View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

l l★K

Federal Reserve Bank of Dallas
2200 N. PEARL ST.
DALLAS, TX 75201-2272

August 18, 2003

Notice 03-43

TO: The Chief Executive Officer of each
financial institution and others concerned
in the Eleventh Federal Reserve District
SUBJECT
Requests for Comment on the Proposed Implementation
of the New Basel Capital Accord
DETAILS
The four federal bank and thrift regulatory agencies have requested comment on two
interagency documents related to the proposed implementation of the new Basel Capital Accord
in the United States. Earlier this month, the agencies approved issuance of the two documents for
public comment.
The new Accord, which is being developed by the Basel Committee on Banking
Supervision, builds on and, for certain banks, would replace the Basel Capital Accord of 1988,
which is the framework for capital adequacy standards for large, internationally active banks and
the basis for the risk-based capital adequacy standards now in place for all U.S. banks and bank
holding companies.
The first document, an Advance Notice of Proposed Rulemaking, sets forth for public
comment the agencies’ current views on a proposed framework for implementing the revised
Capital Accord in the United States. Under the proposal, internationally active banks meeting
certain criteria would be subject to the advanced internal ratings-based approach for credit risk
and the advanced measurement approaches (AMA) for operational risk.
The second document contains two sections. The first section sets forth draft
supervisory guidance on internal ratings-based systems for corporate credits, and the second
describes draft supervisory expectations for operational risk management.

For additional copies, bankers and others are encouraged to use one of the following toll-free numbers in contacting the Federal
Reserve Bank of Dallas: Dallas Office (800) 333-4460; El Paso Branch Intrastate (800) 592-1631, Interstate (800) 351-1012;
Houston Branch Intrastate (800) 392-4162, Interstate (800) 221-0363; San Antonio Branch Intrastate (800) 292-5810.

-2The Board must receive comments by November 3, 2003. Please address comments
to Jennifer J. Johnson, Secretary, Board of Governors of the Federal Reserve System, 20th Street
and Constitution Avenue, N.W., Washington, DC 20551. However, because paper mail in the
Washington area and at the Board is subject to delay, please consider submitting your comments
electronically to regs.comments@federalreserve.gov.
All comments regarding the proposed framework for implementing the revised
Capital Accord should refer to Docket No. R–1154. All comments regarding the internal ratingsbased systems for corporate credits and the draft supervisory expectations for operational risk
management should refer to Docket No. OP–1153.
ATTACHMENT
A copy of the agencies’ notice as it appears on pages 45900–88, Vol. 68, No. 149 of
the Federal Register dated August 4, 2003, is attached.
MORE INFORMATION
For more information regarding the revised Capital Accord, please contact Barbara
Bouchard, Assistant Director, (202) 452-3072; David Adkins, Supervisory Financial Analyst,
(202) 452-5259; or Mark Van Der Weide, Counsel, (202) 452-2263, at the Board. For more
information regarding the corporate internal ratings guidance, please contact David Palmer,
Supervisory Financial Analyst, at the Board, (202) 452-2904. For more information regarding the
AMA guidance, please contact T. Kirk Odegard, Supervisory Financial Analyst, at the Board,
(202) 530-6225.
Paper copies of this notice or previous Federal Reserve Bank notices can be printed
from our web site at www.dallasfed.org/banking/notices/index.html.

Monday,
August 4, 2003

Part II

Department of the
Treasury
Office of the Comptroller of the
Currency
12 CFR Part 3

Federal Reserve System
12 CFR Parts 208 and 225

Federal Deposit Insurance
Corporation
12 CFR Part 325

Department of the Treasury
Office of Thrift Supervision
12 CFR Part 567
Risk-Based Capital Guidelines;
Implementation of New Basel Capital
Accord; Internal Ratings-Based Systems
for Corporate Credit and Operational
Risk Advanced Measurement Approaches
for Regulatory Capital; Proposed Rule and
Notice

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

PO 00000

Frm 00001

Fmt 4737

Sfmt 4737

E:\FR\FM\04AUP2.SGM

04AUP2

45900

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
Other banking organizations that meet
the criteria, standards, and requirements
also would be eligible to use the
advanced approaches. Under the
advanced approaches, banking
organizations would use internal
estimates of certain risk components as
key inputs in the determination of their
regulatory capital requirements.

DEPARTMENT OF THE TREASURY
Office of the Comptroller of the
Currency
12 CFR Part 3
[Docket No. 03–14]
RIN Number 1557–AC48

FEDERAL RESERVE SYSTEM

DATES: Comments must be received no
later than November 3, 2003.

12 CFR Parts 208 and 225

ADDRESSES:

[Regulations H and Y; Docket No. R–1154]

FEDERAL DEPOSIT INSURANCE
CORPORATION
12 CFR Part 325
RIN 3064–AC73

DEPARTMENT OF THE TREASURY
Office of Thrift Supervision
12 CFR Part 567
[No. 2003–27]
RIN 1550–AB56

Risk-Based Capital Guidelines;
Implementation of New Basel Capital
Accord
AGENCIES: Office of the Comptroller of
the Currency, Treasury; Board of
Governors of the Federal Reserve
System; Federal Deposit Insurance
Corporation; and Office of Thrift
Supervision, Treasury.
ACTION: Advance notice of proposed
rulemaking.
SUMMARY: The Office of the Comptroller
of the Currency (OCC), the Board of
Governors of the Federal Reserve
System (Board), the Federal Deposit
Insurance Corporation (FDIC), and the
Office of Thrift Supervision (OTS)
(collectively, the Agencies) are setting
forth for industry comment their current
views on a proposed framework for
implementing the New Basel Capital
Accord in the United States. In
particular, this advance notice of
proposed rulemaking (ANPR) describes
significant elements of the Advanced
Internal Ratings-Based approach for
credit risk and the Advanced
Measurement Approaches for
operational risk (together, the advanced
approaches). The ANPR specifies
criteria that would be used to determine
banking organizations that would be
required to use the advanced
approaches, subject to meeting certain
qualifying criteria, supervisory
standards, and disclosure requirements.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

Comments should be
directed to: OCC: Please direct your
comments to: 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; fax number (202)
874–4448; or Internet address:
regs.comments@occ.treas.gov. Due to
delays in paper mail delivery in the
Washington area, we encourage the
submission of comments by fax or email whenever possible. Comments may
be inspected and photocopied at the
OCC’s Public Information Room, 250 E
Street, SW., Washington, DC. You may
make an appointment to inspect
comments by calling (202) 874–5043.
Board: Comments should refer to
Docket No. R–1154 and may be mailed
to Ms. Jennifer J. Johnson, Secretary,
Board of Governors of the Federal
Reserve System, 20th Street and
Constitution Avenue, NW., Washington,
DC 20551. However, because paper mail
in the Washington area and at the Board
of Governors is subject to delay, please
consider submitting your comments by
e-mail to
regs.comments@federalreserve.gov., or
faxing them to the Office of the
Secretary at (202) 452–3819 or (202)
452–3102. Members of the public may
inspect comments in Room MP–500 of
the Martin Building between 9 a.m. and
5 p.m. weekdays pursuant to § 261.12,
except as provided by § 261.14, of the
Board’s Rules Regarding Availability of
Information, 12 CFR 261.12 and 261.14.
FDIC: Written comments should be
addressed to Robert E. Feldman,
Executive Secretary, Attention:
Comments, Federal Deposit Insurance
Corporation, 550 17th Street, NW.,
Washington, DC 20429. Commenters are
encouraged to submit comments by
facsimile transmission to (202) 898–
3838 or by electronic mail to
Comments@FDIC.gov. Comments also
may be hand-delivered to the guard
station at the rear of the 550 17th Street
Building (located on F Street) on
business days between 8:30 a.m. and 5
p.m. Comments may be inspected and
photocopied at the FDIC’s Public
Information Center, Room 100, 801 17th

PO 00000

Frm 00002

Fmt 4701

Sfmt 4702

Street, NW., Washington, DC between 9
a.m. and 4:30 p.m. on business days.
OTS: Send comments to Regulation
Comments, Chief Counsel’s Office,
Office of Thrift Supervision, 1700 G
Street, NW., Washington, DC 20552,
Attention: No. 2003–27. Delivery: Hand
deliver comments to the Guard’s desk,
east lobby entrance, 1700 G Street, NW.,
from 9 a.m. to 4 p.m. on business days,
Attention: Regulation Comments, Chief
Counsel’s Office, Attention: No. 2003–
27. Facsimiles: Send facsimile
transmissions to FAX Number (202)
906–6518, Attention: No. 2003–27. Email: Send e-mails to
regs.comments@ots.treas.gov, Attention:
No. 2003–27, and include your name
and telephone number. Due to
temporary disruptions in mail service in
the Washington, DC area, commenters
are encouraged to send comments by fax
or e-mail, if possible.
FOR FURTHER INFORMATION CONTACT:
OCC: Roger Tufts, Senior Economic
Advisor (202–874–4925 or
roger.tufts@occ.treas.gov), Tanya Smith,
Senior International Advisor (202–874–
4735 or tanya.smith@occ.treas.gov), or
Ron Shimabukuro, Counsel (202–874–
5090 or
ron.shimabukuro@occ.treas.gov).
Board: Barbara Bouchard, Assistant
Director (202/452–3072 or
barbara.bouchard@frb.gov), David
Adkins, Supervisory Financial Analyst
(202/452–5259 or
david.adkins@frb.gov), Division of
Banking Supervision and Regulation, or
Mark Van Der Weide, Counsel (202/
452–2263 or
mark.vanderweide@frb.gov), Legal
Division. For users of
Telecommunications Device for the Deaf
(‘‘TDD’’) only, contact 202/263–4869.
FDIC: Keith Ligon, Chief (202/898–
3618 or kligon@fdic.gov), Jason Cave,
Chief (202/898–3548 or jcave@fdic.gov),
Division of Supervision and Consumer
Protection, or Michael Phillips, Counsel
(202/898–3581 or mphillips@fdic.gov).
OTS: Michael D. Solomon, Senior
Program Manager for Capital Policy
(202/906–5654); David W. Riley, Project
Manager (202/906–6669), Supervision
Policy; or Teresa A. Scott, Counsel
(Banking and Finance) (202/906–6478),
Regulations and Legislation Division,
Office of the Chief Counsel, Office of
Thrift Supervision, 1700 G Street, NW.,
Washington, DC 20552.
SUPPLEMENTARY INFORMATION:
I. Executive Summary
A. Introduction
B. Overview of the New Accord
C. Overview of U.S. Implementation
The A–IRB Approach for Credit Risk
The AMA for Operational Risk
Other Considerations

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
D. Competitive Considerations
II. Application of the Advanced Approaches
in the United States
A. Threshold Criteria for Mandatory
Advanced Approach Organizations
Application of Advanced Approaches at
Individual Bank/Thrift Levels
U.S. Banking Subsidiaries of Foreign
Banking Organizations
B. Implementation for Advanced Approach
Organizations
C. Other Considerations
General Banks
Majority-Owned or Controlled Subsidiaries
Transitional Arrangements
III. Advanced Internal Ratings-Based
Approach (A–IRB)
A. Conceptual Overview
Expected Losses versus Unexpected Losses
B. A–IRB Capital Calculations
Wholesale Exposures: Definitions and
Inputs
Wholesale Exposures: Formulas
Wholesale Exposures: Other
Considerations
Retail Exposures: Definitions and Inputs
Retail Exposures: Formulas
A–IRB: Other Considerations
Purchased Receivables
Credit Risk Mitigation Techniques
Equity Exposures
C. Supervisory Assessment of A–IRB
Framework
Overview of Supervisory Framework
U.S. Supervisory Review
IV. Securitization
A. General Framework
Operational Criteria
Differences Between the General A–IRB
Framework and the A–IRB Approach for
Securitization Exposures
B. Determining Capital Requirements
General Considerations
Capital Calculation Approaches
Other Considerations
V. AMA Framework for Operational Risk
A. AMA Capital Calculation
Overview of the Supervisory Criteria
B. Elements of an AMA Framework
VI. Disclosure
A. Overview
B. Disclosure Requirements
VII. Regulatory Analysis
A. Executive Order 12866
B. Regulatory Flexibility Act
C. Unfunded Mandates Reform Act of 1995
D. Paperwork Reduction Act
List of Acronyms

current U.S. risk-based capital
requirements are based on an
internationally agreed framework for
capital measurement that was
developed by the Basel Committee on
Banking Supervision (Basel Supervisors
Committee or BSC) and endorsed by the
G–10 Governors in 1988.2 The
international framework (1988 Accord)
accomplished several important
objectives. It strengthened capital levels
at large, internationally active banks and
fostered international consistency and
coordination. The 1988 Accord also
reduced disincentives for banks to hold
liquid, low-risk assets. Moreover, by
requiring banks to hold capital against
off-balance-sheet exposures, the 1988
Accord represented a significant step
forward for regulatory capital
measurement.
Although the 1988 Accord has been a
stabilizing force for the international
banking system, the world financial
system has become increasingly more
complex over the past fifteen years. The
BSC has been working for several years
to develop a new regulatory capital
framework that recognizes new
developments in financial products,
incorporates advances in risk
measurement and management
practices, and more precisely assesses
capital charges in relation to risk. On
April 29, 2003, the BSC released for
public consultation a document entitled
‘‘The New Basel Capital Accord’’ (New
Accord) that sets forth proposed
revisions to the 1988 Accord. The BSC
will accept industry comment on the
New Accord through July 31, 2003 and
expects to issue a final revised Accord
by the end of 2003. The BSC expects
that the New Accord would have an
effective date for implementation of
December 31, 2006.
Accordingly, the Agencies are
soliciting comment on all aspects of this
ANPR, which is based on certain
proposals in the New Accord.
Comments will assist the Agencies in

I. Executive Summary

regulatory capital as a percentage of both on- and
off-balance-sheet credit exposures with some gross
differentiation based on perceived credit risk. The
Agencies’ capital rules may be found at 12 CFR Part
3 (OCC), 12 CFR Parts 208 and 225 (Board), 12 CFR
Part 325 (FDIC), and 12 CFR Part 567 (OTS).
2 The BSC was established in 1974 by the centralbank governors of the Group of Ten (G–10)
countries. Countries are represented on the BSC by
their central bank and also by authorities with bank
supervisory responsibilities. Current member
countries are Belgium, Canada, France, Germany,
Italy, Japan, Luxembourg, the Netherlands, Spain,
Sweden, Switzerland, the United Kingdom, and the
United States. The 1988 Accord is described in a
document entitled ‘‘International Convergence of
Capital Measurement and Capital Standards.’’ This
document and other documents issued by the BSC
are available through the Bank for International
Settlements website at www.bis.org.

A. Introduction
In the United States, banks, thrifts,
and bank holding companies (banking
organizations or institutions) are subject
to minimum regulatory capital
requirements. Specifically, U.S. banking
organizations must maintain a
minimum leverage ratio and two
minimum risk-based ratios.1 The
1 The leverage ratio measures regulatory capital as
a percentage of total on-balance-sheet assets as
reported in accordance with generally accepted
accounting principles (GAAP) (with certain
adjustments). The risk-based ratios measure

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

PO 00000

Frm 00003

Fmt 4701

Sfmt 4702

45901

reaching a determination on a number
of issues related to how the New Accord
would be proposed to be implemented
in the United States. In addition, in light
of the public comments submitted on
the ANPR, the Agencies will seek
appropriate modifications to the New
Accord.
B. Overview of the New Accord
The New Accord encompasses three
pillars: minimum regulatory capital
requirements, supervisory review, and
market discipline. Under the first pillar,
a banking organization must calculate
capital requirements for exposure to
both credit risk and operational risk
(and market risk for institutions with
significant trading activity). The New
Accord does not change the definition
of what qualifies as regulatory capital,
the minimum risk-based capital ratio, or
the methodology for determining capital
charges for market risk. The New
Accord provides several methodologies
for determining capital requirements for
both credit and operational risk. For
credit risk there are two general
approaches; the standardized approach
(essentially a package of modifications
to the 1988 Accord) and the internal
ratings-based (IRB) approach (which
uses an institution’s internal estimates
of key risk drivers to derive capital
requirements). Within the IRB approach
there is a foundation methodology, in
which certain risk component inputs are
provided by supervisors and others are
supplied by the institutions, and an
advanced methodology (A–IRB), where
institutions themselves provide more
risk inputs.
The New Accord provides three
methodologies for determining capital
requirements for operational risk; the
basic indicator approach, the
standardized approach, and the
advanced measurement approaches
(AMA). Under the first two
methodologies, capital requirements for
operational risk are fixed percentages of
specified, objective risk measures (for
example, gross income). The AMA
provides the flexibility for an institution
to develop its own individualized
approach for measuring operational risk,
subject to supervisory oversight.
The second pillar of the New Accord,
supervisory review, highlights the need
for banking organizations to assess their
capital adequacy positions relative to
overall risk (rather than solely to the
minimum capital requirement), and the
need for supervisors to review and take
appropriate actions in response to those
assessments. The third pillar of the New
Accord imposes public disclosure
requirements on institutions that are
intended to allow market participants to

E:\FR\FM\04AUP2.SGM

04AUP2

45902

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

assess key information about an
institution’s risk profile and its
associated level of capital.
The Agencies do not expect the
implementation of the New Accord to
result in a significant decrease in
aggregate capital requirements for the
U.S. banking system. Individual banking
organizations may, however, face
increases or decreases in their minimum
risk-based capital requirements because
the New Accord is more risk sensitive
than the 1988 Accord and the Agencies’
existing risk-based capital rules (general
risk-based capital rules). The Agencies
will continue to analyze the potential
impact of the New Accord on both
systemic and individual bank capital
levels.
C. Overview of U.S. Implementation
The Agencies believe that the
advanced risk and capital measurement
methodologies of the New Accord are
the most appropriate approaches for
large, internationally active banking
organizations. As a result, large,
internationally active banking
organizations in the United States
would be required to use the A–IRB
approach to credit risk and the AMA to
operational risk. The Agencies are
proposing to identify three types of
banking organizations: institutions
subject to the advanced approaches on
a mandatory basis (core banks);
institutions not subject to the advanced
approaches on a mandatory basis, but
that choose voluntarily to apply those
approaches (opt-in banks); and
institutions that are not mandatorily
subject to and do not apply the
advanced approaches (general banks).
Core banks would be those with total
banking (and thrift) assets of $250
billion or more or total on-balance-sheet
foreign exposure of $10 billion or more.
Both core banks and opt-in banks
(advanced approach banks) would be
required to meet certain infrastructure
requirements (including complying with
specified supervisory standards for
credit risk and operational risk) and
make specified public disclosures before
being able to use the advanced
approaches for risk-based regulatory
capital calculation purposes.3
General banks would continue to
apply the general risk-based capital
rules. Because the general risk-based
capital rules include a buffer for risks
not easily quantified (for example,
operational risk and concentration risk),
general banks would not be subject to an
3 The

Agencies continue to reserve the right to
require higher minimum capital levels for
individual institutions, on a case-by-case basis, if
necessary to address particular circumstances.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

additional direct capital charge for
operational risk.
Under this proposal, some U.S.
banking organizations would use the
advanced approaches while others
would apply the general risk-based
capital rules. As a result, the United
States would have a bifurcated
regulatory capital framework. That is,
U.S. capital rules would provide two
distinct methodologies for institutions
to calculate risk-weighted assets (the
denominator of the risk-based capital
ratios). Under the proposed framework,
all U.S. institutions would continue to
calculate regulatory capital, the
numerator of the risk-based capital
ratios, as they do now. Importantly, U.S.
banking organizations would continue
to be subject to a leverage ratio
requirement under existing regulations,
and Prompt Corrective Action (PCA)
legislation and implementing
regulations would remain in effect.4 It is
recognized that in some cases, under the
proposed framework, the leverage ratio
would serve as the most binding
regulatory capital constraint.
Implementing the capital framework
described in this ANPR would raise a
number of significant practical and
conceptual issues about the role of
economic capital calculations relative to
regulatory capital requirements. The
capital formulas described in this
ANPR, as well as the economic capital
models used by banking organizations,
assume the ability to assign precisely
probabilities to future credit and
operational losses that might occur. The
term ‘‘economic capital’’ is often used to
refer to the amount of capital that
should be allocated to an activity
according to the results of such an
exercise. For example, a banking
organization might compute the amount
of income, reserves, and capital that it
would need to cover the 99.9th
percentile of possible credit losses
associated with a given type of lending.
The desired degree of certainty of
covering losses is related to several
factors including, for example, the
banking organization’s target credit
rating. The higher the loss percentile the
institution wishes to provide protection
against, the less likely the capital held
by the institution would be insufficient
to cover losses, and the higher would be
the institution’s credit rating.
While the Agencies intend to move to
a framework where regulatory capital is
more closely aligned to economic
capital, the Agencies do not intend to
4 Thus, for example, to be in the well-capitalized
PCA category a bank must have at least a 10 percent
total risk-based capital ratio, a 6 percent Tier I riskbased capital ratio, and a 5 percent leverage ratio.
The other PCA categories also would not change.

PO 00000

Frm 00004

Fmt 4701

Sfmt 4702

place sole reliance on the results of
economic capital calculations for
purposes of computing minimum
regulatory capital requirements.
Banking organizations face risks other
than credit and operational risks, and
the assumed loss distributions
underlying banking organizations’
economic capital calculations are
subject to the risk of error.
Consequently, the Agencies continue to
view the leverage ratio tripwires
contained in existing PCA and other
regulations as important components of
the regulatory capital framework.
The A–IRB Approach for Credit Risk
Under the A–IRB approach for credit
risk, an institution’s internal assessment
of key risk drivers for a particular
exposure (or pool of exposures) would
serve as the primary inputs in the
calculation of the institution’s minimum
risk-based capital requirements.
Formulas, or risk weight functions,
specified by supervisors would use the
banking organization’s estimated inputs
to derive a specific dollar amount
capital requirement for each exposure
(or pool of exposures). This dollar
capital requirement would be converted
into a risk-weighted assets equivalent by
multiplying the dollar amount of the
capital requirement by 12.5—the
reciprocal of the 8 percent minimum
risk-based capital requirement.
Generally, banking organizations using
the A–IRB approach would assign assets
and off-balance-sheet exposures into
one of three portfolios: wholesale
(corporate, interbank, and sovereign),
retail (residential mortgage, qualifying
revolving, and other), and equities.
There also would be specific treatments
for securitization exposures and
purchased receivables. Certain assets
that do not constitute a direct credit
exposure (for example, premises,
equipment, or mortgage servicing rights)
would continue to be subject to the
general risk-based capital rules and risk
weighted at 100 percent. A brief
overview of each A–IRB portfolio
follows.
Wholesale (Corporate, Interbank, and
Sovereign) Exposures
Wholesale credit exposures comprise
three types of exposures: corporate,
interbank, and sovereign. Generally, the
meaning of interbank and sovereign
would be consistent with the general
risk-based capital rules. Corporate
exposures are exposures to privatesector companies; interbank exposures
are primarily exposures to banks and
securities firms; and sovereign
exposures are those to central
governments, central banks, and certain

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
other public-sector entities (PSEs).
Within the wholesale exposure category,
in addition to the treatment for general
corporate lending, there would be four
sub-categories of specialized lending
(SL). These are project finance (PF),
object finance (OF), commodities
finance (CF), and commercial real estate
(CRE). CRE is further subdivided into
low-asset-correlation CRE, and highvolatility CRE (HVCRE).
For each wholesale exposure, an
institution would assign four
quantitative risk drivers (inputs): (1)
Probability of default (PD), which
measures the likelihood that the
borrower will default over a given time
horizon; (2) loss given default (LGD),
which measures the proportion of the
exposure that will be lost if a default
occurs; (3) exposure at default (EAD),
which is the estimated amount owed to
the institution at the time of default; and
(4) maturity (M), which measures the
remaining economic maturity of the
exposure. Institutions generally would
be able to take into account credit risk
mitigation techniques (CRM), such as
collateral and guarantees (subject to
certain criteria), by adjusting their
estimates for PD or LGD. The wholesale
A–IRB risk weight function would use
all four risk inputs to produce a specific
capital requirement for each wholesale
exposure. There would be a separate,
more conservative risk weight function
for certain acquisition, development,
and construction loans (ADC) in the
HVCRE category.
Retail Exposures
Within the retail category, three
distinct risk weight functions are
proposed for three product areas that
exhibit different historical loss
experiences and different asset
correlations.5 The three retail subcategories would be: (1) Exposures
secured by residential mortgages and
related exposures; (2) qualifying
revolving exposures (QRE); and (3) other
retail exposures. QRE would include
unsecured revolving credits (such as
credit cards and overdraft lines), and
other retail would include most other
types of exposures to individuals, as
well as certain exposures to small
businesses. The key inputs to the three
retail risk weight functions would be a
banking organization’s estimates of PD,
LGD, and EAD. There would be no
explicit M component to the retail A–
IRB risk weight functions. Unlike
5 Asset correlation is a measure of the tendency
for the financial condition of a borrower in a
banking organization’s portfolio to improve or
degrade at the same time as the financial condition
of other borrowers in the portfolio improve or
degrade.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

wholesale exposures, for retail
exposures, an institution would assign a
common set of inputs (PD, LGD, and
EAD) to predetermined pools of
exposures, which are typically referred
to as segments, rather than to individual
exposures.6 The inputs would be used
in the risk weight functions to produce
a capital charge for the associated pool
of exposures.
Equity Exposures
Banking organizations would use a
market-based internal model for
determining capital requirements for
equity exposures in the banking book.
The internal model approach would
assess capital based on an estimate of
loss under extreme market conditions.
Some equity exposures, such as
holdings in entities whose debt
obligations qualify for a zero percent
risk weight, would continue to receive
a zero percent risk weight under the A–
IRB approach to equities. Certain other
equity exposures, such as those made
through a small business investment
company (SBIC) under the Small
Business Investment Act or a
community development corporation
(CDC) or a community and economic
development entity (CEDE), generally
would be risk weighted at 100 percent
under the A–IRB approach to equities.
Banking organizations that are subject to
the Agencies’ market risk capital rules
would continue to apply those rules to
assess capital against equity positions
held in the trading book.7 Banking
organizations that are not subject to the
market risk capital rules would treat
equity positions in the trading account
as if they were in the banking book.
Securitization Exposures
Under the A–IRB treatment for
securitization exposures, a banking
organization that originates a
securitization would first calculate the
A–IRB capital charge that would have
been assessed against the underlying
exposures as if the exposures had not
been securitized. This capital charge
divided by the size of the exposure pool
6 When the PD, LGD, and EAD parameters are
assigned separately to individual exposures, it may
be referred to as a ‘‘bottom-up’’ approach. When
those parameters are assigned to predetermined sets
of exposures (pools or segments), it may be referred
to as a ‘‘top-down’’ approach.
7 The market risk capital rules were implemented
by the banking agencies in 1996. The market risk
capital rules apply to any banking organization
whose trading activity (on a consolidated
worldwide basis) equals 10 percent or more of total
assets, or $1 billion or more. The market risk capital
rules are found at 12 CFR Part 3, Appendix B
(OCC), 12 CFR Parts 208 and 225, Appendix E
(Board), and 12 CFR Part 325, Appendix C (FDIC).
The OTS, to date, has not adopted the market risk
capital rules.

PO 00000

Frm 00005

Fmt 4701

Sfmt 4702

45903

is referred to as KIRB. If an originating
banking organization retains a position
in a securitization that obligates the
banking organization to absorb losses up
to or less than KIRB, the banking
organization would deduct the retained
position from capital as is currently
required under the general risk-based
capital rules. The general risk-based
capital rules, however, require a dollarfor-dollar risk-based capital deduction
for certain residual interests retained by
originating banking organizations in
asset securitization transactions
regardless of amount. The A–IRB
framework would no longer require
automatic deduction of such residual
interests. The amount to be deducted
would be capped at KIRB for most
exposures. For a position in excess of
the KIRB threshold, the originating
banking organization would use an
external-ratings-based approach (if the
position has been rated by an external
rating agency or a rating can be inferred)
or a supervisory formula to determine
the capital charge for the position.
Non-originating banking organizations
that invest in a securitization exposure
generally would use an external-ratingsbased approach (if the exposure has
been rated by an external rating agency
or a rating can be inferred). For unrated
liquidity facilities that banking
organizations provide to securitizations,
capital requirements would be based on
several factors, including the asset
quality of the underlying pool and the
degree to which other credit
enhancements are available. These
factors would be used as inputs to a
supervisory formula. Under the A–IRB
approach to securitization exposures,
banking organizations also would be
required in some cases to hold
regulatory capital against securitizations
of revolving exposures that have early
amortization features.
Purchased Receivables
Purchased receivables, that is, those
that are purchased from another
institution either through a one-off
transaction or as part of an ongoing
program, would be subject to a two-part
capital charge: one part is for the credit
risk arising from the underlying
receivables and the second part is for
dilution risk. Dilution risk refers to the
possibility that contractual amounts
payable by the underlying obligors on
the receivables may be reduced through
future cash payments or other credits to
the accounts made by the seller of the
receivables. The framework for
determining the capital charge for credit
risk permits a purchasing organization
to use a top-down (pool) approach to
estimating PDs and LGDs when the

E:\FR\FM\04AUP2.SGM

04AUP2

45904

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

purchasing organization is unable to
assign an internal risk rating to each of
the purchased accounts. The capital
charge for dilution risk would be
calculated using the wholesale risk
weight function with some additional
specified risk inputs.
The AMA for Operational Risk
Under the A–IRB approach, capital
charges for credit risk would be directly
calibrated solely for such risk and, thus,
unlike the 1988 Accord, would not
implicitly include a charge for
operational risk. As a result, the
Agencies are proposing that banking
organizations operating under the A–
IRB approach also would have to hold
regulatory capital for exposure to
operational risk. The Agencies are
proposing to define operational risk as
the risk of losses resulting from
inadequate or failed internal processes,
people, and systems, or external events.
Under the AMA, each banking
organization would be able to use its
own methodology for assessing
exposure to operational risk, provided
the methodology is comprehensive and
results in a capital charge that is
reflective of the operational risk
experience of the organization. The
operational risk exposure would be
multiplied by 12.5 to determine a riskweighted assets equivalent, which
would be added to the comparable
amounts for credit and market risk in
the denominator of the risk-based
capital ratios. The Agencies will be
working closely with institutions over
the next few years as operational risk
measurement and management
techniques continue to evolve.
Other Considerations
Boundary Issues
With the introduction of an explicit
regulatory capital charge for operational
risk, an issue arises about the proper
treatment of losses that can be attributed
to more than one risk factor. For
example, where a loan defaults and the
banking organization discovers that the
collateral for the loan was not properly
secured, the banking organization’s
resulting losses would be attributable to
both credit and operational risk. The
Agencies recognize that these types of
boundary issues are important and have
significant implications for how banking
organizations would compile loss data
sets and compute regulatory capital
charges.
The Agencies are proposing the
following standard to govern the
boundary between credit and
operational risk: A loss event that has
characteristics of credit risk would be

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

incorporated into the credit risk
calculations for regulatory capital (and
would not be incorporated into
operational risk capital calculations).
This would include credit-related fraud
losses. Thus, in the above example, the
loss from the loan would be attributed
to credit risk (not operational risk) for
regulatory capital purposes. This
separation between credit and
operational risk is supported by current
U.S. accounting standards for the
treatment of credit risks.
With regard to the boundary between
the trading book and the banking book,
for institutions subject to the market risk
rules, positions currently subject to
those rules include all positions held in
the trading account consistent with
GAAP. The New Accord proposed
additional criteria for positions
includable in the trading book for
purposes of market risk capital
requirements. The Agencies encourage
comment on these additional criteria
and whether the Agencies should
consider adopting such criteria (in
addition to the GAAP criteria) in
defining the trading book under the
Agencies’ market risk capital rules. The
Agencies are seeking comment on the
proposed treatment of the boundaries
between credit, operational, and market
risk.
Supervisory Considerations
The advanced approaches introduce
greater complexity to the regulatory
capital framework and would require a
high level of sophistication in the
banking organizations that implement
the advanced approaches. As a result,
the Agencies propose to require core
and opt-in banks to meet certain
infrastructure requirements and comply
with specific supervisory standards for
credit risk and for operational risk. In
addition, banking organizations would
have to satisfy a set of public disclosure
requirements as a prerequisite for
approval to using the advanced
approaches. Supervisory guidance for
each credit risk portfolio type, as well
as for operational risk, is being
developed to ensure a sufficient degree
of consistency within the supervisory
framework, while also recognizing that
internal systems will differ between
banking organizations. The goal is to
establish a supervisory framework
within which all institutions must
develop their internal systems, leaving
exact details to each institution. In the
case of operational risk in particular, the
Agencies recognize that measurement
methodologies are still evolving and
flexibility is needed.
It is important to note that supervisors
would not look at compliance with

PO 00000

Frm 00006

Fmt 4701

Sfmt 4702

requirements, or standards alone.
Supervisors also would evaluate
whether the components of an
institution’s advanced approaches are
consistent with the overall objective of
sound risk management and
measurement. An institution would
have to use appropriately the advanced
approaches across all material business
lines, portfolios, and geographic regions.
Exposures in non-significant business
units as well as asset classes that are
immaterial in terms of size and
perceived risk profile may be exempted
from the advanced approaches with
supervisory approval. These immaterial
portfolios would be subject to the
general risk-based capital rules.
Proposed supervisory guidance for
corporate credit exposures and for
operational risk is provided separately
from this ANPR in today’s Federal
Register. The draft supervisory guidance
for corporate credit exposures is entitled
‘‘Supervisory Guidance on InternalRatings-Based Systems for Corporate
Credit.’’ The guidance includes
specified supervisory standards that an
institution’s internal rating system for
corporate exposures would have to
satisfy for the institution to be eligible
to use the A–IRB approach for credit
risk. The draft operational risk guidance
is entitled ‘‘Supervisory Guidance on
Operational Risk Advanced
Measurement Approaches for
Regulatory Capital.’’ The operational
risk guidance includes identified
supervisory standards for an
institution’s AMA framework for
operational risk. The Agencies
encourage commenters to review and
comment on the draft guidance pieces
in conjunction with this ANPR. The
Agencies intend to issue for public
comment supervisory guidance on retail
credit exposures, equity exposures, and
securitization exposures over the next
several months.
Supervisory Review
As mentioned above, the second pillar
of the New Accord focuses on
supervisory review to ensure that an
institution holds sufficient capital given
its overall risk profile. The concepts of
Pillar 2 are not new to U.S. banking
organizations. U.S. institutions already
are required to hold capital sufficient to
meet their risk profiles, and supervisors
may require that an institution hold
more capital if its current levels are
deficient or some element of its business
practices suggest the need for more
capital. The Agencies also have the right
to intervene when capital levels fall to
an unacceptable level. Given these longstanding elements of the U.S.
supervisory framework, the Agencies

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
are not proposing to introduce specific
requirements or guidelines to
implement Pillar 2. Instead, existing
guidance, rules, and regulations would
continue to be enforced and
supplemented as necessary as part of
this proposed new regulatory capital
framework. However, all institutions
operating under the advanced
approaches would be expected by
supervisors to address specific
assumptions embedded in the advanced
approaches (such as diversification in
credit portfolios), and would be
evaluated for their ability to account for
deviations from the underlying
assumptions in their own portfolios.
Disclosure
An integral part of the advanced
approaches is enhanced public
disclosure practices and improved
transparency. Under the Agencies’
proposal, specific disclosure
requirements would be applicable to all
institutions using the advanced
approaches. These disclosure
requirements would encompass capital,
credit risk, equities, credit risk
mitigation, securitization, market risk,
operational risk, and interest rate risk in
the banking book.
D. Competitive Considerations
It is essential that the Agencies gain
a full appreciation of the possible
competitive equity concerns that may be
presented by the establishment of a new
capital framework. The creation of a
bifurcated capital framework in the
United States—one set of capital
standards applicable to large,
internationally active banking
organizations (and those that choose to
apply the advanced approaches), and
another set of standards applicable to all
other institutions—has created concerns
among some parties about the potential
impact on competitive equity between
the two sets of banking organizations.
Similarly, differences in supervisory
application of the advanced approaches
(both within the United States and
abroad) among large, internationally
active institutions may pose competitive
equity issues among such institutions.
The New Accord relies upon
compliance with certain minimum
operational and supervisory
requirements to promote consistent
interpretation and uniformity in
application of the advanced approaches.
Nevertheless, independent supervisory
judgment will be applied on a case-bycase basis. These processes, albeit
subject to detailed and explicit
supervisory guidance, contain an
inherent amount of subjectivity and
must be assessed by supervisors on an

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

ongoing basis. This supervisory
assessment of the internal processes and
controls leading to an institution’s
internal ratings and other estimates
must maintain the high level of internal
risk measurement and management
processes contemplated in this ANPR.
The BSC’s Accord Implementation
Group (AIG), in which the Agencies
play an active role, will seek to ensure
that all jurisdictions uniformly apply
the same high qualitative and
quantitative standards to internationally
active banking institutions. However, to
the extent that different supervisory
regimes implement these standards
differently, there may be competitive
dislocations. One concern is that the
U.S. supervisory regime will impose
greater scrutiny in its implementation
standards, particularly given the
extensive on-site presence of bank
examiners in the United States.
Quite distinct from the need for a
level playing field among
internationally active institutions are
the competitive concerns of those
institutions that do not elect to adopt or
may not qualify for the advanced
approaches. Some banking
organizations have expressed concerns
that small or regional banks would
become more likely to be acquired by
larger organizations seeking to lever
capital efficiencies. There also is a
qualitative concern about the impact of
being considered a ‘‘second tier’’
institution (one that does not implement
the advanced approaches) by the
market, rating agencies, or sophisticated
customers such as government or
municipal depositors and borrowers.
Finally, there is the question of what, if
any, competitive distortions might be
introduced by differences in regulatory
capital minimums between the
advanced approaches and the general
risk-based capital rules for loans or
securities with otherwise similar risk
characteristics, and the extent to which
such distortions may be mitigated in an
environment in which well-managed
banking organizations continue to hold
excess capital.8
Because the advanced framework
described in this ANPR is more risksensitive than the 1988 Accord and the
general risk-based capital rules, banking
organizations under the advanced
approaches would face increases in
8 The Agencies note that under the general riskbased capital rules some institutions currently are
able to hold less capital than others on some types
of assets (for example, through innovative financing
structures or use of credit risk mitigation
techniques). In addition, some institutions may
hold lower amounts of capital because the market
perceives them as highly diversified, while others
hold higher amounts of capital because of
concentrations of credit risk or other factors.

PO 00000

Frm 00007

Fmt 4701

Sfmt 4702

45905

their minimum risk-based capital
charges on some assets and decreases on
others. The results of a Quantitative
Impact Study (QIS3) the BSC conducted
in late 2002 indicated the potential for
the advanced approaches described in
this document to produce significant
changes in risk-based capital
requirements for specific activities; the
results also varied on an institution-byinstitution basis. The results of QIS3 can
be found at http://www.bis.org and
various results of QIS3 are noted at
pertinent places in this ANPR.
The Agencies do not believe the
results of QIS3 are sufficiently reliable
to form the basis of a competitive
impact analysis, both because the inputs
to the study were provided on a bestefforts basis and because the proposals
in this ANPR are in some cases different
than those that formed the basis of QIS3.
The Agencies are nevertheless
interested in views on how changes in
regulatory capital (for the total of credit
and operational risk) of the magnitude
described in QIS3, if such changes were
in fact realized, would affect the
competitive landscape for domestic
banking organizations.
The Agencies plan to conduct at least
one more QIS, and potentially other
economic impact analyses, to better
understand the potential impact of the
proposed framework on the capital
requirements for individual U.S.
banking organizations and U.S. banking
organizations as a whole. This may
affect the Agencies’ further proposals
through recalibrating the A–IRB risk
weight formulas and making other
modifications to the proposed
approaches if the capital requirements
do not seem consistent with the overall
risk profiles of banking organizations or
safe and sound banking practices.
If competitive effects of the New
Accord are determined to be significant,
the Agencies would need to consider
potential ways to address those effects
while continuing to seek to achieve the
objectives of the current proposal.
Alternatives could potentially include
modifications to the proposed
approaches, as well as fundamentally
different approaches. The Agencies
recognize that an optimal capital system
must strike a balance between the
objectives of simplicity and regulatory
consistency across banking
organizations on the one hand, and the
degree of risk sensitivity of the
regulation on the other. There are many
criteria that must be evaluated in
achieving this balance, including the
resulting incentives for improving risk
measurement and management
practices, the ease of supervisory and
regulatory enforcement, the degree to

E:\FR\FM\04AUP2.SGM

04AUP2

45906

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

which the overall level of regulatory
capital in the banking system is broadly
preserved, and the effects on domestic
and international competition. The
Agencies are interested in commenters’
views on alternatives to the advanced
approaches that could achieve this
balance, and in particular on
alternatives that could do so without a
bifurcated approach.9
The Agencies are committed to
investigate the full scope of possible
competitive impact and welcome all
comments in this regard. Some
questions are suggested below that may
serve to focus commenters’ general
reactions. More specific questions also
are suggested throughout this ANPR.
These questions should not be viewed
as limiting the Agencies’ areas of
interest or commenters’ submissions on
the proposals. The Agencies encourage
commenters to provide supporting data
and analysis, if available.
What are commenters’ views on the
relative pros and cons of a bifurcated
regulatory capital framework versus a single
regulatory capital framework? Would a
bifurcated approach lead to an increase in
industry consolidation? Why or why not?
What are the competitive implications for
community and mid-size regional banks?
Would institutions outside of the core group
be compelled for competitive reasons to optin to the advanced approaches? Under what
circumstances might this occur and what are
the implications? What are the competitive
implications of continuing to operate under
a regulatory capital framework that is not risk
sensitive?
If regulatory minimum capital
requirements declined under the advanced
approaches, would the dollar amount of
capital held by advanced approach banking
organizations also be expected to decline? To
the extent that advanced approach
institutions have lower capital charges on
certain assets, how probable and significant
are concerns that those institutions would
realize competitive benefits in terms of
pricing credit, enhanced returns on equity,
and potentially higher risk-based capital
ratios? To what extent do similar effects
already exist under the current general riskbased capital rules (for example, through
securitization or other techniques that lower
relative capital charges on particular assets
for only some institutions)? If they do exist
now, what is the evidence of competitive
harm?
Apart from the approaches described in
this ANPR, are there other regulatory capital
approaches that are capable of ameliorating
competitive concerns while at the same time
9 In this regard, alternative approaches would
take time to develop, but might present fewer
implementation challenges. Additional work would
be necessary to advance the goal of competitive
equity among internationally active banking
organizations. If consensus on alternative
approaches could not be reached at the BSC, a
departure from the Basel framework also could raise
significant international and domestic issues.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

achieving the goal of better matching
regulatory capital to economic risks? Are
there specific modifications to the proposed
approaches or to the general risk-based
capital rules that the Agencies should
consider?

II. Application of the Advanced
Approaches in the United States
By its terms, the 1988 Accord applied
only to internationally active banks.
Under the New Accord, the scope of
application has been broadened also to
encompass bank holding companies that
are parents of internationally active
‘‘banking groups.’’
A. Threshold Criteria for Mandatory
Advanced Approach Organizations
The Agencies believe that for large,
internationally active U.S. institutions
only the advanced approaches are
appropriate. Accordingly, the Agencies
intend to identify three groups of
banking organizations: (1) Large,
internationally active banking
organizations that would be subject to
the A–IRB approach and AMA on a
mandatory basis (core banks); (2)
organizations not subject to the
advanced approaches on a mandatory
basis, but that voluntarily choose to
adopt those approaches (opt-in banks);
and all remaining organizations that are
not mandatorily subject to and do not
apply the advanced approaches (general
banks).
For purposes of identifying core
banks, the Agencies are proposing a set
of objective criteria for industry
consideration. Specifically, the
Agencies are proposing to treat as a core
bank any banking organization that has
(1) total commercial bank (and thrift)
assets of $250 billion or more, as
reported on year-end regulatory reports
(with banking assets of consolidated
groups aggregated at the U.S. bank
holding company level); 10 or (2) total
on-balance-sheet foreign exposure of
$10 billion or more, as reported on the
year-end Country Exposure Report
(FFIEC 009) (with foreign exposure of
consolidated groups aggregated at the
U.S. bank holding company level).
These threshold criteria are
independent; meeting either condition
would mean an institution is a core
bank.
Once an institution becomes a core
bank it would remain subject to the
advanced approaches on a going
forward basis. If, in subsequent years,
such an institution were to drop below
both threshold levels it would continue
10 For banks this means the December
Consolidated Report of Condition and Income (Call
Report). For thrifts this means the December Thrift
Financial Report.

PO 00000

Frm 00008

Fmt 4701

Sfmt 4702

to be a core bank unless it could
demonstrate to its primary Federal
supervisor that it has substantially and
permanently downsized and should no
longer be a core bank. The Agencies are
proposing an annual test for assessing
banking organizations in reference to
the threshold levels. However, as a
banking organization approaches either
of the threshold levels the Agencies
would expect to have ongoing dialogue
with that organization to ensure that
appropriate practices are in place or are
actively being developed to prepare the
organization for implementation of the
advanced approaches.
Institutions that by expansion or
merger meet the threshold levels must
qualify for use of the advanced
approaches and would be subject to the
same implementation plan requirements
and minimum risk-based capital floors
applicable to core and opt-in banks as
described below. Institutions that seek
to become opt-in banks would be
expected to notify their primary Federal
supervisors well in advance of the date
by which they expect to qualify for the
advanced approaches. Based on the
aforementioned threshold levels, the
Agencies anticipate at this time that
approximately ten U.S. institutions
would be core banks.
Application of Advanced Approaches at
Individual Bank/Thrift Levels
The Agencies are aware that some
institutions might, on a consolidated
basis, exceed one of the threshold levels
for mandatory application of the A–IRB
approach and AMA and, yet, might be
comprised of distinct bank and thrift
charters whose respective sizes fall well
below the thresholds. In those cases, the
Agencies believe that all bank and thrift
institutions that are members of a
consolidated group that is itself a core
bank or an opt-in bank should calculate
and report their risk-based capital
requirements under the advanced
approaches. However, recognizing that
separate bank and thrift charters may, to
a large extent, be independently
managed and have different systems and
portfolios, the Agencies are interested in
comment on the efficacy and burden of
a framework that requires the advanced
approaches to be implemented by (or
pushed down to) each of the separate
subsidiary banks and thrifts that make
up the consolidated group.
U.S. Banking Subsidiaries of Foreign
Banking Organizations
Any U.S. bank or thrift that is a
subsidiary of a foreign bank would have
to comply with the prevailing U.S.
regulatory capital requirements applied
to U.S. banks. Thus, if a U.S. bank or

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
thrift that is owned by a foreign bank
meets the threshold levels for
mandatory application of the advanced
approaches, the U.S. bank or thrift
would be a core bank. If it does not meet
those thresholds, it would have the
choice to opt-in to the advanced
approaches (and be subject to the same
supervisory framework as other U.S.
banking organizations) or to remain a
general bank. A top-tier U.S. bank
holding company that is owned by a
foreign bank also would be subject to
the same threshold levels for core bank
determination and would be subject to
the applicable U.S. bank holding
company capital rules. However,
Federal Reserve SR Letter 01–1 (January
5, 2001) would remain in effect. Thus,
subject to the conditions in SR Letter
01–1, a top-tier U.S. bank holding
company that is owned or controlled by
a foreign bank that is a qualifying
financial holding company generally
would not be required to comply with
the Board’s capital adequacy guidelines.
The Agencies are interested in comment on
the extent to which alternative approaches to
regulatory capital that are implemented
across national boundaries might create
burdensome implementation costs for the
U.S. subsidiaries of foreign banks.

B. Implementation for Advanced
Approach Organizations
As noted earlier, U.S. banking
organizations that apply the advanced
approaches would be required to
comply with supervisory standards
prior to use.
The BSC has targeted December 31,
2006 as the effective date for the
international capital rules based on the
New Accord. The Agencies are
proposing an implementation date of
January 1, 2007. The establishment of a
final effective date in the United States,
however, would be contingent on the
issuance for public comment of a Notice
of Proposed Rulemaking, and
subsequent finalization of any changes
in capital regulations that the Agencies
ultimately decide to adopt.
Because of the need to pre-qualify for
the advanced approaches, banking
organizations would need to take a
number of steps upon the finalization of
any changes to the capital regulations.
These steps would include developing
detailed written implementation plans
for the A–IRB approach and the AMA
and keeping their primary supervisors
advised of these implementation plans
and schedules. Implementation plans
would need to address all supervisory
standards for the A–IRB approach and
the AMA, include objectively
measurable milestones, and demonstrate
that adequate resources would be

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

realistically budgeted and made
available. An institution’s board of
directors would need to approve its
implementation plans.
The Agencies expect core banks to
make every effort to meet the
supervisory standards as soon as
practicable. In this regard, it is possible
that some core banks would not qualify
to use the advanced approaches in time
to meet the effective date that is
ultimately established. For those
banking organizations, the
implementation plan would need to
identify when the supervisory standards
would be met and when the institution
would be ready for implementation. The
Agencies note that developing an
appropriate infrastructure to support the
advanced approaches for regulatory
capital that fully complies with
supervisory conditions and expectations
and the associated supervisory guidance
will be challenging. The Agencies
believe, however, that institutions
would need to be fully prepared before
moving to the advanced approaches.
Use of the advanced approaches
would require the primary Federal
supervisor’s approval. Core banks
unable to qualify for the advanced
approaches in time to meet the effective
date would remain subject to the general
risk-based capital rules existing at that
time. The Agencies would consider the
effort and progress made to meet the
qualifying standards and would
consider whether, under the
circumstances, supervisory action
should be taken against or penalties
imposed on individual core banks that
have not adhered to the schedule
outlined in the implementation plan
they submitted to their primary Federal
supervisor.
Opt-in banks meeting the supervisory
standards could seek to qualify for the
advanced approaches in time to meet
the ultimate final effective date or any
time thereafter. Institutions
contemplating opting-in to the advanced
approaches would need to provide
notice to, and submit an
implementation plan and schedule to be
approved by, their primary Federal
supervisor. As is true of core banks, optin banks would need to allow ample
time for developing and executing
implementation plans.
An institution’s primary Federal
supervisor would have responsibility for
determining the institution’s readiness
for an advanced approach and would be
ultimately responsible, after
consultation with other relevant
supervisors, for determining whether
the institution satisfies the supervisory
expectations for the advanced
approaches. The Agencies recognize

PO 00000

Frm 00009

Fmt 4701

Sfmt 4702

45907

that a consistent and transparent
process to oversee implementation of
the advanced approaches would be
crucial. The Agencies intend to develop
interagency validation standards and
procedures to help ensure consistency.
The Agencies would consult with each
other on significant issues raised during
the validation process and ongoing
implementation.
C. Other Considerations
General Banks
The Agencies expect that the vast
majority of U.S. institutions would be
neither core banks nor opt-in banks.
Most institutions would remain subject
to the general risk-based capital rules.
However, as has been the case since the
1988 Accord was initially implemented
in the United States, the Agencies will
continue to make necessary
modifications to the general risk-based
capital rules as appropriate. In the event
changes are warranted, the Agencies
could implement revisions through
notice and comment procedures prior to
the proposed effective date of the
advanced approaches in 2007.
The Agencies seek comment on
whether changes should be made to the
existing general risk-based capital rules
to enhance their risk-sensitivity or to
reflect changes in the business lines or
activities of banking organizations
without imposing undue regulatory
burden or complication. In particular,
the Agencies seek comment on whether
any changes to the general risk-based
capital rules are necessary or warranted
to address any competitive equity
concerns associated with the bifurcated
framework.
Majority-Owned or Controlled
Subsidiaries
The New Accord generally applies to
internationally active banking
organizations on a fully consolidated
basis. Thus, consistent with the
Agencies’ general risk-based capital
rules, subsidiaries that are consolidated
under U.S. generally accepted
accounting principles (GAAP) typically
should be consolidated for regulatory
capital calculation purposes under the
advanced approaches as well.11 With
regard to investments in consolidated
insurance underwriting subsidiaries, the
New Accord notes that deconsolidation
of assets and deduction of capital is an
11 One notable exception exists at the bank level
where there is an investment in a financial
subsidiary as defined in the Gramm-Leach-Bliley
Act of 1999. For such a subsidiary, assets would
continue to be deconsolidated from the bank’s onbalance-sheet assets, and capital at the subsidiary
level would be deducted from the bank’s capital.

E:\FR\FM\04AUP2.SGM

04AUP2

45908

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

appropriate approach. The Federal
Reserve is actively considering several
approaches to the capital treatment for
investments by bank holding companies
in insurance underwriting subsidiaries.
For example, the Federal Reserve is
currently assessing the merits and
weaknesses of an approach that would
consolidate an insurance underwriting
subsidiary’s assets at the holding
company level and permit excess capital
of the subsidiary to be included in the
consolidated regulatory capital of the
holding company. A deduction would
be required for capital that is not readily
available at the holding company level
for general use throughout the
organization.
The Federal Reserve specifically seeks
comment on the appropriate regulatory
capital treatment for investments by bank
holding companies in insurance
underwriting subsidiaries as well as other
nonbank subsidiaries that are subject to
minimum regulatory capital requirements.

Transitional Arrangements
Core and opt-in banks would be
required to calculate their capital ratios
using the A-IRB and AMA
methodologies, as well as the general
risk-based capital rules, for one year
prior to using the advanced approaches
on a stand-alone basis. In order to begin
this parallel-run year, however, the
institution would have to demonstrate
to its supervisor that it meets the
supervisory standards. Therefore,
banking organizations planning to meet
the January 1, 2007 target effective date
for implementation of the advanced
approaches would have to receive
approval from their primary Federal
supervisor before year-end 2005.
Banking organizations that later adopt
the advanced approaches also would
have a one-year dual calculation period
prior to moving to stand-alone usage of
the advanced approaches.
An institution would be subject to a
minimum risk-based capital floor for
two years following moving to the
advanced approaches on a stand-alone
basis. Specifically, in the first year of
stand-alone usage of the advanced
approaches, an institution’s calculated
risk-weighted assets could not be less
than 90 percent of risk-weighted assets
calculated under the general risk-based
capital rules. In the following year, an
institution’s minimum calculated riskweighted assets could not be less than
80 percent of risk-weighted assets
calculated under the general risk-based
capital rules.12
12 The agencies note that the text above differs
from the floor text in the New Accord, which is
based on 90 and 80 percent of the minimum capital
requirements under the 1988 Accord, rather than on

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

As a consequence, advanced approach
banking organizations would need to
conduct two sets of capital calculations
for at least three years. The preimplementation calculation of A-IRB
and AMA capital would not need to be
made public, but the banking
organization would be required to
disclose risk-based capital ratios
calculated under both advanced and
general risk-based approaches during
the two-year post-implementation
period. The Agencies would not
propose to eliminate the floors after the
two-year transition period for any
institution applying the advanced
approaches until the Agencies are fully
satisfied that the institution’s systems
are sound and accurately assess risk and
that resulting capital levels are prudent.
These transitional arrangements and
the floors established above relate only
to risk-based capital ratios and do not
affect the continued applicability to all
advanced banking organizations of the
leverage ratio and associated PCA
regulations for banks and thrifts.
Importantly, the minimum capital
requirements and the PCA thresholds
would not be changed. Furthermore,
during the implementation period and
before removal of the floors the
Agencies intend to closely monitor the
effect that the advanced approaches
would have on capital levels at
individual institutions and industrywide capital levels. Once the results of
this monitoring process are assessed, the
Agencies may consider modifications to
the advanced approaches to ensure that
capital levels remain prudent.

Do the projected dates provide an adequate
timeframe for core banks to be ready to
implement the advanced approaches? What
other options should the Agencies consider?
The Agencies seek comment on
appropriate thresholds for determining
whether a portfolio, business line, or
geographic exposure would be material.
Considerations should include relative asset
size, percentages of capital, and associated
levels of risk for a given portfolio, business
line, or geographic region.

Given the general principle that the
advanced approaches are expected to be
implemented at the same time across all
material portfolios, business lines, and
geographic regions, to what degree should
the Agencies be concerned that, for example,
data may not be available for key portfolios,
business lines, or regions? Is there a need for
further transitional arrangements? Please be
specific, including suggested durations for
such transitions.

A. Conceptual Overview
The A–IRB framework has as its
conceptual foundation the belief that
any range of possible losses on a
portfolio of credit exposures can be
represented by a probability density
function (PDF) of possible losses over a
one-year time horizon. If known, the
parameters of a PDF can be used to
specify a particular level of capital that
will lower the probability of the
institution’s insolvency due to adverse
credit risk outcomes to a stated
confidence level. With a known or
estimated PDF, the probability of
insolvency can be measured or
estimated directly, based on the level of
reserves and capital available to an
institution.
The A–IRB framework builds off this
concept and reflects an effort to develop
a common set of risk-sensitive formulas
for the calculation of required capital for
credit risk. To a large extent, this
framework resembles more systematic
quantitative approaches to the

risk-weighted assets. The Agencies expect that the
final language of the New Accord would need to be
consistent with this approach. The following
example reflects how the floor in the first year
would be applied by a U.S. banking organizaiton.
If the banking organization’s general risk-based
capital calculation produced risk-weighted assets of
$100 billion in its first year of implementation of
the advanced approaches, then its risk weighted
assets in that year could not be less than $90 billion.
If the advanced approach calculation produced riskweighted assets of $75 billion (a decrease of one
quarter compared to the general risk-based capital
rules), the organization would not calculate riskbased capital ratios on the basis of that $75 billion;
rather, its risk-weighted assets would be $90 billion.
Consequently, its minimum total risk-based capital
charge would be $7.2 billion, and it would need $9
billion to satisfy PCA well-capitalized criteria.

PO 00000

Frm 00010

Fmt 4701

Sfmt 4702

III. Advanced Internal Ratings-Based
(A–IRB) Approach
This section describes the proposed
A–IRB framework for the measurement
of capital requirements for credit risk.
Under this framework, banking
organizations that meet the A–IRB
infrastructure requirements and
supervisory standards would
incorporate internal estimates of risk
inputs into supervisor-provided capital
formulas for the various debt and equity
portfolios to calculate the capital
requirements for each portfolio. The
discussion below provides background
on the conceptual basis of the A–IRB
approach and then describes the
specific details of the capital formulas
for two of the main exposure categories,
wholesale and retail. Separate sections
follow that describe the A–IRB
treatments of loan loss reserves and
partial charge-offs, the A–IRB treatment
of purchased receivables, the A–IRB
treatment of equity exposures, and the
A–IRB treatment of securitization
exposures. The A–IRB supervisory
requirements and the A–IRB approach
to credit risk mitigation techniques also
are discussed in separate sections.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
measurement of credit risk that many
banking organizations have been
developing. These approaches being
developed by banking organizations
generally rely on a statistical or
probability-based assessment of credit
risk and use inputs broadly similar to
those required under the A–IRB
approach. Like the value-at-risk (VaR)
model that forms the basis for the
market risk capital rules, the output of
these statistical approaches to credit risk
is typically an estimate of loss threshold
on a credit exposure or pool of credit
exposures that is highly unlikely to be
exceeded by actual credit-related losses
on the exposure or pool.
Many banking organizations now use
such a credit VaR amount as the basis
for an internal assessment of the
economic capital necessary to cover
credit risk. In this context, it is common
for banking organizations’ internal
credit risk models to consider a one-year
loss horizon, and to focus on a high loss
threshold confidence level (that is, a
loss threshold that has a small
probability of being exceeded), such as
the 99.95th percentile. This is because
banking organizations typically seek to
hold an amount of economic capital for
credit risk whose probability of being
exceeded is broadly consistent with the
institution’s external credit rating and
its associated default probability. For
example, the one-year historical
probability of default for AA-rated firms
is less than 5 basis points (0.05 percent).
There is a great deal of variation
across banking organizations in the
specifics of their credit risk
measurement approaches. It is
important to recognize that the A–IRB
approach is not intended to allow
banking organizations to use all aspects
of their own models to estimate
regulatory capital for credit risk. The A–
IRB approach has been developed as a
single, common methodology that all
advanced approach banking
organizations would use, and consists of
a set of formulas (or functions) and a
single set of assumptions regarding
critical parameters for the formulas. The
A–IRB approach draws on the same
conceptual underpinnings as the credit
VaR approaches that banking
organizations have developed
individually, but likely differs in many
specifics from the approach used by any
individual institution.
The specific A–IRB formulas require
the banking organization first to
estimate certain risk inputs, which the
organization may do using a variety of
techniques. The formulas themselves,
into which the estimated risk inputs are
inserted, are broadly consistent with the
most common statistical approaches for

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

measuring credit risk, but also are more
straightforward to calculate than those
typically employed by banking
organizations (which often require
computer simulations). In particular, an
important property of the A–IRB
formulas is portfolio invariance. That is,
the A–IRB capital requirement for a
particular exposure generally does not
depend on the other exposures held by
the banking organization; as with the
general risk-based capital rules, the total
credit risk capital requirement for a
banking organization is simply the sum
of the credit risk capital requirements
on individual exposures or pools of
exposures.13
As with the existing credit VaR
models, the output of the A–IRB
formulas is an estimate of the amount of
credit losses over a one-year period that
would only be exceeded a small
percentage of the time. In the case of the
A–IRB formulas, this nominal
confidence level is set to 99.9 percent.
This means that within the context of
the A–IRB modeling assumptions a
banking organization’s overall credit
portfolio capital requirement can be
thought of as an estimate of the 99.9th
percentile of potential losses on that
portfolio over a one-year period. In
practice, however, this 99.9 percent
nominal target likely overstates the
actual level of confidence because the
A–IRB framework does not explicitly
address portfolio concentration issues or
the possibility of errors in estimating
PDs, LGDs, or EADs. The choice of the
99.9th percentile reflects a desire on the
part of the Agencies to align the
regulatory capital standard with the
default probabilities typically associated
with maintaining low investment grade
ratings (that is, BBB) even in periods of
economic adversity and to ensure
neither a substantial increase or
decrease in overall required capital
levels among A–IRB banking
organizations compared with the capital
levels that would be required under the
general risk-based capital rules. It also
recognizes that the risk-based capital
rules count a broader range of
instruments as eligible capital (for
example, certain subordinated debt)
13 The theoretical underpinnings for obtaining
portfolio-invariant capital charges within credit
VaR models are provided in the paper ‘‘A RiskFactor Model Foundation for Ratings-Based Bank
Capital Rules,’’ by Michael Gordy, forthcoming in
the Journal of Financial Intermediation. The A–IRB
formulas are derived as an application of these
results to a single-factor CreditMetrics-style mode.
For mathematical details of this model, see M.
Gordy, ‘‘A comparative Anatomy of Credit Risk
Models.’’ Journal of Banking and Finance, January
2000, or H.R. Koyluogu and A. Hickman,
‘‘Reconcilable Differences.’’ Risk, October 1998.

PO 00000

Frm 00011

Fmt 4701

Sfmt 4702

45909

than do internal economic capital
methodologies.
Expected Losses Versus Unexpected
Losses
The diagram below shows a
hypothetical loss distribution for a
portfolio of credit exposures over a oneyear horizon. The loss distribution is
represented by the curve, and is drawn
in such a way that it depicts a higher
proportion of losses falling below the
mean value than falling above the mean.
The average value of credit losses is
referred to as expected loss (EL). The
losses that exceed the expected level are
labeled unexpected loss (UL). An
overarching policy question concerns
whether the proposed design of the A–
IRB capital requirements should reflect
an expectation that institutions would
allocate capital to cover both EL and a
substantial portion of the range of
possible UL outcomes, or only the UL
portion of the range of possible losses
(that is, from the EL point out to the
99.9th percentile).
The Agencies recognize that some
institutions, in their comment letters on
earlier BSC proposals and in discussion
with supervisory staffs, have
highlighted the view that regulatory
capital should not be allocated for EL.
They emphasize that EL is normally
incorporated into the interest rate and
spreads charged on specific products,
such that EL is covered by net interest
margin and provisioning. The
implication is that supervisors would
review provisioning policies and the
adequacy of reserves as part of a
supervisory review, much as they do
today, and would require additional
reserves and/or regulatory capital for EL
in cases where reserves were deemed
insufficient. However, the Agencies are
concerned that the accounting
definition of general reserves differs
significantly across countries, and that
banking practices with respect to the
recognition of impairment also are very
different. Thus, the Agencies are
proposing to include EL in the
calibration of the risk weight functions.
The Agencies also note that the
current regulatory definition of capital
includes a portion of general reserves.
That is, general reserves up to 1.25
percent of risk-weighted assets are
included in the Tier 2 portion of total
capital. If the risk weight functions were
calibrated solely to UL, it could be
argued that the definition of capital
would also need to be revisited. In the
United States, such a discussion would
require a review of the provisioning
practices of institutions under GAAP
and of the distinctions drawn between
specific and general provisions.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

The framework described in this
ANPR calibrates the risk-based capital
requirements to the sum of EL plus UL,
which raises significant calibration
issues. Those calibration issues would
be treated differently if the calibration
were based only on the estimate of UL.
That is, decisions with respect to
significant policy variables that are
described below hinge crucially on the
initial decision to base the calibration
on EL plus UL, rather than UL only.
These issues include, for example, the
appropriate mechanism for
incorporating any future margin income
(FMI) that is associated with particular
business lines, as well as the
appropriate method for incorporating
general and specific reserves into the
risk-based capital ratios.
A final overarching assumption of the
A–IRB framework is the role of asset
correlations. Within the A–IRB capital
formulas (as in the credit VaR models of
many banking organizations), asset
correlation parameters provide a
measure of the extent to which changes
in the economic value of separate
exposures are presumed to move
together. A higher asset correlation
between a particular asset and other
assets in the same portfolio implies a
greater likelihood that the asset will
decline in value at the same time as the
portfolio as a whole declines in value.
Because this means a greater chance that
the asset will be a contributor to high
loss scenarios, its capital requirement
under the A–IRB framework also is
higher.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

Specifically, the A–IRB capital
formulas described in detail below are
based on the assumption that
correlation in defaults across borrowers
is attributable to their common
dependence on one or more systematic
risk factors. The basis for this
assumption is the observation that a
banking organization’s borrowers are
generally susceptible to adverse changes
in the global economy. These systematic
factors are distinct from the borrowerspecific, or idiosyncratic, risk factors
that determine the probability that a
specific loan will be repaid. Like other
risk-factor models, the A–IRB
framework assumes that these borrowerspecific factors represent idiosyncratic
sources of risk, and thus (unlike the
systematic risk-factors) are diversified in
a large lending portfolio.
The A–IRB approach allows for much
improved sensitivity to many of the
loan-level determinants of economic
capital (such as PD and LGD), but does
not explicitly address how an
exposure’s economic capital might vary
with the degree of concentration in the
overall portfolio to specific industries or
regions, or even to specific borrowers.
That is, it neither rewards nor penalizes
differences across banking organizations
in diversification or concentration
across industry, geography, and names.
To introduce such rewards and
penalties in an appropriate manner
would necessarily entail far greater
operational complexity for both
regulatory and financial institutions.

PO 00000

Frm 00012

Fmt 4701

Sfmt 4702

In contrast, the portfolio models of
credit risk employed by many banking
organizations are quite sensitive to all
forms of diversification. That is, the
economic capital charge assigned to a
loan within such a model will depend
on the portfolio as a whole. In order to
apply a portfolio model to the
calibration of A–IRB capital charges, it
would be necessary to identify the
assumptions needed so that a portfolio
model would yield economic capital
charges that do not depend on portfolio
characteristics. Recent advances in the
finance literature demonstrate that
economic capital charges are portfolioinvariant if (and only if) two
assumptions are imposed.14 First, the
portfolio must be infinitely fine-grained.
Second, there must be only a single
systematic risk factor.
Infinite granularity, while never
literally attained, is satisfied in an
approximate sense by the portfolios of
large, internationally active banks.
Analysis of data provided by such
institutions shows that taking account of
single-name concentrations in such
portfolios would lead to only trivial
changes in the total capital requirement.
The single risk-factor assumption would
appear, at first glance, more
troublesome. As an empirical matter,
there undoubtedly are distinct cyclical
factors for different industries and
different geographic regions. From a
substantive perspective, however, the
14 See forthcoming paper by M. Gordy referenced
in footnot number 12 above.

E:\FR\FM\04AUP2.SGM

04AUP2

EP04AU03.000</GPH><Q P=’03’>

45910

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
relevant question is whether portfolios
at large financial institutions are
diversified across the various subsectors of the economy in a reasonably
similar manner. If so, then the portfolio
can be modeled as if there were only a
single factor, namely, the credit cycle as
a whole.
The Agencies seek comment on the
conceptual basis of the A–IRB approach,
including all of the aspects just described.
What are the advantages and disadvantages
of the A–IRB approach relative to
alternatives, including those that would
allow greater flexibility to use internal
models and those that would be more
cautious in incorporating statistical
techniques (such as greater use of credit
ratings by external rating agencies)? The
Agencies also encourage comment on the
extent to which the necessary conditions of
the conceptual justification for the A–IRB
approach are reasonably met, and if not, what
adjustments or alternative approach would
be warranted.
Should the A–IRB capital regime be based
on a framework that allocates capital to EL
plus UL, or to UL only? Which approach
would more closely align the regulatory
framework to the internal capital allocation
techniques currently used by large
institutions? If the framework were
recalibrated solely to UL, modifications to
the rest of the A–IRB framework would be
required. The Agencies seek commenters’
views on issues that would arise as a result
of such recalibration.

B. A–IRB Capital Calculations
A common characteristic of the A–IRB
capital formulas is that they calculate
the actual dollar value of the minimum
capital requirement associated with an
exposure (or, in the case of retail
exposures, a pool of exposures). This
capital requirement must be converted
to an equivalent amount of riskweighted assets in order to be inserted
into the denominator of a banking
organization’s risk-based capital ratios.
Because the minimum risk-based capital
ratio in the United States is 8 percent,
the minimum capital requirement on
any asset would be equal to 8 percent
of the risk-weighted asset amount
associated with that asset. Therefore, in
order to determine the amount of riskweighted assets to associate with a given
minimum capital requirement, it would
be necessary to multiply the dollar
capital requirement generated by the A–
IRB formulas by the reciprocal of 8
percent, or 12.5.
The following subsections of the
ANPR detail the specific features of the
A–IRB capital formulas for two
principal categories of credit exposure:
wholesale and retail. Both of these
subsections include a proposed
definition of the exposure category, a
description of the banking organization-

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

estimated inputs required to complete
the capital calculations, a description of
the specific calculations required to
determine the A–IRB capital
requirement, and tables depicting a
range of representative results.
Wholesale Exposures: Definitions and
Inputs
The Agencies propose that a single
credit exposure category—wholesale
exposures—would encompass most
non-retail credit exposures in the A–IRB
framework. The wholesale category
would include the sub-categories of
corporate, sovereign, and interbank
exposures as well as all types of
specialized lending exposures.
Wholesale exposures would include
debt obligations of corporations,
partnerships, limited liability
companies, proprietorships, and
special-purpose entities (including
those created specifically to finance
and/or operate physical assets).
Wholesale exposures also would
include debt obligations of banks and
securities firms (interbank exposures),
and debt obligations of central
governments, central banks, and certain
public-sector entities (sovereign
exposures). The wholesale exposure
category would not include
securitization exposures, or certain
small-business exposures that are
eligible to be treated as retail exposures.
The Agencies propose that advanced
approach banking organizations would
use the same A–IRB capital formula to
compute capital requirements on all
wholesale exposures with two
exceptions. First, wholesale exposures
to small- and medium-sized enterprises
(SMEs) would use a downward
adjustment to the wholesale A–IRB
capital formula typically based on
borrower size. Second, the A–IRB
capital formula for HVCRE loans
(generally encompassing certain
speculative ADC loans) would use a
higher asset correlation assumption than
other wholesale exposures.
The proposed A–IRB capital
framework for wholesale exposures
would require banking organizations to
assign four key risk inputs for each
individual wholesale exposure: (1)
Probability of default (PD); (2) loss given
default (LGD); (3) exposure at default
(EAD); and (4) effective remaining
maturity (M). In addition, to use the
proposed downward adjustment for
wholesale SMEs described in more
detail below, banking organizations
would be required to provide an
additional input for borrower size (S).

PO 00000

Frm 00013

Fmt 4701

Sfmt 4702

45911

Probability of Default
The first principal input to the
wholesale A–IRB calculation is the
measure of PD. Under the A–IRB
approach, a banking organization would
assign an internal rating to each of its
wholesale obligors (or in other words,
assign each wholesale exposure to an
internal rating grade applicable to the
obligor). The internal rating would have
to be produced by a rating system that
meets the A–IRB infrastructure
requirements and supervisory standards
for wholesale exposures, which are
intended to ensure (among other things)
that the rating system results in a
meaningful differentiation of risk among
exposures. For each internal rating, the
banking organization must associate a
specific one-year PD value. Various
approaches may be used to develop
estimates of PDs; however, regardless of
the specific approach, banking
organizations would be expected to
satisfy the supervisory standards. The
minimum PD that may be assigned to
most wholesale exposures is 3 basis
points (0.03 percent). Certain wholesale
exposures are exempt from this floor,
including exposures to sovereign
governments, their central banks, the
BIS, IMF, European Central Bank, and
high quality multilateral development
banks (MDBs) with strong shareholder
support.
The Agencies intend to apply
standards to the PD quantification
process that are consistent with the
broad guidance outlined in the New
Accord. More detailed discussion of
those points is provided in the draft
supervisory guidance on IRB
approaches for corporate exposures
published elsewhere in today’s Federal
Register.
Loss Given Default
The second principal input to the A–
IRB capital formula for wholesale
exposures is LGD. Under the A–IRB
approach, banking organizations would
estimate an LGD for each wholesale
exposure. An LGD estimate for a
wholesale exposure should provide an
assessment of the expected loss in the
event of default of the obligor, expressed
as a percentage of the institution’s
estimated total exposure at default. The
LGD for a defaulted exposure would be
estimated as the expected economic loss
rate on that exposure taking into
account, where appropriate, recoveries,
workout costs, and the time value of
money. Banking organizations would
estimate LGDs as the loss severities
expected to prevail when default rates
are high, unless they have information
indicating that recoveries on a particular

E:\FR\FM\04AUP2.SGM

04AUP2

45912

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

class of exposure are unlikely to be
affected to an appreciable extent by
cyclical factors. As with estimates of
other A–IRB inputs, banking
organizations would be expected to be
conservative in assigning LGDs.
Although estimated LGDs should be
grounded in historical recovery rates,
the A–IRB approach is structured to
allow banking organizations to assess
the differential impact of various
factors, including, for example, the
presence of collateral or differences in
loan terms and covenants. The Agencies
expect to impose limitations on the use
of guarantees and credit derivatives in a
banking organization’s LGD estimates.
These limitations are discussed in the
separate section of this ANPR on the A–
IRB treatment of credit risk mitigation
techniques.
Exposure at Default

15 Under the add-on approach, an institution
would determine its EAD for an OTC derivative
contract by adding the current value of the contract
(zero if the current value is negative) and an
estimate of potential future exposure (PFE) on the
contract. The estimated PFE would be equal to the
notional amount of the derivative multiplied by a
supervisor-provided add-on factor that takes into
account the type of instrument and its maturity.
16 Repo-style transactions include reverse
repurchase agreements and repurchase agreements
and securities lending and borrowing.

20:44 Aug 01, 2003

A banking organization would
estimate inputs relative to the following
definition of default and loss. A default
is considered to have occurred with
respect to a particular borrower when
either or both of the following two
events has taken place: (1) The banking
organization determines that the
borrower is unlikely to pay its
obligations to the organization in full,
without recourse to actions by the
organization such as the realization of
collateral; or (2) the borrower is more
than 90 days past due on principal or
interest on any material obligation to the
organization. The Agencies believe that
the use of the concept of ‘‘unlikely to
pay’’ is largely consistent with the
practice of U.S. banking organizations in
assessing whether a loan is on nonaccrual status.
Maturity

The third principal input to the
wholesale A–IRB capital formula is
EAD. The Agencies are proposing that
banking organizations would provide
their own estimate of EAD for each
exposure. The EAD for an exposure
would be defined as the amount legally
owed to the banking organization (net of
any charge-offs) in the event that the
borrower defaults on the exposure. For
on-balance-sheet items, banking
organizations would estimate EAD as no
less than the current drawn amount. For
off-balance-sheet items, except over-thecounter (OTC) derivative transactions,
banking organizations would assign an
EAD equal to an estimate of the long-run
default-weighted average EAD for
similar facilities and borrowers or, if
EADs are highly cyclical, the EAD
expected to prevail when default rates
are high. The EAD associated with OTC
derivative transactions would continue
to be estimated using the ‘‘add-on’’
approach contained in the general riskbased capital rules.15 In addition, there
would be a specific EAD calculation for
the recognition of collateral in the
context of repo-style transactions
subject to a master netting agreement,
the features of which are outlined below
in the section on the A–IRB treatment of
credit risk mitigation techniques.16

VerDate jul<14>2003

Definition of Default and Loss

Jkt 200001

The fourth principal input to the A–
IRB capital formula is effective
remaining maturity (M), measured in
years. If a wholesale exposure is subject
to a determinable cash flow schedule,
the banking organization would
calculate M as the weighted-average
remaining maturity of the expected cash
flows, using the amounts of the cash
flows as the relevant weights. The
banking organization also would be able
to use the nominal remaining maturity
of the exposure if the weighted-average
remaining maturity of the exposure
cannot be calculated. For OTC
derivatives and repo-style transactions
subject to master netting agreements, the
institution would set M equal to the
weighted-average remaining maturity of
the individual transactions, using the
notional amounts of the individual
transactions as the relevant weights.
In all cases, M would be set no greater
than five years and, with few
exceptions, M would be set no lower
than one year. The exceptions apply to
certain transactions that are not part of
a banking organization’s ongoing
financing of a borrower. For wholesale
exposures that have an original maturity
of less than three months—including
repo-style transactions, money market
transactions, trade finance-related
transactions, and exposures arising from
payment and settlement processes—M
may be set as low as one day. For OTC
derivatives and repo-style transactions
subject to a master netting agreement, M
would be set at no less than five days.
As with the assignment of PD
estimates, the Agencies propose to
apply supervisory standards for the
estimation of LGD, EAD, and M that are
consistent with the broad guidance

PO 00000

Frm 00014

Fmt 4701

Sfmt 4702

contained in the New Accord. More
detailed discussion of these issues is
provided in the draft supervisory
guidance on IRB approaches for
corporate exposures published
elsewhere in today’s Federal Register.
The Agencies seek comment on the
proposed definition of wholesale exposures
and on the proposed inputs to the wholesale
A–IRB capital formulas. What are views on
the proposed definitions of default, PD, LGD,
EAD, and M? Are there specific issues with
the standards for the quantification of PD,
LGD, EAD, or M on which the Agencies
should focus?
Wholesale Exposures: Formulas
The calculation of the A–IRB capital
requirement for a particular wholesale
exposure would be accomplished in three
steps:
(1) Calculation of the relevant asset
correlation parameter, which would be a
function of PD (as well as borrower size (S)
for SMEs);
(2) Calculation of a preliminary capital
requirement assuming a maturity of one year,
which would be a function of PD, LGD, EAD,
and the asset correlation parameter
calculated in the first step; and
(3) Application of a maturity adjustment
for differences between the actual effective
remaining maturity of the exposure and the
one-year maturity assumption in the second
step, where the adjustment would be a
function of both PD and M.
These calculations result in the A–IRB
capital requirement, expressed in dollars, for
a particular wholesale exposure. As noted
earlier, this amount would be converted to a
risk-weighted assets equivalent by
multiplying the amount by 12.5, and the riskweighted assets equivalent would be
included in the denominator of the riskbased capital ratios.
Asset Correlation
The first step in the calculation of the A–
IRB capital requirement for a wholesale
exposure is the calculation of the asset
correlation parameter, which is denoted by
the letter ‘‘R’’ in the formulas below. This
asset correlation parameter is not a fixed
amount; rather, the parameter varies as an
inverse function of PD. For all wholesale
exposures except HVCRE exposures, the asset
correlation parameter approaches an upper
bound value of 24 percent for very low PD
values and approaches a lower bound value
of 12 percent for very high PD values. This
reflects the Agencies’ view that borrowers
with lower credit quality (that is, higher PDs)
are likely to be more idiosyncratic in the
factors affecting their likelihood of default
than borrowers with higher credit quality
(lower PDs). Therefore, the higher PD
borrowers are proportionately less influenced
by systematic (sector-wide or economy-wide)
factors common to all borrowers.17
An important practical impact of having
asset correlation decline with increases in PD
17 See Jose Lopez, ‘‘The Empirical Relationship
between Average Asset Correlation, Firm
Probability of Default, and Asset Size.’’ Federal
Reserve Bank of San Francisco Working Paper 02–
05 (June 2002).

E:\FR\FM\04AUP2.SGM

04AUP2

45913

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
is to reduce the speed with which capital
requirements increase as PDs increase, and to
increase the speed with which EL dominates
the total capital charge, thereby tending to
reduce procyclicality in the application of
the wholesale A–IRB capital formulas. The
specific formula for determining the asset
correlation parameter for all wholesale
exposures except HVCRE exposures is as
follows:
R = 0.12 * (1¥EXP(¥50 * PD)) + 0.24 *
[1¥(1¥EXP(¥50 * PD))]
Where:
R denotes asset correlation;
EXP(x) denotes the natural exponential
function; and
PD denotes probability of default.
Capital Requirement With Assumed OneYear Maturity Adjustment
The second step in the calculation of the
A–IRB capital requirement for a particular
wholesale exposure is the calculation of the
capital requirement that would apply to the
exposure assuming a one-year effective
remaining maturity. The specific formula to
calculate this one-year-maturity capital
requirement is as follows:
K1 = EAD * LGD * N[(1¥R)∧¥0.5 * G(PD)
+ (R/(1¥ R))∧0.5 * G(0.999)]
Where:
K1 denotes the one-year-maturity capital
requirement;
EAD denotes exposure at default;
LGD denotes loss given default;

N(x) denotes the standard normal cumulative
distribution function;
R denotes asset correlation;
G(x) denotes the inverse of the standard
normal cumulative distribution function;
and 18
PD denotes probability of default.
There are several important aspects of this
formula. First, it rises in a straight-line
fashion with increases in EAD, meaning that
a doubling of the exposure amount would
result in a doubling of the capital
requirement. It also rises in a straight-line
fashion with increases in LGD, which
similarly implies that a loan with an LGD
estimate twice that of an otherwise identical
loan would have twice the capital
requirement of the other loan. This also
implies that as LGD or EAD estimates
approach zero, the capital requirement would
likewise approach zero. The remainder of the
formula is a function of PD, asset correlation
(R), which is itself a function of PD, and the
target loss percentile amount of 99.9 percent
discussed earlier.
Maturity Adjustment
The third stage in the calculation of the A–
IRB capital requirement for a particular
wholesale exposure is the application of a
maturity adjustment to reflect the exposure’s
actual effective remaining maturity (M). The
A–IRB maturity adjustment multiplies the
one-year-maturity capital requirement (K1) by
a factor that depends on both M and PD. The
fact that the A–IRB maturity adjustment

depends on PD reflects the Agencies’ view
that there is a greater proportional need for
maturity adjustments for high-quality
exposures (those with low PDs) because there
is a greater potential for such exposures to
deteriorate in credit quality than for
exposures whose credit quality is lower. The
specific formula for applying the maturity
adjustment and generating the A–IRB capital
requirement is as follows:
K = K1 * [1 + (M¥2.5) * b]/[(1¥1.5 * b)],
where b = (0.08451¥0.05898 * LN(PD))2
and:
K denotes the A–IRB capital requirement;
K1 denotes the one-year-maturity capital
requirement;
M denotes effective remaining maturity;
LN(x) denotes the natural logarithm; and
PD denotes probability of default.
In this formula, the value ‘‘b’’ effectively
determines the slope of the maturity
adjustment and is itself a function of PD.
Note that if M is set equal to one, the
maturity adjustment also equals one and K
will therefore equal K1.
To provide a more concrete sense of the
range of capital requirements under the
wholesale A–IRB framework, the following
table presents the A–IRB capital
requirements (K) for a range of values of both
PD and M. In this table LGD is assumed to
equal 45 percent. For comparison purposes,
the general risk-based capital rules assign a
capital requirement of 8 percent for most
commercial loans.

CAPITAL REQUIREMENTS
[In percentage points]
Effective remaining maturity (M)
PD
1 month
0.05 percent .....................................................................................................
0.10 percent .....................................................................................................
0.25 percent .....................................................................................................
0.50 percent .....................................................................................................
1.00 percent .....................................................................................................
2.00 percent .....................................................................................................
5.00 percent .....................................................................................................
10.00 percent ...................................................................................................
20.00 percent ...................................................................................................

0.50
1.00
2.17
3.57
5.41
7.65
11.91
17.67
26.01

The impact of the A–IRB capital
formulas on minimum risk-based capital
requirements for wholesale exposures
would, of course, depend on the actual
values of PD, LGD, EAD, and M that
banking organizations would use as
inputs to the wholesale formulas.
Subject to the caveats noted earlier,
evidence from QIS3 suggested an
average reduction in credit risk capital
requirements for corporate exposures of
about 26 percent for twenty large U.S.
banking organizations.

SME Adjustment

18 The N(x) and G(x) functions are widely used in
statistics and are commonly available in computer

spreadsheet programs. A description of these
functions may be found in the Help function of

VerDate jul<14>2003

21:05 Aug 01, 2003

Jkt 200001

For loans to SMEs not eligible for
retail A–IRB treatment, the proposed
calculation of the A–IRB capital
requirement has one additional
element—a downward adjustment based
on borrower size (S). This adjustment
would effectively lower the A–IRB
capital requirement on wholesale
exposures to SMEs with annual sales (or
total assets) of less than $50 million.
The Agencies believe the measure of
borrower size should be based on
annual sales (rather than total assets),
unless the banking organization can

PO 00000

Frm 00015

Fmt 4701

Sfmt 4702

1 year
0.92
1.54
2.89
4.40
6.31
8.56
12.80
18.56
26.84

3 years

5 years

1.83
2.71
4.44
6.21
8.29
10.56
14.75
20.50
28.65

2.74
3.88
5.99
8.03
10.27
12.56
16.69
22.45
30.47

demonstrate that it would be more
appropriate for the banking organization
to use the total assets of the borrower as
its measure of borrower size. The
borrower size adjustment would be
made to the asset correlation parameter
(R), as shown in the following formula:
RSME = R¥0.04 * [1¥(S¥ 5)/45]
Where
RSME denotes the size-adjusted asset
correlation;
R denotes asset correlation; and

most spreadsheet programs or in basic statistical
textbooks.

E:\FR\FM\04AUP2.SGM

04AUP2

45914

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

S denotes borrower size (expressed in
millions of dollars).
The maximum reduction in the asset
correlation parameter based on this
formula is 4 percent, and is achieved
when borrower size is $5 million. For
all borrower sizes below $5 million,
borrower size would be set equal to $5
million. The adjustment shrinks to zero
as borrower size approaches $50
million. The broad rationale for this
adjustment is the view that the credit

condition of SMEs will be influenced
relatively more by idiosyncratic factors
than is the case for larger firms, and,
thus, SMEs would be less likely to
deteriorate simultaneously with other
exposures. This greater susceptibility to
idiosyncratic factors would imply lower
asset correlation. The evidence in favor
of this view is mixed, particularly after
considering that the A–IRB framework
already incorporates a negative
relationship between asset correlation
and PD. The following table illustrates

the practical effect of the SME
adjustment by depicting the capital
requirements (K) across a range of PDs
and borrower sizes. As in the previous
table, LGD is assumed to equal 45
percent. For this table, M is assumed to
be equal to three years. Note that the last
column is identical to the three-year
maturity column in the preceding table
because the SME adjustment is phased
out for borrowers of $50 million or more
in size.

CAPITAL REQUIREMENTS
[In percentage points]
Borrower size (S)
PD
$5 million
0.05 percent ...................................................................................................
0.10 percent ...................................................................................................
0.25 percent ...................................................................................................
0.50 percent ...................................................................................................
1.00 percent ...................................................................................................
2.00 percent ...................................................................................................
5.00 percent ...................................................................................................
10.00 percent .................................................................................................
20.00 percent .................................................................................................

Subject to the caveats mentioned
above, evidence from QIS3 suggested an
average reduction in credit risk-based
capital requirements for corporate SME
exposures of about 39 percent for
twenty large U.S. banking organizations.
If the Agencies include a SME adjustment,
are the $50 million threshold and the
proposed approach to measurement of
borrower size appropriate? What standards
should be applied to the borrower size
measurement (for example, frequency of
measurement, use of size buckets rather than
precise measurements)?
Does the proposed borrower size
adjustment add a meaningful element of risk
sensitivity sufficient to balance the costs
associated with its computation? The
Agencies are interested in comments on
whether it is necessary to include an SME
adjustment in the A–IRB approach. Data
supporting views is encouraged.
Wholesale Exposures: Other Considerations
Specialized Lending
The specialized lending (SL) asset class
encompasses exposures for which the
primary source of repayment is the income
generated by the specific asset(s) being
financed, rather than the financial capacity of
a broader commercial enterprise. The SL
category encompasses four broad exposure
types:
• Project finance (PF) exposures finance
large, complex, expensive installations that
produce goods or services for sale, such as
power plants, chemical processing plants,
mines, or transportation infrastructure, where
the source of repayment is primarily the
revenues generated by sale of the goods or
services by the installations.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

1.44
2.14
3.54
4.97
6.63
8.40
11.70
16.76
24.67

• Object finance (OF) exposures
finance the acquisition of (typically
moveable) physical assets, such as ships
or aircraft, where the source of
repayment is primarily the revenues
generated by the assets being financed,
often through rental or lease contracts
with third parties.
• Commodities finance (CF)
exposures are structured short-term
financings of reserves, inventories, or
receivables of exchange-traded
commodities, such as crude oil, metals,
or agricultural commodities, where the
source of repayment is the proceeds of
the sale of the commodity.
• Commercial real estate (CRE)
exposures finance the construction or
acquisition of real estate (including land
as well as improvements) where the
prospects for repayment and recovery
depend primarily on the cash flows
generated by the lease, rental, or sale of
the real estate.19 The broad CRE
category is further divided into two
groups: low-asset-correlation CRE and
HVCRE.20
19 CRE exposures are typically non-recourse
exposures, often to special purpose vehicles, and
are distinguishable from corporate exposures that
are collateralized by real estate for which the
prospects for repayment and recovery depend
primarily on the financial performance of the
broader commercial enterprise that is the obligor.
20 To describe a loan portfolio as having a
relatively high asset correlation means that any
defaults that occur in that portfolio are relatively
likely to occur at the same time, and for this reason
the portfolio is likely to exhibit greater variability

PO 00000

Frm 00016

Fmt 4701

Sfmt 4702

$20 million
1.57
2.33
3.83
5.37
7.17
9.11
12.73
18.05
26.08

$35 million
1.70
2.51
4.13
5.79
7.72
9.83
13.74
19.30
27.40

≥ $50 million
1.83
2.71
4.44
6.21
8.29
10.56
14.75
20.50
28.65

Most of the issues raised below for
comment are described in substantially
greater detail, in the context of CRE
exposures, in a white paper entitled
‘‘Loss Characteristics of CRE Loan
Portfolios,’’ released by the Federal
Reserve Board on June 10, 2003.
Commenters are encouraged to read the
white paper in conjunction with this
section.
A defining characteristic of SL
exposures (including CRE) is that the
risk factors influencing actual default
rates are likely to influence LGDs as
well. This is because both the
borrower’s ability to repay an exposure
and the banking organization’s recovery
on an exposure in the event of default
are likely to depend on the same
underlying factors, such as the net cash
flows of the property being financed.
in aggregate default rates. For two portfolios with
the same EL, the portfolio with more highly variable
aggregate default rates warrants higher capital to
cover UL (‘‘bad-tail events’’) with the same level of
confidence. Describing a portfolio as having a
relatively high asset correlation does not imply that
loans in that portfolio have relatively high PD, LGD,
or EL. In particular, loans in high asset correlation
portfolios may well have very low PDs and LGDs
and therefore ELs); conversely, loans in low asset
correlation portfolios may have very high PDs and
LGDs (and ELs). For any two loans from a portfolio
with a given asset correlation (or from two different
portfolios with the same asset correlation), the loan
with the lower EL should be assigned a lower risk
weight. For any two loans with the same EL, the
loan from the portfolio with the lower asset
correlation should incur a lower capital charge,
because bad-tail events are less likely to occur in
that portfolio.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
This suggests a positive correlation
between observed default frequencies
and observed loss rates on defaulted
loans, with both declining during
periods of favorable economic
conditions and both increasing during
unfavorable economic periods. While
cyclicality in LGDs may be significant
for a number of lending activities, the
Agencies believe that cyclicality is
likely to be the norm for SL portfolios,
and that a banking organization’s
procedures for estimating LGD inputs
for SL exposures should assess and
quantify this cyclicality in a
comprehensive and systematic fashion.
The Agencies invite comment on ways to
deal with cyclicality in LGDs. How can risk
sensitivity be achieved without creating
undue burden?

For core and opt-in banks that may
not be able to provide sufficiently
reliable estimates of PD, LGD, and M for
each SL exposure, the New Accord
offers a Supervisory Slotting Criteria
(SSC) approach. Under this approach,
rather than estimating the loan-level risk
parameters, banking organizations
would use slotting criteria to map their
internal risk rating grades to one of five
supervisory rating grades: Strong, Good,
Satisfactory, Weak, and Default. In

addition, supervisory risk weights
would be assigned to each of these
supervisory rating grades. To assist
banking organizations in implementing
these supervisory rating grades, for
reference purposes the New Accord
associates each with an explicit range of
external rating grades. If the SSC
approach were allowed in the United
States, the Agencies would have to
develop slotting criteria that would take
into account factors such as market
conditions; financial ratios such as debt
service coverage or loan-to-value ratios;
cash flow predictability; strength of
sponsor or developer; and other factors
likely to affect the PD and/or LGD of
each loan.
The Agencies invite comment on the
merits of the SSC approach in the United
States. The Agencies also invite comment on
the specific slotting criteria and associated
risk weights that should be used by
organizations to map their internal rating
grades to supervisory rating grades if the SSC
approach were to be adopted in the United
States.

Under the A–IRB approach, a banking
organization would estimate the risk
inputs for each SL exposure and then
calculate the A–IRB capital charge for
the exposure by substituting the
estimated PD, LGD, EAD, and M into

45915

one of two risk weight functions. The
first risk weight function is the
wholesale risk weight function and
applies to all PF, OF, and CF exposures,
as well as to all low-asset-correlation
CRE exposures (including in-place
commercial properties). The second risk
weight function applies to all HVCRE
exposures. It also is the same as the
wholesale risk weight function, except
that it incorporates a higher asset
correlation parameter. The asset
correlation equation for HVCRE is as
follows:
R = 0.12 × (1¥EXP (¥50 × PD)) + 0.30
× [EXP (¥50 × PD)]
Where
R denotes asset correlation;
EXP denotes the natural exponential
function; and
PD denotes probability of default.
The following table presents the A–
IRB capital requirement (K) for a range
of values of both PD and M. In this
table, LGD is assumed to equal 45
percent. This LGD is used for
consistency with the similar table above
for wholesale exposures and should not
be construed as an indication that 45
percent is a typical LGD for HVCRE
exposures.

HVCRE CAPITAL REQUIREMENTS
[In percentage points]
Effective remaining maturity
PD
1 year
0.05 percent .................................................................................................................................
0.10 percent .................................................................................................................................
0.25 percent .................................................................................................................................
0.50 percent .................................................................................................................................
1.00 percent .................................................................................................................................
2.00 percent .................................................................................................................................
5.00 percent .................................................................................................................................
10.00 percent ...............................................................................................................................
20.00 percent ...............................................................................................................................

All ADC loans would be treated as
HVCRE exposures, unless the borrower
has ‘‘substantial equity’’ at risk or the
property is pre-sold or sufficiently preleased. In part, this reflects some
empirical evidence suggesting that most
ADC loans have relatively high asset
correlations. It also, however, reflects a
longstanding supervisory concern that
CRE lending to finance speculative
construction and development is
vulnerable to, and may worsen,
speculative swings in CRE markets,
especially when there is little borrower
equity at risk. Such lending was a major
factor causing the stress experienced by
many banks in the early 1990s, not only

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

in the United States but in other
countries as well.
Under the New Accord, SL loans
financing the construction of one- to
four-family residential properties (single
or in subdivisions) are included with
other ADC loans in the high asset
correlation category. However, loans
financing the construction of pre-sold
one- to four-family residential
properties would be eligible to be
treated as low-asset-correlation CRE
exposures. In some cases the loans may
finance the construction of subdivisions
or other groups of houses, some of
which are pre-sold while others are not.
Under the New Accord, each national
supervisory authority is directed to

PO 00000

Frm 00017

Fmt 4701

Sfmt 4702

1.24
2.05
3.74
5.52
7.53
9.55
13.12
18.59
26.84

3 years
2.46
3.61
5.76
7.79
9.89
11.79
15.12
20.54
28.65

5 years
3.68
5.16
7.77
10.07
12.25
14.02
17.11
22.49
30.47

recognize and incorporate into its
implementation of the New Accord the
high asset correlation determinations of
other national supervisory authorities
for loans made in their respective
jurisdictions. Thus, when the Agencies
designate certain CRE properties as
HVCRE, foreign banking organizations
making extensions of credit to those
properties also would be expected to
treat them as HVCRE. Similarly, when
non-U.S. supervisory authorities
designate certain CRE as HVCRE, U.S.
banking organizations that extend credit
to those properties would be expected to
treat them as HVCRE.

E:\FR\FM\04AUP2.SGM

04AUP2

45916

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

The Agencies invite the submission of
empirical evidence regarding the (relative or
absolute) asset correlations characterizing
portfolios of ADC loans, as well as comments
regarding the circumstances under which
such loans would appropriately be
categorized as HVCRE.
The Agencies also invite comment on the
appropriateness of exempting from the highasset-correlation category ADC loans with
substantial equity or that are pre-sold or
sufficiently pre-leased. The Agencies invite
comment on what standard should be used
in determining whether a property is
sufficiently pre-leased when prevailing
occupancy rates are unusually low.
The Agencies invite comment on whether
high-asset-correlation treatment for one- to
four-family residential construction loans is
appropriate, or whether they should be
included in the low-asset-correlation
category. In cases where loans finance the
construction of a subdivision or other group
of houses, some of which are pre-sold while
others are not, the Agencies invite comment
regarding how the ‘‘pre-sold’’ exception
should be interpreted.
The Agencies invite comment on the
competitive impact of treating defined
classes of CRE differently. What are
commenters’ views on an alternative
approach where there is only one risk weight
function for all CRE? If a single risk weight
function for all CRE is considered, what
would be the appropriate asset correlation to
employ?

Lease Financings
Under the wholesale A–IRB
framework, some lease financings
require special consideration. A
distinction is made for leases that
expose the lessor to residual value risk,
namely the risk of the fair value of the
assets declining below the banking
organization’s estimate of residual risk
at lease inception. If a banking
organization has exposure to residual
value risk, it would assign a 100 percent
risk weight to the residual value amount
and determine a risk-weighted asset
equivalent for the lease’s remaining net
investment (net of residual value
amount) using the same methodology as
for any other wholesale exposure. The
sum of these components would be the
risk-weighted asset amount for a
particular lease. Where a banking
organization does not have exposure to
residual value risk, the lease’s net
investment would be subject to a capital
charge using the same methodology
applied to any other wholesale
exposure.
This approach would be used
regardless of accounting classification as
a direct finance, operating or leveraged
lease. For leveraged leases, when the
banking organization is the equity
participant it would net the balance of
the non-recourse debt against the
discounted lease payment stream prior

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

to applying the risk weight. If the
banking organization is the debt
participant, the exposure would be
treated as any other wholesale exposure.
The Agencies are seeking comment on the
wholesale A–IRB capital formulas and the
resulting capital requirements. Would this
approach provide a meaningful and
appropriate increase in risk sensitivity in the
sense that the results are consistent with
alternative assessments of the credit risks
associated with such exposures or the capital
needed to support them? If not, where are
there material inconsistencies?
Does the proposed A–IRB maturity
adjustment appropriately address the risk
differences between loans with differing
maturities?

Retail Exposures: Definitions and Inputs
The second major exposure category
in the A–IRB framework is the retail
exposure category. This category
encompasses the vast majority of credit
exposures to individual consumers. The
Agencies also are considering whether
certain SME exposures should be
eligible for retail A–IRB treatment. The
retail exposure category has three
distinct sub-categories: (1) Residential
mortgages (and related exposures); (2)
qualifying revolving exposures (QREs);
and (3) other retail exposures. There are
separate A–IRB capital formulas for
each of these three sub-categories to
reflect different levels of associated risk.
The Agencies propose that the
residential mortgage exposure subcategory be defined to include loans
secured by first or subsequent liens on
one-to four-family residential
properties, including term loans and
revolving lines of credit secured by
home equity. There would be no upper
limit on the size of the exposure that
could be included in the residential
mortgage exposure sub-category, but the
borrower would have to be an
individual and the banking organization
should generally manage the exposure
as part of a pool of similar exposures.
Residential mortgage exposures that are
individually internally rated and
managed similarly to commercial
exposures, rather than managed and
internally rated as pools, would be
treated under the wholesale A–IRB
framework.
The second sub-category of retail
exposures is qualifying revolving
exposures (QREs). The Agencies
propose to define QREs as exposures to
individuals that are revolving,
unsecured, uncommitted, less than
$100,000, and managed as part of a pool
of similar exposures. In practice, QREs
will include primarily exposures where
customers’ outstanding borrowings are
permitted to fluctuate based on their

PO 00000

Frm 00018

Fmt 4701

Sfmt 4702

own decisions to borrow and repay, up
to a limit established by the banking
organization. Most credit card exposures
to individuals and overdraft lines on
individual checking accounts would be
QREs.
The third sub-category of retail
exposures, other retail exposures,
includes two types of exposures. First,
it encompasses all exposures to
individuals for non-business purposes
that are generally managed as part of a
pool of similar exposures and that do
not meet the conditions for inclusion in
the first two sub-categories of retail
exposures. The Agencies are not
proposing to establish a fixed upper
limit on the size of exposures to
individuals that are eligible for the other
retail treatment. In addition, the
Agencies are proposing that the other
retail sub-category include certain SME
exposures that are managed on a pool
basis similar to retail exposures. These
exposures could be to a company or to
an individual. The Agencies are
considering an individual borrower
exposure threshold of $1 million for
such exposures. For the purpose of
assessing compliance with the
individual borrower exposure threshold,
the banking organization would
aggregate all exposures to a particular
borrower on a fully consolidated basis.
Credit card accounts with balances
between $100,000 and $1 million would
be considered other retail exposures
rather than QRE, even if the accounts
are extended to or guaranteed by an
individual and used exclusively for
small business purposes.
The Agencies are interested in comment on
whether the proposed $1 million threshold
provides the appropriate dividing line
between those SME exposures that banking
organizations should be allowed to treat on
a pooled basis under the retail A–IRB
framework and those SME exposures that
should be rated individually and treated
under the wholesale A–IRB framework.

One of the most significant
differences between the wholesale and
retail A–IRB categories is that the risk
inputs for retail exposures do not have
to be assigned at the level of an
individual exposure. The Agencies
recognize that banking organizations
typically manage retail exposures on a
portfolio or pool basis, where each
portfolio or pool contains exposures
with similar risk characteristics.
Therefore, a key characteristic of the
retail A–IRB framework is that the risk
inputs for retail exposures would be
assigned to portfolios or pools of
exposures rather than to individual
exposures.
It is important to highlight that within
each of the three sub-categories of retail

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
exposures, the retail A–IRB framework
is intended to provide banking
organizations with substantial flexibility
to use the retail portfolio segmentation
that they believe is most appropriate for
their activities. In determining how to
group their retail exposures within each
sub-category into portfolio segments for
the purpose of assigning A–IRB risk
inputs, the Agencies believe that
banking organizations should use a
segmentation approach that is
consistent with their approach for
internal risk assessment purposes and
that classifies exposures according to
predominant risk characteristics.
As general principles for
segmentation, banking organizations
should group exposures in each of the
three retail sub-categories into portfolios
or pools according to the sub-category’s
principal risk drivers, and would have
to be able to demonstrate that the
resultant segmentation effectively
differentiates and rank orders risk and
provides reasonably accurate and
consistent quantitative estimates of PD,
LGD, and EAD. With the exceptions
noted below, the Agencies are not
proposing that institutions must
consider any particular risk drivers or
employ any minimum number of
portfolios or pools in any of the three
retail sub-categories. The only specific
limitations that the Agencies would
propose in regard to the portfolio
segmentation of retail exposures are (1)
banking organizations generally would
not be permitted to combine retail
exposures from multiple countries into
the same portfolio segment (because of
differences in national legal systems and
bankruptcy regimes), and (2) banking
organizations would need to separately
segment delinquent retail exposures.
The inputs to the retail A–IRB capital
formulas differ slightly from the inputs
to the wholesale A–IRB capital
formulas. Measures of PD, LGD, and
EAD remain important elements, but
there is no M input to the retail A–IRB
capital formulas. Rather, the retail A–
IRB capital formulas implicitly
incorporate average maturity effects in
general, such as in the residential
mortgage sub-category.
Aside from the applicable definition
of default, discussed below, the
definitions of PD, LGD, and EAD for
retail exposures are generally equivalent
to those for wholesale exposures. One
additional element of potential
flexibility for banking organizations in
the retail context needs to be
highlighted. The Agencies recognize
that certain banking organizations that
may qualify for the advanced
approaches segment their retail
portfolios for management purposes by

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

EL, rather than by separately measuring
PD and LGD, as required under the A–
IRB framework. Therefore, the Agencies
propose that banking organizations be
permitted substantial flexibility in
translating measures of EL into the
requisite PD and LGD inputs. For nonrevolving portfolio segments, EL
generally would equal the product of PD
and LGD, so that if a banking
organization has an estimate of EL and
either PD or LGD, it would be able to
infer an estimate of the other required
input.
In addition, the Agencies are
proposing that if one or the other of PD
and LGD did not tend to vary
significantly across portfolio segments,
the banking organization would be
permitted to apply a general estimate of
that input to multiple segments and to
use that general estimate, together with
segment-specific estimates of EL, to
infer segment-specific estimates of the
other required input. The Agencies note,
however, that this proposal offers
substantial flexibility to institutions and
may, in fact, be overly flexible (for
example, because LGDs on residential
mortgages tend to be quite cyclical). For
these loans, the above method of
inferring PDs or LGDs from a long-run
average EL would not necessarily result
in PD being estimated as a long-run
average, and LGD would not necessarily
reflect the loss rate expected to prevail
when default rates are high. Banking
organizations using an EL approach to
retail portfolio segmentation would
have to ensure that the A–IRB capital
requirement under this method is at
least as conservative as a PD/LGD
method in order to minimize any
potential divergences between capital
requirements computed under the PD/
LGD approach versus an EL approach.
As in the wholesale A–IRB
framework, a floor of 3 basis points
(0.03 percent) applies to the PD
estimates for all retail exposures (that is,
the minimum PD is 3 basis points). In
addition, for residential mortgage
exposures other than those guaranteed
by a sovereign government, a floor of 10
percent on the LGD estimate would
apply, based on the view that LGDs
during periods of high default rates are
unlikely to fall below this level if
measured appropriately. Along with the
overall monitoring of the
implementation of the advanced
approaches and the determination
whether to generally relax the floors
established during the initial
implementation phases (that is, the 90
and 80 percent floors discussed above),
the Agencies intend to review the need
to retain PD and LGD floors for retail
exposures following the first two years

PO 00000

Frm 00019

Fmt 4701

Sfmt 4702

45917

of implementation of the A–IRB
framework.
The Agencies are proposing the
following data requirements for retail
A–IRB. Banking organizations would
have to have a minimum of five years
of data history for PD, LGD, and EAD,
and preferably longer periods so as to
include a complete economic cycle.
Banking organizations would not have
to give equal weight to all historical
factors if they can demonstrate that the
more recent data are better predictors of
the risk inputs. Also, banking
organizations would have to have a
minimum of three years of experience
with their portfolio segmentation and
risk management systems.
Definition of Default and Loss
The retail definition of default and
loss being proposed by the Agencies
differs significantly from that proposed
for the wholesale portfolio. Specifically,
the Agencies propose to use the
definitions of loss recognition embodied
in the Federal Financial Institutions
Examination Council (FFIEC) Uniform
Retail Credit Classification and Account
Management Policy.21 All residential
mortgages and all revolving credits
would be charged off, or charged down
to the value of the property, after a
maximum of 180 days past due; other
credits would be charged off after a
maximum of 120 days past due.
In addition, the Agencies are
proposing to define a retail default to
include the occurrence of any one of the
three following events if it occurs prior
to the respective 120- or 180-day FFIEC
policy trigger: (1) A full or partial
charge-off resulting from a significant
decline in credit quality of the exposure;
(2) a distressed restructuring or workout
involving forbearance and loan
modification; or (3) a notification that
the obligor has sought or been placed in
bankruptcy. Finally, for retail exposures
(as opposed to wholesale exposures) the
definition of default may be applied to
a particular facility, rather than to the
obligor. That is, default on one
obligation would not require a banking
organization to treat all other obligations
of the same obligor as defaulted.
Undrawn Lines
The treatment of undrawn lines of
credit, in particular those associated
with credit cards, merits specific
discussion. Banking organizations
would be permitted to incorporate
undrawn retail lines in one of two ways.
First, banking organizations could
21 The FFIEC Uniform Retail Credit Classification
and Account Management Policy was issued on
June 12, 2000. It is available on the FFIEC Web site
at www.FFIEC.gov.

E:\FR\FM\04AUP2.SGM

04AUP2

45918

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

incorporate them into their EAD
estimates directly, by assessing the
likelihood that undrawn balances would
be drawn at the time of an event of
default. Second, banking organizations
could incorporate them into LGD
estimates by assessing the size of
potential losses in default (including
those arising from both currently drawn
and undrawn balances) as a proportion
of the current drawn balance. In the
latter case, it is possible that the
relevant LGD estimates would exceed
100 percent. While the proposed EAD
approach for undrawn wholesale and
retail lines is the same, the Agencies are
aware that the sheer volume of credit
card undrawn lines and the ratio of
undrawn lines to outstanding balances
create issues for undrawn retail lines
that differ from undrawn wholesale
lines not only in degree but also in kind.
An additional issue arises in
connection with the undrawn lines
associated with credit card accounts
whose drawn balances (but not
undrawn balances) have been
securitized. To the extent that banking
organizations remain exposed to the risk
that such undrawn lines will be drawn,
but such undrawn lines are not
themselves securitized, then there is a
need for institutions to hold regulatory
capital against such undrawn lines. The
Agencies propose that a banking
organization would be required to hold
capital against the full amount of any
undrawn lines regardless of whether
drawn amounts are securitized. This
presumes that the institution itself is
exposed to the credit risk associated
with future draws.
The Agencies are interested in comments
and specific proposals concerning methods
for incorporating undrawn credit card lines
that are consistent with the risk
characteristics and loss and default histories
of this line of business.
The Agencies are interested in further
information on market practices in this
regard, in particular the extent to which
banking organizations remain exposed to
risks associated with such accounts. More
broadly, the Agencies recognize that
undrawn credit card lines are significant in
both of the contexts discussed above, and are
particularly interested in views on the
appropriate retail A–IRB treatment of such
exposures.

Future Margin Income
In the New Accord, the retail A–IRB
treatment of QREs includes a unique
additional input that arises because of
the large amount of expected losses
typically associated with QREs. As
noted above, the A–IRB approach would
require banking organizations to hold
regulatory capital against both EL and
UL. Banking organizations typically

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

seek to cover expected losses through
interest income and fees for all of their
business activities, and the Agencies
recognize that this practice is a
particularly important aspect of the
business model for QREs.
The Agencies are including in this
proposal, for the QRE sub-category only,
that future margin income (FMI) be
permitted to offset a portion of the A–
IRB retail capital charge relating to EL.
For this purpose, the Agencies propose
to define the amount of eligible FMI for
the QRE sub-category as the amount of
income anticipated to be generated by
the relevant exposures over the next
twelve months that can reasonably be
assumed to be available to cover
potential credit losses on the exposures
after covering expected business
expenses, and after subtracting a
cushion to account for potential
volatility in credit losses (UL). FMI
would not be permitted to include
anticipated income from new accounts
and would have to incorporate
assumptions about income from existing
accounts that are in line with the
banking organization’s historical
experience. The amount of the cushion
to account for potential volatility in
credit losses would be set equal to two
standard deviations of the banking
organization’s annualized loss rate on
the exposures. The Agencies would
expect banking organizations to be able
to support their estimates of eligible
FMI on the basis of historical data and
would disallow the use of FMI in the
QRE capital formula if this is not the
case. The step needed to recognize
eligible FMI is discussed below.
Permitting a FMI offset to the A–IRB
capital requirement for QREs could have
a significant impact on the level of
minimum regulatory capital at
institutions adopting the advanced
approaches. The Agencies would need
to fully assess and analyze the impact of
such an FMI offset on institutions’ riskbased capital ratios prior to final
implementation of the A–IRB approach.
Furthermore, the Agencies anticipate
the need to issue additional guidance
setting out more specific expectations in
this regard.
For the QRE sub-category of retail
exposures only, the Agencies are seeking
comment on whether or not to allow banking
organizations to offset a portion of the A–IRB
capital requirement relating to EL by
demonstrating that their anticipated FMI for
this sub-category is likely to more than
sufficiently cover EL over the next year.
The Agencies are seeking comment on the
proposed definitions of the retail A–IRB
exposure category and sub-categories. Do the
proposed categories provide a reasonable
balance between the need for differential
treatment to achieve risk-sensitivity and the

PO 00000

Frm 00020

Fmt 4701

Sfmt 4702

desire to avoid excessive complexity in the
retail A–IRB framework? What are views on
the proposed approach to inclusion of SMEs
in the other retail category?
The Agencies are also seeking views on the
proposed approach to defining the risk
inputs for the retail A–IRB framework. Is the
proposed degree of flexibility in their
calculation, including the application of
specific floors, appropriate? What are views
on the issues associated with undrawn retail
lines of credit described here and on the
proposed incorporation of FMI in the QRE
capital determination process?
The Agencies are seeking comment on the
minimum time requirements for data history
and experience with portfolio segmentation
and risk management systems: Are these time
requirements appropriate during the
transition period? Describe any reasons for
not being able to meet the time requirements.

Retail Exposures: Formulas
The retail A–IRB capital formulas are
very similar to the wholesale A–IRB
formulas, and are based on the same
underlying concepts. However, because
there is no M adjustment associated
with the retail A–IRB framework, the
retail A–IRB capital calculations
generally involve fewer steps than the
wholesale A–IRB capital calculations.
As with the wholesale A–IRB
framework, the output of the retail A–
IRB formulas is a minimum capital
requirement, expressed in dollars, for
the relevant pool of exposures. The
capital requirement would be converted
into an equivalent amount of riskweighted assets by multiplying the
capital requirement by 12.5. The two
key steps in implementing the retail A–
IRB capital formulas are (1) assessing
the relevant asset correlation parameter,
and (2) calculating the minimum capital
requirement for the relevant pool of
exposures.
Residential Mortgages and Related
Exposures
For residential mortgage and related
exposures, the retail A–IRB capital
formula requires only one step. This is
because the asset correlation parameter
for such exposures is fixed at 15
percent, regardless of the PD of any
particular pool of exposures. The fixed
asset correlation parameter reflects the
Agencies’ view that the arguments for
linking the asset correlation to PD, as
occurs in the wholesale A–IRB
framework and in the other two subcategories of retail exposures, are not as
relevant for residential mortgage-related
exposures, whose performance is
significantly influenced by broader
trends in the housing market for
borrowers of all credit qualities. The
assumed asset correlation of 15 percent
also seeks implicitly to reflect the higher
average maturity associated with

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
Where
K denotes the capital requirement;
EAD denotes exposure at default;
LGD denotes loss given default;
N(x) denotes the standard normal
cumulative distribution function;
G(x) denotes the inverse of the standard
normal cumulative distribution
function; and
PD denotes probability of default.

residential mortgage exposures and is
therefore higher than would likely be
the case if a specific maturity
adjustment were also included in the
retail A–IRB framework. The proposed
retail A–IRB capital formula for
residential mortgage and related
exposures is as follows:
K = EAD * LGD * N[1.08465 * G(PD) +
0.4201 * G(0.999)]

45919

The following table depicts a range of
representative capital requirements (K)
for residential mortgage and related
exposures based on this formula. Three
different illustrative LGD assumptions
are shown: 15 percent, 35 percent, and
55 percent. For comparison purposes,
the current capital requirement on most
first mortgage loans is 4 percent and on
most home equity loans is 8 percent.

CAPITAL REQUIREMENTS
[In percentage points]
LGD
PD
15 percent
0.05 percent .................................................................................................................................
0.10 percent .................................................................................................................................
0.25 percent .................................................................................................................................
0.50 percent .................................................................................................................................
1.00 percent .................................................................................................................................
2.00 percent .................................................................................................................................
5.00 percent .................................................................................................................................
10.00 percent ...............................................................................................................................
20.00 percent ...............................................................................................................................

Subject to the caveats noted earlier,
evidence from QIS3 suggested that
advanced approach banking
organizations would experience a
reduction in credit risk capital
requirements for residential mortgage
exposures of about 56 percent.
Private Mortgage Insurance
The Agencies wish to highlight one
issue associated with the A–IRB capital
requirements for the residential
mortgage sub-category relating to the
treatment of private mortgage insurance
(PMI). Most PMI arrangements
effectively provide partial compensation
to the banking organization in the event
of a mortgage default. Accordingly, the
Agencies consider that it may be
appropriate for banking organizations to
recognize such effects in the LGD
estimates for individual mortgage
portfolio segments, consistent with the
historical loss experience on those
segments during periods of high default
rates. Such an approach would avoid
requiring banking organizations to
quantify specifically the effect of PMI on
a loan-by-loan basis; rather, they could
estimate the effect of PMI on an average
basis for each segment. This approach
effectively ignores the risk that the
mortgage insurers themselves could
default.
The Agencies seek comment on the
competitive implications of allowing PMI
recognition for banking organizations using
the A–IRB approach but not allowing such
recognition for general banks. In addition, the
Agencies are interested in data on the
relationship between PMI and LGD to help

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

assess whether it may be appropriate to
exclude residential mortgages covered by
PMI from the proposed 10 percent LGD floor.
The Agencies request comment on whether
or the extent to which it might be appropriate
to recognize PMI in LGD estimates.
More broadly, the Agencies are interested
in information regarding the risks of each
major type of residential mortgage exposure,
including prime first mortgages, sub-prime
mortgages, home equity term loans, and
home equity lines of credit. The Agencies are
aware of various views on the resulting
capital requirements for several of these
product areas, and wish to ensure that all
appropriate evidence and views are
considered in evaluating the A–IRB treatment
of these important exposures.
The risk-based capital requirements for
credit risk of prime mortgages could well be
less than one percent of their face value
under this proposal. The Agencies are
interested in evidence on the capital required
by private market participants to hold
mortgages outside of the federally insured
institution and GSE environment. The
Agencies also are interested in views on
whether the reductions in mortgage capital
requirements on mortgage loans
contemplated here would unduly extend the
federal safety net and risk contributing to a
credit-induced bubble in housing prices. In
addition, the Agencies are also interested in
views on whether there has been any
shortage of mortgage credit under the general
risk-based capital rules that would be
alleviated by the proposed changes.

Qualifying Revolving Exposures
The second sub-category of retail
exposures is QREs. The calculation of
A–IRB capital requirements for QREs
would require three steps: (1)
Calculation of the relevant asset

PO 00000

Frm 00021

Fmt 4701

Sfmt 4702

0.17
0.30
0.61
1.01
1.65
2.64
4.70
6.95
9.75

35 percent
0.41
0.70
1.41
2.36
3.86
6.17
10.97
16.22
22.75

55 percent
0.64
1.10
2.22
3.70
6.06
9.70
17.24
25.49
35.75

correlation parameter, (2) calculation of
the minimum capital requirement
assuming no offset for eligible FMI, and
(3) application of the offset for eligible
FMI. These steps would be performed
for each QRE portfolio segment
individually.
As in the case of wholesale exposures,
it is assumed that the asset correlation
for QREs declines as PD rises. This
reflects the view that pools of borrowers
with lower credit quality (higher PD) are
less likely to experience simultaneous
defaults than pools of higher credit
quality (lower PD) borrowers, because
with higher PD borrowers defaults are
more likely to result from borrowerspecific or idiosyncratic factors. In the
case of QREs, the asset correlation
approaches an upper bound value of 11
percent for very low PD values and
approaches a lower bound value of 2
percent for very high PD values. The
specific formula for determining the
asset correlation parameter for QREs is
as follows:
R = 0.02 * (1–EXP(¥50 * PD)) + 0.11 *
[1–(1–EXP(¥50 * PD))]
Where
R denotes asset correlation;
EXP denotes the natural exponential
function; and
PD denotes probability of default.
The second step in the A–IRB capital
calculation for QREs would be the
calculation of the capital requirement
assuming no FMI offset. The specific
formula to calculate this amount is as
follows:

E:\FR\FM\04AUP2.SGM

04AUP2

45920

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

KNo FMI = EAD * LGD * N[(1–R)∧¥0.5
* G(PD) + (R/(1–R))∧0.5 * G(0.999)]
Where
KNo FMI denotes the capital requirement
assuming no FMI offset;
EAD denotes exposure at default;
LGD denotes loss given default;
N(x) denotes the standard normal
cumulative distribution function;
R denotes asset correlation;
G(x) denotes the inverse of the standard
normal cumulative distribution
function; and
PD denotes probability of default.
Future Margin Income Adjustment
The result of this calculation
effectively includes both an EL and a UL
component. As already discussed, for
QREs only, the Agencies are considering
the possibility of allowing institutions
to offset a portion of the EL portion of
the capital requirement using eligible
FMI. Up to 75 percent of the EL portion
of the capital requirement could
potentially be offset in this fashion. The
specific calculation for determining the
capital requirement (K) after application
of the potential offset for eligible FMI is
as follows.
K = KNo FMI¥eligible FMI offset
Where
K denotes the capital requirement after
application of an offset for eligible
FMI;
KNo FMI denotes the capital requirement
assuming no FMI offset;
Eligible FMI offset equals:
0.75 * EL if estimated FMI equals or
exceeds the expected 12-month loss
amount plus two standard
deviations of the annualized loss
rate, or zero otherwise;
EL denotes expected loss (EL = EAD *
PD * LGD);
FMI denotes future margin income;
EAD denotes exposure at default;
PD denotes probability of default; and
LGD denotes loss given default.
If eligible FMI did not exceed the
required minimum, then recognition of
eligible FMI would be disallowed.
The Agencies are interested in views on
whether partial recognition of FMI should be
permitted in cases where the amount of
eligible FMI fails to meet the required
minimum. The Agencies also are interested
in views on the level of portfolio
segmentation at which it would be

appropriate to perform the FMI calculation.
Would a requirement that FMI eligibility
calculations be performed separately for each
portfolio segment effectively allow FMI to
offset EL capital requirements for QREs?

The following table depicts a range of
representative capital requirements (K)
for QREs based on these formulas. In
each case, it is assumed that the
maximum offset for eligible FMI has
been applied. The LGD is assumed to
equal 90 percent, consistent with
recovery rates for credit card portfolios.
The table shows capital requirements
with recognition of FMI and without
recognition of FMI but using the same
formula in other respects. As PDs
increase, the proportion of total required
capital held against EL after deducting
the 75 percent offset rises at an
increasing rate and the proportion held
against UL declines at an increasing
rate. Offsets from EL, as considered in
this ANPR, would therefore have a
proportionally greater impact on
reducing required capital charges as
default probabilities increase. For
comparison purposes, the current
capital requirement on drawn credit
card exposures is 8 percent and is zero
for undrawn credit lines.

CAPITAL REQUIREMENT
[In percentage points]
With FMI
capital 8%

PD
0.05
0.10
0.25
0.50
1.00
2.00
5.00
10.0
20.0

...................
...................
...................
...................
...................
...................
...................
...................
...................

Without FMI
capital 8%

0.68
1.17
2.24
3.44
4.87
6.21
7.89
11.12
17.23

0.72
1.23
2.41
3.78
5.55
7.56
11.27
17.87
30.73

Subject to the same qualifications
mentioned earlier, the QIS3 results
estimated an increase in credit risk
capital requirements for QREs of about
16 percent.
Other Retail Exposures
The third and final sub-category of
retail A–IRB exposures is other retail
exposures. This sub-category
encompasses a wide variety of different
exposures including auto loans, student
loans, consumer installment loans, and
some SME loans. Two steps would be

required to calculate the A–IRB capital
requirement for other retail exposures:
(1) Calculating the relevant asset
correlation parameter, and (2)
calculating the capital requirement.
Both of these steps would be done
separately for each portfolio segment
included within the other retail subcategory.
As for wholesale exposures and QREs,
the asset correlation parameter for other
retail exposures declines as PD rises. In
the case of other retail exposures, the
asset correlation parameter approaches
an upper bound value of 17 percent for
very low PD values and approaches a
lower bound value of 2 percent for very
high PD values. The specific formula for
determining the asset correlation for
other retail exposures is as follows:
R = 0.02 * (1¥EXP(¥35 * PD)) + 0.17
* [1¥(1¥EXP(¥35 * PD))]
Where
R denotes asset correlation;
EXP denotes the natural exponential
function; and
PD denotes probability of default.
The second step in the A–IRB capital
calculation for other retail exposures
would be the calculation of the capital
requirement (K). The specific formula to
calculate this amount is as follows:
K = EAD * LGD * N[(1¥R)∧¥0.5 *
G(PD) + (R / (1¥R))∧0.5 * G(0.999)]
Where
K denotes the capital requirement;
EAD denotes exposure at default;
LGD denotes loss given default;
PD denotes probability of default;
N(x) denotes the standard normal
cumulative distribution function;
G(x) denotes the inverse of the standard
normal cumulative distribution
function; and
R denotes asset correlation.
The following table depicts a range of
representative capital requirements (K)
for other retail exposures based on this
formula. Three different LGD
assumptions are shown—25 percent, 50
percent, and 75 percent—in order to
depict a range of potential outcomes
depending on the characteristics of the
underlying retail exposure. For
comparison purposes, the current
capital requirement on most of the
exposures likely to be included in the
other retail sub-category is 8 percent.

CAPITAL REQUIREMENTS
[In percentage points]
LGD
PD
25 percent
0.05 percent .................................................................................................................................

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

PO 00000

Frm 00022

Fmt 4701

Sfmt 4702

E:\FR\FM\04AUP2.SGM

0.33

04AUP2

50 percent
0.66

75 percent
0.99

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

45921

CAPITAL REQUIREMENTS—Continued
[In percentage points]
LGD
PD
25 percent
0.10 percent .................................................................................................................................
0.25 percent .................................................................................................................................
0.50 percent .................................................................................................................................
1.00 percent .................................................................................................................................
2.00 percent .................................................................................................................................
5.00 percent .................................................................................................................................
10.00 percent ...............................................................................................................................
20.00 percent ...............................................................................................................................

Subject to the qualifications described
earlier, QIS3 estimated a 25 percent
reduction in credit risk-based capital
requirements for the other retail
category.
The Agencies are seeking comment on the
retail A–IRB capital formulas and the
resulting capital requirements, including the
specific issues mentioned. Are there
particular retail product lines or retail
activities for which the resulting A–IRB
capital requirements would not be
appropriate, either because of a misalignment
with underlying risks or because of other
potential consequences?

A–IRB: Other Considerations
As described earlier, the A–IRB
capital requirement includes
components to cover both EL and UL.
Because banking organizations have
resources other than capital to cover EL,
the Agencies propose to recognize
certain of these measures as potential
offsets to the A–IRB capital requirement,
subject to the limitations set forth
below. The use of eligible FMI for QREs
is one such potential mechanism that
has already been discussed.
Loan Loss Reserves
A second important mechanism
involves the allowance for loan and
lease losses (ALLL), also referred to as
general loan loss reserves. Under the
general risk-based capital rules, an
amount of the ALLL is eligible for
inclusion as an element of Tier 2
capital, up to a limit equal to 1.25
percent of gross risk-weighted assets.
Loan loss reserves above this limit are
deducted from risk-weighted assets, on
a dollar-for-dollar basis. The New
Accord proposes to retain the 1.25
percent limit on the eligibility of loan
loss reserves as an element of Tier 2
capital. However, the New Accord also
contains, and the Agencies are
proposing for comment, a feature that
would allow the amount of the ALLL
(net of associated deferred tax) above
this 1.25 percent limit to be used to

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

offset the EL portion of A–IRB capital
requirements in certain circumstances.
The offset would be limited to that
amount of EL that exceeds the 1.25
percent limit. For example, if the 1.25
percent limit equals $100, the ALLL
equals $125, and the EL portion of the
A–IRB capital requirement equals $110,
then $10 of the capital requirement may
be directly offset ($110¥$100). The
additional amount of the ALLL not
included in Tier 2 capital and not
included as a direct offset against the A–
IRB capital requirement ($125¥$110 =
$15 in the example) would continue to
be deducted from risk-weighted assets.
It is important to recognize that this
treatment would likely result in a
significantly more favorable treatment of
such excess ALLL amounts than simply
deducting them from risk-weighted
assets. Under the proposal, banking
organizations would be allowed to
multiply the eligible excess ALLL by a
factor of 12.5 because the minimum
total capital requirement is 8 percent of
risk-weighted assets. In effect, this
treatment is 12.5 times more favorable
than the treatment contained in the
general risk-based capital rules, which
allow only a deduction against riskweighted assets on a dollar-for-dollar
basis. In addition, it is important to note
that a dollar-for-dollar offset against the
A–IRB capital requirement is also more
favorable than the inclusion of ALLL
below the 1.25 percent limit in Tier 2
capital, because the latter has no impact
on Tier 1 capital ratios, while the former
does.
The Agencies recognize the existence of
various issues in regard to the proposed
treatment of ALLL amounts in excess of the
1.25 percent limit and are interested in views
on these subjects, as well as related issues
concerning the incorporation of expected
losses in the A–IRB framework and the
treatment of the ALLL generally. Specifically,
the Agencies invite comment on the domestic
competitive impact of the potential
difference in the treatment of reserves
described above.

PO 00000

Frm 00023

Fmt 4701

Sfmt 4702

0.56
1.06
1.64
2.35
3.08
3.94
5.24
8.55

50 percent
1.11
2.13
3.28
4.70
6.15
7.87
10.48
17.10

75 percent
1.67
3.19
4.92
7.05
9.23
11.81
15.73
25.64

Another issue the Agencies wish to
highlight is the inclusion within the
New Accord of the ability for banking
organizations to make use of ‘‘general
specific’’ provisions as a direct offset
against EL capital requirements. Such
provisions are not specific to particular
exposures but are specific to particular
categories of exposures and are not
allowed as an element of Tier 2 capital.
While several other countries make use
of such provisions, the Agencies do not
believe existing elements of the ALLL in
the United States qualify for such
treatment.
The Agencies seek views on this issue,
including whether the proposed U.S.
treatment has significant competitive
implications. Feedback also is sought on
whether there is an inconsistency in the
treatment of general specific provisions (all
of which may be used as an offset against the
EL portion of the A–IRB capital requirement)
in comparison to the treatment of the ALLL
(for which only those amounts of general
reserves exceeding the 1.25 percent limit may
be used to offset the EL capital charge).

Charge-Offs
Another potential offset to the EL
portion of the A–IRB capital
requirements is the use of partial
charge-offs, where a portion of an
individual exposure is written off.
Given the A–IRB definition of default, a
partial charge-off would cause an
exposure to be classified as a defaulted
exposure (that is, PD=100%), in which
case the A–IRB capital formulas ensure
that the resulting capital requirement on
the defaulted exposure is equal to EAD
* LGD, where EAD is defined as the
gross exposure amount prior to the
partial charge-off. All of this capital
requirement can be considered to be
covering EL.
The New Accord proposes that for
such partially charged-off exposures, the
banking organization be allowed to use
the amount of the partial charge-off to
offset the EL component of that asset’s
capital charge on a dollar-for-dollar
basis. In addition, to the extent that the

E:\FR\FM\04AUP2.SGM

04AUP2

45922

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

partial charge-off on a defaulted
exposure exceeds the EL capital charge
on that exposure, the amount of this
surplus could be used to offset the EL
capital charges on other defaulted assets
in the same portfolio (for example,
corporates, banks, residential mortgages,
etc.), but not for any other purpose.
An implication of this aspect of the
New Accord is that if a defaulted loan’s
charge-off were at least equal to its
expected loss, no additional capital
requirement would be incurred on that
exposure. For example, consider a $100
defaulted exposure having an LGD of 40
percent, implying an expected loss of
$40, equal to the IRB capital charge. If
the charge-off were equal to $40, under
the New Accord approach, there would
be no additional capital required against
the resultant $60 net position. The
Agencies do not believe this is a
prudent or acceptable outcome, since
this position is not riskless and a
banking organization could be forced to
recognize additional charge-offs if the
recoveries turn out to be less than
expected.
To prevent this possibility, the
Agencies propose that, for defaulted
exposures, the A–IRB capital charge
(inclusive of any EL offsets for chargeoffs) be calculated as the sum of (a) EAD
* LGD less any charge-offs and (b) 8
percent of the carrying value of the loan
(that is, the gross exposure amount
(EAD) less any charge-offs).
Also, the charged off amounts in
excess of the EAD * LGD product would
not be permitted to offset the EL capital
requirements for other exposures. In
effect, the proposed A–IRB capital
charge on a defaulted exposure adds a
buffer for defaulted assets and results in
a floor equal to 8 percent of the
remaining book value of the exposure if
the banking organization has taken a
charge-off equal to or greater than the
EAD * LGD. Importantly, this treatment
would not apply to a defaulted exposure
that has been restructured and where
the obligor has not yet defaulted on the
restructured credit. Upon any
restructuring, whether associated with a
default or otherwise, the A–IRB capital
charge would be based on the EAD, PD,
LGD, and M applicable to the exposure
after it has been restructured. The
existence of any partial charge-offs
associated with the pre-restructured
credit would affect the A–IRB capital
charge on the restructured exposure
only through its impact on the postrestructured exposure’s EAD, PD, and/or
LGD.
Purchased Receivables
This section describes the A–IRB
treatment for wholesale and retail credit

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

exposures acquired from another
institution (purchased receivables). The
purchase of such receivables may
expose the acquiring banking
organization to potential losses from
two sources: credit losses attributable to
defaults by the underlying receivables
obligors, and losses attributable to
dilution of the underlying receivables.22
The total A–IRB capital requirement for
purchased receivables would be the sum
of (a) a capital charge for credit risk, and
(b) a separate capital charge for dilution
risk, when dilution is a material factor.
Capital Charge for Credit Risk
The New Accord’s proposed
treatment of purchased loans would
treat a purchase discount as equivalent
to a partial charge-off, and for this
reason it could imply a zero capital
charge against certain exposures. In
general, a zero capital charge would
emerge whenever the difference
between a loan’s face value and
purchase price (the purchase discount)
was greater than or equal to its LGD, as
might be the case with a secondary
market purchase of deeply distressed
debt. Again, the Agencies believe that a
zero capital charge in such a
circumstance is unwarranted because
the position is not riskless.
The Agencies propose that for a credit
exposure that is purchased or acquired
from another party, the A–IRB capital
charge would be calculated as if the
exposure were a direct loan to the
underlying obligor in the amount of the
loan’s carrying value to the purchasing
banking organization with other
attributes of the loan agreement (for
example, maturity, collateral,
covenants) and, hence, LGD, remaining
unchanged. This treatment would apply
regardless of whether the carrying value
to the purchasing banking organization
was less than, equal to, or greater than
the underlying instrument’s face value.
Thus, if a loan having a principal
amount equal to $100 and associated PD
and LGD of 10 percent and 40 percent
was purchased for $80, the capital
charge against the purchased loan
would be calculated as if that loan had
an EAD equal to $80, PD equal to 10
percent, and LGD equal to 40 percent.
In general, the same treatment would
apply to pools of purchased receivables.
22 Dilution refers to the possibility that the
contractual amounts payable by the receivables
obligors may be reduced through future cash or
non-cash credits to the accounts of these obligors.
Examples include offsets or allowances arising from
returns of goods sold, disputes regarding product
quality, possible debts of the originator/seller to a
receivables obligor, and any payment or
promotional discounts offered by the originator/
seller (for example, a credit for cash payments
within 30 days).

PO 00000

Frm 00024

Fmt 4701

Sfmt 4702

However, under the conditions detailed
below, an alternative top-down
approach (similar to that used for retail
exposures) may be applied to pools of
purchased receivables if the purchasing
banking organization can only estimate
inputs to the capital function (PD, LGD,
EAD, and M) on a pool or aggregate
basis.
Top-Down Method for Pools of
Purchased Receivables
Under the top-down approach,
required capital would be determined
using the appropriate A–IRB capital
formula (that is, for wholesale
exposures, the wholesale capital
function, and for retail exposures, the
appropriate retail capital function) in
combination with estimates of PD, LGD,
EAD, and M developed for pools of
receivables. In estimating the pool
parameters, the banking organization
first would determine EL for the
purchased receivables pool, expressed
(in decimal form) at an annual rate
relative to the amount currently owed to
the banking organization by the obligors
in the receivables pool. The estimated
EL would not take into account any
assumptions of recourse or guarantees
from the seller of the receivables or
other parties. If the banking organization
can decompose EL into PD and LGD
components, then it would do so and
use those components as inputs into the
capital function. If the institution
cannot decompose EL, then it would use
the following split: PD would equal the
estimated EL, and LGD would be 100
percent. Under the top-down approach,
EAD would equal the carrying amount
of the receivables and for wholesale
exposures, M would equal the exposureweighted average effective maturity of
the receivables in the pool.23
Treatment of Undrawn Receivables
Purchase Commitments
Capital charges against any undrawn
portions of receivables purchase
facilities (‘undrawn purchase
commitments’) also would be calculated
using the top-down methodology. The
EL (and/or PD and LGD) parameters
would be determined on the basis of the
current pool of eligible receivables using
the pool-level estimation methods
described above. For undrawn
commitments under revolving purchase
facilities, the New Accord specifies that
the EAD would be set at 75 percent of
the undrawn line. This treatment
reflects a concern that relevant
23 If a banking organization can estimate the
exposure-weighted average size of the pool it also
would use the firm-size adjustment (S) in the
wholesale framework.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
historical data for estimating such EADs
reliably is not available at many banking
organizations. For other undrawn
purchase commitments, EAD would be
estimated by the banking organization
providing the facility and would be
subject to the same operational
standards that are applicable to
undrawn wholesale credit lines. The
level of M associated with undrawn
purchase commitments would be the
average effective maturity of receivables
eligible for purchase from that seller, so
long as the facility contains effective
arrangements for protecting the banking
organization against an unanticipated
deterioration. In the absence of such
protections, the M for an undrawn
commitment would be calculated as the
sum of (a) the longest-dated potential
receivable under the purchase
agreement, and (b) the remaining
maturity of the facility.
The Agencies seek comment on the
proposed methods for calculating credit risk
capital charges for purchased receivables.
Are the proposals reasonable and
practicable?
For committed revolving purchase
facilities, is the assumption of a fixed 75
percent conversion factor for undrawn lines
reasonable? Do banking organizations have
the ability (including relevant data) to
develop their own estimate of EADs for such
facilities? Should banking organizations be
permitted to employ their own estimated
EADs, subject to supervisory approval?

A banking organization may only use
the top-down approach with approval of
its primary Federal supervisor. In
addition, the purchased receivables
would have to have been purchased
from unrelated, third party sellers and
the organization may not have
originated the credit exposures either
directly or indirectly. The receivables
must have been generated on an arm’s
length basis between the seller and the
obligor (intercompany accounts
receivable and receivables subject to
contra-accounts between firms that buy
and sell to each other would not
qualify). Also, the receivables may not
have a remaining maturity of greater
than one year, unless they are fully
secured. The Agencies propose that the
bottom-up method would have to be
used for receivables to any single
obligor, or to any group of related
obligors, that aggregate to more than $1
million.
Capital Charge for Dilution Risk
When dilution is a material risk
factor,24 purchased receivables would
be subject to a separate capital charge
for that risk. The dilution capital charge
24 If

dilution risk is immaterial there would be no
additional capital charge.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

may be calculated at the level of each
individual receivable and then
aggregated, or, for a pool of receivables,
at the level of the pool as a whole. The
capital charge for dilution risk would be
calculated using the wholesale A–IRB
formula and the following parameters:
EAD would be equal to the gross
amount of receivable(s) balance(s); LGD
would be 100 percent; M would be the
(exposure weighted-average) effective
remaining maturity of the exposure(s);
and PD would be the expected dilution
loss rate, defined as total expected
dilution losses over the remaining term
of the receivable(s) divided by EAD.25
Expected dilution losses would be
computed on a stand-alone basis; that is,
under the assumption of no recourse or
other support from the seller or thirdparty guarantors.
The following table illustrates the
dilution risk capital charges (per dollar
of EAD) implied by this approach for a
hypothetical pool of purchased
receivables having a remaining maturity
of one year or less. As can be seen, the
proposal implies capital charges for
dilution risk that are many multiples of
expected dilution losses.

CAPITAL REQUIREMENTS
[In percentage points]

Expected dilution loss rate

Dilution risk
capital charge
(per dollar of
EAD, percent)

0.05 percent ..........................
0.10 percent ..........................
0.25 percent ..........................
0.50 percent ..........................
1.00 percent ..........................
2.00 percent ..........................
5.00 percent ..........................
10.00 percent ........................

2.05
3.42
6.41
9.77
14.03
19.03
28.45
41.24

The Agencies seek comment on the
proposed methods for calculating dilution
risk capital requirements. Does this
methodology produce capital charges for
dilution risk that seem reasonable in light of
available historical evidence? Is the
wholesale A–IRB capital formula appropriate
for computing capital charges for dilution
risk?
In particular, is it reasonable to attribute
the same asset correlations to dilution risk as
are used in quantifying the credit risks of
wholesale exposures within the A–IRB
framework? Are there alternative method(s)
for determining capital charges for dilution
risk that would be superior to that set forth
above?

Minimum Requirements
The Agencies propose to apply
standards for the estimation of risk
25 If the remaining term exceeds one year, the
expected dilution loss rate would be specified at an
annual rate.

PO 00000

Frm 00025

Fmt 4701

Sfmt 4702

45923

inputs and expected dilution losses and
for the control and risk management
systems associated with purchased
receivables programs that are consistent
with the general guidance contained in
the New Accord. These standards will
aim to ensure that risk input and
expected dilution loss estimates are
reliable and consistent over time, and
reflect all relevant information that is
available to the acquiring banking
organization. The minimum operational
requirements are intended to ensure that
the acquiring banking organization has
a valid legal claim to cash proceeds
generated by the receivables pool, that
the pool and cash proceeds are closely
monitored and controlled, and that
systems are in place to identify and
address seller, servicer, and other
potential risks. A more detailed
discussion of these requirements will be
provided when the Agencies release
draft examination guidance dealing with
purchased receivables programs.
The Agencies seek comment on the
appropriate eligibility requirements for using
the top-down method. Are the proposed
eligibility requirements, including the $1
million limit for any single obligor,
reasonable and sufficient?
The Agencies seek comment on the
appropriate requirements for estimating
expected dilution losses. Is the guidance set
forth in the New Accord reasonable and
sufficient?

Risk Mitigation
For purposes of reducing the capital
charges for credit risk or dilution risk
with respect to purchased receivables,
purchase discounts, guarantees, and
other risk mitigants may be recognized
through the same framework used
elsewhere in the A–IRB approach.
Credit Risk Mitigation Techniques
The New Accord takes account of the
risk-mitigating effects of both financial
and nonfinancial collateral, as well as
guarantees, including credit derivatives.
For these risk mitigants to be recognized
for regulatory capital purposes, the
banking organization must have in place
operational procedures and risk
management processes that ensure that
all documentation used in
collateralizing or guaranteeing a
transaction is binding on all parties and
legally enforceable in all relevant
jurisdictions. The banking organization
must have conducted sufficient legal
review to verify this conclusion, must
have a well-founded legal basis for the
conclusion, and must reconduct such a
review as necessary to ensure
continuing enforceability.

E:\FR\FM\04AUP2.SGM

04AUP2

45924

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

Adjusting LGD for the Effects of
Collateral

could not be based solely upon the
collateral’s estimated market value.

A banking organization would be able
to take into account the risk-mitigating
effect of collateral in its internal
estimates of LGD, provided the
organization has established internal
requirements for collateral management,
operational procedures, legal certainty,
and risk management processes that
ensure that:
(1) The legal mechanism under which
the collateral is pledged or transferred
ensures that the banking organization
has the right to liquidate or take legal
possession of the collateral in a timely
manner in the event of the default,
insolvency, or bankruptcy (or other
defined credit event) of the obligor and,
where applicable, the custodian holding
the collateral;
(2) The banking organization has
taken all steps necessary to fulfill legal
requirements to secure the
organization’s interest in the collateral
so that it has and maintains an
enforceable security interest;
(3) The banking organization has clear
and robust procedures for the timely
liquidation of collateral to ensure
observation of any legal conditions
required for declaring the default of the
borrower and prompt liquidation of the
collateral in the event of default;
(4) The banking organization has
established procedures and practices for
(i) conservatively estimating, on a
regular ongoing basis, the market value
of the collateral, taking into account
factors that could affect that value (for
example, the liquidity of the market for
the collateral and obsolescence or
deterioration of the collateral), and (ii)
where applicable, periodically verifying
the collateral (for example, through
physical inspection of collateral such as
inventory and equipment); and
(5) The banking organization has in
place systems for requesting and
receiving promptly additional collateral
for transactions whose terms require
maintenance of collateral values at
specified thresholds.
In reflecting collateral in the LGD
estimate, the banking organization
would need to consider the extent of
any dependence between the risk of the
borrower and that of the collateral or
collateral provider. The banking
organization’s assessment of LGD would
have to address in a conservative way
any significant degrees of dependence,
as well as any currency mismatch
between the underlying obligation and
the collateral. The LGD estimates would
have to be grounded in historical
recovery rates on the collateral and

Repo-Style Transactions Subject to
Master Netting Agreements
Repo-style transactions include
reverse repurchase agreements and
repurchase agreements and securities
lending and borrowing transactions,
including those executed on an
indemnified agency basis.26 Many of
these transactions are conducted under
a bilateral master netting agreement or
equivalent arrangement. The effects of
netting arrangements generally would
be recognized where the banking
organization takes into account the riskmitigating effect of collateral through an
adjustment to EAD. To qualify for the
EAD adjustment treatment, the repostyle transaction would have to be
marked-to-market daily and be subject
to a daily margin maintenance
requirement. Further, the repo-style
transaction would have to be
documented under a qualifying master
netting agreement that would have to:
(1) Provide the non-defaulting party
the right to terminate and close out
promptly all transactions under the
agreement upon an event of default,
including in the event of insolvency or
bankruptcy of the counterparty;
(2) Provide for the netting of gains and
losses on transactions (including the
value of any collateral) terminated and
closed out under the agreement so that
a single net amount is owed by one
party to the other;
(3) Allow for the prompt liquidation
or setoff of collateral upon the
occurrence of an event of default; and
(4) Be, together with the rights arising
from the provisions required in (1) to (3)
above, legally enforceable in each
relevant jurisdiction upon the
occurrence of an event of default and
regardless of the counterparty’s
insolvency or bankruptcy.
Where a banking organization’s repostyle transactions do not meet these
requirements, it would not be able to
use the EAD adjustment method. Rather,
for each individual repo-style
transaction it would estimate an LGD
that takes into account the collateral
received. It would use the notional
amount of the transaction for EAD; it
would not take into account netting
effects for purposes of determining
either EAD or LGD.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

26 Some banking organizations, particularly those
that are custodians, lend, as agent, their customers’
securities on a collateralized basis. Typically, the
agent banking organization indemnifies the
customer againts risk of loss in the event the
borrowing counterparty defaults. Where such
indemnites are provided, the agent banking
organization has the same risks it would have if it
had entered into the transaction as principal.

PO 00000

Frm 00026

Fmt 4701

Sfmt 4702

The method for determining EAD for
repo-style transactions, described
below, is essentially the determination
of an unsecured loan equivalent
exposure amount to the counterparty.
Thus, no collateral effects for these
transactions would be recognized
through LGD; rather, the applicable LGD
would be the one the banking
organization would estimate for an
unsecured exposure to the counterparty.
To determine EAD, the banking
organization would add together its
current exposure to the counterparty
under the master netting arrangement
and a measure for PFE to the
counterparty under the master netting
arrangement. The current exposure
would be the sum of the market values
of all securities and cash lent, sold
subject to repurchase, or pledged as
collateral to the counterparty under the
master netting agreement, less the sum
of the market values of all securities and
cash lent, sold subject to repurchase, or
pledged as collateral by the
counterparty. The PFE calculation
would be based on the market price
volatilities of the securities delivered to,
and the securities received from, the
counterparty, as well as any foreign
exchange rate volatilities associated
with any cash or securities delivered or
received.
Banking organizations would use a
VaR-type measure for determining PFE
for repo-style transactions subject to
master netting agreements. Banking
organizations would be required to use
a 99th percentile, one-tailed confidence
interval for a five-day holding period
using a minimum one-year historical
observation period of price data.
Banking organizations would have to
update their data sets no less frequently
than once every three months and
reassess them whenever market prices
are subject to material changes. The
illiquidity of lower-quality instruments
would have to be taken into account
through an upward adjustment in the
holding period where the five-day
holding period would be inappropriate
given the instrument’s liquidity. No
particular model would be prescribed
for the VaR-based measure, but the
model would have to capture all
material risks for included transactions.
Banking organizations using a VaRbased approach to measuring PFE
would be permitted to take into account
correlations in the price volatilities
among instruments delivered to the
counterparty, among instruments
received from the counterparty, as well
as between the two sets of instruments.
The VaR-based approach for calculating
PFE for repo-style transactions would be
available to all banking organizations

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
that received supervisory approval for
an internal market risk model under the
market risk capital rules. Other banking
organizations could apply separately for
supervisory approval to use their
internal VaR models for calculation of
PFE for repo-style transactions.
A banking organization would use the
following formula to determine EAD for
each counterparty with which it has a
master netting agreement for repo-style
transactions.
EAD = max {0, [(è E ¥ èC) + (VaR
output from internal market risk
model × multiplier)]}
Where:
E denotes the current value of the
exposure (that is, all securities and
cash delivered to the counterparty);
and
C denotes the current value of the
collateral received (that is, all
securities and cash received from
the counterparty).
The multiplier in the above formula
would be determined based on the
results of the banking organization’s
backtesting of the VaR output. To
backtest the output, the banking
organization would be required to
identify on an annual basis twenty
counterparties that include the ten
largest as determined by the banking
organization’s own exposure
measurement approach and ten others
selected at random. For each day and for
each of the twenty counterparties, the
banking organization would compare
the previous day’s VaR estimate for the
counterparty portfolio to the change in
the current exposure of the previous
day’s portfolio. This change represents
the difference between the net value of
the previous day’s portfolio using
today’s market prices and the net value
of that portfolio using the previous day’s
market prices. Where this difference
exceeds the previous day’s VaR
estimate, an exception would have
occurred.
At the end of each quarter, the
banking organization would identify the
number of exceptions it has observed for
its twenty counterparties over the most
recent 250 business days, that is, the
number of exceptions in the most recent
5000 observations. Depending on the
number of exceptions, the output of the
VaR model would be scaled up using a
multiplier as provided in the table
below.
Zone

Number of
exceptions

Green Zone ...
Yellow Zone ..

0–99 ..............
100–119 ........
120–139 ........
140–159 ........

VerDate jul<14>2003

20:44 Aug 01, 2003

Multiplier
None (=1)
2.0
2.2
2.4

Jkt 200001

Zone

Number of
exceptions

Red Zone ......

160–179 ........
180–199 ........
200 or more ..

Multiplier
2.6
2.8
3.0

The Agencies seek comments on the
methods set forth above for determining
EAD, as well as on the proposed backtesting
regime and possible alternatives banking
organizations might find more consistent
with their internal risk management
processes for these transactions. The
Agencies also request comment on whether
banking organizations should be permitted to
use the standard supervisory haircuts or own
estimates haircuts methodologies that are
proposed in the New Accord.

Guarantees and Credit Derivatives
The Agencies are proposing that
banking organizations reflect the credit
risk mitigating effects of guarantees and
credit derivatives through adjusting the
PD or the LGD estimate (but not both)
of the underlying obligation that is
protected. The banking organization
would be required to assign the
borrower and guarantor to an internal
rating in accordance with the minimum
requirements set out for unguaranteed
(unhedged) exposures, both prior to the
adjustments and on an ongoing basis.
The organization also would be required
to monitor regularly the guarantor’s
condition and ability and willingness to
honor its obligation. For guarantees on
retail exposures, these requirements
would also apply to the assignment of
an exposure to a pool and the estimation
of the PD of the pool.
For purposes of reflecting the effect of
guarantees in regulatory capital
requirements, the Agencies are
proposing that a banking organization
have clearly specified criteria for
adjusting internal ratings or LGD
estimates—or, in the case of retail
exposures, for allocating exposures to
pools to reflect use of guarantees and
credit derivatives—that take account of
all relevant information. The adjustment
criteria would have to require a banking
organization to (i) meet all minimum
requirements for an unhedged exposure
when assigning borrower or facility
ratings to guaranteed/hedged exposures;
(ii) be plausible and intuitive; (iii)
consider the guarantor’s ability and
willingness to perform under the
guarantee; (iv) consider the extent to
which the guarantor’s ability and
willingness to perform and the
borrower’s ability to repay may be
correlated (that is, the degree of wrongway risk); and (v) consider the payout
structure of the credit protection and
conservatively assess its effect on the
level and timing of recoveries. The

PO 00000

Frm 00027

Fmt 4701

Sfmt 4702

45925

banking organization also would be
required to consider any residual risk to
the borrower that may remain—for
example, a currency mismatch between
the credit protection and the underlying
exposure.
Banking organizations would be
required to make adjustments to alter
PD or LGD estimates in a consistent way
for a given type of guarantee or credit
derivative. In all cases, the adjusted risk
weight for the hedged obligation could
not be less than the risk weight
associated with a comparable direct
exposure on the protection provider. As
a practical matter, this guarantor risk
weight floor on the risk weight of the
hedged obligation would require a
banking organization first to determine
the risk weight on the hedged obligation
using the adjustment it has made to the
PD or LGD estimate to reflect the hedge.
The banking organization would then
compare that risk weight to the risk
weight assigned to a direct obligation of
the guarantor. The higher of the two risk
weights would then be used to
determine the risk-weighted asset
amount of the hedged obligation.
Notwithstanding the guarantor risk
weight floor, the proposed approach
gives institutions a great deal of
flexibility in their methodology for
recognizing the risk-reducing effects of
guarantees and credit derivatives. At the
same time, the approach does not
differentiate between various types of
guarantee structures, which may have
widely varying characteristics, that a
banking organization may use. For
example, a company to company
guarantee, such as a company’s
guarantee of an affiliate or a supplier, is
fundamentally different from a
guarantee obtained from an unrelated
third party that is in the business of
extending financial guarantees.
Examples of the latter type of guarantee
include standby letters of credit,
financial guarantee insurance, and
credit derivatives. These products tend
to be standardized across institutions
and, thus, arguably should be
recognized for capital purposes in a
consistent fashion across institutions.
The problem of inconsistent treatment
could be exacerbated in the case of
protection in the form of credit
derivatives, which are tradable and
which further can be distinguished by
their characteristic of allowing a
banking organization to have a recovery
claim on two parties, the obligor and the
derivative counterparty, rather than just
one.
Industry comment is sought on whether a
more uniform method of adjusting PD or LGD
estimates should be adopted for various types
of guarantees to minimize inconsistencies in

E:\FR\FM\04AUP2.SGM

04AUP2

45926

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

treatment across institutions and, if so, views
on what methods would best reflect industry
practices. In this regard, the Agencies would
be particularly interested in information on
how banking organizations are currently
treating various forms of guarantees within
their economic capital allocation systems and
the methods used to adjust PD, LGD, EAD,
and any combination thereof.

Double Default Effects
The Agencies are proposing that
neither the banking organization’s
criteria nor rating process for
guaranteed/hedged exposures be
allowed to take into account so-called
‘‘double default’’ effects—that is, the
joint probability of default of the
borrower and guarantor. As a result of
not being able to recognize double
default probabilities, the adjusted risk
weight for the hedged obligation could
not be less than the risk weight
associated with a direct exposure on the
protection provider. The Agencies are
seeking comment on the proposed
nonrecognition of double default effects.
On June 10, 2003, the Federal Reserve
released a white paper on this issue
entitled, ‘‘Treatment of Double Default
and Double Recovery Effects for Hedged
Exposures Under Pillar I of the
Proposed New Basel Capital Accord.’’
Commenters are encouraged to take into
account the white paper in formulating
their responses to the ANPR.
The Agencies also are interested in
obtaining commenters’ views on
alternative methods for giving
recognition to double default effects in
a manner that is operationally feasible
and consistent with safety and
soundness. With regard to the latter,
commenters are requested to bear in
mind the concerns outlined in the
double default white paper, particularly
in connection with concentrations,
wrong-way risk (especially in stress
periods), and the potential for regulatory
capital arbitrage. In this regard,
information is solicited on how banking
organizations consider double default
effects on credit protection
arrangements in their economic capital
calculations and for which types of
credit protection arrangements they
consider these effects.
Requirements for Recognized
Guarantees and Credit Derivatives
The Agencies are not proposing any
restrictions on the types of eligible
guarantors or credit derivative
providers. Rather, a banking
organization would be required to have
clearly specified criteria for those
guarantors they will accept as eligible
for regulatory capital purposes. It is
proposed that guarantees and credit
derivatives recognized for regulatory

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

capital purposes: (1) Be required to
represent a direct claim on the
protection provider; (2) explicitly
reference specific exposures or classes
thereof; (3) be evidenced in writing
through a contract that is irrevocable by
the guarantor; (4) not have a clause that
would (i) allow the protection provider
unilaterally to cancel the credit
protection (other than in the event of
nonpayment or other default by the
protection buying banking organization)
or (ii) increase the effective cost of
credit protection as the credit quality of
the underlying obligor deteriorates; (5)
be in force until the underlying
obligation is satisfied in full (to the
amount and tenor of the guarantee); and
(6) be legally enforceable against the
guarantor in a jurisdiction where the
guarantor has assets to attach and
enforce a judgment.
The Agencies view the risk mitigating
benefits of conditional guarantees—that
is, guarantees that prescribe certain
conditions under which the guarantor
would not be obliged to perform—as
particularly difficult to quantify. The
Agencies are proposing that as a general
matter such guarantees would not be
recognized under the A–IRB approach.
In certain circumstances, however,
conditional guarantees could be
recognized where the banking
organization can demonstrate that its
assignment criteria fully reflect the
reduction in credit risk mitigation
arising from the conditionality and that
the guarantee provides a meaningful
degree of credit protection.
Additional Requirements for
Recognized Credit Derivatives
The Agencies are proposing that
credit derivatives, whether in the form
of credit default swaps or total return
swaps, recognized for A–IRB risk-based
capital purposes meet additional
criteria. The credit events specified by
the contracting parties would be
required to include at a minimum: (i)
Failure to pay amounts due under the
terms of the underlying obligation; (ii)
bankruptcy, insolvency, or inability of
the obligor to pay its debt; and (iii)
restructuring of the underlying
obligation that involves forgiveness or
postponement of principal, interest, or
fees that results in a credit loss.
With regard to restructuring events,
the Agencies note that the New Accord
suggests that a banking organization
may not need to include restructuring
credit events when it has complete
control over the decision of whether or
not there will be a restructuring of the
underlying obligation. This would
occur, for example, where the hedged
obligation requires unanimous consent

PO 00000

Frm 00028

Fmt 4701

Sfmt 4702

of the creditors for a restructuring. The
Agencies have concerns that this
approach could have the incidental
effect of dictating terms in underlying
obligations in ways that over time could
diverge from creditors’ business needs.
The Agencies also question whether
such clauses actually eliminate
restructuring risk on the underlying
obligation, particularly as many credit
derivatives hedge only a small portion
of a banking organization’s exposure to
the underlying obligation.
The Agencies invite comment on this
issue, as well as consideration of an
alternative approach whereby the notional
amount of a credit derivative that does not
include restructuring as a credit event would
be discounted. Comment is sought on the
appropriate level of discount and whether
the level of discount should vary on the basis
of, for example, whether the underlying
obligor has publicly outstanding rated debt or
whether the underlying obligor is an entity
whose obligations have a relatively high
likelihood of restructuring relative to default
(for example, a sovereign or PSE). Another
alternative that commenters may wish to
discuss is elimination of the restructuring
requirement for credit derivatives with a
maturity that is considerably longer—for
example, two years—than that of the hedged
obligation.

Consistent with the New Accord, the
Agencies are proposing not to recognize
credit protection from total return swaps
where the hedging banking organization
records net payments received on the
swap as net income, but does not record
offsetting deterioration in the value of
the hedged obligation either through
reduction in fair value or by an addition
to reserves. The Agencies are
considering imposing similar nonrecognition on credit default swaps
where mark-to-market gains in value are
recognized in income and, thus, in Tier
1 capital, but no offsetting deterioration
in the hedged obligation is recorded.
(This situation generally would not arise
where both the hedged obligation and
the credit default swap are recorded in
the banking book because under GAAP
increases in the swap’s value are
recorded in the Other Comprehensive
Income account, which is not included
in regulatory capital.)
Comment is sought on this matter, as well
as on the possible alternative treatment of
recognizing the hedge in these two cases for
regulatory capital purposes but requiring that
mark-to-market gains on the credit derivative
that have been taken into income be
deducted from Tier 1 capital.

Mismatches in Credit Derivatives
Between Reference and Underlying
Obligations
The Agencies are proposing to
recognize credit derivative hedges for

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
A–IRB capital purposes only where the
reference obligation on which the
protection is based is the same as the
underlying obligation except where: (1)
the reference obligation ranks pari
passu with or is more junior than the
underlying obligation, and (2) the
underlying obligation and reference
obligation share the same obligor and
legally enforceable cross-default or
cross-acceleration clauses are in place.
Treatment of Maturity Mismatch
The Agencies are proposing to
recognize on a discounted basis
guarantees and credit derivatives that
have a shorter maturity than the hedged
obligation. A guarantee or credit
derivative with less than one-year
remaining maturity that does not have a
matching maturity to the underlying
obligation, however, would not be
recognized. The formula for discounting
the amount of a maturity-mismatched
hedge that is recognized is proposed as
follows:
Pa = P * t/T
Where:
Pa denotes the value of the credit
protection adjusted for maturity
mismatch;
P denotes the amount of the credit
protection;
t denotes the lesser of T and the
remaining maturity of the hedge
arrangement, expressed in years;
and
T denotes the lesser of five and the
remaining maturity of the
underlying obligation, expressed in
years.
The Agencies have concerns that the
proposed formulation does not appropriately
reflect distinctions between bullet and
amortizing underlying obligations. Comment
is sought on the best way of making such a
distinction, as well as more generally on
alternative methods for dealing with the
reduced credit risk coverage that results from
a maturity mismatch.

Treatment of Counterparty Risk for
Credit Derivative Contracts
The Agencies are proposing that the
EAD for derivative contracts included in
either the banking book or trading book
be determined in accordance with the
rules for calculating the credit
equivalent amount for such contracts set
forth under the general risk-based
capital rules. The Agencies are
proposing to include in the types of
derivative contracts covered under these
rules credit derivative contracts
recorded in the trading book.
Accordingly, where a banking
organization buys or sells a credit
derivative through its trading book, a
counterparty credit risk capital charge

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

would be imposed based on the
replacement cost plus the following
add-on factors for PFE:

45927

requirements for using an internal
model would have to be met on an
ongoing basis. An advanced approach
banking organization that is
Total return or
Protection
Protection
transitioning into an internal models
credit default
buyer
seller
approach to equity exposures or that
swap
(percent)
(percent)
fails to demonstrate compliance with
the minimum operational requirements
Qualifying Reffor using an internal models approach to
erence
Obligation* .....
5
**5 equity exposures would be required to
Non-Qualifying
develop a plan for compliance, obtain
Reference
approval of the plan from its primary
Obligation* .....
10
**10 Federal supervisor, and implement the
*The definition of qualifying would be the plan in a timely fashion. In addition, a
same as for the ‘‘qualifying’’ category for the banking organization’s primary Federal
treatment of specific risk for covered debt po- supervisor would have the authority to
sitions under the market risk capital rules.
**The protection seller of a credit default impose additional operational
swap would only be subject to the add-on fac- requirements on a case-by-case basis.
tor where the contract is subject to close-out Until it is fully compliant with all
upon the insolvency of the protection buyer applicable requirements, the banking
while the underlying obligor is still solvent.
organization would apply a minimum
The Agencies also are considering
300 percent risk weight to all publicly
applying a counterparty credit risk
traded equity investments (that is,
charge on all credit derivatives that are
equity investments that are traded on a
marked-to-market, including those
nationally recognized securities
recorded in the banking book. Such a
exchange) and a minimum 400 percent
treatment would promote consistency
risk weight to all other equity
with other OTC derivatives, which are
investments.
assessed the same counterparty credit
Positions Covered
risk charge regardless of where they are
booked.
All equity exposures held in the
Further, the Agencies note that, if
banking book, along with any equity
credit derivatives booked in the banking exposures in the trading book that are
book are not assessed a counterparty
not currently subject to a market risk
credit risk charge, banking organizations capital charge, would be subject to the
would be required to exclude these
A-IRB approach for equity exposures. In
derivatives from the net current
general, equity exposures are
exposure of their other derivative
distinguished from other types of
exposures to a counterparty for
exposures based on the economic
purposes of determining regulatory
substance of the exposure. Equity
capital requirements. On balance, the
exposures would include both direct
Agencies believe a better approach
and indirect ownership interests,
would be to align the net derivative
whether voting or non-voting, in the
exposure used for capital purposes with assets or income of a commercial
that used for internal risk management
enterprise or financial institution that
purposes to manage counterparty risk
are not consolidated or deducted for
exposure and collateralization thereof.
regulatory capital purposes. Holdings in
This approach would suggest imposing
funds containing both equity
a counterparty risk charge on all credit
investments and non-equity investments
derivative exposures that are marked to
would be treated either as a single
market, regardless of where they are
investment based on the majority of the
booked.
fund’s holdings or, where possible, as
separate and distinct investments in the
The Agencies are seeking industry views
on the PFE add-ons proposed above and their fund’s component holdings based on a
applicability. Comment is also sought on
‘‘look-through approach’’ (that is, based
whether different add-ons should apply for
on the individual component holdings).
different remaining maturity buckets for
An instrument generally would be
credit derivatives and, if so, views on the
considered to be an equity exposure if
appropriate percentage amounts for the addit (1) would qualify as Tier 1 capital
ons in each bucket.
under the general risk-based capital
Equity Exposures
rules if issued by a banking
Banking organizations using the A–
organization; (2) is irredeemable in the
IRB approach for any credit exposure
sense that the return of invested funds
would be required to use an internal
can be achieved only by the sale of the
models market-based approach to
investment or sale of the rights to the
calculate regulatory capital charges for
investment or in the event of the
equity exposures. Minimum
liquidation of the issuer; (3) conveys a
quantitative and qualitative
residual claim on the assets or income

PO 00000

Frm 00029

Fmt 4701

Sfmt 4702

E:\FR\FM\04AUP2.SGM

04AUP2

45928

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

of the issuer; and (4) does not embody
an obligation on the part of the issuer.
An instrument that embodies an
obligation on the part of the issuer
would be considered an equity exposure
if the instrument meets any of the
following conditions: (1) The issuer may
defer indefinitely the settlement of the
obligation; (2) the obligation requires, or
permits at the issuer’s discretion,
settlement by the issuance of a fixed
number of the issuer’s equity interests;
(3) the obligation requires, or permits at
the issuer’s discretion, settlement by the
issuance of a variable number of the
issuer’s equity interests, and all things
being equal, any change in the value of
the obligation is attributable to,
comparable to, and in the same
direction as, the change in value of a
fixed number of the issuer’s equity
shares; or (4) the holder has the option
to require that the obligation be settled
by issuance of the issuer’s equity
interests, unless the banking
organization’s primary Federal
supervisor has opined in writing that
the instrument should be treated as a
debt position.
Debt obligations and other securities,
derivatives, or other instruments
structured with the intent of conveying
the economic substance of equity
ownership would be considered equity
exposures for purposes of the A-IRB
capital requirements. For example,
options and warrants on equities and
short positions in equity securities
would be characterized as equity
exposures. If a debt instrument is
convertible into equity at the option of
the holder, it would be deemed equity
upon conversion. If such debt is
convertible at the option of the issuer or
automatically by the terms of the
instrument, it would be deemed equity
from inception. In addition, instruments
with a return directly linked to equities
would be characterized as equity
exposures under most circumstances. A
banking organization’s primary Federal
supervisor would have the discretion to
allow a debt characterization of such an
equity-linked instrument, however, if
the instrument is directly hedged by an
equity holding such that the net
position does not involve material
equity risk to the holder. Equity
instruments that are structured with the
intent of conveying the economic
substance of debt holdings, or
securitization exposures would not be
considered equity exposures. For
example, some issuances of term
preferred stock may be more
appropriately characterized as debt.
In all cases, the banking
organization’s primary Federal
supervisor would have the discretion to

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

recharacterize debt holdings as equity
exposures or equity holdings as debt or
securitization exposures for regulatory
capital purposes.
The Agencies encourage comment on
whether the definition of an equity exposure
is sufficiently clear to allow banking
organizations to make an appropriate
determination as to the characterization of
their assets.

Materiality
As noted above, a banking
organization that is required or elects to
use the A–IRB approach for any credit
portfolio would also generally be
required to use the A–IRB approach for
its equity exposures. However, if the
aggregate equity holdings of a banking
organization are not material in amount,
the organization would not be required
to use the A–IRB approach to equity
exposures. For this purpose, a banking
organization’s equity exposures
generally would be considered material
if their aggregate carrying value,
including holdings subject to exclusions
and transitional provisions (as described
below), exceeds 10 percent of the
organization’s Tier 1 and Tier 2 capital
on average during the prior calendar
year. To address concentration
concerns, however, the materiality
threshold would be lowered to 5 percent
of the banking organization’s Tier 1 and
Tier 2 capital if the organization’s equity
portfolio consists of less than ten
individual holdings. Banking
organizations would risk weight at 100
percent equity exposures exempted
from the A–IRB equity treatment under
a materiality threshold.
Comment is sought on whether the
materiality thresholds set forth above are
appropriate. Exclusions from the A–IRB
Equity Capital Charge

Zero and Low Risk Weight Investments
The New Accord provides that
national supervisors may exclude from
the A–IRB capital charge those equity
exposures to entities whose debt
obligations qualify for a zero risk weight
under the New Accord’s standardized
approach for credit risk. Entities whose
debt obligations qualify for a zero risk
weight generally include (i) sovereigns
rated AAA to AA–; (ii) the BIS; (iii) the
IMF; (iv) the European Central Bank; (v)
the European Community; and (vi) highquality multilateral development banks
(MDBs) with strong shareholder
support.27 The Agencies intend to
27 These are, at present, the World Bank group
comprised of the International Bank for
Reconstruction and Development and the
International Finance Corporation, the Asian
Development Bank, the African Development Bank,
the European Bank for Reconstruction and

PO 00000

Frm 00030

Fmt 4701

Sfmt 4702

exclude from the A-IRB equity capital
charge equity investments in these
entities. Instead, these investments
would be risk weighted at zero percent
under the A-IRB approach.
In addition, the Agencies are
proposing to exempt from the A-IRB
equity capital charge investments in
non-central government public-sector
entities (PSEs) that are not traded
publicly and generally are held as a
condition of membership. Examples of
such holdings include stock of a Federal
Home Loan Bank or a Federal Reserve
Bank. These investments would be riskweighted as they would be under the
general risk-based capital rules—20
percent or zero percent, respectively, in
the examples.
Comment is sought on whether other types
of equity investments in PSEs should be
exempted from the A–IRB capital charge on
equity exposures, and if so, the appropriate
criteria for determining which PSEs should
be exempted.

Legislated Program Equity Exposures
Under the New Accord, national
supervisors may exclude from the A–
IRB capital charge on equity exposures
certain equity exposures made under
legislated programs that involve
government oversight and restrictions
on the types or amounts of investments
that may be made (legislated program
equity exposures). Under the New
Accord, a banking organization would
be able to exclude from the A–IRB
capital charge on equity exposures
legislated program equity exposures in
an amount up to 10 percent of the
banking organization’s Tier 1 plus Tier
2 capital.
The Agencies propose that equity
investments by a banking organization
in a small business investment company
(SBIC) under section 302(b) of the Small
Business Investment Act of 1958 would
be legislated program equity exposures
eligible for the exclusion from the A–
IRB equity capital charge in an amount
up to 10 percent of the banking
organization’s Tier 1 plus Tier 2 capital.
A banking organization would be
required to risk weight at 100 percent
any amounts of legislated program
equity exposures that qualify for this
exclusion from the A–IRB equity capital
charge.
The Agencies seek comment on what
conditions might be appropriate for this
partial exclusion from the A–IRB equity
capital charge. Such conditions could
include limitations on the size and types of
Development, the Inter-American Development
Bank, the European Investment Bank, the Islamic
Development Bank, the Nordic Investment Bank,
the Caribbean Development Bank, and the Council
of Europe Development Bank.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
businesses in which the banking organization
invests, geographical limitations, or
limitations on the size of individual
investments.

U.S. banking organizations also make
investments in community development
corporations (CDCs) or community and
economic development entities (CEDEs)
that promote the public welfare. These
investments receive favorable tax
treatment and investment subsidies that
make their risk and return
characteristics markedly different (and
more favorable to investors) than equity
investments in general. Recognizing this
more favorable risk-return structure and
the importance of these investments to
promoting important public welfare
goals, the Agencies are proposing the
exclusion of all such investments from
the A–IRB equity capital charge. Unlike
the exclusion for SBIC exposures, the
exclusion of CDC and CEDE investments
would not be subject to a percentage of
capital limit. All CDC and CEDE equity
exposures would receive a 100 percent
risk weight.
The Agencies seek comment on whether
any conditions relating to the exclusion of
CDC/CEDE investments from the A-IRB
equity capital charge would be appropriate.
These conditions could serve to limit the
exclusion to investments in such entities that
meet specific public welfare goals or to limit
the amount of such investments that would
qualify for the exclusion from the A–IRB
equity capital charge. The Agencies also seek
comment on whether any other classes of
legislated program equity exposures should
be excluded from the A–IRB equity capital
charge.

Grandfathered Investments
Equity exposures held as of the date
of adoption of the final A–IRB capital
rule governing equity exposures would
be exempt from the A–IRB equity
capital charge for a period of ten years
from that date. A banking organization
would be required to risk weight these
holdings during the ten-year period at
100 percent. The investments that
would be considered grandfathered
would be equal to the number of shares
held as of the date of the final rule, plus
any shares that the holder acquires
directly as a result of owning those
shares, provided that any additional
shares do not increase the holder’s
proportional ownership share in the
company.
For example, if a banking organization
owned 100 shares of a company on the
date of adoption of the final rule, and
the issuer thereafter declared a pro rata
stock dividend of 5 percent, the entire
post-dividend holdings of 105 shares
would be exempt from the A–IRB equity
capital charge for a period of ten years
from the date of the adoption of the final

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

rule. However, if additional shares are
acquired such that the holder’s
proportional share of ownership
increases, the additional shares would
not be grandfathered. Thus, if a banking
organization owned 100 shares of a
company on the date of adoption of the
final rule and subsequently acquired an
additional 50 shares, the original 100
shares would be exempt from the A–IRB
equity capital charge for the ten-year
period from the date of adoption of the
final rule, but the additional 50 shares
would be immediately subject to the A–
IRB equity capital charge.
Description of Quantitative Principles
The primary focus of the A–IRB
approach to equity exposures is to
assess capital based on an internal
estimate of loss under extreme market
conditions on an institution’s portfolio
of equity holdings or, in simpler forms,
its individual equity investments. The
methodology or methodologies used to
compute the banking organization’s
estimated loss should be those used by
the institution for internal risk
management purposes. The model
should be fully integrated into the
banking organization’s risk management
infrastructure.
A banking organization’s use of
internal models would be subject to
supervisory approval and ongoing
review by the institution’s primary
Federal supervisor. Given the unique
nature of equity portfolios and
differences in modeling techniques, the
supervisory model review process
would be, in many respects, institutionspecific. The sophistication and nature
of the modeling technique used for a
particular type of equity exposure
should correspond to the banking
organization’s exposure, concentration
in individual equity issues of that type,
and the particular risk of the holding
(including any optionality). Institutions
would have to use an internal model
that is appropriate for the risk
characteristics and complexity of their
equity portfolios. The model would
have to be able to capture adequately all
of the material risks embodied in equity
returns, including both general market
risk and idiosyncratic (that is, specific)
risk of the institution’s equity portfolio.
In their evaluations of institutions’
internal models, the Agencies would
consider, among other factors, (a) the
nature of equity holdings, including the
number and types of equities (for
example, public, private, long, short);
(b) the risk characteristics and makeup
of institutions’ equity portfolio
holdings, including the extent to which
publicly available price information is
obtainable on the exposures; and (c) the

PO 00000

Frm 00031

Fmt 4701

Sfmt 4702

45929

level and degree of concentration.
Institutions with equity portfolios
containing holdings with values that are
highly nonlinear in nature (for example,
equity derivatives or convertibles)
would have to employ an internal
model designed to appropriately capture
the risks associated with these
instruments.
The Agencies recognize that the type
and sophistication of internal modeling
systems will vary across institutions due
to differences in the nature and
complexity of business lines in general
and equity exposures in particular.
Although the Agencies intend to use a
VaR methodology as a benchmark for
the internal model approach, the
Agencies recognize that some
institutions employ models for internal
risk management and capital allocation
purposes that, given the nature of their
equity holdings, can be more risksensitive than some VaR models. For
example, some institutions employ
rigorous historical scenario analysis and
other techniques in assessing the risk of
their equity portfolios. It is not the
Agencies’ intention to dictate the form
or operational details of banking
organizations’ risk measurement and
management practices for their equity
exposures. Accordingly, the Agencies
do not expect to prescribe any particular
type of model for computing A-IRB
capital charges for equity exposures.
For purposes of evaluating the A–IRB
equity capital charges produced by a
banking organization’s selected
methodology, the Agencies would
expect to use as a benchmark a VaR
methodology using a 99.0 percent (onetailed) confidence level of estimated
maximum loss over a quarterly time
horizon using a long-term sample
period. Moreover, A–IRB equity capital
charges would have to produce risk
weights for equity exposures that are at
least equal to a 200 percent risk weight
for publicly traded equity exposures,
and a 300 percent risk weight for all
other equity exposures.
VaR-based internal models must use a
historical observation period that
includes a sufficient amount of data
points to ensure statistically reliable and
robust loss estimates relevant to the
long-term risk profile of the institution’s
specific holdings. The data used to
represent return distributions should
reflect the longest sample period for
which data are available and should
meaningfully represent the risk profile
of the banking organization’s specific
equity holdings. The data sample
should be long-term in nature and, at a
minimum, should encompass at least
one complete equity market cycle
relevant to the institution’s holdings,

E:\FR\FM\04AUP2.SGM

04AUP2

45930

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

including both increases and decreases
in relevant equity values over a longterm data period. The data used should
be sufficient to provide conservative,
statistically reliable, and robust loss
estimates that are not based purely on
subjective or judgmental considerations.
The parameters and assumptions used
in a VaR model must be subject to a
rigorous and comprehensive regime of
stress-testing. Banking organizations
utilizing VaR models would be required
to subject their internal model and
estimation procedures, including
volatility computations, to either
hypothetical or historical scenarios that
reflect worst-case losses given
underlying positions in both public and
private equities. At a minimum, banking
organizations that use a VaR model
would be required to employ stress tests
to provide information about the effect
of tail events beyond the level of
confidence assumed in the internal
models approach.
Banking organizations using non-VaR
internal models that are based on stress
tests or scenario analyses would have to
estimate losses under worst-case
modeled scenarios. These scenarios
would have to reflect the composition of
the organization’s equity portfolio and
should produce capital charges at least
as large as those that would be required
to be held against a representative
market index under a VaR approach. For
example, for a portfolio consisting
primarily of publicly held equity
securities that are actively traded,
capital charges produced using
historical scenario analyses would have
to be greater than or equal to capital
charges produced by a baseline VaR
approach for a major index that is
representative of the institution’s
holdings.
The measure of an equity exposure on
which A–IRB capital requirements
would be based would be the value of
the equity presented in a banking
organization’s financial statements. For
investments held at fair value, the
exposure amount would be equal to the
fair value presented in the balance
sheet. For investments held at the lower
of cost or market value, the exposure
amount would be equal to the cost or
market value presented in the balance
sheet.
The loss estimate derived from the
internal model would constitute the
A–IRB capital charge to be assessed
against the equity exposure. The A–IRB
equity capital charge would be
incorporated into an institution’s riskbased capital ratio through the
calculation of risk-weighted equivalent
assets. To convert the A–IRB equity
capital charge into risk-weighted

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

equivalent assets, a banking
organization would multiply the capital
charge by a factor of 12.5.
Consistent with the general risk-based
capital rules, 45 percent of the positive
change in value held in the tax-adjusted
separate component of equity—that is,
45 percent of revaluation gains on
available-for-sale (AFS) equity
securities—would be includable in Tier
2 capital under the A–IRB framework.
Comment is specifically sought on whether
the measure of an equity exposure under AFS
accounting continues to be appropriate or
whether a different rule for the inclusion of
revaluation gains should be proposed.

C. Supervisory Assessment of A–IRB
Framework
A banking organization would have to
satisfy all the A–IRB infrastructure
requirements and supervisory standards
before it would be able to use the A–IRB
approach for calculating capital
requirements for credit risk. This
section describes key elements of the
framework on which the Agencies
propose to base the A–IRB qualifying
requirements for U.S. banking
organizations. The Agencies intend to
provide more detailed implementation
guidance in regard to these issues for
wholesale and retail exposures, as well
as for equity and securitization
exposures. As noted earlier, draft
guidance for corporate exposures that
identifies associated supervisory
standards was published elsewhere in
today’s Federal Register.
Overview of Supervisory Framework
Many of the supervisory standards are
focused on requirements for a banking
organization’s internal risk rating
system. Emphasis is placed on a
banking organization’s ability to rank
order and quantify risk in a consistent,
reliable and valid manner. In sum, a
banking organization’s internal risk
rating system would have to provide for
a meaningful differentiation of the
riskiness of borrowers, as well as the
risks inherent in individual
transactions. To ensure the reliability of
these estimates, internal risk rating
systems would need to be subject to
review by independent control units.
Data sources and estimation methods
used by banking organizations would
need to be sufficiently robust to support
the production of consistent
quantitative assessments of risk over
time. Finally, to ensure that ratings are
not derived solely for regulatory capital
purposes, internal risk rating systems
and quantification methods would need
to form an integral part of the
management of the institution, as
outlined below.

PO 00000

Frm 00032

Fmt 4701

Sfmt 4702

It is important to emphasize that the
Agencies believe that meeting the
A–IRB infrastructure requirements and
supervisory standards will require
significant efforts by banking
organizations. The A–IRB supervisory
standards will effectively ‘‘raise the bar’’
in regard to sound credit risk
management practices.
Rating System Design
The design of an internal risk rating
system is key to its effectiveness. By
definition, a rating system comprises all
of the processes that support the
assessment of credit risk, the assignment
of internal risk ratings, and the
quantification of default and loss
estimates. Banking organizations would
be able to rely on one or more systems
for assessing their credit risk exposures.
When this is the case, the banking
organization would have to demonstrate
that each system used for A–IRB capital
purposes complies with the supervisory
standards.
The Agencies believe that banking
organizations’ internal rating systems
should accurately and consistently
differentiate degrees of risk. For
wholesale exposures, banking
organizations would need to have a
two-dimensional rating system that
separately assesses the risk of borrower
default, as well as transaction-specific
factors that focus on the amount that
would likely be collected in the event of
default. Such factors may include
whether an exposure is collateralized,
its seniority, and the product type. In
contrast to the individual evaluation
required for wholesale exposures, retail
exposures would be assessed on a pool
basis. Banking organizations would
need to group their retail exposures into
portfolio segments based on the risk
characteristics that they consider
relevant—for example borrower
characteristics such as credit scores or
transaction characteristics such as
product or collateral type. Delinquent or
defaulted exposures would need to be
separated from those that are current.
Banking organizations would be
required to define clearly their
wholesale rating categories and retail
portfolio segments. The clarity and
transparency of the ratings criteria are
critical to ensuring that ratings are
assigned in a consistent and reliable
manner. The Agencies believe it is
important for banking organizations to
document the operating procedures for
their internal risk rating system in
writing. For example, the
documentation should describe which
parties within the organization would
have the authority to approve
exceptions. Further, the documentation

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
would have to clearly specify the
frequency of review, as well as describe
the oversight to be provided by
management of the ratings process.
Banking organizations using the
A–IRB approach would need to be able
to generate sound measurements of the
key risk inputs to the A–IRB capital
formulas. Banking organizations would
be able to rely on data based either on
internal experience or generated by an
external source, as long as the banking
organization can demonstrate the
relevance of external data to its own
experience.
In assigning a rating to an obligor, a
banking organization must assess the
risk of default, taking into account
possible adverse events that might
increase the obligor’s likelihood of
default. The A–IRB supervisory
standards in the supervisory guidance
provide banking organizations with a
degree of flexibility in determining
precisely how to reflect adverse events
in obligor ratings. However, banking
organizations are required to clearly
articulate the approach chosen, and to
articulate the implications for capital
planning and for capital adequacy
during times of systematic economic
stress. The Agencies recognize that
banking organizations’ internal risk
rating systems may include a range of
statistical models or other methods to
assign borrower or facility ratings or to
estimate key inputs. The burden of
proof would remain on the banking
organization as to whether a specific
model or procedure satisfies the
supervisory standards.
Risk Rating System Operations
The risk rating system would have to
form an integral part of the loan
approval process wherein ratings are
assigned to all borrowers, guarantors, or
facilities depending upon whether the
extension of credit is wholesale or retail
in nature. Any deviations from policies
that govern the assignment of ratings
must be clearly documented and
monitored.
Data maintenance is another key
aspect of risk rating system operations.
Banking organizations would be
expected to collect and store data on key
borrower and facility characteristics.
The data would have to be sufficiently
detailed to allow for future
reconsideration of the way in which
obligors and facilities have been
allocated to grades. Furthermore,
banking organizations would have to
collect, retain, and disclose data on
aspects of their internal ratings as
described under the disclosure section
of this proposal.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

Banking organizations would be
required to have in place sound stress
testing processes for use in the
assessment of capital adequacy. Stress
testing would have to involve
identifying possible events or future
changes in economic conditions that
could have unfavorable effects on a
banking organization’s credit exposures.
Specifically, institutions would need to
assess the effect of certain specific
conditions on their A–IRB regulatory
capital requirements. The choice of test
to be employed would remain with the
individual banking organization
provided the method selected is
meaningful and reasonably
conservative.
Corporate Governance and Oversight
The Agencies view the involvement of
the board of directors and management
as critical to the successful
implementation of the A–IRB approach.
The board of directors and management
would be responsible for maintaining
effective internal controls over the
banking organization’s information
systems and processes for assessing
adequacy of regulatory capital and
determining regulatory capital charges
consistent with this ANPR. All
significant aspects of the rating and
estimation processes would have to be
approved by the banking organization’s
board of directors or a designated
committee thereof and senior
management. These parties would need
to be fully aware of whether the system
complies with the supervisory
standards, makes use of the necessary
data, and produces reliable quantitative
estimates. Ongoing management reports
would have to accurately capture the
performance of the rating system.
Oversight would also need to involve
independent credit risk control units
responsible for ensuring the
performance of the rating system, the
accuracy of the ratings and parameter
estimates, and overall compliance with
supervisory standards and capital
regulations. The Agencies believe it is
critical that such units remain
functionally independent from the
personnel and management responsible
for originating credit exposures. Among
other responsibilities, the control units
should be charged with testing and
monitoring the appropriateness of the
rating scale, verifying the consistent use
of ratings for a given exposure type
across the organization, and reviewing
and documenting any changes to be
made to the system.
Use of Internal Ratings
To qualify to use the A–IRB
framework, a banking organization’s

PO 00000

Frm 00033

Fmt 4701

Sfmt 4702

45931

rating systems would have to form an
integral part of its day-to-day credit risk
management process. The Agencies
expect that banking organizations would
rely on their internal risk rating systems
when making decisions about whether
to extend credit as well as in their
ongoing monitoring of credit exposures.
For example, ratings information would
have to be incorporated into other key
processes, such as reserving
determinations and when allocating
economic capital internally.
Risk Quantification
Ratings quantification is the process
of assigning values to the key risk
components of the A–IRB approach: PD,
LGD, EAD and M. With the exception of
M, the risk components are
unobservable and must be estimated.
The estimates would have to be
consistent with sound practice and
supervisory standards. Banking
organizations’ rating system review and
internal audit functions would need to
serve as control mechanisms that ensure
the process of rating assignments and
quantification are functioning according
to policy and that non-compliance or
weaknesses are identified.
Validation of Internal Estimates
An equally important element would
be a robust system for validating the
accuracy and consistency of a banking
organization’s rating system, as well as
the estimation of risk components. The
standards in the supervisory guidance
require that banking organizations use a
broad range of validation tools,
including evaluation of developmental
evidence, ongoing monitoring of rating
and quantification processes,
benchmarking against alternative
approaches, and comparison of
outcomes with estimates. Details of the
validation process would have to be
consistent with the operation of the
banking organization’s rating system
and data would have to be maintained
and updated to support oversight and
validation work. Banking organizations
would have to have well-articulated
standards for situations where
deviations of realized values from
expectations become significant enough
to call the validity of the estimates into
question. Rating systems with
appropriate data and oversight feedback
mechanisms should create an
environment that promotes integrity and
improvements in the rating system over
time.
U.S. Supervisory Review
The primary Federal supervisor
would be responsible for evaluating an
institution’s initial and ongoing

E:\FR\FM\04AUP2.SGM

04AUP2

45932

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

compliance with the infrastructure
requirements and supervisory standards
for approval to use the A-IRB approach
for regulatory capital purposes. As
noted, the Agencies will be developing
and issuing specific implementation
guidance describing the supervisory
standards for wholesale, retail, equity
and securitization exposures. The
Agencies will issue the draft
implementation guidance for each
portfolio for public comment to ensure
that there is an opportunity for banking
organizations and others to provide
feedback on the Agencies’ expectations
in regard to A–IRB systems.
The Agencies seek comment on the extent
to which an appropriate balance has been
struck between flexibility and comparability
for the A-IRB requirements. If this balance is
not appropriate, what are the specific areas
of imbalance, and what is the potential
impact of the identified imbalance? Are there
alternatives that would provide greater
flexibility, while meeting the overall
objectives of producing accurate and
consistent ratings?
The Agencies also seek comment on the
supervisory standards contained in the draft
guidance on internal ratings-based systems
for corporate exposures. Do the standards
cover all of the key elements of an A–IRB
framework? Are there specific practices that
appear to meet the objectives of accurate and
consistent ratings but that would be ruled out
by the supervisory standards related to
controls and oversight? Are there particular
elements from the corporate guidance that
should be modified or reconsidered as the
Agencies draft guidance for other types of
credit?
In addition, the Agencies seek comment on
the extent to which these proposed
requirements are consistent with the ongoing
improvements banking organizations are
making in credit-risk management processes.

IV. Securitization
A. General Framework
This section describes the calculation
of A–IRB capital requirements for
securitization exposures. A
securitization exposure is any on- or offbalance-sheet position created by
aggregating and then tranching the risks
of a pool of assets, commitments, or
other instruments (underlying
exposures) into multiple financial
interests where, typically, the pooled
risks are not shared pro rata. The pool
may include one or more underlying
exposures. Examples include all
exposures arising from traditional and
synthetic securitizations, as well as
partial guarantee arrangements where
credit losses are not divided
proportionately among the parties (often
referred to as tranched cover). Assetand mortgage-backed securities
(including those privately issued and

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

those issued by GSEs such as Fannie
Mae and Freddie Mac), credit
enhancements, liquidity facilities, and
credit derivatives that have the
characteristics noted above would be
considered securitization exposures.
With ongoing advances in financial
engineering, the Agencies recognize that
securitization exposures having similar
risks can take different legal forms. For
this reason, both the designation of
positions as securitization exposures
and the calculation of A–IRB capital
requirements for securitization
exposures would be guided by the
economic substance of a given
transaction, rather than by its legal form.
Operational Criteria
Banking organizations would have to
satisfy certain operational criteria to be
eligible to use the A–IRB approach to
securitization exposures. Moreover, all
banking organizations that use the A–
IRB approach for the underlying
exposures that have been securitized
would have to apply the A–IRB
treatment for securitization exposures.
Minimum operational criteria would
apply to traditional and synthetic
securitizations. The Agencies propose to
establish supervisory criteria for
determining when, for risk-based capital
purposes, a banking organization may
treat exposures that it has originated
directly or indirectly as having been
securitized and, hence, not subject to
the same capital charge as if the banking
organization continued to hold the
assets. The Agencies anticipate these
supervisory criteria will be substantially
equivalent to the criteria contained in
the New Accord (paragraphs 516–520).
Broadly, these criteria are intended to
ensure that securitization transactions
transfer significant credit risk to third
parties and, in the case of traditional
securitizations, that each transaction
qualifies as a true sale under applicable
accounting standards.
The supervisory criteria also would
describe the types of clean-up calls that
may be incorporated within transactions
qualifying for the A–IRB securitization
treatment.28 Specifically, any clean-up
call would have to meet the following
conditions: (a) Its exercise is at the
discretion of the originating banking
organization; (b) it does not serve as a
credit enhancement; and (c) it is only
exercisable when 10 percent or less of
the original underlying portfolio or
reference portfolio value remains. If a
clean-up call does not meet all of these
28 In general terms, a clean-up call is an option
that permits an originating banking organization to
call the securitization exposures (for example, assetor mortgage-backed securities) before all of the
underlying exposures have been repaid.

PO 00000

Frm 00034

Fmt 4701

Sfmt 4702

criteria, the originating banking
organization would have to treat the
underlying exposures as if they had not
been securitized.
The Agencies seek comment on the
proposed operational requirements for
securitizations. Are the proposed criteria for
risk transference and clean-up calls
consistent with existing market practices?

Differences Between General A–IRB
Approach and the A–IRB Approach for
Securitization Exposures
In contrast to the proposed A–IRB
framework for traditional loans and
commitments, the A–IRB securitization
framework does not rely on a banking
organization’s own internal assessments
of the PD and LGD of a securitization
exposure. For securitization exposures
backed by pools of multiple assets, such
assessments require implicit or explicit
estimates of correlations among the
losses on those assets. Such correlations
are extremely difficult to estimate and
validate in an objective manner and on
a going-forward basis. For this reason,
the A–IRB framework generally would
not permit a banking organization to use
its internal risk assessments of PD or
LGD when such assessments depend,
implicitly or explicitly, on estimates of
correlation effects. The A–IRB treatment
of securitization exposures would rely
principally on two sources of
information, when available: (i) An
assessment of the securitization
exposure’s credit risk made by an
external rating agency; and (ii) the
A–IRB capital charge that would have
been assessed against the underlying
exposures had the exposures not been
securitized (the pool’s A–IRB capital
charge), along with other information
about the transaction.
B. Determining Capital Requirements
General Considerations
Because the information available to a
banking organization about a
securitization exposure often reflects the
organization’s role in a securitization
transaction, the Agencies are proposing
that the method of calculating the
A–IRB capital requirement for a
securitization exposure may depend on
whether a banking organization is an
originator or a third-party investor in
the securitization transaction. In
general, a banking organization would
be considered an originator of a
securitization if the organization
directly or indirectly originated the
underlying exposures or serves as the
sponsor of an asset-backed commercial
paper (ABCP) conduit or similar

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
program.29 If a banking organization is
not deemed an originator of a
securitization transaction, then it would
be considered an investor in the
securitization.
There are several methods for
determining the A–IRB capital
requirement for a securitization
exposure: the Ratings-Based Approach
(RBA), the Alternative RBA, the
Supervisory Formula Approach (SFA),
the Look-Through Approach, deduction
from Tier 1 capital, and deduction from
total capital. The following table
summarizes conditions under which a
banking organization would apply each
of these methods. In this table, KIRB
denotes the ratio of (a) the pool’s A–IRB
capital charge to (b) the notional or loan
equivalent amount of underlying
exposures in the pool.
Steps for Determining A–IRB Capital
Requirements for Securitization
Exposures
For an investing banking organization:
1. Deduct from total capital any
credit-enhancing interest-only strips
2. When an external or inferred rating
exists, apply the RBA
3. When an external or inferred rating
does not exist, do the following:
a. Subject to supervisory review and
approval, if the investing banking
organization can determine KIRB, then
calculate required capital as would an
originating banking organization using
the steps described in 2.a. below
b. Otherwise, deduct the exposure
from total capital
For an originating banking
organization:*
1. Deduct from Tier 1 capital any
increase in capital resulting from the
securitization transaction and deduct
from total capital any credit-enhancing
interest-only strips (net of deductions
from Tier 1 capital due to increases in
capital)
2. When an A–IRB approach exists for
the underlying exposures do the
following:
a. If KIRB can be determined:
i. For a securitization exposure (or
portion thereof) that is at or below KIRB,
deduct the exposure from total capital
ii. For a securitization exposure (or
portion thereof) that is above KIRB:
1. Apply the RBA whenever an
external or inferred rating is available
29 A banking organization is generally considered
a sponsor of an ABCP conduit or similar program
if, in fact or in substance, it manages or advises the
conduit program, places securities into the market
for the program, or provides liquidity support or
credit enhancements to the program.
* In addition to the capital treatments delineated,
an originating banking organization’s total A–IRB
capital charge with regard to any single
securitization transaction is subject to a maximum
or ceiling, as described later in this section.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

2. Otherwise, apply the SFA
b. If KIRB cannot be determined:
i. Apply the Look-Through Approach
if the exposure is an eligible liquidity
facility, subject to supervisory approval
ii. Otherwise, deduct the exposure
from total capital
3. When an A–IRB approach does not
exist for the underlying exposures do
the following:
a. Apply the Look-Through Approach
if the exposure is an eligible liquidity
facility, subject to supervisory approval
b. Otherwise, apply the Alternative
RBA
Deductions of Gain-on-Sale or Other
Accounting Elements That Result in
Increases in Equity Capital
Any increase in equity capital
resulting from a securitization
transaction (for example, a gain
resulting from FAS 140 accounting
treatment of the sale of assets) would be
deducted from Tier 1 capital. Such
deductions are intended to offset any
gain on sale or other accounting
treatments (‘‘gain on sale’’) that result in
an increase in an originating banking
organization’s shareholders’ equity and
Tier 1 capital. Over time, as banking
organizations, from an accounting
perspective, realize the increase in
equity that was booked at origination of
a securitization transaction through
actual receipt of cash flows, the amount
of the required deduction would be
reduced accordingly.
Banking organizations would have to
deduct from total capital any onbalance-sheet credit-enhancing interestonly strips (net of any increase in the
shareholders’ equity deducted from Tier
1 capital as described in the previous
paragraph).30 Credit-enhancing interestonly strips are defined in the general
risk-based capital rules and include
items, such as excess spread, that
represent subordinated cash flows of
future margin income.
Maximum Capital Requirement
Where an A–IRB approach exists for
the underlying exposures, an originating
banking organization’s total A–IRB
capital charge for exposures associated
with a given securitization transaction
would be subject to a maximum or
ceiling. This maximum A–IRB capital
charge would equal the pool’s A–IRB
capital charge plus any required
deductions, as described in the
preceding paragraphs. The aim of this
treatment is to ensure that an
institution’s effective A–IRB capital
30 Deductions other than of increases in equity
capital are to be taken 50 percent from Tier 1 capital
and 50 percent from Tier 2 capital.

PO 00000

Frm 00035

Fmt 4701

Sfmt 4702

45933

charge generally would not be greater
after securitization than before, while
also addressing the Agencies’ safety and
soundness concerns with respect to
credit-enhancing interest-only strips
and other capitalized assets.31
The proposed maximum A–IRB
capital requirement effectively would
reverse one aspect of the general riskbased capital rules for securitization
exposures referred to as residual
interests. Under the general risk-based
capital rules, banking organizations are
required to hold a dollar in capital for
every dollar in residual interest,
regardless of the capital requirement on
the underlying exposures. One of the
reasons the Agencies adopted the
‘‘dollar-for-dollar’’ capital treatment for
residual interests is that in many
instances the relative size of the
exposure retained by the originating
banking organization reveals additional
market information about the quality of
the underlying exposures and deal
structure that may not have been
captured in the capital requirement on
the underlying exposures, had those
exposures remained on the banking
organization’s balance sheet. The
Agencies will continue to review the
proposal for safety and soundness
considerations and may consider
retaining the current dollar-for-dollar
capital treatment for residual interests,
especially in those instances where an
originator retains first loss and other
deeply subordinated interests in
amounts that significantly exceed the
pool’s A–IRB capital charge plus
required deductions.
Comments are invited on the
circumstances under which the retention of
the treatment in the general risk-based capital
rules for residual interests for banking
organizations using the A–IRB approach to
securitization would be appropriate.
Should the Agencies require originators to
hold dollar-for-dollar capital against all
retained securitization exposures, even if this
treatment would result in an aggregate
amount of capital required of the originator
that exceeded the pool’s A–IRB capital
charge plus any applicable deductions?
Please provide the underlying rationale.

Investors
Third-party investors generally do not
have access to detailed, ongoing
information about the credit quality of
the underlying exposures in a
securitization. In such cases, investors
often rely upon credit assessments made
by external rating agencies. For a
securitization exposure held by an
investing banking organization, and
31 The maximum capital, requirement also
applies to investing banking organizations that
receive approval to use the SFA.

E:\FR\FM\04AUP2.SGM

04AUP2

45934

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

where an A–IRB treatment for the
underlying exposures exists, the
institution would use the Ratings-Based
Approach (RBA) described below if the
securitization exposure is externally
rated or if an inferred rating is available
(as defined in the RBA discussion
below). When neither an external rating
nor an inferred rating is available, an
investing banking organization would
compute the A–IRB capital charge for
the exposure using the methodology
described below for originating
institutions (subject to supervisory
review and approval). Otherwise, the
securitization exposure would be
deducted 50 percent from Tier 1 capital
and 50 percent from Tier 2 capital. The
Agencies anticipate that investing
banking organizations would apply the
RBA in the vast majority of situations.
Originators
This section presumes that an A–IRB
approach exists for the underlying
exposures. If no A–IRB treatment exists
for the underlying exposures, then an
originating banking organization
(originator) would use the Alternative
RBA discussed below.
In contrast to third-party investors,
banking organizations that originate
securitizations are presumed to have
much greater access to information
about the current credit quality of the
underlying exposures. In general, when
an originator retains a securitization
exposure, the A–IRB securitization
framework would require the institution
to calculate, on an ongoing basis, the
underlying exposure pool’s A–IRB
capital requirement had the underlying
exposures not been securitized (the
pool’s A–IRB capital charge), which
would be based on the notional dollar
amount of underlying exposures (the
size of the pool). The pool’s A–IRB
capital charge would be calculated
using the top-down or bottom-up
method applicable to the type(s) of
underlying exposure(s).32 As noted
above, the pool’s A–IRB capital charge
divided by the size of the pool is
denoted KIRB.
An originator also would be expected
to know: (a) Its retained securitization
exposure’s nominal size relative to the
size of the pool (the exposure’s
‘‘thickness,’’ denoted T); and (b) the
notional amount of all more junior
securitization exposures relative to the
size of the pool (the exposure’s ‘‘credit
enhancement level,’’ denoted L). The
32 For the purpose of determining the A–IRB
capital requirement for a securitization exposure,
the top-down method could be used regardless of
the maturity of the underlying exposures, provided
the other eligibility criteria for employing the topdown approach are satisfied.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

retained securitization exposure’s A–
IRB capital requirement depends on the
relationship between KIRB, T, and L. If
an originator cannot determine KIRB,
any retained securitization exposure
would be deducted from capital. For
eligible liquidity facilities (defined
below in the Look Through Approach)
provided to ABCP programs where a
banking organization lacks the
information necessary to calculate KIRB,
the Look-Through Approach described
below would be applied on a temporary
basis and subject to supervisory
approval.
Positions Below KIRB
An originating banking organization
would deduct from capital any retained
securitization exposure (or part thereof)
that absorbs losses at or below the level
of KIRB (that is, an exposure for which
L+T ≤ KIRB).33 This means that an
originating banking organization would
be given no risk-based capital relief
unless it sheds at least some exposures
below KIRB. Deduction from capital
would be required regardless of the
securitization exposure’s external rating.
This deduction treatment is in contrast
to the A–IRB capital treatment for
investors, who would be able to look to
the external (or inferred) rating of a
securitization exposure regardless of
whether the exposure was below KIRB.
While this disparate treatment of
originators and investors may be viewed
as inconsistent with the principle of
equal capital for equal risk, the Agencies
believe it is appropriate in order to
provide incentives for originating banks
to shed highly subordinated
securitization exposures. Such
exposures contain the greatest credit
risks. Moreover, these risks are difficult
to evaluate, and risk quantifications
tend to be highly sensitive to modeling
assumptions that are difficult to validate
objectively. The proposal to prevent an
originator from using the RBA for
securitization exposures below KIRB
reflects, in part, a concern by the
Agencies that the market discipline
underpinning an external credit rating
may be less effective when the rating
applies to a retained, non-traded
securitization exposure and is sought by
an originator primarily for regulatory
capital purposes.
The Agencies note that the specific
securitization exposures retained by an
originator that are subject to deduction
treatment could change over time in
response to variations in the credit
33 If an originator holds a securitization exposure
that straddles KIRB, the exposure must be
decomposed into two separate positions—one that
is above KIRB and another that is at or below KIRB.

PO 00000

Frm 00036

Fmt 4701

Sfmt 4702

quality of the underlying exposures. For
example, if the pool’s A–IRB capital
charge were to increase after the
inception of a securitization, additional
portions of securitization exposures
held by an originator may fall below
KIRB and, thus, become subject to
deduction. Therefore, when an
originator retains a first-loss
securitization exposure well in excess of
KIRB, the originator’s A–IRB capital
requirement on the exposure could
climb rapidly in the event of any
marked deterioration of the underlying
exposures. In general, an originator
could minimize variability in future
capital charges by minimizing the size
of any retained first-loss securitization
exposures.
Positions Above KIRB
When an originating banking
organization retains a securitization
exposure, or part thereof, that absorbs
losses above the KIRB amount (that is,
an exposure for which L + T > KIRB)
and the banking organization has not
already met the maximum capital
requirement for securitization exposures
described previously, the A–IRB capital
requirement for the exposure would be
calculated as follows. For securitization
exposures having an external or inferred
rating, the organization would calculate
its A–IRB capital requirement using the
RBA. However, if neither an external
rating nor an inferred rating is available,
an originator would be able to use the
SFA, subject to supervisory review and
approval. Otherwise, the organization
would deduct the securitization
exposure from total capital.
The Agencies seek comment on the
proposed treatment of securitization
exposures held by originators. In particular,
the Agencies seek comment on whether
originating banking organizations should be
permitted to calculate A–IRB capital charges
for securitizations exposures below the KIRB
threshold based on an external or inferred
rating, when available.
The Agencies seek comment on whether
deduction should be required for all nonrated positions above KIRB. What are the
advantages and disadvantages of the SFA
approach versus the deduction approach?

Capital Calculation Approaches
The Ratings-Based Approach (RBA)
The RBA builds upon the widespread
acceptance of external ratings by thirdparty investors as objective assessments
of a securitization exposure’s standalone credit risk. Certain minimum
requirements would have to be satisfied
in order for a banking organization to
rely on an external credit rating for
determining its A–IRB capital charge for
a securitization exposure. To be

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
recognized for regulatory capital
purposes, the external credit rating on a
securitization exposure would have to
be public and reflect the entire amount
of credit risk exposure the banking
organization has with regard to all
payments owed to it under the
exposure. In particular, if a banking
organization is owed both principal and
interest on a securitization exposure, the
external rating on the exposure would
have to fully reflect the credit risk
associated with both payment streams.
The Agencies propose to establish
criteria to ensure the integrity of
external ratings processes and banking
organizations’ use of these ratings under
the RBA. These criteria are expected to
be consistent with the proposed
guidance provided in the New Accord
(paragraph 525).
In certain circumstances, an ‘‘inferred
rating’’ may be used for risk weighting
a non-rated securitization exposure.
Similar to the general risk-based capital
rules, to qualify for use of an inferred
rating, a non-rated securitization
exposure would have to be senior in all
respects to a subordinate rated position
within the same securitization
transaction. Further, the junior rated
tranche would have to have an
equivalent or longer remaining maturity
than the non-rated exposure. Where
these conditions are met, the non-rated
exposure would be treated as if it had
the same rating (an ‘‘inferred rating’’) as
that of the junior rated tranche. External
and inferred ratings would be treated
equivalently.
Under the RBA, the capital charge per
dollar of a securitization exposure
would depend on: (i) The external rating
(or inferred rating) of the exposure, (ii)
whether the rating reflects a long-term

or short-term assessment of the
exposure’s credit risk, and (iii) a
measure of the effective number—or
granularity—of the underlying
exposures (N).34 For a securitization
exposure rated AA or AAA, the RBA
capital charge also would depend on a
measure of the exposure’s relative
seniority in the overall transaction (Q).35
Tables 1 and 2 below present the risk
weights that would result from the RBA
when a securitization exposure’s
external rating (or inferred rating)
represents a long-term or short-term
credit rating, respectively. In both
tables, the risk weights in column 2
would apply to AA and AAA-rated
securitization exposures when the
effective number of exposures (N) is 100
or more, and the exposure’s relative
seniority (Q) is greater than or equal to
0.1 + 25/N. If the underlying exposures
are retail exposures, N would be treated
as infinite and the minimum qualifying
value of Q would be 0.10. The Agencies
anticipate that these risk weights would
apply to AA and AAA-rated tranches of
most retail securitizations. Column 4
would apply only to securitizations
involving non-retail exposures for
which N is less than 6, and column 3
would apply in all other situations.
Within each table, risk weights
increase as external rating grades
decline. Under the Base Case (column
3), for example, the risk weights range
from 12 percent for AAA-rated
exposures to 650 percent for exposures
rated BB¥. This pattern of risk weights
is broadly consistent with analyses
employing standard credit risk models
and a range of assumptions regarding
correlation effects and the types of
exposures being securitized.36 These
analyses imply that, compared with a

45935

corporate bond having a given level of
stand-alone credit risk (for example, as
measured by its expected loss rate), a
securitization tranche having the same
level of stand-alone risk—but backed by
a reasonably granular and diversified
pool—will tend to exhibit more
systematic risk.37 This effect is most
pronounced for below-investment grade
tranches, and is the primary reason why
the RBA risk weights increase rapidly as
ratings deteriorate over this range—
much more rapidly than for similarly
rated corporate bonds. Similarly, for
highly granular pools, the risk weights
expected to apply to most AA and AAArated securitization exposures (7 percent
and 10 percent, respectively) decline
steeply relative to the risk weight
applicable to A-rated exposures (20
percent, column 3)—again, more so than
might be the case for similarly rated
corporate bonds. The decline in risk
weights as ratings improve over the
investment grade range is less
pronounced for the Base Case and for
tranches backed by non-granular pools
(column 4).
For securitization exposures rated
below BB¥, the proposed A–IRB
treatment—deduction from capital—
would be somewhat more conservative
than suggested by credit risk modeling
analyses. However, the Agencies believe
this more conservative treatment would
be appropriate in light of modeling
uncertainties and the tendency for
securitization exposures in this range, at
least at the inception of the
securitization transaction, to be nontraded positions retained by an
originator because they cannot be sold
at a reasonable price.

TABLE 1.—ABS RISK WEIGHTS BASED ON LONG-TERM EXTERNAL CREDIT ASSESSMENTS
External rating (illustrative)

Thick tranches backed by
highly granular pools

Base case

AAA ..................................................................................
AA ....................................................................................
A ......................................................................................
BBB+ ................................................................................
BBB ..................................................................................
BBB¥ ..............................................................................
BB+ ..................................................................................
BB ....................................................................................
BB¥ ................................................................................
Below BB¥ .....................................................................

7% .....................................
10% ...................................
N/A .....................................
N/A .....................................
N/A .....................................
N/A .....................................
N/A .....................................
N/A .....................................
N/A .....................................
N/A .....................................

12% ...................................
15% ...................................
20% ...................................
50% ...................................
75% ...................................
100% .................................
250% .................................
425% .................................
650% .................................
Deduction ..........................

34 N is defined more formally in the discussion
below of the Supervisory Formula Approach.
35 Q is defined as the total size of all
securitization exposures rated at least AA¥ that are
pari passu or junior to the exposure of interest,
measured relative to the size of the pool and

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

expressed as a decimal. Thus, for a securitization
transaction having an AAA-rated tranche in the
amount of 70 percent of the pool, an AAA-rated
tranche of 10 percent, a BBB-rated tranche of 10
percent, and a non-rated tranche of 10 percent, the
values of Q associated with these positions would
be 0.80, 0.10, 0, and 0, respectively.

PO 00000

Frm 00037

Fmt 4701

Sfmt 4702

Tranches backed by nongranular pools
20%
25%
35%
50%
75%
100%
250%
425%
650%
Deduction

36 See Vladislav Peretyatkin and William
Perraudin, ‘‘Capital for Asset-Backed Securities,’’
Bank of England, February 2003.
37 See, for example, Michael Pykhtin and Ashish
Dev, ‘‘Credit Risk in Asset Securitizations:
Analytical Model,’’ Risk (May 2002) S16–S20.

E:\FR\FM\04AUP2.SGM

04AUP2

45936

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
TABLE 2.—ABS RISK WEIGHTS BASED ON SHORT-TERM EXTERNAL CREDIT ASSESSMENTS
Thick tranches backed by
highly granular pools

External rating (illustrative)
A–1/P–1 ...........................................................................
A–2/P–2 ...........................................................................
A–3/P–3 ...........................................................................
All other ratings ...............................................................

The Agencies seek comment on the
proposed treatment of securitization
exposures under the RBA. For rated
securitization exposures, is it appropriate to
differentiate risk weights based on tranche
thickness and pool granularity?
For non-retail securitizations, will
investors generally have sufficient
information to calculate the effective number
of underlying exposures (N).
What are views on the thresholds, based on
N and Q, for determining when the different
risk weights apply in the RBA?
Are there concerns regarding the reliability
of external ratings and their use in
determining regulatory capital? How might
the Agencies address any such potential
concerns?
Unlike the A–IRB framework for wholesale
exposures, there is no maturity adjustment
within the proposed RBA. Is this reasonable
in light of the criteria to assign external
ratings?

7%
N/A
N/A
N/A

.....................................
.....................................
.....................................
.....................................

Base case
12% ...................................
20% ...................................
75% ...................................
Deduction ..........................

The Supervisory Formula Approach
(SFA)
As noted above, when an explicit A–
IRB approach exists for the underlying
exposures, originating and investing
banking organizations would be able to
apply the SFA to non-rated exposures
above the KIRB threshold, subject to
supervisory approval and review. The
Agencies anticipate that, in addition to
its application to liquidity facilities and
to other traditional and synthetic
securitization exposures, the SFA would
be used when calculating A–IRB capital
requirements for tranched guarantees
(for example, a loan for which a
guarantor assumes a first-loss position
that is less than the full amount of the
loan).
Under the SFA, the A–IRB capital
charge for a securitization tranche
would depend on six institutionsupplied inputs: 38 the notional amount

h = (1 − KIRB / LGD)

Tranches backed by nongranular pools
20%
35%
75%
Deduction

of underlying exposures that have been
securitized (E), the A–IRB capital charge
had the underlying exposures not been
securitized (KIRB); the tranche’s credit
enhancement level (L); the tranche’s
thickness (T); the pool’s effective
number of exposures (N); and the pool’s
exposure-weighted average loss-givendefault (LGD). In general, the estimates
of N and LGD would be developed as a
by-product of the process used to
determine KIRB.
The SFA capital charge for a given
securitization tranche would be
calculated as the notional amount of
underlying exposures that have been
securitized (E), multiplied by the greater
of: (i) 0.0056 * T or (ii) the following
expression: 39
K[L + T]¥K[L] + {(0.05 * d * KIRB *
e¥20(L¥KIRB)/KIRB) *
(1¥e¥20T/KIRB)},
where,40

N

c = KIRB / (1 − h )
v=

(LGD − KIRB) KIRB

+ 0.25 (1 − LGD) KIRB
N

 v + KIRB2

(1 − KIRB) KIRB − v
− c2  +
f =
(1 − h) ∗1000
 1− h

g=

(1 − c) c

f
a = g∗c

− 1

b = g ∗ (1 − c)
d = 1 − (1 − h ) ∗ (1 − Beta [KIRB;
[KIRB a, b])

Although visually daunting, the above
supervisory formula is easily
programmable within standard
spreadsheet packages, and its various

components have intuitive
interpretations.
Part (i), noted above, of the SFA
effectively imposes a 56 basis point
minimum or floor A–IRB capital charge

per dollar of tranche exposure. While
acknowledging that such a floor is not
risk-sensitive, the Agencies believe that
some minimum prudential capital
charge is nevertheless appropriate. The

38 When the banking organization holds only a
proportional interest in the tranche, that position’s
A–IRB capital charge equals the prorated share of
the capital charge for the entire tranche.

39 The SFA applies only to exposures above KIRB.
When a securitization tranche straddles KIRB, for
the purpose of applying the SFA the tranche should
be decomposed into a position at or below KIRB
and another above KIRB. The latter would be the
position to which the SFA is actually applied.

40 In these expressions, Beta[X; a, b] refers to the
cumulative beta distribution with parameters a and
b evaluated at X. The cumulative beta distribution
function is available in Excel as the function
BETADIST.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

PO 00000

Frm 00038

Fmt 4701

Sfmt 4702

E:\FR\FM\04AUP2.SGM

04AUP2

EP04AU03.004</MATH>

K[x] = (1 − h ) ∗ (x ∗ (1 − Beta [x; a, b]) + c ∗ Beta[x; a + 1, b]).

45937

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

43 See Michael Pykhtin and Ashish Dev, ‘‘Coarsegranied CDOs,’’ Risk (January 2003) 113–116.

PO 00000

Frm 00039

Fmt 4701

Sfmt 4702

spread account or overcollateralization)
that provides credit enhancement to the
tranche of interest. Credit-enhancing
interest-only strips would not be
included in the calculation of L.
Thickness (T). This input would be
measured (in decimal form) as the ratio
of (a) the notional amount of the tranche
of interest to (b) the notional or loan
equivalent amount of underlying
exposures in the pool (E).
Effective number of exposures (N).
This input would be calculated as



 ∑ EAD i 


N= i
2
∑ EADi

2

i

where EADi represents the exposure-atdefault associated with the i-th
underlying exposure in the pool.
Multiple underlying exposures to the
same obligor would be consolidated
(that is, treated as a single exposure). If
the pool contains any underlying
exposures that are themselves
securitization exposures (for example,
one or more asset-backed securities),
each of these would be treated as a
single exposure for the purpose of
measuring N.44
Exposure-weighted average LGD. This
input would be calculated (in decimal
form) as

LGD =

∑ LGDi ⋅ EADi
i
∑ EADi
i

where LGDi represents the average LGD
associated with all underlying
exposures to the i-th obligor. In the case
of re-securitization (a securitization of
securitization exposures), an LGD of 100
percent would be assumed for any
underlying exposure that was itself a
securitization exposure.45
44 Within the supervisory formula, the probability
distribution of credit losses associated with the pool
of underlying exposures is approximated by treating
the pool as if it consisted of N homogeneous
exposures, each having an A–IRB capital charge of
KIRB/N. The proposed treatment of N implies, for
example, that a pool containing one ABS tranche
backed by 1 million effective loans behaves more
like a single loan having an A–IRB capital charge
of KIRB than a pool of 1 million loans, each having
an A–IRB capital charge of KIRB/1,000,000.
45 As noted above, the A–IRB securitization
framework does not permit banking organizations to
use their own internal estimates of LGDs (and PDs)
for securitization exposures because such
quantification requires implicit or explicit estimates
of loss correlations among the underlying
exposures. Recall that LGDs should be measured as
the loss rates expected to prevail when default rates
are high. While setting LGDs equal to 100 percent
is reasonable for certain types of ABSs, such as
highly subordinated or thin tranches, this level of

E:\FR\FM\04AUP2.SGM

Continued

04AUP2

EP04AU03.010</MATH>

41 See Vladislav Peretyatkin and William
Perraudin, ‘‘Capital for Asset-Backed Securities,’’
Bank of England, February 2003.
42 The conceptual basis for specification of K[x]
is developed in Michael B. Gordy and David Jones,
‘‘Random Tranches,’’ Risk (March 2003) 78–83.

addition, this specification embodies
the well-known result that a pool’s total
systematic risk (that is, KIRB) tends to
be redistributed toward more senior
tranches as the effective number of
underlying exposures in the pool (N)
declines.43 The importance of pool
granularity depends on the pool’s
average loss-rate-given-default, as
increases in LGD also tend to shift
systematic risk toward senior tranches
when N is small. For highly granular
pools, such as securitizations of retail
exposures, LGD would have no
influence on the SFA capital charge.
The Agencies propose to establish
criteria for determining E, KIRB, L, T, N,
and LGD that are consistent with those
suggested in the New Accord. A
summary of these requirements is
presented below.
E. This input would be measured (in
dollars) as the A–IRB estimate of the
exposures in the underlying pool of
securitized exposures, as if they were
held directly by the banking
organization, rather than securitized.
This amount would reflect only those
underlying exposures that have actually
been securitized to date. Thus, for
example, E would exclude undrawn
lines associated with revolving credit
facilities (for example, credit card
accounts).
KIRB. This input would be measured
(in decimal form) as the ratio of (a) the
pool’s A–IRB capital requirement to (b)
the notional or loan equivalent amount
of the underlying exposures in the pool
(E). The pool’s A–IRB capital
requirement would be calculated in
accordance with the applicable A–IRB
standard for the type of underlying
exposure. This calculation would
incorporate the effect of any credit risk
mitigant that is applied to the
underlying exposures (either
individually or to the entire pool), and
hence benefits all of the securitization
exposures. Consistent with the
measurement of E, the estimate of KIRB
would reflect only the underlying
exposures that have been securitized.
For example, KIRB generally would
exclude the A–IRB capital charges
against the undrawn portions of
revolving credit facilities.
Credit enhancement level (L). This
input would be measured (in decimal
form) as the ratio of (a) the notional
amount of all securitization exposures
subordinate to the tranche of interest to
(b) the notional or loan equivalent
amount of underlying exposures in the
pool (E). L would incorporate any
funded reserve account (for example,

EP04AU03.005</MATH>

floor has been proposed at 56 basis
points partly on the basis of empirical
analyses suggesting that, across a broad
range of modeling assumptions and
exposure types, this level provides a
reasonable lower bound on the capital
charges implied by standard credit risk
models for securitization tranches
meeting the standards for an external
rating of AAA.41 This floor also is
consistent with the lowest capital
charge available under the RBA.
Part (ii) of the SFA also is a blend of
credit risk modeling results and
supervisory judgment. The function
denoted K[x] represents a pure modelbased estimate of the pool’s aggregate
systematic or non-diversifiable credit
risk that is attributable to a first-loss
position covering pool losses up to and
including x. Because the tranche of
interest (defined in terms of a credit
enhancement level L, and thickness T)
covers losses between L and L+T, its
total systematic risk can be represented
as K[L + T]¥K[L], which are the first
two terms in (1). The term in braces
within (1) represents a supervisory addon to the pure model-based result. This
add-on is intended primarily to avoid
potential behavioral distortions
associated with what would otherwise
be a discontinuity in capital charges for
relatively thin mezzanine tranches lying
just below and just above KIRB: all
tranches at or below KIRB would be
deducted from capital, whereas a very
thin tranche just above KIRB would
incur a pure model-based capital charge
that could vary between zero and one,
depending upon the number of effective
underlying exposures in the pool (N).
The add-on would apply primarily to
positions just above KIRB, and its
quantitative effect would diminish
rapidly as the distance from KIRB
widens.
Most of the complexity of the
supervisory formula is a consequence of
attempting to make K[x] as consistent as
possible with the parameters and
assumptions of the A–IRB framework
that would apply to the underlying
exposures if held directly by a banking
organization.42 The specification of K[x]
assumes that KIRB is an accurate
measure of the pool’s total systematic
credit risk, and that a securitization
merely redistributes this systematic risk
among its various tranches. In this way,
K[x] embodies precisely the same asset
correlations as are assumed elsewhere
within the A–IRB framework. In

45938

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

Simplified method for computing N
and LGD. Under the conditions
provided below, banking organizations
would be able to employ simplified
methods for calculating N and the
exposure-weighted average LGD. When

the underlying exposures are retail
exposures, the SFA may be
implemented by setting h = 0 and v =
0, subject to supervisory approval and
review. When the share of the pool
associated with the largest exposure, C1,

is no more than 0.03 (or 3 percent of the
pool), the banking organization would
be able to set LGD = 0.50 and N equal
to:

−1



 C − Cl 
N =  ClCm +  m
 max{1 − mC l , 0} ,
1
m
−





LGD may be conservative for other types of ABSs.
However, the Agencies believe that the complexity
and burden assoicated with a more refined
treatment of LGDs would outweigh any
improvement in the overall risk sensitivity of A–

IRB capital charges for originators, owing to the
combined effects of (a) the dollar-for-dollar A–IRB
capital charge on positions at or below KIRB, and
(b) the maximum or cap on an originator’s total A–
IRB capital charge.

The Agencies seek comment on the
proposed SFA. How might it be simplified
without sacrificing significant risk
sensitivity? How useful are the alternative
simplified computation methodologies for N
and LGD

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

PO 00000

Frm 00040

Fmt 4701

Sfmt 4702

Agencies propose that the risk weight be
set equal to the risk weight applicable
under the general risk-based capital
rules for banking organizations not
using the A–IRB approach (that is, to the
underlying assets or obligors after
consideration of collateral or guarantees
or, if applicable, external ratings).
The Agencies seek comment on the
proposed treatment of eligible liquidity
facilities, including the qualifying criteria for
such facilities. Does the proposed LookThrough Approach—to be available as a
temporary measure—satisfactorily address
concerns that, in some cases, it may be
impractical for providers of liquidity
facilities to apply either the ‘‘bottom-up’’ or
‘‘top-down’’ approach for calculating KIRB?
It would be helpful to understand the degree
to which any potential obstacles are likely to
persist.
Feedback also is sought on whether
liquidity providers should be permitted to
calculate A–IRB capital charges based on
their internal risk ratings for such facilities in
combination with the appropriate RBA risk
weight. What are the advantages and
disadvantages of such an approach, and how
might the Agencies address concerns that the
supervisory validation of such internal
ratings would be difficult and burdensome?
Under such an approach, would the lack of
any maturity adjustment with the RBA be
problematic for assigning reasonable risk
weights to liquidity facilities backed by
relatively short-term receivables, such as
trade credit?

Other Considerations
Capital Treatment Absent an A–IRB
Approach—The Alternative RBA
For originating banking organizations
when there is not a specific A–IRB
treatment for an underlying exposure or
group of underlying exposures, the
Agencies propose that a securitization
exposure’s A–IRB capital charge be
based exclusively on the exposure’s
external or inferred credit rating using
46 The level of m is to be set by each banking
organization.
47 The Alternative RBA does not apply to eligible
liquidity facilities, which may use the Look-

E:\FR\FM\04AUP2.SGM

04AUP2

EP04AU03.006</MATH>

The Look-Through Approach for
Eligible Liquidity Facilities
ABCP conduits and similar programs
sponsored by U.S. banking
organizations are major sources of
funding for financial and non-financial
companies. Liquidity facilities
supporting these programs are
considered to be securitization
exposures of the banking organizations
providing the liquidity, and generally
would be treated under the rules
proposed for originators. As a general
matter, the Agencies expect that banking
organizations using the A–IRB approach
would apply the SFA when determining
the A–IRB capital requirement for
liquidity facilities provided to ABCP
conduits and similar programs.
However, if it would not be practical for
a banking organization to calculate KIRB
for the underlying exposures using a
top-down or a bottom-up approach, the
banking organization may be allowed to
use the Look-Through Approach,
described below, for determining the A–
IRB capital requirement, subject to
supervisory approval and only for a
temporary period of time to be
determined in consultation with the
organization’s primary Federal
supervisor.
Because the Look-Through Approach
has limited risk sensitivity, the Agencies

propose that its applicability be
restricted to liquidity facilities that are
structured to minimize the extent to
which the facilities provide credit
support to the conduit. The LookThrough Approach would only be
available to liquidity facilities that meet
the following criteria:
(a) The facility documentation clearly
identifies and limits the circumstances
under which it may be drawn. In
particular, the facility must not be able
to cover losses already sustained by the
pool of underlying exposures (for
example, to acquire assets from the pool
at above fair value) or be structured
such that draw-down is highly probable
(as indicated by regular or continuous
draws);
(b) The facility is subject to an asset
quality test that prevents it from being
drawn to cover underlying exposures
that are in default;
(c) The facility cannot be drawn after
all applicable (specific and programwide) credit enhancements from which
the liquidity facility would benefit have
been exhausted;
(d) Repayment of any draws on the
facility (that is, assets acquired under a
purchase agreement or loans made
under a lending agreement) may not
represent a subordinated obligation of
the pool or be subject to deferral or
waiver; and
(e) Reduction in the maximum drawn
amount, or early termination of the
facility, occurs if the quality of the pool
falls below investment grade.
Under the Look-Through Approach,
the liquidity facility’s A–IRB capital
charge would be computed as the
product of (a) 8 percent, (b) the
maximum potential drawdown under
the facility, (c) the applicable credit
conversion factor (CCF), and (d) the
applicable risk weight. The CCF would
be set at 50 percent if the liquidity
facility’s original maturity is one year or
less, and at 100 percent if the original
maturity is more than one year. The

provided that the banking organization
can measure Cm, which denotes the
share of the pool corresponding to the
largest ‘‘m’’ exposures (for example, a 15
percent share corresponds to a value of
0.15).46 Alternatively, when only C1 is
available and this amount is no more
than 0.03, then the banking organization
would be able to set LGD = 0.50 and N
= 1/ C1.

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
the Alternative RBA.47 Under the
Alternative RBA, a risk weight of 20
percent is applied to exposures rated
AAA to AA¥, 50 percent to exposures
rated A+ to A¥, and 100 percent to
exposures rated BBB+ to BBB-.
Securitization exposures having ratings
below investment grade, or that are nonrated, would be deducted from riskbased capital on a dollar-for-dollar
basis.
Should the A–IRB capital treatment for
securitization exposures that do not have a
specific A–IRB treatment be the same for
investors and originators? If so, which
treatment should be applied—that used for
investors (the RBA) or originators (the
Alternative RBA)? The rationale for the
response would be helpful.

Structures With Early Amortization
Provisions
Many securitizations of revolving
credit facilities (for example, credit card
accounts) contain provisions that call
for the securitization to be wound down
if the excess spread falls below a certain
threshold.48 This decrease in excess
spread can, in some cases, be caused by
deterioration in the credit quality of the
underlying exposures. An early
amortization event can increase a
banking organization’s capital needs if
any new draws on the revolving
facilities would need to be financed by
the banking organization itself using onbalance-sheet sources of funding. The
payment allocations used to distribute
principal and finance charge collections
during the amortization phase of these
structures also can expose a banking
organization to greater risk of loss than
in other securitization structures. To
account for the risks that early
amortization structures pose to
originating banking organizations, the
capital treatment described below
would apply to securitizations of
revolving credit facilities containing
such features.
In addition to the A–IRB capital
charge an originating banking
organization would incur on the
securitization exposures it retains, an
originator would be required to hold
capital against all or a portion of the
Through Approach as described above.
Additionally, the securitization exposures subject to
the Alternative RBA are not limited by the
maximum capital requirement discussed above.
48 Excess spread is defined as gross finance charge
collections and other income received by the trust
or special purpose entity (SPE) minus certificate
interest, servicing fees, charge-offs, and other senior
trust or SPE expenses.
29 A banking organization is generally considered
a sponsor of an ABCP conduit or similar program
if, in fact or in substance, it manages or advises the
conduit program, places securities into the market
for the program, or provides liquidity support or
credit enhancements to the program.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

investors’ interest in a securitization
when (i) the organization sells
exposures into a securitization that
contains an early amortization feature,
and (ii) the underlying exposures sold
are of a revolving nature. The A–IRB
capital charge attributed to the
originator that is associated with the
investors’ interest is calculated as the
product of (a) the A–IRB capital charge
that would be imposed on the entire
investors’ interest if it were held by the
originating banking organization, and
(b) an applicable CCF.
In general, the CCF would depend on
whether the early amortization feature
repays investors through a controlled or
non-controlled mechanism, and
whether the underlying exposures
represent uncommitted revolving retail
facilities that are unconditionally
cancellable without prior notice (for
example, credit card receivables) or
other credit lines (for example,
revolving corporate facilities).
An early amortization provision
would be considered controlled if,
throughout the duration of the
securitization transaction, including the
amortization period, there is a pro rata
sharing of interest, principal, expenses,
losses, and recoveries based on the
balances of receivables outstanding at
the beginning of each month. Further,
the pace of repayment may not be any
more rapid than would be allowed
through straight-line amortization over a
period sufficient for 90 percent of the
total debt outstanding at the beginning
of the early amortization period to have
been repaid or recognized as in default.
In addition to these criteria, banking
organizations with structures containing
controlled early amortization features
would also have to have appropriate
plans in place to ensure that there is
sufficient capital and liquidity available
in the event of an early amortization.
When these conditions are not met, the
early amortization provision would be
treated as non-controlled.
Determination of CCFs for Controlled
Early Amortization Structures
The following method for determining
CCFs applies to a securitization of
revolving credit facilities containing a
controlled early amortization
mechanism. When the pool of
underlying exposures includes
uncommitted retail credit lines (for
example, credit card receivables), an
originator would first compare the
securitization’s three-month average
excess spread against the following two
reference levels:
A. The point at which the banking
organization would be required to trap

PO 00000

Frm 00041

Fmt 4701

Sfmt 4702

45939

excess spread under the terms of the
securitization; and
B. The excess spread level at which
an early amortization would be
triggered.
In cases where a transaction does not
require excess spread to be trapped, the
first trapping point would be deemed to
be 4.5 percentage points greater than the
excess spread level at which an early
amortization is triggered.
The banking organization would
divide the distance between the two
points described above into four equal
segments. For example if the spread
trapping point is 4.5 percent and the
early amortization trigger is zero
percent, then 4.5 percent would be
divided into four equal segments of
112.5 basis points each. The following
conversion factors, based on illustrative
segments, would apply to the investors’
interest.

CONTROLLED EARLY AMORTIZATION OF
UNCOMMITTED RETAIL CREDIT LINES
3-month average excess
spread
450 basis points (bp) or more ..
Less than 450 bp to 337.5 bp ..
Less than 337.5 bp to 225 bp ..
Less than 225 bp to 112.5 bp ..
Less than 112.5 bp ...................

Credit Conversion Factor (CCF)
(percent)
0
1
2
20
40

All other securitizations of revolving
facilities (that is, those containing
underlying exposures that are
committed or non-retail) having
controlled early amortization features
would be subject to a CCF of 90 percent.
Determination of CCFs for NonControlled Early Amortization
Structures
The process for determining CCFs
when a securitization of revolving credit
facilities contains a non-controlled early
amortization mechanism would be the
same as that described above for
controlled early amortization structures,
except that different CCFs would apply
to the various excess spread segments.
For non-controlled structures, the
following conversion factors, based on
illustrative segments, would apply:

NON-CONTROLLED EARLY AMORTIZATION
OF
UNCOMMITTED RETAIL
CREDIT LINES
3-month average excess
spread
450 basis points (bp) or more ..
Less than 450 bp to 337.5 bp ..

E:\FR\FM\04AUP2.SGM

04AUP2

Credit Conversion Factor (CCF)
(percent)
0
5

45940

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

NON-CONTROLLED EARLY AMORTIZA- Overlapping Credit Enhancements or
TION
OF
UNCOMMITTED RETAIL Liquidity Facilities
CREDIT LINES—Continued
In some ABCP or similar programs, a
banking organization may provide
Credit Con- multiple facilities that may be drawn
3-month average excess
version Facunder varying circumstances. The
spread
tor (CCF)
Agencies do not intend that a banking
(percent)
organization incur duplicative capital
Less than 337.5 bp to 225 bp ..
10 requirements against these multiple
Less than 225 bp to 112.5 bp ..
50 exposures as long as, in the aggregate,
Less than 112.5 bp ...................
100
multiple advances are not permitted
against the same collateral. Rather, a
All other securitizations of revolving
banking organization would be required
credit facilities (that is, those containing to hold capital only once for the
underlying exposures that are
exposure covered by the overlapping
committed or non-retail) having nonfacilities (whether they are general
liquidity facilities, eligible liquidity
controlled early amortization
facilities, or the facilities serve as credit
mechanisms would be subject to a CCF
enhancements). Where the overlapping
of 100 percent. In other words, no risk
facilities are subject to different
transference would be recognized for
conversion factors, the banking
these structures; an originator’s A–IRB
organization would attribute the
capital charge would be the same as if
overlapping part to the facility with the
the underlying exposures had not been
highest conversion factor. However, if
securitized.
different banking organizations provide
The Agencies seek comment on the
overlapping facilities, each institution
proposed treatment of securitization of
would hold capital against the entire
revolving credit facilities containing early
maximum amount of its facility. That is,
amortization mechanisms. Does the proposal
there may be some duplication of
satisfactorily address the potential risks such capital charges for overlapping facilities
transactions pose to originators?
provided by multiple banking
Comments are invited on the interplay
organizations.
between the A–IRB capital charge for
securitization structures containing early
amortization features and that for undrawn
lines that have not been securitized. Are
there common elements that the Agencies
should consider? Specific examples would be
helpful.
Are proposed differences in CCFs for
controlled and non-controlled amortization
mechanisms appropriate? Are there other
factors that the Agencies should consider?

Market-Disruption Eligible Liquidity
Facilities
A banking organization would be able
to apply a 20 percent CCF to an eligible
liquidity facility that can be drawn only
in the event of a general market
disruption (that is, where a capital
market instrument cannot be issued at
any price), provided that any advance
under the facility represents a senior
secured claim on the assets in the pool.
A banking organization using this
treatment would recognize 20 percent of
the A–IRB capital charge required for
the facility through use of the SFA. If
the market disruption eligible liquidity
facility is externally rated, a banking
organization would be able to rely on
the external rating under the RBA for
determining the A–IRB capital
requirement provided the organization
assigns a 100 percent CCF rather than a
20 percent CCF to the facility.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

Servicer Cash Advances
Subject to supervisory approval,
servicer cash advances that are
recoverable would receive a zero
percent CCF. This treatment would
apply when servicers, as part of their
contracts, may advance cash to the pool
to ensure an uninterrupted flow of
payments to investors, provided the
servicer is entitled to full
reimbursement and this right is senior
to other claims on cash flows from the
pool of underlying exposures.
When providing servicer cash advances,
are banking organizations obligated to
advance funds up to a specified recoverable
amount? If so, does the practice differ by
asset type? Please provide a rationale for the
response given.

Credit Risk Mitigation
For securitization exposures covered
by collateral or guarantees, the credit
risk mitigation rules discussed earlier
would apply. For example, a banking
organization may reduce the A–IRB
capital charge when a credit risk
mitigant covers first losses or losses on
a proportional basis. For all other cases,
a banking organization would assume
that the credit risk mitigant covers the
most senior portion of the securitization
exposure (that is, that the most junior
portion of the securitization exposure is
uncovered).

PO 00000

Frm 00042

Fmt 4701

Sfmt 4702

V. AMA Framework for Operational
Risk
This section describes features of the
proposed AMA framework for
measuring the regulatory capital
requirement for operational risk. Under
this framework, a banking organization
meeting the AMA supervisory standards
would use its internal operational risk
measurement system to calculate its
regulatory capital requirement for
operational risk. The discussion below
provides background information on
operational risk and the conceptual
underpinnings of the AMA, followed by
a discussion of the AMA supervisory
standards.49
The Agencies’ general risk-based
capital rules do not currently include an
explicit capital charge for operational
risk, which is defined as the risk of loss
resulting from inadequate or failed
processes, people, and systems or from
external events. When developing the
general risk-based capital rules, the
Agencies recognized that institutions
were exposed to non-credit related risks,
including operational risk.
Consequently, the Agencies built a
‘‘buffer’’ into the general risk-based
capital rules to implicitly cover other
risks such as operational risk. With the
introduction of the A–IRB framework
for credit risk in this ANPR, which
results in a more risk-sensitive
treatment of credit risk, there is no
longer an implicit capital buffer for
other risks.
The Agencies recognize that
operational risk is a key risk in financial
institutions, and evidence indicates that
a number of factors are driving increases
in operational risk. These include the
recent experience of a number of highprofile, high-severity losses across the
banking industry highlighting
operational risk as a major source of
unexpected losses. Because the
regulatory capital buffer for operational
risk would be removed under the
proposal, the Agencies are now seeking
comment on a risk-sensitive capital
framework for the largest, most complex
institutions that would include an
explicit risk-based capital requirement
for operational risk. The Agencies
propose to require banking
organizations using the A–IRB approach
for credit risk also to use the AMA to
compute capital charges for operational
risk.
49 For a more detailed discussion of the concepts
set forth in this ANPR and definitions of relevant
terms, see the accompanying interagency
‘‘Supervisory Guidance on Operational Risk
Advanced Measurement Approaches for Regulatory
Capital’’ (supervisory guidance) published
elsewhere in today’s Federal Register.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
The Agencies are proposing the AMA to
address operational risk for regulatory capital
purposes. The Agencies are interested,
however, in possible alternatives. Are there
alternative concepts or approaches that might
be equally or more effective in addressing
operational risk? If so, please provide some
discussion on possible alternatives.

A. AMA Capital Calculation
The AMA capital requirement would
be based on the measure of operational
risk exposure generated by a banking
organization’s internal operational risk
measurement system. In calculating the
operational risk exposure, an AMAqualified institution would be expected
to estimate the aggregate operational
risk loss that it faces over a one-year
period at a soundness standard
consistent with a 99.9 percent
confidence level. The institution’s AMA
capital requirement for operational risk
would be the sum of EL and UL, unless
the institution can demonstrate that an
EL offset would meet the supervisory
standards for operational risk. The
institution would have to use a
combination of internal loss event data,
relevant external loss event data,
business environment and internal
control factors, and scenario analysis in
calculating its operational risk exposure.
The institution also would be allowed to
recognize the effect of risk dependency
(for example, correlation) and, to a
limited extent, the effect of insurance as
a risk mitigant.
As with the proposed A–IRB capital
requirement for credit risk, the
operational risk exposure would be
converted to an equivalent amount of
risk-weighted assets for the calculation
of an institution’s risk-based capital
ratios. An AMA-qualified institution
would multiply the operational risk
exposure generated by its analytical
framework by a factor of 12.5 to convert
the exposure to a risk-weighted assets
equivalent. The resulting figure would
be added to the comparable figures for
credit and market risk in calculating the
institution’s risk-based capital
denominator.
Does the broad structure that the Agencies
have outlined incorporate all the key
elements that should be factored into the
operational risk framework for regulatory
capital? If not, what other issues should be
addressed? Are any elements included not
directly relevant for operational risk
measurement or management? The Agencies
have not included indirect losses (for
example, opportunity costs) in the definition
of operational risk against which institutions
would have to hold capital; because such
losses can be substantial, should they be
included in the definition of operational risk?

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

Overview of the Supervisory Criteria
Use of the AMA would be subject to
supervisory approval. A banking
organization would have to demonstrate
that it has satisfied all supervisory
standards before it would be able to use
the AMA for risk-based capital
purposes. The supervisory standards are
briefly described below. Because an
institution would have significant
flexibility to develop its own
methodology for calculating its riskbased capital requirement for
operational risk, it would be necessary
for supervisors to ensure that the
institution’s methodology is
fundamentally sound. In addition,
because different institutions may adopt
different methodologies for assessing
operational risk, the requirement to
satisfy supervisory standards offers
some assurance to institutions and their
supervisors that all AMA-qualified
institutions would be subject to a
common set of standards.
While the supervisory standards are
rigorous, institutions would have
substantial flexibility in terms of how
they satisfy the standards in practice.
This flexibility is intended to encourage
an institution to adopt a system that is
responsive to its unique risk profile,
foster improved risk management, and
allow for future innovation. The
Agencies recognize that operational risk
measurement is evolving rapidly and
wish to encourage continued evolution
and innovation. Nevertheless, the
Agencies also acknowledge that this
flexibility would make cross-institution
comparisons more difficult than if a
single supervisory approach were to be
mandated for all institutions. The
supervisory standards outlined below
are intended to allow flexibility while
also being sufficiently objective to
ensure consistent supervisory
assessment and enforcement of
standards across institutions.
The Agencies seek comment on the extent
to which an appropriate balance has been
struck between flexibility and comparability
for the operational risk requirement. If this
balance is not appropriate, what are the
specific areas of imbalance and what is the
potential impact of the identified imbalance?
The Agencies are considering additional
measures to facilitate consistency in both the
supervisory assessment of AMA frameworks
and the enforcement of AMA standards
across institutions. Specifically, the Agencies
are considering enhancements to existing
interagency operational and managerial
standards to directly address operational risk
and to articulate supervisory expectations for
AMA frameworks. The Agencies seek
comment on the need for and effectiveness of
these additional measures.
The Agencies also seek comment on the
supervisory standards. Do the standards

PO 00000

Frm 00043

Fmt 4701

Sfmt 4702

45941

cover the key elements of an operational risk
framework?

An institution’s operational risk
framework would have to include an
independent operational risk
management function, line of business
oversight, and independent testing and
verification. Both the institution’s board
of directors and management would
have to have responsibilities in
establishing and overseeing this
framework. The institution would have
to have clear policies and procedures in
place for identifying, measuring,
monitoring, and controlling operational
risk.
An institution would have to establish
an analytical framework that
incorporates internal operational loss
event data, relevant external loss event
data, assessments of the business
environment and internal control
factors, and scenario analysis. The
institution would have to have
standards in place to capture all of these
elements. The combination of these
elements would determine the
institution’s quantification of
operational risk and related regulatory
capital requirement.
The supervisory standards for the
AMA have both quantitative and
qualitative elements. Effective
operational risk quantification is critical
to the objective of a risk-sensitive
capital requirement. Consequently, a
number of the supervisory standards are
aimed at ensuring the integrity of the
process by which an institution arrives
at its estimated operational risk
exposure.
It is not sufficient, however, to focus
solely on operational risk measurement.
If the Agencies are to rely on
institutions to determine their riskbased capital requirements for
operational risk, there would have to be
assurances that institutions have in
place sound operational risk
management infrastructures. In
addition, risk management elements
would be critical inputs into the
quantification of operational risk
exposure, that is, operational risk
quantification would have to take into
account such risk management elements
as the quality of an institution’s internal
controls. Likewise, the AMA capital
requirement derived from an
institution’s quantification methodology
would need to offer incentives for an
institution to improve its operational
risk management practices. Ultimately,
the Agencies believe that better
operational risk management will
enhance operational risk measurement,
and vice versa.

E:\FR\FM\04AUP2.SGM

04AUP2

45942

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

Corporate Governance
An institution’s operational risk
framework would have to include an
independent firm-wide operational risk
management function, line of business
management oversight, and
independent testing and verification
functions. While no specific
management structure would be
mandated, all three components would
have to be evident.
The institution’s board of directors
would have to oversee the development
of the firm-wide operational risk
framework, as well as major changes to
the framework. Management roles and
accountability would have to be clearly
established. The board and management
would have to ensure that appropriate
resources have been allocated to support
the operational risk framework.
The independent firm-wide
operational risk management function
would be responsible for overseeing the
operational risk framework at the firm
level to ensure the development and
consistent application of operational
risk policies, processes, and procedures
throughout the institution. This
function would have to be independent
from line of business management and
the testing and verification functions.
The firm-wide operational risk
management function would have to
ensure appropriate reporting of
operational risk exposures and loss data
to the board and management.
Lines of business would be
responsible for the day-to-day
management of operational risk within
each business unit. Line of business
management would have to ensure that
internal controls and practices within
their lines of business are consistent
with firm-wide policies and procedures
that support the management and
measurement of the institution’s
operational risk.
The Agencies are introducing the concept
of an operational risk management function,
while emphasizing the importance of the
roles played by the board, management, lines
of business, and audit. Are the
responsibilities delineated for each of these
functions sufficiently clear and would they
result in a satisfactory process for managing
the operational risk framework?

Operational Risk Management Elements
An institution would have to have
policies and procedures that clearly
describe the major elements of its
operational risk framework, including
identifying, measuring, monitoring, and
controlling operational risk.
Management reports would need to be
developed to address both firm-wide
and line of business results. These
reports would summarize operational

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

risk exposure, operational loss
experience, and relevant assessments of
business environment and internal
control factors, and would have to be
produced at least quarterly. Operational
risk reports, which summarize relevant
firm-wide operational risk information,
would also have to be provided
periodically to senior management and
the board. An institution’s internal
control system and practice would have
to be adequate in view of the complexity
and scope of its operations. In addition,
an institution would be expected to
meet or exceed minimum supervisory
standards as set forth in the Agencies’
supervisory policy statements and other
guidance.
B. Elements of an AMA Framework
An institution would have to
demonstrate that it has adequate
internal loss event data, relevant
external loss event data, assessments of
business environments and internal
control factors, and scenario analysis to
support its operational risk management
and quantification framework. These
inputs would need to be consistent with
the regulatory definition of operational
risk. The institution would have to have
clear standards for the collection and
modification of operational risk inputs.
There are a number of standards that
banking organizations would have to
meet with respect to internal
operational loss data. Institutions would
have to have at least five years of
internal operational risk loss data
captured across all material business
lines, events, product types, and
geographic locations.50 An institution
would have to establish thresholds
above which all internal operational
losses would be captured. The New
Accord introduces seven loss event type
classifications; the Agencies are not
proposing that an institution would be
required to internally manage its
operational risk according to these
specific loss event type classifications,
but nevertheless it would have to be
able to map its internal loss data to
these loss event categories. The
institution would have to provide
consistent treatment for the timing of
reporting an operational loss in its
internal data systems. As highlighted
earlier in this ANPR, credit losses
caused or exacerbated by operational
risk events would be treated as credit
losses for regulatory capital purposes;
these would include fraud-related credit
losses.
50 With supervisory approval, a shorter initial
observation period may be acceptable for
institutions that are newly authorized to use an
AMA methodology.

PO 00000

Frm 00044

Fmt 4701

Sfmt 4702

An institution would have to establish
and adhere to policies and procedures
that provide for the use of relevant
external loss data in the operational risk
framework. External data would be
particularly relevant where an
institution’s internal loss history is not
sufficient to generate an estimate of
major unexpected losses. Management
would have to systematically review
external data to ensure an
understanding of industry experience.
The Agencies seek comment on the use
of external data and its optimal function
in the operational risk framework.
While internal and external data
provide an important historic picture of
an institution’s operational risk profile,
it is important that institutions take a
forward-looking view as well.
Consequently, an institution would
have to incorporate assessments of the
business environment and internal
control factors (for example, audit
scores, risk and control assessments,
risk indicators, etc.) into its AMA
capital assessment. In addition, an
institution would have to periodically
compare its assessment of these factors
with actual operational loss experience.
Another element of the AMA
framework is scenario analysis. Scenario
analysis is a systematic process of
obtaining expert opinions from business
managers and risk management experts
to derive reasoned assessments of the
likelihood and impact of plausible
operational losses consistent with the
regulatory soundness standard. While
scenario analysis may rely, to a large
extent, on internal or, especially,
external data (for example, where an
institution looks to industry experience
to generate plausible loss scenarios), it
is particularly useful where internal and
external data do not generate a sufficient
assessment of the institution’s
operational risk profile.
An institution would be required to
have a comprehensive analytical
framework that provides an estimate of
the aggregate operational loss that it
faces over a one-year period at a
soundness standard consistent with a
99.9 percent confidence level. The
institution would have to document the
rationale for all assumptions
underpinning its chosen analytical
framework, including the choice of
inputs, distributional assumptions, and
weighting of quantitative and qualitative
elements. The institution would also
have to document and justify any
subsequent changes to these
assumptions.
An institution’s operational risk
analytical framework would have to use
a combination of internal operational
loss event data, relevant external

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
operational loss event data, business
environment and control factors, as well
as scenario analysis. The institution
would have to combine these elements
in the manner that most effectively
enables it to quantify its operational risk
exposure. The institution would have to
develop an analytical framework that is
appropriate to its business model and
risk profile.
Regulatory capital for operational risk
would be based on the sum of EL and
UL. There may be instances where an
EL offset could be recognized, but the
Agencies believe that this is likely to be
difficult given existing supervisory and
accounting standards. The Agencies
have considered both reserving and
budgeting as potential mechanisms for
EL offsets. The use of reserves may be
hampered by accounting standards,
while budgeting raises concerns about
availability over a one-year time horizon
to act as a capital replacement
mechanism. The Agencies are interested
in specific examples of how business
practices might be used to offset EL in
the operational risk framework.
An institution would have to
document how its chosen analytical
framework accounts for dependence (for
example, correlation) among operational
losses across and within business lines.
The institution would have to
demonstrate that its explicit and
embedded dependence assumptions are
appropriate, and where dependence
assumptions are uncertain, the
institution would have to use
conservative estimates.
An institution would be able to
reduce its operational risk exposure by
no more than 20 percent to reflect the
impact of risk mitigants such as
insurance. Institutions would have to
demonstrate that qualifying risk
mitigants meet a series of criteria
(described in the supervisory guidance)
to assess whether the risk mitigants are
sufficiently capital-like to warrant a
reduction of the operational risk
exposure.
The Agencies seek comment on the
reasonableness of the criteria for recognition
of risk mitigants in reducing an institution’s
operational risk exposure. In particular, do
the criteria allow for recognition of common
insurance policies? If not, what criteria are
most binding against current insurance
products? Other than insurance, are there
additional risk mitigation products that
should be considered for operational risk?

An institution using an AMA for
regulatory capital purposes would have
to use advanced data management
practices to produce credible and
reliable operational risk estimates.
These practices are comparable to the
data maintenance requirements set forth

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

under the A–IRB approach for credit
risk.
The institution would have to test and
verify the accuracy and appropriateness
of the operational risk framework and
results. Testing and verification would
have to be done independently of the
firm-wide risk management function
and the lines of business.

45943

Disclosure requirements would apply
to the bank holding company
representing the top consolidated level
of the banking group. Individual banks
within the holding company or
consolidated group would not generally
be required to fulfill the disclosure
requirements set out below. An
exception to the general rule would be
that individual banks and thrifts within
a group would still be required to
disclose Tier 1 and total capital ratios
and their components (that is, Tier 1,
Tier 2, and Tier 3 capital), as is the case
today. In addition, all banks and thrifts
would continue to be required to submit
appropriate information to regulatory
authorities (for example, Report of
Condition of Income (Call Reports) or
Thrift Financial Reports).51

The Agencies are proposing a set of
disclosure requirements that would
allow market participants to assess key
pieces of information regarding a
banking group’s capital structure, risk
exposures, risk assessment processes,
and ultimately, the capital adequacy of
the institution. Failure to meet these
minimum disclosure requirements, if
not corrected, would render a banking
organization ineligible to use the
advanced approaches or would
otherwise cause the banking
organization to forgo potential capital
benefits arising from the advanced
approaches. In addition, other
supervisory measures may be taken if
appropriate.
Management would have some
discretion to determine the appropriate
medium and location of the required
disclosure. Disclosures made in public
financial reports (for example, in
financial statements or Management’s
Discussion and Analysis included in
periodic reports or SEC filings) or other
regulatory reports (for example, FR Y–
9C Reports), could fulfill the applicable
disclosure requirements and would not
need to be repeated elsewhere. For those
disclosures that are not made under
accounting or other requirements, the
Agencies are seeking comment on the
appropriate means of providing this
data to market participants. Institutions
would be encouraged to provide all
related information in one location; at a
minimum, institutions would be
required to provide a cross reference to
the location of the required disclosures.
The Agencies intend to maximize a
banking organization’s flexibility
regarding where to make the required
disclosures while ensuring that the
information is readily available to
market participants without
unnecessary burden. To balance these
contrasting objectives, the Agencies are
considering requiring banking
organizations to provide a summary
table on their public websites that
indicate where all disclosures may be
found. Such an approach also would
allow institutions to cross-reference
other web addresses (for example, those
containing public financial reports or
regulatory reports or other risk-oriented
disclosures) where certain of the
disclosures are located.
Given longstanding requirements for
robust quarterly disclosure in the
United States, and recognizing the
potential for rapid change in risk
profiles, the Agencies intend to require
that the disclosures be made on a

51 In order to meet supervisory responsibilities,
the Agencies plan to collect more detailed
information through the supervisory process or

regulatory reports. Much of this information may be
proprietary and accordingly would not be made
public.

VI. Disclosure
Market discipline is a key component
of the New Accord. The disclosure
requirements summarized below seek to
enhance the public disclosure practices,
and thereby the transparency, of
advanced approach organizations.
Commenters are encouraged to consult
the New Accord for specifics on the
disclosure requirements under
consideration. The Agencies view
enhanced market discipline as an
important complement to the advanced
approaches to calculating minimum
regulatory capital requirements, which
would be heavily based on internal
methodologies. Increased disclosures,
especially regarding a banking
organization’s use of the A–IRB
approach for credit risk and the AMA
for operational risk, would allow a
banking organization’s private sector
investors to more fully evaluate the
institution’s financial condition, risk
profile, and capital adequacy. Given
better information, private shareholders
and debt holders can better influence
the funding and capital costs of a
banking organization. Such actions
would enhance market discipline and
supplement supervisory oversight of the
organization’s risk-taking and
management.
A. Overview

PO 00000

Frm 00045

Fmt 4701

Sfmt 4702

E:\FR\FM\04AUP2.SGM

04AUP2

45944

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

quarterly basis. However, qualitative
disclosures that provide a general
summary of a banking organization’s
risk management objectives and
policies, reporting system, and
definitions would be able to be
published on an annual basis, provided
any significant changes to these are
disclosed in the interim. When
significant events occur, banking
organizations would be required to
publish material information as soon as
practicable rather than at the end of the
quarter.
The risks to which banking
organizations are exposed and the
techniques that they use to identify,
measure, monitor, and control those
risks are important factors that market
participants consider in their
assessment of an institution.
Accordingly, banking organizations
would be required to have a formal
disclosure policy approved by the board
of directors that addresses the
institution’s approach for determining
the disclosures it will make. The policy
also would have to address the
associated internal controls and
disclosure controls and procedures. The
board of directors and senior
management would have to ensure that
appropriate verification of the
disclosures takes place and that
effective internal controls and
disclosure controls and procedures are
maintained.
Consistent with sections 302 and 404
of the Sarbanes-Oxley Act of 2002,
management would have to certify to
the effectiveness of internal controls
over financial reporting and disclosure
controls and procedures, and the
banking organization’s external auditor
would have to attest to management’s
assertions with respect to internal
controls over financial reporting. The
scope of these reports would need to
include all information included in
regulatory reports and the disclosures
outlined in this ANPR. Section 36 of the
Federal Deposit Insurance Act has
similar requirements. Accordingly,
banking organizations would have to
implement a process for assessing the
appropriateness of their disclosures,
including validation and frequency.
Unless otherwise required by
accounting or auditing standards, or by
other regulatory authorities, the
proposed requirements do not mandate
that the new disclosures be audited by
an external auditor for purposes of
opining on whether the financial
statements are presented in accordance
with GAAP.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

B. Disclosure Requirements
Banking organizations would be
required to provide disclosures related
to scope of application, capital
structure, capital adequacy, credit risk,
equities in the banking book, credit risk
mitigation, asset securitization, market
risk, operational risk and interest rate
risk in the banking book. The disclosure
requirements are summarized below.
The required disclosures pertaining to
the scope of application of the advanced
approaches would include a description
of the entities found in the consolidated
banking group. Additionally, banking
organizations would be required to
disclose the methods used to
consolidate them, any major
impediments on the transfer of funds or
regulatory capital within the banking
group, and specific disclosures related
to insurance subsidiaries.
Capital structure disclosures would
provide summary information on the
terms and conditions of the main
features of capital instruments issued by
the banking organization, especially in
the case of innovative, complex, or
hybrid capital instruments. Quantitative
disclosures include the amount of Tier
1, Tier 2, and Tier 3 capital, deductions
from capital, and total eligible capital.
Capital adequacy disclosures would
include a summary discussion of the
banking organization’s approach to
assessing the adequacy of its capital to
support current and future activities.
These requirements also include a
breakdown of the capital requirements
for credit, equity, market, and
operational risks. Banking organizations
also would be required to disclose their
Tier 1 and total capital ratios for the
consolidated group, as well as those of
significant bank or thrift subsidiaries.
For each separate risk area, a banking
organization would describe its risk
management objectives and policies.
Such disclosures would include an
explanation of the banking
organization’s strategies and processes;
the structure and organization of the
relevant risk management function; the
scope and nature of risk reporting and/
or measurement systems; and the
policies for hedging and/or mitigating
risk and strategies and processes for
monitoring the continuing effectiveness
of hedges/mitigants.
The credit risk disclosure regime is
intended to enable market participants
to assess the credit risk exposure of A–
IRB banking organizations and the
overall applicability of the A–IRB
framework, without revealing
proprietary information or duplicating
the role of the supervisor in validating

PO 00000

Frm 00046

Fmt 4701

Sfmt 4702

the framework the banking organization
has put into place.
Credit risk disclosures would include
breakdowns of the banking
organization’s exposures by type of
credit exposure, geographic distribution,
industry or counterparty type
distribution, residual contractual
maturity, amount and type of impaired
and past due exposures, and
reconciliation of changes in the
allowances for exposure impairment.
Banking organizations would provide
disclosures discussing the status of the
regulatory acceptance process for the
adoption of the A–IRB approach,
including supervisory approval of such
transition. The disclosures would
provide an explanation and review of
the structure of internal rating systems
and relation between internal and
external ratings; the use of internal
estimates other than for A–IRB capital
purposes; the process for managing and
recognizing credit risk mitigation; and,
the control mechanisms for the rating
system including discussion of
independence, accountability, and
rating systems review. Required
qualitative disclosures would include a
description of the internal ratings
process and separate disclosures
pertaining to the banking organization’s
wholesale, retail and equity exposures.
There would be two categories of
quantitative disclosures for credit risk:
those that focus on the analysis of risk
and those that focus on the actual
results. Risk assessment disclosures
would include the percentage of total
credit exposures to which A–IRB
disclosures relate. Also, for each
portfolio except retail, the disclosures
would have to provide (1) a presentation
of exposures across a sufficient number
of PD grades (including default) to allow
for a meaningful differentiation of credit
risk,52 and (2) the default weightedaverage LGD for each PD, and the
amount of undrawn commitments and
weighted average EAD.53 For retail
portfolios, banking organization would
provide either 54 (a) disclosures outlined
52 Where banking organizations are aggregating
PD grades for the purposes of disclosure, this would
be a representative breakdown of the distribution of
PD grades used in the A–IRB approach.
53 Banking organizations need only provide one
estimate of EAD for each portfolio. However, where
banking organizations believe it is helpful, in order
to give a more meaningful assessment of risk, they
may also disclose EAD estimates across a number
of EAD categories, against the undrawn exposures
to which these relate.
54 Banking organizations would normally be
expected to follow the disclosures provided for the
non-retail portfolios. However, banking
organizations would be able to adopt EL grades at
the basis of disclosure where they believe this can
provide the reader with a meaningful differentiation
of credit risk. Where banking organizations are

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
above on a pool basis (that is, the same
as for non-retail portfolios), or (b)
analysis of exposures on a pool basis
against a sufficient number of EL grades
to allow for a meaningful differentiation
of credit risk.
Quantitative disclosures pertaining to
historical results would include actual
losses (for example, charge-offs and
specific provisions) in the preceding
period for each portfolio and how this
differs from past experience and a
discussion of the factors that affected
the loss experience in the preceding
period. In addition, disclosures would
include banking organizations’
estimates against actual outcomes over a
longer period.55 At a minimum, this
would include information on estimates
of losses against actual losses in each
portfolio over a period sufficient to
allow for a meaningful assessment of the
performance of the internal rating
processes. Banking organizations would
further be expected to decompose this to
provide analysis of PD, LGD and EAD
estimates against estimates provided in
the quantitative risk assessment
disclosures above.56
Disclosures for banking book equity
positions would include both balance
sheet and fair values, and the types and
nature of investments. The total
cumulative realized gains or losses
arising from sales and liquidations
would be disclosed, together with total
unrealized gains/losses and any
amounts included in Tier 1 and/or Tier
2 capital. Details on the equity capital
requirements would also be disclosed.
Disclosures relating to credit risk
mitigation would include a description
of the policies and processes for netting
and collateral valuation and
management, and the types of collateral
accepted by the bank. Banking
organizations would also be expected to
aggregating internal grades (either PD/LGD or EL)
for the purposes of disclosure, this should be a
representative breakdown of the distribution of
those grades used in the IRB approach.
55 For banking organizations implementing the
A–IRB and AMA in 2007, the disclosures would be
required from year-end-2008; in the meantime,
early adoption would be encouraged. The phased
implementation is to allow banking organizations
sufficient time to build up a longer run of data that
will make these disclosures meaningful. For
banking organizations that may adopt the advanced
approaches at a later date, they would also be
subject to a one-year phase in period after which
the disclosures would be required.
56 Banking organizations would have to provide
this further decomposition where it would allow
users greater insight into the reliability of the
estimates provided in the quantitative disclosures:
risk assessment. In particular, banking
organizations should provide this information
where there are material differences between the
PD, LGD or EAD estimates given by banking
organizations compared in actual outcomes over the
long run. Banking organizations should also
provide explanations for such differences.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

include information about the main
types of guarantor or credit derivative
counterparties, and any risk
concentrations arising from the use of a
mitigation technique.
Securitization disclosures would
summarize a banking organization’s
accounting policies for securitization
activities and the current year’s
securitization activity. Further, banking
organizations would be expected to
disclose the names of the external credit
rating providers used for securitizations.
They would also provide details of the
outstanding exposures securitized by
the banking organization and subject to
the securitization framework, including
impairments and losses, exposures
retained or purchased broken down into
risk weight bands, and aggregate
outstanding amounts of securitized
revolving exposures.
Disclosures for market risk would
include a description of the models,
stress testing, and backtesting used in
assessing market risk, as well as
information on the scope of supervisory
acceptance. Quantitative disclosures
would include the aggregate VaR, the
high, mean, and low VaR values over
the reporting period, and a comparison
of VaR estimates with actual outcomes.
A key disclosure under the
operational risk framework would be a
description of the AMA the banking
organization uses, including a
discussion of relevant internal and
external factors considered in the
banking organization’s measurement
approach. In addition, the banking
organization would disclose the
operational risk charge before and after
any reduction in capital resulting from
the use of insurance or other potential
risk mitigants.
Finally, disclosures relating to interest
rate risk in the banking book would
include the nature of that risk, key
assumptions made, and the frequency of
risk measurement. They would also
include the increase or decline in
earnings or economic value for upward
and downward rate shocks according to
management’s method for measuring
interest rate risk in the banking book.
The Agencies seek comment on the
feasibility of such an approach to the
disclosure of pertinent information and also
whether commenters have any other
suggestions regarding how best to present the
required disclosures.
Comments are requested on whether the
Agencies’ description of the required formal
disclosure policy is adequate, or whether
additional guidance would be useful.
Comments are requested regarding whether
any of the information sought by the
Agencies to be disclosed raises any particular
concerns regarding the disclosure of
proprietary or confidential information. If a

PO 00000

Frm 00047

Fmt 4701

Sfmt 4702

45945

commenter believes certain of the required
information would be proprietary or
confidential, the Agencies seek comment on
why that is so and alternatives that would
meet the objectives of the required
disclosure.
The Agencies also seek comment regarding
the most efficient means for institutions to
meet the disclosure requirements.
Specifically, the Agencies are interested in
comments about the feasibility of requiring
institutions to provide all requested
information in one location and also whether
commenters have other suggestions on how
to ensure that the requested information is
readily available to market participants.

VII. Regulatory Analysis
Federal agencies are required to
consider the costs, benefits, or other
effects of their regulations for various
purposes described by statute or
executive order. In particular, an
executive order and several statutes may
require the preparation of detailed
analyses of the costs, benefits, or other
effects of rules, depending on threshold
determinations as to whether the
rulemaking in question triggers the
substantive requirements of the
applicable statute or executive order.
For the reasons described above, the
proposed and final rules that the
Agencies may issue to implement the
New Accord would represent a
significant change to their current
approach to the measurement of
regulatory capital ratios, and the
supervision of institutions’ internal risk
management processes with respect to
capital allocations. First, in this ANPR,
core and opt-in banks would rely on
their own analyses to derive some of the
principal inputs that would determine
their regulatory capital requirements.
Core and opt-in banks would incur new
costs to create and refine their internal
systems and to attract and train the staff
expertise necessary to develop, oversee,
manage and test those systems. Second,
the measured regulatory capital ratios
(although not the minimums) would
likely change, perhaps substantially for
core and opt-in banks. Third, the
Agencies’ approach to supervising
capital adequacy would become
bifurcated; that is, general banks would
continue to use the general risk-based
capital rules, either in their current form
or as modified. As a result, there may be
significant differences in the regulatory
capital assigned to a particular type of
asset depending on whether the bank is
a core, opt-in, or general bank. To the
extent that an institution’s product mix
would be directly affected by a change
in the landscape of regulatory capital
requirements, this might also affect the
customers of those institutions due to

E:\FR\FM\04AUP2.SGM

04AUP2

45946

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

the changes in pricing and market
strategies.
The economic impact that would be
created by these possibly unforeseen
competitive effects is difficult to
estimate, and the Agencies encourage
comment. In particular, the Agencies are
interested in comments on the
competitive impact that a change in the
regulatory capital regime applied to
large institutions would have relative to
the competitive position of smaller
institutions that remain subject to the
general risk-based capital rules.
Conversely, if the regulatory burden of
the more prescriptive A–IRB approach
applied to core institutions were so
large as to offset the potential for a
lower measured capital requirement for
certain exposures, then the competitive
position of large institutions, with
respect to both their domestic and
international competitors, might be
worsened. The Agencies are also
interested in comments that address the
competitive position of regulated
institutions in the United States with
respect to financial service providers,
both domestic and foreign, that are not
subject to the same degree of regulatory
oversight.
None of the Agencies has yet made
the threshold determinations required
by executive order or statute with
respect to this ANPR. Because the
proposed approaches to assessing
capital adequacy described in this
ANPR are new, the Agencies currently
lack information that is sufficiently
specific or complete to permit those
determinations to be made or to prepare
any economic analysis that may
ultimately be required. Therefore, this
section of the ANPR describes the
relevant executive order and statutes,
and asks for comment and information
that will assist in the determination of
whether such analyses would be
necessary before the Agencies published
proposed or final rules.
Quantitative information would be
the most useful to the Agencies.
However, commenters may also provide
estimates of costs, benefits, or other
effects, or any other information they
believe would be useful to the Agencies
in making the determinations. In
addition, commenters are asked to
identify or estimate start-up, or nonrecurring, costs separately from costs or
effects they believe would be ongoing.
A. Executive Order 12866
Executive Order 12866 requires
preparation of an economic analysis for
agency actions that are ‘‘significant
regulatory actions.’’ ‘‘Significant
regulatory actions’’ include, among
other things, regulations that ‘‘have an

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

annual effect on the economy of $100
million or more or adversely affect in a
material way the economy, a sector of
the economy, productivity, competition,
jobs, the environment, public health or
safety, or state, local, or tribal
governments or communities. * * *’’ 57
Regulatory actions that satisfy one or
more of these criteria are called
‘‘economically significant regulatory
actions.’’ E.O. 12866 applies to the OCC
and the OTS, but not the Board or the
FDIC. If the OCC or the OTS determines
that the rules implementing the New
Accord comprise an ‘‘economically
significant regulatory action,’’ then the
agency making that determination
would be required to prepare and
submit to the Office of Management and
Budget’s (OMB) Office of Information
and Regulatory Affairs (OIRA) an
economic analysis that includes:
• A description of the need for the
rules and an explanation of how they
will meet the need;
• An assessment of the benefits
anticipated from the rules (for example,
the promotion of the efficient
functioning of the economy and private
markets) together with, to the extent
feasible, a quantification of those
benefits;
• An assessment of the costs
anticipated from the rules (for example,
the direct cost both to the government
in administering the regulation and to
businesses and others in complying
with the regulation, and any adverse
effects on the efficient functioning of the
economy, private markets (including
productivity, employment, and
competitiveness)), together with, to the
extent feasible, a quantification of those
costs; and
• An assessment of the costs and
benefits of potentially effective and
reasonably feasible alternatives to the
planned regulation (including
improving the current regulation and
reasonably viable nonregulatory
actions), and an explanation why the
planned regulatory action is preferable
to the identified potential alternatives.58
57 Executive Order 12866 (Sept. 30, 1993), 58 FR
51735 (Oct. 4, 1993), as amended by Executive
Order 13258, 67 FR 9385 (referred to hereafter as
E.O. 12866). For the complete text of the definition
of ‘‘significant regulatory action,’’ see E.O. 12866 at
§ 3(f). A ‘‘regulatory action’’ is ‘‘any substantive
action by an agency (normally published in the
Federal Register) that promulgates or is expected to
lead to the promulgation of a final rule or
regulation, including notices of inquiry, advance
notices of proposed rulemaking, and notices of
proposed rulemaking.’’ E.O. 12866 at § (e).
58 The components of the economic analysis are
set forth in E.O. 12866 § 6(a)(3)(C)(i)–(iii). For a
description of the methodology that OMB
recommends for preparing an economic analysis,
see Office of Management and Budget, ‘‘Economic
Analysis of Federal Regulations Under Executive

PO 00000

Frm 00048

Fmt 4701

Sfmt 4702

For purposes of determining whether
this rulemaking would constitute an
‘‘economically significant regulatory
action,’’ as defined by E.O. 12866, and
to assist any economic analysis that E.O.
12866 may require, the OCC and the
OTS encourage commenters to provide
information about:
• The direct and indirect costs, for
core banks and those banks who intend
to qualify as opt-in banks, of compliance
with the approach described in this
ANPR and the related supervisory
guidance;
• The costs, for general banks, of
adopting the approach;
• The effects on regulatory capital
requirements for core, opt-in, and
general banks;
• The effects on competitiveness, in
both domestic and international
markets, for core, opt-in, and general
banks. This would include the possible
effects on the customers served by these
U.S. institutions through changes in the
mix of product offerings and prices;
• The economic benefits of the
approach for core, opt-in, or general
banks, as measured by lower regulatory
capital ratios, and a potentially more
efficient allocation of capital. This
might also include estimates of savings
associated with regulatory capital
arbitrage transactions that are currently
undertaken in order to optimize return
on capital under the current capital
regime. That is, what estimates might
exist to quantify the improvements in
market efficiency from no longer
pursuing regulatory capital arbitrage
transactions?
• The features of the A–IRB approach
that provide an incentive for a bank to
seek to qualify to use it, that is, to
become an opt-in bank.
The OCC and the OTS also encourage
comment on any alternatives to the
regulatory approaches described in the
ANPR that the Agencies should
consider.
B. Regulatory Flexibility Act
The Regulatory Flexibility Act (RFA)
generally requires agencies to prepare a
‘‘regulatory flexibility analysis’’ unless
the head of the agency certifies that a
regulation will not ‘‘have a significant
economic impact on a substantial
number of small entities.’’ 59 The RFA
applies to all of the Agencies.
The Agencies understand that the
RFA has been construed to require
Order 12866’’ (January 11, 1996). This publication
is available on OMB’s Web site at http://
www.whitehouse.gov/omb/inforeg/riaguide.html.
OMB recently published revisions to this
publication for comment. See 68 FR 5492 (February
3, 2003).
59 The RFA is codified at 5 U.S.C. 601 et seq.

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules
consideration only of the direct impact
on small entities.60 The Small Business
Administration (SBA) has said: ‘‘The
courts have held that the RFA requires
an agency to perform a regulatory
flexibility analysis of small entity
impacts only when a rule directly
regulates them,’’ that is, when it directly
applies to them.61 Since the proposed
approach would directly apply to only
a limited number of large banking
organizations, it would appear that the
Agencies may certify that the issuance
of this ANPR would not have significant
economic impact on a substantial
number of small entities.
Do the potential advantages of the A–
IRB approach, as measured by the
specific capital requirements on lowerrisk loans, create a competitive
inequality for small institutions, which
are effectively precluded from adopting
the A–IRB due to stringent qualification
standards? Conversely, would small
institutions that remain on the general
risk-based capital rules be at a
competitive advantage from specific
capital requirements on higher risk
assets vis-à-vis advanced approach
institutions? How might the Agencies
estimate the effect on credit availability
to small businesses or retail customers
of general banks?
C. Unfunded Mandates Reform Act of
1995
The Unfunded Mandates Reform Act
of 1995 (UMRA) requires preparation of
a written budgetary impact statement
before promulgation of any rule likely to
result in a ‘‘Federal mandate’’ that ‘‘may
result in the expenditure by State, local,
and tribal governments, in the aggregate,
or by the private sector, of $100,000,000
or more (adjusted annually for inflation)
in any 1 year.’’ 62 A ‘‘Federal mandate’’
includes any regulation ‘‘that would
impose an enforceable duty upon the
private sector. * * *’’ If a budgetary
impact statement is required, the UMRA
further requires the agency to identify
and consider a reasonable number of
60 With respect to banks, the Small Business
Administration (SBA) has defined a small entity to
be a bank with total assets of $150 million or less.
13 CFR § 121.201.
61 SBA Office of Advocacy, A Guide for
Government Agencies, ‘‘How to Comply with the
Regulatory Flexibility Act (May 2003), at 20
(emphasis added). See also Mid-Tex Electric
Cooperative, Inc. v. FERC. 773 F.2d 327, 340–43
(D.C. Cir. 1985) (‘‘[W]e conclude that an agency may
properly certify that no regulatory flexibility
analysis is necessary when it determines that the
rule will not have a significant economic impact on
a substantial number of small entities that are
subject to the requirements of the rule.’’) (emphasis
added) (construing language in the RFA that was
unchanged by subsequent statutory amendments).
62 The Unfunded Mandates Reform Act is
codified at 2 U.S.C. 1532 et seq.

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

regulatory alternatives before
promulgating the rule in question. The
UMRA applies to the OCC and the OTS,
but not the Board or the FDIC.
The OCC and the OTS have asked for
comments and information from core
and opt-in banks on compliance costs
and, generally, on alternative regulatory
approaches, for purposes of evaluating
what actions they need to take in order
to comply with E.O. 12866. That same
information (with cost information
adjusted annually for inflation) is
relevant to those agencies’
determination of whether a budgetary
impact statement is necessary pursuant
to the UMRA. Commenters are therefore
asked to be mindful of the UMRA
requirements when they provide
information about compliance costs and
in suggesting alternatives to the
approach described in this ANPR.
D. Paperwork Reduction Act
Each of the Agencies is subject to the
Paperwork Reduction Act of 1995
(PRA).63 The PRA requires burden
estimates that will likely be based on
some of the same information that is
necessary to prepare an economic
analysis under E.O. 12866 or an
estimate of private sector expenditures
pursuant to the UMRA.
In particular, an agency may not
‘‘conduct or sponsor’’ a collection of
information without conducting an
analysis that includes an estimate of the
‘‘burden’’ imposed by the collection. A
collection of information includes,
essentially, the eliciting of identical
information—whether through
questions, recordkeeping requirements,
or reporting requirements—from ten or
more persons. ‘‘Burden’’ means the
‘‘time, effort, or financial resources
expended by persons to generate,
maintain, or provide information’’ to the
agency. The rulemaking initiated by this
ANPR will likely impose requirements,
either in the regulations themselves or
as part of interagency implementation
guidance, that are covered by the PRA.
In order to estimate burden, the
Agencies will need to know, for
example, the cost—in terms of time and
money—that mandatory and opt-in
banks would have to expend to develop
and maintain the systems, procedures,
and personnel that compliance with the
rules would require. With this in mind,
to assist in their analysis of the
treatment of retail portfolios and other
exposures, the Agencies intend to
request from U.S. institutions additional
quantitative data for which confidential
treatment may be requested in

PO 00000

63 44

U.S.C. § 3501 et seq.

Frm 00049

Fmt 4701

Sfmt 4702

45947

accordance with the Agencies’
applicable rules.
While it will be difficult to identify
those requirements with precision
before a proposed rule is issued, this
notice and the draft supervisory
guidance published elsewhere in
today’s Federal Register generally
describes aspects of the Agencies’
implementation of the New Accord
where new reporting and recordkeeping
requirements would be likely.
Commenters are asked to provide any
estimates they can reasonably derive
about the time, effort, and financial
resources that will be required to
provide the Agencies with the requisite
plans, reports, and records that are
described in this notice and in the
supervisory guidance. Commenters also
are requested to identify any activities
that will be conducted as a result from
the capital and methodological
standards in the framework presented in
this ANPR that would impose new
recordkeeping or reporting burden.
Commenters should specify whether
certain capital and methodological
standards would necessitate the
acquisition or development of new
compliance/ information systems or the
significant modification of existing
compliance/information systems.
List of Acronyms
ABCP Asset-Backed Commercial Paper
ADC Acquisition, Development, and
Construction
AFS Available-for-Sale (securities)
AIG Accord Implementation Group
A–IRB Advanced Internal RatingsBased (approach for credit risk)
ALLL Allowance for Loan and Lease
Losses
AMA Advanced Measurement
Approach (for operational risk)
ANPR Advance Notice of Proposed
Rulemaking
BIS Bank for International Settlements
BSC Basel Committee on Banking
Supervision
CCF Credit Conversion Factor
CDC Community Development
Corporations
CEDE Community and Economic
Development Entity
CF Commodities Finance
CRE Commercial Real Estate
CRM Credit Risk Mitigation
EAD Exposure at Default
EL Expected Loss
FFIEC Federal Financial Institutions
Examination Council
FMI Future Margin Income
GAAP Generally Accepted Accounting
Principles
HVCRE High Volatility Commercial
Real Estate
IMF International Monetary Fund

E:\FR\FM\04AUP2.SGM

04AUP2

45948

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Proposed Rules

IRB Internal Ratings-Based
KIRB Capital for Underlying Pool of
Exposures (securitizations)
LGD Loss Given Default
M Maturity
MDB Multilateral Development Bank
OF Object Finance
OTC Over-the-Counter (derivatives)
PCA Prompt Corrective Action
(regulation)
PD Probability of Default
PDF Probability Density Function
PF Project Finance
PFE Potential Future Exposure
PMI Private Mortgage Insurance
PRA Paperwork Reduction Act
PSE Public-Sector Entity
QIS3 Third Quantitative Impact Study
QRE Qualifying Revolving Exposures
R Asset Correlation
RBA Ratings-Based Approach
(securitizations)
RFA Regulatory Flexibility Act

VerDate jul<14>2003

20:44 Aug 01, 2003

Jkt 200001

S Borrower-Size
SBIC Small Business Investment
Company
SFA Supervisory Formula Approach
(securitizations)
SL Specialized Lending
SME Small-to Medium-Sized
Enterprise
SPE Special Purpose Entity
SSC Supervisory Slotting Criteria
UL Unexpected Loss
UMRA Unfunded Mandates Reform
Act
VaR Value at Risk (model)

PO 00000

Dated: July 17, 2003.
John D. Hawke, Jr.,
Comptroller of the Currency.
By order of the Board of Governors of the
Federal Reserve System, July 21, 2003.
Jennifer J. Johnson,
Secretary of the Board.
Dated at Washington, DC, this 11th day of
July, 2003.
By order of the Board of Directors.
Federal Deposit Insurance Corporation.
Robert E. Feldman,
Executive Secretary.
Dated: July 18, 2003.
By the Office of Thrift Supervision.
James E. Gilleran,
Director.
[FR Doc. 03–18977 Filed 8–1–03; 8:45 am]
BILLING CODE 4810–33–P; 6210–01–P; 6714–01–P;
6720–01–P

Frm 00050

Fmt 4701

Sfmt 4702

E:\FR\FM\04AUP2.SGM

04AUP2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
DEPARTMENT OF THE TREASURY
Office of the Comptroller of the
Currency
[Docket No. 03–15]

FEDERAL RESERVE SYSTEM
[Docket No. OP–1153]

FEDERAL DEPOSIT INSURANCE
CORPORATION
DEPARTMENT OF THE TREASURY
Office of Thrift Supervision
[No. 2003–28]

Internal Ratings-Based Systems for
Corporate Credit and Operational Risk
Advanced Measurement Approaches
for Regulatory Capital
AGENCIES: Office of the Comptroller of
the Currency (OCC), Treasury; Board of
Governors of the Federal Reserve
System (Board); Federal Deposit
Insurance Corporation (FDIC); and
Office of Thrift Supervision (OTS),
Treasury.
ACTION: Draft supervisory guidance with
request for comment.
SUMMARY: The OCC, Board, FDIC, and
OTS (the Agencies) are publishing for
industry comment two documents that
set forth draft supervisory guidance for
implementing proposed revisions to the
risk-based capital standards in the
United States. These proposed
revisions, which would implement the
New Basel Capital Accord in the United
States, are published as an advance
notice of proposed rulemaking (ANPR)
elsewhere in today’s Federal Register.
Under the advanced approaches for
credit and operational risk described in
the ANPR, banking organizations would
use internal estimates of certain risk
components as key inputs in the
determination of their regulatory capital
requirements. The Agencies believe that
supervisory guidance is necessary to
balance the flexibility inherent in the
advanced approaches with high
standards that promote safety and
soundness and encourage comparability
across institutions.
The first document sets forth Draft
Supervisory Guidance on Internal
Ratings-Based Systems for Corporate
Credit (corporate IRB guidance). This
document describes supervisory
expectations for institutions that intend
to adopt the advanced internal ratingsbased approach (A–IRB) for credit risk
as set forth in today’s ANPR. The
corporate IRB guidance is intended to
provide supervisors and institutions

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

with a clear description of the essential
components and characteristics of an
acceptable A–IRB framework. The
guidance focuses specifically on
corporate credit portfolios; further
guidance is expected at a later date on
other credit portfolios (including, for
example, retail and commercial real
estate portfolios).
The second document sets forth Draft
Supervisory Guidance on Operational
Risk Advanced Measurement
Approaches for Operational Risk (AMA
guidance). This document outlines
supervisory expectations for institutions
that intend to adopt an advanced
measurement approach (AMA) for
operational risk as set forth in today’s
ANPR.
The Agencies are seeking comments
on the supervisory standards set forth in
both documents. In addition to seeking
comment on specific aspects of the
supervisory guidance set forth in the
documents, the Agencies are seeking
comment on the extent to which the
supervisory guidance strikes the
appropriate balance between flexibility
and specificity. Likewise, the Agencies
are seeking comment on whether an
appropriate balance has been struck
between the regulatory requirements set
forth in the ANPR and the supervisory
standards set forth in these documents.
DATES: Comments must be received no
later than November 3, 2003.
ADDRESSES: Comments should be
directed to:
OCC: Please direct your comments to:
Office of the Comptroller of the
Currency, 250 E Street, SW., Public
Information Room, Mailstop 1–5,
Washington, DC 20219, Attention:
Docket No. 03–15; fax number (202)
874–4448; or Internet address:
regs.comments@occ.treas.gov. Due to
delays in paper mail delivery in the
Washington area, we encourage the
submission of comments by fax or email whenever possible. Comments may
be inspected and photocopied at the
OCC’s Public Information Room, 250 E
Street, SW., Washington, DC. You may
make an appointment to inspect
comments by calling (202) 874–5043.
Board: Comments should refer to
Docket No. OP–1153 and may be mailed
to Ms. Jennifer J. Johnson, Secretary,
Board of Governors of the Federal
Reserve System, 20th Street and
Constitution Avenue, NW., Washington,
DC, 20551. However, because paper
mail in the Washington area and at the
Board of Governors is subject to delay,
please consider submitting your
comments by e-mail to
regs.comments@federalreserve.gov, or
faxing them to the Office of the

PO 00000

Frm 00001

Fmt 4701

Sfmt 4703

45949

Secretary at 202/452–3819 or 202/452–
3102. Members of the public may
inspect comments in Room MP–500 of
the Martin Building between 9 a.m. and
5 p.m. on weekdays pursuant to
§ 261.12, except as provided in § 261.14,
of the Board’s Rules Regarding
Availability of Information, 12 CFR
261.12 and 261.14.
FDIC: Written comments should be
addressed to Robert E. Feldman,
Executive Secretary, Attention:
Comments, Federal Deposit Insurance
Corporation, 550 17th Street, NW.,
Washington, DC, 20429. Commenters
are encouraged to submit comments by
facsimile transmission to (202) 898–
3838 or by electronic mail to Comments
@FDIC.gov. Comments also may be
hand-delivered to the guard station at
the rear of the 550 17th Street Building
(located on F Street) on business days
between 8:30 a.m. and 5 p.m. Comments
may be inspected and photocopied at
the FDIC’s Public Information Center,
Room 100, 801 17th Street, NW.,
Washington, DC between 9 a.m. and
4:30 p.m. on business days.
OTS: Send comments to Regulation
Comments, Chief Counsel’s Office,
Office of Thrift Supervision, 1700 G
Street, NW., Washington, DC 20552,
Attention: No. 2003–28. Delivery: Hand
deliver comments to the Guard’s desk,
east lobby entrance, 1700 G Street, NW.,
from 9 a.m. to 4 p.m. on business days,
Attention: Regulation Comments, Chief
Counsel’s Office, Attention: No. 2003–
28. Facsimiles: Send facsimile
transmissions to FAX Number (202)
906–6518, Attention: No 2003–28. email: Send e-mails to
regs.comments@ots.treas.gov, Attention:
No. 2003–28, and include your name
and telephone number. Due to
temporary disruptions in mail service in
the Washington, DC area, commenters
are encouraged to send comments by fax
or e-mail, if possible.
FOR FURTHER INFORMATION CONTACT:
OCC: Corporate IRB guidance: Jim
Vesely, National Bank Examiner, Large
Bank Supervision (202/874–5170 or
james.vesely@occ.treas.gov); AMA
guidance: Tanya Smith, Senior
International Advisor, International
Banking & Finance (202/874–4735 or
tanya.smith@occ.treas.gov).
Board: Corporate IRB guidance: David
Palmer, Supervisory Financial Analyst,
Division of Banking Supervision and
Regulation (202/452–2904 or
david.e.palmer@frb.gov); AMA
guidance: T. Kirk Odegard, Supervisory
Financial Analyst, Division of Banking
Supervision and Regulation (202/530–
6225 or thomas.k.odegard@frb.gov). For
users of Telecommunications Device for

E:\FR\FM\04AUN2.SGM

04AUN2

45950

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

the Deaf (‘‘TDD’’) only, contact 202/
263–4869.
FDIC: Corporate IRB guidance and
AMA guidance: Pete D. Hirsch, Basel
Project Manager, Division of
Supervision and Consumer Protection
(202/898–6751 or phirsch@fdic.gov).
OTS: Corporate IRB guidance and
AMA guidance: Michael D. Solomon,
Senior Program Manager for Capital
Policy (202/906–5654); David W. Riley,
Project Manager (202/906–6669),
Supervision Policy; Teresa A. Scott,
Counsel (Banking and Finance) (202/
906–6478); or Eric Hirschhorn, Principal
Financial Economist (202/906–7350),
Regulations and Legislation Division,
Office of the Chief Counsel, Office of
Thrift Supervision, 1700 G Street, NW.,
Washington, DC 20552.
Document 1: Draft Supervisory
Guidance on Internal Ratings-Based
Systems for Corporate Credit
Table of Contents
I. Introduction
A. Purpose
B. Overview of Supervisory Expectations
1. Ratings Assignment
2. Quantification
3. Data Maintenance
4. Control and Oversight Mechanisms
C. Scope of Guidance
D. Timing
II. Ratings for IRB Systems
A. Overview
B. Credit Ratings
1. Rating Assignment Techniques
a. Expert Judgment
b. Models
c. Constrained Judgment
C. IRB Ratings System Architecture
1. Two-Dimensional Rating System
a. Definition of Default
b. Obligor Ratings
c. Loss Severity Ratings
2. Other Considerations of IRB Rating
System Architecture
a. Timeliness of Ratings
b. Multiple Ratings Systems
c. Recognition of the Risk Mitigation
Benefits of Guarantees
3. Validation Process
a. Ratings System Developmental Evidence
b. Ratings System Ongoing Validation
c. Back Testing
III. Quantification of IRB Systems
A. Introduction
1. Stages of the Quantification Process
2. General Principles for Sound IRB
Quantification
B. Probability of Default (PD)
1. Data
2. Estimation
3. Mapping
4. Application
C. Loss Given Default (LGD)
1. Data
2. Estimation
3. Mapping
4. Application
D. Exposure at Default (EAD)
1. Data

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

2. Estimation
3. Mapping
4. Application
E. Maturity (M)
F. Validation
Appendix to Part III: Illustrations of the
Quantification Process
IV. Data Maintenance
A. Overview
B. Data Maintenance Framework
1. Life Cycle Tracking
2. Rating Assignment Data
3. Example Data Elements
C. Data Element Functions
1. Validation and Refinement
2. Developing Parameter Estimates
3. Applying Rating System Improvements
Historically
4. Calculating Capital Ratios and Reporting
to the Public
5. Supporting Risk Management
D. Managing data quality and integrity
1. Documentation and Definitions
2. Electronic Storage
3. Data Gaps
V. Control and Oversight Mechanisms
A. Overview
B. Independence in the Rating Approval
Process
C. Transparency
D. Accountability
1. Responsibility for Assigning Ratings
2. Responsibility for Rating System
Performance
E. Use of Ratings
F. Rating System Review (RSR)
G. Internal Audit
1. External Audit
H. Corporate Oversight

I. Introduction
A. Purpose
This document describes supervisory
expectations for banking organizations
(institutions) adopting the advanced
internal ratings-based approach (IRB) for
the determination of minimum
regulatory risk-based capital
requirements. The focus of this
guidance is corporate credit portfolios.
Retail, commercial real estate,
securitizations, and other portfolios will
be the focus of later guidance. This draft
guidance should be considered with the
advance notice of proposed rulemaking
(ANPR) on revisions to the risk-based
capital standard published elsewhere in
today’s Federal Register.
The primary objective of IRB is to
enhance the sensitivity of regulatory
capital requirements to credit risk. To
accomplish that objective, IRB harnesses
a bank’s own risk rating and
quantification capabilities. In general,
the IRB approach reflects and extends
recent developments in risk
management and banking supervision.
However, the degree to which any
individual bank will need to modify its
own credit risk management practices to
deliver accurate and consistent IRB risk

PO 00000

Frm 00002

Fmt 4701

Sfmt 4703

parameters will vary from institution to
institution.
This guidance is intended to provide
supervisors and institutions with a clear
description of the essential components
and characteristics of an acceptable IRB
framework. Toward that end, this
document sets forth IRB system
supervisory standards that are
highlighted in bold and designated by
the prefix ‘‘S.’’ Whenever possible, these
supervisory standards are principlebased to enable institutions to
implement the framework flexibly.
However, when prudential concerns or
the need for standardization override
the desire for flexibility, the supervisory
standards are more detailed. Ultimately,
institutions must have credit risk
management practices that are
consistent with the substance and spirit
of the standards in this guidance.
The IRB conceptual framework
outlined in this document is intended
neither to dictate the precise manner by
which institutions should seek to meet
supervisory expectations, nor to provide
technical guidance on how to develop
such a framework. As institutions
develop their IRB systems in
anticipation of adopting them for
regulatory capital purposes, supervisors
will be evaluating, on an individual
bank basis, the extent to which
institutions meet the standards outlined
in this document. In evaluating
institutions, supervisors will rely on
this supervisory guidance as well as
examination procedures, which will be
developed separately. This document
assumes that readers are familiar with
the proposed IRB approach to
calculating minimum regulatory capital
articulated in the ANPR.
B. Overview of Supervisory Expectations
Rigorous credit risk measurement is a
necessary element of advanced risk
management. Qualifying institutions
will use their internal rating systems to
associate a probability of default (PD)
with each obligor grade, as well as a loss
given default (LGD) with each credit
facility. In addition, institutions will
estimate exposure at default (EAD) and
will calculate the effective remaining
maturity (M) of credit facilities.
Qualifying institutions will be
expected to have an IRB system
consisting of four interdependent
components:
• A system that assigns ratings and
validates their accuracy (Chapter 1),
• A quantification process that
translates risk ratings into IRB
parameters (Chapter 2),
• A data maintenance system that
supports the IRB system (Chapter 3),
and,

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
• Oversight and control mechanisms
that ensure the system is functioning as
intended and producing accurate ratings
(Chapter 4).
Together these rating, quantification,
data, and oversight mechanisms present
a framework for defining and improving
the evaluation of credit risk.
It is expected that rating systems will
operate dynamically. As ratings are
assigned, quantified and used, estimates
will be compared with actual results
and data will be maintained and
updated to support oversight and
validation efforts and to better inform
future estimates. The rating system
review and internal audit functions will
serve as control mechanisms that ensure
that the process of ratings assignment
and quantification function according to
policy and design and that
noncompliance and weaknesses are
identified, communicated to senior
management and the board, and
addressed. Rating systems with
appropriate data and oversight feedback
mechanisms foster a learning
environment that promotes integrity in
the rating system and continuing
refinement.
IRB systems need the support and
oversight of the board and senior
management to ensure that the various
components fit together seamlessly and
that incentives to make the system
rigorous extend across line, risk
management, and other control groups.
Without strong board and senior
management support and involvement,
rating systems are unlikely to provide
accurate and consistent risk estimates
during both good and bad times.
The new regulatory minimum capital
requirement is predicated on an
institution’s internal systems being
sufficiently advanced to allow a full and
accurate assessment of its risk
exposures. Under the new framework,
an institution could experience a
considerable capital shortfall in the
most difficult of times if its risk
estimates are materially understated.
Consequently, the IRB framework
demands a greater level of validation
work and controls than supervisors have
required in the past. When properly
implemented, the new framework holds
the potential for better aligning
minimum capital requirements with the
risk taken, pushing capital requirements
higher for institutions that specialize in
riskier types of lending, and lower for
those that specialize in safer risk
exposures.
Supervisors will evaluate compliance
with the supervisory standards for each
of the four components of an IRB
system. However, evaluating
compliance with each of the standards

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

individually will not be sufficient to
determine an institution’s overall
compliance. Rather, supervisors and
institutions must also evaluate how well
the various components of an
institution’s IRB system complement
and reinforce one another to achieve the
overall objective of accurate measures of
risk. In performing their evaluation,
supervisors will need to exercise
considerable supervisory judgment,
both in evaluating the individual
components and the overall IRB
framework. A summary of the key
supervisory expectations for each of the
IRB components follows.
Ratings Assignment
The first component of an IRB system
involves the assignment and validation
of ratings (see Chapter 1). Ratings must
be accurately and consistently applied
to all corporate credit exposures and be
subject to initial and ongoing validation.
Institutions will have latitude in
designing and operating IRB rating
systems subject to five broad standards:
Two-dimensional risk-rating system—
IRB institutions must be able to make
meaningful and consistent
differentiations among credit exposures
along two dimensions—obligor default
risk and loss severity in the event of a
default.
Rank order risks—IRB institutions
must rank obligors by their likelihood of
default, and facilities by the loss
severity expected in default.
Calibration—IRB obligor ratings must
be calibrated to values of the probability
of default (PD) parameter and loss
severity ratings must be calibrated to
values of the loss given default (LGD)
parameter.
Accuracy—Actual long-run actual
default frequencies for obligor rating
grades must closely approximate the
PDs assigned to those grades and
realized loss rates on loss severity
grades must closely approximate the
LGDs assigned to those grades.
Validation process—IRB institutions
must have ongoing validation processes
for rating systems that include the
evaluation of developmental evidence,
process verification, benchmarking, and
the comparison of predicted parameter
values to actual outcomes (back-testing).
Quantification
The second component of an IRB
system is a quantification process (see
Chapter 2). Since obligor and facility
ratings may be assigned separately from
the quantification of the associated PD
and LGD parameters, quantification is
addressed as a separate process. The
quantification process must produce
values not only for PD and LGD but also

PO 00000

Frm 00003

Fmt 4701

Sfmt 4703

45951

for EAD and for the effective remaining
maturity (M). The quantification of
those four parameters is expected to be
the result of a disciplined process. The
key considerations for effective
quantification are as follows:
Process—IRB institutions must have a
fully specified process covering all
aspects of quantification (reference data,
estimation, mapping, and application).
Documentation—The quantification
process, including the role and scope of
expert judgment, must be fully
documented and updated periodically.
Updating—Parameter estimates and
related documentation must be updated
regularly.
Review—A bank must subject all
aspects of the quantification process,
including design and implementation,
to an appropriate degree of independent
review and validation.
Constraints on Judgment—Judgmental
adjustments may be an appropriate part
of the quantification process, but must
not be biased toward lower risk
estimates.
Conservatism—Parameter estimates
must incorporate a degree of
conservatism that is appropriate for the
overall robustness of the quantification
process.
Data Maintenance
The third component of an IRB
system is an advanced data management
system that produces credible and
reliable risk estimates (see Chapter 3).
The broad standard governing an IRB
data maintenance system is that it
supports the requirements for the other
IRB system components, as well as the
institution’s broader risk management
and reporting needs. Institutions will
have latitude in managing their data,
subject to the following key data
maintenance standards:
Life Cycle Tracking—Institutions
must collect, maintain, and analyze
essential data for obligors and facilities
throughout the life and disposition of
the credit exposure.
Rating Assignment Data—Institutions
must capture all significant quantitative
and qualitative factors used to assign the
obligor and loss severity rating.
Support of IRB System—Data
collected by institutions must be of
sufficient depth, scope, and reliability
to:
• Validate IRB system processes,
• Validate parameters,
• Refine the IRB system,
• Develop internal parameter
estimates,
• Apply improvements historically,
• Calculate capital ratios,
• Produce internal and public reports,
and

E:\FR\FM\04AUN2.SGM

04AUN2

45952

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

• Support risk management.

D. Timing

Control and Oversight Mechanisms
The fourth component of an IRB
system is comprised of control and
oversight mechanisms that ensure that
the various components of the IRB
system are functioning as intended (see
Chapter 4). Given the various uses of
internal risk ratings, including their
direct link to regulatory capital
requirements, there is enormous,
sometimes conflicting, pressure on
banks’ internal rating systems. Control
structures are subject to the following
broad standards:
Interdependent System of Controls—
IRB institutions must implement a
system of interdependent controls that
include the following elements:
• Independence,
• Transparency,
• Accountability,
• Use of ratings,
• Rating system review,
• Internal audit, and
• Board and senior management
oversight.
Checks and Balances—Institutions
must combine the various control
mechanisms in a way that provides
checks and balances for ensuring IRB
system integrity.
The system of oversight and controls
required for an effective IRB system may
operate in various ways within
individual institutions. This guidance
does not prescribe any particular
organizational structure for IRB
oversight and control mechanisms.
Banks have broad latitude to implement
structures that are most effective for
their individual circumstances, as long
as those structures support and enhance
the institution’s ability to satisfy the
supervisory standards expressed in this
document.
C. Scope of Guidance
This draft guidance reflects work
performed by supervisors to evaluate
and compare current practices at
institutions with the concepts and
requirements for an IRB framework. For
instances in which a range of practice
was observable, examples are provided
on how certain practices may or may
not qualify. However, in many other
instances, practices were at such an
early stage of development that it was
not feasible to describe specific
examples. In those cases, requirements
tend to be principle-based and without
examples. Given that institutions are
still in the early stages of developing
qualifying IRB systems, it is expected
that this guidance will evolve over time
to more explicitly take into account new
and improving practices.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

S. An IRB system must be operating
fully at least one year prior to the
institution’s intended start date for the
advanced approach.
As noted in the ANPR, the significant
challenge of implementing a fully
complying IRB system requires that
institutions and supervisors have
sufficient time to observe whether the
IRB system is delivering risk-based
capital figures with a high level of
integrity. The ability to observe the
institution’s ratings architecture,
validation, data maintenance and
control functions in a fully operating
environment prior to implementation
will help identify how well the IRB
system design functions in practice.
This will be particularly important
given that in the first year of
implementation institutions will not
only be subject to the new minimum
capital requirements, but will also be
disclosing risk-based capital ratios for
the public to rely upon in the
assessment of the institution’s financial
health.
II. Ratings for IRB Systems
A. Overview
This chapter describes the design and
operation of risk-rating systems that will
be acceptable in an internal ratingsbased (IRB) framework. Banks will have
latitude in designing and operating IRB
rating systems, subject to five broad
standards:
Two-dimensional risk-rating system—
IRB institutions must be able to make
meaningful and consistent
differentiations among credit exposures
along two dimensions—obligor default
risk and loss severity in the event of a
default.
Rank order risks—IRB institutions
must rank obligors by their likelihood of
default, and facilities by the loss
severity expected in default.
Calibration—IRB obligor ratings must
be calibrated to values of the probability
of default (PD) parameter and loss
severity ratings must be calibrated to
values of the loss given default (LGD)
parameter.
Accuracy—Actual long-run actual
default frequencies for obligor rating
grades must closely approximate the
PDs assigned to those grades and actual
loss rates on loss severity grades must
closely approximate the LGDs assigned
to those grades.
Validation process—IRB institutions
must have ongoing validation processes
for rating systems that include the
evaluation of developmental evidence,
process verification, benchmarking, and

PO 00000

Frm 00004

Fmt 4701

Sfmt 4703

the comparison of predicted parameter
values to actual outcomes (back-testing).
B. Credit Ratings
In general, a credit rating is a
summary indicator of the relative risk
on a credit exposure. Credit ratings can
take many forms. The most widely
known credit ratings are the public
agency ratings, which are expressed as
letters; bank internal ratings tend to be
expressed as whole numbers—for
example, 1 through 10. Some rating
model outputs are expressed in terms of
probability of default or expected
default frequency, in which case they
may be more than relative measures of
risk. Regardless of the form, meaningful
credit ratings share two characteristics:
• They group credits to discriminate
among possible outcomes.
• They rank the perceived levels of
credit risk.
Banks have used credit ratings of
various types for a variety of purposes.
Some ratings are intended to rank
obligors by risk of default and some are
intended to rank facilities1 by expected
loss, which incorporates risk of default
and loss severity. Bank rating systems
that are geared solely to expected loss
will need to be amended to meet the
two-dimensional requirements of the
IRB approach.
Rating Assignment Techniques
Banks use different techniques, such
as expert judgment and models, to
assign credit risk ratings. For banks
using the IRB approach, how ratings are
assigned is important because different
techniques will require different
validation processes and control
mechanisms to ensure the integrity of
the rating system. To assist the
discussion of rating architecture
requirements, described below are some
of the current rating assignment
techniques. Any of these techniques—
expert judgment, models, constrained
judgment, or a combination thereof—
could be acceptable within an IRB
system, provided the bank meets the
standards outlined in this document.
Expert Judgment
Historically, banks have used expert
judgment to assign ratings to
commercial credits. With this
technique, an individual weighs
relevant information and reaches a
conclusion about the appropriate risk
rating. Presumably, the rater makes
informed judgments based on
knowledge gained through experience
and training.
1 Facilities—loans, lines, or other separate
extensions of credit to an obligor.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
The key feature of expert-judgment
systems is flexibility. The prevalence of
judgmental rating systems reflects the
view that the determinants of default are
too complicated to be captured by a
single quantitative model. The quality of
management is often cited as an
example of a risk determinant that is
difficult to assess through a quantitative
model. In order to foster internal
consistency, banks employing expert
judgment rating systems typically
provide narrative guidelines that set out
ratings criteria. However, the expert
must decide how narrative guidelines
apply to a given set of circumstances.
The flexibility possible in the
assignment of judgmental ratings has
implications for the types of ratings
review that are feasible. As part of the
ratings validation process, banks will
attempt to confirm that raters follow
bank policy. However, two individuals
exercising judgment can use the same
information to support different ratings.
Thus, the review of an expert judgment
rating system will require an expert who
can identify the impact of policy and
the impact of judgment on a rating.
Models
In recent years, models have been
developed for use in rating commercial
credits. In a model-based approach,
inputs are numeric and provide
quantitative and qualitative information
about an obligor. The inputs are
combined using mathematical equations
to produce a number that is translated
into a categorical rating. An important
feature of models is that the rating is
perfectly replicable by another party,
given the same inputs.
The models used in credit rating can
be distinguished by the techniques used
to develop them. Some models may rely
on statistical techniques while others
rely on expert-judgment techniques.
Statistical models. Statistically
developed models are the result of
statistical optimization, in which welldefined mathematical criteria are used
to choose the model that has the closest
fit to the observed data. Numerous
techniques can be used to build
statistical models; regression is one
widely recognized example. Regardless
of the specific statistical technique, a
knowledgeable independent reviewer
will have to exercise judgment in
evaluating the reasonableness of a
model’s development, including its
underlying logic, the techniques used to
handle the data, and the statistical
model building techniques.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

45953

Expert-derived models.2 Several
banks have built rating models by
asking their experts to decide what
weights to assign to critical variables in
the models. Drawing on their
experience, the experts first identify the
observable variables that affect the
likelihood of default. They then reach
agreement on the weights to be assigned
to each of the variables. Unlike
statistical optimization, the experts are
not necessarily using clear, consistent
criteria to select the weights attached to
the variables. Indeed, expert-judgment
model building is often a practical
choice when there is not enough data to
support a statistical model building.
Despite its dependence on expert
judgment, this method can be called
model-based as long as the result—the
equation, most likely with linear
weights—is used as the basis to rate the
credits. Once the equation is set, the
model shares the feature of replicability
with statistically derived models.
Generally, independent credit experts
use judgment to evaluate the
reasonableness of the development of
these models.

weight to attach to the model result. The
appeal of this approach is that the
model combines many pieces of
information into a single output, which
simplifies analysis, while the rater
retains flexibility regarding the use of
the model output.
Model-based ratings with judgmental
overrides. When banks use rating
models, individuals are generally
permitted to override the results under
certain conditions and within tolerance
levels for frequency. Credit-rating
systems in which individuals can
override models raise many of the same
issues presented separately by pure
judgment and model-based systems. If
overrides are rare, the system can be
evaluated largely as if it is a modelbased system. If, however, overrides are
prevalent, the system will be evaluated
more like a judgmental system.
Since constrained judgment systems
combine features of both expert
judgment and model-based systems,
their evaluation will require the skills
required to evaluate both of these other
systems.

Constrained Judgment
The alternatives just described
present the extremes, but in practice,
many banks use rating systems that
combine models with judgment. Two
approaches are common.
Judgmental systems with quantitative
guidelines or model results as inputs.
Historically, the most common
approach to rating has involved
individuals exercising judgment about
risks, subject to policy guidelines
containing quantitative criteria such as
minimum values for particular financial
ratios. Banks develop quantitative
criteria to guide individuals in assigning
ratings, but often believe that those
criteria do not adequately reflect the
information needed to assign a rating.
One version of this constrained
judgment approach features a model
output as one among several criteria that
an individual may consider in assigning
ratings. The individual assigning the
rating is responsible for prioritizing the
criteria, reconciling conflicts between
criteria, and if warranted, overriding
some criteria. Even if individuals
incorporate model results as one of the
factors in their ratings, they will
exercise judgment in deciding what

Two-Dimensional Rating System
S. IRB risk rating systems must have
two rating dimensions—obligor and loss
severity ratings.
S. IRB obligor and loss severity ratings
must be calibrated to values of the
probability of default (PD) and the loss
given default (LGD), respectively.
Regardless of the type of rating
system(s) used by an institution, the IRB
approach imposes some specific
requirements. The first requirement is
that an IRB rating system must be twodimensional. Banks will assign obligor
ratings, which will be associated with a
PD. They will also either assign a loss
severity rating, which will be associated
with LGD values, or directly assign LGD
values to each facility. The process of
assigning the obligor and loss severity
ratings—hereafter referred to as the
rating system—is discussed below, and
the process of calibrating obligor and
loss severity ratings to PD and LGD
parameters is discussed in Chapter 2.
S. Banks must record obligor defaults
in accordance with the IRB definition of
default.

2 Some banks have developed credit rating
models that they refer to as ‘‘scorecards,’’ but they
have used expert judgment to derive the weights.
While they are models, they are not scoring models
in the now conventional use of the term. In its
conventional use, the term scoring model is
reserved for a rating model derived using statistical
techniques.

PO 00000

Frm 00005

Fmt 4701

Sfmt 4703

C. IRB Ratings System Architecture

Definition of Default
The consistent identification of
defaults is fundamental to any IRB
rating system. For IRB purposes, a
default is considered to have occurred
with regard to a particular obligor when
either or both of the two following
events have taken place:
• The obligor is past due more than
90 days on any material credit

E:\FR\FM\04AUN2.SGM

04AUN2

45954

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

obligation to the banking group.
Overdrafts will be considered as being
past due once the customer has
breached an advised limit or been
advised of a limit smaller than current
outstandings.
• The bank considers that the obligor
is unlikely to pay its credit obligations
to the banking group in full, without
recourse by the bank to actions such as
liquidating collateral (if held).
Any obligor (or its underlying credit
facilities) that meets one or more of the
following conditions is considered
unlikely to pay and therefore in default:
• The bank puts the credit obligation
on non-accrual status.
• The bank makes a charge-off or
account-specific provision resulting
from a significant perceived decline in
credit quality subsequent to the bank
taking on the exposure.
• The bank sells the credit obligation
at a material credit-related economic
loss.
• The bank consents to a distressed
restructuring of the credit obligation
where this is likely to result in a
diminished financial obligation caused
by the material forgiveness, or
postponement, of principal, interest or
(where relevant) fees.
• The bank has filed for the obligor’s
bankruptcy or a similar order in respect
of the obligor’s credit obligation to the
banking group.
• The obligor has sought or has been
placed in bankruptcy or similar
protection where this would avoid or
delay repayment of the credit obligation
to the banking group.
While most conditions of default
currently are identified by bank
reporting systems, institutions will need
to augment data capture systems to
collect those default circumstances that
may not have been traditionally
identified. These include facilities that
are current and still accruing but where
the obligor declared or was placed in
bankruptcy. They must also capture so
called ‘‘silent defaults’’—defaults when
the loss on a facility was avoided by
liquidating collateral.
Loan sales on which a bank
experiences a material loss due to credit
deterioration are considered a default.
Material credit related losses are defined
as XX. (The agencies seek comment on
how to define ‘‘material’’ loss in the
case of loans sold at a discount). Banks
should ensure that they have adequate
systems to identify such transactions
and to maintain adequate records so that
reviewers can assess the adequacy of the
institution’s decision-making process in
this area.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Obligor Ratings
S. Banks must assign discrete obligor
grades.
While banks may use models to
estimate probabilities of default for
individual obligors, the IRB approach
requires banks to group the obligors into
discrete grades. Each obligor grade, in
turn, must be associated with a single
PD.
S. The obligor-rating system must
result in a ranking of obligors by
likelihood of default.
The proper operation of the obligorrating system will feature a ranking of
obligors by likelihood of default. For
example, if a bank uses a rating system
based on a 10-point scale, with 1
representing obligors of highest
financial strength and 10 representing
defaulted obligors, grades 2 through 9
should represent groups of everincreasing risk. In a rating system in
which risk increases with the grade, an
obligor with a grade 4 is riskier than an
obligor with a grade 2, but need not be
twice as risky.
S. Separate exposures to the same
obligor must be assigned to the same
obligor rating grade.
As noted above, the IRB framework
requires that the obligor rating be
distinct from the loss severity rating,
which is assigned to the facility.
Collateral and other facility
characteristics should not influence the
obligor rating. For example, in a 1-to-10
rating system, where risk increases with
the number grade, a defaulted borrower
with a fully cash-secured transaction
should be rated a 10—defaulted—
regardless of the remote expectation of
loss. Likewise, a borrower whose
financial condition warrants the highest
investment grade rating should be rated
a 1 even if the bank’s transactions are
subordinate to other creditors and
unsecured. Since the rating is assigned
to the obligor and not the facility,
separate exposures to the same obligor
must be assigned to the same obligor
rating grade.
At the bottom of any IRB system
rating scale is a default grade. Once an
obligor is considered to be in default for
IRB purposes, that obligor must be
assigned a default grade until such time
as its financial condition and
performance improve sufficiently to
clearly meet the bank’s internal rating
definition for one of its non-default
grades. Once an obligor is in default on
any material credit obligation to the
subject bank, all of its facilities at that
institution are considered to be in
default.
S. In assigning an obligor to a rating
category, the bank must assess the risk

PO 00000

Frm 00006

Fmt 4701

Sfmt 4703

of obligor default over a period of at
least one year.
S. Obligor ratings must reflect the
impact of financial distress.
In assigning an obligor to a rating
category, the bank must assess the risk
of obligor default over a period of at
least one year. This use of a one-year
assessment horizon does not mean that
a bank should limit its consideration to
outcomes for that obligor that are most
likely over that year; the rating must
take into account possible adverse
events that might increase an obligor’s
likelihood of default.
Rating Philosophy—Decisions
Underlying Ratings Architecture
S. Banks must adopt a ratings
philosophy. Policy guidelines should
describe the ratings philosophy,
particularly how quickly ratings are
expected to migrate in response to
economic cycles.
S. A bank’s capital management
policy must be consistent with its
ratings philosophy in order to avoid
capital shortfalls in times of systematic
economic stress.
In the IRB framework, banks assign
obligors to groups that are expected to
share common default frequencies. That
general description, however, still
leaves open different possible
implementations, depending on how the
bank defines the set of possible adverse
events that the obligor might face. A
bank must decide whether obligors are
grouped by expected common default
frequency over the next year (a so-called
point-in-time rating system) or by an
expected common default frequency
over a wider range of possible stress
outcomes (a so-called through-the-cycle
rating system). Choosing between a
point-in-time system and a through-thecycle system yields a rating philosophy.
In point in time rating systems,
obligors are assigned to groups that are
expected to share a common default
frequency in a particular year. Point-intime ratings change from year to year as
borrowers’ circumstances change,
including changes due to the economic
possibilities faced by the borrowers.
Since the economic circumstances of
many borrowers reflect the common
impact of the general economic
environment, the transitions in point-intime ratings will reflect that systematic
influence. A Merton-style probability of
default prediction model is commonly
believed to be an example of a point-intime approach to rating (although that
may depend on the specific
implementation of the model).
Through-the-cycle rating systems do
not ask the question, what is the
probability of default over the next year.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
Instead, they assign obligors to groups
that would be expected to share a
common default frequency if the
borrowers in them were to experience
distress, regardless of whether that
distress is in the next year. Thus, as the
descriptive title suggests, this rating
philosophy abstracts from the near-term
economic possibilities and considers a
richer assessment of the possibilities.
Like point-in-time ratings, through the
cycle ratings will change from year to
year due to changes in borrower
circumstance. However, since this rating
philosophy abstracts from the
immediate economic circumstance and
considers the implications of
hypothetical stress circumstances, year
to year transitions in ratings will be less
influenced by changes in the actual
economic environment. The ratings
agencies are commonly believed to use
through-the-cycle rating approaches.
Current practice in many banks in the
U.S. is to rate obligors using an
approach that combines aspects of both
point-in-time and through the cycle
approaches. The explanation provided
by banks that combine those approaches
is that they want rating transitions to
reflect the directional impact of changes
in the economic environment, but that
they do not want all of the volatility in
ratings associated with a point-in-time
approach.
Regardless of which ratings
philosophy a bank chooses, an IRB bank
must articulate clearly its approach and
the implications of that choice. As part
of the choice of rating philosophy, the
bank must decide whether the same
ratings philosophy will be employed for
all of the bank’s portfolios. And
management must articulate the
implications that the bank’s ratings
philosophy has on the bank’s capital
planning process. If a bank chooses a
ratings philosophy that is likely to result
in ratings transitions that reflect the
impact of the economic cycle, its capital
management policy must be designed to
avoid capital shortfalls in times of
systematic economic stress.
Obligor-Rating Granularity
S. An institution must have at least
seven obligor grades that contain only
non-defaulted borrowers and at least
one grade to which only defaulted
borrowers are assigned.
The number of grades used in a rating
system should be sufficient to
reasonably ensure that management can
meaningfully differentiate risk in the
portfolio, without being so large that it
limits the practical use of the rating
system. To determine the appropriate
number of grades beyond the minimum
seven non-default grades, each

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

institution must perform its own
internal analysis.
S. An institution must justify the
number of obligor grades used in its
rating system and the distribution of
obligors across those grades.
The mere existence of an exposure
concentration in a grade (or grades) does
not, by itself, reflect weakness in a
rating system. For example, banks may
focus on a particular type of lending,
such as asset-based lending, in which
the borrowers may have similar default
risk. Banks with such focused lending
activities may use close to the minimum
number of obligor grades, while banks
with a broad range of lending activities
should have more grades. However,
banks with a high concentration of
obligors in a particular grade are
expected to perform a thorough analysis
that supports such a concentration.
A significant concentration within an
obligor grade may be suspected if the
financial strength of the borrowers
within that grade varies considerably. If
obligors seem unduly concentrated,
then management should ask
themselves the following questions:
• Are the criteria for each grade clear?
Those rating criteria may be too vague
to allow raters to make clear
distinctions. Ambiguity may be an issue
throughout the rating scale or it may be
limited to the most commonly used
ratings.
• How diverse are the obligors? That
is how many market segments (for
example, large commercial, middle
market, private banking, small business,
geography, etc.) are significantly
represented in the bank’s borrower
population? If a bank’s commercial loan
portfolio is not concentrated in one
market segment, its risk rating
distribution is not likely to be
concentrated.
• How broad are the bank’s internal
rating categories compared to those of
other lenders? The bank may be able to
learn enough from publicly available
information to adjust its rating criteria.
Some banks use ‘‘modifiers’’ to
provide more risk differentiation to a
given rating system. A risk rating
modified with a plus, minus or other
indicator does not constitute a separate
grade unless the bank has developed a
distinct rating definition and criteria for
the modified grade. In the absence of
such distinctions, grades such as 5, 5+,
and 5¥ are viewed as a single grade for
regulatory capital purposes regardless of
the existence of the modifiers.
Loss Severity Ratings
S. Banks must rank facilities by the
expected severity of the loss upon
default.

PO 00000

Frm 00007

Fmt 4701

Sfmt 4703

45955

The second dimension of an IRB
system is the loss severity rating, which
is calibrated to LGD. A facility’s LGD
estimate is the loss the bank is likely to
incur in the event that the obligor
defaults, and is expressed as a
percentage of exposure at the time of
default. LGD estimates can be assigned
either through the use of a loss severity
rating system or they can be directly
assigned to each facility.
LGD analysis is still in very early
stages of development relative to default
risk modeling. Academic research in
this area is relatively sparse, data are not
abundant, and industry practice is still
widely varying and evolving. Given the
lack of data and the lack of research into
LGD modeling, some banks are likely, as
a first step, to segment their portfolios
by a handful of available characteristics
and determine the appropriate LGDs for
those segments. Over time, banks’ LGD
methodologies are expected to evolve.
Long-standing banking experience and
existing research on LGD, while
preliminary, suggests that collateral
values, seniority, industry, etc. are
predictive of loss severity.
S. Banks must have empirical support
for LGD rating systems regardless of
whether they use an LGD grading
system or directly assign LGD estimates.
Whether a bank chooses to assign
LGD values directly or, alternatively, to
rate facilities and then quantify the LGD
for the rating grades, the key
requirement is that it will need to
identify facility characteristics that
influence LGD. Each of the loss severity
rating categories must be associated
with an empirically supported LGD
estimate. In much the same way an
obligor-rating system ranks exposures
by the probability of default, a facility
rating system must rank facilities by the
likely loss severity.
Regardless of the method used to
assign LGDs (loss severity grades or
direct LGD estimation), data used to
support the methodology must be
gathered systematically. For many
banks, the quality and quantity of data
available to support the LGD estimation
process will have an influence on the
method they choose.
Stress Condition LGDs
S. Loss severity ratings must reflect
losses expected during periods with a
relatively high number of defaults.
Like obligor ratings, which group
obligors by expected default frequency,
loss severity ratings assign facilities to
groups that are expected to experience
a common loss severity. However, the
different treatment accorded to PD and
LGD in the model used to calculate IRB
capital requirements mandates an

E:\FR\FM\04AUN2.SGM

04AUN2

45956

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

asymmetric treatment of obligor and
loss severity ratings. Obligor ratings
assign obligors to groups that are
expected to experience common default
frequencies across a number of years,
some of which are years of general
economic stress and some of which are
not. In contrast, loss severity ratings (or
estimates) must pertain to losses
expected during periods with a high
number of defaults—particular years
that can be called stress conditions. For
cases in which loss severities do not
have a material degree of cyclical
variability, use of a long-run default
weighted average is appropriate,
although stress condition LGD generally
exceeds these averages.
Loss Severity Rating/LGD Granularity
S. Banks must have a sufficiently fine
loss severity grading system or
prediction model to avoid grouping
facilities with widely varying LGDs
together.
While there is no stated minimum
number of loss severity grades, the
systems that provide LGD estimates
must be flexible enough to adequately
segment facilities with significantly
varying LGDs. Banks should have a
sufficiently fine LGD grading system or
LGD prediction model to avoid grouping
facilities with widely varying LGDs
together. For example, a bank using a
loss severity rating-scale approach that
has credit products with a variety of
collateral packages or financing
structures would be expected to have
more LGD grades than those institutions
with fewer options in their credit
products.
Other Considerations of IRB Rating
System Architecture
Timeliness of Ratings
S. All risk ratings must be updated
whenever new relevant information is
received, but must be updated at least
annually.
A bank must have a policy that
requires a dynamic ratings approach
ensuring that obligor and loss severity
ratings reflect current information. That
policy must also specify minimum
financial reporting and collateral
valuation requirements. For example, at
the time of servicing events, banks
typically receive updated financial
information on obligors. For cases in
which loss severity grades or estimates
are dependent on collateral values or
other factors that change periodically,
that policy must take into account the
need to update these factors.
Banks’ policies may include an
alternative rating update timetable for
exposures below a de minimus amount

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

that is justified by the lack of materiality
of the potential impact on capital. For
example, some banks use triggering
events to prompt an update of their
ratings on de minimus exposures rather
than adhering to a specific timetable.
Multiple Ratings Systems
Some banks may develop one riskrating system that can be used across the
entire commercial loan portfolio.
However, a bank can choose to deploy
any number of rating systems as long as
all exposures are assigned PD and LGD
values. A different rating system could
be used for each business line and each
rating system could use a different
rating scale. A bank could also use a
different rating system for each business
line with each system using a common
rating scale. Rating models could be
used for some portfolios and expert
judgment systems for others. An
institution’s complexity and
sophistication, as well as the size and
range of products offered, will affect the
types and numbers of rating systems
employed.
While using a number of rating
systems is feasible, such a practice
might make it more difficult to meet
supervisory standards. Each rating
system must conform to the standards in
this guidance and must be validated for
accuracy and consistency. The
requirement that each rating systems be
calibrated to parameter values imposes
the ultimate constraint, which is that
ratings be applied consistently.
Recognition of the Risk Mitigation
Benefits of Guarantees
S. Banks reflecting the risk-mitigating
effect of guarantees must do so by either
adjusting PDs or LGDs, but not both.
S. To recognize the risk-mitigating
effects of guarantees, institutions must
ensure that the written guarantee is
evidenced by an unconditional and
legally enforceable commitment to pay
that remains in force until the debt is
satisfied in full.
Adjustments for guarantees must be
made in accordance with specific
criteria contained in the bank’s credit
policy. The criteria should be plausible
and intuitive, and should address the
guarantor’s ability and willingness to
meet its obligations. Banks are expected
to gather evidence that confirms the
risk-mitigating effect of guarantees.
Other forms of written third-party
support (for example, comfort letters or
letters of awareness) that are not legally
binding should not be used to adjust PD
or LGD unless a bank can demonstrate
through analysis of internal data the
risk-mitigating effect of such support.
Banks may not adjust PDs or LGDs to

PO 00000

Frm 00008

Fmt 4701

Sfmt 4703

reflect implied support or verbal
assurances.
Regardless of the method used to
recognize the risk-mitigating effects of
guarantees, a bank must adopt an
approach that is applied consistently
over time and across the portfolio.
Moreover, the onus is on the bank to
demonstrate that its approach is
supported by logic and empirical
results. While guarantees may provide
grounds for adjusting PD or LGD, they
cannot result in a lower risk weight than
that assigned to a similar direct
obligation of the guarantor.3
Validation Process
S. IRB rating system architecture must
be designed to ensure rating system
accuracy.
As part of their IRB rating system
architecture, banks must implement a
process to ensure the accuracy of their
rating systems. Rating system accuracy
is defined as the combination of the
following outcomes:
• The actual long-run average default
frequency for each rating grade is not
significantly greater than the PD
assigned to that grade.
• The actual stress-condition loss
rates experienced on defaulted facilities
are not significantly greater than the
LGD estimates assigned to those
facilities.
Some differences across individual
grades between observed outcomes and
the estimated parameter inputs to the
IRB equations can be expected. But if
systematic differences suggest a bias
toward lowering regulatory capital
requirements, the integrity of the rating
system (of either the PD or LGD
dimensions or of both) becomes suspect.
Validation is the set of activities
designed to give the greatest possible
assurances of ratings system accuracy.
S. Banks must have ongoing
validation processes that include the
review of developmental evidence,
ongoing monitoring, and the
comparison of predicted parameter
values to actual outcomes (back-testing).
Validation is an integral part of the
rating system architecture. Banks must
have processes designed to give
3 The probability that an obligor and a guarantor
(who supports the obligor’s debt) will both default
on a debt is lower than the probability that either
the obligor or the guarantor will default. This
favorable risk-mitigation effect is known as the
reduced likelihood of ‘‘double default.’’ In
determining their rating criteria and procedures,
banks are not permitted to consider possible
favorable effects of imperfect expected correlation
between default events for the borrower and
guarantor for purposes of regulatory capital
requirements. Thus, the adjusted risk weight cannot
reflect the risk mitigation of double default. The
ANPR solicits public comment on the doubledefault issues.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
reasonable assurances of their rating
systems’ accuracy. The ongoing process
to confirm and ensure rating system
accuracy consists of:
• The evaluation of developmental
evidence,
• Ongoing monitoring of system
implementation and reasonableness
(verification and benchmarking), and
• Back-testing (comparing actual to
predicted outcomes).
IRB institutions are expected to
employ all of the components of this
process. However, the data to perform
comprehensive back-testing will not be
available in the early stages of
implementing an IRB rating system.
Therefore, banks will have to rely more
heavily on developmental evidence,
quality control tests, and benchmarking
to assure themselves and other
interested parties that their rating
systems are likely to be accurate. Since
the time delay before rating systems can
be back-tested is likely to be an
important issue—because of the rarity of
defaults in most years and the bunching
of defaults in a few years—the other
parts of the validation process will
assume greater importance. If rating
processes are developed in a learning
environment in which banks attempt to
change and improve ratings, back
testing may be delayed even further.
Validation in its early stages will
depend on bank management’s
exercising informed judgment about the
likelihood of the rating system
working—not simply on empirical tests.
Ratings System Developmental
Evidence
The first source of support for the
validity of a bank’s rating system is
developmental evidence. Evaluating
developmental evidence involves
making a reasonable assessment of the
quality of the rating system by analyzing
its design and construction.
Developmental evidence is intended to
answer the question, Could the rating
system be expected to work reasonably
if it is implemented as designed? That
evidence will have to be revisited
whenever the bank makes a change to
its rating system. If a bank adopts a
rating system and does not make
changes, this step will not have to be
revisited. However, since rating systems
are likely to change over time as the
bank learns about the effectiveness of
the system and incorporates the results
of those analyses, the evaluation of
developmental evidence is likely to be
an ongoing part of the process. The
particular steps taken in evaluating
developmental evidence will depend on
the type of rating system.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Generally, the evaluation of
developmental evidence will include a
body of expert opinion. For example,
developmental evidence in support of a
statistical rating model must include
information on the logic that supports
the model and an analysis of the
statistical model-building techniques. In
contrast, developmental evidence in
support of a constrained-judgment
system that features guidance values of
financial ratios might include a
description of the logic and evidence
relating the values of the ratios to past
default and loss outcomes.
Regardless of the type of rating
system, the developmental evidence
will be more persuasive when it
includes empirical evidence on how
well the ratings might have worked in
the past. This evidence should be
available for a statistical model since
such models are chosen to maximize the
fit to outcomes in the development
sample. In addition, statistical models
should be supported by evidence that
they work well outside the development
sample. Use of ‘‘holdout’’ sample
evidence is a good model-building
practice to ensure that the model is not
merely a statistical quirk of the
particular data set used to build the
model.
Empirical developmental evidence of
rating effectiveness will be more
difficult to produce for a judgmental
rating system. Such evidence would
require asking raters how they would
have rated past credits for which they
did not know the outcomes. Those
retrospective ratings could then be
compared to the outcomes to determine
whether the ratings were correct on
average. Conducting such tests,
however, will be difficult because
historical data sets may not include all
of the information that an individual
would have actually used in making a
judgment about a rating.
The sufficiency of the developmental
evidence will itself be a matter of
informed expert opinion. Even if the
rating system is model-based, an
evaluation of developmental evidence
will entail judging the merits of the
model-building technique. Although no
bright line tests are feasible because
expert judgment is essential to the
evaluation of rating system
development, experts will be able to
draw conclusions about whether a wellimplemented system would be likely to
perform satisfactorily.
Ratings System Ongoing Validation
The second source of analytical
support for the validity of a bank rating
system is the ongoing analysis intended
to confirm that the rating system is

PO 00000

Frm 00009

Fmt 4701

Sfmt 4703

45957

being implemented and continues to
perform as intended. Such analysis
involves process verification and
benchmarking.
Process Verification
Verification activities address the
question, Are the ratings being assigned
as intended? Specific verification
activities will depend on the rating
approach. If a model is used for rating,
verification analysis begins by
confirming that the computer code used
to deploy the model is correct. The
computer code can be verified in a
number of established ways. For
example, a qualified expert can
duplicate the code or check the code
line by line. Process verification for a
model will also include confirmation
that the correct data are being used in
the model.
For expert-judgment and constrainedjudgment systems, verification requires
other individual reviewers to evaluate
whether the rater followed rating policy.
The primary requirements for
verification of ratings assigned by
individuals are:
• A transparent rating process,
• A database with information used
by the rater, and
• Documentation of how the
decisions were made.
The specific steps will depend on
how much the process incorporates
specific guidelines and how much the
exercise of judgment is allowed. As the
dependence on specific guidelines
increases, other individuals can more
easily confirm that guidelines were
followed by reference to sufficient
documentation. As the dependence on
judgment rises, the ratings review
function will have to be staffed
increasingly by experts with appropriate
skills and knowledge about the rating
policies of the bank.
Ratings process verification also
includes override monitoring. If
individuals have the ability to override
either models or policies in a
constrained-judgment system, the bank
should have both a policy stating the
tolerance for overrides and a monitoring
system for identifying the occurrence of
overrides. A reporting system capturing
data on reasons for overrides will
facilitate learning about whether
overrides improve accuracy.
Benchmarking
S. Banks must benchmark their
internal ratings against internal, market
and other third-party ratings.
Benchmarking is the set of activities
that uses alternative tools to draw
inferences about the correctness of
ratings before outcomes are actually

E:\FR\FM\04AUN2.SGM

04AUN2

45958

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

known. The most important type of
benchmarking of a rating system is to
ask whether another rater or rating
method attaches the same rating to a
particular obligor or facility. Regardless
of the rating approach, the benchmark
can be either a judgmental or a modelbased rating. Examples of such
benchmarking include:
• Ratings reviewers who completely
re-rate a sample of credits rated by
individuals in a judgmental system.
• An internally developed model is
used to rate credits rated earlier in a
judgmental system.
• Individuals rate a sample of credits
rated by a model.
• Internal ratings are compared
against results from external agencies or
external models.
Because it will take considerable time
before outcomes will be available, using
alternative ratings as benchmarks will
be a very important validation device.
Such benchmarking must be applied to
all rating approaches, and the
benchmark can be either a model or
judgment. At a minimum, banks must
establish a process in which a
representative sample of its internal
ratings is compared to third-party
ratings (e.g., independent internal raters,
external rating agencies, models, or
other market data sources) of the same
credits.
Benchmarking also includes activities
designed to draw broader inferences
about whether the rating system—as
opposed to individual ratings—is
working as expected. The bank can look
for consistency in ranking or
consistency in the values of rating
characteristics for similarly rated
credits. Examples of such benchmarking
activities include:
• Analyzing the characteristics of
obligors that have received common
ratings.
• Monitoring changes in the
distribution of ratings over time.
• Calculating a transition matrix
calculated from changes in ratings in a
bank’s portfolio and comparing it to
historical transition matrices from
internal bank data or publicly available
ratings.
While benchmarking activities allow
for inferences about the correctness of
the ratings system, they are the not same
thing as back-testing. The benchmark
itself is a prediction and may be in
error. If benchmarking evidence
suggests a pattern of rating differences,
it should lead the bank to investigate the
source of the differences. Thus, the
benchmarking process illustrates the
possibility of feedback from ongoing
validation to model development,

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

underscoring the characterization of
validation as a process.

Validation of the quantification process
is covered in section VI.

Back Testing

A. Introduction

S. Banks must develop statistical tests
to back-test their IRB rating systems.
S. Banks must establish internal
tolerance limits for differences between
expected and actual outcomes.
S. Banks must have a policy that
requires remedial actions be taken when
policy tolerances are exceeded.
The third component of a validation
process is back-testing, which is the
comparison of predictions with actual
outcomes. Back-testing of IRB systems is
the empirical test of the accuracy of the
parameter values, PD and LGD,
associated with obligor and loss severity
ratings, respectively. For IRB rating
systems, back-testing addresses the
combined effectiveness of the
assignment of obligor and loss severity
ratings and the calibration of the
parameters PD and LGD attached to
those ratings.
At this time, there is no generally
agreed-upon statistical test of the
accuracy of IRB systems. Banks must
develop statistical tests to back-test their
IRB rating systems. In addition, banks
must have a policy that specifies
internal tolerance limits for comparing
back-testing results. Importantly, that
policy must outline the actions that
would be taken whenever policy limits
are exceeded.
As a combined test of ratings
effectiveness, back-testing is a
conceptual bridge between the ratings
system architecture discussed in this
chapter and the quantification of
parameters, discussed in Chapter 2. The
final section of Chapter 2 discusses
back-testing as one type of quantitative
test required to validate the
quantification of parameter values.

Stages of the Quantification Process

III. Quantification of IRB Systems
Ratings quantification is the process
of assigning numerical values to the four
key components for internal ratingsbased assessments of credit-risk capital:
probability of default (PD), the expected
loss given default (LGD), the expected
exposure at default (EAD), and maturity
(M). Section I establishes an organizing
framework for considering IRB
quantification and develops general
principles that apply to the entire
process. Sections II through IV cover
specific principles or supervisory
standards that apply to PD, LGD, and
EAD respectively. The maturity
component, which is much less
dependent on statistical estimates and
the use of data, receives somewhat
different treatment in section V.

PO 00000

Frm 00010

Fmt 4701

Sfmt 4703

With the exception of maturity, the
risk components are unobservable and
must be estimated. The estimation must
be consistent with sound practice and
supervisory standards. In addition, a
bank must have processes to ensure that
these estimates remain valid.
Calculation of risk components for
IRB involves two sets of data: the bank’s
actual portfolio data, consisting of
current credit exposures assigned to
internal grades, and a ‘‘reference data
set,’’ consisting of a set of defaulted
credits (in the case of LGD and EAD
estimation) or both defaulted and nondefaulted credits (in the case of PD
estimation). The bank estimates a
relationship between the reference data
set and probability of default, loss
severity, or exposure; then this
estimated relationship is applied to the
actual portfolio data for which capital is
being assessed.
Quantification proceeds through four
logical stages: obtaining reference data;
estimating the reference data’s
relationship to the parameters; mapping
the correspondence between the
reference data and the portfolio’s data;
and applying the relationship between
reference data and parameters to the
portfolio’s data. (Readers may find it
helpful to refer to the appendix to this
chapter, which illustrates how this fourstage framework can be applied to
ratings quantification approaches in
practice.) An evaluation of any bank’s
IRB quantification process focuses on
understanding how the bank
implements each stage for each of the
key parameters, and on assessing the
adequacy of the bank’s approach.
Data—First, the bank constructs a
reference data set, or source of data,
from which parameters can be
estimated.
Reference data sets include internal
data, external data, and pooled internal/
external data. Important considerations
include the comparability of the
reference data to the current credit
portfolio, whether the sample period
‘‘appropriately’’ includes periods of
stress, and the definition of default used
in the reference data. The reference data
must be described using a set of
observed characteristics; consequently,
the data set must contain variables that
can be used for this characterization.
Relevant characteristics might include
external debt ratings, financial
measures, geographic regions, or any
other factors that are believed to be

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
related in some way to PD, LGD, or
EAD. More than one reference data set
may be used.
Estimation—Second, the bank applies
statistical techniques to the reference
data to determine a relationship
between characteristics of the reference
data and the parameters (PD, LGD, or
EAD).
The result of this step is a model that
ties descriptive characteristics of the
obligor or facility in the reference data
set to PD, LGD, or EAD estimates. In this
context, the term ‘models’ is used in the
most general sense; a model may be
simple, such as the calculation of
averages, or more complicated, such as
an approach based on advanced
regression techniques. This step may
include adjustments for differences
between the IRB definition of default
and the default definition in the
reference data set, or adjustments for
data limitations. More than one
estimation technique may be used to
generate estimates of the risk
components, especially if there are
multiple sets of reference data or
multiple sample periods.
Mapping—Third, the bank creates a
link between its portfolio data and the
reference data based on common
characteristics.
Variables or characteristics that are
available for the current portfolio must
be mapped to the variables used in the
default, loss-severity, or exposure
model. (In some cases, the bank
constructs the link for a representative
exposure in each internal grade, and the
mapping is then applied to all credits
within a grade.) An important element
of mapping is making adjustments for
differences between reference data sets
and the bank’s portfolio. The bank must
create a mapping for each reference data
set and for each combination of
variables used in any estimation model.
Application—Fourth, the bank
applies the relationship estimated for
the reference data to the actual portfolio
data.
The ultimate aim of quantification is
to attribute a PD, LGD, or EAD to each
exposure within the portfolio, or to each
internal grade if the mapping was done
at the grade level. This step may include
adjustments to default frequencies or
loss rates to ‘‘smooth’’ the final
parameter estimates. If the estimates are
applied to individual transactions, the
bank must in some way aggregate the
estimates at the grade level. In addition,
if multiple data sets or estimation
methods are used, the bank must adopt
a means of combining the various
estimates.
A number of examples are given in
this chapter to aid exposition and

VerDate jul<14>2003

20:32 Aug 01, 2003

Jkt 200001

interpretation. None of the examples is
sufficiently detailed to incorporate all
the considerations discussed in this
chapter. Moreover, technical progress in
the area of quantification is rapid. Thus,
banks should not interpret an example
that is consistent with the standard
being discussed, and that resembles the
bank’s current practice, as creation of a
‘‘safe harbor’’ or as an indication that
the bank’s practice will be approved asis. Banks should consider this guidance
in its entirety when determining
whether systems and practices are
adequate.
General Principles for Sound IRB
Quantification
Several core principles apply to all
elements of the overall ratings
quantification process; those general
principles are discussed in this
introductory section. Each of these
principles is, in effect, a supervisory
standard for IRB systems. Other
supervisory standards, specific to
particular elements or parameters, are
discussed in the relevant sections.
Supervisory evaluation of IRB
quantification requires consideration of
all of these principles and standards,
both general and specific. Particular
practical approaches to ratings
quantification may be highly consistent
with some standards, and less so with
others. In any particular case, an
ultimate assessment relies on the
judgment of supervisors to weigh the
strengths and weaknesses of a bank’s
chosen approach, using these
supervisory standards as a guide.
S. IRB institutions must have a fully
specified process covering all aspects of
quantification (reference data,
estimation, mapping, and application).
The quantification process, including
the role and scope of expert judgment,
must be fully documented and updated
periodically.
A fully specified quantification
process must describe how all four
stages (data, estimation, mapping, and
application) are implemented for each
parameter. Documentation promotes
consistency and allows third parties to
review and replicate the entire process.
Examples of third parties that might use
the documentation include ratingsystem reviewers, auditors, and bank
supervisors. Periodic updates to the
process must be conducted to ensure
that new data, analytical techniques,
and evolving industry practice are
incorporated into the quantification
process.
S. Parameter estimates and related
documentation must be updated
regularly.

PO 00000

Frm 00011

Fmt 4701

Sfmt 4703

45959

The parameter estimates must be
updated at least annually, and the
process for doing so must be
documented in bank policy. The update
should also evaluate the judgmental
adjustments embedded in the estimates;
new data or techniques may suggest a
need to modify those adjustments.
Particular attention should be given to
new business lines or portfolios in
which the mix of obligors is believed to
have changed substantially. A material
merger, acquisition, divestiture, or exit
clearly raises questions about the
continued applicability of the process
and should trigger an intensive review
and updating.
The updating process is particularly
relevant for the reference data stage
because new data become available all
the time. New data must be
incorporated, into the PD, LGD, and
EAD estimates, using a well-defined
process.
S. A bank must subject all aspects of
the quantification process, including
design and implementation, to an
appropriate degree of independent
review and validation.
An independent review is an
assessment conducted by persons not
accountable for the work being
reviewed. The reviewers may be either
internal or external parties. The review
serves as a check that the quantification
process is sound and works as intended;
it should be broad-based, and must
include all of the elements of the
quantification process that lead to the
ultimate estimates of PD, LGD, and
EAD. The review must cover the full
scope of validation: evaluation of the
integrity of data inputs, analysis of the
internal logic and consistency of the
process, comparison with relevant
benchmarks, and appropriate backtesting based on actual outcomes.
S. Judgmental adjustments may be an
appropriate part of the quantification
process, but must not be biased toward
lower estimates of risk.
Judgment will inevitably play a role
in the quantification process and may
materially affect the estimates.
Judgmental adjustments to estimates are
often necessary because of some
limitations on available reference data
or because of inherent differences
between the reference data and the
bank’s portfolio data. The bank must
ensure that adjustments are not biased
toward optimistically low parameter
estimates for PD, LGD, and EAD.
Individual assumptions are less
important than broad patterns;
consistent signs of judgmental decisions
that lower parameter estimates
materially may be evidence of bias.

E:\FR\FM\04AUN2.SGM

04AUN2

45960

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

The reasoning and empirical support
for any adjustments, as well as the
mechanics of the calculation, must be
documented. The bank should conduct
sensitivity analysis to demonstrate that
the adjustment procedure is not biased
toward reducing capital requirements.
The analysis must consider the impact
of any judgmental adjustments on
estimates and risk weights, and must be
fully documented.
S. Parameter estimates must
incorporate a degree of conservatism
that is appropriate for the overall
robustness of the quantification process.
In estimating values of PD, LGD, and
EAD should be as precise and accurate
as possible. However, estimates of PD,
LGD and EAD are statistics, and thus
inherently subject to uncertainty and
potential error. It is often possible to be
reasonably confident that a risk
component or other parameter lies
within a particular range, but greater
precision is difficult to achieve. Aspects
of the ratings quantification process that
are apt to introduce uncertainty and
potential error include the following:
The estimation of coefficients of
particular variables in a regressionbased statistical default or severity
model.
• The calculation of average default
or loss rates for particular categories of
credits in external default databases.
• The mapping between portfolio
obligors or facilities and reference data
when the set of common characteristics
does not align exactly.
A general principle of the IRB
approach is that a bank must adjust
estimates conservatively in the presence
of uncertainty or potential error. In
many cases this corresponds to
assigning a final parameter estimate that
increases required capital relative to the
best estimate produced through soundpractice estimation techniques. The
extent of this conservative adjustment
should be related to factors such as the
relevance of the reference data, the
quality of the mapping, the precision of
the statistical estimates, and the amount
of judgment used throughout the
process. Margins of conservatism need
not be added at each step; indeed, that
could produce an excessively
conservative result. The overall margin
of conservatism should adequately
account for all uncertainties and
weaknesses; this is the general
interpretation of requirements to
incorporate appropriate degrees of
conservatism. Improvements in the
quantification process (use of better
data, estimation techniques, and so on)
may reduce the appropriate degree of
conservatism over time.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Estimates of PD, LGD, EAD, or other
parameters or coefficients should be
presented with an accompanying sense
of the statistical precision of the
estimates; this facilitates an assessment
of the appropriate degree of
conservatism.
B. Probability of Default (PD)
Data
To estimate PD accurately, a bank
must have a comprehensive reference
data set with observations that are
comparable to the bank’s current
portfolio of obligors. Clearly, the data
set used for estimation should be similar
to the portfolio to which such estimates
will be applied. The same comparability
standard applies to both internal and
external data sets.
To ensure ongoing applicability of the
reference data, a bank must assess the
characteristics of its current obligors
relative to the characteristics of obligors
in the reference data. Such variables
might include qualitative and
quantitative obligor information,
internal and external rating, rating
dates, and line of business or geography.
To this end, a bank must maintain
documentation that fully describes all
explanatory variables in the data set,
including any changes to those variables
over time. A well-defined and
documented process must be in place to
ensure that the reference data are
updated as frequently as is practical, as
fresh data become available or portfolio
changes make necessary.
S. The sample for the reference data
must be at least five years, and must
include periods of economic stress
during which default rates were
relatively high.
To foster more robust estimation,
banks should use longer time series
when more than five years of data are
available. However, the benefits of using
a longer time series (longer than five
years) may have to be weighed against
a possible loss of data comparability.
The older the reference data, the less
similar they are likely to be to the bank’s
current portfolio; striking the correct
balance is a matter of judgment.
Reference obligors must not differ from
the current portfolio obligors
systematically in ways that seem likely
to be related to obligor default risk.
Otherwise, the derived PD estimates
may not be applicable to the current
portfolio.
Note that this principle does not
simply restate the requirement for five
years of data: periods of stress during
which default rates are relatively high
must be included in the data sample.
Exclusion of such periods biases PD

PO 00000

Frm 00012

Fmt 4701

Sfmt 4703

estimates downward and unjustifiably
lowers regulatory capital requirements.
Example. A bank’s reference data set
covers the years 1987 through 2001. Each
year includes identical data elements, and
each year is similarly populated. For its grade
PD estimates, the bank relies upon data from
a sub-sample covering 1992 through 2001.
The bank provides no justification for
dropping the years from 1987 through 1991.
The bank contends that it is not necessary to
include those data, as the reference sample
they use for estimation satisfies the five-year
requirement. This practice is not consistent
with the standard because the bank has not
supported its decision to ignore available
data. The fact that the excluded years include
a recession would raise particular concerns.

S. The definition of default within the
reference data must be reasonably
consistent with the IRB definition of
default.
Regardless of the source of the
reference data, a bank must apply the
same default definition throughout the
quantification processes. This fosters
consistent estimation across parameters
and reduces the potential for undesired
bias. In addition, consistent application
of the same definition across banks will
permit true horizontal analysis by
supervisors and engaged market
participants.
This standard applies to both internal
and external reference data. For internal
data, a bank’s default definition is
expected to be consistent with the IRB
definition going forward. Banks will be
expected to make appropriate
adjustments to their data systems such
that all defaults as defined for IRB are
captured by the time a bank fully
implements its IRB system. For any
historical or external data that do not
fully comply with the IRB definition of
default, a bank must make conservative
adjustments to reflect such
discrepancies. Larger discrepancies
require larger adjustments for
conservatism.
Example. To identify defaults in its
historical data, a bank applies a consistent
definition of ‘‘placed on nonaccrual.’’ This
definition is used in the bank’s quantification
exercises to estimate PD, LGD, and EAD. The
bank recognizes that use of the nonaccrual
definition fails to capture certain defaults as
identified in the IRB rules. Specifically, the
bank indicates that the following kinds of
defaulted facilities would not have been
placed on nonaccrual: (1) Credit obligations
that were sold at a material credit-related
economic loss, and (2) distressed
restructurings. To be consistent with the
standard, the bank must make a wellsupported adjustment to its grade PD
estimates to reflect the difference in the
default definitions.

Estimation
Estimation of PD is the process by
which characteristics of the reference

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
data are related to default frequencies.4
The relevant characteristics that help to
determine the likelihood of default are
referred to as ‘‘drivers of default’’.
Drivers might include variables such as
financial ratios, management expertise,
industry, and geography.
S. Estimates of default rates must be
empirically based and must represent a
long-run average.
Estimates must capture average
default experience over a reasonable
mix of high-default and low-default
years of the economic cycle. The
average is labeled ‘‘long-run’’ because a
long observation period would span
both peaks and valleys of the economic
cycle. The emphasis should not be on
time-span; the long-run average concept
captures the breadth, not the length, of
experience.
If the reference data are characterized
by internal or external rating grades, one
estimation approach is to calculate the
mean of one-year realized default rates
for each grade, giving equal weight to
each year’s realized default rate. PD
estimates generally should be calculated
in this manner.
Another approach is to pool obligors
in a given grade over a number of years
and then calculate the mean default
rate. In this case, each year’s default rate
is weighted by the number of obligors.
This approach may underestimate
default rates. For example, if lending
declines in recessions so that obligors
are fewer in those years than in others,
weighting by number of obligors would
dilute the effect of the recession year on
the overall mean. The obligor-weighted
calculation, or another approach, will be
allowed only if the bank can
demonstrate that this approach provides
a better estimate of the long-run average
PD. At a minimum, this would involve
comparing the results of both methods.
Statistical default prediction models
may also play a role in PD estimation.
For example, the characteristics of the
reference data might include financial
ratios or a distance-to-default measure,
as defined by a specific implementation
of a Merton-style structural model.
For a model-based approach to meet
the requirement that ultimate grade PD
estimates be long-run averages, the
reference data used in the default model
must meet the long-run requirement.
4 The New Basel Capital Accord produced by the
Basel Committee on Banking Supervision discusses
three techniques for PD estimation. IRB banks are
not constrained to select from among these three
techniques; they have broad flexibility to
implement appropriate approaches to
quantification. The three Basel techniques are best
regarded not as a complete taxonomy of the
possible approaches to PD estimation, but rather as
illustrations of a few of the many possible
approaches.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

For example, a model can be used to
relate financial ratios to likelihood of
default based on the outcome for the
firms—default or non-default. Such a
model must be calibrated to capture the
default experience over a reasonable
mix of good and bad years of the
economic cycle. The same requirement
would hold for a structural model;
distance to default must be calibrated to
default frequency using long-run
experience. This applies to both internal
and vendor models, and a bank must
verify that this requirement is met.
Example 1. A bank uses external data from
a rating agency to estimate PD. The PD
estimate for each agency grade is calculated
as the mean of yearly realized default rates
over a time period (1980 through 2001) that
includes several recessions and high-default
years. The bank provides support that this
time period adequately represents long-run
experience. This illustrates an estimation
method that is consistent with the standard.
Example 2a. Like the institution in
example 1, a bank maps internal ratings to
agency grades. The estimates for the agency
grades are set indirectly, using the default
probabilities from a default prediction model.
The bank does so because although it links
internal and agency grades, the bank views
the default model’s results as more predictive
than the historical agency default experience.
For each agency grade, the bank calculates a
PD estimate as the mean of the model-based
default probabilities for the agency-rated
obligors. In order to meet the long-run
requirement, the bank calculates the
estimates over the seven years from 1995
through 2001. The bank demonstrates that
this time period includes a reasonable mix of
high-default and low-default experience.
This estimation method is consistent with
the standard.
Example 2b. In a variant of example 2a, a
bank uses the mean default frequency per
agency rating grade for a single year, such as
2001. Empirical evidence shows that the
mean default frequency for agency grades
varies substantially from year to year. A
single year thus does not reflect the full range
of experience, because a long-run average
should be relatively stable year to year. Such
instability makes this estimation method
unacceptable.
Example 2c. Another bank calculates the
agency grade PD estimates as the median
default probability of companies in that
grade. The bank does so without
demonstrating that the median is a better
statistical estimator than the mean. This
estimation method is not consistent with the
standard. A median gives less weight to
obligors with high estimated default
probabilities than a simple mean does. The
difference between mean and median can be
material because distributions of credits
within grades often are substantially skewed
toward higher default probabilities: the
riskier obligors within a grade tend to have
individual default probabilities that are
substantially worse than the median, while
the least risky have default probabilities only
somewhat better than the median.

PO 00000

Frm 00013

Fmt 4701

Sfmt 4703

45961

S. Judgmental adjustments may play
an appropriate role in PD estimation,
but must not be biased toward lower
estimates.
The following examples illustrate
how supervisors will evaluate
adjustments:
Example 1. A bank uses the last five years
of internal default history to estimate grade
PDs. However, they recognize that the
internal experience does not include any
high-default years. In order to remedy this
and still take advantage of its experience, the
bank uses external agency data to adjust the
estimates upward. Using the agency data, the
bank calculates the ratio between the longrun average and the mean default rate per
grade over the last five years. The bank
assumes that the relationship observed in the
agency data applies to its portfolio, and
adjusts the estimates for the internal data
accordingly. This practice is consistent with
the standard.
Example 2. A bank uses internal default
experience to estimate grade PDs. However,
the bank has historically failed to recognize
defaults when the loss on the default
obligation was avoided by seizing collateral.
The bank makes no adjustment for such
missing defaults. The realized default rate
using the more inclusive definition would be
higher than that observed by the bank (and
loss severity rates would be correspondingly
lower). This practice would not be consistent
with the standard, unless the bank
demonstrates that the necessary adjustment
is immaterial.

Mapping
Mapping is the process of establishing
a correspondence between the bank’s
current obligors and the reference
obligor data used in the default model.
Hence, mapping involves identifying
how default-related characteristics of
the current portfolio correspond to the
characteristics of reference obligors.
Such characteristics might include
financial and nonfinancial variables,
and assigned ratings or grades.
Mapping can be thought of as taking
each obligor in the bank’s portfolio and
characterizing it as if it were part of the
reference data. There are two broad
approaches to the mapping process:
Obligor mapping: Each portfolio
obligor is mapped to the reference data
based on its individual characteristics.
For example, if a bank applies a default
model, a default probability will be
generated for each obligor. That
individual default probability is then
used to assign each obligor to a
particular internal grade, based on the
bank’s established criteria. To obtain a
final estimate of the grade PD in the
subsequent application stage, the bank
averages the default probabilities of
individual obligors within each grade.
Grade mapping: Characteristics of the
obligors within an internal grade are

E:\FR\FM\04AUN2.SGM

04AUN2

45962

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

averaged or otherwise summarized to
construct a ‘‘typical’’ or representative
obligor for each grade. Then, the bank
maps that representative obligor to the
reference data. For example, if the bank
uses a default model, the default
probability associated with that typical
obligor will serve as the grade PD in the
application stage. Alternatively, the
bank may map the typical obligor to a
particular external rating grade based on
quantitative and qualitative
characteristics, and assign the long-run
default rate for that rating to the internal
grade in the application stage.
Either grade mapping or obligor
mapping can be part of the
quantification process; either method
can produce a single PD estimate for
each grade in the application stage.
However, in the absence of other
compelling considerations, banks
should use obligor mapping for two
reasons:
• First, default probabilities are
nonlinear under many estimation
approaches. As a result, the default
probability of the typical obligor—the
result of a grade mapping approach—is
often lower than the mean of the
individual obligor default probabilities
from the obligor mapping approach. For
example, consider a bank that maps to
the S&P scale and uses historical S&P
bond default rates. For ease of
illustration, suppose that one internal
grade contains only three obligors that
individually map to BB, BB¥, and B+.
The historical default rates for these
three grades are 1.07, 1.76, and 3.24
percent, respectively (based on 1981–
2001 data). Using obligor mapping,
those rates would be assigned directly to
the three obligors, yielding a mean PD
of 2.02 percent for the grade. Using
grade mapping, the grade PD would be
only 1.76, because the grade’s typical
obligor is rated BB¥.
• Second, a hypothetical obligor with
a grade’s average characteristics may not
represent well the risks presented by the
grade’s typical obligor. For example, a
bank might observe that obligors with
high leverage and low earnings
variability have about the same default
risk as obligors with low leverage and
high earnings variability. These two
types of obligors might both end up in
the same grade, for example, Grade 6. If
so, the typical obligor in Grade 6 would
have moderate leverage and moderate
earnings variability—a combination that
might fail to reflect any of the
individual obligors in Grade 6, and that
could easily result in a PD for the grade
that is too low.
A bank electing to use grade mapping
instead of obligor mapping should be
especially careful in choosing a

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

‘‘typical’’ obligor for each grade. Doing
so typically requires that the bank
examine the actual distribution of
obligors within each grade, as well as
the characteristics of those obligors.
Banks should be aware that different
measures of central tendency (such as
mean, median, or mode) will give
different results, and that these different
results may have a material effect on a
grade’s PD; they must be able to justify
their choice of a measure. Banks must
have a clear and consistent policy
toward the calculation.
S. The mapping must be based on a
robust comparison of available data
elements that are common to the
portfolio and the reference data.
Sound mapping practice uses all
common elements that are available in
the data as the basis for mapping. If a
bank chooses to ignore certain common
variables or to weight some variables
more heavily than others, those choices
must be supported. Mapping should
also take into account differences in
rating philosophy (for example, pointin-time or through-the-cycle) between
any ratings embedded in the reference
data set and the bank’s own rating
regime.
A mapping should be plausible, and
should be consistent with the rating
philosophy established by the bank as
part of its obligor rating policy. For a
bank that uses grade mapping, levels
and ranges of key variables within each
internal grade should be close to values
of similar variables for corresponding
obligors within the reference data.
The standard allows for use of a
limited set of common variables that are
predictive of default risk, in part to
permit flexibility in early years when
data may be far from ideal.
Nevertheless, banks will eventually be
expected to use variables that are widely
recognized as the most reliable
predictors of default risk in mapping
exercises. In the meantime, banks
relying on data elements that are weak
predictors must compensate by making
their estimates more conservative. For
example, leverage and cash flow are
widely recognized to be reliable
predictors of corporate default risk.
Borrower size is also predictive, but less
so. A mapping based solely on size is by
nature less reliable than one based on
leverage, cash flow, and size.
Example 1. In estimating PD, a bank relies
on observed default rates on bonds in various
agency grades for PD quantification. To map
its internal grades to the agency grades, the
bank identifies variables that together explain
much of the rating variation in the bond
sample. The bank then conducts a statistical
analysis of those same variables within its
portfolio of obligors, using a multivariate

PO 00000

Frm 00014

Fmt 4701

Sfmt 4703

distance calculation to assign each portfolio
obligor to the external rating whose
characteristics it matches most closely (for
example, assigning obligors to ratings so that
the sum of squared differences between the
external grade averages and the obligor’s
characteristics is minimized). This practice is
broadly consistent with the standard.
Example 2. A bank uses grade mapping to
link portfolio obligors to the reference data
set described by agency ratings. The bank
looks at publicly rated portfolio obligors
within an internal grade to determine the
most common external rating, does the same
for all grades, and creates a correspondence
between internal and external ratings. The
strength of the correspondence is a function
of the number of externally rated obligors
within each grade, the distribution of those
external ratings within each grade and the
similarity of externally rated obligors in the
grade to those not externally rated. This
practice is broadly consistent with this
standard, but would require a comparison of
rating philosophies and may require
adjustments and the addition of margins of
conservatism.

S. A mapping process must be
established for each reference data set
and for each estimation model.
Banks should never assume that a
mapping is self-evident. Even a rating
system that has been explicitly designed
to replicate external agency ratings may
or may not be effective in producing a
replica; formal mapping is still
necessary. Indeed, in such a system the
kind of analysis involved in mapping
may help identify inconsistencies in the
rating process itself.
A mapping process is needed even
where the reference obligors come from
internal historical experience. Banks
must not assume that internal data do
not require mapping, because changes
in bank strategy or external economic
forces may alter the composition of
internal grades or the nature of the
obligors in those grades over time.
Mappings must be reaffirmed regardless
of whether rating criteria or other
aspects of the ratings system have
undergone explicit changes during the
period covered by the reference data set.
Banks often use multiple reference
data sets, and then combine the
resulting estimates to get a grade PD. A
bank that does that must conduct a
rigorous mapping process for each data
set.
Supervisors expect all meaningful
characteristics of obligors to be factored
directly into the rating process; this
should include characteristics like the
obligor’s industry or physical location.
But in some circumstances, certain
effects related to industry, geography, or
other factors are not reflected in rating
assignments or default estimates. In
such cases, it may be appropriate for
banks to capture the impact of the

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
omissions by using different mappings
for different business lines or types of
obligors. Supervisors expect this
practice to be transitional; banks will
eventually be required to incorporate
the omitted effects into the rating
system and the estimation process as
they are uncovered and documented,
rather than adjusting the mapping.
Example 1. The bank maps its internal
grades carefully to one rating agency, and
then assumes a correspondence to another
agency’s scale despite known differences in
the rating methods of the two agencies. The
bank then applies a mean of the grade default
rates from these two public debt-rating
agencies to its internal grades. This practice
is not consistent with the standard, because
the bank should map to each agency’s scale
separately.
Example 2. A bank uses internal historical
data as its reference data. The bank computes
a mean default rate for each grade as the
grade PD for capital purposes, and asserts
that mapping is unnecessary because ‘‘its
strong credit culture ensures that a 4 is
always a 4.’’ This practice is not consistent
with the standard, because no mapping has
been done; there is no assurance that a
representative obligor in a grade today is
comparable to an obligor in that same grade
in the past.

S. The mapping must be updated and
independently validated regularly.
The appropriate mapping between a
bank’s portfolio and the reference data
may change over time. For example,
relationships between internal grades
and external agency grades may change
during the economic cycle because of
differences in rating philosophy.
Similarly, distance-to-default measures
for obligors in a given grade may not be
constant over time. These likely changes
make it imperative that the bank update
all mappings regularly.
Sound validation practices may
include tests for internal consistency
such as ‘‘reverse mapping.’’ Using this
technique, a bank evaluates obligors
from the reference data set as if they
were subject to the bank’s rating system
(that is, part of the bank’s current
portfolio). The bank’s mapping is then
applied to these reverse-mapped
obligors to see whether the mapped
characterization of the reference obligor
is consistent with that of the initial
evaluation.5 Another valuable technique
is to apply different mapping methods
and compare the results. For example,
mappings based on financial ratio
comparisons can be rechecked using
5 For example, suppose a bank asserts that its
Grade 3 corresponds to an S&P rating of A.
Applying reverse mapping, the bank would take a
sample of A-rated obligors from the reference data,
run them through the bank’s rating process (perhaps
a simplified version), and check to see that those
obligors usually receive a grade of 3 on the bank’s
internal scale.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

mappings based on available external
ratings.
Example. A bank mapped its internal
grades to the rating scale of one public debtrating agency in 1992. Since then, the bank
has completed a major acquisition of another
large bank and significantly changed its
business mix in other ways. The bank
continues to use the same mapping, without
reassessing its validity. This practice is not
consistent with the standard.

Application
In the application stage, the bank
applies the PD estimation method to the
current portfolio of obligors using the
mapping process. It obtains final PD
estimates for each rating grade, which
will be used to calculate minimum
regulatory capital. To arrive at those
estimates, a bank may adjust the raw
results derived from the estimation
stage. For example, it might aggregate
individual obligor default probabilities
to the rating grade level, or smooth
results because a rating grade’s PD
estimate was higher than a lower quality
grade. The bank must explain and
support all adjustments when
documenting its quantification process.
Example. A bank uses external data to
estimate long-run average PDs for each grade.
The resulting PD estimate for Grade 2 is
slightly higher than the estimate for Grade 3,
even though Grade 2 is supposedly of higher
credit quality. The bank uses statistics to
demonstrate that this anomaly occurred
because defaults are rare in the highest
quality rating grades. The bank judgmentally
adjusts the PD estimates for grades 2 and 3
to preserve the expected relationship
between obligor grade and PD, but requires
that total risk-weighted assets across both
grades using the adjusted PD estimates be no
less than total risk-weighted assets based on
the unadjusted estimates, using a typical
distribution of obligors across the two grades.
Such an adjustment during the application
stage is consistent with this guidance.

S. IRB institutions that aggregate the
default probabilities of individual
portfolio obligors when calculating PD
estimates for internal grades must have
a clear policy governing the aggregation
process.
As noted above, mapping may be
grade-based or obligor-based. Gradebased mappings naturally provide a
single PD per grade, because the
estimated default model is applied to
the representative obligor for each
grade. In contrast, obligor-based
mappings must aggregate in some
manner the individual PD estimates to
the grade level. The expectation is that
the grade PD estimate will be calculated
as the mean. The bank will be allowed
to calculate this estimate differently
only if it can demonstrate that the
alternative method provides a better

PO 00000

Frm 00015

Fmt 4701

Sfmt 4703

45963

estimate of the long-run average PD. To
obtain this evidence, the bank must at
least compare the results of both
methods.
S. IRB institutions that combine
estimates from multiple sets of reference
data must have a clear policy governing
the combination process, and must
examine the sensitivity of the results to
alternative combinations.
Because a bank should make use of as
much information as possible when
mapping, it will usually use multiple
data sets. The manner in which the data
or the estimates from those multiple
data sets are combined is extremely
important. A bank must document its
justification for the particular
combination methods selected. Those
methods must be subject to appropriate
approval and oversight.
The data may come from the same
basic data source but from different time
periods or from different data sources
altogether. For example, banks often
combine internal data with external
data, use external data from different
sample periods, or combine results from
corporate-bond default databases with
results from equity-based models of
obligor default. Different combinations
will produce different PD estimates. The
bank should investigate alternative
combinations and document the impact
on the estimates. When ultimate results
are highly sensitive to how estimates
from different data sources are
combined, the bank must choose among
the alternatives conservatively.
C. Loss Given Default (LGD)
The LGD estimation process is similar
to the PD estimation process. The bank
identifies a reference data set of
defaulted credits and relevant
descriptive characteristics. Once the
bank obtains these data sets (with the
facility characteristics), it must select a
technique to estimate the economic loss
per dollar of exposure at default, for a
defaulted exposure with a given array of
characteristics. The bank’s portfolio
must then be mapped, so that the model
can be applied to generate an estimate
of LGD for each portfolio transaction or
severity grade.
Data
Unlike reference data sets used for PD
estimation, data sets for severity
estimation contain only exposures to
defaulting obligors. At least two broad
categories of data are necessary to
produce LGD estimates.
First, data must be available to
calculate the actual economic loss
experienced for each defaulted facility.
Such data may include the market value
of the facility at default, which can be

E:\FR\FM\04AUN2.SGM

04AUN2

45964

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

used to proxy a recovery rate.
Alternatively, economic loss may be
calculated using the exposure at the
time of default, loss of principal,
interest, and fees, the present value of
subsequent recoveries and related
expenses (or the costs as calculated
using an approved allocation method),
and the appropriate discount rate.
Second, factors must be available to
group the defaulted facilities in
meaningful ways. Characteristics that
are likely to be important in predicting
loss rates include whether or not the
facility is secured and the type and
coverage of collateral if the facility is
secured, seniority of the claim, general
economic conditions, and obligor’s
industry. Although these factors have
been found to be significant in existing
academic and industry studies, a bank’s
quantification of LGD certainly need not
be limited to these variables. For
example, a bank might expand its loss
severity research by examining many
other potential drivers of severity
(characteristics of an obligor that might
help the bank predict the severity of a
loss), including obligor size, line of
business, geographic location, facility
type, obligor ratings (internal or
external), historical internal severity
grade, or tenor of the relationship.
A bank must ensure that the reference
data remains applicable to its current
portfolio of facilities. It must implement
established processes to ensure that
reference data sets are updated when
new data become available. All data
sources, variables, and the overall
processes concerning data collection
and maintenance must be fully
documented, and that documentation
should be readily available for review.
S. The sample period for the reference
data must be at least seven years, and
must include periods of economic stress
during which defaults were relatively
high.
Seven years is the minimum sample
period for the LGD reference data. A
longer sample period is desirable,
because more default observations will
be available for analysis and may serve
to refine severity estimates. In any case,
a bank must select a sample period that
includes episodes of economic stress,
which are defined as periods with a
relatively high number of defaults.
Inclusion of stress periods increases the
size and potentially the breadth of the
reference data set. According to some
empirical studies, the average loss rate
is higher during periods of stress.
Example. A bank intends to rely primarily
on internal data when quantifying all
parameter estimates, including LGD. Its
internal data cover the period 1994 through
2000. The bank will continue to extend its

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

data set as time progresses. Its current policy
mandates that credits be resolved within two
years of default, and the data set contains the
most recent data available. Although the
current data set satisfies the seven-year
requirement, the bank is aware that it does
not include stress periods. In comparing its
loss estimates with rates published in
external studies for similarly stratified data,
the bank observes that its estimates are
systematically lower. To be consistent with
the standard, the bank must take steps to
include stress periods in its estimates.

S. The definition of default within the
reference data must be reasonably
consistent with the IRB definition of
default.
This standard parallels a similar
standard in the section on PD. The
following examples illustrate how it
applies in the case of LGD.
Example 1. For LGD estimation, a bank
includes in its default data base only
defaulted facilities that actually experience a
loss, and excludes credits for which no loss
was recorded because liquidated collateral
covered the loss (effectively applying a ‘‘loss
given loss’’ concept). This practice is not
consistent with the standard because the
bank’s default definition for LGD is narrower
than the IRB definition.
Example 2. A bank relies on external data
sources to estimate LGD because it lacks
sufficient internal data. One source uses
‘‘bankruptcy filing’’ to indicate default while
another uses ‘‘missed principal or interest
payment,’’ and the two sources result in
significantly different loss estimates for the
severity grades defined by the bank. The
bank’s practice is not consistent with the
standard, and the bank should determine
whether the definitions used in the reference
data sets differ substantially from the IRB
definition. If so, and the differences are
difficult to quantify, the bank should seek
other sources of reference data. For more
minor differences, the bank may be able to
make appropriate adjustments during the
estimation stage.

Estimation
Estimation of LGD is the process by
which characteristics of the reference
data are related to loss severity. The
relevant characteristics that help
explain how severe losses tend to be
upon default might include variables
such as seniority, collateral, facility
type, or business line.
S. The estimates of loss severity must
be empirically based and must reflect
the concept of ‘‘economic loss.’’
Loss severity is defined as economic
loss, which is different from accounting
measures of loss. Economic loss
captures the value of recoveries and
direct and indirect costs discounted to
the time of default, and it should be
measured for each defaulted facility.
The scope of the cash flows included in
recoveries and costs is meant to be
broad. Workout costs that can be clearly

PO 00000

Frm 00016

Fmt 4701

Sfmt 4703

attributed to certain facilities or types of
facilities must be reflected in the bank’s
LGD assignments for those exposures.
When such allocation is not practical,
the bank may assign those costs using
factors based on broad averages.
A bank must establish a discount rate
that reflects the time value of money
and the opportunity cost of funds to
apply to recoveries and costs. The
discount rate must be no less than the
contract interest rate on new
originations of a type similar to the
transaction in question, for the lowestquality grade in which a bank originates
such transactions.6 Where possible, the
rate should reflect the fixed rate on
newly originated exposures with term
corresponding to the average resolution
period of defaulting assets.
Ideally, severity should be measured
once all recoveries and costs have been
realized. However, a bank may not
resolve a defaulted obligation for many
years following default. For practical
purposes, banks may choose to close the
period of observation before this final
resolution occurs—that is, at a point in
time when most costs have been
incurred and when recoveries are
substantially complete. Banks that do so
should estimate the additional costs and
recoveries that would likely occur
beyond this period and include them in
the LGD estimates. A bank must
document its choice of the period of
observation, and how it estimated
additional costs and recoveries beyond
this period.
LGD for each type of exposure must
be the loss per default (expressed as a
percentage of exposure at default)
expected during periods when default
rates are relatively high. This expected
loss rate is referred to as ‘‘stresscondition LGD.’’ For cases in which loss
severities do not have a material degree
of cyclical variability, use of the longrun default-weighted average is
appropriate, although stress-condition
LGD generally exceeds this average.
The drivers of severity can be linked
to loss estimates in a number of ways.
One approach is to segment the
reference defaults into groups that do
not overlap. For example, defaults could
be grouped by business line,
predominant collateral type, and loanto-value coverage. The LGD estimate for
each category is the mean loss
calculated over the category’s defaulted
facilities. Loss must be calculated as the
default-weighted average (where
individual defaults receive equal
weight) rather than the average of
6 The appropriate discount rate for IRB purposes
may differ from the contract rate required under
FAS 114 for accounting purposes.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
annual loss rates, and must be based on
results from periods during which
default rates were relatively numerous if
loss rates are materially cyclical.
Banks can also draw estimates of LGD
from a statistical model. For example,
they can build a regression model of
severity using data on loss severity and
some quantitative measures of the loss
drivers. Any model must meet the
requirements for model validation
discussed in Chapter 1. Other methods
for computing LGD could also be
appropriate.
Example 1. A bank has internal data on
defaulted facilities, including information on
business line, facility type, seniority, and
predominant collateral type (if the facility is
secured). The data allow for a reasonable
calculation of economic loss. The data span
eight years and include three years that can
be termed high-default years. After analyzing
the economic cycle using internal and
external data, the bank concludes that the
data show no evidence of material cyclical
variability in loss severities, and that the
default data span enough experience to allow
estimation of a long-run average. On the basis
of preliminary analysis, the bank determines
that the drivers of loss severity for large
corporate facilities are similar to those for
middle-market loans, and that the two groups
can be estimated as a pool. Again on the basis
of preliminary analysis, the bank segments
this pool by seniority and by six collateral
groupings, including unsecured. These
groupings contain enough defaults to allow
reasonably precise estimates. The loss
severity estimates are then calculated by
averaging loss rates within each segment.
This practice is consistent with the standard.
Example 2. A bank uses internal data in
which information on security and seniority
is lacking. The bank groups corporate and
middle-market defaulted facilities into a
single pool and calculates the LGD estimate
as the mean loss rate. No adjustments for the
lack of data are made in the estimation or
application steps. This practice is
unacceptable because there is ample external
evidence that security and seniority matter in
these segments. A bank with such limited
internal default data must incorporate
external or pooled data into the estimation.
Example 3. A bank determines that a
business unit—for example, a unit dedicated
to a particular type of asset-based lending—
forms a homogeneous pool for the purposes
of estimating loss severity. That is, although
the facilities in this pool may differ in some
respects, the bank determines that they share
a similar loss experience in default. The bank
must provide reasonable support for this
pooling through analysis of lending practices
and available internal and external data. In
this example, the mean of a single segment
is consistent with the standard.

S. Judgmental adjustments may play
an appropriate role in LGD estimation,
but must not be biased toward lower
estimates.
It is difficult to make general
statements about good and bad practices

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

in this area, because adjustments can
take many different forms. The
following examples illustrate how
supervisors would be likely to evaluate
particular adjustments observed in
practice.
Example 1. A bank divides observed
defaults into segments according to collateral
type. One of the segments has too few
observations to produce a reliable estimate.
Relying on external data and judgment, the
bank determines that the segment’s estimated
severity of loss falls somewhere between the
estimates for two other categories. This
segment’s severity is set judgmentally to be
the mean of the estimates for the other
segments. This practice is consistent with the
standard.
Example 2. A bank does not know when
recoveries (and related costs) occurred in a
portfolio segment; therefore, it cannot
properly discount the segment’s cash flows.
However, the bank has sufficient internal
data to calculate economic loss for defaulted
facilities in another portfolio segment. The
bank can support the assumption that the
timing of cash flows for the two segments is
comparable. Using the available data and
informed judgment, the bank estimates that
the measured loss without discounting
should be grossed up to account for the time
value of money and the opportunity cost of
funds. This practice is consistent with the
standard.
Example 3. A bank segments internal
defaults in a business unit by some factors,
including collateral. Although the available
internal and external evidence indicates a
higher LGD, the bank judgmentally assigns a
loss estimate of 2 percent for facilities
secured by cash collateral. The basis for this
adjustment is that the lower estimate is
justified by the expectation that the bank
would do a better job of following policies for
monitoring cash collateral in the future. Such
an adjustment is generally not appropriate
because it is based on projections of future
performance rather than realized experience.
This practice is not consistent with the
standard.

Mapping
LGD mapping follows the same
general principles that PD mapping
does. A mapping must be plausible and
must be based on a comparison of
severity-related data elements common
to both the reference data and the
current portfolio. The mapping
approach is expected to be unbiased,
such that the exercise of judgment does
not consistently lower LGD estimates.
The default definitions in the reference
data and the current portfolio of obligors
should be comparable. The mapping
process must be updated regularly, welldocumented, and independently
reviewed.
S. A bank must conduct a robust
comparison of available common
elements in the reference data and the
portfolio.
Mapping involves matching facilityspecific data elements available in the

PO 00000

Frm 00017

Fmt 4701

Sfmt 4703

45965

current portfolio to the factors in the
reference data set used to estimate
expected loss severity rates. Examples of
factors that influence loss rates include
collateral type and coverage, seniority,
industry, and location.
At least three kinds of mapping
challenges may arise. First, even if
similarly named variables are available
in the reference data and portfolio data,
they may not be directly comparable.
For example, the definition of particular
collateral types, or the meaning of
‘‘secured,’’ may vary from one
application to another. Hence, a bank
must ensure that linked variables are
truly similar. Although adjustments to
enhance comparability can be
appropriate, they must be rigorously
developed and documented. Second,
levels of aggregation may vary. For
example, the reference data may only
broadly identify collateral types, such as
financial and nonfinancial. The bank’s
information systems for its portfolio
might supply more detail, with a wide
variety of collateral type identifiers. To
apply the estimates derived from the
reference data, the internal data must be
regrouped to match the coarser level of
aggregation in the reference data. Third,
reference data often do not include
workout costs and will often use
different discounting. Judgmental
adjustments for such problems must be
well-documented and, as much as
possible, empirically based.
S. A mapping process must be
established for each reference data set
and for each estimation model.
Mapping is never self-evident. Even
when reference data are drawn from
internal default experience, a bank must
still link the characteristics of the
reference data with those of the current
portfolio.
Different data sets and different
approaches to severity estimation may
be entirely appropriate, especially for
different business segments or product
lines. Each mapping process must be
specified and documented.
Application
At the application stage, banks apply
the LGD estimation framework to their
current portfolio of credit exposures.
Doing so might require them to
aggregate individual LGD estimates into
broader averages (for example, into
discrete severity grades) or to combine
estimates in various ways.
The inherent variability of recovery,
due in part to unanticipated
circumstances, demonstrates that no
facility type is wholly risk-free,
regardless of structure, collateral type,
or collateral coverage. The existence of

E:\FR\FM\04AUN2.SGM

04AUN2

45966

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

recovery risk dictates that application of
a zero percent LGD is not acceptable.
S. IRB institutions that aggregate LGD
estimates for severity grades from
individual exposures within those
grades must have a clear policy
governing the aggregation process.
Banks with discrete severity grades
compute a single estimate of LGD for a
representative exposure within each of
those grades. If a bank with a discrete
scale of severity grades maps those
grades to the reference data using grade
mapping, there will be a single estimate
of LGD for each grade, and the bank
does not need to aggregate further.
However, if the bank maps at the
individual transaction level, the bank
may then choose to aggregate those
individual LGD estimates to the grade
level and use the grade LGD in capital
calculations. Because different methods
of aggregation are possible, a bank must
have a clear policy regarding how
aggregation should be accomplished; in
general, simple averaging is preferred.
(This standard is irrelevant for banks
that choose to assign LGD estimates
directly to individual exposures rather
than grades, because aggregation is not
required in that case.)
S. An IRB institution must have a
policy describing how it combines
multiple sets of reference data.
Multiple data sets may produce
superior estimates of loss severity, if the
results are appropriately combined.
Combining such sets differently usually
produces different estimates of LGD. As
a matter of internal policy, a bank
should investigate alternative
combinations, and document the impact
on the estimates. If the results are highly
sensitive to the manner in which
different data sources are combined, the
bank must choose conservatively among
the alternatives.
D. Exposure at Default (EAD)
Compared with PD and LGD
quantification, EAD quantification is
less advanced. As such, it is addressed
in somewhat less detail in this guidance
than are PD and LGD quantification.
Banks should continue to innovate in
the area EAD estimation, refining and
improving practices in EAD
measurement and prediction.
Additional supervisory guidance will be
provided as more data become available
and estimation techniques evolve.
A bank must provide an estimate of
expected EAD for each facility in its
portfolio. EAD is defined as the bank’s
expected gross dollar exposure of the
facility upon the obligor’s default. For
fixed exposures like term loans, EAD is
equal to the current amount
outstanding. For variable exposures

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

such as loan commitments or lines of
credit, exposure is equal to current
outstandings plus an estimate of
additional drawings up to the time of
default. This additional drawdown,
identified as loan equivalent exposure
(LEQ) in many institutions, is typically
expressed as a percentage of the current
total committed but undrawn amount.
EAD can thus be represented as:
EAD = current outstanding + LEQ ×
(total committed¥current
outstanding)
As it is the LEQ that must be estimated,
LEQ is the focus of this guidance.
Even though EAD estimation is less
sophisticated than PD and LGD
estimation, a bank still develops EAD
estimates by working through the four
stages that produce the other types of
quantification: The bank must use a
reference data set; it must apply an
estimation technique to produce an
expected total dollar exposure at default
for a facility with a given array of
characteristics; it must map its current
portfolio to the reference data; and, by
applying the estimation model, it must
generate an EAD estimate for each
portfolio facility or facility-type, as the
case may be.
Data
Like reference data sets used for LGD
estimation, LEQ data sets contain only
exposures to defaulting obligors. In
many cases, the same reference data
may be used for both LGD and LEQ. In
addition to relevant descriptive
characteristics (referred to as ‘‘drivers’’)
that can be used in estimation, the
reference data must include historical
information on the exposure (both
drawn and undrawn amounts) as of
some date prior to default, as well as the
drawn exposure at the date of default.
As discussed below under
‘‘Estimation,’’ LEQ estimates may be
developed using either a cohort method
or a fixed-horizon method. The bank’s
reference data set must be structured so
that it is consistent with the estimation
method the bank applies. Thus, the data
must include information on the total
commitment, the undrawn amount, and
the exposure drivers for each defaulted
facility, either at fixed calendar dates for
the cohort method or at a fixed interval
prior to the default date for the fixedhorizon method.
The reference data must contain
variables that enable the bank to group
the exposures to defaulted obligors in
meaningful ways. Obligor and facility
risk ratings are commonly believed to be
significant characteristics for predicting
additional drawdown. Since less
empirical research has been done on

PO 00000

Frm 00018

Fmt 4701

Sfmt 4703

EAD estimation, little is known about
other potential drivers of EAD. Among
the many possibilities, banks may
consider time from origination, time to
expiration or renewal, economic
conditions, risk rating changes, or
certain types of covenants. Some
potential drivers may be linked to a
bank’s credit risk management skills,
while others may be exogenous.
Industry practice is likely to improve as
banks extend their research to identify
other meaningful drivers of EAD.
A bank must ensure continued
applicability of the reference data to its
current portfolio of facilities. The
reference data must include the types of
variable exposures found in a bank’s
current portfolio. The definitions of
default and exposure in the reference
data should be consistent with the IRB
definition of default, and consistent
with the definitions used for PD and
LGD quantification. Established
processes must be in place to ensure
that reference data sets are updated
when new data are available. All data
sources, variables, and the overall
processes governing data collection and
maintenance must be fully documented,
and that documentation should be
readily available for review.
Seven years of data are required for
EAD (or LEQ) estimation. The sample
should include periods during which
default rates were relatively high, and
ideally cover a complete economic
cycle.
Estimation
To derive LEQ estimates,
characteristics of the reference data are
related to additional drawings preceding
a default event. The estimation process
must be capable of producing a
plausible estimate of LEQ to support the
EAD calculation for each facility. Two
broad types of estimation methods are
used in practice, the cohort method and
the fixed-horizon method.
Under the cohort method, a bank
groups defaults into discrete calendar
periods (such as a year or a quarter). The
bank then estimates the relationship
between the drivers as of the start of that
calendar period, and EAD or LEQ for
each exposure to a defaulter. For each
exposure category (that is, for each
combination of exposure drivers
identified by the bank), the LEQ
estimate is calculated as the mean
additional drawing for facilities in that
category. To combine results for
multiple periods into a single long-run
average, the period-by-period means
should be weighted by the proportion of
defaults occurring in each period.
Under the fixed-horizon method, for
each exposure to a defaulted obligor the

E:\FR\FM\04AUN2.SGM

04AUN2

45967

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Application
In the application stage, the estimated
relationship between drivers and LEQ is
applied to the bank’s actual portfolio.
To ensure that estimated EAD is at least
as large as the currently drawn amount
for all exposures, LEQs must not be
negative. Multiple reference data sets
may be used for LEQ estimation and
combined at the application stage; those
combinations should be rigorously
developed, approved, and documented.
Any smoothing or use of expert
judgment to adjust the results should be
well-justified and clearly documented.
This includes any adjustment for
definitions of default that do not meet
the supervisory standards. The less
robust the process, the more
conservative the result should be.
Some facility types may be treated as
exceptions, and assigned an LEQ that
does not vary with characteristics such
as line of business or risk rating. Such
exceptional treatment should be clearly
justified, and the justification should be
fully documented.
EAD may be particularly sensitive to
changes in the way banks manage
individual credits. For example, a
change in policy regarding covenants
may have a significant impact on LEQ.
When such changes take place, the bank
should consider them when making its
estimates—and it should do so from a
conservative point of view. Policy
changes likely to significantly increase
LEQ should prompt immediate
increases in LEQ estimates. If a bank’s
policy changes seem likely to reduce
LEQ, estimates should be reduced only
after the bank accumulates a significant
amount of actual experience under the
new policy to support the reductions.
E. Maturity (M)
A bank must assign a value of
effective remaining maturity (M) to each
credit exposure in its portfolio. In
general, M is the weighted-average
number of years to receipt of the cash
flows the bank expects under the
contractual terms of the exposure,
where the weights are equal to the
fraction of the total undiscounted cash
flow to be received at each date.
Mathematically, M is given by:

PO 00000

M = ∑ t × wt
t

Frm 00019

Fmt 4701

Sfmt 4703

where wt is the fraction of the total cash
flow received at time t, that is:

wt = Ct / ∑ Ct
t

Ct is the undiscounted cash flow
received at time t, with t measured in
years from the date of the calculation of
M.
Effective maturity, sometimes referred
to as ‘‘average life,’’ need not be a whole
number, and often is not. For example,
if 33 percent of the cash flow is
expected at the end of one year (t=1)
and the other 67 percent two years from
today (t=2), then M is calculated as:
M = (1×0.33) + (2×0.67) = 1.67
for an effective maturity of 1.67 years.
This value of M would be used in the
IRB capital calculation.
The relevant cash flows are the future
payments the bank expects to receive
from the obligor, regardless of form;
they may include payments of interest
or fees, principal repayments, or other
types of payments depending on the
structure of the transaction. For
exposures whose cash flow schedule is
virtually predetermined unless the
obligors defaults (fixed-rate loans, for
example), the calculation of the
weighted-average remaining maturity is
straightforward, using the scheduled
timing and amounts of the individual
undiscounted cash flows. These cash
flows should be the contractually
expected payments; the bank should not
take into account the possibility of
delayed or reduced cash flows due to
potential future default.
Cash flows associated with other
types of credit exposures may be
somewhat less certain. In such cases,
the bank must establish a method of
projecting expected cash flows. In
general, the method used for any
exposure should be the same as the one
used by the bank for purposes of
valuation or risk management. The
method must be well-documented and
subject to independent review and
approval. A bank must demonstrate that
the method used is standard industry
practice, that it is widely used within
the bank for purposes other than
regulatory capital calculations, or both.
To be conservative, a bank may set M
equal to the maximum number of years
the obligor could take to fully discharge
the contractual obligation (provided that
the maximum is not longer than five
years, as noted below). In many cases,
this maximum will correspond to the
stated or nominal maturity of the
instrument. Banks must make this
conservative choice (maximum nominal
maturity) if the timing and amounts of

E:\FR\FM\04AUN2.SGM

04AUN2

EN04AU03.009</GPH>

Mapping
If the same variables that drive
exposure in the reference data are also
available for facilities in the portfolio,
mapping may be relatively easy.
However, the bank must still review the
definitions to ensure that variables that
seem to be the same actually are. If the
relevant variables are not available in a
bank’s current portfolio information
system, the bank will encounter the
same mapping complexities that it does
when mapping for PD and LGD in
similar circumstances. A bank should
have well-documented policies that
govern the mapping. Any exceptions to
mapping policy should be reviewed,

justified and fully documented.
Mapping may be done for each exposure
or for broad categories of exposure; the
latter would be analogous to the ‘‘grade
mapping’’ discussed earlier in this
chapter.

EN04AU03.008</GPH>

bank compares additional drawdowns
to the total commitment but undrawn
amount that existed at the start of a
fixed interval prior to the date of the
default (the horizon). For example, the
bank might base its estimates on a
reference data set that supplies the
actual exposure at default along with
the drawn and undrawn amounts (as
well as relevant drivers) at a date a fixed
number of months prior to the date of
each default, regardless of the actual
calendar date on which the default
occurred. Estimates of LEQ are
computed from the average drawdowns
that occur over the fixed-horizon
interval, for whatever combinations of
the driving variables the bank has
determined are relevant for explaining
and predicting exposure at default.
Evidence may indicate that LEQ
estimates are positively correlated with
economic downturns; that is, it may be
that LEQs increase during high-default
periods. If so, the higher drawdowns
that occur during high-default periods
are denoted ‘‘stress-condition LEQs,’’
analogous to the ‘‘stress-condition
LGDs’’ discussed earlier in this chapter.
For any exposure type whose LEQ
estimates exhibit material cyclicality, a
bank must use the stress-condition LEQ
for purposes of calculating EAD.
In general, all available data should be
used; particular observations or time
periods should not be excluded from the
data sample. Any adjustments a bank
makes to the estimation results should
be justified and fully documented. The
analysis should be refreshed
periodically as new data become
available, and a bank should have a
process in place to ensure that advances
in analytical techniques and industry
practice are considered as they emerge
and are incorporated as appropriate.
LEQ estimates should be updated at
least annually. Detailed documentation,
ongoing validation, and adequate
oversight are fundamental controls that
support a sound estimation process.

45968

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

the cash flows on the exposure cannot
be projected with a reasonable degree of
confidence.
Certain over-the-counter derivatives
contracts and repurchase transactions
may be subject to master netting
agreements. In such cases, the bank may
compute a single value of M for the
transactions as a group by weighting
each individual transaction’s effective
maturity by that transaction’s share of
the total notional value subject to the
netting agreement, and summing the
result across all of the transactions.
For IRB capital calculations, the value
of M for any exposure is subject to
certain upper and lower limits,
regardless of the actual effective
maturity of the exposure. In all cases,
the value of M should be no greater than
5 years. If an exposure clearly has an
effective maturity that exceeds this
upper limit, the bank may simply use a
value of M=5 rather than calculating the
actual effective maturity.
For most exposures, the value of M
must be no less than one year. For
certain short-term exposures (repo-style
transactions, money market
transactions, trade finance-related
transactions, and exposures arising from
payment and settlement processes) that
are not part of a bank’s ongoing
financing of a borrower and that have an
original maturity of less than three
months, M may be set as low as one day.
For over-the-counter derivative and
repurchase-style transactions subject to
a master netting agreement, weighted
average maturity must be set at no less
than five days.
F. Validation
Values of PD, LGD, and EAD are
estimates with implications for credit
risk and the future performance of a
bank’s credit portfolio under IRB; in
essence, they are forecasts. ‘‘Validation’’
of these estimates describes the full
range of activities used to assess their
quality as forecasts of default rates, loss
severity rates, and exposures at default.
Chapter 1 discusses validation of IRB
systems in general; this section focuses
specifically on ratings quantification,
which includes the assignment of PD to
obligor grades and the assignment of
LGD, EAD, and M to exposures.
S. A validation process must cover all
aspects of IRB quantification.
Banks must have a process for
validating IRB quantification; their
policies must state who is accountable
for validation, and describe the actions
that will proceed from the different
possible results. Validation should focus
on the three estimated IRB parameters
(PD, LGD, and EAD). Although the
established validation process should

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

result in an overall assessment of IRB
quantification for each parameter, it also
must cover each of the four stages of the
quantification process as described in
preceding sections of this chapter (data,
estimation, mapping, and application).
The validation process must be fully
documented, and must be approved by
appropriate levels of the bank’s senior
management. The process must be
updated periodically to incorporate new
developments in validation practices
and to ensure that validation methods
remain appropriate; documentation
must be updated whenever validation
methods change.
Banks should use a variety of
validation approaches or tools; no single
validation tool can completely and
conclusively assess IRB quantification.
Three broad types of tools that are
useful in this regard are evaluation of
the conceptual soundness of the
approach to quantification (evaluation
of logic), comparison to other sources of
data or estimates (benchmarking), and
comparisons of actual outcomes to
predictions (back-testing). Each of these
types of tools has a role to play in
validation, although the role varies
across the four stages of quantification.
Evaluation of logic is essential in
validating all stages of the quantification
process. The quantification process
requires banks to adopt methods, choose
variables, and make adjustments; each
of these actions requires an exercise of
judgment. Validation should ensure that
these judgments are plausible and
informed.
A bank should also validate estimates
by comparing them with relevant
external sources, a process broadly
described as benchmarking. ‘‘External’’
in this context refers to anything other
than the specific reference data,
estimation approach, or mapping under
consideration. Reference data can be
compared with other data sources;
choices of variables can be compared
with similar choices made by others;
estimation results can be compared with
the results of alternative estimation
methods using the same reference data.
Other data sources may show that
default and severity rates across the
economy or the banking system are high
or low relative to other periods, or may
reveal unusual effects in parts of the
quality spectrum.
Effective validation must compare
actual results with predictions. Such
comparisons, often referred to as ‘‘backtesting,’’ are valuable comprehensive
tests of the rating system and its
quantification. However, they are only
one element of the broader validation
regime, and should not be a bank’s only
method of validation. Because they test

PO 00000

Frm 00020

Fmt 4701

Sfmt 4703

the results of the rating system as a
whole, they are unlikely to identify
specific reasons for any divergence
between expectations and realizations.
Rather they will indicate only that
further investigation is necessary.
By applying back-testing to the
reference data set as it is updated with
new data, a bank can improve the
estimation process. To further improve
the process, a bank must regularly
compare realized default rates, loss
severities, and exposure-at-default
experience from its portfolio with the
PD, LGD, and EAD estimates on which
capital calculations are based.
Realizations should be compared with
expected ranges based on the estimates.
These expected ranges should take into
account the bank’s rating philosophy
(the relative weight given to current and
stress conditions in assigning ratings).
Depending on that philosophy, year-byyear realized default rates and loss
severities may be expected to differ
significantly from the long-run average.
If a bank adjusts final estimates to be
conservative, it should likely do its
back-testing on the unadjusted
estimates.
A bank’s quantitative testing methods
and other validation techniques should
be robust to economic cycles. A sound
validation process should take business
cycles into account, and any
adjustments for stages of the cycle
should be clearly specified in advance
and fully documented as part of the
validation policy. The fact that a year
has been ‘‘unusual’’ should not be taken
as a reason to abandon the bank’s
standard validation practices.
S. A bank must comprehensively
validate parameter estimates at least
annually, must document the results,
and must report these results to senior
management.
A full and comprehensive annual
validation is a minimum for effective
risk management under IRB. More
frequent validation may be appropriate
for certain parts of the IRB system and
in certain circumstances; for example,
during high-default periods, banks
should compute realized default and
loss severity rates more frequently,
perhaps quarterly. They must document
the results of validation, and must
report them to appropriate levels of
senior risk management.
S. The validation policy must outline
appropriate remedial responses to the
results of parameter validation.
The goal of validation should be to
continually improve the rating process
and its quantification. To this end, the
bank should establish thresholds or
accuracy tolerances for validation
results. Results that breach thresholds

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
should bring an appropriate response;
that response should depend on the
results and should not necessarily be to
adjust the parameter estimates. When
realized default, severity, or exposures
rates diverge from expected ranges,
those divergences may point to issues in
the estimation or mapping elements of
quantification. They may also indicate
potential problems in other parts of the
ratings assignment process. The bank’s
validation policy must describe (at least
in broad terms) the types of responses
that should be considered when
relevant action thresholds are crossed.
Appendix to Part III: Illustrations of the
Quantification Process
This appendix provides examples to show
how the logical framework described in this
guidance, with its four stages (data,
estimation, mapping, and application),
applies when analyzing typical current bank
practices. The framework is broadly
applicable—for PD or LGD or EAD; using
internal, external, or pooled reference data;
for simple or complex estimation methods—
although the issues and concerns that arise
at each stage depend on a bank’s approach.
These examples are intended only to
illustrate the logic of the four-stage IRB
quantification framework, and should not be
taken to endorse the particular techniques
presented in the examples. In fact, certain
aspects of the examples are not consistent
with the standards outlined in this guidance.
Example 1: PD Estimation From Bond Data
• A bank establishes a correspondence
between its internal grades and external
rating agency grades; the bank has
determined that its Grade 4 is equivalent to
3⁄4 BB and 1⁄4 B on the Standard and Poor’s
scale.
• The bank regularly obtains published
estimates of mean default frequencies for
publicly rated BB and B obligors in North
America from 1970 through 2002.
• The BB and B historical default
frequencies are weighted 75/25, and the
result is a preliminary PD for the bank’s
internal Grade 4 credits.
• However, the bank then increases the PD
by 10 percent to account for the fact that the
S&P definition of default is more lenient than
the IRB definition.
• The bank makes a further adjustment to
ensure that the resulting grade PD is greater
than the PD attributed to Grade 3 and less
than the PD attributed to Grade 5.
• The result is the final PD estimate for
Grade 4.
Process Analysis for Example 1
Data—The reference data set consists of
issuers of publicly rated debt in North
America over the period 1970 through 2002.
The data description is very basic: each
issuer in the reference data is described only
by its rating (such as AAA, AA, A, BBB, and
so on).
Estimation—The bank could have
estimated default rates itself using a database
purchased from Standard and Poor’s, but

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

since these estimates would just be the mean
default rates per year for each grade, the bank
could just as well (and in this example does)
use the published historical default rates
from S&P; in essence, the estimation step has
been outsourced to S&P. The 10 percent
adjustment of PD is part of the estimation
process in this case because the adjustment
was made prior to the application of the
agency default rates to the internal portfolio
data.
Mapping—The bank’s mapping is an
example of a grade mapping; internal Grade
4 is linked to the 75/25 mix of BB and B.
Based on the limited information presented
in the example, this step should be explored
further. Specifically, how did the bank
determine the 75/25 mix?
Application—Although the application
step is relatively straightforward in this case,
the bank does make the adjustment of the
Grade 4 PD estimate to give it the desired
relationship to the adjacent grades. This
adjustment is part of the application stage
because it is made after the adjusted agency
default rates are applied to the internal
grades.
Example 2: PD Estimation Using a MertonType Equity-Based Model
• A bank obtains a 20-year database of
North American firms with publicly traded
equity, some of which defaulted during the
20-year period.
• The bank uses the Merton approach to
modeling equity in these firms as a
contingent claim, constructing an estimate of
each firm’s distance-to-default at the start of
each year in the database. The bank then
ranks the firm-years within the database by
distance-to-default, divides the ordered
observations into 20 equal groups or buckets,
and computes a mean historical one-year
default frequency for each bucket. That
default frequency is taken as an estimate of
the applicable PD for any obligor within the
range of distance-to-default values
represented by each of the 20 buckets.
• The bank next looks at all obligors with
publicly traded shares within each of its
internal grades, applies the same Merton-type
model to compute distance-to-default at
quarter-end, sorts these observations into the
20 buckets from the previous step, and
assigns the corresponding PD estimate.
• For each internal grade, the bank
computes the mean of the individual obligor
default probabilities and uses that average as
the grade PD.
Process Analysis for Example 2
Data—The reference data set consists of the
North American firms with publicly traded
equity in the acquired database. The
reference data are described in this case by
a single variable, specifically an identifier of
the specific distance-to-default range from
the Merton model (one of the 20 possible in
this case) into which a firm falls in any year.
Estimation—The estimation step is simple:
the average default rate is calculated for each
distance-to-default bucket. Since the data
cover 20 years and a wide range of economic
conditions, the resulting estimates satisfy the
long-run average requirement.
Mapping—The bank maps selected
portfolio obligors to the reference data set

PO 00000

Frm 00021

Fmt 4701

Sfmt 4703

45969

using the distance-to-default generated by the
Merton model. However, not all obligors can
be mapped, since not all have traded equity.
This introduces an element of uncertainty
into the mapping that requires additional
analysis by the bank: were the mapped
obligors representative of other obligors in
the same grade? The bank would need to
demonstrate comparability between the
publicly traded portfolio obligors and those
not publicly traded. It may be appropriate for
the bank to make conservative adjustments to
its ultimate PD estimates to compensate for
the uncertainty in the mapping. The bank
also would need further analysis to
demonstrate that the implied distance-todefault for each internal grade represented
long-run expectations for obligors assigned to
that grade; this could involve computing the
Merton model for portfolio obligors over
several years of relevant history that span a
wide range of credit conditions.
Application—The final step is aggregation
of individual obligors to the grade level
through calculation of the mean for each
grade, and application of this grade PD to all
obligors in the grade. The bank might also
choose to modify PD assignments further at
this stage, combining PD estimates derived
from other sources, applying adjustments for
cyclicality, introducing an appropriate degree
of conservatism, or making other
adjustments.
Example 3: LGD Estimation From Internal
Default Data
• For each loan in its portfolio, a bank
records collateral coverage as a percentage, as
well as which of four types of collateral
applies.
• A bank has retained data on all defaulted
loans since 1995. For each defaulted loan in
the database, the bank has a record of the
collateral type within the same four broad
categories. However, collateral coverage is
only recorded at three levels (low, moderate,
or high, depending on the ratio of collateral
to exposure at default).
• The bank also records the timing and
discounted value of recoveries net of workout
costs for each defaulted loan in the database.
Cash flows are tracked from the date of
default to a ‘‘resolution date,’’ defined as the
point at which the remaining balance is less
than 5 percent of the exposure at the time of
default. A recovery percentage is computed,
equal to the value of recoveries discounted to
the date of default, divided by the exposure
at default.
• For each cell (each of the 12
combinations of collateral type and
coverage), the bank computes a simple mean
LGD percentage as the mean of one minus the
recovery percentage. One of the categories
has a mean LGD of less than zero (recoveries
have exceeded exposure on average), so the
bank sets the LGD at zero to be conservative.
• The bank assigns an estimate of expected
LGD to each loan in the current portfolio by
using collateral information to slot it into one
of the 12 cells. The bank then applies the
mean historical LGD for that cell and adjusts
the result upward by 10 percent to
compensate for the fact that the loss data
come from a period believed to be unusually
good economic performance.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

Process Analysis for Example 3
Data—The reference data is the collection
of historical defaults with the loss amounts
from the bank’s historical portfolio. The
reference data are described by the two
categorical variables (levels of collateral
coverage and types of collateral). It would be
important to determine whether the defaults
over the past few years are comparable to
defaults from the current portfolio. One
would also want to ask why the bank ignores
potentially valuable information by
converting the continuous data on collateral
coverage into a trimodal categorical variable.
Estimation—Conceptually, the bank is
using a ‘‘loss severity model’’ in which 12
binary variables, one for each loan coverage/
type combination, explain the percentage
loss. The coefficients on the variables are just
the mean loss figures from the reference data.
Mapping—Mapping in this case is fairly
straightforward, since all of the relevant
characteristics of the reference data are also
in the loan system for the current portfolio.
However, the bank should determine
whether the variables are being recorded in
the same way (for example, the same
definitions of collateral types), otherwise
some adjustment might be needed.
Application—The bank is able to apply the
loss model by simply plugging in the relevant
values for the current portfolio (or what
amounts to the same thing, looking up the
cell mean). The bank’s assignment of zero
LGD for one of the cells merits special
attention; while the bank represented this
assignment as conservative, the adjustment
does not satisfy the supervisory requirement
that LGD must exceed zero. A larger upward
adjustment is necessary. Finally, the upward
adjustment of the LGD numbers to account

Data elements must be recorded at
origination and whenever the rating is
reviewed, regardless of whether the
rating is actually changed. Data
elements associated with current and
past ratings must be retained and
include the following:
• Key borrower and facility
characteristics,
• Ratings for obligor and loss severity
grades,

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

for the benign environment in which the
reference data were generated presents one
additional wrinkle. The bank must provide a
well-documented, empirically based analysis
of why a 10 percent upward adjustment is
sufficient.

IV. Data Maintenance
A. Overview
Institutions using the IRB approach
for regulatory capital purposes will need
advanced data management practices to
produce credible and reliable risk
estimates. The guiding principle
governing an IRB data maintenance
system is that it must support the
requirements for the quantification,
validation, control and oversight
mechanisms described in this guidance,
as well as the institution’s broader risk
management and reporting needs. The
precise data elements to be collected
will be dictated by the features and
methodology of the IRB system
employed by the institution. The
necessary data elements will therefore
vary by institution and even among
business lines within an institution.
Institutions will have latitude in
managing their data, subject to the
following key data maintenance
standards:
Life Cycle Tracking—institutions
must collect, maintain, and analyze
essential data for obligors and facilities
throughout the life and disposition of
the credit exposure.

• Key factors used to assign the
ratings,
• Person or model responsible for
assigning the rating,
• Date rating assigned, and
• Overrides to the rating and
authorizing individual.
At disposition, data elements must
include:

PO 00000

Frm 00022

Fmt 4701

Sfmt 4703

Rating Assignment Data—institutions
must capture all significant quantitative
and qualitative factors used to assign the
obligor and loss severity ratings.
Support of IRB System—data
collected by institutions must be of
sufficient depth, scope, and reliability
to:
• Validate IRB system processes,
• Validate parameters,
• Refine the IRB system,
• Develop internal parameter
estimates,
• Apply improvements historically,
• Calculate capital ratios,
• Produce internal and public reports,
and
• Support risk management.
This chapter covers the requirements
for maintaining internal data. Reference
data sets used for estimating IRB
parameters are discussed in Chapter 2.
B. Data Maintenance Framework
Life Cycle Tracking
S. Institutions must collect, maintain,
and analyze essential data for obligors
and facilities throughout the life and
disposition of the credit exposure.
Using a life cycle or ‘‘cradle to grave’’
concept for each obligor and facility
supports front-end validation, backtesting, system refinements and risk
parameter estimates. A depiction of lifecycle tracking follows:

• Nature of disposition: renewal,
repayment, loan sale, default,
restructuring,
• For defaults: exposure, actual
recoveries, source of recoveries, costs of
workouts and timing,
• Guarantor support,
• Sale price for loans sold, and
• Other key elements that the bank
deems necessary.

E:\FR\FM\04AUN2.SGM

04AUN2

EN04AU03.001</GPH>

45970

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
Rating Assignment Data
S. Institutions must capture all
significant quantitative and qualitative
factors used to assign the obligor and
loss severity rating.
Assigning a rating to an obligor
requires the systematic collection of
various borrower characteristics as these
factors are critical to validating the
rating system. Obligors are rated using
various methods, as discussed in
Chapter 1. Each of these methods
presents different challenges for input
collection. For example, in judgmental
rating systems, the factors used in the
ratings decision have not traditionally
been explicitly recorded. For purposes
of an IRB approach, institutions that use
expert and constrained judgment must
record these factors and deliver them to
the data warehouse.
For loss severity estimates,
institutions must record the basic
structural characteristics of facilities
and the factors used in developing the
facility rating or LGD estimate. These
often include the seniority of the credit,
the amount and type of collateral, the
most recent collateral valuation date
and its fair value.
Institutions must also track any
overrides of the obligor or loss severity
rating. Tracking overrides separately
allows risk managers to identify
whether the outcome of such overrides
suggests either problems with rating
criteria, or an improper level of
discretion in adjusting the ratings.
Example Data Elements
For illustrative purposes, the
following section provides examples of
the kinds of data elements institutions
will collect under an IRB data
maintenance framework.
General descriptive obligor and facility
data
The data below could be contained
within a loan record or derived from
various sources within the data
warehouse. Guarantor data requirements
are the same as for the obligor.
Obligor/Guarantor Data
• General data: name, address,
industry
• ID number (unique for all related
parent/sub relationships)
• Rating, date, and rater
• PD percentage corresponding to
rating
General Facility Characteristics
• Facility amounts: committed,
outstanding
• Facility type: Term, revolver, bullet,
amortizing, etc.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

• Purpose: acquisition, expansion,
liquidity, inventory, working capital
• Covenants
• Facility ID number
• Origination and maturity dates
• Last renewal date
• Obligor ID link
• Rating, date and rater
• LGD dollar amount or percentage
• EAD dollar amount or percentage
Rating Assignment Data
The data below provide an example of
the categories and types of data that
institutions must retain in order to
continually validate and improve rating
systems. These data items should tie
directly to the documented criteria that
the institution employs in assigning
ratings, both qualitative and
quantitative. For example, rating criteria
often include ranges of leverage or cash
flow for a particular obligor rating. In
addition, qualitative factors, such as
management effectiveness can be
recorded in numeric form. For example,
a 1 may equate to exceptionally strong
management, and a 5 to very weak. The
rating data elements collected should be
complete enough so that others can
review the relevant factors driving the
rating decisions.
Quantitative Factors in Obligor Ratings
• Asset and sale size
• Key ratios used within rating
criteria:
—profitability,
—cash flow,
—leverage,
—liquidity, and
—other relevant factors.

45971

• Geography
Rating Notations
• Flag for overrides or exceptions
• Authorized individual for changing
rating
Final Disposition Data
Only recently have institutions begun
to collect more complete data about a
loan’s disposition. Many institutions
maintain subsidiary systems for their
problem credits with details recorded, at
times manually, on systems that were
not linked with the institution’s central
loan or risk management systems. The
unlinked data are a significant
hindrance in developing reliable PD,
LGD, and EAD estimates.
In advanced systems, the ‘‘grave’’
portion of obligor and exposure tracking
is an essential component for producing
and validating risk estimates and is an
important feedback mechanism for
adjusting and improving risk estimates
over time. Essential data elements are
outlined below.
Obligor/Guarantor
• Default date
• Circumstances of default (for
example, nonaccrual, bankruptcy
chapters 7–11, nonpayment)
Facility
• Outstandings at default
• Amounts undrawn and outstanding
plus time series prior to and through
default

• Quality of earnings and cash flow
• Management effectiveness,
reliability
• Strategic direction, industry
outlook, position
• Country factors and political risk
• Other relevant factors

Disposition
• Amounts recovered and dates
(including source: cash, collateral,
guarantor, etc.)
• Collection cost and dates
• Discount factors to determine
economic cost of collection
• Final disposition (for example,
restructuring or sale)
• Sales price, if applicable
• Accounting items (charge-offs to
date, purchased discounts)

External Factors in Obligor Ratings

C. Data Element Functions

Qualitative Factors in Obligor Ratings

• Public debt rating and trend
• External credit model score and
trend
Rating Notations
• Flag for overrides or exceptions
• Authorized individual for changing
rating
Key Facility Factors in LGD Ratings
• Seniority
• Collateral type: (cash, marketable
securities, AR, stock, RE, etc.)
• Collateral value and valuation date
• Advance rates, LTV
• Industry

PO 00000

Frm 00023

Fmt 4701

Sfmt 4703

S. Data elements must be of sufficient
depth, scope, and reliability to:
• Validate IRB system processes,
• Validate parameters,
• Refine the IRB system,
• Develop internal parameter
estimates,
• Apply improvements historically,
• Calculate capital ratios,
• Produce internal and public reports,
and
• Support risk management.
Validation and Refinement
The data elements collected by
institutions must be capable of meeting

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

the validation requirements described in
Chapters 1 and 2. These requirements
include validating the institution’s IRB
system processes, including the ‘‘front
end’’ aspects such as assigning ratings
so that any issues can be identified
early. The data must support efforts to
identify whether raters and models are
following rating criteria and policies
and whether ratings are consistent
across portfolios. In addition, data must
support the validation of parameters,
particularly the comparison of realized
outcomes with estimates. Thorough data
on default and disposition
characteristics are of paramount
importance for parameter back-testing.
A rich source of data for validation
efforts provides insights on the
performance of the IRB system, and
contributes to a learning environment in
which refinements can be made to the
system. These potential refinements
include enhancements to rating
assignment controls, processes, criteria

or model coefficients, rating system
architecture and parameter estimates.

For loss severity estimates,
institutions must record the basic
structural characteristics of facilities
and the factors used in developing the
facility rating or LGD estimate. These
often include the seniority of the credit,
the amount and type of collateral, the
most recent collateral valuation date
and its fair value.

To maintain a consistent series of
information for credit risk monitoring
and validation purposes, institutions
need to be able to apply historically
improvements they make to their rating
systems. In the example below, a bank
experiences unexpected and rapid
migrations and defaults in its grade 4
category during 2006. Analysis of the
actual financial condition of borrowers
that defaulted compared with those that
did not suggests the debt-to-EBITDA
range for its expert judgment criteria of
3.0 to 5.5 is too broad. Research
indicates that grade 4 should be
redefined to include only borrowers
with debt-to-EBITDA ratios of 3.0–4.5
and grade 5 as 4.5–6.5. In 2007, the
change is initiated, but prior years’
numbers are not recast (see Exhibit A).
Consequently, a break in the series
prevents the bank from evaluating credit
quality changes over several years and
from identifying whether applying the
new rating criteria historically provides
reasonable results.

Recognizing the need to provide
senior managers and board members
with a consistent risk trend, the new
criteria are applied historically to
obligors in grades 4 and 5 as reflected
in Exhibit B. The original ratings
assigned to the grades are maintained
along with notations describing what

the grade would be under the new rating
criteria. If the precise weight an expert
has given one of the redefined criteria
is unknown, institutions are expected to
make estimates on a best efforts basis.
After the retroactive reallocation
process, the bank observes that the mix
of obligors in grade 5 declined

somewhat over the past several years
while the mix in grade 4 increased
slightly. This contrasts with the trend
identified before the retroactive
reallocation. The result is that the
multiyear transition statistics for grades
4 and 5 provide risk managers a clearer
picture of risk.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Developing Parameter Estimates
As detailed in Chapter 2, institutions
will be developing their PD, LGD, and
EAD parameter estimates using
reference data sets comprised of
internal, pooled, and external data.
Institutions are expected to work toward
eventually using as much of their own
experience as possible in their reference
data sets.
Applying Rating System Improvements
Historically

PO 00000

Frm 00024

Fmt 4701

Sfmt 4703

E:\FR\FM\04AUN2.SGM

04AUN2

EN04AU03.007</GPH>

45972

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

Calculating Capital Ratios and Reporting
to the Public
Data retained by the bank will be
essential for regulatory risk-based
capital calculations and public reporting
under the Pillar 3 disclosures. These
uses underscore the need for a welldefined data maintenance framework
and strong controls over data integrity.
Control processes and data elements
themselves should also be subject to

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

periodic verification and testing by
internal and external auditors.
Supervisors will rely on these processes
and also perform testing as
circumstances warrant.
Supporting Risk Management
The information that can be gleaned
from more extensive data collection will
support a broad range of risk
management activities. Risk
management functions will rely on
accurate and timely data to track credit
quality, make informed portfolio risk
mitigation decisions, and perform
portfolio stress tests. Trends developed
from obligor and facility risk rating data
will be used to support internal capital
allocation models, pricing models,
ALLL calculations, and performance
management measures, among others.
Summaries of these are included in
reports to institutions’ boards of
directors, regulators, and in public
disclosures.
D. Managing Data Quality and Integrity
Because data are collected at so many
different stages involving a variety of
groups and individuals, there are
numerous challenges to ensuring the
quality of the data. For example:
• Data will be retained over long
timeframes,
• Qualitative risk-rating variables will
have subjective elements and will be
open to interpretation, and
• Exposures will be acquired through
mergers and purchases, but without an
adequate and easily retrievable
institutional rating history.

PO 00000

Frm 00025

Fmt 4701

Sfmt 4703

Documentation and Definitions
S. Institutions must document the
process for delivering, retaining and
updating inputs to the data warehouse
and ensuring data integrity.
Given the many challenges presented
by data for an IRB system, the
management of data must be formalized.
Fully documenting how the institution’s
flow of data is managed provides a
means for evaluating whether the data
maintenance framework is functioning
as intended. Moreover, institutions must
be able to communicate to individuals
developing or delivering various data
the precise definition of the items
intended to be collected. Consequently,
a ‘‘data dictionary’’ is necessary to
ensure consistent inputs from
individuals and data vendors and to
allow third parties (such as the rating
system review function, auditors, or
bank supervisors) to evaluate data
quality and integrity.
S. Institutions must develop
comprehensive definitions for the data
elements used within each credit group
or business line (a ‘‘data dictionary’’).
Electronic Storage
S. Institutions must store data in
electronic format to allow timely
retrieval for analysis, validation of risk
rating systems, and required
disclosures.
To meet the significant data
management challenges presented by
the validation and control features of an
IRB system, institutions will need to
store their data electronically.
Institutions will have a variety of
storage techniques and potentially a
variety of systems to create their data

E:\FR\FM\04AUN2.SGM

04AUN2

EN04AU03.002</GPH>

This example is based on applying
ratings historically using data already
collected by the bank. However, for
some rating system refinements,
institutions may identify in the future
drivers of default or loss that might not
have been collected for borrowers or
facilities in the past. That is why
institutions are encouraged to collect
data that they believe may serve as a
stronger predictor of default in the
future. For example, certain elements of
a borrower’s cash flow might currently
be suspected to overstate actual
operational health for a particular
industry. In the future, should an
institution decide to deduct this item
from cash flow with a resulting
downgrade of many obligor ratings, the
institution that collected these data
could apply this rating change for prior
years. This would provide the benefit of
providing a consistent picture of risk
over time and also present opportunities
to validate the new criteria using
historical data. Recognizing that
institutions will not be able to anticipate
fully the data they might find useful in
the future, institutions are expected to
reallocate grades on a best efforts basis
when practical.

45973

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

warehouses. IRB data requirements can
be achieved by melding together
existing accounting, servicing,
processing, workout and risk
management systems, provided the
linkages among these systems are well
documented and include sufficient edit
and integrity checks to ensure the data
can be used reliably.
Institutions without electronic
databases would need to resort to
manual reviews of paper files for
ongoing back-testing and ad hoc
‘‘forensic’’ data mining and would be
unable to perform that work in the
timely and comprehensive manner
required of IRB systems. Forensic
mining of paper files to build an initial
data warehouse from the institution’s
credit history is encouraged. In some
instances, paper research may be
necessary to identify data elements or
factors not originally considered
significant in estimating the risk of a
particular class of obligor or facility.
Data Gaps
Rating histories are often lost or are
irretrievable for loans acquired through
mergers, acquisitions, or portfolio
purchases. Institutions are encouraged
wherever practical to collect any
missing historical rating assignment
driver data and to re-grade the acquired
obligors and facilities for prior periods.
In cases where retrieving historical data
is not practical, institutions may attempt
to create a rating history through a
careful mapping of the legacy system
and the new rating structure. Mapped

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

ratings should be reviewed thoroughly
for accuracy. The level of effort placed
on filling data gaps should be
commensurate with the size of the new
exposures to be newly incorporated into
the institution’s IRB system.
V. Control and Oversight Mechanisms
A. Overview
Banks’ internal rating systems are the
foundation for credit-risk management
practices and play an important role in
pricing, reserving, portfolio
management, performance
measurement, economic capital
modeling, and long-term capital
planning. Banks adopting the IRB
approach will also use their credit-risk
ratings to determine regulatory capital
levels. The pivotal and varied uses of
such risk ratings put enormous,
sometimes conflicting, pressure on
banks’ internal rating systems. The
consequences of inaccurate ratings and
their associated estimates are
significant, particularly as they affect
minimum regulatory capital
requirements.
As risk ratings and their related
parameters become better integrated in
institutions’ decision making,
conflicting incentives arise that, if not
well managed, can lead to overly
optimistic or biased ratings. For
example, sales and marketing staff
(relationship managers or RMs) are
typically compensated according to the
volume of business they generate. That
may predispose the RMs to assign more

PO 00000

Frm 00026

Fmt 4701

Sfmt 4725

favorable ratings in order to achieve
rate-of-return and sales objectives. More
favorable ratings may create the
appearance of higher risk-adjusted
returns and business line profitability.
Banks need to be aware of the full range
of incentive conflicts that arise, and
must develop effective controls to keep
these incentive conflicts in check.
Banks will have latitude in designing
and implementing their control
structures subject to the following
principle:
IRB institutions must implement a
system of controls that includes the
following elements: independence,
transparency, accountability, use of
ratings, rating system review, internal
audit, and board and senior
management oversight. While banks
will have flexibility in how these
elements are combined, they must
incorporate sufficient checks and
balances to ensure that the credit risk
management system is functioning
properly.
Banks additionally will want to
embody the following more generic
principles in their control system:
separation of duties, balancing
incentives, and layers of review. Table
4.1 lists the key components of an IRB
control and oversight system. How these
control mechanisms can best be
combined to reinforce one another is a
key challenge for banks implementing
IRB systems:
Table 4.1 Control and Oversight
Mechanisms

E:\FR\FM\04AUN2.SGM

04AUN2

EN04AU03.003</GPH>

45974

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
As the following examples indicate,
how a bank conducts its business will
influence how it designs its control
structure. A bank using an expertjudgment system will likely establish a
different set of controls than a bank
using mainly models. Recognizing that
its expert-judgment system is less than
fully transparent, a bank could offset
this vulnerability by opting for complete
independence in the rating approval
process and an enhanced rating system
review.
Other considerations would influence
the choice of controls when banks use
models to assign ratings. While the
ratings produced by models are
transparent, a model’s performance
depends on how well the model was
developed, the model’s logic, and the
quality of the data used to implement
the model. Banks that use models to
assign ratings must implement a system
of controls that addresses model
development, testing and
implementation, data integrity and
overrides. These activities would be
covered by a comprehensive and
independent rating system review and
by ongoing spot checks on the accuracy
of model inputs. Other control
mechanisms such as accountability and
audit would also be required.
B. Independence in the Rating Approval
Process
An independent rating process is one
in which the parties responsible for
approving ratings and transactions are
separate from sales and marketing and
in which the persons approving ratings
are principally compensated on riskrating accuracy. As relative
independence increases, the likelihood
of accurate ratings assignments grows
markedly.
S. Ratings must be subject to
independent approval or review.
One way institutions can better
achieve objective and accurate risk
ratings is by ensuring that its rating
approval process is independent.
Institutions that firmly separate sales/
marketing from credit are better able to
manage the conflict between the goal of
high sales volume and the need for good
credit quality. An institution whose
rating process is less independent must
compensate by strengthening other
control and oversight mechanisms. A
significant factor in the evaluation of the
rating system will be the assessment of
whether such compensating controls are
sufficient to offset a less-thanindependent ratings process. While the
overriding objective is to achieve
independence in the rating approval
process, in some instances, the relative
materiality of a portfolio and cost/

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

benefit trade-offs may support a less
rigorous control process.
The degree of independence achieved
in the rating process depends on how an
institution is organized and how it
conducts its lending activities.
Rating Approval Processes
Responsibility for recommending and
approving ratings varies by institution
and, quite often, by portfolio.7 At some
institutions, ratings are assigned and
approved by relationship managers
(RMs); at others, deal teams assign
ratings that are later approved by credit
officers. Still other institutions have
independent credit officers assign and
approve ratings. The culture of an
institution and its business mix
generally determine whether the
business line or credit function is
ultimately responsible for ratings.
The subsections that follow describe
various rating assignment and approval
structures used by banking
organizations and the challenges that
emerge in ensuring objective and
consistent ratings. Any of the following
structures can work as long as ratings
are subject to an independent approval
or review process, and are not unduly
influenced by the line of business:
Relationship Managers. As noted
earlier, relationship managers are
primarily responsible for marketing the
bank’s products and services, and their
compensation is tied to the volume of
business they generate. When RMs also
have responsibility for assigning and
approving ratings, there is an inherent
conflict of interest. Credit quality and
the ability to produce timely and
accurate risk ratings are generally not
major factors in an RM’s compensation,
even when he or she has responsibility
for assigning and approving ratings. In
addition, RMs also may become too
close to the borrower to maintain their
objectivity and remain unbiased. When
banks delegate rating responsibility to
RMs, they must offset the lack of
independence with rigorous controls to
prevent bias from affecting the rating
process. Such controls must operate in
practice, not just on paper, and would
include, at a minimum, a
comprehensive, independent post7 Rating processes vary by institution but
generally involve an ‘‘assignor’’ and an ‘‘approver.’’
For instance, at many organizations the rating
assignor is the person who ‘‘owns’’ the relationship
(such as a ‘‘relationship manager’’) and the rating
approver is an individual with credit authority (a
‘‘credit risk manager’’). In some cases, the rating
assignor and approver are the same. Banks that
separate the rating assignment and approval
processes do so in order to minimize potential
conflicts of interest and the potential for rating
errors.

PO 00000

Frm 00027

Fmt 4701

Sfmt 4703

45975

closing review of ratings by a rating
system review function.
Deal Team. Some major banks employ
a ‘‘deal-team’’ structure for credit
origination and rating assignment. Using
this approach, all members of the
team—credit officers, investment
bankers, underwriters, and others—
contribute to analyzing
creditworthiness, underwriting the deal,
and assigning ratings.
On the one hand, deal teams increase
the access of credit officers to
information on obligors and transactions
early in the underwriting process,
enabling them to make more informed
credit decisions and to influence facility
structure to address obligors’
weaknesses. On the other hand,
participation in the deal team could
compromise the credit officer’s
objectivity. While credit officers
typically report to an independent
credit-risk-management function, they
also have allegiance to the deal team
that reports to executives within the
sales and marketing line of business. In
addition, credit officers may defer to the
members of the team whose
compensation is based on the revenue
and sales volume they generate for the
bank. Banks that maintain deal teams
must ensure that the credit officer’s
independence is safeguarded through
independent reporting lines and welldefined performance measures (e.g.,
adherence to policy, rating accuracy and
timeliness).
Credit Officers. Some banks give sole
responsibility for assigning and
approving ratings to credit officers who
report to an independent credit
function. In addition to assigning and
approving and assigning initial ratings,
credit officers regularly monitor the
condition of obligors and refresh ratings
as necessary. The potential downside of
this structure is that these credit officers
may have limited access to borrower
information. Those credit officers that
have a separate reporting line and
whose compensation is principally
based on their risk-rating accuracy are
typically more independent than RMs or
deal teams.
Models. At some institutions, models
assign ratings directly; at other
institutions, models and judgment are
combined to rate credits. Models
introduce a high degree of
independence to the rating process, but
they too require human oversight and
controls. Banks that use models must
incorporate an independent judgmental
review of the rating assignments to
ensure that all relevant information is
considered and to identify potential
rating errors. Judgmental reviews are
also needed when model outputs are

E:\FR\FM\04AUN2.SGM

04AUN2

45976

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

overridden. In addition, controls are
needed to ensure accuracy of data
inputs. When a bank uses a model to
assign risk ratings, an individual
obligor’s rating is ‘‘transparent.’’
However, the model itself is not
‘‘transparent’’ without a great deal of
effort to document how the model
functions.
C. Transparency
Transparency is the ability of a third
party, such as rating system reviewers,
auditors or bank supervisors, to observe
how the rating system operates and to
understand the pertinent characteristics
of individual ratings.
S. IRB institutions must have a
transparent rating system.
Transparency in a rating system is
achieved through documentation that
covers the following:
• The rating system’s design,
purpose, performance horizon, and
performance standards;
• The rating assignment process,
including procedures for adjustments
and overrides;
• Rating definitions and criteria,
scorecard criteria, and model
specifications;
• Parameter estimates and the process
for their estimation;
• Definition of the data elements to be
warehoused to support controls,
oversight, validation, and parameter
estimation; and
• Specific responsibilities of, and
performance standards for, individuals
and units involved in the rating system
and its oversight.
Transparency allows third parties
(such as rating system review, auditors,
or supervisors) to evaluate whether the
rating system is performing as intended.
Without transparency, it is difficult to
hold people accountable for ratings
errors and to validate the performance
of the system.
S. Rating criteria must be clear and
specific and must include qualitative
and quantitative factors.
To produce transparent individual
ratings, a bank’s policies must contain
clear, detailed ratings definitions. Banks
should specify criteria for each factor
that raters must consider, which may
require unique rating definitions for
certain industries. Banks should
consider criteria for factors such as
liquidity, sales and profitability, debt
service and fixed charge coverage,
minimum equity support, position
within the industry, strength of
management. A rating system with
vague criteria or one merely defined by
PDs or LGDs is not transparent. For
example, the following rating
definitions are not transparent because

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

they require the rater to do too much
interpreting:
Borrower exhibits satisfactory quality
and demonstrates acceptable principal
and interest repayment capacity in the
near term.
Lower tier company in a cyclical
industry. Unbalanced position with
tight liquidity and high leverage.
Declining or erratic profitability and
marginal debt service capacity.
Management is untested.
D. Accountability
‘‘Accountability’’ is holding people
responsible for their actions and
establishing adverse consequences for
inaccurate ratings.
S. Policies must identify the parties
responsible for rating accuracy and
rating system performance.
For accountability to be effective, it
should be both observable and ingrained
in the culture. Persons who assign and
approve rate credits, derive parameter
estimates, or oversee rating systems
must be held accountable for complying
with rating system policies and ensuring
that aspects of the rating system within
their control are as unbiased and
accurate as possible. These persons
must have the tools and resources
necessary to carry out their
responsibilities, and their performance
should be evaluated against clear and
specific objectives documented in
policy.
Responsibility for Assigning Ratings
S. Individuals must be held
accountable for complying with rating
system policies and for assigning
accurate ratings, and their performance
and compensation must be linked to
well-defined measurable performance
standards.
Responsibilities of raters should be
clear, and performance should be
measured against specific objectives.
Performance evaluation and incentive
compensation should be tied to
performance goals. Examples of
performance measures include:
• Number and frequency of rating
errors,
• Significance of errors (for example,
multiple downgrades), and
• Proper and consistent application of
criteria, including override criteria.
Responsibility for Rating System
Performance
Just as individuals will be held
accountable for the accuracy of ratings,
an individual must be held responsible
for the overall performance of the rating
system. This individual must ensure
that the rating system and all of its
component parts—rating assignments,

PO 00000

Frm 00028

Fmt 4701

Sfmt 4703

parameter estimation, data collection,
control and oversight mechanisms—are
functioning as intended. While these
components often are housed within
separate units of the organization, an
individual must be responsible for
ensuring that the parts work together
effectively and efficiently.
E. Use of Ratings
S. Ratings used for regulatory capital
must be the same ratings used to guide
day-to-day credit risk management
activities.
The different uses and applications of
the risk-rating system’s outputs should
promote greater accuracy and
consistency of credit-risk evaluations
across an organization. Ratings and the
associated default, loss, and EAD
estimates need to be incorporated
within the credit-risk management,
internal capital allocation, and
corporate governance functions of IRB
banks.
S. Banks that use parameter estimates
for risk management that are different
from those used for regulatory capital
must provide a well-documented
rationale for the differences.
PD and LGD parameters used for
regulatory capital purposes may not be
appropriate for other uses purposes. For
example, PD estimates used to estimate
reserve needs could reflect current
economic conditions that are different
from the longer term view appropriate
to calculations of regulatory capital.
When banks employ different estimates,
those parameters must be defensible and
supported by the following:
• Qualitative and quantitative
analysis of the logic and rationale for
the difference(s); and
• Senior management approval of the
difference(s).
F. Rating System Review (RSR)
S. Banks must have a comprehensive,
coordinated, independent review
process to ensure that ratings are
accurate and that the rating system is
performing as intended.
Rating system review (RSR) ensures
that the rating system as a whole is
functioning as intended. A broad range
of responsibilities come under RSR’s
purview, as outlined in Table 4.2:

TABLE 4.2.—RESPONSIBILITIES OF
RATING SYSTEM REVIEW
Scope of Review:
Design of the rating system.
Compliance with policies and procedures,
including application of criteria.
Check of all risk-rating grades for accuracy.
Consistency across industries/portfolios/geographies.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
TABLE 4.2.—RESPONSIBILITIES OF
RATING SYSTEM REVIEW—Continued
Model development.
Model use, including inputs and outputs.
Overrides and policy exceptions.
Quantification process.
Back-testing (perform or review).
Actual and predicted ratings transitions.
Benchmarking against third-party data
sources (perform or review).
Adequacy of data maintenance.
Analysis and Reporting:
Identify errors and flaws.
Recommend corrective action.

For each of these responsibilities, RSR
is largely checking and confirming the
work of others and ensuring that the
rating system’s components work well
together. RSR’s testing and review
should identify current and potential
weaknesses and should lead to
recommendations and corrective action
such as
• Adjusting policies and procedures,
• Requiring additional training of
staff,
• Investing in infrastructure
improvements,
• Adjusting rating criteria, and
• Adjusting parameter estimates.
S. Rating system review must report
significant findings to senior
management and the board quarterly.
RSR’s role is to identify issues and
areas of concern and report findings to
the area that is accountable. When
issues are systematic, RSR should bring
them to the attention of senior
management and the board.
The activities of this function could
be distributed across multiple areas or
housed within one unit. Organizations
will choose a structure that fits within
their management and oversight
framework. These units must always
have high standing within the
organization and should be staffed by
individuals possessing the requisite
stature, skills, and experience.
Like internal audit, RSR must be
independent from all in-house designers
and developers (that is, system and
model designers) and raters (that is,
ratings and parameter assigners) in the
risk-rating process. RSR’s independence
eliminates potential conflicts of interest
and gives the group credibility when it
reports findings and conclusions to the
board and senior management.
G. Internal Audit
S. An independent internal audit
function must determine whether rating
system controls function as intended.
S. Internal audit must evaluate
annually whether the bank is in
compliance with the risk-based capital
regulation and supervisory guidance.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Internal audit determines whether the
bank’s system of controls over internal
ratings and the related parameters is
robust. In its evaluation of controls,
internal audit must consider any tradeoffs made between the various
mechanisms and confirm their
continued appropriateness and
relevance. As part of its review of
control mechanisms, audit will evaluate
the depth, scope, and quality of RSR’s
work and will conduct limited testing to
ensure that their conclusions are well
founded. The amount of testing will
depend on whether audit is the primary
or secondary reviewer of that work.
Internal audit will report to the board
and management on whether the bank is
in compliance with the IRB standards.
This report will allow the board and
management to disclose that its rating
processes and the controls surrounding
these processes are in compliance with
the IRB standards. This will be critical
for public disclosure and ongoing work
of supervisors.
External Audit
As part of the process of certifying
financial statements, external auditors
will confirm that the institution’s
capital position is fairly presented. To
verify that actual capital exceeds
regulatory minimums and to confirm
compliance with the IRB rules, the
external auditors must ascertain that the
IRB system is rating credit risk
appropriately and linking these ratings
to appropriate estimates. Auditors must
evaluate the bank’s internal control
functions and its compliance with the
risk-based capital regulation and
supervisory guidance.
H. Corporate Oversight
S. The full board or a committee of
the board must approve key elements of
the IRB system.
Consistent with sound practice, bank
management must ensure that a
corporate culture exists in which
institutional needs are readily identified
and appropriate resources are brought to
bear to rectify shortcomings. In the IRB
context, senior management and the
board of directors must ensure the
objectivity and accuracy of the bank’s
credit-risk management systems and
approach.
Either the full board or a committee
of the board should approve key
elements of the risk-rating system.
Information provided to the board
should be sufficiently detailed to allow
directors to confirm the continuing
appropriateness of the institution’s
rating approach and to verify the
adequacy of the controls supporting the
rating system.

PO 00000

Frm 00029

Fmt 4701

Sfmt 4703

45977

S. Senior management must ensure
that all components of the IRB system,
including controls, are functioning as
intended and comply with the riskbased capital regulation and supervisory
guidance.
Senior management’s oversight
should be even more active than that of
the board of directors. Senior
management should articulate what it
expects of the technical and operational
units of the risk-rating system, as well
as what it expects of the units that
manage the system’s controls. To
oversee the risk-rating system, senior
management must have an extensive
understanding of credit policies,
underwriting standards, lending
practices, and collection and recovery
practices, and must be able to
understand how these factors affect
default and loss estimates. Senior
management should not only oversee
the controls process (its traditional role)
but also should periodically meet with
raters and validators to discuss the
rating system’s performance, areas
needing improvement, and the status of
efforts to improve previously identified
deficiencies.
The depth and frequency of
information provided to the board and
senior management must be
commensurate with their oversight
responsibilities and the condition of the
institution. These reports should
include the following information:
• Risk profile by grade,
• Risk rating migration across grades
with emphasis on unexpected results,
• Changes in parameter estimates by
grade,
• Comparison of realized PD, LGD,
and EAD rates against expectations,
• Reports measuring changes in
regulatory and economic capital,
• Results of capital stress testing, and
• Reports generated by rating system
review, audit, and other control units.
Although all of an institution’s
controls must function smoothly,
independently, and in concert with the
others, the direction and oversight
provided by the board and senior
management are perhaps most
important to ensure that the IRB system
is functioning properly.
Document 2: Draft Supervisory
Guidance on Operational Risk
Advanced Measurement Approaches
for Regulatory Capital
Table of Contents
I. Purpose
II. Background
III. Definitions
IV. Banking Activities and Operational Risk
V. Corporate Governance
A. Board and Management Oversight

E:\FR\FM\04AUN2.SGM

04AUN2

45978

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

B. Independent Firm-wide Risk
Management Function
C. Line of Business Management
VI. Operational Risk Management Elements
A. Operational Risk Policies and
Procedures
B. Identification and Measurement of
Operational Risk
C. Monitoring and Reporting
D. Internal Control Environment
VII. Elements of an AMA Framework
A. Internal Operational Risk Loss Event
Data
B. External Data
C. Business Environment and Internal
Control Factor Assessments
D. Scenario Analysis
VIII. Risk Quantification
A. Analytical Framework
B. Accounting for Dependence
IX. Risk Mitigation
X. Data Maintenance
XI. Testing and Verification
Appendix A: Supervisory Standards for the
AMA

I. Purpose
The purpose of this guidance is to set
forth the expectations of the U.S.
banking agencies for banking
institutions that use Advanced
Measurement Approaches (AMA) for
calculating the operational risk capital
charge under the new capital regulation.
Institutions using the AMA will have
considerable flexibility to develop
operational risk measurement systems
appropriate to the nature of their
activities, business environment, and
internal controls. An institution’s
operational risk regulatory capital
requirement will be calculated as the
amount needed to cover its operational
risk at a level of confidence determined
by the supervisors, as discussed below.
Use of an AMA is subject to supervisory
approval.
This draft guidance should be
considered with the advance notice of
proposed rulemaking (ANPR) on
revisions to the risk-based capital
standard published elsewhere in today’s
Federal Register. As with the ANPR, the
Agencies are seeking industry comment
on this draft guidance. In addition to
seeking comment on all specific aspects
of this supervisory guidance, the
Agencies are seeking comment on the
extent to which the supervisory
guidance strikes the appropriate balance
between flexibility and specificity.
Likewise, the Agencies are seeking
comment on whether an appropriate
balance has been struck between the
regulatory requirements set forth in the
ANPR and the supervisory standards set
forth in this guidance.
II. Background
Effective management of operational
risk is integral to the business of

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

banking and to institutions’ roles as
financial intermediaries. Although
operational risk is not a new risk,
deregulation and globalization of
financial services, together with the
growing sophistication of financial
technology, new business activities and
delivery channels, are making
institutions’ operational risk profiles
(i.e., the level of operational risk across
an institution’s activities and risk
categories) more complex.
This guidance identifies the
supervisory standards (S) that
institutions must meet and maintain to
use an AMA for the regulatory capital
charge for operational risk. The purpose
of the standards is to provide the
foundation for a sound operational risk
framework, while allowing institutions
to identify the most appropriate
mechanisms to meet AMA
requirements. Each institution will need
to consider its complexity, range of
products and services, organizational
structure, and risk management culture
as it develops its AMA. Operational risk
governance processes need to be
established on a firm-wide basis to
identify, measure, monitor, and control
operational risk in a manner comparable
with the treatment of credit, interest
rate, and market risks.
Institutions will be expected to
develop a framework that measures and
quantifies operational risk for regulatory
capital purposes. To do this, institutions
will need a systematic process for
collecting operational risk loss data,
assessing the risks within the
institution, and adopting an analytical
framework that translates the data and
risk assessments into an operational risk
exposure (see definition below). The
analytical framework must incorporate a
degree of conservatism that is
appropriate for the overall robustness of
the quantification process. Because
institutions will be permitted to
calculate their minimum regulatory
capital on the basis of internal
processes, the requirements for data
capture, risk assessment, and the
analytical framework described below
are detailed and specific.
Effective operational risk
measurement systems are built on both
quantitative and qualitative risk
assessment techniques. While the
output of the regulatory framework for
operational risk is a measure of
exposure resulting in a capital number,
the integrity of that estimate depends
not only on the soundness of the
measurement model, but also on the
robustness of the institution’s
underlying risk management processes.
In addition, supervisors view the
introduction of the AMA as an

PO 00000

Frm 00030

Fmt 4701

Sfmt 4703

important tool to further promote
improvements in operational risk
management and controls at large
banking institutions.
This document provides both AMA
supervisory standards and a discussion
of how those standards should be
incorporated into an operational risk
framework. The relevant supervisory
standards are listed at the beginning of
each section and a full compilation of
the standards is provided in Appendix
A. Not every section has specific
supervisory standards. When spanning
more than one section, supervisory
standards are listed only once.
Institutions will be required to meet,
and remain in compliance with, all the
supervisory standards to use an AMA
framework. However, evaluating an
institution’s qualification with each of
the individual supervisory standards
will not be sufficient to determine an
institution’s overall readiness for AMA.
Instead, supervisors and institutions
must also evaluate how well the various
components of an institution’s AMA
framework complement and reinforce
one another to achieve the overall
objectives of an accurate measure and
effective management of operational
risk. In performing their evaluation,
supervisors will exercise considerable
supervisory judgment, both in
evaluating the individual components
and the overall operational risk
framework.
An institution’s AMA methodology
will be assessed as part of the ongoing
supervision process. This will allow
supervisors to incorporate existing
supervisory efforts as much as possible
into the AMA assessments. Some
elements of operational risk (e.g.,
internal controls and information
technology) have long been subject to
examination by supervisors. Where this
is the case, supervisors will make every
effort to leverage off these examination
activities to assess the effectiveness of
the AMA process. Substantive
weaknesses identified in an
examination will be factored into the
AMA qualification process.
III. Definitions
There are important definitions that
institutions must incorporate into an
AMA framework. They are:
• Operational risk: The risk of loss
resulting from inadequate or failed
internal processes, people and systems,
or from external events. The definition
includes legal risk, which is the risk of
loss resulting from failure to comply
with laws as well as prudent ethical
standards and contractual obligations. It
also includes the exposure to litigation
from all aspects of an institution’s

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
activities. The definition does not
include strategic or reputational risks.8
• Operational risk loss: The financial
impact associated with an operational
event that is recorded in the
institution’s financial statements
consistent with Generally Accepted
Accounting Principles (GAAP).
Financial impact includes all out-ofpocket expenses associated with an
operational event but does not include
opportunity costs, foregone revenue, or
costs related to investment programs
implemented to prevent subsequent
operational risk losses. Operational risk
losses are characterized by seven event
factors associated with:
i. Internal fraud: An act of a type
intended to defraud, misappropriate
property or circumvent regulations, the
law or company policy, excluding
diversity/discrimination events, which
involve at least one internal party.
ii. External fraud: An act of a type
intended to defraud, misappropriate
property or circumvent the law, by a
third party.
iii. Employment practices and
workplace safety: An act inconsistent
with employment, health or safety laws
or agreements, from payment of
personal injury claims, or from
diversity/discrimination events.
iv. Clients, products, and business
practices: An unintentional or negligent
failure to meet a professional obligation
to specific clients (including fiduciary
and suitability requirements), or from
the nature or design of a product.
v. Damage to physical assets: The loss
or damage to physical assets from
natural disaster or other events.
vi. Business disruption and system
failures: Disruption of business or
system failures.
vii. Execution, delivery, and process
management: Failed transaction
processing or process management, from
relations with trade counterparties and
vendors.
• Operational risk exposure: An
estimate of the potential operational
losses that the banking institution faces
at a soundness standard consistent with
a 99.9 per cent confidence level over a
one-year period. The institution will
multiply the exposure by 12.5 to obtain
risk-weighted assets for operational risk;
this is added to the risk-weighted assets
for credit and market risk to arrive at the
denominator of the regulatory capital
ratio.
• Business environment and internal
control factor assessments: The range of
tools that provide a meaningful
8 An institution’s definition of risk may
encompass other risk elements as long as the
supervisory definition is met.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

assessment of the level and trends in
operational risk across the institution.
While the institution may use multiple
tools in an AMA framework, they must
all have the same objective of
identifying key risks. There are a
number of existing tools, such as audit
scores and performance indicators that
may be acceptable under this definition.
IV. Banking Activities and Operational
Risk
The above definition of operational
risk gives a sense of the breadth of
exposure to operational risk that exists
in banking today as well as the many
interdependencies among risk factors
that may result in an operational risk
loss. Indeed, operational risk can occur
in any activity, function, or unit of the
institution.
The definition of operational risk
incorporates the risks stemming from
people, processes, systems and external
events. People risk refers to the risk of
management failure, organizational
structure or other human resource
failures. These risks may be exacerbated
by poor training, inadequate controls,
poor staffing resources, or other factors.
The risk from processes stem from
breakdowns in established processes,
failure to follow processes, or
inadequate process mapping within
business lines. System risk covers
instances of both disruption and
outright system failures in both internal
and outsourced operations. Finally,
external events can include natural
disasters, terrorism, and vandalism.
There are a number of areas where
operational risks are emerging. These
include:
• Greater use of automated
technology has the potential to
transform risks from manual processing
errors to system failure risks, as greater
reliance is placed on globally integrated
systems;
• Proliferation of new and highly
complex products;
• Growth of e-banking transactions
and related business applications
expose an institution to potential new
risks (e.g., internal and external fraud
and system security issues);
• Large-scale acquisitions, mergers,
and consolidations test the viability of
new or newly integrated systems;
• Emergence of institutions acting as
large-volume service providers create
the need for continual maintenance of
high-grade internal controls and backup systems;
• Development and use of risk
mitigation techniques (e.g., collateral,
insurance, credit derivatives, netting
arrangements and asset securitizations)
optimize an institution’s exposure to

PO 00000

Frm 00031

Fmt 4701

Sfmt 4703

45979

market risk and credit risk, but
potentially create other forms of risk
(e.g., legal risk); and
• Greater use of outsourcing
arrangements and participation in
clearing and settlement systems mitigate
some risks while increasing others.
The range of banking activities and
areas affected by operational risk must
be fully identified and considered in the
development of the institution’s risk
management and measurement plans.
Since operational risk is not confined to
particular business lines 9, product
types, or organizational units, it should
be managed in a consistent and
comprehensive manner across the
institution. Consequently, risk
management mechanisms must
encompass the full range of risks, as
well as strategies that help to identify,
measure, monitor and control those
risks.
V. Corporate Governance
Supervisory Standards
S 1. The institution’s operational risk
framework must include an
independent firm-wide operational risk
management function, line of business
management oversight, and
independent testing and verification
functions.
The management structure underlying
an AMA operational risk framework
may vary between institutions.
However, within all AMA institutions,
there are three key components that
must be evident—the firm-wide
operational risk management function,
lines of business management, and the
testing and verification function. These
three elements are functionally
independent 10 organizational
components, but should work in
cooperation to ensure a robust
operational risk framework.
A. Board and Management Oversight
Supervisory Standards
S 2. The board of directors must
oversee the development of the firmwide operational risk framework, as
9 Throughout this guidance, terms such as
‘‘business units’’ and ‘‘business lines’’ are used
interchangeably and refer not only to an
institution’s revenue-generating businesses, but also
to corporate staff functions such as human
resources or information technology.
10 For the purposes of AMA, ‘‘functional
independence’’ is defined as the ability to carry out
work freely and objectively and render impartial
and unbiased judgments. There should be
appropriate independence between the firm-wide
operational risk management functions, line of
business management and staff and the testing/
verification functions. Supervisory assessments of
independence issues will rely upon existing
regulatory guidance (e.g. audit, internal control
systems, board of directors/management, etc.)

E:\FR\FM\04AUN2.SGM

04AUN2

45980

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

well as major changes to the framework.
Management roles and accountability
must be clearly established.
S 3. The board of directors and
management must ensure that
appropriate resources are allocated to
support the operational risk framework.
The board is responsible for
overseeing the establishment of the
operational risk framework, but may
delegate the responsibility for
implementing the framework to
management with the authority
necessary to allow for its effective
implementation. Other key
responsibilities of the board include:
• Ensuring appropriate management
responsibility, accountability and
reporting;
• Understanding the major aspects of
the institution’s operational risk as a
distinct risk category that should be
managed;
• Reviewing periodic high-level
reports on the institution’s overall
operational risk profile, which identify
material risks and strategic implications
for the institution;
• Overseeing significant changes to
the operational risk framework; and
• Ensuring compliance with
regulatory disclosure requirements.
Effective board and management
oversight forms the cornerstone of an
effective operational risk management
process. The board and management
have several broad responsibilities with
respect to operational risk:
• To establish a framework for
assessing operational risk exposure and
identify the institution’s tolerance for
operational risk;
• To identify the senior managers
who have the authority for managing
operational risk;
• To monitor the institution’s
performance and overall operational
risk profile, ensuring that it is
maintained at prudent levels and is
supported by adequate capital;
• To implement sound fundamental
risk governance principles that facilitate
the identification, measurement,
monitoring, and control of operational
risk;
• To devote adequate human and
technical resources to operational risk
management; and
• To institute remuneration policies
that are consistent with the institution’s
appetite for risk and are sufficient to
attract qualified operational risk
management and staff.
Management should translate the
operational risk management framework
into specific policies, processes and
procedures that can be implemented
and verified within the institution’s
different business units.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

Communication of these elements will
be essential to the understanding and
consistent treatment of operational risk
across the institution. While each level
of management is responsible for
effectively implementing the policies
and procedures within its purview,
senior management should clearly
assign authority, responsibilities, and
reporting relationships to encourage and
maintain this accountability and ensure
that the necessary resources are
available to manage operational risk.
Moreover, management should assess
the appropriateness of the operational
risk management oversight process in
light of the risks inherent in a business
unit’s activities. The testing and
verification function is responsible for
completing timely and comprehensive
assessments of the effectiveness of
implementation of the institution’s
operational risk framework at the line of
business and firm-wide levels.
Management collectively is also
responsible for ensuring that the
institution has qualified staff and
sufficient resources to carry out the
operational risk functions outlined in
the operational risk framework.
Additionally, management must
communicate operational risk issues to
appropriate staff that may not be
directly involved in its management.
Key management responsibilities
include ensuring that:
• Operational risk management
activities are conducted by qualified
staff with the necessary experience,
technical capabilities and access to
adequate resources;
• Sufficient resources have been
allocated to operational risk
management, in the business lines as
well as the independent firm-wide
operational risk management function
and verification areas, so as to
sufficiently monitor and enforce
compliance with the institution’s
operational risk policy and procedures;
and
• Operational risk issues are
effectively communicated with staff
responsible for managing credit, market
and other risks, as well as those
responsible for purchasing insurance
and managing third-party outsourcing
arrangements.
B. Independent Firm-Wide Risk
Management Function
Supervisory Standards
S 4. The institution must have an
independent operational risk
management function that is responsible
for overseeing the operational risk
framework at the firm level to ensure
the development and consistent

PO 00000

Frm 00032

Fmt 4701

Sfmt 4703

application of operational risk policies,
processes, and procedures throughout
the institution.
S 5. The firm-wide operational risk
management function must ensure
appropriate reporting of operational risk
exposures and loss data to the board of
directors and senior management.
The institution must have an
independent firm-wide operational risk
management function. The roles and
responsibilities of the function will vary
between institutions, but must be
clearly documented. The independent
firm-wide operational risk function
should have organizational stature
commensurate with the institution’s
operational risk profile, while remaining
independent of the lines of business and
the testing and verification function. At
a minimum, the institution’s
independent firm-wide operational risk
management function should ensure the
development of policies, processes, and
procedures that explicitly manage
operational risk as a distinct risk to the
institution’s safety and soundness.
These policies, processes and
procedures should include principles
for how operational risk is to be
identified, measured, monitored, and
controlled across the organization.
Additionally, they should provide for
the collection of the data needed to
calculate the institution’s operational
risk exposure.
Additional responsibilities of the
independent firm-wide operational risk
management function include:
• Assisting in the implementation of
the overall firm-wide operational risk
framework;
• Reviewing the institution’s progress
towards stated operational risk
objectives, goals and risk tolerances;
• Periodically reviewing the
institution’s operational risk framework
to consider the loss experience, effects
of external market changes, other
environmental factors, and the potential
for new or changing operational risks
associated with new products, activities
or systems. This review process should
include an assessment of industry best
practices for the institution’s activities,
systems and processes;
• Reviewing and analyzing
operational risk data and reports; and
• Ensuring appropriate reporting to
senior management and the board.
C. Line of Business Management
Supervisory Standards
S 6. Line of business management is
responsible for the day-to-day
management of operational risk within
each business unit.
S 7. Line of business management
must ensure that internal controls and

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
practices within their line of business
are consistent with firm-wide policies
and procedures to support the
management and measurement of the
institution’s operational risk.
Line of business management is
responsible for both managing
operational risk within the business
lines and ensuring that policies and
procedures are consistent with and
support the firm-wide operational risk
framework. Management should ensure
that business-specific policies,
processes, procedures and staff are in
place to manage operational risk for all
material products, activities, and
processes. Implementation of the
operational risk framework within each
line of business should reflect the scope
of that business and its inherent
operational complexity and operational
risk profile. Line of business
management must be independent of
both the firm-wide operational risk
management and the testing and
verification functions.
VI. Operational Risk Management
Elements
The operational risk management
framework provides the overall
operational risk strategic direction and
ensures that an effective operational risk
management and measurement process
is adopted throughout the institution.
The framework should provide for the
consistent application of operational
risk policies and procedures throughout
the institution and address the roles of
both the independent firm-wide
operational risk management function
and the lines of business. The
framework should also provide for the
consistent and comprehensive capture
of data elements needed to measure and
verify the institution’s operational risk
exposure, as well as appropriate
operational risk analytical frameworks,
reporting systems, and mitigation
strategies. The framework must also
include independent testing and
verification to assess the effectiveness of
implementation of the institution’s
operational risk framework, including
compliance with policies, processes,
and procedures.
In practice, an institution’s
operational risk framework must reflect
the scope and complexity of business
lines, as well as the corporate
organizational structure. Each
institution’s operational risk profile is
unique and requires a tailored risk
management approach appropriate for
the scale and materiality of the risks
present, and the size of the institution.
There is no single framework that would
suit every institution; different
approaches will be needed for different

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

institutions. In fact, many operational
risk management techniques continue to
evolve rapidly to keep pace with new
technologies, business models and
applications.
The key elements in the operational
risk management process include:
• Appropriate policies and
procedures;
• Efforts to identify and measure
operational risk;
• Effective monitoring and reporting;
• A sound system of internal controls;
and
• Appropriate testing and verification
of the operational risk framework.
A. Operational Risk Policies and
Procedures
Supervisory Standards
S 8. The institution must have
policies and procedures that clearly
describe the major elements of the
operational risk management
framework, including identifying,
measuring, monitoring, and controlling
operational risk.
Operational risk management
policies, processes, and procedures
should be documented and
communicated to appropriate staff. The
policies and procedures should outline
all aspects of the institution’s
operational risk management
framework, including:
• The roles and responsibilities of the
independent firm-wide operational risk
management function and line of
business management;
• A definition for operational risk,
including the loss event types that will
be monitored;
• The capture and use of internal and
external operational risk loss data,
including large potential events
(including the use of scenario analysis);
• The development and incorporation
of business environment and internal
control factor assessments into the
operational risk framework;
• A description of the internally
derived analytical framework that
quantifies the operational risk exposure
of the institution;
• An outline of the reporting
framework and the type of data/
information to be included in line of
business and firm-wide reporting;
• A discussion of qualitative factors
and risk mitigants and how they are
incorporated into the operational risk
framework;
• A discussion of the testing and
verification processes and procedures;
• A discussion of other factors that
affect the measurement of operational
risk; and

PO 00000

Frm 00033

Fmt 4701

Sfmt 4703

45981

• Provisions for the review and
approval of significant policy and
procedural exceptions.
B. Identification and Measurement of
Operational Risk
The result of a comprehensive
program to identify and measure
operational risk is an assessment of the
institution’s operational risk exposure.
Management must establish a process
that identifies the nature and types of
operational risk and their causes and
resulting effects on the institution.
Proper operational risk identification
supports the reporting and maintenance
of capital for operational risk exposure
and events, facilitates the establishment
of mechanisms to mitigate or control the
risks, and ensures that management is
fully aware of the sources of emerging
operational risk loss events.
C. Monitoring and Reporting
Supervisory Standards
S 9. Operational risk management
reports must address both firm-wide
and line of business results. These
reports must summarize operational risk
exposure, loss experience, relevant
business environment and internal
control assessments, and must be
produced no less often than quarterly.
S 10. Operational risk reports must
also be provided periodically to senior
management and the board of directors,
summarizing relevant firm-wide
operational risk information.
Ongoing monitoring of operational
risk exposures is a key aspect of an
effective operational risk framework. To
facilitate monitoring of operational risk,
results from the measurement system
should be summarized in reports that
can be used by the firm-wide
operational risk and line of business
management functions to understand,
manage, and control operational risk
and losses. These reports should serve
as a basis for assessing operational risk
and related mitigation strategies and
creating incentives to improve
operational risk management
throughout the institution.
Operational risk management reports
should summarize:
• Operational risk loss experience on
an institution, line of business, and
event-type basis;
• Operational risk exposure;
• Changes in relevant risk and control
assessments;
• Management assessment of early
warning factors signaling an increased
risk of future losses;
• Trend analysis, allowing line of
business and independent firm-wide
operational risk management to assess

E:\FR\FM\04AUN2.SGM

04AUN2

45982

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

and manage operational risk exposures,
systemic line of business risk issues,
and other corporate risk issues;
• Exception reporting; and
• To the extent developed,
operational risk causal factors.
High-level operational risk reports
must also be produced periodically for
the board and senior management.
These reports must provide information
regarding the operational risk profile of
the institution, including the sources of
material risk both from a firm-wide and
line of business perspective, versus
established management expectations.
Supervisory Standards
S 11. An institution’s internal control
structure must meet or exceed minimum
regulatory standards established by the
Agencies.
Sound internal controls are essential
to an institution’s management of
operational risk and are one of the
foundations of safe and sound banking.
When properly designed and
consistently enforced, a sound system of
internal controls will help management
safeguard the institution’s resources,
produce reliable financial reports, and
comply with laws and regulations.
Sound internal controls will also reduce
the possibility of significant human
errors and irregularities in internal
processes and systems, and will assist in
their timely detection when they do
occur.
The Agencies are not introducing any
new internal control standards, but
rather emphasizing the importance of
meeting existing standards. There is a
recognition that internal control systems
will differ among institutions due to the
nature and complexity of an
institution’s products and services,
organizational structure, and risk
management culture. The AMA
standards allows for these differences,
while also establishing a baseline
standard for the quality of the internal
control structure. Institutions will be
expected to at least meet the minimum
interagency standards11 relating to
internal controls as a criterion for AMA
qualification.
11 There are a number of interagency standards
that cover topics relevant to the internal control
structure. These include, for example, the
Interagency Policy Statement on the Internal Audit
Function and Its Outsourcing (March 2003), the
Federal Financial Institution’s Examination
Council’s (FFIEC’s) Business Continuity Planning
Booklet (May 2003), the FFIEC’s Information
Security Booklet (January 2003). In addition, each
Agency has extensive guidance on corporate
governance, internal controls, and monitoring and
reporting in its respective examination policies and
procedures.

19:57 Aug 01, 2003

Jkt 200001

VII. Elements of an AMA Framework
Supervisory Standards

D. Internal Control Environment

VerDate jul<14>2003

The extent to which an institution
meets or exceeds the minimum
standards will primarily be assessed
through current and ongoing
supervisory processes. As noted earlier,
the Agencies will leverage off existing
examination processes, to avoid
duplication in assessing an institution’s
implementation of an AMA framework.
Assessing the internal control
environment is clearly an area where
the supervisory authorities already
focus considerable attention.

S 12. The institution must
demonstrate that it has appropriate
internal loss event data, relevant
external loss event data, assessments of
business environment and internal
controls factors, and results from
scenario analysis to support its
operational risk management and
measurement framework.
S 13. The institution must include the
regulatory definition of operational risk
as the baseline for capturing the
elements of the AMA framework and
determining its operational risk
exposure.
S 14. The institution must have clear
standards for the collection and
modification of the elements of the
operational risk AMA framework.
Operational risk inputs play a
significant role in both the management
and measurement of operational risk.
Necessary elements of an institution’s
AMA framework include internal loss
event data, relevant external loss event
data, results of scenario analysis, and
assessments of the institution’s business
environment and internal controls.
Operational risk inputs aid the
institution in identifying the level and
trend of operational risk, determining
the effectiveness of risk management
and control efforts, highlighting
opportunities to better mitigate
operational risk, and assessing
operational risk on a forward-looking
basis.
To use its AMA framework, an
institution must demonstrate that it has
established a consistent and
comprehensive process for the capture
of all elements of the AMA framework.
The institution must also demonstrate
that it has clear standards for the
collection and modification of all AMA
inputs. While the analytical framework
will generally combine these inputs to
develop the operational risk exposure,
supervisors must have the capacity to
review the individual inputs as well;
specifically, supervisors will need to
review the loss information that is being

PO 00000

Frm 00034

Fmt 4701

Sfmt 4703

provided to the analytical framework
that stems from internal loss event data,
versus the loss event information
provided by external loss event data
capture, scenario analysis, or the
assessments of the business
environment and internal control
factors.
The capture systems must cover all
material business lines, business
activities and corporate functions that
could generate operational risk. The
institution must have a defined process
that establishes responsibilities over the
systems developed to capture the AMA
elements. In particular, the issue of
overriding the data capture systems
must be addressed. Any overrides
should be tracked separately and
documented. Tracking overrides
separately allows management and
supervisors to identify the nature and
rationale, including whether they stem
from simple input errors or, more
importantly, from exclusion because a
loss event was not pertinent for the
quantitative measurement. Management
should have clear standards for
addressing overrides and should clearly
delineate who has authority to override
the data systems and under what
circumstances.
As noted earlier, for AMA
qualification purposes, an institution’s
operational risk framework must, at a
minimum, use the definition of
operational risk that is provided in
paragraph 10 when capturing the
elements of the AMA framework.
Institutions may use an expanded
definition if considered more
appropriate for risk management and
measurement efforts. However, for the
quantification of operational risk
exposure for regulatory capital
purposes, an institution must
demonstrate that the AMA elements are
captured so as to meet the baseline
definition.
A. Internal Operational Risk Loss Event
Data
Supervisory Standards
S 15. The institution must have at
least five years of internal operational
risk loss data 12 captured across all
material business lines, events, product
types, and geographic locations.
S 16. The institution must be able to
map internal operational risk losses to
the seven loss-event type categories.
S 17. The institution must have a
policy that identifies when an
operational risk loss becomes a loss
event and must be added to the loss
12 With supervisory approval, a shorter initial
historical observation period is acceptable for banks
newly authorized to use an AMA methodology.

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
event database. The policy must provide
for consistent treatment across the
institution.
S 18. The institution must establish
appropriate operational risk data
thresholds.
S 19. Losses that have any
characteristics of credit risk, including
fraud-related credit losses, must be
treated as credit risk for regulatory
capital purposes. The institution must
have a clear policy that allows for the
consistent treatment of loss event
classifications (e.g., credit, market, or
operational risk) across the organization.
The key to internal data integrity is
the consistency and completeness with
which loss event data capture processes
are implemented across the institution.
Management must ensure that
operational risk loss event information
captured is consistent across the
business lines and incorporates any
corporate functions that may also
experience operational risk events.
Policies and procedures should be
addressed to the appropriate staff to
ensure that there is satisfactory
understanding of operational risk and
the data capture requirements under the
operational risk framework. Further, the
independent operational risk
management function must ensure that
the loss data is captured across all
material business lines, products types,
event types, and from all significant
geographic locations. The institution
must be able to capture and aggregate
internal losses that cross multiple
business lines or event types. If data is
not captured across all business lines or
from all geographic locations, the
institution must document and explain
the exceptions.
AMA institutions must be able to map
operational risk losses into the seven
loss event categories defined in
paragraph 10. Institutions will not be
required to produce reports or perform
analysis for internal purposes on the
basis of the loss event categories, but
will be expected to use the information
about the event-type categories as a
check on the comprehensiveness of the
institution’s data set.
The institution must have five years
of internal loss data, although a shorter
range of historical data may be allowed,
subject to supervisory approval. The
extent to which an institution collects
operational risk loss event data will, in
part, be dependent upon the data
thresholds that the institution
establishes. There are a number of
standards that an institution may use to
establish the thresholds. They may be
based on product types, business lines,
geographic location, or other
appropriate factors. The Agencies will

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

allow flexibility in this area, provided
the institution can demonstrate that the
thresholds are reasonable, do not
exclude important loss events, and
capture a significant proportion of the
institution’s operational risk losses.
The institution must capture
comprehensive data on all loss events
above its established threshold level.
Aside from information on the gross loss
amount, the institution should collect
information about the date of the event,
any recoveries, and descriptive
information about the drivers or causes
of the loss event. The level of detail of
any descriptive information should be
commensurate with the size of the gross
loss amount. Examples of the type of
information collected include:
• Loss amount;
• Description of loss event;
• Where the loss is reported and
expensed;
• Loss event type category;
• Date of the loss;
• Discovery date of the loss;
• Event end date;
• Management actions;
• Insurance recoveries;
• Other recoveries; and
• Adjustments to the loss estimate.
There are a number of additional data
elements that may be captured. It may
be appropriate, for example, to capture
data on ‘‘near miss’’ events, where no
financial loss was incurred. These near
misses will not factor into the regulatory
capital calculation, but may be useful
for the operational risk management
process.
Institutions will also be permitted and
encouraged to capture loss events in
their operational risk databases that are
treated as credit risk for regulatory
capital purposes, but have an
underlying element of operational risk
failure. These types of events, while not
incorporated into the regulatory capital
calculation, may have implications for
operational risk management. It will be
essential for institutions that capture
loss events that are treated differently
for regulatory capital and management
purposes to demonstrate that (1) loss
events are being captured consistently
across the institution; (2) the data
systems are sufficiently advanced to
allow for this differential treatment of
loss events; and (3) credit, market, and
operational risk losses are being
appropriated in the correct manner for
regulatory capital purposes.
The Agencies have established a clear
boundary between credit and
operational risks for regulatory capital
purposes. If a loss event has any
element of credit risk, it must be treated
as credit risk for regulatory capital
purposes. This would include all credit-

PO 00000

Frm 00035

Fmt 4701

Sfmt 4703

45983

related fraud losses. In addition,
operational risk losses with credit risk
characteristics that have historically
been included in institutions’ credit risk
databases will continue to be treated as
credit risk for the purposes of
calculating minimum regulatory capital.
The accounting guidance for credit
losses provides that creditors recognize
credit losses when it is probable that
they will be unable to collect all
amounts due according to the
contractual terms of a loan agreement.
Credit losses may result from the
creditor’s own underwriting, processing,
servicing or administrative activities
along with the borrower’s failure to pay
according to the terms of the loan
agreement. While the creditor’s
personnel, systems, policies or
procedures may affect the timing or
magnitude of a credit loss, they do not
change its character from credit to
operational risk loss for regulatory
capital purposes. Losses that arise from
a contractual relationship between a
creditor and a borrower are credit losses
whereas losses that arise outside of a
relationship between a creditor and a
borrower are operational losses.
B. External Data
Supervisory Standards
S 20. The institution must have
policies and procedures that provide for
the use of external loss data in the
operational risk framework.
S 21. Management must
systematically review external data to
ensure an understanding of industry
experience.
External data may serve a number of
different purposes in the operational
risk framework. Where internal loss data
is limited, external data may be a useful
input in determining the institution’s
level of operational risk exposure. Even
where external loss data is not an
explicit input to an institution’s data
set, such data provides a means for the
institution to understand industry
experience, and in turn, provides a
means for assessing the adequacy of its
internal data. External data may also
prove useful to inform scenario analysis,
fit severity distributions, or benchmark
the overall operational risk exposure
results.
To incorporate external loss
information into an institution’s
framework, the institution should
collect the following information:
• External loss amount;
• External loss description;
• Loss event type category;
• External loss event date;
• Adjustments to the loss amount
(i.e., recoveries, insurance settlements,

E:\FR\FM\04AUN2.SGM

04AUN2

45984

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

etc) to the extent that they are known;
and
• Sufficient information about the
reporting institution to facilitate
comparison to its own organization.
Institutions may obtain external loss
data in any reasonable manner. There
are many ways to do so; some
institutions are using data acquired
through membership with industry
consortia while other institutions are
using data obtained from vendor
databases or public sources such as
court records or media reports. In all
cases, management will need to
carefully evaluate the data source to
ensure that they are comfortable that the
information being reported is relevant
and reasonably accurate.
C. Business Environment and Internal
Control Factor Assessments
Supervisory Standards
S 22. The institution must have a
system to identify and assess business
environment and internal control
factors.
S 23. Management must periodically
compare the results of their business
environment and internal control factor
assessments against actual operational
risk loss experience.
While internal and external loss data
provide a historical perspective on
operational risk, it is also important that
institutions incorporate a forwardlooking element to the operational risk
measure. In principle, an institution
with strong internal controls in a stable
business environment will have less
exposure to operational risk than an
institution with internal control
weaknesses that is growing rapidly or
introducing new products. In this
regard, institutions will be required to
identify the level and trends in
operational risk in the institution. These
assessments must be current,
comprehensive across the institution,
and identify the critical operational
risks facing the institution.
The business environment and
internal control factor assessments
should reflect both the positive and
negative trends in risk management
within the institution as well as changes
in an institution’s business activities
that increase or decrease risk. Because
the results of the risk assessment are
part of the capital methodology,
management must ensure that the risk
assessments are done appropriately and
reflect the risks of the institution.
Periodic comparisons should be made
between actual loss exposure and the
assessment results.
The framework established to
maintain the risk assessments must be

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

sufficiently flexible to encompass an
institution’s increased complexity of
activities, new activities, changes in
internal control systems, or an increased
volume of information.
D. Scenario Analysis
Supervisory Standards
S 24. Management must have policies
and procedures that identify how
scenario analysis will be incorporated
into the operational risk framework.
Scenario analysis is a systematic
process of obtaining expert opinions
from business managers and risk
management experts to derive reasoned
assessments of the likelihood and
impact of plausible operational losses
consistent with the regulatory
soundness standard. Within an
institution’s operational risk framework,
scenario analysis may be used as an
input or may, as discussed below, form
the basis of an operational risk
analytical framework.
As an input to the institution’s
framework, scenario analysis is
especially relevant for business lines or
loss event types where internal data,
external data, and assessments of the
business environment and internal
control factors do not provide a
sufficiently robust estimate of the
institution’s exposure to operational
risk. In some cases, an institution’s
internal loss history may be sufficient to
provide a reasonable estimate of
exposure to future operational losses. In
other cases, the use of well-reasoned,
scaled external data may itself be a form
of scenario analysis.
The institution must have policies
and procedures that define scenario
analysis and identify its role in the
operational risk framework. The policy
should cover key elements of scenario
analysis, such as the manner in which
the scenarios are generated, the
frequency with which they are updated,
and the scope and coverage of
operational loss events they are
intended to reflect.
VIII. Risk Quantification
A. Analytical Framework
Supervisory Standards
S 25. The institution must have a
comprehensive operational risk
analytical framework that provides an
estimate of the institution’s operational
risk exposure, which is the aggregate
operational loss that it faces over a oneyear period at a soundness standard
consistent with a 99.9 per cent
confidence level.
S 26. Management must document the
rationale for all assumptions

PO 00000

Frm 00036

Fmt 4701

Sfmt 4703

underpinning its chosen analytical
framework, including the choice of
inputs, distributional assumptions, and
the weighting across qualitative and
quantitative elements. Management
must also document and justify any
subsequent changes to these
assumptions.
S 27. The institution’s operational risk
analytical framework must use a
combination of internal operational loss
event data, relevant external operational
loss event data, business environment
and internal control factor assessments,
and scenario analysis. The institution
must combine these elements in a
manner that most effectively enables it
to quantify its operational risk exposure.
The institution can choose the
analytical framework that is most
appropriate to its business model.
S 28. The institution’s capital
requirement for operational risk will be
the sum of expected and unexpected
losses unless the institution can
demonstrate, consistent with
supervisory standards, the expected loss
offset.
The industry has made significant
progress in recent years in developing
analytical frameworks to quantify
operational risk. The analytical
frameworks, which are a part of the
overall operational risk framework, are
based on various combinations of an
institution’s own operational loss
experience, the industry’s operational
loss experience, the size and scope of
the institution’s activities, the quality of
the institution’s control environment,
and management’s expert judgment.
Because these models capture specific
characteristics of each institution, such
models yield unique risk-sensitive
estimates of the institutions’ operational
risk exposures.
While the Agencies are not specifying
the exact methodology that an
institution should use to determine its
operational risk exposure, minimum
supervisory standards for acceptable
approaches have been developed. These
standards have been set so as to assure
that the regulation can accommodate
continued evolution of operational risk
quantification techniques, yet remain
amenable to consistent application and
enforcement across institutions. The
Agencies will require that the
institution have a comprehensive
analytical framework that provides an
estimate of the aggregate operational
loss that it faces over a one-year period
at a soundness standard consistent with
a 99.9 percent confidence level, referred
to as the institution’s operational risk
exposure. The institution will multiply
the exposure estimate by 12.5 to obtain
risk weighted assets for operational risk,

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
and add this figure to risk-weighted
assets for credit and market risk to
obtain total risk-weighted assets. The
final minimum regulatory capital
number will be 8 percent of total riskweighted assets.
The Agencies expect that there will be
significant variation in analytical
frameworks across institutions, with
each institution tailoring its framework
to leverage existing technology
platforms and risk management
procedures. These approaches may only
be used, provided they meet the
supervisory standards and include, as
inputs, internal operational loss event
data, relevant external operational loss
event data, assessments of business
environment and internal control
factors, and scenario analysis. The
Agencies do expect that there will be
some uncertainty and potential error in
the analytical frameworks because of the
evolving nature of operational risk
measurement and data capture.
Therefore, a degree of conservatism will
need to be built into the analytical
frameworks to reflect the evolutionary
status of operational risk and its impact
on data capture and analytical
modeling.
A diversity of analytical approaches is
emerging in the industry, combining
and weighting these inputs in different
ways. Most current approaches seek to
estimate loss frequency and loss severity
to arrive at an aggregate loss
distribution. Institutions then use the
aggregate loss distribution to determine
the appropriate amount of capital to
hold for a given soundness standard.
Scenario analysis is also being used by
many institutions, albeit to significantly
varying degrees. Some institutions are
using scenario analysis as the basis for
their analytical framework, while others
are incorporating scenarios as a means
for considering the possible impact of
significant operational losses on their
overall operational risk exposure.
The primary differences among
approaches being used today relate to
the weight that institutions place on
each input. For example, institutions
with comprehensive internal data may
place less emphasis on external data or
scenario analysis. Another example is
that some institutions estimate a unique
loss distribution for each business line/
loss type combination (bottom-up
approach) while others estimate a loss
distribution on a firm-wide basis and
then use an allocation methodology to
assign capital to business lines (topdown approach).
The Agencies expect internal loss
event data to play an important role in
the institution’s analytical framework,
hence the requirement for five years of

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

internal operational risk loss data.
However, as footnote 5 makes clear, five
years of data is not always required for
the analytical framework. For example,
if a bank exited a business line, the
institution would not be expected to
make use of that business unit’s loss
experience unless it had relevance for
other activities of the institution.
Another example would be where a
bank has made a recent acquisition
where the acquired firm does not have
internal loss event data. In these cases,
the Agencies expect the institution to
make use of the loss data available at the
acquired institution and any internal
loss data from operations similar to that
of the acquired firm, but the institution
will likely have to place more weight
relevant external loss event data, results
from scenario analysis, and factors
reflecting assessments of the business
environment and internal controls.
Whatever analytical approach an
institution chooses, it must document
and provide the rationale for all
assumptions embedded in its chosen
analytical framework, including the
choice of inputs, distributional
assumptions, and the weighting of
qualitative and quantitative elements.
Management must also document and
justify any subsequent changes to these
assumptions. This documentation
should:
• Clearly identify how the different
inputs are combined and weighted to
arrive at the overall operational risk
exposure so that the analytical
framework is transparent. The
documentation should demonstrate that
the analytical framework is
comprehensive and internally
consistent. Comprehensiveness means
that all required inputs are incorporated
and appropriately weighted. At the
same time, there should not be overlaps
or double counting.
• Clearly identify the quantitative
assumptions embedded in the
methodology and provide explanation
for the choice of these assumptions.
Examples of quantitative assumptions
include distributional assumptions
about frequency and severity, the
methodology for combining frequency
and severity to arrive at the overall loss
distribution, and dependence
assumptions between operational losses
across and within business lines.
• Clearly identify the qualitative
assumptions embedded in the
methodology and provide explanations
for the choice of these assumptions.
Examples of qualitative assumptions
include the use of business environment
and control factors as well as scenario
analysis in the approach.

PO 00000

Frm 00037

Fmt 4701

Sfmt 4703

45985

• Where feasible, provide results
based purely on quantitative methods
separately from results that incorporate
qualitative factors. This will provide a
transparent means of determining the
relative importance of quantitative
versus qualitative inputs.
• Where feasible, provide results
based on alternative quantitative and
qualitative assumptions to gauge the
overall model’s sensitivity to these
assumptions.
• Provide a comparison of the
operational risk exposure estimate
generated by the analytical framework
with actual loss experience over time, to
assess the reasonable of the framework’s
outputs.
• Clearly identify all changes to
assumptions, and provide explanations
for such changes.
• Clearly identify the results of an
independent verification of the
analytical framework.
The regulatory capital charge for
operational risk will include both
expected losses (EL) and unexpected
losses (UL). The Agencies have
considered two approaches that might
allow for some recognition of EL; these
approaches are reserving and budgeting.
However, both approaches raise
questions about their ability to act as an
EL offset for regulatory capital purposes.
The current U.S. GAAP treatment for
reserves (or liabilities) is based on an
incurred-loss (liability) model. Given
that EL is looking beyond current losses
to losses that will be incurred in the
future, establishing a reserve for
operational risk EL is not likely to meet
U.S. accounting standards. While
reserves are specific allocations for
incurred losses, budgeting is a process
of generally allocating future income for
loss contingencies, including losses
resulting from operational risk.
Institutions will be required to
demonstrate that budgeted funds are
sufficiently capital-like and remain
available to cover EL over the next year.
In addition, an institution will not be
permitted to recognize EL offsets on
budgeted loss contingencies that fall
below the established data thresholds;
this is relevant as many institutions
currently budget for low severity, high
frequency events that are more likely to
fall below most institutions’ thresholds.
An institution’s analytical framework
complements but does not substitute for
prudent controls. Rather, with improved
risk measurement, institutions are
finding that they can make betterinformed strategic decisions regarding
enhancements to controls and
processes, the desired scale and scope of
the operations, and how insurance and

E:\FR\FM\04AUN2.SGM

04AUN2

45986

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

other risk mitigation tools can be used
to offset operational risk exposure.
B. Accounting for Dependence
Supervisory Standards
S 29. Management must document
how its chosen analytical framework
accounts for dependence (e.g.,
correlations) among operational losses
across and within business lines. The
institution must demonstrate that its
explicit and embedded dependence
assumptions are appropriate, and where
dependence assumptions are uncertain,
the institution must use conservative
estimates.
Management must document how its
chosen analytical framework accounts
for dependence (e.g., correlation)
between operational losses across and
within business lines. The issue of
dependence is closely related to the
choice between a bottom-up or a topdown modeling approach. Under a
bottom-up approach, explicit
assumptions regarding cross-event
dependence are required to estimate
operational risk exposure at the firmwide level. Management must
demonstrate that these assumptions are
appropriate and reflect the institution’s
current environment. If the dependence
assumptions are uncertain, the
institution must choose conservative
estimates. In so doing, the institution
should consider the possibility that
cross-event dependence may not be
constant, and may increase during stress
environments.
Under a top-down approach, an
explicit assumption regarding
dependence is not required. However, a
parametric distribution for loss severity
may be more difficult to specify under
the top-down approach, as it is a
statistical mixture of (potentially)
heterogeneous business line and event
type distributions. Institutions must
carefully consider the conditions
necessary for the validity of top-down
approaches, and whether these
conditions are met in their particular
circumstances. Similar to bottom-up
approaches, institutions using top-down
approaches must ensure that implicit
dependence assumptions are
appropriate and reflect the institution’s
current environment. If historic
dependence assumptions embedded in
top-down approaches are uncertain, the
institution must be conservative and
implement a qualitative adjustment to
the analysis.
IX. Risk Mitigation
Supervisory Standards
S 30. Institutions may reduce their
operational risk exposure results by no

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

more than 20% to reflect the impact of
risk mitigants. Institutions must
demonstrate that mitigation products
are sufficiently capital-like to warrant
inclusion in the adjustment to the
operational risk exposure.
There are many mechanisms to
manage operational risk, including risk
transfer through risk mitigation
products. Because risk mitigation can be
an important element in limiting or
reducing operational risk exposure in an
institution, an adjustment is being
permitted that will directly impact the
amount of regulatory capital that is held
for operational risk. The adjustment is
limited to 20% of the overall
operational risk exposure result
determined by the institution using its
loss data, qualitative factors, and
quantitative framework.
Currently, the primary risk mitigant
used for operational risk is insurance.
There has been discussion that some
securities products may be developed to
provide risk mitigation benefits;
however, to date, no specific products
have emerged that have characteristics
sufficient to be considered capitalreplacement for operational risk. As a
result, securities products and other
capital market instruments may not be
factored in to the regulatory capital risk
mitigation adjustment at this time.
For an institution that wishes to
adjust its regulatory capital requirement
as a result of the risk mitigating impact
of insurance, management must
demonstrate that the insurance policy is
sufficiently capital-like to provide the
cushion that is necessary. A product
that would fall in this category must
have the following characteristics:
• The policy is provided through a
third party 13 that has a minimum
claims paying ability rating of A; 14
• The policy has an initial term of
one year; 15
• The policy has no exclusions or
limitations based upon regulatory action
or for the receiver or liquidator of a
failed bank;
13 Where operational risk is transferred to a
captive or an affiliated insurer such that risk is
retained within the group structure, recognition of
such risk transfer will only be allowed for
regulatory capital purposes where the risk has been
transferred to a third party (e.g., an unaffiliated
reinsurer) that meets the standards set forth in this
section.
14 Rating agencies may use slightly different
rating scales.For the purpose of this supervisory
guidance, the insurer must have a rating that is at
least the equivalent of A under Standard and Poor’s
Insurer Financial Strength Ratings or an A2 under
Moody’s Insurance Financial Strength Ratings.
15 Institutions must decrease the amount of the
adjustment if the remaining term is less than one
year. The institution must have a clear policy in
place that links the remaining term to the
adjustment factor.

PO 00000

Frm 00038

Fmt 4701

Sfmt 4703

• The policy has clear cancellation
and non-renewal notice periods; and
• The policy coverage has been
explicitly mapped to actual operational
risk exposure of the institution.
Insurance policies that meet these
standards may be incorporated into an
institution’s adjustment for risk
mitigation. An institution should be
conservative in its recognition of such
policies, for example, the institution
must also demonstrate that insurance
policies used as the basis for the
adjustment have a history of timely
payouts. If claims have not been paid on
a timely basis, the institution must
exclude that policy from the operational
risk capital adjustment. In addition, the
institution must be able to show that the
policy would actually be used in the
event of a loss situation; that is, the
deductible may not be set so high that
no loss would ever conceivably exceed
the deductible threshold.
The Agencies will not specify how
institutions should calculate the risk
mitigation adjustment. Nevertheless,
institutions are expected to use
conservative assumptions when
calculating adjustments. An institution
should discount (i.e., apply its own
estimates of haircuts) the impact of
insurance coverage to take into account
factors, which may limit the likelihood
or size of claims payouts. Among these
factors are the remaining terms of a
policy, especially when it is less than a
year, the willingness and ability of the
insurer to pay on a claim in a timely
manner, the legal risk that a claim may
be disputed, and the possibility that a
policy can be cancelled before the
contractual expiration.
X. Data Maintenance
Supervisory Standards
S 31. Institutions using the AMA
approach for regulatory capital purposes
must use advanced data management
practices to produce credible and
reliable operational risk estimates.
Data maintenance is a critical factor in
an institution’s operational risk
framework. Institutions with advanced
data management practices should be
able to track operational risk loss events
from initial discovery through final
resolution. These institutions should
also be able to make appropriate
adjustments to the data and use the data
to identify trends, track problem areas,
and identify areas of future risk. Such
data should include not only
operational risk loss event information,
but also information on risk
assessments, which are factored into the
operational risk exposure calculation. In
general, institutions using the AMA

E:\FR\FM\04AUN2.SGM

04AUN2

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices
should have the same data maintenance
standards for operational risk as those
set forth for A–IRB institutions under
the credit risk guidance.
Operational risk data elements
captured by the institution must be of
sufficient depth, scope, and reliability
to:
• Track and identify operational risk
loss events across all business lines,
including when a loss event impacts
multiple business lines.
• Calculate capital ratios based on
operational risk exposure results. The
institution must also be able to factor in
adjustments related to risk mitigation,
correlations, and risk assessments.
• Produce internal and public reports
on operational risk measurement and
management results, including trends
revealed by loss data and/or risk
assessments. The institution must also
have sufficient data to produce
exception reports for management.
• Support risk management activities.
The data warehouse 16 16 must
contain the key data elements needed
for operational risk measurement,
management, and verification. The
precise data elements may vary by
institution and also among business
lines within an institution. An
important element of ensuring
consistent reporting of the data elements
is to develop comprehensive definitions
for each data element used by the
institution for reporting operational risk
loss events or for the risk assessment
inputs. The data must be stored in an
electronic format to allow for timely
retrieval for analysis, verification and
testing of the operational risk
framework, and required disclosures.
Management will need to identify
those responsible for maintaining the
data warehouse. In particular, policies
and processes will need to be developed
for delivering, storing, retaining, and
updating the data warehouse. Policies
and procedures must also cover the edit
checks for data input functions, as well
as the requirements for the testing and
verification function to verify data
integrity. Like other areas of the
operational risk framework, it is critical
that management ensure accountability
for ongoing data maintenance, as this
will impact operational risk
management and measurement efforts.
XI. Testing and Verification
Supervisory Standards
S 32. The institution must test and
verify the accuracy and appropriateness
16 In this document, the terms ‘‘database’’ and
‘‘data warehouse’’ are used interchangeably to refer
to a collection of data arranged for easy retrieval
using computer technology.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

of the operational risk framework and
results.
S 33. Testing and verification must be
done independently of the firm-wide
operational risk management function
and the institution’s lines of business.
The operational risk framework must
provide for regular and independent
testing and verification of operational
risk management policies, processes and
measurement systems, as well as
operational risk data capture systems.
For most institutions, operational risk
verification and testing will primarily be
done by the audit function. Internal and
external audits can provide an
independent assessment of the quality
and effectiveness of the control systems’
design and performance. However,
institutions may use other independent
internal units (e.g. quality assurance) or
third parties. The testing and
verification function, whether internally
or externally performed, should be
staffed by qualified individuals who are
independent from the firm-wide
operational risk management function
and the institution’s lines of business.
The verification of the operational
risk measurement system should
include the testing of:
• Key operational risk processes and
systems;
• Data feeds and processes associated
with the operational risk measurement
system;
• Adjustments to empirical
operational risk capital estimates,
including operational risk exposure;
• Periodic certification of operational
risk models used and their underlying
assumptions; and
• Assumptions underlying
operational risk exposure, data decision
models, and operational risk capital
charge.
The operational risk reporting
processes should be periodically
reviewed for scope and effectiveness.
The institution should have
independent verification processes to
ensure the timeliness, accuracy, and
comprehensiveness of operational risk
reporting systems, both at the firm-wide
and the line of business levels.
Independent verification and testing
should be done to ensure the integrity
and applicability of the operational risk
framework, operational risk exposure/
loss data, and the underlying
assumptions driving the regulatory
capital measurement process.
Appropriate reports, summarizing
operational risk verification and testing
findings for both the independent firmwide risk management function and
lines of business should be provided to
appropriate management and the board

PO 00000

Frm 00039

Fmt 4701

Sfmt 4703

45987

of directors or a designated board
committee.
Appendix A: Supervisory Standards for
the AMA
S 1. The institution’s operational risk
framework must include an independent
firm-wide operational risk management
function, line of business management
oversight, and independent testing and
verification functions.
S 2. The board of directors must oversee
the development of the firm-wide operational
risk framework, as well as major changes to
the framework. Management roles and
accountability must be clearly established.
S 3. The board of directors and
management must ensure that appropriate
resources are allocated to support the
operational risk framework.
S 4. The institution must have an
independent operational risk management
function that is responsible for overseeing the
operational risk framework at the firm level
to ensure the development and consistent
application of operational risk policies,
processes, and procedures throughout the
institution.
S 5. The firm-wide operational risk
management function must ensure
appropriate reporting of operational risk
exposures and loss data to the board of
directors and senior management.
S 6. Line of business management is
responsible for the day-to-day management of
operational risk within each business unit.
S 7. Line of business management must
ensure that internal controls and practices
within their line of business are consistent
with firm-wide policies and procedures to
support the management and measurement of
the institution’s operational risk.
S 8. The institution must have policies
and procedures that clearly describe the
major elements of the operational risk
management framework, including
identifying, measuring, monitoring, and
controlling operational risk.
S 9. Operational risk management reports
must address both firm-wide and line of
business results. These reports must
summarize operational risk exposure, loss
experience, relevant business environment
and internal control assessments, and must
be produced no less often than quarterly.
S 10. Operational risk reports must also
be provided periodically to senior
management and the board of directors,
summarizing relevant firm-wide operational
risk information.
S 11. An institution’s internal control
structure must meet or exceed minimum
regulatory standards established by the
Agencies.
S 12. The institution must demonstrate
that it has appropriate internal loss event
data, relevant external loss event data,
assessments of business environment and
internal controls factors, and results from
scenario analysis to support its operational
risk management and measurement
framework.
S 13. The institution must include the
regulatory definition of operational risk as
the baseline for capturing the elements of the

E:\FR\FM\04AUN2.SGM

04AUN2

45988

Federal Register / Vol. 68, No. 149 / Monday, August 4, 2003 / Notices

AMA framework and determining its
operational risk exposure.
S 14. The institution must have clear
standards for the collection and modification
of the elements of the operational risk AMA
framework.
S 15. The institution must have at least
five years of internal operational risk loss
data 17 captured across all material business
lines, events, product types, and geographic
locations.
S 16. The institution must be able to map
internal operational risk losses to the seven
loss-event type categories.
S 17. The institution must have a policy
that identifies when an operational risk loss
becomes a loss event and must be added to
the loss event database. The policy must
provide for consistent treatment across the
institution.
S 18. The institution must establish
appropriate operational risk data thresholds.
S 19. Losses that have any characteristics
of credit risk, including fraud-related credit
losses, must be treated as credit risk for
regulatory capital purposes. The institution
must have a clear policy that allows for the
consistent treatment of loss event
classifications (e.g., credit, market, or
operational risk) across the organization.
S 20. The institution must have policies
and procedures that provide for the use of
external loss data in the operational risk
framework.
S 21. Management must systematically
review external data to ensure an
understanding of industry experience.
S 22. The institution must have a system
to identify and assess business environment
and internal control factors.
S 23. Management must periodically
compare the results of their business
environment and internal control factor
17 With supervisory approval, a shorter initial
historical observation period is acceptable for banks
newly authorized to use an AMA methodology.

VerDate jul<14>2003

19:57 Aug 01, 2003

Jkt 200001

assessments against actual operational risk
loss experience.
S 24. Management must have policies
and procedures that identify how scenario
analysis will be incorporated into the
operational risk framework.
S 25. The institution must have a
comprehensive operational risk analytical
framework that provides an estimate of the
institution’s operational risk exposure, which
is the aggregate operational loss that it faces
over a one-year period at a soundness
standard consistent with a 99.9 per cent
confidence level.
S 26. Management must document the
rationale for all assumptions underpinning
its chosen analytical framework, including
the choice of inputs, distributional
assumptions, and the weighting across
qualitative and quantitative elements.
Management must also document and justify
any subsequent changes to these
assumptions.
S 27. The institution’s operational risk
analytical framework must use a combination
of internal operational loss event data,
relevant external operational loss event data,
business environment and internal control
factor assessments, and scenario analysis.
The institution must combine these elements
in a manner that most effectively enables it
to quantify its operational risk exposure. The
institution can choose the analytical
framework that is most appropriate to its
business model.
S 28. The institution’s capital
requirement for operational risk will be the
sum of expected and unexpected losses
unless the institution can demonstrate,
consistent with supervisory standards, the
expected loss offset.
S 29. Management must document how
its chosen analytical framework accounts for
dependence (e.g., correlations) among
operational losses across and within business
lines. The institution must demonstrate that
its explicit and embedded dependence

PO 00000

Frm 00040

Fmt 4701

Sfmt 4703

assumptions are appropriate, and where
dependence assumptions are uncertain, the
institution must use conservative estimates.
S 30. Institutions may reduce their
operational risk exposure results by no more
than 20% to reflect the impact of risk
mitigants. Institutions must demonstrate that
mitigation products are sufficiently capitallike to warrant inclusion in the adjustment to
the operational risk exposure.
S 31. Institutions using the AMA
approach for regulatory capital purposes
must use advanced data management
practices to produce credible and reliable
operational risk estimates.
S 32. The institution must test and verify
the accuracy and appropriateness of the
operational risk framework and results.
S 33. Testing and verification must be
done independently of the firm-wide
operational risk management function and
the institution’s lines of business.
Dated: July 17, 2003.
John D. Hawke, Jr.,
Comptroller of the Currency.
By order of the Board of Governors of the
Federal Reserve System, July 21, 2003.
Jennifer J. Johnson,
Secretary of the Board.
Dated at Washington, DC, this 11th day of
July, 2003.
By order of the Board of Directors.
Federal Deposit Insurance Corporation.
Robert E. Feldman,
Executive Secretary.
Dated: July 18, 2003.
By the Office of Thrift Supervision.
James E. Gilleran,
Director.
[FR Doc. 03–18976 Filed 8–1–03; 8:45 am]
BILLING CODE 4810–33–P; 6210–01–P; 6714–01–P;
6720–01–P

E:\FR\FM\04AUN2.SGM

04AUN2