Re: Interim Capital Treatment of ABCP
Program Assets/Permanent Capital Treatment of ABCP Program Assets
The American Securitization Forum thanks
the member agencies (the 'Agencies') of the Federal Financial
Institutions Examination Council for this opportunity to comment on two
related regulatory capital releases published in the Federal Register on
October 1, 2003: (i) the interim final rule on interim capital treatment
of assetbacked commercial paper ("ABCP") program assets (the "Interim
Final Rule") and (ii) the proposed rule on permanent capital treatment
of ABCP program assets (the "Proposed Permanent Rule").
We agree with the Agencies that without
an adjustment to the risk-based capital rules, the regulatory capital
applicable after giving effect to FIN 461 would exceed that
necessary to address the risks of a bank's exposure to ABCP program
assets. We greatly appreciate the hard work of the Agencies to provide
the interim regulatory capital relief set forth in the Interim Final
Rule.
Our primary comments are to the Proposed
Permanent Rule, although one technical comment discussed in Part F
below, relates to both the Interim Final Rule and the Proposed Permanent
Rule. Our points are set forth as follows:
• In Part A we advocate the delay of the
implementation of the proposed treatment of liquidity positions and
facilities with early amortization provisions until the implementation
of the Accord in the U.S.
• In Part B, we advocate the adoption of
an internal bank rating approach for determining required capital for
liquidity positions.
• In Part C, we suggest a more
appropriate conversion factor for liquidity positions than that
proposed.
• We discuss the appropriate requirements
for "eligible" liquidity facilities in Part D.
• Our comments on the early amortization
capital requirements are set forth in Part E.
Finally, we have three technical comments
set forth in Part F. We note that our comments in Parts B through E only
apply if the Agencies do not defer the implementation of the proposed
changes until the adoption of the Accord in the U.S.
A. Proposed should be implemented only
when the revised Accord is adopted
While we understand the Agencies' desire
to move the regulatory capital rules to a more risk-sensitive approach,
we believe that the proposed changes for treatment of liquidity
facilities and revolving transactions with early amortization features
should be made in the context of the U.S. implementation of the revised
Basel Accord (the "Accord"). With the ongoing work in developing the
final framework that will govern risk-based capital requirements for all
securitizations under the Accord, it strikes us an unfair to adopt
portions of the Accord which will impose additional capital requirements
on one subset of providers of funding through one particular type of
securitization. As one of the fundamental premises of the Accord is to
maintain consistency within and throughout jurisdictions in minimum
capital requirements, we believe that it is inappropriate to adopt
portions of the Accord prior to the international implementation date.
Not only will the early adoption of these proposals put U.S. banks at a
competitive disadvantage to foreign banks, it will also add additional
costs to funding through an ABCP conduit, making it a less efficient
funding source for customers. Given the relatively low risk of liquidity
positions for conduit transactions, we believe it is inappropriate that
these commitments attract a higher capital charge than other short-term
commitments, such as a back-stop facility to a corporate borrower for
its commercial paper issuances.
First, the 20% conversion factor is
simply the substitution of one arbitrary line for another (the current
0% conversion factor). In effect, the proposal for liquidity facilities
is an adoption of the Standardized Approach under the Accord-an approach
that the Agencies have themselves rejected in their initial
implementation proposal for the Accord in the U.S. For reasons discussed
in Part C below, we strongly believe that the true risk of liquidity
positions is closer to a 0% conversion factor than a 20% factor. Second,
there are several practical problems in the proposal relating to the
definition of eligible liquidity facilities-several of which are
different than those within the proposed Accord. While we discuss the
specifics of the problems below, we note that any proposal that would
change the standards for liquidity facilities will take time for banks
to implement for existing facilities either through an amendment process
or as these facilities are renewed.
We understand that the Basel Committee on
Banking Supervision received over 200 letters on the latest draft of the
proposed Accord and that the Committee is considering revising the
Accord to eliminate or simplify the SFA in whole or in part for
securitizations in order to replace all or portions of that approach
with a less complex approach. We do not believe that it makes sense to
change the capital requirements now when we know that within the next
couple of years they will be ultimately changed again upon adoption of
the Accord in the U.S., especially given the relatively low risk of
liquidity positions. Therefore, we request that the Agencies preserve
the status quo until the adoption of the revised Accord in the U.S. We
believe that the additional time provided will allow for a more fully
developed, risk sensitive proposal, less disruption for banks and their
customers in the implementation and less potential damage to the U.S.
ABCP conduit market due to the competitive disadvantage as a funding
source within the U.S. and the competitive disadvantage as to which U.S.
banks would be put as compared to their foreign competitors.
B. Adoption of an Internal Bank Rating
Approach
If the Agencies were not willing to delay
the implementation of the revised capital requirements in the Proposed
Permanent Rules mtil the adoption of the revised Accord, we recommend
the adoption of an internal ratings approach for determining the risk
weights applicable to liquidity facilities. The benefits of the adoption
of such an approach are two-fold. First, we believe that a bank's
internal system is the best method for determining the risk of a
liquidity exposure. Second, the early adoption of an internal ratings
based system will give all regulators a chance to become comfortable
with the internal approach for broader adoption at the Accord level.
Our proposal expands to liquidity
commitments the internal approach currently in place in the United
States for credit enhancement positions where the Agencies have already
shown their satisfaction with the ability of ABCP conduit sponsors to
analyze positions constituting and supporting the conduit's asset pool
using a variety of models and methods of analysis that have proven
highly reliable. We expect that if the Agencies were to adopt this
approach there would need to be a reasonable delay in the implementation
of the Proposed Permanent rule to allow for regulators to approve a
bank's internal system. We don't believe that this would require more
than a year given the work that has already been done by the regulators
in reviewing internal systems under the current rules for credit
enhancement exposures.
Under our proposal, banks would be
permitted to produce their own internal ratings generated from one or
more risk assessment models used by recognized external credit
assessment institutions or models and methods of analysis employed in an
internal system, provided that such bank has received specific approval
from its regulator to do so. Approval would be subject to a regulator's
complete satisfaction with a bank's ability to apply such models in a
reliable manner and the regulator's ability to validate it. These
internal ratings would then be used to determine the risk weight for the
liquidity position based on the ratings table proposed under the Accord,
with reductions in the risk weights we requested in our comment on the
advance notice of proposed rulemaking relating to the U.S.
implementation of the Accord (the "ANPR"), a copy of which we submit
herewith.2 Capital would then be determined by applying the appropriate
credit conversion factor, which we discuss below, to the applicable risk
weight from the ratings table.
C. Appropriate Conversion Factors
Whether or not the Agencies are willing
to permit an internal bank rating approach at this time, we believe that
the proposed 20% conversion factor is too high in light of the risks of
these positions. As discussed in part 3(C) of our comment on the ANPR,
we strongly feel that the asset quality tests present in liquidity
facilities, coupled with the presence of significant risk-mitigating
protections inherent in the underlying transactions which together
provide a conduit sponsor with the ability to actively manage a
transaction, significantly reduce the level of exposure by a liquidity
bank to the risks in the related portfolio.
The utilization history of liquidity
commitments (including parallel purchase commitments) of the conduits
administered by 17 banks participating in the preparation of this
comment supports the argument in favor of a lower conversion factor for liquidity commitments. The
conduits administered by these banks issue approximately 80% of all
multi-seller conduit ABCP outstanding as of September 30, 2003. The
results of the survey conducted by Mayer, Brown, Rowe & Maw are set
forth below.
• Conduits for which information was
reported have been in operation for periods ranging from 0.5 years to 20
years, with the mean period of operation being 10.4 years.
• These conduits have funded
securitization transactions with an aggregate principal balance of
$886.9 billion.
• For all transactions, only 148
liquidity draws have been made, for an aggregate amount of $12.1
billion.
• The aggregate amount of drawn
commitments represented 1.36% of the aggregate amount of funding for the
receivables pools.
• Only $593 million in losses have been
experienced on liquidity draws in transactions for which the responding
banks act as sponsor of a conduit, representing approximately 0.067% of
all originations of those banks.
• Annualizing the cumulative loss
percentage by dividing it by the average operating history of the
surveyed conduits results in an annual loss percentage of approximately
0.0064%.
The annualized loss percentage is
equivalent to that for a AAA exposure. Thus, although many conduits
include transactions on average structured to the so called "A" level,
the performance under related liquidity exposures is significantly
better.
Given that the proposed capital charge
against liquidity will put United States banks at a competitive
disadvantage vis a vis foreign banks as well as other funding sources
within the U.S., the need for an appropriately calibrated credit
conversion factor is heightened during this interim period prior to the
adoption of the Accord. Additionally, in order to preserve the
availability of an important risk dispersing technique, namely the
syndication of portions of a liquidity facility to third party
providers, the capital charge for these positions cannot exceed what
banks generally hold as economic capital against a position. While a
sponsor that provides liquidity has a broader interest in an underlying
transaction, the third party liquidity provider's interest is limited to
the particular transaction for which it provides liquidity support. If
that position becomes uneconomic to take on, the liquidity syndication
market is likely to disappear.
We believe that a conversion factor of
between 5 to 10% is the appropriate level to set for minimum capital
requirements for these very safe positions.3
D. Definition of Eligible Liquidity
Facilities
In addition to our concerns over the
appropriate credit conversion factor for liquidity facilities, we have
several issues with the eligibility requirements for liquidity. First,
while a 60 day delinquency standard may be appropriate in the context of
some trade receivables transactions, it is not appropriate in all such
transactions or for other asset classes. For example, a credit card
transaction might have a 120 day delinquency standard. Rather than a
"one-size fits all" definition, we believe it is more appropriate to tie
the limitation to not permitting funding against assets that would be
characterized as "defaulted" by the bank. We note that this is the
standard currently proposed both in the Accord and in the proposed U.S.
implementation of the Accord and see no justification for a different
standard during the interim. Second, we do not believe the limitation
that prohibits draws under the facility that supports a rated security
when that security's rating falls below investment grade is appropriate.
First, if the rating is not the asset quality test used for a liquidity
facility, it is irrelevant to assuring that liquidity positions do not
fund against bad assets. For example, if the relevant asset quality test
reduces the purchase price paid under a liquidity facility
dollar-for-dollar for the amount of defaulted receivables being funded,
whether that transaction is rated AAA or BB, the liquidity position is
protected from funding bad assets. Second, even when a rating is >sed as
the asset quality test for a liquidity facility, the proposed limitation
is unnecessary-as the risk of a position increases more capital will be
required to be held against related exposures. For example, for a
$10,000,000 liquidity position, the required capital under the current
proposal would increase from $32,000 (8% x 20% x 20% x $10,000,000) to
$320,000 (8% x 20% x 200% x $10,000,000) if the rating of the underlying
transaction fell from AAA to BB.
If the Agencies were unwilling to
eliminate investment grade funding threshold for rated securities in its
entirety, we suggest that (i) it should only be applicable in
transactions where the asset quality test ties to a rating of an
exposure (or guarantor), (ii) when it is applicable, the requirement
that liquidity should not fund against defaulted assets should not apply
(as it is irrelevant to the ratings asset quality test), (iii) a more
appropriate trigger should be when a position falls below BB, given that
the average rating of corporate loans held by United States banks is in
the area of BB and (iv) rather than eliminating from
"eligible liquidity facilities" those that permit funding below the
specified trigger, a more appropriate limitation would be to eliminate
the credit conversion factor from the calculation of required capital
for the related liquidity facility once the rating fell below the
specified trigger.
E. Early Amortization Capital
Requirements
If the Agencies were unwilling to delay
the implementation of the capital requirements for revolving retail
transactions until the implementation of the Accord in the U.S., we have
several comments on the proposal set forth in the Proposed Permanent
Rule. First, we believe that the Agencies must recognize and establish
an alternative approach for controlled early amortization transactions
similar to the approach specified in the proposed Accord. Unlike the
proposal in the Accord, however, banks should be able to utilize this
approach so long as they can meet certain more limited and objective,
principles-based criteria. To meet the necessary conditions for
"controlled early amortization" an originator should be required to show
only that: (i) the period for amortization is sufficient for 90% of the
total debt outstanding at the beginning of the amortization period to be
repaid or recognized as in default and (ii) the amortization occurs at a
pace no more rapid than a straight-line amortization.
The Agencies should be clear that the
amortization requirements would apply only to economic pay-out events
and not normal amortization or accumulation periods. The early
amortization capital charge represents a new capital requirement
specifically targeting the credit and liquidity risks associated with
early amortization events - when things go bad. As a result, the
amortization requirements should only apply to the specific economic
early amortization risk. During normal amortization periods, the loans,
by definition, are performing well and liquidity requirements are
incorporated into the bank's liquidity planning process.
Second, we note that while the proposed
amortization rules make sense in the credit card context, it is not
clear that the same application should be used across the board for
other revolving retail assets, For example, some securitizations early
amortization provisions are linked to the size of the
overcollateralization in a transaction. Therefore, the appropriate
triggers in those securitizations should be to the level of
overcollateralization rather than the level of excess spread. The rules
for amortization provisions should provide regulators with sufficient
flexibility to apply appropriate modifications to the amortization rules
when the context requires.
Third, we recommend a simplification of
the conversion factor early amortization capital requirement that would
make implementation much easier and that would prevent the unintended result of
incentivizing a bank to establish lower triggers to avoid capital
charges. The methodology should use the lesser of 4%, or the point at
which the organization would be required to begin trapping excess spread
as the starting reference point. This would allow for broad consistency
across the industry, with four, simple 1% quadrants. This would also
help the test be more operational for originators and verifiable for
examiners. Slight variances in the starting point for trapping excess
spread are not uncommon and not necessarily indicative of significant
risk differentiation in the underlying assets. You will find that
originators may have different spread triggers for transactions from the
same asset pool. A standard starting reference point will make it much
easier for originators to implement without sacrificing much from a risk
perspective. Conversely, the proposal as drafted gives a bank an
incentive to establish lower triggers for trapping excess cash to avoid
the early amortization capital charges associated with higher triggers.
We believe that prudent risk mitigation should drive the establishment
of appropriate triggers, not minimum capital requirements. We also
believe it is important to allow flexibility for non-credit card asset
types to have excess spread start points less than 4% if they can be
justified.
Finally, we recommend conversion factors
for the segments for controlled amortization structures be the same as
proposed in the ANPR (0%, 1%, 2%, 20% and 40%). We also recommend a
reduction to the required conversion factors for non-controlled early
amortization risk. The following conservative conversion factors for
these early amortization structures: 0%, 2%, 4%, 40%, and 80% would
represent a more appropriate risk differential than those currently
proposed.
F. Technical Comments
1. Definition of ABCP Program.
We note that in both the Interim Final
Rule and the Proposed Permanent Rule there are differences between the
definition of an ABCP program by the OTS and OCC, on the one hand, and
the FDIC and FRB, on the other hand. The OTS and OCC have adopted a
definition that is sufficiently broad in scope to cover all types of
conduits affected by FIN 46-bankruptcy remote special purpose entities
that issue commercial paper to fund the assets held by that entity.
Rather than adopting this definition, the FRB and FDIC's proposed
definition describes the "typical" conduit, which is a multi-seller
conduit funding customers through loans or purchases of asset pools. It
is unclear from a definition that describes the typical conduit whether
the FRB and FDIC intended to exclude other types of conduits from
regulatory capital relief. While a vast majority of conduits are
structured as described by this definition, there are a number of
bankruptcy remote conduits, such as structured investment vehicles, which fund the
purchase of securities issued in the capital markets through the
issuance of ABCP. The distortive effects of FIN 46 consolidation apply
equally to all types of conduits. Therefore, we believe that it is
appropriate that the capital relief should apply to all conduits. We
request that the FDIC and FRB adopt the definition of an "asset-backed
commercial paper program" as defined by the OTS and OCC or, at a
minimum, clarify that the definition used in their proposals was not
meant to be exclusive of conduits that do not fall into the "typical"
structure.
2. Determining Risk Weight for Unrated
Transactions
If an internal bank rating approach is
not adopted, we believe that the Proposed Permanent Rule needs to
provide a mechanic for determining the risk weight for liquidity
facilities that support unrated transactions. Liquidity facilities
generally support a pool of receivables and related obligors. We believe
that the appropriate risk weight for the pool should be the weighted
average risk weight of the underlying obligors. This weighted average
risk would reflect the true risks in the portfolio.
3. Extension of Implementation Deadline
We ask that the April 1st deadline be
extended to at least one year past the adoption of the Proposed
Permanent Rules to permit any required changes in liquidity facilities
to be implemented as these facilities come up for renewal rather than
require a potentially conduit wide amendment process within the next
several months. This extension would also permit the adoption of the
internal bank rating approach that we advocate. At a minimum, we propose
that all existing liquidity facilities be deemed to be "eligible"
facilities until the earlier to occur of (i) an amendment to that
facility or (ii) the first renewal date for such facility following the
effective date of the new rules to allow for an orderly implementation
of the new requirements for liquidity facilities in the current market.
* * * *
We appreciate this opportunity to comment
on the these proposals.
Vernon H.C. Wright
Chairman, American Securitization Forum (MBNA America Bank)
Greg Medcraft
Deputy Chairman
American Securitization Forum
(Societe Generale Securities Group)
Dwight Jenkins
Executive Director
American Securitization Forum
Jason H.P. Kravitt
Secretary, American Securitization Forum
(Mayer, Brown, Rowe & Maw LLP)
______________________________________________
1 Interpretation No. 46, "Consolidation of
Variable Interest Entities" issued by the Financial Accounting Standards
Board (FASB) in January 2003.
2 See part I of our comment to the ANPR
for our discussion of the proposed ratings table.
3 We are asking for a revised conversion
factor along with revised risk weights with the goal of some combination
that results in an ultimate calculation of minimum capital that is
appropriate for a particular position.
