production, but if the work was conducted during normal hours
and didn't require any repairs or outside expenditures, the assignment
of a cost for the recovery is difficult to assign.
Insurance Mitigation of Operational Capital
The Agencies have defined insurance as a mitigant for operational
risk. The ANPR limits the insurance to 20% of the gross exposure if the
financial institution can show that the risk mitigants are sufficiently
"capital-like". We would recommend that the 20% be eliminated and
replaced with a requirement that the insurance be shown to be
sufficiently "capital-like" to qualify. The justification for this
change is due to the nature of operational risk. We believe that the
main components of net risk are the operational risks already inherent
in banking and the extent of controls in place to mitigate those risks.
A bank with good controls should be able to obtain insurance for
operational losses at a lower cost than the cost of capital. This type
of flexibility will help ensure that the financial industry can remain
competitive and a low-cost provider of services to its customers.
5. Securitization Data Collection Requirements
The ANPR proposes that rated securitizations have differentiated risk
weights based on tranche thickness and pool granularity This approach
requires that banks collect data for each tranche below the tranche they
own and evaluate the thickness of these tranches. We believe that this
requirement is excessive. We would propose that this requirement be
eased for tranches that are rate AA/Aa or higher. Exempting highly rated
tranches from this requirement will significantly reduce the burden of
data collection that we feel is of limited value.
6. Disclosure Requirements
U.S. Bancorp does not believe that the additional disclosure
requirements listed in the ANPR are necessary to provide the market the
necessary information to evaluate the risk profile of the bank. We
believe that the current disclosure requirements provide ample
information to the market about the risks and controls of the bank. We
believe that substantially increasing the disclosure requirements will
not increase the understanding of the market.
Secondly, we are concerned that excessive disclosure will create more
confusion than understanding by investors and depositors. This confusion
can disadvantage banks relative to non-bank competitors and decrease the
competitiveness of the banking industry.
7. Cost of Implementation
The ANPR and DSG as proposed will result in material expenditures by
U.S. Bancorp to change systems and business processes. Much of the
complexity is driven by the requirements to link data between systems,
acquire new data or systems. Many of the changes are driven by the
complexity of the proposed rules. The increased costs will have a
material impact to the expense base of the bank. Some of these expenses
will be absorbed by U.S. Bancorp because we will not be able to build
them into our pricing structure because our competitors in the
marketplace will not have to meet these rules, whether they are
non-banks or banks not adopting the ANPR rules. Other costs will be able
to be passed through our pricing to our customers who will bear the cost
of this new regulation. However, we believe that most of the expense
will be born by U.S. Bancorp and our shareholders.
We believe that simplification of the proposed rules is an important
step in reducing the cost of improving the capital measures. I encourage
the Agencies to consider the simplifications we have proposed in this
comment letter and also work with the industry in developing further
simplifications of the rules. Making these changes will provide a better
balance to the cost/benefit analysis and best serve the customers of
U.S. Bancorp.
Sincerely,
David M. Moffett
Vice Chairman and Chief Financial Officer
U.S. Bancorp
800 Nicollet Mall
Minneapolis, MN 55402
