[Federal Register Volume 90, Number 220 (Tuesday, November 18, 2025)]
[Proposed Rules]
[Pages 51856-51953]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2025-20211]
[[Page 51855]]
Vol. 90
Tuesday,
No. 220
November 18, 2025
Part III
Federal Reserve System
-----------------------------------------------------------------------
12 CFR Parts 225, 238, and 252
Enhanced Transparency and Public Accountability of the Supervisory
Stress Test Models and Scenarios; Modifications to the Capital Planning
and Stress Capital Buffer Requirement Rule, Enhanced Prudential
Standards Rule, and Regulation LL; Proposed Rule
Federal Register / Vol. 90, No. 220 / Tuesday, November 18, 2025 /
Proposed Rules
[[Page 51856]]
-----------------------------------------------------------------------
FEDERAL RESERVE SYSTEM
12 CFR Parts 225, 238, and 252
[Regulations Y, LL, and YY; Docket No. R-1873]
RIN 7100-AH05
Enhanced Transparency and Public Accountability of the
Supervisory Stress Test Models and Scenarios; Modifications to the
Capital Planning and Stress Capital Buffer Requirement Rule, Enhanced
Prudential Standards Rule, and Regulation LL
AGENCY: Board of Governors of the Federal Reserve System (Board).
ACTION: Notice of Proposed Rulemaking.
-----------------------------------------------------------------------
SUMMARY: The Board is inviting public comment on the models used to
conduct the Board's supervisory stress test, changes to those models to
be implemented in the 2026 stress test, and proposed changes to enhance
the transparency and public accountability of the Board's stress
testing framework (the proposal). The proposal would amend the Policy
Statement on the Scenario Design Framework for Stress Testing,
including to implement guides for additional scenario variables, and
the Stress Testing Policy Statement. The proposal would also codify an
enhanced disclosure process under which the Board would annually
publish comprehensive documentation on the stress test models, invite
public comment on any material changes that the Board seeks to make to
those models, and annually publish the stress test scenarios for
comment. Lastly, the proposal would make changes to the FR Y-14A/Q/M to
remove items that are no longer needed to conduct the supervisory
stress test and to collect additional data to support the stress test
models and improve risk capture.
DATES: Comments must be received on or before January 22, 2026.
ADDRESSES: You may submit comments, identified by Docket No. R-1873 and
RIN 7100-AH05, by any of the following methods:
Agency website: https://www.federalreserve.gov/apps/proposals/. Follow the instructions for submitting comments, including
attachments. Preferred Method.
Mail: Benjamin W. McDonough, Deputy Secretary, Board of
Governors of the Federal Reserve System, 20th Street and Constitution
Avenue NW, Washington, DC 20551.
Hand Delivery/Courier: Same as mailing address.
Other Means: [email protected]. You must include the
docket number in the subject line of the message.
Comments received are subject to public disclosure. In general,
comments received will be made available on the Board's website at
https://www.federalreserve.gov/apps/proposals/ without change and will
not be modified to remove personal or business information including
confidential, contact, or other identifying information. Comments
should not include any information such as confidential information
that would be not appropriate for public disclosure. Public comments
may also be viewed electronically or in person in Room M-4365A, 2001 C
St. NW, Washington, DC 20551, between 9 a.m. and 5 p.m. during Federal
business weekdays.
FOR FURTHER INFORMATION CONTACT: Doriana Ruffino, Assistant Director,
(202) 452-5235, Hillel Kipnis, Assistant Director, (202) 452-2924, John
Simone, Lead Financial Institution Policy Analyst, (202) 245-4256, Ben
Ranish, Principal Economist, (202) 973-6964, Nathan Palmer, Senior
Economist, (202) 785-6089, and Theo Pistner, Financial Institution and
Policy Analyst II, (202) 941-1825, Division of Supervision and
Regulation; William Bassett, Senior Associate Director, (202) 736-5644,
Bora Durdu, Deputy Associate Director, (202) 452-3755, Elena
Afanasyeva, Principal Economist, (202) 736-1971, Levent Altinoglu,
Principal Economist, (202) 721-4503, and Sam Jerow, Senior Financial
Analyst, (202) 245-4299, Division of Financial Stability; Asad Kudiya,
Associate General Counsel, (202) 360-6887, Julie Anthony, Senior
Special Counsel, (202) 658-9400, Jonah Kind, Senior Counsel, (202) 452-
2045, Brian Kesten, Senior Counsel, (202) 843-4079, Katherine Di
Lucido, Senior Attorney, (202) 253-5994, Legal Division. Board of
Governors of the Federal Reserve System, 20th Street and Constitution
Avenue NW, Washington, DC 20551. For users of TDD-TYY, please call 711
from any telephone, anywhere in the United States.
SUPPLEMENTARY INFORMATION:
Table of Contents
I. Introduction
II. Background on Stress Testing Framework, Stress Test Models, and
Scenario Design Framework
A. Stress Testing Framework
B. Prior Supervisory Stress Disclosures and Policy Statements
C. Supervisory Stress Test Modeling Framework
D. Stress Test Models
E. Summary of the Proposal
F. Purpose of the Proposal
III. Overview of the Stress Test Modeling Framework
A. Supervisory Stress Test Models
B. Supervisory Stress Test Scenarios
C. Data Used in Stress Testing
IV. Enhanced Disclosure Process
A. Annual Disclosure of Models
B. Model Changes
C. Material Model Changes
D. Annual Disclosure of Scenarios
E. Stress Capital Buffer Requirement Reconsideration Process
V. Revisions to the Stress Testing Policy Statement
A. Future Supervisory Stress Test Results Disclosures
B. Other Revisions to the Stress Testing Policy Statement
VI. Other Revisions to the Stress Testing and Capital Plan Rules
A. Stress Test Jump-Off Date Change
B. Global Market Shock Date
C. Amendment to the Dividend Add-On Component Calculation
VII. Revisions to the FR Y-14A/Q/M
VIII. Proposed Changes to the Stress Test Modeling Framework
A. Proposed Changes to Stress Test Models
B. Analysis of Proposed Model Changes
IX. Proposed Changes to the Scenario Design Policy Statement
A. Changes to the Background and Overview and Scope Sections
B. Changes to the Content of the Stress Test Scenarios Section
C. Approach for Formulating Macroeconomic Assumptions in the
Baseline Scenario
D. Scenario Narrative: Refinement to the Recession Approach
E. Changes to Construction of Certain Variables in the Severely
Adverse Scenario
F. Scenario Design Principles Derived From Stress Testing
Literature: Severity, Credibility, and Procyclicality
G. Description of Variable Guides in the Severely Adverse
Scenario
H. Global Market Shock
X. Economic Analysis
XI. Administrative Law Matters
A. Paperwork Reduction Act Analysis
B. Regulatory Flexibility Act Analysis
C. Plain Language
D. Providing Accountability Through Transparency Act of 2023
I. Introduction
In December 2024, the Board announced that it would propose
significant changes to improve the transparency of the supervisory
stress test and reduce the volatility of resulting capital
requirements.\1\ The Board noted it planned to propose changes to
disclose and seek public comment on the models that determine the
[[Page 51857]]
hypothetical losses and revenue of banks under stress and ensure that
the public can comment on the hypothetical scenarios used annually for
the test, before the scenarios are finalized. With this proposal, the
Board is inviting public comment on the comprehensive model
documentation for the 2026 stress test, as well as proposed changes to
the models relative to the 2025 stress test. The comprehensive model
documentation is available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. The Board is inviting comment
on the proposed scenarios for the 2026 stress test through a separate
notice.
---------------------------------------------------------------------------
\1\ See Board, Press Release (Dec. 23, 2024), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20241223a.htm.
In February 2025, the Board reiterated its previous announcement
that it would begin the public comment process on changes to the
supervisory stress test. See Board, Press Release (Feb. 5, 2025),
https://www.federalreserve.gov/newsevents/pressreleases/bcreg20250205a.htm.
---------------------------------------------------------------------------
This proposal seeks to improve the transparency and public
accountability of the supervisory stress test, while ensuring that the
stress test remains an effective tool for understanding and assessing
risk and retaining appropriate risk sensitivity and risk capture in
capital requirements.
The Board periodically reviews its regulations, including
transparency efforts surrounding its regulations, to ensure they
continue to achieve their goals in an effective and efficient manner.
In addition to the changes discussed herein, the Board is also
considering the effectiveness of its regulatory capital and capital
planning requirements for large firms to ensure they remain cohesive
and effective, maintain the resilience of the banking sector, and
minimize any unnecessary burden. If appropriate, the Board will make
changes to its rules through the public notice and comment process.
Question 1: The Board seeks comment on all aspects of the proposal.
What, if any, other elements of the supervisory stress test framework
should the Board consider amending to improve the transparency, public
accountability, and effectiveness of the supervisory stress test? For
example, the Board could instead transliterate the models used to
conduct the stress test and codify these transliterations in its
regulations. What would be the advantages and disadvantages of this
approach or other approaches the Board could consider?
II. Background on Stress Testing Framework, Stress Test Models, and
Scenario Design Framework
A. Stress Testing Framework
Congress enacted the Dodd-Frank Wall Street Reform and Consumer
Protection Act (Dodd-Frank Act) in the wake of the 2007-09 financial
crisis.\2\ Section 165 of the Dodd-Frank Act, as amended by section 401
of the Economic Growth, Regulatory Relief, and Consumer Protection
Act,\3\ requires the Board to establish enhanced prudential standards
for nonbank financial companies supervised by the Board and bank
holding companies with $250 billion or more in total consolidated
assets.\4\ The purpose of these enhanced prudential standards is to
prevent or mitigate risks to the financial stability of the United
States that could arise from the material financial distress or
failure, or ongoing activities, of large, interconnected financial
institutions.
---------------------------------------------------------------------------
\2\ Dodd-Frank Wall Street Reform and Consumer Protection Act,
Public Law 111-203, 124 Stat. 1376 (2010).
\3\ Economic Growth, Regulatory Relief, and Consumer Protection
Act, Public Law 115-174, 132 Stat. 1296 (2018).
\4\ See 12 U.S.C. 5365(a). In addition, the International
Lending Supervision Act of 1983 provides the Board with broad
discretionary authority to set minimum capital levels for state
member banks and certain affiliates of insured depository
institutions, including holding companies, supervised by the Board.
See 12 U.S.C. 3902(1); 3907(a); 3909(a). Under section 5(b) of the
Bank Holding Company Act of 1956 (Bank Holding Company Act), the
Board may issue such regulations and orders relating to capital
requirements of bank holding companies as may be necessary for the
Board to carry out the purposes of the Bank Holding Company Act. 12
U.S.C. 1844(b). Foreign banking organizations with a U.S. branch,
agency, or commercial lending company subsidiary are made subject by
the International Banking Act of 1978 (International Banking Act) to
the provisions of the Bank Holding Company Act in the same manner as
bank holding companies, see 12 U.S.C. 3106; therefore, the Board is
also authorized under section 5(b) of the Bank Holding Company Act
to impose these requirements on those foreign banking organizations,
including on their U.S. operations. Similarly, with regard to
savings and loan holding companies, section 10(g) of the Home
Owners' Loan Act authorizes the Board to issue such regulations and
orders relating to capital requirements as the Board deems necessary
and appropriate to carry out the purposes of the Home Owners' Loan
Act. See 12 U.S.C. 1467a(g)(1).
---------------------------------------------------------------------------
Section 165(i)(1) of the Dodd-Frank Act requires the Board to
conduct an annual supervisory stress test of nonbank financial
companies supervised by the Board and bank holding companies with $250
billion or more in total consolidated assets to evaluate whether the
firm has the capital, on a total consolidated basis, necessary to
absorb losses as a result of adverse economic conditions.\5\ Section
401(e) of the Economic Growth, Regulatory Relief, and Consumer
Protection Act requires the Board to conduct periodic stress tests for
bank holding companies with total consolidated assets between $100
billion and $250 billion.\6\ Section 165(i)(1) of the Dodd-Frank Act
requires the Board to publish a summary of the supervisory stress test
results.\7\ In 2012, the Board adopted a final rule implementing the
stress test requirements established in the Dodd-Frank Act.\8\
---------------------------------------------------------------------------
\5\ 12 U.S.C. 5365(i)(1).
\6\ 12 U.S.C. 5365 note (Supervisory Stress Test).
\7\ 12 U.S.C. 5365(i)(1)(B)(v).
\8\ See 77 FR 62378 (Oct. 12, 2012).
---------------------------------------------------------------------------
The Dodd-Frank Act also requires bank holding companies with $250
billion or more in total consolidated assets, as well as nonbank
financial companies supervised by the Board, to conduct company-run
stress tests on a periodic basis.\9\ Under the Board's rules, firms
subject to Category I, II, or III standards must conduct company-run
stress tests.\10\ Company-run stress tests provide forward-looking
information to supervisors to assist in their overall assessments of a
firm's capital adequacy, help to better identify downside risks and the
potential impact of adverse outcomes on the firm`s capital adequacy,
and assist in achieving the financial stability goals of the Dodd-Frank
Act. Further, the company-run stress tests help improve firms' stress
testing practices with respect to their own internal assessments of
capital adequacy and overall capital planning.
---------------------------------------------------------------------------
\9\ 12 U.S.C. 5365(i)(2).
\10\ See 84 FR 59032 (Nov. 1, 2019); 12 CFR 238.142; 12 CFR
252.53. State member banks with average total consolidated assets of
greater than $250 billion must also conduct company-run stress
tests. 12 CFR 252.13.
---------------------------------------------------------------------------
Each June, the Board publishes the results of its annual
supervisory stress test, including each firm's projected capital
ratios, pre-tax net income, losses, revenues, and expenses, under
hypothetical, severely adverse economic and financial conditions.\11\
These disclosures provide the public with valuable information about
each firm's financial condition and the ability of
[[Page 51858]]
each firm to absorb losses considering a stressful economic
environment.
---------------------------------------------------------------------------
\11\ A firm subject to Category I through III standards must
participate in the supervisory stress test every year, while a firm
subject to Category IV standards is generally required to
participate only every other year. See 12 CFR 217.2; 12 CFR 238.10;
12 CFR 252.5; 84 FR 59032 (Nov. 1, 2019). In 2019, the Board adopted
rules establishing four categories of prudential standards for U.S.
banking organizations with total consolidated assets of $100 billion
or more and foreign banking organizations with combined U.S. assets
of $100 billion or more. See 12 CFR 217.2; 12 CFR 238.10; 12 CFR
252.5; 84 FR 59032 (Nov. 1, 2019). Category I standards apply to
U.S. GSIBs and their depository institution subsidiaries. Category
II standards apply to banking organizations with at least $700
billion in total consolidated assets or at least $75 billion in
cross-jurisdictional activity and their depository institution
subsidiaries. Category III standards apply to banking organizations
with total consolidated assets of at least $250 billion or at least
$75 billion in weighted short-term wholesale funding, nonbank
assets, or off-balance sheet exposure and their depository
institution subsidiaries. Category IV standards apply to banking
organizations with total consolidated assets of at least $100
billion that do not meet the thresholds for a higher category and
their depository institution subsidiaries.
---------------------------------------------------------------------------
Following the 2007-09 financial crisis, the Board also made changes
to its capital rule to address weaknesses observed during the
crisis.\12\ These changes included the establishment of a minimum
common equity tier 1 capital requirement and a fixed capital
conservation buffer equal to 2.5 percent of risk-weighted assets.\13\
Large firms also became subject to a countercyclical capital buffer
requirement, and the largest and most systemically important firms--
global systemically important bank holding companies, or GSIBs--became
subject to an additional capital buffer based on a measure of their
systemic risk, the GSIB surcharge.\14\ In 2020, the Board adopted the
stress capital buffer requirement for certain firms.\15\ Because a
firm's stress capital buffer requirement is informed by the firm's
performance under the hypothetical economic conditions modeled by the
supervisory stress test, each firm's stress capital buffer requirement
is tailored to its risk profile.
---------------------------------------------------------------------------
\12\ See generally 12 CFR part 217.
\13\ See 78 FR 62018 (Oct. 11, 2013); 12 CFR 217.11.
\14\ See 80 FR 49082 (Aug. 14, 2015).
\15\ In 2020, the Board finalized a rule to integrate
supervisory stress test results into the capital framework, through
the stress capital buffer requirement. See 85 FR 15576 (Mar. 18,
2020). The stress capital buffer requirement is calculated as the
difference between a firm's starting and lowest projected common
equity tier 1 capital ratio under the severely adverse scenario in
the supervisory stress test plus four quarters of planned common
stock dividends, expressed as a percentage of risk-weighted assets.
See 12 CFR 225.8(f); 12 CFR 238.170(f). The stress capital buffer
requirement framework generally applies to firms with $100 billion
or more in total consolidated assets.
---------------------------------------------------------------------------
Supervisory stress testing and stronger capital requirements have
significantly improved the resilience of the U.S. banking system. Since
2009, the common equity capital ratios of firms subject to the test
have more than doubled, with common equity capital of such firms
increasing by over $1 trillion.\16\ Since 2020, the supervisory stress
test results have also informed a firm's stress capital buffer
requirement. Greater transparency would allow firms to better
understand the capital requirements associated with investment and
expansion of different business lines and would facilitate more
effective long-term capital planning. This, in turn, could enhance
firms' ability to supply credit to households and businesses,
ultimately supporting economic growth and financial stability.
---------------------------------------------------------------------------
\16\ Based on FR Y-9C (Consolidated Financial Statements for
Holding Companies) filings.
---------------------------------------------------------------------------
B. Prior Supervisory Stress Disclosures and Policy Statements
In addition to the annual stress test results disclosure, the Board
has historically published some information about the supervisory
stress test scenarios and models.
Scenarios
The Board's stress test rules provide that the Board will notify
firms, by no later than February 15 of each year, of the scenarios that
the Board will apply to conduct its annual supervisory stress test and
that firms must use to conduct their company-run stress tests.\17\ The
Board also provides a narrative description of the scenarios no later
than February 15 of each calendar year.\18\
---------------------------------------------------------------------------
\17\ See 12 CFR 238.132(b); 12 CFR 238.143(b); 12 CFR 252.14(b);
12 CFR 252.44(b); 12 CFR 252.54(b).
\18\ See, e.g., Board, 2025 Stress Test Scenarios (Feb. 2025),
https://www.federalreserve.gov/publications/files/2025-stress-test-scenarios-20250205.pdf.
---------------------------------------------------------------------------
In 2013, the Board increased the transparency of the scenarios by
finalizing the Policy Statement on the Scenario Design Framework for
Stress Testing (Scenario Design Policy Statement), which articulated
the Board's approach to scenario design for the supervisory and
company-run stress tests, outlining the characteristics of the stress
test scenarios, and explaining the considerations and procedures that
underlie the formulation of these scenarios.\19\ The Scenario Design
Policy Statement also described the baseline and severely adverse
scenarios, the Board's approach for developing these two macroeconomic
scenarios, and the approach for developing any additional components of
the stress test scenarios. The Scenario Design Policy Statement
explained that the severely adverse scenario is designed to reflect
conditions that have characterized post-war U.S. recessions (the
recession approach). Historically, recessions have typically featured
increases in the unemployment rate, contractions in aggregate incomes
and economic activity, and declines in inflation and interest rates.
---------------------------------------------------------------------------
\19\ 12 CFR part 252, Appendix A.
---------------------------------------------------------------------------
In the 2013 Scenario Design Policy Statement, the Board explained
that, in light of the typical co-movement of measures of economic
activity during economic downturns, such as the unemployment rate and
gross domestic product, the Board would first specify a path for the
unemployment rate and then develops paths for other measures of
activity broadly consistent with the course of the unemployment rate in
developing the severely adverse scenario. The 2013 Scenario Design
Policy Statement also stated that economic variables included in the
scenarios may change over time, and that the Board may augment the
recession approach with certain salient risks, which would involve
incorporating features that address aspects of the current economic or
financial market environment that represent higher-than-normal risks to
the condition of the banking system.
In 2019, the Board updated the Scenario Design Policy Statement,
which increased the transparency and predictability of the scenarios by
allowing for a smaller-than-usual increase in unemployment if the
stress test were to occur during an economic downturn, a change that
would pass through to reduced severity of other key scenario variables
due to the deference given to historical correlations. The 2019 update
also introduced a formula with countercyclical features to guide the
evolution of the ratio of housing prices to disposable income in the
scenario, which provided more predictability in the way that the stress
test would treat business lines affected by changes in house prices.
However, the Board believes that the design of scenarios could be made
more transparent and predictable by providing additional guides for
certain macroeconomic variables, and by disclosing additional detailed
information on the methodology used to create the global market shock
component of the severely adverse scenario, as described below.
a. Trading and Counterparty Components
For a subset of firms, the severely adverse scenario also includes
two additional components: the global market shock component and the
largest counterparty default component.\20\ The global market shock
component is a set of hypothetical shocks to a large set of risk
factors reflecting general market distress and heightened uncertainty.
A firm with significant trading activity must consider the global
market shock component as part of its severely adverse scenario and
recognize associated losses in the first quarter of the projection
horizon.\21\ The global
[[Page 51859]]
market shock component is applied to asset positions held by the firms
on a given as-of date.\22\ In addition, for certain large and highly
interconnected firms, the same global market shock component is applied
to counterparty exposures under the largest counterparty default
component.\23\ The largest counterparty default component is intended
to assess the potential losses and capital impact associated with the
default of the largest counterparty of each applicable firm, and the
as-of date aligns with that of the global market shock component.
---------------------------------------------------------------------------
\20\ See 12 CFR 238.143(b)(2)(i); 12 CFR 252.54(b)(2)(i). For
more information on the scenarios and components, see Board, 2025
Stress Test Scenarios (Feb. 2025), https://www.federalreserve.gov/publications/files/2025-stress-test-scenarios-20250205.pdf.
\21\ The global market shock component applies to firms subject
to Category I, II, and III standards that have aggregate trading
assets and liabilities of $50 billion or more, or trading assets and
liabilities equal to or greater than 10 percent of total
consolidated assets. See 12 CFR 238.143(b)(2)(i); 12 CFR
252.54(b)(2)(i).
\22\ Under the Board's current stress test rules, the global
market shock as-of date must occur between October 1 and March 1.
See 12 CFR 238.143(b)(2)(i); 12 CFR 252.14(b)(2)(i); 12 CFR
252.54(b)(2)(i).
\23\ The largest counterparty default component generally
applies to all firms subject to the global market shock component,
as well as firms with substantial processing and custodial
operations.
---------------------------------------------------------------------------
The design and specification of the global market shock component
differs from the design and specification of the severely adverse
scenario in several respects. First, in alignment with U.S. generally
accepted accounting principles (U.S. GAAP), profits and losses from
trading and counterparty credit positions are measured in mark-to-
market accounting terms in the global market shock, while revenues and
losses from traditional banking activities, as generated under
macroeconomic scenarios, are generally measured using the accrual
accounting method. Second, the timing of loss recognition differs
between the global market shock and the severely adverse macroeconomic
scenario. The global market shock affects the mark-to-market value of
trading positions and counterparty credit losses in the first quarter
of the severely adverse scenario. This timing is based on an
observation that market dislocations can happen rapidly and
unpredictably at any time under stressed conditions. In addition, the
severely adverse scenario is applied as of December 31 of each year
(the jump-off date), whereas the global market shock as-of date changes
every year (within the window specified in the Board's stress test
rules) and does not necessarily coincide with the year-end. This timing
is also based on a scenario assumption that market dislocations can
happen rapidly and unpredictably at any time during the scenario
horizon. Recognizing the global market shock in the first quarter helps
ensure that potential losses from trading and counterparty exposures
are incorporated into firms' capital ratios in each quarter of the
severely adverse scenario.
Models
Prior to 2019, the annual stress test results disclosure document
contained an appendix describing the Board's supervisory stress test
models.\24\ In 2019, the Board increased the transparency of the
supervisory stress test models by finalizing the Stress Testing Policy
Statement \25\ and the Enhanced Disclosure of the Models Used in the
Federal Reserve's Supervisory Stress Test (Enhanced Model
Disclosure).\26\ The Stress Testing Policy Statement describes the
Board's policies and procedures that guide the development,
implementation, and validation of the models.\27\ The Stress Testing
Policy Statement also describes the Board's principles for stress test
model design, namely that the system of models used in the supervisory
stress test should result in projections that are (1) independent of
firm projections; (2) forward-looking in that they project future
losses and revenue; (3) consistent and comparable across firms; (4)
generated from simple approaches, where appropriate; (5) robust and
stable; (6) conservative; and (7) able to capture the effect of severe
economic stress. The Board has developed stress test models in
accordance with these principles, which are the foundation for the
stress test modeling decisions described in the comprehensive
documentation of the supervisory stress test models that the Board is
publishing in conjunction with this proposal.
---------------------------------------------------------------------------
\24\ See, e.g., Board, 2018 Supervisory Stress Test Results
(Jun. 2018), https://www.federalreserve.gov/publications/files/2018-dfast-methodology-results-20180621.pdf.
\25\ See 84 FR 6664 (Feb. 28, 2019).
\26\ See 84 FR 6784 (Feb. 28, 2019).
\27\ See 12 CFR 252, Appendix B.
---------------------------------------------------------------------------
The Enhanced Model Disclosure supplemented prior public
descriptions of the stress test models by providing some information
about their structure and by including a list of key variables that
influence the results of each model.\28\ However, the Board believes
more detailed information, beyond what is in the current Enhanced Model
Disclosure, would improve the ability of firms to accurately assess how
changes in their business activities might impact their supervisory
stress test results and, relatedly, their stress capital buffer
requirements and overall capital requirements.
---------------------------------------------------------------------------
\28\ See, e.g., Board, 2025 Supervisory Stress Test Methodology
(Jun. 2025), https://www.federalreserve.gov/publications/files/2025-june-supervisory-stress-test-methodology.pdf.
---------------------------------------------------------------------------
C. Supervisory Stress Test Modeling Framework
The Board's stress test models take macroeconomic variables from
the Board's severely adverse scenario and firm data as inputs to
produce each firm's projected capital ratios over a nine-quarter
horizon. The projected common equity tier 1 capital ratio is used to
inform each firm's stress capital buffer requirement, which becomes
part of a firm's capital conservation buffer.
The stress test models are intended to capture how a firm's
regulatory capital would be affected by the macroeconomic and financial
conditions described in the stress test scenarios, given the
characteristics of the firm's business model and balance sheet
composition. The Board uses a variety of statistical modeling
techniques to produce the stress test results, including multivariate
regression, which uses relationships in historical data to produce
projections of a variable (such as a loss given default). These models
are represented by a set of formulas and coefficients that produce the
projections.
The Board estimates the effect of the severely adverse scenario on
the regulatory capital ratios of firms by projecting revenues,
expenses, and losses for each firm over a nine-quarter projection
horizon (projection horizon). The projection horizon spans nine
quarters to ensure that the firms can continue to provide credit and
serve as financial intermediaries despite several quarters of adverse
economic conditions, as well as to promote the forward-looking nature
of capital planning by firms.
Projected net income, adjusted for the effect of taxes, is combined
with assumptions regarding capital actions and other changes to
regulatory capital to produce post-stress capital ratios. The Board's
approach to modeling supervisory stress test results, including the
calculation of post-stress capital ratios, is generally in alignment
with U.S. GAAP and the regulatory capital framework.\29\ However, the
stress test models may deviate from U.S. GAAP and the regulatory
capital framework, as circumstances warrant.
---------------------------------------------------------------------------
\29\ See generally 12 CFR part 217.
---------------------------------------------------------------------------
The Board established the Stress Testing Policy Statement modeling
principles to ensure that the models are well suited for their purpose
in the regulatory framework. In some cases, the Board's adherence to
the principles limits modeling choices and results in certain common
limitations across similarly constructed component
[[Page 51860]]
models. For instance, consistent with the principles of independence,
consistency and comparability, and simplicity, models are not designed
to capture all firm-specific nuances, future strategic initiatives, or
planned capital actions. Additionally, models may be limited by their
reliance on historic relationships and by the nature of the data
captured in firms' regulatory reports. Detailed assumptions and
limitations for the models are discussed in the comprehensive
documentation, which is available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
Under the Stress Testing Policy Statement, the Board's projections
also assume that a firm's balance sheet remains unchanged throughout
the projection horizon.\30\ This assumption seeks to help ensure that a
firm cannot ``shrink to health'' and that it remains sufficiently
capitalized to accommodate credit demand in a severe downturn.
---------------------------------------------------------------------------
\30\ See 12 CFR 252, Appendix B, section 2.7.
---------------------------------------------------------------------------
D. Stress Test Models
The Board's stress test models comprise twenty-one component models
that, when aggregated, produce projected regulatory capital ratios for
each firm (see Table 1 below). The models can be grouped into four
categories: credit risk, market risk, net revenue, and aggregation.
Credit risk models capture losses associated with retail and wholesale
loans that are held at amortized cost. Market risk models capture
losses associated with trading and counterparty exposures, securities,
and other assets held at fair value. Net revenue models capture income
and expenses, including those related to operational risk, earned or
incurred by a firm. Positive pre-provision net revenue offsets credit
and market risk losses in the calculation of a firm's pre-tax net
income. Aggregation models calculate a firm's pre-tax net income, which
is then adjusted for other elements such as taxes and regulatory
capital deductions to arrive at the projection of a firm's regulatory
capital, which is used to calculate a firm's projected capital ratios.
Additional detail about these component models is provided in Section
III.A of this Supplementary Information and the comprehensive model
documentation available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.\31\
---------------------------------------------------------------------------
\31\ See also Board, 2025 Supervisory Stress Test Methodology
(Jun. 2025), https://www.federalreserve.gov/publications/files/2025-june-supervisory-stress-test-methodology.pdf.
---------------------------------------------------------------------------
[[Page 51861]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.035
[[Page 51862]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.036
E. Summary of the Proposal
---------------------------------------------------------------------------
\32\ The Trading Issuer Default Loss Model, Trading Profit and
Loss Model, Credit Valuation Adjustment Model, and Largest
Counterparty Default Model apply only to a subset of firms. See
Section II.B of this Supplementary Information.
---------------------------------------------------------------------------
The Board is publishing comprehensive documentation on the stress
test models on the Board's website, at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. This model documentation
contains information on the models that together produce the results of
the supervisory stress test. The model documentation includes the
equations, variables, and coefficients used in each model (where
applicable); assumptions and limitations of each model; rationales for
modeling decisions; and discussions of alternative models. Section
VIII.A of this Supplementary Information summarizes changes to the
models, relative to the 2025 stress test, that the Board plans to
implement in the 2026 stress test cycle; section VIII.B of this
Supplementary Information contains an analysis of the potential effects
of these proposed model changes. Detailed documentation on these
changes is also provided on the Board's website, at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. As
part of this proposal, the Board is inviting public comment on the
stress test models and these changes.
In addition, the Board is proposing to codify an enhanced
disclosure process that would build on the previous efforts that the
Board has made to increase the transparency and public accountability
of the supervisory stress test. Under this enhanced disclosure process,
the Board would annually publish comprehensive model documentation on
the stress test models, invite public comment on any material changes
that the Board seeks to make to those models, and annually publish the
stress test scenarios for comment. The Board would also commit to
responding to substantive public comments on any material model changes
before implementing them. The proposal would revise the Stress Testing
Policy Statement to align with this enhanced disclosure process, as
well as to amend the Board's general policy related to disclosing
additional information directly to a firm about that firm's supervisory
stress test results. To accommodate the annual comment process on the
scenarios, the proposal would shift the jump-off date of the
supervisory and company-run stress tests from December 31 to September
30.
Additionally, this proposal would amend the Scenario Design Policy
Statement in several ways. The Board would include in the Scenario
Design Policy Statement detailed descriptions of additional guides that
are used to inform the Board's choice of the values of the scenario
variables along their scenario paths. The guides are designed to
balance the competing objectives of predictability and transparency, on
the one hand, with the severity and relevance of the macroeconomic and
financial market scenarios, on the other hand. Most of the proposed
guides also incorporate features similar to the range of options in the
existing unemployment guide or the automatic adjustment of the house
price path to current housing market conditions in the existing house
price guide. This approach would allow the Board to continue to adjust
the severity of those variables as necessary to avoid inducing greater
procyclicality in the financial system and macroeconomy.
Similarly, the Board is proposing to incorporate additional
information into the Scenario Design Policy Statement about the
framework used to create the global market shock component of the
severely adverse scenario. This information includes, but is not
limited to, details on the logic underlying the severity of the shocks
and a description of the processes used to generate the shock values.
The Board is also proposing to update the global market shock
methodology to simplify the scenario and better align certain elements
of the global market shock with the nature of an ``instantaneous''
shock. The proposal would also make revisions to the stress test rules
to improve the risk capture of the supervisory stress test by widening
the
[[Page 51863]]
as-of date window for the global market shock.
Finally, the proposal would make changes to the FR Y-14A/Q/M
reports to remove items and documentation requirements that are no
longer needed to conduct the supervisory stress test, as well as to
collect additional data to improve risk capture.
F. Purpose of the Proposal
The purpose of this proposal is to provide the public with more
information about the stress test models and scenarios and to help
ensure that the public has an opportunity to comment on the models and
scenarios. While the Board has increased the transparency of the stress
test models over time, disclosing additional information about the
stress test models and their underlying methodologies will further
increase transparency and improve public accountability.
Publishing detailed descriptions of the stress test models for
comment, as well as committing to future enhanced disclosures, has
benefits. First, the increase in transparency would increase public
accountability and instill confidence in the fairness of the
supervisory stress tests. Second, the disclosure process would create a
new mechanism for obtaining feedback from the public, including
academics, financial analysts, and firms, on the design and
specifications of the models, which should lead to model improvements.
Third, a firm would have a better sense of how its risk profile would
factor into its stress capital buffer requirement, which would reduce
the likelihood of unanticipated stress test results and allow for
better capital and business planning by firms. Finally, the public
disclosure of additional information about supervisory stress tests
should strengthen market discipline, because investors, counterparties,
and rating agencies would be able to better assess a firm's risk
profile.\33\ The costs and benefits of this proposal are described more
thoroughly in Section X of this Supplementary Information.
---------------------------------------------------------------------------
\33\ See, e.g., N. Gambetta, M.A. Garc[iacute]a-Benau, & A.
Zorio-Grima, Stress test impact and bank risk profile: Evidence from
macro stress testing in Europe, 61 Intl. Rev. of Econ. & Fin 347-54
(2019); I. Goldstein & Y. Leitner, ``Stress test disclosure: theory,
practice, and new perspectives,'' Handbook of Financial Stress
Testing 208-223 (2022).
---------------------------------------------------------------------------
With respect to the proposed amendments to the Scenario Design
Policy Statement, this proposal also builds on the contents of the
current Scenario Design Policy Statement and would amend it to provide
additional transparency, public accountability, and predictability in
the variable paths. The changes would support the Board in developing
scenarios, inviting comment on those scenarios, incorporating input
from commenters, and maintaining the current schedule for release of
the final scenarios. Despite the increased predictability in the
scenarios, the new framework would remain flexible enough to suitably
assess whether firms can maintain an adequate amount of loss-absorbing
capital to stay above minimum regulatory requirements and continue
financial intermediation during periods of stress, as well as adjust
features that might add to existing procyclicality in the financial
system, as appropriate. In practice, the scenarios resulting from the
revised framework are expected to remain consistent with the current
Scenario Design Policy Statement and should not result, on average over
a typical business cycle, in materially different scenarios than would
have been designed previously.
Additionally, the proposal would simplify the design of the global
market shock component and incorporate additional information on the
development process into the Scenario Design Policy Statement, which
outlines the Board's approaches to designing market shocks, including
important considerations for scenario design, possible approaches to
developing scenarios, and a development strategy for implementing the
preferred approach. Taken together, these changes would improve
transparency, public accountability, and predictability of the
supervisory scenarios, while ensuring the supervisory stress test's
ability to capture changes in the risks in the financial industry over
time.
III. Overview of the Stress Test Modeling Framework
As summarized in Section II.D of this Supplementary Information,
the Board estimates the effect of the scenarios on the regulatory
capital ratios of firms participating in the stress test by projecting
net income and other components of regulatory capital for each firm
over a nine-quarter projection horizon. To do so, the Board uses
twenty-one component models, the macroeconomic variables from the
Board's severely adverse scenario, and firm data. This section provides
an overview of the component models the Board used to run the 2025
supervisory stress test. See Table 1 in Section II.D of this
Supplementary Information.
A. Supervisory Stress Test Models
The Board calculates projected pre-tax net income by combining
projections of pre-provision net revenue,\34\ provisions for credit
losses,\35\ and other gains or losses.\36\ Each component of pre-tax
net income is described below.
---------------------------------------------------------------------------
\34\ Pre-provision net revenue includes, among other items,
income from mortgage servicing rights, losses from operational risk
events, and other real estate owned costs.
\35\ For firms that have adopted Accounting Standards Update
(ASU) 2016-13, the Federal Reserve incorporates its projection of
expected credit losses on securities in the allowance for credit
losses, in accordance with Financial Accounting Standards Board
(FASB), Financial Instruments--Credit Losses (Topic 326). See FASB
ASU No. 2016-13, ``Financial Instruments--Credit Losses (Topic 326):
Measurement of Credit Losses on Financial Instruments.''
\36\ Other gains or losses include losses on held-for-sale
loans, loans measured under the fair-value option, and loan hedges.
---------------------------------------------------------------------------
Pre-Provision Net Revenue
Pre-provision net revenue is defined as net interest income
(interest income minus interest expense) plus noninterest income minus
noninterest expense. Consistent with U.S. GAAP, these projections
include projected losses due to operational risk events and expenses
related to the disposition of other real estate owned.\37\ The Board
projects most components of pre-provision net revenue using models that
relate specific revenue and non-provision-related expenses to the
characteristics of firms and to macroeconomic variables. These include
eight components of interest income, seven components of interest
expense, six components of noninterest income, and three components of
noninterest expense. The Board separately projects losses from
operational risk and other real estate owned expenses. Operational risk
is defined as ``the risk of loss resulting from inadequate or failed
internal processes, people and systems or from external events.'' \38\
Other real estate owned expenses are expenses related to the
disposition of real estate owned properties and stem from losses on
first-lien mortgages.
---------------------------------------------------------------------------
\37\ However, pre-provision net revenue projections do not
include debt valuation adjustments, which are not included in
regulatory capital.
\38\ 12 CFR 217.101 ``Operational risk.''
---------------------------------------------------------------------------
Loan Losses and Provisions on Loans Measured at Amortized Cost
The Board typically projects losses using one of two modeling
approaches: the expected-loss approach or the net charge-off approach.
Generally, under the expected loss approach, expected losses are
estimated by projecting the probability of default, loss given default,
and exposure at default for each quarter of the projection horizon.
Expected losses in each quarter are the product of these three
components. Under the net
[[Page 51864]]
charge-off approach, losses are projected using historical behavior of
net charge-offs as a function of macroeconomic and financial market
conditions and loan portfolio characteristics.\39\
---------------------------------------------------------------------------
\39\ Entire loans or portions of loans may be charged off if a
firm believes that the loan will not be repaid. If an amount that is
charged off is ultimately repaid by the borrower, then that repaid
amount is added to a firm's income as a recovery. Net charge-offs
are total charge-offs less any recoveries.
---------------------------------------------------------------------------
The Board estimates losses for loans measured at amortized cost
separately for different categories of loans, based on the type of
obligor, collateral, and loan structure. The individual loan types
modeled can broadly be divided into (1) retail loans, including various
types of residential mortgages, credit cards, student loans, auto
loans, small business loans, and other consumer loans; and (2)
wholesale loans, such as commercial and industrial loans and commercial
real estate loans. For most loan types, losses in quarter t are
estimated as the product of the projected probability of default in
quarter t, the loss given default in quarter t, and exposure at default
in quarter t.
The probability of default component measures the likelihood that a
borrower enters default status during a given quarter t. The other two
components capture the lender's net loss on the loan if the borrower
enters default. The loss given default component measures the
percentage of the loan balance that the lender will not be able to
recover after the borrower enters default, and the exposure at default
component measures the total expected outstanding loan balance at the
time of default.\40\
---------------------------------------------------------------------------
\40\ When applicable, loan loss models may factor in shared-loss
agreements with the Federal Deposit Insurance Corporation.
---------------------------------------------------------------------------
The Board's definition of default, for stress test modeling
purposes, may vary for different types of loans and may differ from
general industry definitions or classifications. The Board generally
models probability of default as a function of loan characteristics and
economic conditions. The Board typically models loss given default
based on historical data, and modeling approaches vary for different
types of loans. For certain loan types, the Board models loss given
default as a function of borrower, collateral, or loan characteristics
and the macroeconomic variables from the supervisory scenarios. For
other loan types, the Board assumes loss given default is a fixed
percentage of the loan balance for all loans in a category. The
approach to modeling exposure at default also varies by loan type and
depends on whether the loan is a term loan or a line of credit.
For certain retail loan categories, projections capture the
historical behavior of net charge-offs as a function of macroeconomic
and financial market conditions and loan portfolio characteristics. The
Board then uses these stress test models to project future charge-offs
consistent with the evolution of macroeconomic conditions under the
severely adverse scenario. To project losses, the projected net charge-
off rate is applied to projected loan balances.
Losses on loans are then projected to flow into net income through
provisions for loan and lease losses (for simplicity, provisions for
loan losses). Provisions for loan losses reflect funds set aside to
cover loan losses that a firm expects to incur in a predetermined
future window. Provisions for loan losses feed into the allowance for
loan losses, which serves as a contra asset on a firm's balance sheet.
The charged-off amount of a loan reduces the outstanding balance of the
loan while also reducing the allowance for loan losses (that is,
charge-offs do not reduce a firm's total assets). Generally, provisions
for loan losses for each projected quarter in the supervisory stress
test equal projected losses on loans for the quarter plus the change in
the allowance for loan losses needed to cover the subsequent four
quarters of expected loan losses. This calculation incorporates the
allowance for loan losses established by the firm as of the jump-off
date of the stress test exercise.
Current Expected Credit Losses Framework
On January 1, 2020, most large and mid-sized U.S. banks adopted the
Current Expected Credit Losses (CECL) standard for calculating
allowances.\41\ CECL superseded the incurred loss accounting standard,
which was a backward-looking measure that enabled firms to calculate
allowances based on historical loss data and current economic
conditions. CECL, by contrast, is a forward-looking measure that
requires firms to estimate lifetime losses based on reasonable
estimates of future economic conditions. In October 2024, the Board
announced that it would continue to evaluate future enhancements to the
supervisory stress test approach for the incorporation of CECL.\42\
---------------------------------------------------------------------------
\41\ See FASB ASU No. 2016-13, ``Financial Instruments--Credit
Losses (Topic 326): Measurement of Credit Losses on Financial
Instruments.''
\42\ See Q(DST0030) (Oct. 9, 2024) and Q(DST0029) (Dec. 15,
2023), https://www.federalreserve.gov/publications/ccar-qas/comprehensive-capital-analysis-and-review-questions-and-answers.htm.
---------------------------------------------------------------------------
The Board is not proposing to implement CECL into the supervisory
stress testing framework as a part of this proposal. The allowance
calculation framework currently used in the supervisory stress test is
already forward-looking: it projects loan loss provisions four quarters
ahead. This approach aligns with the Board's modeling principle of
simplicity as it requires fewer assumptions than would be required to
determine provisions under CECL. In addition, in aggregate, the
cumulative loan loss provisions under the supervisory severely adverse
scenario are similar to provision projections submitted by the firms
that have adopted CECL. Should the Board decide to implement CECL into
the supervisory stress testing framework, it would seek public comment
prior to implementation, as it would likely be a material model change
as defined in this proposal.
Question 2: What factors should the Board consider when determining
whether to implement CECL into the supervisory stress testing framework
and why?
Question 3: What would be the advantages and disadvantages of
incorporating CECL into the supervisory stress testing framework?
Losses on Loans Measured on a Fair Value Basis
Certain loans are accounted for on a fair value basis instead of on
an amortized cost basis. If a loan is accounted for using the fair
value option, it is marked to market, and the accounting value of the
loan changes as a function of changes in market risk factors and
fundamentals. Similarly, loans that are held for sale are accounted for
at the lower of cost or market value. The stress test models for these
asset classes project gains and losses over the nine-quarter projection
horizon, net of any hedges, using the scenario-specific path of
interest rates and credit spreads. The Board uses different models to
estimate gains and losses on wholesale loans and retail loans that are
accounted for on a fair value basis since these loans have different
risk characteristics. However, these models all generally project gains
and losses over the nine-quarter projection horizon, net of hedges, by
applying the scenario-specific interest rate and credit spread shocks
to loan yields.
Losses on Securities
A firm's balance sheet typically contains holdings of two types of
securities related to investment activities: available-for-sale and
held-to-
[[Page 51865]]
maturity. Available-for-sale and held-to-maturity securities are
generally held at fair value and amortized cost, respectively, on a
firm's balance sheet. The Board estimates two types of losses on
securities related to investment activities.\43\
---------------------------------------------------------------------------
\43\ This portfolio does not include securities held for
trading. Losses on these securities are projected by the Trading
Profit and Loss Model that projects gains and losses on trading
exposures.
---------------------------------------------------------------------------
For debt securities classified as available-for-sale, projected
fluctuations in the fair value of the securities due to changes in
interest rates and other factors will result in unrealized gains or
losses that are recognized in capital for some firms through other
comprehensive income. Under U.S. GAAP, unrealized gains and losses on
available-for-sale debt securities are reflected in accumulated other
comprehensive income and do not flow through net income.\44\ Under the
regulatory capital rule, accumulated other comprehensive income must be
incorporated into common equity tier 1 capital for certain firms.
Unrealized gains and losses are calculated as the difference between
each security's fair value and its amortized cost. The amortized cost
of each available-for-sale debt security is equivalent to the purchase
price of the debt security, which is periodically adjusted if the debt
security was purchased at a price other than par or face value, has a
principal repayment, or has an impairment recognized in earnings.\45\
---------------------------------------------------------------------------
\44\ Unrealized gains and losses on equity securities are
recognized in net income and affect regulatory capital for all
firms. See FASB ASU No. 2016-01, ``Financial Instruments--Overall
(Subtopic 825-10): Recognition and Measurement of Financial Assets
and Financial Liabilities.''
\45\ The fair value of each available-for-sale security is
projected over the nine quarter projection horizon using either a
present value calculation, a full revaluation using a security-
specific discounted cash flow model, or a duration-based approach,
depending on the asset class.
---------------------------------------------------------------------------
Credit losses on available-for-sale and held-to-maturity securities
may be also recorded. Except for certain government-backed obligations,
both available-for-sale and held-to-maturity securities are at risk of
incurring credit losses.\46\ The stress test models project security-
level credit losses, using as an input the projected fair value for
each security over the nine-quarter projection horizon under the
severely adverse scenario. Credit losses on securities are included in
the projection of provisions.
---------------------------------------------------------------------------
\46\ Certain government-backed securities, such as U.S.
Treasuries, U.S. government agency obligations, U.S. government
agency or government-sponsored enterprise mortgage-backed
securities, federally backed student loan asset-backed securities,
and pre-refunded municipal bonds, are assumed not to be subject to
credit losses.
---------------------------------------------------------------------------
Projected other comprehensive income gains or losses from
available-for-sale debt securities are computed directly from the
projected change in fair value, taking into account credit losses and
applicable interest-rate hedges on securities. All debt securities held
in the available-for-sale portfolio are subject to other comprehensive
income losses.
Losses on Private Equity Exposures
The Board projects the value of private equity investments in
response to the severely adverse scenario of the supervisory stress
test.\47\ The Private Equity Model assigns losses and recoveries based
on changes in fair value, recognized in net income for all positions,
regardless of their individual accounting elections. While U.S. GAAP
allows for private equity to be carried under a variety of accounting
measures, the different accounting methods are generally not reflective
of fundamental risk differences--fair value is typically realized upon
the orderly sale of a given private equity investment, irrespective of
its accounting treatment during the holding period.\48\
---------------------------------------------------------------------------
\47\ The Board projects private equity losses only for firms
that are required to submit FR Y-14Q, Schedule F.24 (Private Equity)
because private equity exposures are reported on that schedule.
Currently, Schedule F.24 is required to be reported by firms subject
to Category I through III standards that have, on average, aggregate
trading assets and liabilities of $50 billion or more, or aggregate
trading assets and liabilities equal to 10 percent or more of total
consolidated assets. As discussed in Section XI.A of this
Supplementary Information, the Board is proposing to modify the
threshold for Schedule F.24 to align with other banking book
schedules.
\48\ Unlike a bond or loan, private equity investments generally
cannot be redeemed by holding to maturity and are therefore
fundamentally exposed to market risk at exit.
---------------------------------------------------------------------------
Losses on Trading Exposures
The trading stress test models cover a wide range of a firm's
exposure to asset classes such as public equity, foreign exchange,
interest rates, commodities, securitized products, traded credit (for
example, municipal securities, auction rate securities, corporate
credit, and sovereign credit), and other fair-value assets. Loss
projections are constructed by applying the market risk factor
movements specified in the global market shock component \49\ to market
values of firm-provided positions and risk factor sensitivities.\50\
The global market shock only applies to a subset of firms, as described
in Section II.B.a of this Supplementary Information. In addition, the
global market shock component is applied to firm counterparty exposures
to generate losses due to changes in credit valuation adjustment, which
is a change to the market value of an exposure (for example, a
derivative) to account for the risk that the counterparty defaults on
its obligation. Trading and credit valuation adjustment losses are
calculated only for firms subject to the global market shock component.
In contrast to the nine-quarter evolution of losses for other parts of
the supervisory stress test, and as previously described, these losses
are estimated and applied in the first quarter of the projection
horizon. This timing is based on the observation that market
dislocations can happen rapidly and unpredictably any time under stress
conditions. It also ensures that potential losses from trading and
counterparty exposures are incorporated into a firm's capital ratio at
all points in the projection horizon.
---------------------------------------------------------------------------
\49\ See Section II.B.a of this Supplementary Information.
\50\ The supervisory trading models are also used to calculate
gains or losses on firms' portfolios of hedges on credit valuation
adjustment exposures.
---------------------------------------------------------------------------
The Board separately estimates the risk of losses arising from the
default of issuers of debt securities held for trading. These losses
account for concentration risk in corporate, sovereign, agency, and
municipal credit positions. In contrast to the trading losses described
above, these losses are applied in each of the nine quarters of the
projection horizon to capture the risk that several quarters of
stressful economic conditions may cause additional issuers of debt
securities to default, which aligns with the Board's principle of
conservatism from the Stress Testing Policy Statement.
Largest Counterparty Default Losses
The largest counterparty default component is applied to firms with
substantial trading or custodial operations. This component captures
the risk of loss due to the unexpected default of the counterparty
whose default on derivatives and securities financing transactions,
with exposures revalued by applying the global market shock component,
would generate the largest stressed losses for a firm. Consistent with
the Board's modeling principles and with the losses associated with the
global market shock component, losses associated with the largest
counterparty default component are recognized in the first quarter of
the projection horizon.
Balance Projections and the Calculation of Regulatory Capital Ratios
As described above, the Board assumes that a firm takes actions to
maintain its current level of assets, including its investment
securities, trading assets, and loans, over the
[[Page 51866]]
projection horizon. The Board also assumes that a firm's risk-weighted
assets and leverage ratio denominators remain unchanged over the
projection horizon, except that the Board will account for changes
primarily related to the calculation of regulatory capital or due to
changes to the Board's regulations.\51\
---------------------------------------------------------------------------
\51\ See 12 CFR 252, Appendix B, section 3.4; Board, Press
Release (Mar. 4, 2020), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20200304a.htm.
---------------------------------------------------------------------------
The Board includes five regulatory capital ratios in the
supervisory stress test: (1) common equity tier 1 risk-based capital,
(2) tier 1 risk-based capital, (3) total risk-based capital, (4) tier 1
leverage, and (5) supplementary leverage. A firm's post-stress
regulatory capital ratios are projected in accordance with the Board's
regulatory capital rule using the Board's projections of pre-tax net
income and other scenario-dependent components of the regulatory
capital ratios. Pre-tax net income and the other scenario-dependent
components of the regulatory capital ratios are combined with
additional information, including assumptions about taxes and capital
distributions, to project post-stress measures of regulatory capital.
In those calculations, the Board adjusts pre-tax net income to account
for taxes and other components of net income, such as income
attributable to minority interests, to arrive at after-tax net income.
The Board calculates the change in equity capital over the projection
horizon by combining projected after-tax net income with changes in
other comprehensive income, assumed capital distributions, and other
components of equity capital. The path of regulatory capital measures
over the projection horizon is calculated by combining the projected
change in equity capital with the firm's starting capital position and
accounting for other adjustments to regulatory capital specified in the
Board's regulatory capital framework.\52\ The denominator of each
firm's risk-based capital ratios is based on a firm's standardized
approach for calculating risk-weighted assets on the jump-off date of
the supervisory stress test, and may change for each quarter of the
projection horizon to account for adjustments specified in the capital
rule (for example, adjustments due to the thresholds for deducting
certain deferred tax assets).
---------------------------------------------------------------------------
\52\ The regulatory capital framework specifies that regulatory
capital ratios account for items subject to adjustment or deduction
in regulatory capital, limits the recognition of certain assets that
are less loss-absorbing, and imposes other restrictions. See
generally 12 CFR part 217.
---------------------------------------------------------------------------
B. Supervisory Stress Test Scenarios
The Board conducts the supervisory stress test using two
scenarios--the baseline and severely adverse. The severely adverse
scenario describes a hypothetical set of conditions designed to assess
the strength and resilience of firms in a severely adverse economic
environment and includes 28 variables that are disclosed by the Board
each year prior to the supervisory stress test. Some variables describe
economic developments within the United States while others describe
developments in foreign countries.\53\ These variables serve as an
input to the calculation of supervisory stress test results for all
firms. As discussed above, for a subset of firms, the severely adverse
scenario also includes two additional components: the global market
shock component and the largest counterparty default component. The
scenarios and associated components are developed solely for
supervisory stress testing purposes and do not represent economic
forecasts of the Board.
---------------------------------------------------------------------------
\53\ For a description of the macroeconomic variables applicable
to the 2025 supervisory stress test, see Board, 2025 Stress Test
Scenarios (Feb. 2025), https://www.federalreserve.gov/publications/files/2025-stress-test-scenarios-20250205.pdf.
---------------------------------------------------------------------------
Geographic Variation of Macroeconomic Variables
While the Board projects the paths of macroeconomic variables at
the national level, the Board uses regional-level (that is, state- and/
or county-level) macroeconomic variables in the stress test models to
project losses on certain loans held for investment at amortized
cost.\54\ In general, model outputs are demonstrably impacted by the
macroeconomic environment, as both probability of default and loss
given default increase during periods of economic stress. Importantly,
the macroeconomic environment can also vary notably across geography,
in addition to across time. For instance, during the 2007-2009 crisis
period, housing prices fell more sharply in certain geographies
compared to others. Accordingly, historical loss rates in many loan
categories were higher during this period in geographies where housing
prices fell more sharply.
---------------------------------------------------------------------------
\54\ Specifically, the Board uses regional-level macroeconomic
variables in the First Lien Model, the Home Equity Model, the Credit
Card Model, the Auto Model, and the Commercial Real Estate Model.
---------------------------------------------------------------------------
Therefore, to account for the impacts of different macroeconomic
environments across geographies on historical loan performance, the
Board calibrates model parameters in certain stress test models using
regional macroeconomic variables as opposed to national macroeconomic
variables. For example, the unemployment rate used in an applicable
model may be the state level unemployment rate, while the house price
index values used in the model may be the county-level house price
indices or, in the case of loans in counties where a house price index
is not projected, a state-level house price index.\55\ Analysis
performed by the Board demonstrates that a certain model's statistical
fit and sensitivity to the macroeconomic environment may perform better
when using regional-level variables compared to when using only
national-level variables. The use of regional-level variables is
described in each applicable model section of the comprehensive model
documentation.
---------------------------------------------------------------------------
\55\ Certain variables do not vary based on geography. For
example, interest rates are typically set by national and not
regional markets. For these variables, the Board uses the national-
level paths in the models.
---------------------------------------------------------------------------
However, because the severely adverse scenario only includes
national-level variable paths, the Board derives the paths of regional-
level variables from the paths of national-level variables. The Board
employs a simple approach to calculating the paths of regional-level
variables in that these variables have the same percentage change (in
the case of an index variable) or level change (in the case of non-
index variables) as the national-level variables, but the starting
points are the regional-level values, not the national-level values.
For example, the projected path of the house price index is assumed to
have the same percentage change in a given quarter as the percentage
change of the national house price index,\56\ and the projected path of
unemployment rate is assumed to have the same level change in a given
quarter as the level change of the national unemployment rate.\57\ The
use of percentage changes for home price indices and level changes for
unemployment rates avoids accentuating differences in the macroeconomic
environment observed immediately prior to the beginning of the
scenario, which could lead to large discrepancies in projected variable
paths across geographies during the severely adverse scenario.
---------------------------------------------------------------------------
\56\ The house price index used in the supervisory stress test
scenarios is set to be equal to 100 in January 2000. This choice of
index month is arbitrary and does not reflect any underlying
economic importance of this period.
\57\ For example, if the national unemployment rate increases by
0.5 percentage points in a given quarter, the state-level
unemployment rate would be projected to increase by 0.5 percentage
points in that quarter as well.
---------------------------------------------------------------------------
These simple, uniform policies for allocating changes to the
national
[[Page 51867]]
macroeconomic environment at the regional level ensure that loans to
borrowers in certain geographies are not unduly favored or penalized.
While it is plausible that certain geographies may experience more
volatility than others in terms of the macroeconomic environment, the
Board does not estimate such volatility to differentiate scenarios
across geography, to avoid making assumptions about the severity of a
hypothetical recession across different regions.
The Board also uses historical regional data to produce model
projections. While the regional scenarios are projected based on the
national path, the Board retains variation in the historical regional
macroeconomic variables.\58\ The Board may also use historical regional
macroeconomic variables in the models to calculate the appreciation in
house prices since origination (which may be needed to calculate loan-
to-value ratios), or the Board may use regional macroeconomic variables
to calculate year-over-year changes in the variables. Alternatively,
the Board could replace all historical values with their national
equivalent when projecting losses, thus applying a truly uniform
treatment across geographies. While this alternative would have the
benefit of maximizing geographic consistency, it would ignore
meaningful variation in the historical environment and thereby reduce
the predictive power of the model. For instance, if a given geography
has had higher house price appreciation since its origination date
compared to the national average, without incorporating these
historical values into the macroeconomic data used to project losses
the model would understate the level of equity the borrower has as of
the beginning of the projection period. The Board has therefore
developed this hybrid approach to estimating losses in the supervisory
stress test, in which it applies a uniform treatment to projected
values of macroeconomic variables across geographies, while also
retaining historical differences across geographies. This methodology
allows for the incorporation of all available historical data needed to
produce accurate projections, while avoiding the need to make
assumptions about which geographies will have more or less severe
macroeconomic paths during a hypothetical recession. Further discussion
of how the Board's models account for geographic variation in
variables, including a proposed change to the Board's modeling
approach, is included in the comprehensive model documentation,
available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
---------------------------------------------------------------------------
\58\ The historical regional unemployment rate and house price
index data are seasonally adjusted using the X11 procedure when a
seasonally adjusted version of these series is not available from
the source data. Seasonal adjustment is applied for consistency and
comparability with the published national scenario variables. For
more information about the X11 procedure developed by the U.S.
Census Bureau, see Shiskin J., Young A., and Musgrave, J., 1967. The
X-11 Variant of the Census Method II Seasonal Adjustment Program.
U.S. Department of Commerce, Bureau of the Census.
---------------------------------------------------------------------------
Question 4: What are the advantages and disadvantages of the
Board's treatment of regional (i.e., state and county) macroeconomic
variables in the credit risk models?
Question 5: What alternatives should the Board consider to the
approach outlined above for defining state and county macroeconomic
variables based on the national variables included in the scenarios?
What would be the advantages and disadvantages of these alternatives?
Auxiliary Variables
In addition to the 28 variables that the Board discloses each year,
the Board also generates paths for a limited number of other variables
that are used in the supervisory stress test. These variables, known as
auxiliary variables, are not disclosed by the Board because their paths
are based on the paths of the 28 disclosed variables (that is, the
paths are contingent upon movements in the 28 disclosed variables). For
example, the path of Mexico's gross domestic product (GDP) growth rate
is a function of the GDP growth rate paths of other country blocs that
are disclosed. Some models use these auxiliary variables, as described
in the applicable model sections of the comprehensive model
documentation available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.\59\
---------------------------------------------------------------------------
\59\ Detailed descriptions of the process for creating the paths
of auxiliary variables are included in the applicable model
documentation.
---------------------------------------------------------------------------
C. Data Used in Stress Testing
Input Data
The Board generally develops and implements the models with data it
collects on regulatory reports as well as proprietary third-party
industry data. Most of the data used in the supervisory stress test
projections are collected through the Capital Assessments and Stress
Testing regulatory report (FR Y-14), which includes a set of annual (FR
Y-14A), quarterly (FRY-14Q), and monthly (FRY-14M) schedules.\60\
---------------------------------------------------------------------------
\60\ The FR Y-14 report forms and instructions are available on
the Board's website at https://www.federalreserve.gov/apps/reportforms/default.aspx.
---------------------------------------------------------------------------
A firm must submit detailed loan and securities information for all
material portfolios on the FR Y-14Q and FR Y-14M. The definition of a
material portfolio for purposes of FR Y-14 reporting is based on a
firm's size and complexity.\61\ Portfolio categories are defined in the
FR Y-14M and FR Y-14Q reporting instructions. Each firm has the option
to submit the relevant data schedule for a given portfolio that does
not meet the materiality threshold as defined in the instructions. If a
firm does not submit data on its immaterial portfolio(s), the Board
will assign to that portfolio the median loss rate estimated across the
set of firms with material portfolios. This loss assumption adheres to
the principle of simplicity, as well as the principle of consistency
and comparability, from the Stress Testing Policy Statement.
---------------------------------------------------------------------------
\61\ Specifically, the definition of a material portfolio varies
depending upon a firm's categorization in the risk-based category
framework adopted by the Board for determining prudential standards.
See 12 CFR 238.10; 12 CFR 252.5.
---------------------------------------------------------------------------
While each firm is responsible for ensuring the completeness and
accuracy of data provided in the FR Y-14 reports, the Board makes
efforts to validate firm-reported data and requests resubmissions of
data where errors are identified. If data quality remains deficient
after resubmission, the Board applies conservative assumptions to a
particular portfolio or to specific data, depending on the severity of
deficiencies. For example, if the Board deems the quality of a firm's
submitted data too deficient to produce a stress test model estimate
for a particular portfolio, then the Board assigns a high loss rate
(for example, 90th percentile) or a conservative pre-provision net
revenue rate (for example, 10th percentile) to the portfolio balances
based on supervisory stress test projections of portfolio losses or
pre-provision net revenue for other firms.\62\ If data that are direct
inputs to stress test models are missing or reported erroneously but
the problem is isolated in such a way that the existing supervisory
framework can still be used, the Board assigns a conservative value
(for example, 10th or 90th percentile) to the specific data based on
all available data reported by firms. These assumptions are consistent
with the Board's principle of conservatism and policies on the
treatment of immaterial portfolios and missing or erroneous
[[Page 51868]]
data, as described in the Stress Testing Policy Statement.
---------------------------------------------------------------------------
\62\ Prior to assigning a conservate loss or revenue rate to
produce a firm's stress test results, the Board consults with a firm
that submits deficient data in order to determine whether the
applicable data issue can be remedied.
---------------------------------------------------------------------------
Additionally, certain stress test model projections rely on data
from the Consolidated Financial Statements for Holding Companies
regulatory report (FR Y-9C), which contains consolidated income
statement and balance sheet information for each firm subject to the
stress test. The FR Y-9C also includes off-balance sheet items and
other supporting schedules, such as the components of risk-weighted
assets and regulatory capital, that may be used in the stress test
models.
In limited circumstances, the Board also uses data provided by
third parties in the development and execution of the supervisory
stress test. The comprehensive model documentation identifies these
instances. The scenario data discussed above is also an input into the
stress test projections.
Data Preparation and Adjustments
a. Data Preparation
The data inputs the Board uses may not be initially suitable for
use in the stress test models. In these cases, the Board takes several
steps to prepare the data for use in the stress test models. The
specific steps for each model are discussed in the applicable model
descriptions within the comprehensive model documentation, though
generally data are prepared for use in the models for two purposes: to
remove outliers from the sample and to seasonally adjust the data.
These adjustments help ensure that the model results are reasonable.
The Board may remove outliers or data that are not applicable to
the model from the sample to facilitate more usable results. For
example, if a commercial real estate loan has a unusually high loan-to-
value (LTV) ratio (over 150 percent at origination), then data for that
loan are not included in the Commercial Real Estate Model because its
inclusion may produce unreliable results. Additionally, if first lien
mortgages are insured by the Federal Housing Administration or
Department of Veterans Affairs, then they are excluded from the First
Lien Model because these loans would not generate losses in the
supervisory stress test, as they are assumed to be fully insured by the
U.S. government. In both examples, the model output is more sensible
and more reflective of a firm's risk profile because of these
adjustments.
The Board also may seasonally adjust data, where appropriate. For
example, the vacancy rate of hotel commercial real estate exposures may
fluctuate on a seasonal cycle, with the vacancy rate moving higher or
lower in certain months based on a somewhat predictable pattern.
Because the vacancy rate can be an important variable for calculating
losses on hotel commercial real estate loans, this rate is seasonally
adjusted to ensure that the Commercial Real Estate Model produces more
stable results.
These types of data preparation steps help ensure that the Board's
models produce more reasonable results and that they align with the
principles in the Stress Testing Policy Statement in that they generate
consistent and robust projections. The Board therefore expects to
continue to use these data preparation steps, where appropriate, as
they are integral to the supervisory stress test process.
b. Data Adjustments
Data inputs are integral to generating the output of the stress
test models, which is a key component of a firm's stress capital buffer
requirement. The Board's Stress Testing Policy Statement notes that the
Board does not use data submitted by one or some of the firms unless
comparable data can be collected from all the firms that have material
exposure in a given area when generating supervisory stress test
projections.\63\ However, situations may arise where adjustments to a
firm's data would make the results more reasonable, and therefore
better calibrate a firm's stress capital buffer requirement to its risk
profile. The Board expects to continue to make these adjustments going
forward, where appropriate. Examples of when the Board may apply these
adjustments are described below.
---------------------------------------------------------------------------
\63\ See 12 CFR 252, Appendix B, section 2.8.
---------------------------------------------------------------------------
For example, the Board may apply a data adjustment where there is
missing or deficient firm-provided data, or where a firm uses
divestiture accounting. As described above, if the Board deems the
quality of a firm's submitted data too deficient to produce a stress
test model estimate for a particular portfolio, then the Board assigns
a conservative loss rate (for example, 90th percentile) or a
conservative pre-provision net revenue rate (for example, 10th
percentile) to the portfolio balances based on supervisory stress test
projections of portfolio losses or pre-provision net revenue for other
firms. If data that are direct inputs to stress test models are missing
or reported erroneously but the problem is isolated in such a way that
the existing supervisory framework can still be used, the Board assigns
a conservative value to the specific data based on all available data
reported by firms.
Additionally, when a firm sells assets or businesses, it may use
divestiture accounting in its financial statements until the sale is
consummated. Under divestiture accounting, a firm may list divested
assets as discontinued operations, classify them as held for sale or
available for sale instead of held for investment or held to maturity,
and report revenues as income from discontinued operations. The
accounting classification can be important for the supervisory stress
test as it may determine which model stresses the assets or income. For
example, in the 2025 supervisory stress test, the Board adjusted
certain input data that had been reclassified due to divestiture
accounting to improve projections of loan losses and related income to
ensure consistent treatment across firms with similar risks.
IV. Enhanced Disclosure Process
The Board is proposing to codify an enhanced disclosure process
under which the Board would annually publish comprehensive
documentation on the stress test models, invite public comment on any
material changes that the Board seeks to make to those models, and
annually publish the stress test scenarios for comment.
A. Annual Disclosure of Models
Under the proposal, the Board would annually publish comprehensive
documentation on the stress test models, similar to the comprehensive
documentation the Board is publishing with this proposal at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. The
Board would be required to publish this comprehensive documentation by
May 15 of the year in which the stress test is performed, and the
models described in the documentation would be used to produce the
stress test results disclosed by the Board by June 30 of that year. In
addition, the Board would seek public comment, and respond to such
public comment, on any material changes to the models before
implementing those changes in a stress test. Material model changes are
discussed in more detail in Section IV.B of this Supplementary
Information. To implement this enhanced disclosure process, the Board
is proposing to revise Regulations YY and LL, as well as the Stress
Testing Policy Statement.
For example, if the Board did not seek to make any material model
changes to its stress test models for the 2027 supervisory stress test,
then it would publish the comprehensive model documentation used in the
2027 stress test cycle by May 15, 2027. This
[[Page 51869]]
documentation would identify any changes (relative to the models used
in the 2026 stress test), including technical, non-material changes to
the models to improve performance. This process would allow the public
to review the changes, as well as comprehensive documentation on the
models used in the 2027 stress test cycle, before the release of the
stress test results.
As an alternative example, if the Board sought to implement a
material model change (as discussed in Section IV.B of this
Supplementary Information) in the 2027 supervisory stress test, then
the Board would seek comment on the proposed change, consider and
respond to public feedback, and, then implement, defer, or reject the
material model change for the 2027 stress test cycle. If the Board
sought to implement the material model change in the 2027 stress test,
the Board would republish updated model documentation before or
simultaneously with the annual publication of comprehensive model
documentation (i.e., by May 15, 2027). This process for material model
changes would increase the transparency of the Board's stress testing
model framework and ensure that the public has the opportunity to
comment on material model changes before they are used in the next
stress test cycle.
Question 6: How else could the Board enhance the transparency and
public accountability of its stress test models? For instance, what
additional information regarding the stress test models, if any, should
the Board provide, and why?
Question 7: How else could the Board facilitate public
participation in model development? For example, the Board could invite
comment on all model changes, rather than only material model changes,
before implementing them in the stress test. Under such an approach,
the Board could make an exception for technical or other types of
ministerial changes. Such a process would limit the Board's flexibility
to revise models due to unforeseen events and circumstances. What are
the advantages and disadvantages of this expanded approach or other
approaches to facilitate public participation in model development? How
should the Board balance transparency and public accountability with
model dynamism and operational burden?
Question 8: What are the advantages and disadvantages of inviting
public comment, and committing to responding to comments, on material
model changes before the Board implements them in the subsequent stress
test?
Question 9: What are the advantages and disadvantages of publishing
the comprehensive model documentation by May 15 of each stress test
cycle? For example, does this timeline provide enough time for the
public to review any changes made by the Board to confirm they are not
material? Should the Board consider publishing the comprehensive model
documentation earlier at an earlier date, such as April 5, or a later
date, such as June 30? What would be the advantages or disadvantages of
publishing the comprehensive model documentation earlier or later?
Question 10: The Board is not currently publishing the results of
its internal model validation process. What would be the advantages and
disadvantages of publishing these results or providing more information
about its internal model validation process?
B. Model Changes
The proposed rule would define a ``model change'' to mean ``the
introduction of a new model or a conceptual change to an existing
model.'' \64\ Conceptual changes to existing models would include
changes to model assumptions, incorporation of a new statistical
technique to estimate loss, or the addition or deletion of any model
components or sub-components that currently inform a firm's stress
capital buffer requirement.
---------------------------------------------------------------------------
\64\ As discussed in Section II.D of this Supplementary
Information, there are twenty-one component models that comprise the
stress test models. A ``new model'' would mean a model that fully
replaces one of these twenty-one component models or is added to the
modeling suite (e.g., a 22nd component model). For purposes of
assessing materiality, as discussed in Section IV.C of this
Supplementary Information, model changes would not be aggregated or
netted across the component models.
---------------------------------------------------------------------------
Model changes would not include changes resulting from updates or
adjustments to input data, such as firm data, third-party vendor data,
and scenario data, including any re-estimation based on this data, as
well as changes related to the mechanical implementation of federal,
state, or local laws that are directly embedded in a stress test model
(e.g., the federal statutory tax rate).\65\ As is current practice, the
Board would continue to implement model changes related to changes in
accounting definitions or regulatory capital rules and model parameter
re-estimation based on newly available data with immediate effect.
These types of adjustments would not be considered model changes since
they do not substantively change the form of the stress test models as
described in the documentation. For example, the Board re-estimates
many of its models with updated data each year when it runs the
supervisory stress test. This re-estimation may result in changes to
the statistical coefficients produced by some of the models, even
though the Board has made no conceptual changes to the models. Under
the proposed definition of model change, such re-estimation would not
be viewed as a model change because the resulting changes stem solely
from updated data and not from a conceptual change to the models. In
contrast, the introduction or revision of a legal requirement that
causes a conceptual change to a model could be considered a model
change, and the Board would seek public comment before implementing
such a change if it met the proposed definition of a material model
change.
---------------------------------------------------------------------------
\65\ Re-estimation comprises updates to model parameters based
on consideration of different input data (e.g., incorporating the
most recent year's data as a model input, or incorporating data from
new stress test entrants or from mergers).
---------------------------------------------------------------------------
Question 11: What other types of changes to the supervisory stress
testing framework could the Board consider including in the definition
of ``model change''? What are the advantages and disadvantages of
broadening or narrowing the definition of ``model change''? For
example, should the Board define ``model changes'' to include changes
that result from new or updated input data, or changes that result from
using a new, third-party data source?
C. Material Model Changes
Each year, the Board refines and enhances its stress test models to
reflect advances in modeling techniques, respond to model validation
findings, incorporate richer and more detailed data, or identify more
stable models or models with improved performance, particularly under
stressful economic conditions. These changes may include re-
specification of models based on performance testing, benchmarking, and
other targeted changes used to produce projections.\66\ This process is
an important aspect of the modeling framework to help ensure that the
stress test models capture changes in borrower and lender behavior and
bank business practices. These model changes also help ensure that the
models are able to remain dynamic (i.e., can be enhanced to capture
emerging risks), produce
[[Page 51870]]
reasonable results, identify salient risks at firms, and maintain an
optimal level of robustness and stability.
---------------------------------------------------------------------------
\66\ Benchmarking is the process of evaluating a model's
performance by comparing its outputs and other performance metrics
against a specific standard, baseline, or the output and performance
of other comparable models or relevant data sources.
---------------------------------------------------------------------------
In addition, the Board must sometimes make changes to its stress
test models while it is running the stress test in response to
unforeseen events or circumstances to ensure that model output is
reasonable. For example, during the COVID-19 pandemic, the vacancy
rates for hotel properties were unprecedented and the Board made
certain adjustments to yield sensible commercial real estate loan
losses in the model output. Without making these in-cycle changes, the
results of the stress test would have been irrational and led to stress
capital buffer requirements that were not commensurate with applicable
firms' risk profile.
Under the proposed enhanced disclosure process, if these changes
are not material, as defined below, the Board would publish these model
changes by May 15 of the year in which the stress test is performed. To
balance the benefit of public feedback with the operational and
resource costs of seeking such feedback and to allow the Board to make
timely model adjustments to ensure reasonable results, the Board would
not formally invite public comment on these non-material model changes
before implementing them in the stress test; however, the Board
welcomes public feedback on these and all other aspects of the stress
test models once they are published. Notably, the Board would not
implement any in-cycle adjustments that are considered material model
changes prior to seeking public comment on the adjustment. In addition,
the Board would review and respond to all substantive public comments
on material model changes before implementing them in the stress test.
As discussed above, the Board is proposing to publish for comment
all material model changes and respond to all substantive comments on
such material model changes before implementing them in the stress
test. For example, if the Board sought to implement a new statistical
technique that would result in a material model change, then the Board
would seek public comment prior to implementing either of those
changes.\67\ The Board is proposing to define a ``material model
change'' as a model change that could have, in the Board's estimation,
an impact on the post-stress common equity tier 1 capital ratio of any
firm, or on the average post-stress common equity tier 1 capital ratios
of all firms required to participate in the upcoming stress test cycle,
based on the prior year's severely adverse scenario and prior year's
input data, equal to (i) a change of 20 basis points or more in the
projected common equity tier 1 ratio of any firm participating in the
upcoming stress test cycle; or (ii) a change of 10 basis points or more
in the average of the absolute value of each firm's change in projected
common equity tier 1 ratio.\68\ The Board proposes to apply this
definition of a material model change across both Regulation YY and
Regulation LL, such that the individual materiality threshold would
apply to all firms required to participate in the next stress test
under either regulation, and such that the Board's estimation of
whether a change meets the aggregate materiality threshold would be
determined across all firms required to participate in the next stress
test under either regulation.
---------------------------------------------------------------------------
\67\ For purposes of assessing materiality, model changes would
not be aggregated or netted across the component models. For
example, if the Board proposed a change to both the Pre-Provision
Net Revenue Model and Corporate Model in the same stress test cycle,
the Board would estimate the effects of each change separately for
purposes of determining materiality. Similarly, for purposes of
assessing materiality, model changes would not be aggregated or
netted within component models. For example, if the Board proposed
two changes to a component model, the Board would evaluate the
materiality of each change separately.
\68\ The Board would take the absolute value of each firm's
change in projected common equity tier 1 ratio, then average those
values. If the average is 10 basis points or greater, the change
would constitute a material model change.
---------------------------------------------------------------------------
The Board is proposing to use the threshold of a 20 basis point
change in the common equity tier 1 capital ratio for individual firms
in the definition of material model change because that threshold would
ensure that the public would be able to comment on any change that
would be likely to affect a firm's stress capital buffer requirement.
Considering the history of recent model changes, a threshold of 20
basis points would generally capture model changes that involve
conceptual enhancements to model specifications, such as to incorporate
improved modeling techniques or to capture emerging risks, while
scoping out those that are simpler model refinements, such as those
implemented to ensure that the models maintain consistency given
changing requirements (e.g., refinements made to accommodate the
transition from the London Interbank Offered Rate to SOFR). Therefore,
changes of smaller magnitudes would be unlikely to impact a firm's
stress capital buffer requirement, particularly if the proposed two-
year averaging approach to calculate a firm's stress capital buffer
requirement is adopted.\69\ If the two-year averaging approach is not
finalized or not finalized as proposed, the Board would consider a
lower individual materiality threshold of 10 basis points, which would
ensure that the public would be able to comment on any change that
would be likely to affect a firm's stress capital buffer requirement
without two-year averaging.
---------------------------------------------------------------------------
\69\ See 90 FR 16843 (Apr. 22, 2025).
---------------------------------------------------------------------------
The Board is proposing the threshold of a 10 basis point average
change in the absolute value of the change to each firm's projected
common equity tier 1 capital ratio in case a model change has minimal
individual impacts, but has a notable aggregate impact on firms
required to participate in the upcoming stress test. The Board selected
10 basis points for this aggregate prong because a model change of this
size would be likely to impact the aggregate projected common equity
tier 1 capital ratio, which is a salient measure of the health of the
banking system. A change that satisfies either of these materiality
thresholds would be considered a material model change.
Question 12: What are the advantages and disadvantages of this
definition of a material model change? What alternative quantitative
thresholds for materiality, if any, should the Board consider, and why?
For example, in assessing the materiality of a model change, as
described in the Stress Testing Policy Statement, the Federal Reserve
currently considers a change to be highly material if it would result
in a change in the common equity tier 1 capital ratio of 50 basis
points or more for one or more firms, relative to the model used in
prior years' supervisory exercises. What would be the advantages and
disadvantages of this or other alternative standards?
Question 13: What alternative definitions of materiality, if any,
should the Board consider? For example, the Board could consider the
impact of a change on a firm's pre-tax net income, rather than its
common equity tier 1 ratio. What are the advantages and disadvantages
of such alternative definitions?
Question 14: Under the proposal, for purposes of assessing the
materiality of a model change, the Board would not aggregate or net the
impact across or within component models. What forms of netting or
aggregation, if any, would be most appropriate and why? What would be
the advantages and disadvantages of netting or aggregating model
changes across or within component models to assess materiality? If the
Board were to net or
[[Page 51871]]
aggregate model changes, should the Board consider alternative
materiality thresholds? For example, the Board could consider an
alternative definition of materiality that considers the aggregate
impact of all of the model changes the Board intends to implement in a
future stress test cycle. Alternatively, the Board could aggregate the
impacts of all model changes to a given suite of models (e.g., credit
risk models) instead of considering the individual impacts of model
changes to the Auto Loan Model and the Commercial Real Estate Model.
Question 15: What are the advantages and disadvantages of inviting
and responding to public comment on material model changes before
implementing those changes? The proposal does not currently specify the
length of the comment period. What are the advantages and disadvantages
of a set length for the comment period (e.g., 30-day, 60-day, etc.)?
When considering the appropriate length of the comment period, how
should the Board evaluate trade-offs, for instance, between ensuring
that the public has ample time to consider and comment on material
model changes and ensuring that the stress test results are released by
June 30?
Question 16: If the Board does not adopt its proposal to calculate
a firm's stress capital buffer requirement by averaging stress test
results over two consecutive years, should the Board consider a lower
threshold to determine materiality, such as 10 basis points for the
individual firm threshold instead of the proposed 20 basis points? What
would be the advantages and disadvantages of a lower threshold?
D. Annual Disclosure of Scenarios
Under the proposal, the Board would annually publish for comment
the proposed stress test scenarios by October 15 of the calendar year
prior to the stress test, for at least a 30-day period. The timing of
the release and duration of the comment period will allow for
sufficient time to respond to comments and finalize the scenarios
within the current window for publishing final scenarios by February 15
in each annual stress test cycle.\70\ The disclosure of the annual
scenarios for comment, along with the implementation of additional
scenario variable guides and revisions to the Scenario Design Policy
Statement, would meaningfully improve the transparency, public
accountability, and predictability of the annual stress tests.
---------------------------------------------------------------------------
\70\ Trading or other components of the scenarios, and any
additional scenarios used by the Board, would continue to be
communicated by March 1 of the calendar year in which the stress
test is performed. 12 CFR 238.132(b); 12 CFR 238.143(b)(2)(i); 12
CFR 252.14(b)(2)(i); 12 CFR 252.44(b); 12 CFR 252.54(b)(2)(i).
---------------------------------------------------------------------------
The publication of macroeconomic scenarios in October would use
nowcasts, which are projections under baseline conditions, to determine
the jump-off points for the proposed scenario variable paths. The final
scenarios would be updated to include actual data. The paths of
scenario variables may be adjusted to some extent between the initial
scenario publication and the finalized scenario to reflect these
updated values.
By designing and publishing the guides described in Section IX.G of
this Supplementary Information, the Board expects that the annual
severely adverse scenarios will generally be more consistent and
predictable year-to-year. As a result, the Board weighed whether
publishing the annual scenarios for comment in a typical year would
contribute meaningful additional accountability that would improve the
stress test program, and whether the Board should limit publication of
the annual scenarios for comment to situations where the Board is
proposing to incorporate a salient risk into the scenarios that is not
described in this proposal. However, in the interest of enhancing
transparency and public accountability, the Board determined to
maintain its current practice of publishing its annual scenarios and,
further, to propose changes to Regulations LL and YY formalizing this
disclosure process.
Additionally, the Board plans to maintain its current practice of
disclosing the final scenarios only after firms' portfolios are fixed,
as disclosure of the final scenarios prior to the jump-off date of the
stress test could incentivize firms to modify their businesses to
minimize losses in the supervisory stress test without changing the
actual risk profile of the firms. Therefore, the Board is proposing to
move the jump-off date of the stress test from December 31 to September
30. This proposed change is discussed in greater detail in Section VI.A
of this Supplementary Information.
Finally, as described in Section VI.B of this Supplementary
Information, the Board is proposing to change the as-of date window for
the global market shock to occur between October 1 of the calendar year
two years prior to the year in which the stress test is performed to
October 1 of the year prior to the year in which the stress test is
performed. Therefore, the Board anticipates that the global market
shock as-of date will have already occurred for most future proposals
regarding the initial disclosure of the stress test scenarios. However,
the Board has not yet announced the global market shock as-of date for
the 2026 stress test and so cannot provide the exact relative shock
values for certain global market shock variables since the relative
shock values are a function of the actual data on the as-of date.
For relative shocks associated with the 2026 global market shock,
the data on the global market shock as-of date would be applied to
determine relative shock values, which will be disclosed as part of the
finalized scenarios. For example, if the Board proposes a shock to the
BBB corporate spread of 200 basis points and the BBB corporate spread
market level on the global market shock as-of-date is 400 basis points,
then the relative shock to the BBB corporate spread would be 200/400,
or 50 percent, for the 2026 global market shock.
Question 17: How should the Board publish the annual scenario for
comment? For example, the Board could publish the scenario on the
Board's website or include the text and supporting materials in a
Federal Register notice. Alternatively, the Board could consider
codifying each annual scenario as a part of Regulation YY. What would
be the advantages and disadvantages of these options or other
alternatives?
Question 18: What are the advantages and disadvantages of
publishing the annual scenarios for comment prior to the jump-off date
of the annual stress test cycle?
Question 19: What are the advantages and disadvantages of a 30-day
comment period? Should the Board consider an alternative comment period
length? If so, how long should the comment period be (e.g., 45 days, 60
days, etc.)? When considering the appropriate length of the comment
period, how should the Board evaluate trade-offs, for instance, between
ensuring that the public has ample time to consider and comment on
annual scenarios and ensuring that the stress test scenarios can be
finalized before February 15?
Question 20: How should the Board analyze comments received from
the public on proposed scenarios? What types of information would be
helpful to commenters in order to understand how the Board incorporates
comments received on proposed scenarios before finalizing the annual
scenarios?
E. Stress Capital Buffer Requirement Reconsideration Process
Under the Board's capital plan rule, a firm may request
reconsideration of the calculation of its preliminary stress
[[Page 51872]]
capital buffer requirement within 15 calendar days of receiving notice
of the preliminary requirement.\71\ A request for reconsideration may
include a request for an informal hearing on the firm's request for
reconsideration; the Board may, in its sole discretion, order an
informal hearing if the Board finds that a hearing is appropriate or
necessary to resolve disputes regarding material issues of fact.\72\
The Board is not proposing to change this reconsideration process.\73\
However, the Board is requesting public input on potential enhancements
to the stress capital buffer requirement reconsideration process. In
particular, the Board seeks public input on the following question:
---------------------------------------------------------------------------
\71\ 12 CFR 225.8(h)(2)(i); 12 CFR 225.8(i)(2); 12 CFR
238.170(h)(2)(i); 12 CFR 238.170(i)(2).
\72\ 12 CFR 225.8(i)(3)(ii); 12 CFR 225.8(i)(4); 12 CFR
238.170(i)(3)(ii); 12 CFR 238.170(i)(4).
\73\ Model adjustments made in response to a reconsideration
request granted by the Board would not be considered model changes
under the proposed enhanced disclosure process.
---------------------------------------------------------------------------
Question 21: What enhancements, if any, should the Board consider
making to its reconsideration request process? For example, the Board
could allow firms more time to request reconsideration of their
results, broaden or narrow the grounds for and scope of review, and/or
modify existing reconsideration request requirements in light of the
publication of the comprehensive model documentation and proposed
enhanced disclosure process. What would be the advantages and
disadvantages of these enhancements? What other changes, if any, should
the Board consider making to the reconsideration requirements and
procedures? What would be the advantages and disadvantages of those
changes?
V. Revisions to the Stress Testing Policy Statement
The Board is also proposing certain changes to the Stress Testing
Policy Statement to (i) amend the section related to disclosure of
information related to the stress test; and (ii) to align the Stress
Testing Policy Statement with the proposed enhanced disclosure process.
A. Future Supervisory Stress Test Results Disclosures
The Board's Stress Testing Policy Statement states that, in
general, the Board does not share information regarding supervisory
stress test results with firms that is not made available to the
broader public. However, providing additional details to a firm about
its own results could provide the firm with additional visibility into
its stressed revenue and loss projections, including any underlying
risks, and improve the firm's understanding of its stress capital
buffer requirement. For example, additional results information would
allow a firm to better understand how the stress test translates their
balance sheet and income information into projected losses and revenue,
which could help them better plan their business and understand the
risk of their exposures. To provide additional transparency, the Board
is therefore proposing to revise the Stress Testing Policy Statement to
clarify that the Board will generally disclose information directly to
a firm about the firm's supervisory stress test results that is not
available to the broader public, so long as the Board discloses similar
information to the other firms participating in a given stress test
cycle. For example, the Board may provide a firm's common equity tier 1
capital ratio during all quarters of the projection horizon. Providing
firm-specific results directly to the affected firms even when that
information is not disclosed to the broader public would allow firms to
better understand their results while preventing potentially sensitive
information about a firm from being shared with competitors. The Board
would continue to disclose the supervisory stress test results to the
public.
Question 22: What are the advantages and disadvantages of revising
the Stress Testing Policy Statement to clarify that the Board will
generally share non-public information about a firm's results directly
with a firm (provided that the Board is disclosing similar information
to other participating firms)?
B. Other Revisions to the Stress Testing Policy Statement
In addition, the Board is proposing to revise the Stress Testing
Policy Statement to align it with the proposed enhanced disclosure
process. For example, the Board is proposing to state that, during
model development, it invites, evaluates, and responds to substantive
public input on the stress test models. The Board is also proposing to
revise the Stress Testing Policy Statement to clarify that its public
disclosures about the stress test will now include comprehensive
descriptions of the models and changes to those models.
Question 23: What other changes could the Board make to the Stress
Testing Policy Statement to reflect the enhanced transparency of the
supervisory stress test or to supplement the Board's efforts to make
the supervisory stress test more transparent and to facilitate public
participation? What are the advantages and disadvantages of such
changes?
VI. Other Revisions to the Stress Testing and Capital Plan Rules
The Board is also proposing to revise the stress testing and
capital plan rules to reflect the Board's efforts to disclose more
information about the stress test scenarios.
A. Stress Test Jump-Off Date Change
The Board is separately seeking comment on the proposed scenarios
for use in the 2026 supervisory stress test. In general, disclosure of
the proposed scenarios prior to the jump-off date of the supervisory
stress test could incent firms to temporarily modify their businesses
to affect the results of the stress test without changing the actual
risk profile of the firms. The Board recognizes that the increased
transparency around scenario design resulting from the disclosure of
additional guides and a macroeconomic model used in that process would
allow firms to anticipate the trajectories of key scenario variables.
Using this information, firms could adjust their portfolios to specific
aspects of the proposed scenarios in ways that would reduce measured
losses without reducing the actual riskiness of the portfolios. Such
changes to firm business profiles could also result in greater than
typical quarter-to-quarter variability in the banking books of firms.
To address this potential risk associated with increased
transparency, the Board proposes to modify the jump-off date of the
supervisory and company-run stress tests from December 31 to September
30, while leaving unchanged the other dates associated with publication
of the final scenario and stress test results.\74\ With respect to the
capital planning rules, the Board proposes accomplishing this change
through revision to the definition of ``planning horizon'' in
Regulation Y and Regulation LL. This change would allow the Board to
publish the scenario for comment after the jump-off date of the stress
test, preventing firms from adjusting their exposures based on the
stress test. However, this proposed change would introduce an
additional quarter of staleness to the stress test and
[[Page 51873]]
stress test results. This change would also affect firms' capital plan
submissions. Although the due date for firms' annual capital plan
submissions would be unchanged, because of the proposed update to the
definition of planning horizon, firms' capital plans would not project
out as far. While the Board weighs these risks and considers adjusting
the stress test jump-off date, the Board seeks input from the public
regarding whether these risks are outweighed by the value to firms and
the public by publishing scenarios prior to the jump-off date of the
supervisory and company-run stress tests. Therefore, the Board seeks
public comment on whether to propose such modifications to limit the
ability of firms to adjust their balance sheets in response to the
proposed scenario prior to the jump-off date of the stress test.
---------------------------------------------------------------------------
\74\ The Board has experience operating the annual supervisory
stress test with a September 30 jump-off date. See, e.g., Board,
2015 Supervisory Scenarios for Annual Stress Tests Required under
the Dodd-Frank Act Stress Testing Rules and the Capital Plan Rule
(Oct. 23, 2014), https://www.federalreserve.gov/newsevents/pressreleases/files/bcreg20141023a1.pdf.
---------------------------------------------------------------------------
Question 24: What are the advantages and disadvantages of retaining
a jump-off date that would occur after the publication of the annual
scenario for comment? Should the Board consider retaining the December
31 jump-off date in order to promote transparency? Are there additional
risks or trade-offs that the Board should consider?
Question 25: What would be the advantages and disadvantages of
modifying the jump-off date of the stress test from December 31 to
September 30? If the Board were to modify the jump-off date, what other
changes should the Board consider making to the stress test timeline?
For example, what would be the advantages and disadvantages if the
Board were to change the timing of a firm's capital plan submission?
What would be the advantages and disadvantages of these changes?
Question 26: Should the Board consider modifying the jump-off date
of the stress test to a later date, rather than an earlier date, in
order to accommodate a public comment period?
B. Global Market Shock Date
The global market shock (GMS) is applied to market risk positions
held by the firms on a given as-of date, which, under the Board's
stress test rule, currently occurs between October 1 of the previous
year and March 1 of the year of a given stress test cycle.\75\ Under
the Board's regulations, the GMS can apply to both the supervisory
stress test and the company-run stress test for applicable firms. For
the supervisory stress test and the company-run stress test, the Board
must generally provide each affected firm with a description of the GMS
and with the specific GMS as-of date by March 1 of the year in which
the stress test occurs.\76\ For the company-run stress test, the Board
generally must also notify each affected firm by December 31 of year
preceding the stress test that the firm is required to include
additional components or scenarios in its company-run stress test.\77\
---------------------------------------------------------------------------
\75\ See, e.g., 12 CFR 238.143(b)(2)(i); 12 CFR 252.14(b)(2)(i);
12 CFR 252.54(b)(2)(i).
\76\ See, e.g., 12 CFR 238.132(b); 12 CFR 238.143(b)(2)(i); 12
CFR 252.14(b)(2)(i); 12 CFR 252.44(b); 12 CFR 252.54(b)(2)(i).
\77\ See, e.g., 12 CFR 238.143(b)(4)(i); 12 CFR 252.14(b)(4)(i);
12 CFR 252.54(b)(4)(i).
---------------------------------------------------------------------------
The Board selects a cycle-specific as-of date each year and,
typically, announces it to firms about two weeks later to ensure the
firms retain necessary data. The as-of date is expected to change from
year to year to avoid creating potential incentives for firms to take
temporary trading positions. However, there is a comprehensive date
selection process that, in practice, shortens the actual window during
which the GMS as-of date is generally selected. A wider date range
would allow the Board to capture a broader range of market risks across
different time periods, thereby improving the risk capture of the
global market shock. The Board therefore proposes revising the date
range for the GMS as-of date to occur between (inclusive of) October 1
of the calendar year two years prior to the year in which the stress
test is performed to (exclusive of) October 1 of the calendar year one
year prior to the year in which the stress test is performed. By way of
example, this change would mean that for the 2026 supervisory stress
test, the GMS as-of date could fall on any date between October 1,
2024, through September 30, 2025. The Board proposes using this date
range because it would allow the Board to choose from a full year's
worth of potential GMS as-of dates. Additionally, the proposed range
would include only dates prior to the release of the given stress test
cycle's GMS for notice and comment. Therefore, firms subject to the GMS
would not be able to use their knowledge of the GMS as-of date to
update their balance sheet positions or adjust their portfolios to
minimize stress losses without a commensurate reduction in risk
profile.
In conjunction with the proposal to change the GMS as-of date
window, the Board also proposes to change the date by which the Board
needs to notify affected firms of this as-of date from March 1 of the
year in which the stress test occurs. Unless it determines otherwise,
the Board must notify affected firms of the GMS as-of date by October
15 of the year prior to the year in which the stress test is performed.
The Board would continue to provide firms with a description of the
GMS, as finalized, by March 1 of the calendar year in which the stress
test occurs. Additionally, to conform to the proposed changes to the
stress test timeline, the Board proposes to change the date by which
the Board must notify firms that they are required to include
additional components or scenarios in their company-run stress test
from December 31 to September 30 of the year preceding the stress test.
This change would ensure that firms are aware of the components to
which they would be subject prior to the annual publication of the
scenarios for notice and comment.
Question 27: What are the advantages and disadvantages of modifying
the window for the GMS as-of date in the stress test from October 1 of
the calendar year one year prior to the year in which the stress test
is performed through March 1 of the year in which the stress test is
performed, to a date that is no earlier than October 1 of calendar year
two years prior to the year in which the stress test is performed and
that precedes October 1 of the calendar year one year prior to the year
in which the stress test is performed? What alternative GMS as-of date
ranges, if any, should the Board consider, and why? In addition to
changing the GMS as-of date window, what other changes, if any, should
the Board consider making to the stress test timeline? What effects, if
any, would changing the window for the GMS as-of date have on any other
aspects of the stress test or the stress test timeline?
Question 28: What are the advantages and disadvantages of the
proposed dates by which the Board would notify firms of the GMS as-of
date, provide a description of any associated components, and notify
firms of any additional components that they are required to include in
their supervisory and company-run stress tests? What alternative dates,
if any, should the Board consider for these activities and why? For
example, to better ensure that more stakeholders provide input into the
proposed GMS, the Board could wait until the scenarios are final before
notifying firms which components they must include in their company-run
run stress tests.
Question 29: The GMS only considers a firm's positions on one as-of
date and only under one set of shocks. Should the Board consider
alternative approaches to further increase the risk capture of the GMS,
such as applying the GMS to more than one as-of date or more than
[[Page 51874]]
one set of shocks for a given stress test? What would be the advantages
and disadvantages of these alternative approaches? What other
approaches should the Board consider to improve the risk capture of the
GMS and why?
C. Amendment to the Dividend Add-On Component Calculation
The dividend add-on component of the stress capital buffer
requirement currently comprises planned dividends in the fourth through
seventh quarters of the planning (or projection) horizon of the
supervisory stress test.\78\ Under the current framework, the planned
dividends that are incorporated in the stress capital buffer
requirement align with the effective date of the stress capital buffer
requirement (that is, October 1 generally is the first day of the
fourth quarter of the existing planning horizon) and last for the one-
year period through which the stress capital buffer requirement is
expected to be effective (that is, through the seventh quarter of the
existing planning horizon, after which the following year's stress
capital buffer requirement would be expected to take effect).
---------------------------------------------------------------------------
\78\ See 12 CFR 225.8(d)(16); 12 CFR 238.130. The planning (or
projection) horizon for the supervisory stress test is nine
consecutive quarters starting on the jump-off date of the
supervisory stress test.
---------------------------------------------------------------------------
As part of this rulemaking, the Board is proposing to change the
jump-off date of the stress test from December 31 to September 30. To
maintain alignment between the dividend add-on component of the stress
capital buffer requirement and the one-year period during which the
requirement typically is effective, the Board proposes to change the
dividend add-on component to cover dividends issued in quarters five
through eight, instead of quarters four through seven, of the planning
horizon of the supervisory stress test. This change involves updates to
the capital plan rules, at Regulation Y and Regulation LL, to any
references to the relevant quarters of the planning horizon.\79\ This
proposed revision is intended to maintain the alignment between the
dividend add-on component and the one-year period during which the
stress capital buffer requirement generally is effective, assuming the
proposal to move the jump-off date of the stress test to September 30
is adopted. If this aspect of the proposal is not adopted, then the
Board would not adjust the planning horizon period for planned
dividends.
---------------------------------------------------------------------------
\79\ 12 CFR 225.8(f)(2)(i)(C)(1); 12 CFR 225.8(f)(4); 12 CFR
225.8(h)(2)(ii)(A); 12 CFR 225.8(h)(2)(ii)(B); 12 CFR 225.8(k)(2);
12 CFR 238.170(f)(2)(i)(C)(1); 12 CFR 238.170(f)(4); 12 CFR
238.170(h)(2)(ii)(A); 12 CFR 238.170(h)(2)(ii)(B); 12 CFR
238.170(k)(2).
---------------------------------------------------------------------------
Such a change to the planning horizon period has likewise been
proposed as part of the Board's proposed Modifications to the Capital
Plan Rule and Stress Capital Buffer Requirement, issued in April 2025,
in order to similarly maintain alignment between that proposal's
updates to the effective date of a firm's stress capital buffer
requirement and the dividend add-on component.\80\ Should both
proposals be finalized as proposed, the Board would expect to adjust
the dividend add-on component of the stress capital buffer requirement
to maintain alignment between the dividend add-on component and the
one-year period in which the stress capital buffer requirement
generally is effective. In such an instance, the Board would expect to
change the dividend add-on component to cover dividends issued in
quarters six through nine of the planning horizon of the supervisory
stress test.
---------------------------------------------------------------------------
\80\ See 90 FR 16843 (Apr. 22, 2025).
---------------------------------------------------------------------------
Question 30: What would be the advantages and disadvantages of the
proposed change to the dividend add-on component of the stress capital
buffer requirement?
VII. Revisions to the FR Y-14A/Q/M
To reduce regulatory reporting burden, support the proposed model
changes, and improve risk capture, the Board is proposing several
revisions to the FR Y-14A/Q/M. To reduce regulatory reporting burden,
the Board is proposing to remove items and documentation requirements
that are no longer needed to conduct the supervisory stress test. For
example, the proposal would remove certain FR Y-14 supporting
documentation requirements that are no longer needed to assess a firm's
FR Y-14 submission. The Board also proposes to collect additional data
that would support the supervisory stress test models and improve risk
capture. For example, to capture data in a manner that aligns better
with the treatment of private equity under the macroeconomic scenario,
the proposal would include revisions for reporting private equity
exposures and associated hedges. Additionally, to broaden the
consideration of hedges and revenue and loss sharing agreements in the
stress test, the Board is proposing revisions that would capture more
data on various types of hedges or revenue and loss sharing agreements.
Lastly, the Board is proposing several minor revisions to clarify the
FR Y-14 instructions or align with the proposed changes to the stress
test timeline. The proposed revisions are described in Section XI.A of
this Supplementary Information.
VIII. Proposed Changes to the Stress Test Modeling Framework
The Board is proposing to use the models described in the documents
posted on the Board's website to generate results for the 2026
supervisory stress test. Included in these descriptions are some model
specifications that were not used to conduct the 2025 supervisory
stress test but are proposed to be used for the 2026 supervisory stress
test. These proposed model changes are summarized in Section
VIII.A.\81\ In addition, a detailed description of and rationale for
each of these proposed model changes is provided in a separate document
posted on the Board's website with the comprehensive model
documentation. Section VIII.B of this Supplementary Information
provides an analysis of the potential effects of the proposed changes.
Based on this analysis, implementing the proposed model changes and
proposed revisions to the global market shock scenario design in the
2024 and 2025 stress tests would have, independent of other factors,
increased the aggregate projected common equity tier 1 (CET1) stress
ratio, on average, by 29 basis points. This is equivalent to a
reduction in stress capital buffer requirements of approximately 23
basis points or approximately 2.2 percent of current required capital.
---------------------------------------------------------------------------
\81\ These proposed changes would constitute ``model changes''
under the proposed definition of ``model change,'' as discussed in
Section IV.B of this Supplementary Information.
---------------------------------------------------------------------------
A. Proposed Changes to Stress Test Models
The Board is proposing several changes to the supervisory stress
test models for the 2026 stress test, which are discussed in more
detail in the Model Changes document provided on the Board's website,
at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. More significant proposed changes to the Pre-provision Net
Revenue and Operational Risk Models are described within the
comprehensive model documentation, also available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. The
Board is requesting comment on the proposed changes, together with the
model documentation.
With respect to the credit risk models, the Board is proposing to
change how it uses geography in scenario variables (First Lien, Home
Equity, Credit Cards,
[[Page 51875]]
Auto, and Commercial Real Estate Models); change how it treats
foreclosures under judicial supervision (First Lien and Home Equity
Models); change how it calculates loss given default for international
loans (Commercial Real Estate and Corporate Models); change how it
includes losses attributable to accrued interest and carrying costs
(First Lien and Home Equity Models); change how it uses multipliers in
the Provisions Model; revise the mortgage loss given default model in
the First Lien Model; revise the bank card model in the Credit Card
Model; change how it projects losses on auto leases in the Auto Model;
and update the probability of default, loss given default, and exposure
at default components in the Corporate Model.
With respect to the market risk models, the Board is proposing to
update several of its market risk models for the 2026 stress test,
including to simplify the Yield Curve Model; adjust its process for
projecting credit valuation adjustments for derivative positions in the
Credit Valuation Adjustment Model; lower the loss given default
assumption amount and loan equivalent factor parameter in the Fair
Value Option Model; update and simplify the Securities Model; and
exclude additional counterparties in the Largest Counterparty Default
Model.
With respect to the net revenue models, the Board is proposing an
alternative suite of pre-provision net revenue component models that
depart from the current panel regression-based approach. This
alternative suite is described in the Pre-provision Net Revenue Model
documentation, available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm. The Board is also proposing
to discontinue the current regression model used to project operational
risk losses and instead project losses with a distributional model.
This alternative model is described in the Operational Risk Model
documentation, also available at https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
Aggregate impacts on regulatory capital of the model changes
described above are small (see Table 2). Across risk stripes, the
proposed model changes would reduce credit, market, and operational
losses, which would be balanced by the effects of the proposed model
changes to the Pre-provision Net Revenue Model. Across firm categories,
GSIBs would observe modest increases in aggregate projected CET1 stress
ratio under the proposed changes. Firms subject to Category II-III
standards would also observe a modest increase in their projected CET1
stress ratio.
Question 31: The Board invites public comment on these proposed
model changes. What other changes, if any, should the Board consider
implementing in the 2026 stress test cycle, either instead of or in
addition to the proposed changes?
Question 32: What other information or data should the Board
consider to assess the quantitative economic impact of the proposed
model changes and why?
B. Analysis of Proposed Model Changes
To further enhance the transparency of the stress test models, this
section analyzes the potential effects of the proposed model changes
described in Section VIII.A of this Supplementary Information, and the
liquidity horizon revisions to the global market shock scenario design
described in Section IX.H of this Supplementary Information, that
inform the Board's determination of firms' stress capital buffer
requirements.
In aggregate, the stress test model and scenario changes are not
expected to materially change capital requirements for firms subject to
the supervisory stress test, across various stress scenarios and jump-
off conditions at the start of the test. To illustrate the effect of
these proposed model changes, this analysis averaged the impact of
these changes on the CET1 stress ratio for a balanced sample of 30
firms subject to the 2024 stress test and expected to participate in
the 2026 stress test, then aggregated the averages.\82\ The analysis
estimates that the proposed model and scenario changes, independent of
other models and components, could have resulted in an increase of 29
basis points in the average aggregate CET1 stress ratio. This is
equivalent to a reduction in stress capital buffer requirements of
approximately 23 basis points or approximately 2.2 percent of current
required capital. The analysis estimates that the model changes would
reduce stress capital buffer requirements by approximately 13 basis
points, and that the revisions to the global market shock scenario
design, described in Section IX.H of this Supplementary Information,
would reduce stress capital buffer requirements by approximately 10
basis points. For U.S. GSIBs, the analysis estimates a decline of 25
basis points of stress capital buffer requirements.
---------------------------------------------------------------------------
\82\ This analysis used the 2024 and 2025 scenarios,
respectively, and the same data used for those years' stress tests.
The estimated impact of these changes remains highly sensitive to
the stress test scenario and firm-specific data for each year. While
the precise impact will vary each year based on stress test
scenarios and specific firm data, Board analysis across a range of
conditions shows that capital requirements should remain essentially
unchanged.
---------------------------------------------------------------------------
As the U.S. banking system's 13.0 percent CET1 capital ratio (8.2
percent leverage ratio) is well within the estimated optimal range in
the literature,\83\ the net benefit of modest changes to the overall
level of banking system capital is small.\84\ However, as discussed
further below, the proposed model changes have varied effects on
capital requirements across loss type and firm category.
---------------------------------------------------------------------------
\83\ For discussions of optimal bank capital, see generally
Basel Committee, ``An Assessment of the Long-Term Economic Impact of
Stronger Capital and Liquidity Requirements'' (Aug. 2010), https://www.bis.org/publ/bcbs173.pdf (``BCBS 2010 study''); see also I.
Fender & U. Lewrick, Adding it All Up: The Macroeconomic Impact of
Basel III and Outstanding Reform Issues, BIS Working Paper No. 591
(Nov. 2016) (``Fender and Lewrick (2016)''), https://www.bis.org/publ/work591.pdf; D. Miles et al., Optimal Bank Capital, 123 The
Econ J. 1, 29 Table 10 (Mar. 2013) (``Miles et al. (2013)''),
https://academic.oup.com/ej/article/123/567/1/5080596; M. Brooke et
al., Measuring the Macroeconomic Costs and Benefits of Higher UK
Bank Capital Requirements, Bank of England, Financial Stability
Paper No. 35 (Dec. 2015) (``Brooke et al. (2015)''), https://www.bankofengland.co.uk/-/media/boe/files/financial-stability-paper/2015/measuring-the-macroeconomic-costs-and-benefits-of.pdf; S.
Firestone et al., An Empirical Economic Assessment of the Costs and
Benefits of Bank Capital in the United States, 101 Federal Reserve
Bank of St. Louis Rev. 203, 203-30 (2019) (``Firestone et al.
(2019)''), https://doi.org/10.20955/r.101.203-30; B. Soederhuizen,
et al., Optimal Capital Ratios for Banks in the Euro Area, 69 J.
Fin. Stability, Art. No. 101164 (Dec. 2023) (``Soederhuizen et al.
(2023)''), https://doi.org/10.1016/j.jfs.2023.101164; J. Barth & S.
Matteo Miller, Benefits and Costs of a Higher Bank `Leverage
Ratio','' 38 J. Fin. Stability 37, 37-52 (Oct. 2018) (``Barth and
Miller (2018)''), https://doi.org/10.1016/j.jfs.2018.07.001; J.
Dagher et al., Benefits and Costs of Bank Capital, IMF Staff
Discussion Note SND/16/04 (Mar. 2016) (``Dagher et al. (2016)''),
https://www.imf.org/external/pubs/ft/sdn/2016/sdn1604.pdf.
\84\ Ratios are based on the aggregate of all FR Y-9C filers as
of Q1 2025, which generally excludes holding companies with less
than $3 billion in consolidated assets and depository institutions
without parent holding companies. The aggregate CET1 ratio
additionally excludes holding companies that have opted in to the
Community Bank Leverage Ratio requirement, and reflects standardized
risk-weighted assets.
---------------------------------------------------------------------------
Based on this analysis, the proposed model changes are expected to
result in more risk-sensitive capital requirements, independent of
their effect on the level of requirements. Specifically, implementation
of the proposed model changes would render the models more stable,
likely reducing misalignment between firms' losses under stress and
their respective stress capital buffer requirements. To the extent that
the stress capital buffer requirements are affected by these proposed
model changes and are a part of firms' most-binding capital
constraint,\85\ the proposed model
[[Page 51876]]
changes would thereby improve the risk sensitivity--and efficiency and
effectiveness--of capital requirements.
---------------------------------------------------------------------------
\85\ The capital requirements of firms with stress losses plus
dividend add-ons reliably below the 2.5 percent capital conservation
buffer would be unaffected by the proposed model changes.
---------------------------------------------------------------------------
This analysis recognizes that the limited overall effect on
stressed CET1 capital ratios masks significant variation across the
different loss drivers. As shown in Table 2 below, the proposed model
changes could result in less severe credit, market, and operational
loss estimates--which would be driven by overhauling the wholesale
corporate probability of default model and discontinuing the
macroeconomic regression approach for operational risk loss estimation,
as described further in the Corporate Model and Operational Risk Model
descriptions. However, the proposed changes to the Pre-provision Net
Revenue Model would offset these loss reductions. By reducing the
reliance of net revenue projections on recent outcomes and relying more
on firm projections of net noninterest income, the projections of net
revenue would be more consistent with a stress scenario and would
better align with firms' projections.
[GRAPHIC] [TIFF OMITTED] TP18NO25.037
Table 3 below provides a separate analysis of estimates of stress
losses across firm types that are subject to the stress capital buffer
requirement. The analysis shows the reduction in hypothetical stress
losses is concentrated at larger firms.
[GRAPHIC] [TIFF OMITTED] TP18NO25.038
IX. Proposed Changes to the Scenario Design Policy Statement
The Board is also proposing to make several changes to the Scenario
Design Policy Statement. While many of these proposed changes are
technical in nature, this section identifies substantive changes and
requests comment on those proposed changes.
Question 33: The Board invites comment on all aspects of the
technical and substantive proposed revisions to the Scenario Design
Policy Statement. What are the advantages and disadvantages of these
proposed changes? What would be the advantages and disadvantages if the
Board were to consider describing the Board's expectations for
additional components of the scenario design framework?
A. Changes to the Background and Overview and Scope Sections
The Board is proposing to make limited changes to the first two
sections of the Scenario Design Policy Statement, which address
background and overview and scope topics, respectively. In the
background section, the Board would clarify that the stress tests
primarily focus on credit risk, operational risk, and market risk. The
inclusion of operational risk in this list helps clarify the Board's
continued focus on designing a supervisory tool that makes a valuable
forward-looking assessment of large financial companies' capital
adequacy under hypothetical economic and financial market conditions.
The Board would also clarify that it expects to provide only two
different sets of macroeconomic scenarios for both the supervisory and
company-run stress tests. These two sets of macroeconomic scenarios are
the baseline and severely adverse scenario. This change would clarify
the quantity of macroeconomic scenarios the Board expects to provide,
consistent with the removal of a separate adverse scenario.\86\
---------------------------------------------------------------------------
\86\ 84 FR 59032, 59061 (Nov. 1, 2019).
---------------------------------------------------------------------------
In the overview and scope section, the Board would make conforming
edits to the description of the organization of the Scenario Design
Policy Statement to reflect the changes discussed earlier in this
proposal.
Question 34: What additional changes, if any, should the Board
consider making to these sections, and why? What would be the
advantages and disadvantages of providing more than two scenarios? What
are the
[[Page 51877]]
advantages and disadvantages of the Board's continued focus on credit,
operational, and market risk?
B. Changes to the Content of the Stress Test Scenarios Section
The Board is proposing to make two general changes to this section,
which describes the Board's expectations for the content of the
published stress test scenarios.
First, as described below, this section would be amended to clarify
that the Board expects to generally publish two different macroeconomic
scenarios: the baseline and severely adverse scenarios. This section
would also be revised to clarify that the Board expects to invite
comment on severely adverse scenarios.
Second, as described in Section IX.H of this Supplementary
Information, the Board is proposing to make certain changes related to
the global market shock component. See Section IX.H of this
Supplementary Information for a discussion of those changes.
Question 35: What additional changes, if any, should the Board
consider making to these sections, and why?
C. Approach for Formulating Macroeconomic Assumptions in the Baseline
Scenario
The Board is proposing to provide additional details describing the
process by which the Board would set the paths of the variables in the
baseline and severely adverse scenarios. In particular, the amendments
reflect that the Board would post on the Board's website a description
of the macroeconomic model utilized to support the construction of the
baseline and severely adverse scenarios in the annual stress test. By
posting a description of this model (the ``macroeconomic model for
stress testing'') on the Board's website, the Board expects to improve
the transparency, public accountability, and predictability around the
Board's scenario design framework, particularly with respect to the
baseline scenario and certain variables in the severely adverse
scenario. The Board recognizes that, while these enhancements are
consistent with the Board's goal of increased transparency in the
supervisory stress test, they may constrain the design of the scenario
paths for some variables to follow those prescribed by the
macroeconomic model for stress testing. Nevertheless, the Board expects
that other aspects of the proposed changes to the Scenario Design
Policy Statement will preserve sufficient flexibility to allow the
Board to adjust the severity of the annual scenario based on relevant
indicators of economic and financial conditions and other emergent
procyclical factors. Importantly, the Board uses these models to
generate paths for the scenario variables only. These models are used
solely for stress testing purposes and the output is not a forecast of
the Board.
Question 36: What are the advantages and disadvantages of adopting
a macroeconomic model for stress testing to guide the selection of
certain variables in the baseline and severely adverse scenarios?
Question 37: What additional changes, if any, should the Board
consider making to this section, and why?
D. Scenario Narrative: Refinement to the Recession Approach
A number of considerations contribute to the Board's formulation of
the severely adverse scenario. As a starting point, the basic approach
adopted by the Board is the recession approach--the notion that the
Board will construct a scenario informed by the historical paths of
macroeconomic and financial market variables across post-war U.S.
recessions. However, different recessions have differed in important
respects, and a simple recreation of a given episode or an average over
all recessions would fail to reproduce important potential stressors to
firms' balance sheets. Hence, in applying the recession approach, the
Board develops a specific narrative characterizing the hypothetical
recession represented by the scenario to help inform the specific paths
for scenario variables. This narrative combined with data are then
modified to account for the Board's stress testing principle of
conservatism alongside other considerations offered by the literature
on stress testing including a goal to develop sufficient severity and
credibility of the scenarios, and a goal to not add sources of
procyclicality to the financial system, as described below.\87\ This
section gives an overview of these considerations and other details,
providing a common structure for the discussion outlined in the guides
for individual variables under this framework, in Section IX.G of this
Supplementary Information.
---------------------------------------------------------------------------
\87\ 12 CFR 252, Appendix B.
---------------------------------------------------------------------------
The Recession Approach
The Board intends to continue to use a recession approach to
develop the severely adverse scenario. Under the recession approach,
the Board expects to specify the future paths of variables to reflect
conditions that characterize post-war U.S. recessions, generating
either a typical or specific recreation of a post-war U.S. recession.
The Board chose this approach in developing past scenarios, and in the
Scenario Design Policy Statement, because it has observed that the
conditions that typically occur in recessions--such as increasing
unemployment, declining asset prices, and contracting loan demand--can
put significant stress on firms' balance sheets. This stress can occur
through a variety of channels, including higher loss provisions due to
increased delinquencies and defaults, losses on trading positions
through sharp moves in market prices, and lower bank income through
reduced loan originations. For these reasons, the Board expects that
the paths of economic and financial variables in the severely adverse
scenario should, at a minimum, resemble the paths of those variables
observed during a recession. The guide for each variable in this
framework reviews the movements of that variable across past recessions
and bases the formulation of its scenario path on that analysis. While
the recession approach provides a starting point for the formulation of
the scenario, recessions are not all the same. The length and depth of
recessions differ, as do the parts of the economy and financial markets
that are most affected, so the Board must include other considerations
in its scenario design.
The Scenario Narrative
Because recessions have differed in cause, character, and
consequence--from oil price shocks and housing slumps to asset-price
busts and pandemics, from short to long, and from mild to moderate to
severe--the Board augments the basic recession approach with an annual
scenario narrative. The annual scenario narrative provides qualitative
direction on how the Board builds that year's severely adverse
scenario.
While some specifics of the narrative may be adjusted on a year-to-
year basis to reflect developments in the macroeconomic and financial
environment, the overall narrative motivating scenario design will be
that of a sharp recession triggered by an adverse shock to financial
markets. Under the proposal, the Board expects that the macroeconomic
scenario used in the Board's annual supervisory severely adverse
scenario will begin with a sudden and significant increase in
uncertainty and associated rapid deterioration in risk appetite that
cause a spike in financial market volatility and a sharp decline in
many U.S. and
[[Page 51878]]
foreign financial assets. The resulting turmoil would disrupt funding
markets and lead to widespread deleveraging, including forced sales of
illiquid assets at fire sale prices by a range of financial firms and
some temporary breakdowns in the typical correlations between financial
asset prices. (Such sharp changes in financial conditions have been
observed previously in response to the outbreak of COVID-19 or regional
wars, the failure or distress of a large financial institution, or
sudden shifts in the economic policies in advanced economies.)
Under the Board's recession approach, the Board expects that,
although financial market functioning returns to normal within a few
months of the initial shock, uncertainty remains high and risk appetite
remains low for an extended period. The sustained flight to quality
would be expected to push down risk-free interest rates but keep credit
conditions tight and financial asset prices depressed for several
quarters. The market dysfunction would cause a contraction in the
supply of credit from other types of financial intermediaries that
would create demands on banks to provide substantial liquidity to
existing customers with formal credit lines. Banks would also make ad
hoc decisions to support customers without formal arrangements when
doing so could lead to lower losses on their existing loans.\88\ This
shift in demand for credit toward banks from other financial
intermediaries would lead to banks' balance sheets remaining constant
even as overall credit demand declines.\89\ This feature of the
scenario is supported by the stress testing principle of
conservatism.\90\ To that end, maintaining higher capital requirements
during periods of economic expansion ensures that stress tested firms
employ sufficient capital to absorb losses and support the economy
during a downturn.
---------------------------------------------------------------------------
\88\ For example, in June 2020 the Federal Financial
Institutions Examination Council issued interagency guidance to bank
examiners stating, ``examiners will not subject a . . . modified
loan to adverse classification solely because the value of the
underlying collateral has declined . . ., provided that the borrower
has ability to repay . . .'' See Interagency COVID-19 Examiner
Guidance, https://www.federalreserve.gov/newsevents/pressreleases/files/bcreg20200623a1.pdf.
\89\ Commercial and industrial loans grew 20 percent in 2007 as
credit markets seized at the beginning of the 2007-2009 financial
crisis. See M. Bech & Tara Rice, Profits and Balance Sheet
Developments at U.S. Commercial Banks in 2008, 95 Fed. Rsrv. Bull.
A57-97 (2009), https://www.federalreserve.gov/pubs/Bulletin/2009/articles/bankprofit/default.htm. For COVID-19, see H. Ennis & A.
Jarque, Bank Lending in the Time of COVID, Federal Reserve Bank of
Richmond Economic Brief No. 21-05 (Feb. 2021).
\90\ 12 CFR 252, Appendix B.
---------------------------------------------------------------------------
In the scenario, the news from financial markets would cause near-
immediate decisions by consumers to curtail spending and by businesses
to cut payroll and cancel planned investments, leading to a demand-
driven contraction in economic activity putting downward pressure on
inflation. The initial disruption to spending and employment along with
tightening credit conditions would trigger a negative feedback loop
that results in further declines in payrolls, investment, and spending
in subsequent quarters. With businesses shrinking or failing in the
scenario, demand for commercial real estate would decrease
significantly relative to supply, leading to large declines in
commercial property prices. Meanwhile, rising household financial
distress would lead to increased supply of homes for sale and reduced
household formation, which would depress residential real estate
markets.
The financial market dysfunction and deepening recession in the
United States would spill over to its major trading partners, including
the euro area, United Kingdom, Japan, and Developing Asia. Those areas
would experience declines in economic activity commensurate with the
global slowdown running from 2008 to 2010. Consistent with existing
stress testing principles, this scenario assumes that permanent
government stabilization programs (e.g., unemployment insurance) and
monetary policy in the United States and elsewhere would function
normally, but that there would be no extraordinary measures taken by
fiscal or financial authorities to support the economy or financial
markets during this time. The specific implications of this narrative
for scenario variables are detailed in each guide, but the narrative
interacts importantly with the recession approach: financial recessions
often exhibit different properties than other recessions, as they are
often steeper, deeper, and more drawn-out than typical, non-financial
recessions.\91\ Adopting this scenario narrative reflects a principle
of conservatism, and is in line with recommendations from the stress
testing literature, as discussed in Section IX.F of this Supplementary
Information.
---------------------------------------------------------------------------
\91\ See, e.g., C. Reinhart & K. Rogoff, This Time Is Different:
Eight Centuries of Financial Folly (2009).
---------------------------------------------------------------------------
Question 38: The Board invites comment on all aspects of how the
Board designs the scenario narrative in the annual stress test. What
are the advantages and disadvantages of adopting this financial
recession approach? What other approaches, if any, should the Board
consider adopting, and why? What adjustments, if any, to the financial
recession approach should the Board consider adopting, and why?
Adding Salient Risks to the Severely Adverse Scenario
Consistent with the Scenario Design Policy Statement, under this
proposal, the Board expects that the severely adverse scenario would be
developed to reflect the current level of vulnerabilities or risks to
the banking sector that are apparent in relevant indicators of economic
and financial conditions. The Board anticipates that the proposed
guides for certain scenario variables described below provide an
appropriate range of values to design the severely adverse scenario in
most years. The waxing and waning of relevant indicators of economic
and financial conditions will inform the Board's decisions about where
to set the value of those parameters within those ranges for each
variable.
The Board continues to expect that there will be some important
instances when it will be appropriate to augment the recession approach
with salient risks and to set variables values inside of, and in some
cases, outside of the ranges and values provided in the guides in the
Scenario Design Policy Statement. As a result, each year, the Board
will consider particular risks to the financial system and to the
domestic and international macroeconomic outlook identified by its
economists, bank supervisors, and financial market experts. The Board,
using its internal analysis and supervisory information and in
consultation with the Federal Deposit Insurance Corporation and the
Office of the Comptroller of the Currency, will then determine whether
any of those risks appear significantly more elevated than usual or,
conversely, whether risks are unusually low at a particular time, such
that they cannot be appropriately reflected by choosing values within
the ranges of the proposed guides. In those cases, which it expects to
be infrequent, the Board will make appropriate adjustments to the paths
of specific economic variables. These adjustments will not always be
reflected in the general severity of the recession and, thus, all
macroeconomic variables; rather, the adjustments will sometimes apply
to a subset of variables to reflect co-movements in these variables
that are historically less typical.
To assist the public in assessing the use of salient risks in the
scenario, the Board considered the following examples. A stress test
initiated in a
[[Page 51879]]
period of unusually high uncertainty and rapid deterioration in
economic and financial conditions, such as the first quarter of 2009 or
the first quarter of 2020, likely would prove challenging for the
ranges in this proposed framework. In each case, the prevailing
conditions made it plausible that key variables would settle beyond the
range of their previous peak or trough values, on which the guides for
the variables in the severely adverse scenario are calibrated. Although
the unemployment guide remained flexible enough to respond to the spike
in the unemployment rate to nearly 15 percent during the first months
of the COVID-19-related business closures in 2020, the paths of other
variables may have needed to be adjusted more severely if the economy
had not recovered as quickly as it did.
As another example, the Board may become increasingly concerned
about vulnerabilities related to a particular asset class that was
experiencing rapid and persistent price increases supported by
increasingly leveraged investors. Those circumstances existed in the
housing market in the early 2000s and may have tested the credibility
of a guide framework based solely on past performance of home prices,
given that up until then, the price index for homes the Board uses for
stress testing had rarely experienced a decline.\92\
---------------------------------------------------------------------------
\92\ The Board uses the Price Index for Owner-Occupied Real
Estate, Z.1 Release (Financial Accounts of the United States),
Federal Reserve Board (series FL075035243.Q).
---------------------------------------------------------------------------
Sometimes, the salient risk may arise within an asset class. The
Board most recently incorporated this type of salient risk in the 2024
stress test scenario. That year, the Board noted unusually high
vulnerabilities in types of commercial properties that could be most at
risk for a sustained drop in income and asset values due to the
prevalence of remote work.\93\
---------------------------------------------------------------------------
\93\ See Board, 2024 Stress Test Scenarios, ``Additional Key
Features of the Severely Adverse Scenario,'' at 12-13 (Feb. 2024),
https://www.federalreserve.gov/publications/files/2024-stress-test-scenarios-20240215.pdf.
---------------------------------------------------------------------------
The Board is proposing two changes to its consideration of salient
risks in the severely adverse scenario. First, the Board would remove
paragraph 4.2.4(d) from the Scenario Design Policy Statement. Removing
this paragraph could help improve the transparency of the scenario
design process by limiting the Board's expectations for considering
risks of uncertain significance. While this approach would reduce the
Board's ability to test for emerging and untested risks in the
financial system through the severely adverse scenario, the Board
expects that the remaining components of the Board's supervisory stress
test should be sufficient to establish a credible severely adverse
scenario.
Second, where the Board does consider salient risks in designing
the severely adverse scenario, the Board will endeavor to disclose and
explain the Board's reasoning in the Board's publication of the annual
stress test scenarios, and subsequently adjust those aspects of the
scenario, if necessary, in response to those comments.
Question 39: What are the advantages and disadvantages of the
Board's approach to considering salient risks? What additional or
alternative approaches, if any, should the Board consider for the
consideration of salient risks? What additional or alternative
circumstances should the Board take into account when evaluating
whether to consider salient risks, if any?
E. Changes to Construction of Certain Variables in the Severely Adverse
Scenario
As noted above, the Board finalized changes to the Scenario Design
Policy Statement in 2019 that established a guide that it would use in
setting the size of the maximum change in the unemployment rate and the
timing of its peak. The Scenario Design Policy Statement also
introduced a guide to govern the size of the maximum decline in house
prices in the severely adverse scenario. This proposal maintains those
features of the guides for those two variables, introduces guides that
will be used to set the changes in the values, and the timing of those
changes, for more variables in the severely adverse scenario, and
provides additional context for the path of each variable before it
reaches the maximum change. In addition, the Board is separately
disclosing a specific macroeconomic model that it proposes to use to
translate the paths of certain variables that are set using the
proposed guides into internally consistent projections for the
remaining variables, such as the 3-month Treasury bill rate, GDP,
Disposable Personal Income (DPI), and inflation.
In addition to updating existing guides for the unemployment rate
and house prices, the Board is proposing to establish a guide for each
of the following variables: equity prices; the VIX index; 5-year
Treasury yields; 10-year Treasury yields; BBB corporate bond yields;
mortgage rates; commercial real estate prices; and certain
international scenario values. These include all but one of the
remaining financial market variables typically included in the domestic
severely adverse scenario disclosure each year (the exception being the
3-month Treasury bill rate, as discussed below).
The Board uses guides to inform its determination of the behavior
of these financial market variables in the severely adverse scenario,
rather than model predictions, for several reasons. Although the
parameters of the guides are calibrated based on an analysis of
historical changes in those variables during recessions and the
resulting set of scenario paths typically would be consistent with
historical co-movements in those variables, using explicit forward-
looking models of these variables to determine scenario paths would be
inconsistent with several stress testing principles, such as simplicity
and transparency, as described below.
Under a model-driven approach to determine the paths of these
variables, each model would require the Board to identify, design,
test, explain, and publish additional assumptions, variables, formulas,
and parameters that would drive the results of the model. Models of
financial market variables can be particularly unreliable during
periods of severe stress like the environment envisioned by the
hypothetical severely adverse scenario.\94\ Thus, the model-driven
approach to determining these variables would contrast with the stress
testing principle of using simpler and more transparent approaches,
where appropriate.
---------------------------------------------------------------------------
\94\ T.C. Green & S. Figlewski, Market Risk and Model Risk for a
Financial Institution Writing Options, 54 J. Fin. 1465-99 (Dec.
1999).
---------------------------------------------------------------------------
The Board believes that the guide-based approach also better
achieves the stress testing principle of using a stable process that is
reliably able to capture the impact of economic stress. These simple,
transparent guides also will allow the Board to use its judgment at
times when it is necessary to account for conditions that are plausible
even if they have not been observed previously, consistent with the
stress testing principle of conservatism. Finally, the guides better
preserve the Board's ability to adjust the severity of the stress test
to avoid adding to procyclical forces, when doing so is appropriate and
consistent with fostering financial stability. The Board's judgment
about the appropriateness of the annual stress test scenarios will
reflect changes in the specific risks or vulnerabilities that the
Board, in consultation with the other federal banking agencies,
determines should be considered in the annual stress tests.\95\
---------------------------------------------------------------------------
\95\ See 84 FR 6651, 6656 (Feb. 28, 2019).
---------------------------------------------------------------------------
[[Page 51880]]
The paths for the remaining variables in the domestic scenario--
GDP, DPI, inflation, and the 3-month Treasury rate--will be informed by
the Board's macroeconomic model for stress testing.\96\ In contrast to
the guide-based approach described above for certain variables, the
Board uses a model-driven approach for these remaining variables
because they are particularly suited to model projections that are
simple to produce and explain. As explained in the model documentation
available on the Board's website, that model uses a set of well-studied
longer-run economic relationships that have proven to be useful in a
variety of economic conditions and modeling frameworks. These include
Okun's Law, a Phillips Curve, and an inertial Taylor Rule.\97\ The
Board acknowledges that increasing the predictability of the paths of
scenario variables in this way could reduce the dynamism of the stress
test or incent firms to optimize their portfolios in ways that reduce
capital requirements, perhaps without a commensurate reduction in risk.
However, the guides and the model are constructed to remain flexible
enough to ensure that the Board can adjust the severely adverse
scenario to capture emerging risks and changes in the level of systemic
risk since the previous stress test in a timely fashion. This
flexibility includes the ability to increase scenario severity when
systemic risks may have built up during robust economic expansions or
periods when risk appetite is high or to avoid adding sources of
procyclicality through the stress test. The proposal continues to
ensure that the scenarios maintain a minimum severity level, even when
economic and financial conditions are strained. Setting a floor for the
severity of the scenario is appropriate because risks that built up
during an economic expansion can persist at financial intermediaries
during downturns and because firms that are under stress sometimes take
imprudent risks that they believe will facilitate recovery.\98\
---------------------------------------------------------------------------
\96\ This approach is consistent with how the Board has designed
recent stress test scenarios. See id. at 6659.
\97\ See https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
\98\ See J. Peek & E. Rosengren, Unnatural Selection, Perverse
Incentives and the Misallocation of Credit in Japan, 95 Am. Econ.
Rev. 1144-66 (2005).
---------------------------------------------------------------------------
The Board also considered that employing the guides or the
macroeconomic model for stress testing sometimes may reduce the
severity of some aspects of the scenario relative to what the currently
less-constrained scenario design process would achieve, and in other
cases it may result in higher severity for some aspects of the scenario
than might otherwise be the case. The flexibility in the guides should
be sufficient for the Board to account for those eventualities by
choosing offsetting values across multiple guides that create the
appropriate overall severity of the scenario.
Question 40: What are the advantages and disadvantages of using
guides and the macroeconomic model for stress testing to guide the
setting of scenario variables in the severely adverse scenario? What,
if any, alternatives to using a macroeconomic model to set the
projection paths of other variables should the Board consider?
F. Scenario Design Principles Derived from Stress Testing Literature:
Severity, Credibility, and Procyclicality
In designing the guides for the construction of the severely
adverse scenario presented in this framework, the Board is informed by
the stress testing literature, which provides certain principles for
scenario design,\99\ which are also reflected in the Board's Stress
Testing Policy Statement.\100\ First, the literature emphasizes the
need for adequately severe scenarios, even when the economy and
financial system are in a stressed condition--complementing the Board's
principle of conservatism.\101\ Second, the literature offers insights
on how historical data should inform the design of an adequately severe
scenario, augmenting the Board's recession approach. Third, the
literature highlights the need for stress tests to avoid adding to
other sources of procyclicality in the financial system. In explaining
the paths for variables in the severely adverse scenario, the guides
provide specific applications of these principles, while this
introduction provides an overview of their general meaning and
rationale.
---------------------------------------------------------------------------
\99\ Some of the well-known contributions are T. Schuermann,
Stress Testing Banks, 30 International Journal of Forecasting 717-28
(2014) (``Schermann (2014)''); and N. Liang, Well-Designed Stress
Test Scenarios Are Important for Financial Stability, Brookings
Institution Paper (Feb. 2, 2018) (``Liang (2018)''), https://www.brookings.edu/articles/well-designed-stress-test-scenarios-are-important-for-financial-stability.
\100\ See 12 CFR 252, Appendix B.
\101\ Id.
---------------------------------------------------------------------------
The first principle derived from the literature concerns the need
for sufficiently severe scenarios. Plainly, insufficient stress test
severity can lead to adverse outcomes. Inadequately assessed risks lead
to an underassessment of the associated credit losses and capital
needs--the basic source of failures of many financial institutions
during the 2007-2009 financial crisis which the Board's stress tests
are meant to avoid. Frame et al. (2015) provide an in-depth analysis of
how the assessment of risks (or stress test) conducted by the Office of
the Federal Housing Enterprise Oversight (OFHEO) actually contributed
to the failures of the Federal National Mortgage Association (Fannie
Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac).\102\
Importantly, stress tests must be adequately severe both in good times
and in bad.
---------------------------------------------------------------------------
\102\ See S. Frame, C. Gerardi, & P. Willen, The Failure of
Supervisory Stress Testing: Fannie Mae, Freddie Mac, and OFHEO,
Federal Reserve Bank of Boston Working Paper No. 15-4 (2015),
https://www.bostonfed.org/publications/research-department-working-paper/2015/the-failure-of-supervisory-stress-testing-fannie-mae-freddie-mac-and-ofheo.aspx. OFHEO was the federal regulator of the
government-sponsored mortgage agencies, Fannie Mae and Freddie Mac.
---------------------------------------------------------------------------
In the context of stress testing during crises, in particular,
there are additional arguments against insufficient stress test
severity. Schuermann (2014) and Judge (2022) argue that insufficiently
severe stress test scenarios can erode credibility and trust and impede
timely and adequate policy responses to ongoing crisis developments,
thereby exacerbating a downturn.\103\ Bernanke (2013) also highlights
that stress tests in times of crisis should provide anxious investors
with credible information about prospective losses.\104\ This
literature points to the importance of sufficiently severe scenarios
for the health of the financial system, including by maintaining
credibility with the public and financial markets.
---------------------------------------------------------------------------
\103\ K. Judge, ``Stress Testing During Times of War,'' Handbook
of Financial Stress Testing (2022) (``Judge (2022)'').
\104\ B. Bernanke, ``Stress testing banks: What have we
learned?,'' Speech at the ``Maintaining Financial Stability: Holding
a Tiger by the Tail'' Conference (2013) (``Bernanke (2013)''),
www.federalreserve.gov/newsevents/speech/bernanke20130408a.htm.
---------------------------------------------------------------------------
Further evidence for the importance of sufficiently stressful
scenarios to maintaining public credibility comes from past U.S. stress
tests. For example, the rapid deterioration in the U.S. economy in
early 2009 led to realized unemployment rates that approached the peak
of the unemployment rate path in the severely adverse scenario used for
the Supervisory Capital Assessment Program (SCAP) in 2009.\105\ In
fact, the scenario peak for the unemployment rate hypothesized would
reach only 8.9 percent at the end of 2009, but as of
[[Page 51881]]
March 2009 the unemployment rate measured 8.5 percent and ultimately
the unemployment rate peaked at 10 percent in October of 2009.\106\
Because the results of the SCAP determined the amount of capital that
firms needed to raise in financial markets or through the Treasury's
Capital Assistance Program, a scenario that turned out to be
insufficiently severe could have left some firms undercapitalized and
failed to achieve the goal of stabilizing the financial system.\107\
---------------------------------------------------------------------------
\105\ See, e.g., E. Andrews & E. Dash, ``Government Offers
Details of Bank Stress Test,'' N.Y. Times (Feb. 25, 2009), https://archive.nytimes.com/www.nytimes.com/indexes/2009/02/26/todayspaper/index.html.
\106\ A similar concern related to insufficient scenario
severity followed the announcement of the European Union's stress
tests in 2018, with the criticism that the assumptions were milder
than conditions in the 2007-2009 financial crisis. See F. Guarascio,
``EU's 2018 Stress Test too Mild, Spared Weaker States--Auditors'',
Reuters (Jul. 10, 2019), https://www.reuters.com/article/business/
eus-2018-bank-stress-test-too-mild-spared-weaker-states-auditors-
idUSKCN1U5113/
#:~:text=The%20auditors%20said%20last%20year's,their%20risk%20rather%
20than%20size.
\107\ An explanation of the synergy between the SCAP and CAP is
available here: Supervisory Capital Assessment Program & Capital
Assistance Program (SCAP and CAP), U.S. Department of the Treasury,
https://home.treasury.gov/data/troubled-assets-relief-program/bank-investment-programs/scap-and-cap.
---------------------------------------------------------------------------
This example helps demonstrate the importance of the principle of
severity when considering historical data and current conditions in the
construction of an adequately severe scenario. While unemployment rates
are discussed at length in the unemployment guide below, the maximum
level of 8.9 percent specified in the 2009 SCAP, at the time, was well
beyond the level reached in most post-war recessions. At the time the
scenario was issued, a projected increase to 8.9 percent was thus very
severe compared to outcomes over the past quarter century, but
nonetheless proved lower than the actual realized peak in 2009.
That experience reinforces the need for the framework to support
variable paths that exceed levels observed in the historical data.
Choosing a historical scenario has a price--``it does not test for
anything new.'' \108\ While the recession approach dictates that
variable movements follow historical recessions, when current
conditions are already extreme, a credible scenario may replicate
historical recessions in terms of the size of movements previously
observed, leading to levels of variables that may exceed historical
levels. Several of the guides in this framework allow, at times, for
variables to exceed their historical range, either in levels or in
changes, in order to maintain adequate severity.
---------------------------------------------------------------------------
\108\ See Schuermann (2014), supra note 99.
---------------------------------------------------------------------------
Ultimately, no single scenario can account for all potential
contingencies. Therefore, the severely adverse scenario used in the
Board's annual stress test must be sufficiently severe to ensure that
banks will be resilient to a range of alternative and plausible
scenarios that could generate net losses that are of similar
magnitudes.\109\
---------------------------------------------------------------------------
\109\ See Liang (2018), supra note 99.
---------------------------------------------------------------------------
At the same time, the Board recognizes that the severity of the
annual stress tests potentially can have unintended effects on firms'
operations. For instance, the academic literature finds that stress
tests improve financial stability by reducing riskier bank
lending.\110\ Ensuring that firms are appropriately capitalized for the
risks they are taking is a goal of stress testing; however, if those
effects are not well aligned with the true riskiness of a particular
type of loan, then stress tests could unintentionally reduce banks'
credit supply. For instance, some evidence exists that counties in
which stress tested banks had high market share may have experienced a
lower supply of credit to small and young businesses, which are
generally considered riskier than established businesses but can
generate a disproportionate share of growth in employment and
income.\111\ However, other research concludes that businesses largely
offset the reduction in loans from banks that participate in the stress
tests with other sources of credit. Those sources include loans from
smaller banks not in the stress tests,\112\ debt issuance in capital
markets, or loans from nonbank financial institutions.\113\ Moreover,
these potential unintended effects on credit supply by stress tested
firms must be weighed against the benefits, discussed above, that more
credible stress tests bring to the economy and the financial system. By
ensuring that firms have sufficient quantity and quality of loss-
absorbing capital to cover the risks that they are taking, the stress
tests ensure the resilience and stability of the banking sector even in
circumstances when stresses take unexpected forms.
---------------------------------------------------------------------------
\110\ V. Acharya, A. Berger, & R. Roman, Lending implications of
U.S. bank stress tests: Costs or benefits?, 34 J. Fin.
Intermediation 58-90 (2018).
\111\ See S. Doerr, Stress Tests, Entrepreneurship, and
Innovation, 25 Rev. of Fin. 1609-1637 (Sep. 2021), https://doi.org/10.1093/rof/rfab007.
\112\ See K. Cort[eacute]s et al., Stress tests and small
business lending, 136 J. Fin. Econ. 260-279 (2021) (``Cort[eacute]s
(2021)'').
\113\ See J. Berrospide & R. Edge, Bank capital buffers and
lending, firm financing and spending: What can be learned from five
years of stress test results?, 57 J. Fin. Intermediation 1010-61
(2024) (``Berrospide (2024)''); T. Davydiuk, T. Marchuk, & S. Rosen,
Direct lenders in the U.S. middle market, 162 J. Fin. Econ. (2024)
103946 (``Davydiuk (2024)'').
---------------------------------------------------------------------------
The balance of those advantages and disadvantages of scenario
severity can change over time. Losses at financial institutions are
more likely to arise when the economy slows. Profits are more robust
during periods of economic growth, in turn increasing resources
available to cover future losses. In other words, capital is naturally
procyclical, having an underlying tendency towards a positive
correlation with financial conditions. Moreover, when underlying
conditions are favorable and firm losses are low, firms sometimes
project forward an expectation for low losses, paving the way to take
more risk.\114\ Conversely, when conditions are bad, firms may
overcompensate and restrict credit even to otherwise creditworthy
borrowers, exacerbating the downturn. Thus, firms' behavior may amplify
underlying procyclicality.
---------------------------------------------------------------------------
\114\ See A. Berger & G. Udell, The institutional memory
hypothesis and the procyclicality of bank lending behavior, 13 J.
Fin. Intermediation 458-495 (2004) (``Berger (2004)''); A.
Greenspan, ``Challenges facing community banks,'' Remarks before the
Independent Community Bankers of America (Mar. 8, 2000) (``Greenspan
(2000)''), https://www.federalreserve.gov/boarddocs/speeches/2000/20000308.htm.
---------------------------------------------------------------------------
Stress tests could, through different designs, either amplify or
mitigate this procyclicality. If stress tests are always more severe in
bad times, despite an expectation that conditions could soon improve,
then this severity would add undue stress to the financial system,
reducing financial intermediation with negative implications for the
macroeconomy. That said, the purpose of the stress test scenarios is
not to serve as an explicit countercyclical offset to the financial
system, but rather to ensure that the firms are properly capitalized to
withstand severe economic and financial conditions. Hence, the Board
adopts a middle path, seeking to specify the severely adverse scenario
to avoid adding sources of procyclicality to the financial system,
neither explicitly mitigating any existing procyclical tendencies nor
magnifying them. Indeed, Kohn and Liang (2019) argue that the ability
to adjust elements that potentially add procyclicality can be a major
benefit of stress tests as ``banks with forward-looking, less-
procyclical capital buffers will not pull back as much when a downturn
occurs.'' \115\
---------------------------------------------------------------------------
\115\ D. Kohn & N. Liang, Understanding the Effects of the U.S.
Stress Tests, Brookings Institute (Jul. 2019), https://www.brookings.edu/articles/understanding-the-effects-of-the-u-s-stress-tests/.
---------------------------------------------------------------------------
In summary, in formulating the guides presented in this framework,
the Board embraces three principles suggested by the literature: the
importance of severity, the importance of credibility,
[[Page 51882]]
and the importance of not adding to procyclicality.
Stress Testing Literature and the Principle of Flexibility
When considering these principles in light of the recession
approach and the scenario narrative, the Board identified the
importance of maintaining flexibility in the guides. While the Board
intends to increase the transparency, public accountability, and
predictability of stress tests through this proposal, these goals
should not come at the expense of the overall effectiveness of the
Board's stress tests.
For instance, predictability and transparency could be achieved
with a completely specified, entirely formulaic scenario that leaves no
flexibility. However, simple, fixed guides may not achieve at least one
of the goals of severity, credibility, or not adding to procyclicality.
A guide that always increased unemployment to a fixed level, say 10
percent, may not be credible or severe were the unemployment rate
already at or close to that level. A guide that always increased
unemployment by a fixed amount, say 4 percent, could add to
procyclicality by implying lower losses when unemployment was low in
good times and higher losses when unemployment was high in bad times.
More sophisticated formulations might improve on simple rules by
accounting for the factors affecting firms' balance sheets and overall
economic and financial conditions. For many types of economic
indicators used in the Board's scenario framework, however, a fixed
rule for the design of a scenario variable that satisfied the
principles related to procyclicality and severity laid out above could
require a complex structure that would violate the Board's principle of
simplicity.\116\
---------------------------------------------------------------------------
\116\ Alongside conservatism, simplicity is one of the Board's
principles for supervisory stress testing. See 12 CFR 252, Appendix
B.
---------------------------------------------------------------------------
A lack of simplicity is not, however, the only concern with a
framework that eliminates flexibility. Unexpected shocks occur, like
oil embargoes, national house price collapses, and pandemics. Moreover,
the implications of these shocks are often not readily captured in
concurrent data, especially their future effects on the economy and
financial stability in the United States, and so on firms' future
financial condition. Maintaining a degree of flexibility would allow
the scenarios to adapt to evolving conditions while adhering to the
principles outlined above.
In specifying the guides in this framework, the Board seeks to
maintain flexibility by specifying ranges for the peak or trough value,
the timing of that value, or the speed of adjustment for many of the
variables. The amount of flexibility in the guides, as measured by the
size of ranges specified, is calibrated to be as narrow as possible
while adhering to the principles laid out above and is based on
research and analysis of the behavior of those variables during past
recessions, consistent with the recession approach, or periods of
stress in financial markets. In addition to suggesting typical ranges
within which scenarios will vary, the Board seeks to provide
explanations of how the guide flexibility would be applied in different
economic and financial conditions.
Generally speaking, the Board would design a more severe path for
the scenario variables when it judged the level of systemic risks to be
high, and a less severe path for the scenario variables when it judged
systemic risks to be low. In some cases, the level of systemic risk can
be tied to the level of specific indicators. For instance, when the
unemployment rate is very high, the level of risk aversion also tends
to be high, and that causes firms to reduce risk across their various
business lines. In other cases, the Board would consider overall
assessments by economists, supervisors, and financial market experts to
assess the level of systemic risks, which typically incorporate many of
the specific indicators mentioned in the discussions of individual
guides below, when it is difficult to do so using individual or small
sets of scenario variables.\117\
---------------------------------------------------------------------------
\117\ For examples of relevant statistical analyses, see, e.g.,
V. Acharya et al., Measuring Systemic Risk, 30 Rev. of Fin. Studies
2-47 (Jan. 2017); T. Adrian & M. Brunnermeier, CoVaR, 106 Am. Econ.
Rev. 1705-41 (Jul. 2016).
---------------------------------------------------------------------------
Therefore, the Board expects that it may choose to have similar
severities for variable values in an annual scenario for those
variables where the Board retains discretion within established ranges
of the proposed guides. This expectation reflects the Board's
consistent view that annual scenarios are not forecasts of potential
future outcomes in the baseline or in a hypothetical stress
environment. Establishing variable values with similar severity levels
enhances the transparency and predictability of the annual scenarios,
and reflects an expectation that these variables are likely to
experience stress concurrently in a hypothetical stress scenario. As
discussed below, if the Board were to determine that a specific salient
risk should be addressed in a particular annual stress test, it would
provide a specific assessment of that risk and the rationale for an
alternative calibration of the variable's severity in the scenario
disclosure for comment.
While flexibility allows scenarios to adapt to fast-evolving
conditions, the guides in this framework are based on long-lasting
structural features of the economy. Macroeconomic history, however,
features many examples where new data have contradicted long-held
beliefs about underlying structural relationships. Also, the financial
system is constantly evolving, presenting new risks and
vulnerabilities. The relatively narrow ranges in the guides may not
always allow for a fulsome response by the scenarios to significant
developments. Therefore, the Board also sets out expectations for
circumstances that could require additional flexibility in setting the
specifications of the variables in the stress tests, so that the public
can anticipate where the Board could adopt a specification that differs
from those identified in the guides in this proposal. For instance, if
events occur that alter the historical severity of a given variable,
the Board could incorporate that data in its evaluation of the
appropriate path for a given variable in annual scenarios that occur
following such an event. The Board continuously monitors the
macroeconomy and the financial system. If ongoing developments warrant,
the Board may revisit this framework and adjust guides.
Finally, the increased predictability and transparency of the
scenario as specified in this framework may allow firms to adjust their
portfolios to reduce capital requirements, perhaps without a
commensurate reduction in risk. While the Board acknowledges this
possibility, the Board expects that the principle of severity embraced
in this framework will produce scenarios that adequately test such
risks. Flexibility is maintained to allow scenarios to adapt to
evolving conditions, not to reduce predictability and transparency.
Indeed, the ranges of flexibility specified, especially when considered
alongside the guidance offered regarding the conditions under which
that flexibility might be employed, result in highly transparent and
predictable scenario paths. Overall, the Board finds that the degree of
flexibility and the goals of transparency and predictability are well
balanced by this proposal, given the other requirements for designing
effective and credible stress tests.
Summary of Scenario Design Principles
In formulating the guides presented in this framework, the Board is
proposing to continue to use a recession approach,
[[Page 51883]]
where the severely adverse scenario reflects conditions that
characterize post-war U.S. recessions. To implement this approach, the
Board adopts a specific scenario narrative in which a severe shock to
financial markets propagates through the economy and results in a
severe, prolonged recession most similar to that of the 2007-2009
financial crisis. The Board provides a qualitative description of the
scenario informing the hypothetical recession that the scenario
reflects. In choosing specific scenario paths, the Board recognizes a
need for the scenario to be adequately severe and credible, and to
avoid adding to procyclicality.\118\ Finally, in this pursuit, the
guides maintain a degree of flexibility to adapt to evolving economic
and financial conditions. The Board continues to expect that there will
be some important instances when it will be appropriate to augment the
recession approach with salient risks and to set variables' values
inside of, and in some cases, outside of the ranges and values provided
in the guides in the Scenario Design Policy Statement.
---------------------------------------------------------------------------
\118\ Assumptions that are meant to avoid adding procyclicality
may add a degree of uncertainty to the path of the stress test
scenario, relative to an assumption that is neutral to current
economic conditions. However, the proposed variable guides and the
model used to design the macroeconomic scenario would promote the
predictability of the scenario and would help reduce year-to-year
volatility of the stress test and the resulting capital
requirements. This flexibility is particularly useful for the Board
when the economy enters a recession and the credit quality of the
banks' borrowers deteriorates, because a less-flexible scenario
design framework could result in a significantly larger increase in
capital requirements and hence a further drag on economic activity
relative to the previous year than would the proposed framework.
---------------------------------------------------------------------------
Question 41: What are the advantages and disadvantages of selecting
the scenario design principles described in this section? Are there
other principles that the Board should weigh along with these
principles? Should the Board develop guidance for how it would weigh
these principles in selecting values in annual scenario narratives?
Question 42: What considerations should the Board evaluate when
determining whether to set a given scenario variable independently of
other variables in the annual scenario, or at similar levels of
severity across multiple variables?
Common Components of Scenario Path Guides
The guides in this framework set out paths for each variable over
the 13 quarters in the severely adverse scenario. The stress test
requires projections of 13 quarters worth of losses to determine
capital ratios at the end of 9 quarters of the scenario, because loss
provisions in quarter 9 are affected by firm performance in quarters 10
to 13. To describe these paths, most guides adopt a simple framework
involving the following four parameters: the jump-off; the peak or
trough; the timing of the peak or trough; and the trajectory from jump-
off to peak or trough. The purpose of publishing these components is to
increase the transparency and public accountability of the stress test
scenario by communicating how the variable would behave throughout the
scenario period. In calibrating these parameters, the guides explain
their rationale in applying the recession approach along with the
scenario narrative and the three principles for scenario design
described above. These parameters are described as follows:
Jump-off: Jump-off values are important for informing the overall
state of the economy in the scenarios, often affecting the specific
levels achieved by the other parameters of the variable guide and
informing the exercise of flexibility as specified in the guides. In
the scenario, the jump-off value is the value of the variable in the
quarter preceding the scenario. For most variables, the jump-off value
is easily determined from published data at the time the scenario is
released to the public. However, for some variables the jump-off value
is not available prior to the date that the Board must finalize the
annual scenarios for publication, so an estimate is used; these details
are described in the individual guides. A separate issue involves
choosing the appropriate historical jump-off date in the Board's
analysis underlying the calibration of the guides. In many cases,
stresses developed over time and a specific jump-off date or quarter
for a particular period of stress may not be clearly identifiable. For
instance, the 2007-2009 financial crisis had multiple newsworthy
events--the suspension of redemptions from money market mutual funds by
BNP Paribas in August 2007, the failure of Bear Stearns in February
2008, and the bankruptcy of Lehman Brothers in September 2008.
Therefore, the Board uses a range of quarters around the beginning of
an identified recession or period of market stress to determine the
jump-off values. The Board determined that using the most extreme value
of the variable in the four quarters before, and the first quarter of,
the National Bureau of Economic Research (NBER) recession date or
documented financial stress event as the starting point for the
analysis supporting the calibration of the severity of the guides was
most consistent with the Board's stress testing principle of
conservatism. Each guide provides further details on selection of
relevant reference periods.
Peak or trough: The paths in the guide specify that each variable
in the scenario will either increase or decrease from its jump-off
value. If it increases, it will reach a maximum or peak value during
the scenario. If it decreases, it will reach a minimum or trough value
during the scenario. For example, during the scenario, unemployment
initially increases to a peak value, while house prices decrease to a
trough value. Each guide provides details on how the Board expects to
determine the level of this peak or trough and the rationale for this
determination. In general, more extreme values are more stressful, and
the specific levels of the peak or trough often depend on the jump-off
values in line with the principles of severity, credibility, and not
adding to procyclicality.
Trajectories from jump-off to peak or trough: This parameter
describes the values between the jump-off and peak or trough with a
straight line (linear) function, a nonlinear function, or by specifying
the proportion of the change from jump-off to peak or trough that will
obtain in each of the intervening quarters. Two further notes on
trajectories: first, trajectories are frequently described as either
frontloaded, meaning that larger changes occur earlier in the
trajectory, or backloaded, meaning that larger changes occur later in
the trajectory. Depending on the variable, frontloading and backloading
affect the overall severity of the scenario by having stressful changes
earlier or lasting longer. The individual guides discuss this issue.
Second, while several of the guides specify precise mathematical
formulas for trajectories, for example linear (straight line)
trajectories, rounding conventions--such as rounding to the first
decimal place--for the published scenario may result in small
differences from the result specified by the underlying formula. These
rounding conventions result in small changes to scenario variables that
tend not to affect overall severity. Instead, such rounding conventions
are meant to help simplify the communication of the scenario to the
public.
The Board also considered the appropriate trajectory of variables
after they reach the peak or trough and the appropriate end value. This
analysis confirmed that the range of end values used in past stress
tests are generally supported by historical analysis combined with the
stress testing
[[Page 51884]]
principle of conservatism. The end value describes the value of the
variable in the last (13th) quarter of the scenario. In applying the
recession approach to calibrating end-values, the Board considers the
values of a variable within a 10-15 quarter window after the beginning
of the recession or other identified financial stress event, instead of
simply taking the value of the variable in the 13th quarter. This range
of values allows the Board to better assess outliers or other
interactions between the data and the annual scenario narrative than
other calibration methods. This flexibility also helps accommodate
choices that account for the highly variable lengths of historical
recessions. The Board expects that for most variables determined by
guides, the recovery trajectories between the peak or trough and end
value typically should follow a roughly linear path that proportionally
allocates the change across the relevant time remaining to the end of
the scenario. A roughly linear recovery reflects a preference for
simplicity and transparency. For variables determined by the Board's
macroeconomic model for stress testing, the end values and related
trajectory from the peak or trough generally will be determined by the
model.
Timing of peak or trough: The guides for each variable set out the
quarter of the scenario in which the variable path reaches its peak or
trough. Generally, these occur earlier for fast moving variables and
later for slow moving variables. Depending on the variable, either
earlier or later timing may be more stressful, and there may be some
flexibility in the timing of the peak or trough.
In developing this framework, the Board considered a number of
alternative specifications, both for specific variables and for the
overall approach. Some of these alternatives are described in greater
detail within the discussion of each proposed guide in Section IX.G of
this Supplementary Information.
As described in the Scenario Design Policy Statement, the Board
considered alternatives to the recession approach for the overall
design of the severely adverse scenario, including a probabilistic
approach. The probabilistic approach would construct a baseline
forecast from a large-scale macroeconomic model and identify a scenario
that would have a specific probabilistic likelihood, given the baseline
forecast. The Board believes that, at this time, the recession approach
is better suited for developing the severely adverse scenario than a
probabilistic approach because it guarantees a recession of some
specified severity. In contrast, the probabilistic approach requires
the choice of an extreme tail outcome--relative to baseline--to
characterize the severely adverse scenario (e.g., a five percent or a
one percent tail outcome). In practice, this choice is difficult as
adverse economic outcomes are typically thought of in terms of how
variables evolve in an absolute sense rather than how far from the
baseline they lie in the probability space. In this sense, a scenario
featuring a recession may be somewhat clearer and more straightforward
to communicate. Finally, the probabilistic approach relies on estimates
of uncertainty around the baseline scenario and such estimates are in
practice model-dependent.
The Board also considered two types of alternative specifications
for each of the guides. First, the Board considered a more-prescriptive
approach, in which the guides set a typical peak or trough value and a
specific quarter in which that value would obtain, usually either at
the most severe end of the range specified in the proposed guide or at
the mid-point of the range. A guide set at the most severe end of the
range would be consistent with the principle of conservatism and
provide a high degree of transparency and predictability. In contrast,
the lack of flexibility in such a guide would reduce the ability of the
Board to respond appropriately to risks that are apparent in relevant
indicators of economic and financial conditions and could potentially
add to procyclical forces during economic booms or stressful periods. A
guide benchmarked to the midpoint of the range might not be credible
during periods of high vulnerability, while still being too severe when
stresses were already present.
Second, the Board considered that guides could have larger ranges
for the potential peak or trough values or the timing of the peak or
trough than the proposed guides. Larger ranges would increase the
Board's ability to capture risks that are apparent in relevant
indicators of economic and financial conditions and to adjust to
procyclical forces but would be less predictable and transparent. In
general, the Board expects the lower end of the range chosen for the
proposed guides to represent the least amount of stress that would be
deemed credible, while the higher end of the ranges already reflects
the most severe plausible realizations of the variable. The proposed
ranges for the guides are benchmarked to historical experience while
still providing some ability to move beyond the upper or lower end of
the historical range if circumstances dictate. In consideration of
these factors and the principles discussed above in this section,
therefore, the Board expects that the disadvantages from the loss of
transparency and predictability from guides with larger ranges
generally would be larger than the advantages stemming from more
flexibility in the wider ranges of such guides.
In each case, the proposed and some specific examples of
alternative guides are both discussed. While the Board views the
alternative guides as reasonable, the proposed guides have significant
advantages over the considered alternatives. However, the purpose of
the alternative guide discussion is to invite comment on a reasonable
alternative considered by the Board and to transparently lay out the
Board's present decision making in not adopting it.
Question 43: What are the advantages and disadvantages of the
alternative guides? Should the Board consider adopting any of the
alternative guides? What, if any, other guides should the Board
consider in addition to the alternative guides considered?
G. Description of Variable Guides in the Severely Adverse Scenario
Unemployment Rate
The stress test scenarios set out trajectories for several
variables, including the unemployment rate of the civilian non-
institutional population aged 16 and over (unemployment rate).\119\ As
described in the previous sections, the Board intends to use a
recession approach to develop the severely adverse scenario. The most
common features of recessions are increases in the unemployment rate
and contractions in aggregate incomes and economic activity. For this
and the following reasons, the Board intends to use the unemployment
rate as the primary basis for specifying the severely adverse scenario.
First, the unemployment rate is likely the most representative single
summary indicator of adverse economic conditions. Second, in comparison
to GDP, labor market data have traditionally featured more prominently
than GDP in the set of indicators that the NBER reviews to inform its
recession dates.\120\ Third and finally, the growth rate of potential
output can cause the size of the decline
[[Page 51885]]
in GDP to vary between recessions. While changes in the unemployment
rate can also vary over time due to demographic factors, this seems to
have more limited implications over time relative to changes in
potential output growth. The unemployment rate used in the severely
adverse scenario will reflect an unemployment rate that has been
observed in severe post-war U.S. recessions, measuring severity by
changes in the unemployment rate and GDP.\121\
---------------------------------------------------------------------------
\119\ The Board uses the quarterly average of seasonally
adjusted monthly unemployment rates for the civilian, non-
institutional population aged 16 years and older series from the
Bureau of Labor Statistics (series LNS14000000).
\120\ More recently, a monthly measure of GDP has been added to
the list of indicators.
\121\ Even though all recessions feature increases in the
unemployment rate and contractions in incomes and economic activity,
the size of this change has varied over post-war U.S. recessions.
Table 5 documents the variability in the depth of post-war U.S.
recessions. There is no universal agreement on how to categorize
recession severity. For the purposes of this guide, the following
categorization is employed: Recessions where the decline in real GDP
and the increase in the unemployment rate are less than 1.5 percent
or 1.5 percentage points, respectively, are considered mild;
recessions where the decline in real GDP is 2.5 percent or more, or
the increase in the unemployment rate is 3 percentage points or
more, are considered severe; all other recessions are considered
moderate.
---------------------------------------------------------------------------
The Board uses a quarterly average of the monthly unemployment rate
data in the stress test scenarios. The Board uses a quarterly average
of unemployment for several reasons. Unemployment and, importantly,
related variables such as disposable income (discussed below) can
feature volatility at higher frequencies unrelated to underlying market
conditions (e.g., unexpected weather events or a baseline level of
statistical variation in the survey responses); quarterly averages
smooth out the volatility that is present at monthly frequencies.
Overall, using quarterly averages strikes a balance between being
sensitive enough to capture broader economic trends and stable enough
to avoid overreaction to short-term fluctuations. The Scenario Design
Policy Statement outlines certain information regarding the peak level
and timing of the peak level of the unemployment rate for the severely
adverse scenario.\122\ This proposed guide conforms with and expands on
that statement, providing greater predictability, transparency, and
specificity with regards to the trajectory to peak value. The remainder
of this section is outlined as follows. An overview of the unemployment
guide components is given in Table 4. This is followed by a reiteration
of the Scenario Design Policy Statement which describes the peak
component of the unemployment rate and its timing. After that, a
discussion of the trajectory to peak value is provided.
---------------------------------------------------------------------------
\122\ Peak level represents the maximum value achieved during
the scenario.
---------------------------------------------------------------------------
The purpose of publishing these components is to increase the
predictability, public accountability, and transparency of the stress
test scenario by communicating how the variable will behave throughout
the scenario period.
[GRAPHIC] [TIFF OMITTED] TP18NO25.039
[[Page 51886]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.040
a. Peak Value and Timing of Peak
The Board is proposing to retain the guide established in the
Scenario Design Policy Statement, with some additional explanations
provided here. The Board anticipates that the severely adverse scenario
will feature an unemployment rate that increases between 3 to 5
percentage points from its initial level over the course of 6 to 8
calendar quarters.\123\ The initial level will be set based on the
conditions at the time that the scenario is designed. However, if a 3
to 5 percentage point increase in the unemployment rate does not raise
the level of the unemployment rate to at least 10 percent--the average
level to which it has increased in severe recessions--the path of the
unemployment rate in most cases will be specified so as to raise the
unemployment rate to at least 10 percent.
---------------------------------------------------------------------------
\123\ Six to eight quarters is the average number of quarters
for which a severe recession lasts plus the average number of
subsequent quarters over which the unemployment rate continues to
rise. The variable length of the timeframe reflects the different
paths to the peak unemployment rate depending on the severity of the
scenario.
---------------------------------------------------------------------------
This methodology is intended to generate scenarios that feature
stressful outcomes but do not add to procyclicality in the financial
system and macroeconomy.\124\ When the economy is in the early stages
of a recovery, the unemployment rate in a baseline scenario generally
trends downward, resulting in a larger difference between the path of
the unemployment rate in the severely adverse scenario and the baseline
scenario, resulting in a severely adverse scenario that is relatively
more intense. Conversely, in a sustained strong expansion--when the
unemployment rate may be below the level consistent with full
employment--unemployment
[[Page 51887]]
in a baseline scenario generally trends upward, resulting in a smaller
difference between the path of the unemployment rate in the severely
adverse scenario and the baseline scenario, resulting in a severely
adverse scenario that is relatively less intense. Historically, a 3 to
5 percentage point increase in the unemployment rate is reflective of
stressful conditions. As illustrated in Table 5, over the last half-
century, the U.S. economy has experienced five severe post-war
recessions. In all of these recessions excluding COVID-19, the
unemployment rate increased 3 to 5 percentage points, and in the three
most recent of these recessions excluding COVID-19, the unemployment
rate reached a level between 8 percent and 11 percent.\125\
---------------------------------------------------------------------------
\124\ For a discussion on the benefits of adequate severity,
see, e.g., Judge 2022, supra note 103. For a discussion on the
benefits of avoiding adding sources of procyclicality to the
financial system, see, e.g., D. Kohn & N. Liang, Understanding the
Effects of the U.S. Stress Tests, Brookings Institute (Jul. 2019),
https://www.brookings.edu/articles/understanding-the-effects-of-the-u-s-stress-tests/.
\125\ The unemployment rate was 8 percent in 1975Q1, 11 percent
in 1982Q4, and 9 percent in 2009Q2.
---------------------------------------------------------------------------
Under this method, if the initial unemployment rate were low--as it
would be after a sustained long expansion--the unemployment rate in the
scenario would increase to a level as high as what has been seen in
past severe recessions. However, if the initial unemployment rate were
already high--as would be the case in the early stages of a recovery--
the unemployment rate would exhibit a change as large as what has been
seen in past severe recessions.
The Board expects that the typical increase in the unemployment
rate in the severely adverse scenario will be about 4 percentage
points. However, as discussed in Section IX.F of this Supplementary
Information, the Board expects to calibrate the increase in
unemployment based on its views of the status of cyclical systemic
risk. More specifically, the Board would be more likely to set the
unemployment rate at the higher end of the range if the Board expects
that cyclical systemic risks are high (as it would be after a sustained
long expansion), and alternatively would be more likely to set the
unemployment rate to the lower end of the range if cyclical systemic
risks are low (as it would be in the earlier stages of a recovery),
provided doing so remained consistent with the goal of ensuring that
firms were properly capitalized to withstand severe economic and
financial conditions. This may result in a scenario that is slightly
more intense than normal if the Board expects that cyclical systemic
risks were increasing in a period of robust expansion.\126\
---------------------------------------------------------------------------
\126\ Note, however, that the severity of the scenario would not
reach an implausible level: even at the upper end of the range of
unemployment-rate increases, the path of the unemployment rate would
still be consistent with severe post-war U.S. recessions. However,
historical values need not serve as a binding upper bound for the
scenario peaks as discussed in the introductory section of this
proposal.
---------------------------------------------------------------------------
Conversely, it would also allow the Board to specify a scenario
that is slightly less intense than normal in an environment where
systemic risks appeared subdued, such as in the early stages of a
recovery. This choice would consider that the scenario does not add
unduly to remaining stress, thereby exacerbating the initial adverse
shock, and it would be particularly appropriate if the Board judges
that firms are already taking steps to reduce their risk--for instance,
by potentially restricting lending to otherwise qualified borrowers.
The Board expects that, in general, it would adopt a change in the
unemployment rate of less than 4 percentage points when systemic risks
are low or receding. This might be the case when, along with other
factors, the unemployment rate at the start of the scenarios is
elevated but the labor market is judged to be strengthening and higher-
than-usual credit losses stemming from previously elevated unemployment
rates were already realized--or are in the process of being realized--
and thus removed from firms' balance sheets.\127\ However, even at the
lower end of the range of unemployment-rate increases, the scenario
would still be expected to feature an increase in the unemployment rate
similar to what has been seen in about half of the severe recessions of
the past 50 years.
---------------------------------------------------------------------------
\127\ Evidence of a strengthening labor market could include
declines in weekly initial claims for unemployment, a declining
unemployment rate, steadily expanding nonfarm payroll employment, or
improving labor force participation. Evidence that credit losses are
being realized could include elevated charge-offs on loans and
leases, loan-loss provisions in excess of gross charge-offs, or
losses being realized in securities portfolios that include
securities that are subject to credit risk.
---------------------------------------------------------------------------
As indicated previously, if a 3 to 5 percentage point increase in
the unemployment rate does not raise the level of the unemployment rate
to 10 percent--the average level to which it has increased in severe
recessions--the path of the unemployment rate will be specified so as
to raise the unemployment rate to 10 percent. Setting a floor for the
unemployment rate at 10 percent recognizes the fact that not only do
cyclical systemic risks build up at financial intermediaries during
robust expansions, but also that these risks are easily obscured by a
buoyant environment.\128\
---------------------------------------------------------------------------
\128\ See supra note 114.
---------------------------------------------------------------------------
In setting the increase in the unemployment rate, the Board will
consider the extent to which analysis by economists, supervisors, and
financial market experts finds cyclical systemic risks to be elevated
(but difficult to be captured more precisely in one of the scenario's
other variables).\129\ In addition, the Board--in light of potential
impending shocks to the economy and financial system--expects to also
take into consideration the extent to which a scenario of some
increased severity might be necessary for the results of the stress
test and the associated supervisory actions to sustain public
confidence in financial institutions. Some indicators that would inform
the Board's decision would be the growth rate of real GDP and its
trajectory in recent quarters as well as leading economic indicators,
such as equity prices as these measures provide a broader perspective
on the state and direction of the economy. Consistent with the Scenario
Design Policy Statement, the Board is mindful of sources of
procyclicality in the financial system and in designing the severely
adverse scenario. While the Board designs the stress test scenarios to
promote the proper capitalization of firms, the scenarios are not
intended to serve as an explicit countercyclical offset to the
financial system.\130\
---------------------------------------------------------------------------
\129\ For relevant analyses, see supra note 117.
\130\ See 12 CFR 252, Appendix A.
---------------------------------------------------------------------------
Alternative Peak Guide Options
In preparing this proposal, the Board considered a guide that would
choose a peak level for unemployment that is 4 percentage points higher
than the jump-off value or 10 percent, whichever is higher. This
alternative has the advantage of being simpler, more predictable, and
more transparent than the guide choice. The Board views this
alternative guide to be less desirable as it is less flexible and may
end up being inadequately severe. Furthermore, such lack of flexibility
could potentially add to scenario procyclicality. For example, in
periods with already highly elevated unemployment rates above 7
percent, this alternative could result in unemployment rates of
historically high levels at times when economic conditions were already
depressed.
Instead, the current guide, specifying the greater of an increase
of 3 to 5 percentage points or 10 percent, acknowledges that the Board
would be unlikely to consider larger changes in unemployment when its
rate is already highly elevated. As discussed in Section IX.F of this
Supplementary Information, when the underlying conditions are favorable
and firm losses are low, firms may project these tendencies forward,
[[Page 51888]]
paving the way to take more risk.\131\ Similarly, as discussed
previously, the ability to adjust elements that potentially add
procyclicality can be a major benefit of stress tests.\132\
---------------------------------------------------------------------------
\131\ See Berrospide (2024) and Davydiuk (2024), supra note 113;
Cort[eacute]s (2021), supra note 112.
\132\ See Berger (2004) and Greenspan (2000), supra note 114.
---------------------------------------------------------------------------
b. Trajectory to Peak
The Board anticipates that the severely adverse scenario would
feature a trajectory to the peak unemployment rate that initially
increases quickly with slower incremental increases. The trajectory to
peak will have a concave parabolic path starting from the value in the
economy at the beginning of the scenario and reaching a peak at between
6-8 quarters.\133\ This approach for the trajectory to peak reflects
several considerations. First, this trajectory to peak features larger
increases in unemployment in the early quarters of the scenario,
reflecting a rapid and deep deterioration in labor market conditions,
in line with the scenario narrative discussed above and consistent with
the principle that the severely adverse scenario be highly stressful as
a rapid increase gives firms less time to adapt to changes. Second,
this trajectory to peak is consistent with theoretical economic models
which often share the feature that the response of unemployment to a
shock features initially large increases in unemployment with
decreasing incremental changes up to the peak.\134\ Empirically, this
general pattern can be seen, for example, in the impulse response
function illustrated in the first panel of Figure 2 in the FEDS Note
that evaluates empirical regularities in variable co-movement in stress
test scenarios.\135\ Third, while all recessions have differences in
their specific paths of the unemployment rate, a concave trajectory to
peak is broadly consistent with the data from severe recessions. One
indicator is to look at second differences, which are the change in
changes, an approximation of the acceleration of a variable.\136\
Concave paths have negative second differences. The second differences
of the unemployment rate are negative on average for severe recessions,
indicating a generally concave path with decreasing changes up to the
peak.
---------------------------------------------------------------------------
\133\ A concave curve is one with the property that any straight
line drawn between two points on the curve lies on or below the
curve. A parabolic path is a curve, x(t), that can be written as:
x(t) = a(t[caret]2) + b(t) + c for some constants a, b, and c. In
this case, concavity implies a < 0. If x0 is the jump-off value,
xpeak is the peak value, and tpeak is the peak quarter, then the
parameters for the path are given by the following equations: a =
(x0-xpeak)/tpeak\2\, b = 2*(xpeak-x0)/tpeak, and c = x0. Published
scenario values may differ somewhat from this formula because of
rounding conventions.
\134\ See, e.g., Panel A of Figure 12 in N.
Petrosky[hyphen]Nadeau & L. Zhang, Solving the Diamond-Mortensen-
Pissarides model accurately, 8 Quantitative Economics 611-50 (Jul.
2017).
\135\ See E. Afanasyeva et al., Evaluating Empirical
Regularities in Variable Comovement in Stress Test Scenarios, FEDS
Notes (Sep. 19, 2025), https://doi.org/10.17016/2380-7172.3885.
\136\ Given a time series x(t), the first difference is defined
as y(t) = x(t) - x(t-1) and measures changes from one period to the
next. The second difference is then defined as z(t) = y(t) - y(t-1)
= (x(t) - x(t-1)) - (x(t-1) - x(t-2)) and measures the change in the
rate of change, otherwise described as acceleration.
---------------------------------------------------------------------------
Finally, a trajectory with frontloading of increases in the
unemployment rate has been a characteristic of all recent severely
adverse scenarios, except for the second round of bank stress tests in
September 2020.\137\
---------------------------------------------------------------------------
\137\ This additional round of stress tests was performed due to
the continued uncertainty from the COVID-19 event. As the scenarios
were designed for the unique COVID-19 event, the Board does not
anticipate future stress testing to closely follow this unique
episode.
---------------------------------------------------------------------------
House Prices
The stress test scenarios set out trajectories for several
variables, including house prices as measured by the Price Index for
Owner-Occupied Real Estate (HPI).\138\ The Scenario Design Policy
Statement outlined information regarding the formulation of house
prices in the severely adverse scenario. This guide conforms with and
expands on that statement, providing further information on the data
used in the construction of the house price path in the severely
adverse scenario, including the timing of the trough value and the
trajectory to the trough value.\139\
---------------------------------------------------------------------------
\138\ Specifically, the Price Index for Owner-Occupied Real
Estate, Z.1 (Financial Accounts of the United States), Federal
Reserve Board series FL075035243.Q, divided by 1000.
\139\ Trough value represents the minimum value achieved during
the scenario.
---------------------------------------------------------------------------
Firms subject to the supervisory stress test have a substantial
exposure to the residential real estate market.\140\ Given firms'
direct exposures, and the broader impact of the housing sector on
household balance sheets and the macroeconomy, the Board's methodology
for supervisory stress tests incorporates house prices into a number of
models.\141\ Moreover, house price build-ups sometimes precede episodes
of banking stress, with a notable example being the 2007-2009 financial
crisis. By incorporating house prices into macroeconomic scenarios,
supervisory stress tests help ensure that firms subject to the stress
test are prepared for a range of market outcomes, including periods of
large declines in house prices directly affecting loan performance and
firms' balance sheets. This helps maintain the overall stability and
resilience of the financial system.
---------------------------------------------------------------------------
\140\ Regarding the importance of house prices to insured
depository institutions generally, in 2025Q1, mortgages and
mortgage-backed securities comprised more than 20 percent of FDIC
insured firms' assets (based on the ratio of Loans Secured by Real
Estate, 1-4 Family Residential Mortgages, plus Mortgage-backed
Securities, divided by Total Assets. Table II-A: Aggregate Condition
and Income Data, All FDIC-Insured Institutions, FDIC Quarterly 2025,
Volume 19(2), p.7, https://www.fdic.gov/quarterly-banking-profile/fdic-quarterly-2025-volume-19-number-2.pdf).
\141\ See Board, 2025 Supervisory Stress Test Methodology (Jun.
2025), https://www.federalreserve.gov/publications/files/2025-june-supervisory-stress-test-methodology.pdf.
---------------------------------------------------------------------------
The Board uses a quarterly average frequency for this data in the
supervisory stress test scenario. Instead of using the monthly
frequency at which the underlying data is available, the Board uses a
quarterly average of house prices in the stress test scenario for
several reasons. House prices and, importantly, related variables such
as disposable income (discussed more below) can feature volatility at
higher frequencies unrelated to underlying market conditions. For
example, extreme weather can affect the demand for home purchases and
employment during a particular month, and thus the prices paid in home
transactions and income that month, notwithstanding market conditions.
Therefore, quarterly averages smooth out month-to-month volatility.
Overall, using quarterly averages strikes a balance between being
sensitive enough to capture market trends and stable enough to avoid
overreaction to short-term fluctuations in prices.
In determining the appropriate level of scenario severity, the
Board adheres to the scenario design principles discussed in Section
IX.F of this Supplementary Information. While doing so, the Board also
strives to avoid introducing additional sources of procyclicality into
the financial system. In the context of house prices, these principles
are applied in calibrating the key aspects of the guide: the trough
value, the timing of the trough value, and the trajectory to trough
value. This approach helps ensure that the house price guide aligns
with the established stress testing literature while mitigating
potential systemic risks for the financial system. This guide
description is outlined as follows. An overview of the house prices
guide is given in Table 6. This is followed by a reiteration of the
Scenario Design Policy Statement which describes the trough value used
in the
[[Page 51889]]
construction of house prices. After that, this guide provides a
supplementary discussion of the construction of house prices in the
severely adverse scenario, followed by a discussion of the other
components of the trajectory of house prices.
---------------------------------------------------------------------------
\142\ Regarding New England, see J. Jordan, Problem Loans at New
England banks, 1989 to 1992: Evidence of Aggressive Loan Policies,
New England Econ. Rev. 23-38 (Jan. 1998); J. Jordan, Resolving a
Banking Crisis: What Worked in New England, New England Econ. Rev.
49-62 (Sep. 1998). Regarding California, see G. Zimmerman, Factors
Influencing Community Bank Performance in California, Federal
Reserve Bank of San Francisco Econ. Rev., 26-40 (1996), https://www.frbsf.org/wp-content/uploads/26-42.pdf. For a popular media
account, see D. Wood, ``California Real Estate Crunch Puts Pressure
on Bank Profits,'' Christian Science Monitor (Oct. 11, 1991).
Regarding Texas, while a number of factors, including nonperformance
of commercial and industrial loans, contributed to the Texas banking
crisis of the 80s, excesses in residential real estate were a strong
contributing factor. See J. Duca, M. Weiss, & E. Organ, ``Texas Real
Estate: From the 1980s' Oil Bust to the Shale Oil Boom,'' Ten-Gallon
Economy: Sizing Up Economic Growth in Texas 109-18 (2014); J.
O'Keefe, The Texas Banking Rrisis: Causes and Consequences 1980-
1989, 3 FDIC Banking Rev. 1 (Jul. 1990), https://fraser.stlouisfed.org/files/docs/publications/texasbankcrisis_1980_1989.pdf.
[GRAPHIC] [TIFF OMITTED] TP18NO25.041
a. Trough Value Component of the Guide
The Board is proposing to retain the guide established in the
Scenario Design Policy Statement to inform the trough of house prices
in the scenario, with additional explanations provided here. In most
circumstances, the Board expects that the ratio of HPI to nominal per
capita DPI (HPI-DPI ratio) falls by at least 25 percent or enough to
bring the ratio down to the trough reached in the wake of the 2007-2009
financial crisis, which occurred in the first quarter of 2012,
whichever is greater.
Data- and Scenario-Based Rationale for the Trough Value
Declining house prices, which are an important source of stress to
a firm's balance sheet, are not a steadfast feature of recessions, and
the historical relationship of national house prices with the
unemployment rate is not strong. Simply adopting their typical path in
a severe recession would likely underestimate risks stemming from the
housing sector. This can be seen when considering regional housing
recessions, which have occurred with greater frequency. Three examples
include New England and California in the early 1990s, and Texas in the
1980s. While regional house price indices featured only moderate
decreases, the ratios of price to income fell precipitously. Further,
in each case, the regional housing recession precipitated a regional
banking crisis.\142\
Assessing the procyclicality of house price paths over time is
complicated by the fact that house prices--in contrast to the
unemployment rate--have historically trended upward over time.
Therefore, instead of specifying the path of house prices directly, the
Board expects to consider the ratio of the nominal HPI to nominal per
capita DPI. The HPI-DPI ratio does not exhibit an upward trend and, as
such, provides an alternative way to assess the procyclicality of the
scenarios' house price paths. Moreover, the HPI-DPI ratio is a commonly
used valuation metric for the housing sector.\143\
---------------------------------------------------------------------------
\143\ While different authors have considered different measures
of house prices or income, there is wide agreement in the literature
that price to income ratios are an important gauge of the state of
the housing market. On the long-run stability of housing expenditure
shares, see M. Davis & F. Ortalo-Magn[eacute], Household
Expenditures, Wages, Rents, 14 Rev. of Econ. Dynamics 248-261
(2011). For an analysis of the importance of price-to-income ratios
for mortgage delinquencies, see K. Gazi & C. Vojtech, Bank Failures,
Capital Buffers, and Exposure to the Housing Market Bubble, 52 Real
Estate Econ. 1470-1505 (2024). For a macroeconomic model and
discussion, see C. Leung & E. Tang, The Dynamics of the House Price-
to-Income Ratio: Theory and Evidence, 41 Contemporary Econ. Policy
61-78 (2023). Other references considering price-to-income ratios in
financial stability include E. Pavlidis et al., Episodes of
Exuberance in Housing Markets: in Search of the Smoking Gun, 53 The
J. of Real Estate Fin. and Econ. 419-49 (2016); and K. Case & R.
Shiller, Is there a Bubble in the Housing Market?, Brookings Papers
on Economic Activity, No. 2003.2, 299-362 (2003).
---------------------------------------------------------------------------
Under most circumstances, the Board expects the decline in the HPI-
DPI ratio in the severely adverse scenario to be 25 percent from its
starting value or enough to bring the ratio down to its trough during
the 2007-2009 financial crisis, whichever is the larger decline. The
maximum trough level specified in this guide is motivated by the data,
corresponding to the level achieved in the wake of the 2007-2009
financial crisis, which reached a trough in the first quarter of 2012.
The minimum decline specified in this guide for the HPI-DPI ratio from
its starting value, a 25 percent decline,is motivated by the data as
well--such a fall reflects the average peak to trough fall in this
ratio across the three national housing recessions identified by the
Board, as shown in Table 7.\144\ While the average across housing
recessions is heavily influenced by the steep decline in the 2007-2009
financial crisis, similar magnitude falls have occurred with greater
frequency when considering the
---------------------------------------------------------------------------
\144\ The national house-price retrenchments that occurred over
the periods 1980-1985, 1989-1996, 2006-2011 are referred to in this
document as housing recessions. The date ranges of housing
recessions are based on the timing of house-price retrenchments.
These dates were also associated with sustained declines in real
residential investment, and the precise timings of housing
recessions would likely be slightly different were they to be
classified based on real residential investment in addition to house
prices. The ratios described in Table 7 are calculated based on
nominal HPI and HPI-DPI ratios indexed to 100 in 2000:Q1.
---------------------------------------------------------------------------
[[Page 51890]]
aforementioned regional housing recessions.\145\
---------------------------------------------------------------------------
\145\ See infra note 148.
---------------------------------------------------------------------------
The minimum decline of 25 percent ensures adequate scenario
severity, maintaining the credibility of the stress test while at the
same time constraining the trough from becoming unduly contractionary
and deviating too far from historically observed levels.\146\ Applying
a larger value of a minimum decline (e.g., the 2007-2009 peak-to-trough
fall of more than 40 percent) could result in a trough level that is
unjustifiably far away from most historical movements, especially if it
were applied during a period in which the HPI-DPI ratio were already at
a low level. Alternately, specifying a maximum trough level higher than
that experienced during the 2007-2009 financial crisis might not allow
the Board to adequately test firms' resilience to potential shocks when
home valuations are as elevated as they were in the mid-2000s.
---------------------------------------------------------------------------
\146\ If a future stress event causes the HPI-DPI to fall
significantly below the 2007-2009 financial crisis trough, or
perhaps just to that level, the Board will consider an update of the
trough calibration to reflect that new empirical evidence in
subsequent future tests.
---------------------------------------------------------------------------
The construction of this part of the house prices guide reflects
the goal of avoiding adding sources of procyclicality in the financial
system. Accordingly, the severely adverse scenario will feature smaller
variable movements when those variables are less extreme, and the
severely adverse scenario will feature larger variable movements when
those variables are more extreme, generally up to a level at least as
extreme as the 2007-2009 financial crisis.
The recession approach provides further justification for the
proposed calibration of the severity of the trough of house prices.
While national house prices and national unemployment do not exhibit a
strong relationship in the data, research shows that unemployment in a
household has a large effect on default rates, and that increases in
local unemployment are correlated with decreases in local house
prices.\147\ Similarly, regional housing recessions often feature
increases in regional unemployment.\148\ Hence, the recession approach
suggests that a scenario with a high peak level of unemployment should
also feature a low nadir in house prices.
---------------------------------------------------------------------------
\147\ On the relationship between unemployment and
delinquencies, see K. Gerardi et al., Can't Pay or Won't Pay?
Unemployment, Negative Equity, and Strategic Default, 31 The Rev. of
Fin. Studies, 1098-1131 (2018). On the Relationship Between Local
Unemployment and House Prices, see L. Gan, P. Wang, & Q. Zhang,
Market Thickness and the Impact of Unemployment on Housing Market
Outcomes, 98 Journal of Monetary Economics 27-49 (2018); and M.
Dvorkin & H. Shell, The Recent Evolution of U.S. Local Labor
Markets, Federal Reserve Bank of St. Louis Economic Synopses 1-3,
Issue 15 (2016).
\148\ For example, regarding the three regional housing
recessions mentioned above, the unemployment rate in New England
increased from 3.0 percent in January of 1988 to 8.2 percent in
1992, the unemployment rate in California increased from 5.2 percent
in January of 1990 to 9.8 percent in December of 1992, and the
unemployment rate in Texas increased from 5.8 percent in August of
1984 to 9.3 percent in October of 1986 according to the Bureau of
Labor Statistics.
[GRAPHIC] [TIFF OMITTED] TP18NO25.042
b. Additional Guide Parameters and Rationale
This subsection begins with a description of the construction of
the house price series. This is followed by a description of the timing
of the trough of HPI-DPI. The subsection concludes with information
regarding the trajectory to trough.
Construction of House Prices From HPI-DPI
Unlike the guides for some other variables, such as unemployment
and equity prices, this guide does not directly specify a path for
house prices in the severely adverse scenario. Instead, this guide
specifies a path for the HPI-DPI ratio. The scenario projection for
house prices is then calculated from this ratio using paths for DPI and
population, as calculated by the macroeconomic model for stress testing
that the Board has developed specifically to aid in communicating the
stress test scenario to the public specified on the Board's website.
The scenario projection for population is the same as that contemplated
in the Baseline Scenario Guide, as described in Section IX.C of this
Supplementary Information and in section 4.1 of the Scenario Design
Policy Statement. The scenario projection for house prices is then
calculated as the HPI-DPI path, discussed in this guide, multiplied by
nominal disposable income divided by population.
Trough Value Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The Board expects that the trough of HPI-DPI typically should occur
between quarter 8 and quarter 10 of the severely adverse scenario, as
explained below.
To support this range for the timing of the trough in house prices,
the Board applied the recession approach and used the timing of
unemployment peaks to calibrate the timing of the trough of HPI-DPI.
This benchmarking to the unemployment peak was necessary because house
prices have more protracted cyclical dynamics than other scenario
variables described in this framework. The three major house price
retrenchments indicated in Table 7 featured peak-to-trough durations
for HPI-DPI of between 19 and 30 quarters. The full implications of
such a protracted decline cannot be adequately assessed by including
only a portion of that decline within the nine-quarter horizon of the
annual stress tests, because the resilience of firms would be impacted
importantly by investors' perceptions of the expected future
[[Page 51891]]
losses.\149\ Moreover, the practical difficulties presented by the
difference between the length of historical housing cycles and the
length of the stress test scenario is an example of why the Board
expects to maintain the flexibility to use scenarios that are not
exactly like historical scenarios.\150\ Together, these two notions,
one practical and the other principled, require the Board to consider a
more careful approach to reading the historical record in its
determination of the timing of the trough value for HPI-DPI.
---------------------------------------------------------------------------
\149\ Supervisory stress tests consider results from the nine
quarters following the jump-off quarter. This and other guides
specify a 13-quarter path because the calculation of provisions for
losses are forward looking; that is, they depend on estimated losses
in the subsequent four quarters. Therefore, they require values for
some macroeconomic variables to extend beyond the nine quarters that
are counted in the stress test.
\150\ See Schuermann (2014), supra note 99.
---------------------------------------------------------------------------
Because the length of the severely adverse scenario cannot
replicate the duration of historical housing recessions, the Board
identified the subperiods within past housing recessions that featured
the greatest declines in HPI-DPI to support its calibration of the
trough within the scenario. This choice reflects the principle of
severity. The Board considered three window lengths when calculating
periods of maximum declines in HPI-DPI: 6, 9, and 13 quarters.\151\ The
calculations in Table 7 include the trough-quarter of such windows,
along with the percentage decline in HPI-DPI over each window.
---------------------------------------------------------------------------
\151\ These three window lengths were considered as they span
the set that would satisfy the limited duration of the scenario and
the need for severity discussed above.
---------------------------------------------------------------------------
Under the recession approach, the Board calibrates other variables
to be consistent with the scenario path for unemployment. To compare
the maximum decline in the HPI-DPI ratio with the peak in unemployment,
the table also includes the timing of the peak quarter for unemployment
along with the difference in timing between the peak unemployment rate
and the end of the window. For example, when considering the period
2005Q4-2012Q1 (Column 3, Table 8), the 6-quarter window with the
greatest change in HPI-DPI is 2007Q2-2008Q4 (Row 2, Column 3). This
window featured a fall in the HPI-DPI ratio of 24.1 percent. The end of
this window, 2008Q4 is 4 quarters before the unemployment rate peaked
in 2009Q4.
On average, the quarter of the maximum decline in HPI-DPI over 6-
quarter windows precedes the quarter of peak unemployment by 1.67
quarters. The unemployment guide features a range for the peak in
unemployment with a midpoint in quarter 7. Therefore, to be consistent
with some years' contemplated path for unemployment, a 6-quarter window
for the decline in HPI-DPI would have to start with the scenario jump-
off quarter rather than the first quarter of the scenario. Hence, the
Board deemed a trough timing for HPI-DPI of 6 quarters as too short.
More promisingly, the relationship between the peak of unemployment
and the trough of the HPI-DPI ratio flips at longer horizons. The
unemployment peak quarter precedes the quarter of the maximum declines
in HPI-DPI over 9 and 13 quarter windows by an average of 0.67 and
2.33, respectively. Therefore, trough timings of both 9 and 13 quarters
would be broadly consistent with the length of the scenario and the
timing of the unemployment peak within it. Of these two options, the
Board deems that the trough timing of HPI-DPI should occur around
quarter 9 for two auxiliary reasons: First, an interior trough time
allows for some subsequent recovery, mirroring the movement of
unemployment and other variables in this framework. Second, a shorter
duration to trough, all else equal, will result in a more severe
scenario, consistent with the principal of conservatism.
In addition, the maximum changes in HPI-DPI for the 6, 9, and 13
quarter subperiods associated with the 2007-2009 financial crisis are
close to or larger than 25 percent. Hence, this subperiod analysis also
further supports the calibration of the trough level in this guide.
Turning to a comparison with past scenarios, the selection of a
range of quarter 8 to 10 for the trough of HPI-DPI in the severely
adverse scenario is broadly consistent with the timing of past
scenarios. In 2019 to 2022, the severely adverse scenario featured a
trough in quarter 9. In 2023 to 2025, the severely adverse scenario
featured a trough in quarter 7, as the Board assessed valuation
pressures in residential real estate to be very elevated and wanted to
ensure that the banking system remained resilient to a sudden
correction in the housing market. Although that calibration of the
guide would require the Board to explain its rationale for choosing an
earlier trough going forward, the analysis presented above about the
typical timing of house price troughs suggests that a trough between
quarters 8 and 10 of the scenario usually would be sufficiently and
credibly stressful. In choosing the timing of the trough, the Board
expects to choose an earlier trough when the level of systemic risks is
high or rising and a later trough when the level of systemic risks is
low or declining. Housing market indicators such as recent trends in
HPI-to-DPI ratios, house price growth, the growth rate of mortgage
lending, or changes in mortgage lending standards are factors in that
determination. Conversely, when vulnerabilities or risks related to
residential real estate and related lending are low or decreasing, the
Board could consider a later trough.
[[Page 51892]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.043
Trajectory to Trough
This guide specifies a trajectory to trough featuring 20 percent of
the decline in the first quarter, 20 percent of the decline in the
second quarter, and a linear trajectory to trough thereafter, subject
to the rounding conventions mentioned in Section IX.F of this
Supplementary Information. As shown in Table 8, housing recessions tend
to be protracted. While the Board follows the recession approach, the
other principles from the stress testing literature suggest that a
careful reading of the data is warranted. To this end, when considering
the windows with the most rapid declines in Table 8 above, further
analysis shows that each housing recession featured quarters with
declines near 20 percent. In an application of the principle of
conservatism, the Board finds that two quarters of 20 percent declines
broadly fits the scenario narrative of a rapid decline in economic
conditions and sentiment, while meeting the other principles set out in
this guide; frontloaded declines are relatively more severe, so are
consistent with the principles of conservatism, severity, and the need
to consider possibilities somewhat outside the historical evidence. The
specification of linear declines thereafter was chosen in the interest
of simplicity.
---------------------------------------------------------------------------
\152\ Source: (1) Quarterly percent change in disposable
personal income (current dollars), expressed at an annualized rate,
Bureau of Economic Analysis; (2) Commercial Real Estate Price Index,
Z.1 Release (Financial Accounts of the United States), Federal
Reserve Board; (3) Federal Reserve staff calculations.
---------------------------------------------------------------------------
Moreover, a rapid decline in house prices is consistent with the
recession approach, in which other variables in the scenario are guided
by the scenario trajectory for the unemployment rate, which features
rapid initial deterioration. In addition, rather than having HPI-DPI
decline throughout the 13 quarter scenario as might be justified given
the historical record, the Board expects that house prices in the
severely adverse scenario will feature a moderate recovery after their
trough--again, consistent with the recession approach where variables
follow from the general movements of the unemployment rate, which
itself recovers after its trough--a feature which moderates the
severity of the initial decreases in house prices. Turning to past
scenarios, a moderately frontloaded trajectory to trough strikes a
balance between recent scenarios. Scenarios from 2023 to 2025 featured
strongly frontloaded declines, with more than 40 percent of the drop
happening in the first quarter, and increasingly smaller drops to the
trough. Frontloading the decline in this manner is consistent with the
principle of conservatism and the advice from stress testing literature
to consider features that are outside of historical experience when
vulnerabilities are elevated. The Board made a different decision with
house price scenarios in 2021 and 2022, which featured a less stressful
trajectory of initially small declines followed by a
[[Page 51893]]
period of larger declines while the economy was recovering from the
COVID-19 recession. Hence, a moderately frontloaded trajectory falls
between these earlier and later scenarios. The Board sees the reduction
in flexibility in this component of the house price path as partially
offset by the additional predictability and simplification that it
provides.
The Board expects that a scenario consistent with the level,
timing, and trajectory to the trough of house prices specified by this
guide will be at least somewhat more severe than the average of past
housing recessions and sufficiently close to the house price correction
associated with the 2007-2009 financial crisis.
Commercial Real Estate Prices
The stress test scenarios set out trajectories for several
variables, including commercial real estate prices as reported in the
Board's Z.1 statistical release.\153\ The Commercial Real Estate Price
Index aggregates price indices across office, retail, industrial and
other types of properties.
---------------------------------------------------------------------------
\153\ The source for the data is the Commercial Real Estate
Price Index, Z.1 Release (Financial Accounts of the United States),
Federal Reserve Board. This index is based on quarterly change of
the Value Weighted Costar U.S. Composite Index Excluding
Multifamily.
---------------------------------------------------------------------------
In the supervisory stress test, commercial real estate prices
capture a key part of the risks to firms from their commercial real
estate exposures, which are reported by firms on FR Y-14Q, Schedule
H.2. Most firms subject to the supervisory stress test have a
substantial exposure to the commercial real estate market. Moreover,
commercial real estate price build-ups often precede episodes of market
stress. By incorporating commercial real estate prices into
macroeconomic scenarios, supervisory stress tests help ensure that
firms subject to the stress test are prepared for a range of market
conditions, including periods of large decline in commercial real
estate prices directly affecting the firms' balance sheets. This helps
maintain the overall stability and resilience of the financial system.
In determining the appropriate level of scenario severity, the
Board adheres to the scenario design principles discussed in the
earlier Section IX.F of this Supplementary Information. While doing so,
the Board also strives to avoid introducing additional sources of
procyclicality into the financial system. In the context of commercial
real estate prices, these principles are applied in calibrating three
key aspects of the guide: the trough value, the timing of the trough
value, and the trajectory to trough value. This approach ensures that
the commercial real estate price guide aligns with the established
stress testing literature while mitigating potential systemic risks for
the financial system.
The rest of this section is organized as follows. First, Table 9
includes an overview of the Board's proposed guide for setting
commercial real estate prices in the severely adverse scenario. The
next subsection provides the data- and scenario-based rationale for the
calibration of the trough component. Afterward follows a discussion of
the alternative trough option, comparing the implementation and caveats
to the proposed guide description. Finally, additional guide parameters
for the trough timing and trajectory to trough value, and the rationale
for their calibration are discussed.
[GRAPHIC] [TIFF OMITTED] TP18NO25.044
a. Trough Value Component of the Guide
The proposed guide stipulates that at the trough, commercial real
estate prices will drop between 30 percent and 45 percent from the
jump-off value. The choice of the specific magnitude of drop within
this range will be determined based on the overall level of cyclical
systemic risk and an assessment of relevant indicators in the market as
reflected by a range of commercial real estate indicators such as the
level and change over preceding years in commercial real estate prices,
commercial real estate capitalization rate (cap rate), lending
standards on commercial real estate loans, rents, and vacancy rates,
among other indicators. The Board generally judges valuation pressures
and the implied level of risk by looking at where recent observations
of these relevant indicators are within their distributions.
Data- and Scenario-Based Rationale for the Trough Value
In line with the scenario design principles for setting the
scenario severity, as discussed earlier in Section IX.F of this
Supplementary Information, the proposed guide takes into account the
dynamics of a variety of commercial real estate market indicators,
including but not limited to the growth rates of commercial real estate
prices, changes in bank lending standards in the commercial real estate
segment, and the commercial real estate capitalization rate over the
past several years. The consideration of several years of history for
this variable is due to the slower-moving nature of commercial real
estate markets, in contrast with market volatility (measured by the
Chicago Board Options Exchange's CBOE Volatility Index (VIX)), stock
market prices, and corporate bond spreads, as described in those guides
below. The long-lived nature of these assets and substantial upfront
financial investment involved can loosen the connections between their
current observed valuations and financial conditions at firms and in
broader financial markets. For instance, lending practices adopted in a
period of declining prices, such as 2023 and 2024, can cloud immediate
price signals. Additionally, the complexity of these connections and
the
[[Page 51894]]
breadth of property types make it difficult to track developments in
the commercial real estate sector with a single quantitative indicator
or a very limited set of indicators that would constitute a basis for
the commercial real estate guidance. Therefore, the proposed guide
establishes a range of price decline values that determine the
magnitude of the price decline to the trough, as well as its
characteristics.
The proposed calibration of the range of decline (30 to 45 percent)
to the trough for the commercial real estate price index is determined
to account for commercial real estate price behavior in severe post-war
U.S. recessions and to allow for increases in severity after economic
expansions, in line with the principles outlined in the policy
statement as well as those discussed earlier in this section. First,
the range is centered around the value observed during the 2007-2009
financial crisis, when commercial real estate prices dropped about 39
percent from the peak in 2007Q3 to the trough in 2009Q4 (Table 10).
Second, the extent of commercial real estate price upswings provides a
guide for their subsequent unwinding and another target for the range.
As mentioned in the Board's policy statement, cyclical vulnerabilities
rise during more robust expansions. Looking back at the most recent
commercial real estate cycle upswing in 2013-2024, the median four-year
commercial real estate price growth rate in this period is about 30
percent, which the Board uses to calibrate the lower part of the range.
Setting a floor for the decline in commercial real estate prices of 30
percent recognizes the fact that, not only do cyclical systemic risks
build up at financial intermediaries during robust expansions, but also
a minimum level of risk exists even in an already stressed environment.
Separately, the Board opts for 45 percent as the higher end of the
range, as a similar value (43 percent, as measured by the four-year
growth rate of the commercial real estate price index between 2011Q3 to
2015Q3) was observed in the 2013-2024 commercial real estate cycle. The
upper end of this range is also set to be larger than the 39 percent
decrease experienced during the 2007-2009 financial crisis to allow for
scenarios that feature commercial real estate price declines that are
larger than what have been seen historically. Adequate severity
requires a guide to be able to go somewhat beyond historical
experiences when initial conditions warrant. Furthermore, certain
sectors within the commercial real estate market have already
experienced larger declines than 39 percent in the post-COVID-19
period, further justifying a range of potential declines that can
address risks that are apparent in relevant indicators of economic and
financial conditions as they arise.
[GRAPHIC] [TIFF OMITTED] TP18NO25.045
In its formulation of the annual scenarios, the Board could
consider the overall level of cyclical systemic risk or various
indicators related to commercial real estate markets to determine the
appropriate decline in commercial real estate prices in the scenario.
As discussed in Section IX.F of this Supplementary Information, the
Board expects to calibrate the decline in commercial real estate prices
based on its views of the status of cyclical systemic risk.
---------------------------------------------------------------------------
\154\ Source: Commercial Real Estate Price Index, Z.1 Release
(Financial Accounts of the United States), Federal Reserve Board
(series FL075035503.Q divided by 1000).
---------------------------------------------------------------------------
Specifically, the Board would be more likely to set the commercial
real estate price trough value at the higher end of the range if the
Board expects that cyclical systemic risks are high (as it would be
after a sustained long expansion), and alternatively would be more
likely to set the trough value to the lower end of the range if
cyclical systemic risks are low (as it would be in the earlier stages
of a recovery), provided doing so remained consistent with the goal of
ensuring that firms were properly capitalized to withstand severe
economic and financial conditions. This may result in a scenario that
is more intense than normal if the Board expects that cyclical systemic
risks were increasing in a period of sustained robust expansion.
Conversely, it would also allow the Board to specify a scenario
that is less intense than normal in an environment where systemic risks
appeared subdued, such as in the early stages of an expansion. This
choice would consider that the scenario does not add unduly to
remaining stress, thereby exacerbating the initial adverse shock, and
it would be particularly appropriate if the Board judges that firms are
already taking steps to reduce their risk--for instance, by potentially
restricting lending to otherwise qualified borrowers. Factors such as
whether underlying commercial real estate market conditions have
started to normalize and higher-than-usual credit losses stemming from
previous commercial real estate price declines were either already
realized--or are in the process of being realized--and thus removed
from firms' balance sheets would contribute to the assessment of
cyclical systemic risks.\155\
---------------------------------------------------------------------------
\155\ A commercial real estate market normalization could occur
when lending standards stop tightening, commercial real estate price
levels stabilize, and the capitalization rate moves toward the
middle of its historical range or higher. Evidence that credit
losses are being realized could include elevated charge-offs on
loans and leases or loan-loss provisions in excess of gross charge-
offs.
---------------------------------------------------------------------------
Figure 1 illustrates how the proposed guide (range between solid
lines) performs compared to past scenarios (shown as dots). As seen in
this figure, the proposed guide fully brackets the declines featured in
previous scenarios. Thus, the proposed guide is likely to result in
similar stress test severity as
[[Page 51895]]
before this revision to the policy statement.
[GRAPHIC] [TIFF OMITTED] TP18NO25.046
Alternative Trough Guide Option
The Board considered an alternative trough option in which
commercial real estate prices fall 35 percent from the jump-off value,
or reversal of prior 4 years of price increases up to 45 percent,
whichever results in a larger decline. The calibration of the
alternative guide relies on the similar observations shown in Table 10
and used for the calibration of the proposed guide. Specifically, the
alternative guide caps the decline in the commercial real estate prices
to a range between 35 and 45 percent. However, to determine the
specific decline in this range, in contrast to the proposed guide which
considers a variety of commercial real estate-market indicators and
allows for weighing them against each other, this alternative focuses
on only one dimension of potential risks in the commercial real estate
market--price pressures accumulated over the previous 4 years--and
formalizes the decline to the trough based on this indicator.
---------------------------------------------------------------------------
\156\ Z.1 Release (Financial Accounts of the United States),
Federal Reserve Board; Federal Reserve staff estimates.
---------------------------------------------------------------------------
The alternative guide stipulates that commercial real estate prices
will decline to the trough from the jump-off value by 35 percent or by
an amount needed to offset the four-year commercial real estate price
growth preceding the jump-off quarter. Hence, the alternative minimum
decline could be somewhat more severe compared to the proposed guide.
That said, the decline is capped at 45 percent to constrain the trough
calibration within historically plausible bounds. The choice of four
years (rather than, for example, the one-year look back used in the
equity price guide) to span the relevant accumulation period of price
pressures for this guide stems from a slower-moving nature of the
commercial real estate cycle, in contrast to faster moving variables
(like VIX or stock prices). At the same time, choosing a longer look-
back time period, such as five years, for example, would often produce
commercial real estate growth rates above 45 percent, thus triggering
the 45 percent maximum threshold of the guide too frequently and
resulting in excessive scenario severity relative to historically
observed events, particularly at the beginning of market
corrections.\157\
---------------------------------------------------------------------------
\157\ In the 2014-2024 period, for example, 5-year growth rates
above 45 percent occur in 2014Q4, 2015Q1, 2016Q3, 2016Q4, 2017Q1.
---------------------------------------------------------------------------
The commercial real estate price troughs set in past annual stress
test scenarios and the prescription of the alternative guide could be
noticeably different. In the data, previous commercial real estate
price changes in annual stress test scenarios, the key factor in
determining the prescription for the alternative guide, are not always
highly correlated with other commercial real estate indicators that the
Board would have used to gauge the extent of salient risks at the time.
For instance, from 2021-2023 (post-COVID-19 pandemic) the
alternative guide would prescribe troughs at 35 percent below jump-off
values, while the proposed guide would prescribe troughs similar to
those of past scenarios, around 40 percent below jump-off. The
alternative guide thus would not have accounted for the unusually small
number of commercial real estate sales that occurred during that period
and the upward biases in transaction-based commercial real estate
[[Page 51896]]
price indices created by the strategic behavior of owners, lenders, and
buyers in those conditions.\158\ Once commercial real estate prices had
declined considerably by 2024 and transaction volumes increased, the
shallower trough calibration for this alternative guide aligns with the
Board's choice for the severely adverse scenario.\159\ This example
illustrates that focusing on only one quantitative indicator (four-year
commercial real estate price growth) may be too narrow to determine an
adequate severity for the magnitude of decline from the start of the
stress test scenario to its trough (i.e., start-to-trough decline) for
this variable.
---------------------------------------------------------------------------
\158\ See, e.g., Board, Financial Stability Report (May 2023)
(discussing recent changes in commercial real estate prices
potentially understating the extent of weakness across the sector),
https://www.federalreserve.gov/publications/files/financial-stability-report-20230508.pdf; Remarks by Gov. Michelle Bowman,
Financial Stability in Uncertain Times (Oct. 11, 2023) (highlighting
the vulnerabilities from high vacancy rates in the office sector),
https://www.federalreserve.gov/newsevents/speech/bowman20231011a.htm.
\159\ The April 2025 Board Financial Stability Report discusses
the stability of commercial real estate prices and stronger position
of the commercial real estate market. Board, Financial Stability
Report (Apr. 2025), https://www.federalreserve.gov/publications/files/financial-stability-report-20250425.pdf.
---------------------------------------------------------------------------
Therefore, a guide that weighs a broader range of indicators and
how conditions differ by property type could provide a fuller, more
adequate framework for the Board to choose an appropriate level of
stress for commercial real estate exposures in future stress test
scenarios. Consequently, the Board would consider the overall level of
cyclical systemic risk, which is informed by a range of indicators
related to commercial real estate markets, in its formulation of the
annual scenarios as discussed in this section.
Although the proposed and the alternative guides are both
discussed, and the Board views the alternative guide as reasonable, it
may be insufficient to capture the complexity of the commercial real
estate market relative to the proposed guide. In addition, the
implementation of the alternative guide for commercial real estate
would be complicated by the lack of a real-time commercial real estate
price indicator. Typically, the data are available with a 4-month lag,
which means that the final quarter or two of data required to compute
the value of the guide would be based on a projection rather than
reported data. The purpose of the alternative guide discussion is to
invite comment on a reasonable alternative considered by the Board and
to transparently lay out the Board's present arguments for choosing the
proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Trough Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the trough and its timing signify the magnitude and timing
of the most severe point in this trajectory. The Board considers the
dynamics of commercial real estate prices using the official NBER
recession dates augmented by one year prior to the beginning of the
recession and one year after the end of the recession to compute
summary statistics for validating the timing of the trough for
commercial real estate prices in this guide. The Board considers such
additional data points because of the slow-moving nature of the
commercial real estate cycles, as referenced earlier in this section,
in comparison with the fast-moving and forward-looking behavior of
equity prices, corporate bond spreads, and VIX, for which the moves
following the Lehman Brothers bankruptcy during the 2007-2009 financial
crisis are most consistent with the scenario narrative adopted in this
proposed policy statement.
The guide stipulates that the trough level in the scenario would be
reached in quarters 8 to 10. This range is consistent with the slower-
moving nature of commercial real estate price cycles, the practice in
previous severely adverse scenarios, and the behavior in previous
periods of financial stress or recession. In the stress episode
surrounding the 2007-2009 financial crisis, the commercial real estate
price trough was in quarter 9 (Table 10). The usual process of slow
adjustment of commercial real estate prices and the ambiguity in the
measurement of those prices described earlier in this section motivates
the Board to reserve a range in which the adjustment occurs. Keeping
the magnitude of the trough constant, a more delayed trough timing
generally results in less severity of the overall path, as a less
abrupt worsening in conditions and credit quality gives firms more time
to adjust to the shock. Thus, a range in the timing (quarter 8 to 10)
is an additional lever (together with the trough magnitude range) to
avoiding the addition of sources of procyclicality in the stress test.
The Board would likely consider a delayed timing of the trough when the
cyclical vulnerabilities are lower, and an earlier trough timing when
the Board deems it appropriate to increase scenario severity, as
described in this section in relation to the choice of price decline.
Trajectory to Trough Value
To reach the trough value, the guide prescribes a smooth roughly-
linear transition from the jump-off point to the trough. This
prescription is consistent with the linear models often used in the
statistical modeling of macroeconomic series.\160\ Commercial real
estate prices are slower-moving, even in crisis times, so there is less
evidence of the frontloading seen in faster-moving variables such as
the VIX or BBB spreads. Moreover, the breadth of property types and
lags in real-time data availability contribute to the difficulty of
tracking the developments in this sector. As discussed above,
transactions-based prices may have biases based on the strategic
behavior of the parties involved. Given these circumstances,
considering more complicated trajectories may inject unnecessary
volatility into the exercise, counter to the principles laid out on
effective stress testing in Quarles (2019).\161\
---------------------------------------------------------------------------
\160\ See, e.g., M. Marcellino, J. Stock, & M. Watson, A
Comparison of Direct and Iterated Multistep AR Methods for
Forecasting Macroeconomic Time Series, 135 J. of Econometrics 449-
526 (2006) (discussing the popular linear time series models used
for forecasting macroeconomic time series).
\161\ See ``Stress Testing: A Decade of Continuity and Change,''
Remarks by Vice Chair for Supervision Randal K. Quarles at the
``Stress Testing: A Discussion and Review'' conference (Jul. 9,
2019), https://www.federalreserve.gov/newsevents/speech/quarles20190709a.htm.
---------------------------------------------------------------------------
The trajectories prescribed in previous scenarios are consistent
with the proposed guidance that commercial real estate price declines
are not frontloaded. The two exceptions are for the scenarios during
2017 and 2018, where the largest declines occur in the second quarter
of the scenario. In these years' scenarios, to test the resilience of
the banking system to strong economic conditions and commercial real
estate price increases in prior years, the Board chose scenarios which
called for deeper and earlier declines in commercial real estate prices
than considered in prior years' stress test scenarios. Notwithstanding
these exceptions, the smoother decline specified by the proposed guide
is more in line with historical behavior of the series and has the
benefit of reducing volatility.
Equity Prices
The stress test scenarios set out trajectories for several
variables, including equity prices proxied by the U.S. Dow Jones Total
Stock Market
[[Page 51897]]
Index (DWCF).\162\ This index includes about 3,700 stocks trading on
U.S. exchanges that account for 95 percent of the total market
capitalization.
Along with commercial real estate prices, housing prices, and the
VIX, equity prices are an essential gauge for asset prices that affect
the U.S. economy and the financial conditions of financial and
nonfinancial firms. Equity prices are generally recognized as a leading
indicator of future economic conditions broadly, including economic
growth and inflation.\163\ Therefore, testing the ability of a firm to
withstand a steep decline in equity prices helps ensure that these
firms are properly capitalized to withstand severe economic and
financial conditions.
In the supervisory stress test scenarios, equity prices are
converted to quarterly frequency using the quarter-end value. The
Board's use of this aggregation method in the severely adverse
scenario, rather than average or maximum value in the quarter used for
other variables, is a deliberate choice that reflects how equity prices
might impact the balance sheets of financial institutions. Quarter-end
values provide a clear, specific point-in-time snapshot of market
conditions, which is crucial for assessing firms' balance sheets and
market risk exposures. For trading books and fair-value estimates for
assets that firms hold, quarter-end prices provide the most up-to-date
mark-to-market valuation, which is critical for stress testing. Equity
markets are typically more liquid than debt markets or markets for real
estate, which means the most recent prices are less likely to be
affected by technical factors instead of economic fundamentals and
expectations about future conditions than in bonds or property markets.
Using quarter-end values also makes it easier to compare stress
scenarios with historical data, which is often reported on a quarter-
end basis. Finally, many equity options expire at the end of quarters,
making quarter-end prices particularly relevant for assessing option-
related risks.
In determining the appropriate level of scenario severity, the
Board adheres to the scenario design principles discussed in the
earlier Section IX.F of this Supplementary Information. While doing so,
the Board also strives to avoid introducing additional sources of
procyclicality into the financial system. In the context of equity
prices, these principles are applied in calibrating three key aspects
of the guide: the trough value, the timing of the trough value, and the
trajectory to trough. This approach helps ensure that the equity price
guide aligns with the established stress testing literature while
mitigating potential systemic risks for the financial system.
The rest of this section is organized as follows. First, Table 11
summarizes all of the equity prices guide components. This is followed
by a detailed description of the guide's trough component, including
the data- and scenario-based rationale for the calibration of the
trough component and a discussion of the alternative trough option.
Finally, additional guide parameters and the rationale for their
calibration are discussed.
[GRAPHIC] [TIFF OMITTED] TP18NO25.047
a. Trough Value Component of the Guide
The proposed guide stipulates that the decline in equity prices
from the jump-off value (i.e., the value of the equity price index at
the end of the quarter immediately preceding the start of the scenario)
will vary around 50 percent with an additional amount that offsets one
half of the price growth over the prior year, up to 10 percent. These
declines imply that equity prices would fall to a trough level that is
between 40 and 60 percent below the jump-off value. More formally, this
calibration implies that at the trough of the scenario path, equity
prices fall by
---------------------------------------------------------------------------
\162\ Specifically, the Board uses the U.S. Dow Jones Total
Stock Market (Float Cap) Index (DWCF): End-of-quarter value via
Bloomberg Finance L.P.; this index encompasses a wider universe of
stocks than the S&P 500 Composite.
\163\ In the academic literature, stock prices are well-known to
be fast-moving or forward-looking variables that react to shocks
quickly. One prominent example is the study by B. Bernanke, J.
Boivin, & P. Eliasz, Measuring the Effects of Monetary Policy: a
Factor-Augmented Vector Autoregressive (FAVAR) Approach, 120 Q. J.
of Econ. 387-422 (2005) (classifying stock market prices as fast-
moving variables that respond to shocks on impact).
---------------------------------------------------------------------------
[[Page 51898]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.048
Data- and Scenario-Based Rationale for the Trough Value
In line with the scenario design principles for setting the
scenario severity, discussed earlier in Section IX.F of this
Supplementary Information, the rationale behind the choice of the
neutral value of 50 percent comes from the data, as several recessions
in the sample featured a decline of this magnitude. In particular, the
equity price declines in the 1973 recession and the 2001 recession were
46 percent, whereas the decline in the 2007-2009 financial crisis
measured 48 percent (Table 12). The equity price decline in the 2007-
2009 financial crisis is most analogous to the scenario narrative,
which starts with a substantial adverse shock to risk appetite and
uncertainty and leads to a period of market disfunction followed by
very high unemployment. Other financial stress episodes have seen
maximum equity price declines of less than 50 percent, but in those
instances the declines were not exacerbated by market dysfunction as
considered in the scenario narrative.
The adjustment portion of this guide responds to the possibility
that economic or financial conditions at the beginning of the annual
stress test cycle might warrant a decline in equity prices that is
smaller or larger than 50 percent. This flexibility reduces the
likelihood that the calibration of the trough would unduly amplify
rising or falling valuation pressures in equity prices over the past
year. When the stock market does well (or poorly) in the prior year,
the guide stipulates that equity prices fall by more (respectively,
less), with the exact amount determined by one half of the prior year's
price change. The use of half instead of, for example, full price
change results in troughs that are less likely to be unduly severe.
This calibration of the guide is based on historical equity market
valuations. However, when recent price moves are not consistent with
fundamentals or longer-term trends, the Board could deviate from the
proposed guide and use price growth over a longer horizon.
[GRAPHIC] [TIFF OMITTED] TP18NO25.049
The choice of 10 percentage points as the upper bound for the
absolute value of the year-to-year variation in this scenario variable,
or equivalently the choice of effective bounds (between 40 and 60
percent) on the trough decline, is rooted in the data and is similar to
changes that have been used in past severely adverse scenarios. The
upper end of the range would allow the Board to meaningfully increase
scenario severity when equity market valuations are likely to be high
or rising (as they were during the dot-com era) to ensure that firms
are resilient to outsized losses if valuations return to more normal
levels. The lower end of the range would allow the Board to reduce
scenario severity if equity valuation pressures recently declined, as
might be the case following a stock market correction or early in an
economic recovery.\166\ Setting a floor for the decline in equity
prices of 40 percent recognizes the fact that, not only do cyclical
systemic risks build up at financial intermediaries during robust
expansions, but a minimum level of risk exists even in an already
stressed environment.
---------------------------------------------------------------------------
\164\ The Board uses the DWCF for the scenarios because this
index encompasses a wider universe of stocks compared with the S&P
500 Composite. That said, the quantitative differences between the
two measures are rather small. For instance, the implied declines in
the dotcom episode would be 45.6 percent for both the Dow Jones time
series and the S&P 500 Composite time series. Also, the overall
correlation of the one-year growth rate computed for both time
series on their common sample (1988Q1-2024Q4) is 0.99. Therefore, to
cover a larger sample of historical episodes, the Board uses the S&P
500 Composite time series to compute statistics in columns (1) and
(2) and uses the DWCF to compute statistics in column (3).
\165\ DWCF: End-of-quarter value via Bloomberg Finance L.P. and
S&P 500 Composite via Bloomberg Finance L.P.
\166\ Assessing equity market valuations requires some judgment
as to the indicators that are used. Two commonly referenced
indicators are the equity price to expected earnings ratio and the
equity risk premium, which is the estimated expected return on
equities minus the 10-year Treasury yield. These measures rely on
projections of future earnings and other economic indicators that
require additional judgments. Therefore, the Board has chosen to
increase transparency and predictability by specifying this guide
based on directly observable equity price changes and will typically
use the guide rather than relying on judgmental assessments of other
indicators of underlying valuation pressures.
---------------------------------------------------------------------------
Figure 2 illustrates how the proposed guide performs relative to
the 2014-2025 stress test cycles, comparing the guide-implied decline
with those of past stress test scenarios and realized changes in equity
prices. Overall, the troughs implied by the proposed guide (solid line)
are similar to past scenario
[[Page 51899]]
troughs. However, deviations between the proposed guide and past
scenarios have been distributed across lower or higher severity,
implying that the proposed guide and the previous more-judgmental
process can provide similar average severity across multiple years of
stress tests. Indeed, the decline in equity prices in past stress test
scenarios during 2014-2025 averages 52 percent, whereas the proposed
guide's prescription of the declines for the same period averages 55
percent. The slightly higher average decline is consistent with the
principle of adequate severity discussed in Section IX.F of this
Supplementary Information.
[GRAPHIC] [TIFF OMITTED] TP18NO25.050
Alternative Trough Guide Option
---------------------------------------------------------------------------
\167\ Bloomberg Finance L.P. (ticker: ``DWCF'') and Federal
Reserve staff estimates.
---------------------------------------------------------------------------
The Board considered an alternative in which the trough would be a
50 percent decline from the jump-off value in equity prices
unconditional on the previous year's price change and jump-off
conditions. The 50 percent value is chosen based on the same reasons as
the midpoint of the proposed guide. Although this alternative option is
fully transparent and predictable, it has several weaknesses.
On average, the proposed guide would prescribe troughs that would
have been somewhat lower than the alternative if it had been
operational over the past 12 years: 55 percent for the proposed guide
on average vs 50 percent for the alternative. However, although a 50
percent decline matches the judgmental average, it means that the test
would be more severe each year than the decline observed during the
2007-2009 financial crisis.
Furthermore, as the alternative guide is not sensitive to the jump-
off conditions, the resulting troughs could be either excessive or
insufficient in severity, thus exacerbating procyclicality in ways the
proposed guide does not. This weakness would be particularly
detrimental to the credibility of the stress test during long bull
markets (as the United States has experienced during the stress testing
era) or periods of protracted decline in equity prices as the stress
test would be serially under- or over-stating the likely risks.
The Board also considered a wider range in the proposed guide. An
upper bound of 15 percentage points for the variable change relative to
the midpoint of 50 percent would imply a much wider range of 35 to 65
percent declines at the trough. A 65 percent decline has not been
observed in the post-war US data, whereas a 35 percent equity price
decline could be insufficiently severe to maintain credibility of the
test at times of heightened uncertainty. An upper bound of 5 percentage
points for the variable change from the midpoint would cover the
relevant historical benchmarks but would provide a narrow range:
between 45 and 55 percent decline at the trough. This choice would
substantially limit the Board's ability to match the severity of the
equity price decline with the recent performance in equity markets so
might inadvertently add to procyclical forces in financial markets. A
choice of 10 percentage points as the upper bound on the change
relative to the 50 percent midpoint strikes a balance between an overly
narrow and an overly wide adjustment window.
Although the proposed and the alternative guides are both
discussed, and the Board views the alternative guide as reasonable, the
alternative guide's inability to respond to recent changes in equity
valuations would be
[[Page 51900]]
a significant limitation compared with the proposed guide. The purpose
of the alternative guide discussion is to invite comment on a
reasonable alternative considered by the Board and to transparently lay
out the Board's present arguments for choosing the proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Trough Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the trough and its timing signify the magnitude and timing
of the most severe point in this trajectory. The guide stipulates that
the trough level in the scenario would be reached in quarter 3 or
quarter 4, which is consistent with historical observations (Table 12).
For instance, in the stress episode surrounding the 2007-2009 financial
crisis, the trough for equity markets occurred three quarters after the
bankruptcy of Lehman Brothers in 2008Q3.\168\ That timing also accords
with the scenario narrative, in which a sudden and significant increase
in uncertainty and rapid deterioration in risk appetite leads to a
spike in financial market volatility and a sharp decline in U.S.
financial assets during the first quarter of the scenario.
---------------------------------------------------------------------------
\168\ Note that in the case of fast-moving variables (such as
equity prices or the VIX), the Board times the onset of the stress
period during the 2007-2009 financial crisis based on the Lehman
Brothers bankruptcy rather than the NBER recession timing.
---------------------------------------------------------------------------
Trajectory to Trough Value
To reach the trough value, the guide prescribes that between 60 and
70 percent of the decline occurs in the first quarter of the scenario,
10 to 20 percent of the decline occurs in the second quarter, with the
remaining decline being realized about equally in the remaining
quarter(s) to trough. This trajectory is consistent with the scenario
narrative in which a severe recession is triggered by a large financial
shock in the first quarter followed by a return to normal market
functioning in subsequent quarters.
These expected trajectory benchmarks reduce the variation in
trajectories relative to previous scenarios: across past severely
adverse scenarios, the median first quarter decline in equity prices
was 68.3 percent of the total decline, so a range between 60 and 70
percent is in line with the midpoint of past scenario choices. Also,
across past severely adverse scenarios, the median second quarter
decline in equity prices was 18.4 percent of the total decline, which
is also within the range of 10 to 20 percent specified in this guide.
Such a frontloaded decline is also consistent with the status of equity
prices in the index of leading economic indicators and the empirical
evidence from periods of equity market weakness.\169\ Across episodes
of stock market distress, the average share of the decline realized in
the two quarters preceding the trough amounts to 63 percent, with one
episode measuring a much higher 88 percent in one quarter (in 1962) and
most measuring 50 percent or more for these two quarters (for example,
52 percent in the 2007-2009 financial crisis).\170\
---------------------------------------------------------------------------
\169\ In the academic literature, stock prices are well-known to
be fast-moving or forward-looking variables that react to shocks
fast. See infra note 163.
\170\ The episodes of stock market distress include the
recessions of 1969, 1973, 2001, the 2007-2009 financial crisis as
well as the stock market decline in 1962.
---------------------------------------------------------------------------
In specifying a range for the timing of the trough and the
proportion of declines in each quarter along the trajectory to the
trough the Board maintains the option to create more or less severe
scenarios if it wishes to avoid adding to existing procyclical factors
or for another reason. An earlier trough with higher frontloading of
the declines generally would be more severe. The Board could consider
an earlier trough timing or higher frontloading when economic and
financial market conditions are buoyant, such as when equity prices
have increased by more than the maximum 10 percent adjustment to the
trough level. A delayed trough timing and lower frontloading generally
would decrease the scenario severity. The Board could consider delayed
timing of the trough or smaller frontloading when equity prices at
jump-off are depressed but have been increasing, or are projected to
increase, and firms have de-risked and begun to recognize related
losses.
VIX
The stress test scenarios set out trajectories for several
variables, including the VIX, that is, the Chicago Board Options
Exchange's CBOE Volatility Index. The VIX is an index measuring implied
volatility based on a portfolio of options of the Standard and Poor's
500 (S&P 500).\171\ The VIX is calculated and distributed by the
Chicago Board Options Exchange.\172\
---------------------------------------------------------------------------
\171\ The S&P 500 is a stock market index tracking the stock
performance of 500 leading companies listed on stock exchanges in
the United States.
\172\ Chicago Board Options Exchange via Bloomberg Finance L.P.
(ticker: ``VIX Index'').
---------------------------------------------------------------------------
The VIX is often referred to as the ``fear index'' because it
measures the market's expectation of future volatility. Furthermore,
equity market volatility has been often used as an indicator of the
price of risk, along with the spreads, which can depress economic
activity when elevated.\173\
---------------------------------------------------------------------------
\173\ The role of equity market volatility as an indicator of
the price of risk (along with the spreads) is discussed in T.
Adrian, N. Boyarchenko, & D. Giannone, Vulnerable Growth, 109 Am.
Econ. Rev. 1263-89 (2019). Relatedly, the National Financial
Conditions Index (NFCI) of the Federal Reserve Bank of Chicago
constructs a risk subcomponent that is based on co-movement between
volatility measures and spreads. See S. Brave & A. Butters,
Diagnosing the Financial System: Financial Conditions and Financial
Stress, 8 International Journal of Central Banking 191-239 (2012)
---------------------------------------------------------------------------
In the supervisory stress test models that use the macroeconomic
scenario, the VIX can act as an indicator of stress for a wide range of
important assets and income streams even if those business lines are
not specifically linked to the VIX index. By incorporating the VIX into
scenarios, stress tests help ensure that firms are prepared for a wide
range of market conditions, including periods of extreme volatility and
uncertainty and any associated economic downturn.\174\ This helps
maintain the overall stability and resilience of the financial system.
---------------------------------------------------------------------------
\174\ See, e.g., N. Bloom, The Impact of Uncertainty Shocks, 77
Econometrica 623-85 (2009); S. Baker, N. Bloom, & S. Davis,
Measuring Economic Policy Uncertainty, 131 Q. J. of Econ. 1593-1636
(2016).
---------------------------------------------------------------------------
In the supervisory stress test scenarios, the VIX is converted to
quarterly frequency using the maximum close-of-day value in any quarter
and expressed in percent. The Board's use of this aggregation method in
the scenarios, rather than average or quarter-end values as used for
other variables, is a deliberate choice to have at least one scenario
variable that reflects the unique nature of market volatility and its
impact on financial institutions. This approach ensures firms are
tested against the most extreme, potentially destabilizing market
conditions, even if short-lived. Short-term and sharp increases in the
VIX can reflect markets' initial response to changes in risk appetite
or the economic outlook that then have longer-lasting adverse effects
on the broader economy, such as reduced employment.\175\ Moreover, the
use of the maximum close-of-day values captures the non-linear effects
of volatility spikes on financial instruments, risk models, and
liquidity,
[[Page 51901]]
while also testing firms' ability to handle rapid market movements,
margin calls, and behavioral factors during peak stress.
---------------------------------------------------------------------------
\175\ See, e.g., A. Chomicz-Grabowska & L. Orlowski, Financial
Market Risk and Macroeconomic Stability Variables: Dynamic
Interactions and Feedback Effects, 44 J. of Econ. & Fin. 655-69
(2020).
---------------------------------------------------------------------------
In determining the appropriate level of scenario severity, the
Board adheres to scenario design principles discussed in Section IX.F
of this Supplementary Information. While doing so, the Board also
strives to avoid introducing additional sources of procyclicality into
the financial system. In the context of the VIX, these principles are
applied in calibrating three key aspects of the guide: the peak value,
the timing of the peak value, and the trajectory to peak. This approach
ensures that the VIX guide aligns with the established stress testing
literature while mitigating potential systemic risks for the financial
system.
The rest of this section is organized as follows. First, Table 13
provides an overview of the VIX guide components, which is followed by
the guide description of the peak component. A data- and scenario-based
rationale for the calibration of the peak component follows in the next
subsection. Next is a discussion of an alternative peak option,
comparing the implementation and caveats to the proposed guide option.
Finally, additional guide parameters and the rationale for their
calibration are discussed.
---------------------------------------------------------------------------
\176\ Theoretically, there is no upper bound on the VIX; i.e.,
it is not constrained by 100 percent (or any other ceiling value).
However, a value surpassing 100 percent would require extraordinary
levels of daily market volatility and has never been observed in the
historical sample, spanning 1990Q1-2025Q1.
[GRAPHIC] [TIFF OMITTED] TP18NO25.051
a. Peak Value Component of the Proposed Guide
The VIX will increase to a level between 65 percent and 75 percent
or by at least 10 percentage points from the jump-off value, whichever
results in a higher level.\176\
Data- and Scenario-Based Rationale for the Peak Value
In line with the scenario design principles for setting the
scenario severity, discussed in Section IX.F of this Supplementary
Information, the VIX guide calibrates the minimum level to be between
65 percent and 75 percent. This choice is consistent with the
historical observations during periods of stress (Table 14). In
particular, the proposed range for the peak value of the guide is
calibrated based mainly on the range of VIX realizations across four
recent recessions or episodes of financial stress. The minimum value of
65 also reflects a judgment that the stress test always must be
consistent with the goal of promoting financial stability, which means
that markets and the public must continue to view the stress test as
sufficiently severe to maintain confidence, especially during periods
of high uncertainty and volatility.\177\ Thus, the lower end of the
range for the guide is chosen to be modestly above the average VIX peak
of 61 percent (first column). Moreover, setting a floor for the
increase in the VIX of 65 percent recognizes the fact that, not only do
cyclical systemic risks build up at financial intermediaries during
robust expansions, but a minimum level of risk exists even in an
already stressed environment. The higher end of the range is close to
the maximum value across those periods, 83 percent, which was observed
during the COVID-19 pandemic (third column).
---------------------------------------------------------------------------
\177\ See Judge (2022), supra note 103.
---------------------------------------------------------------------------
[[Page 51902]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.052
The minimum increment of 10 percentage points would only be
relevant if the jump-off occurred during a period of already-high
volatility (for example, in the 2007-2009 financial crisis, when the
peak was 81 percent, or in the COVID-19 pandemic, when it was 83
percent). In such an instance, the guide allows for the possibility
that conditions could worsen further, given the other aspects of the
severely adverse scenario, such as the increase in unemployment and
decline in house prices from the baseline. This assumption ensures that
the VIX scenario peak is adequately severe. Limiting the increase to 10
percentage points ensures, however, that the peak does not deviate too
far from historically observed levels and become unduly contractionary.
---------------------------------------------------------------------------
\178\ Source: Data for the VIX are from the Chicago Board
Options Exchange via Bloomberg Finance L.P. (ticker: ``VIX Index'')
and span the period 1990Q1-2025Q1.
---------------------------------------------------------------------------
Figure 3 plots historical VIX data, past scenario peaks, and this
guide (solid lines). On average across the past stress test scenarios
(2014-2025), the VIX has been approximately 30 percent at the jump-off
quarter, i.e., the data observation serving as a starting point for the
scenario. The implied increase from the initial condition to the peak
can be quite large--in such instances where the VIX is around 30
percent at the jump-off quarter, the increase to the peak value would
be between 35-45 percentage points. Such rapid increases in the VIX are
consistent with what occurred during the four stress episodes
considered in this calibration. On average across those episodes, which
start in 1990Q1 when data for the VIX became available, the VIX
increases by approximately 39 percentage points from the onset of a
stress event, which is one quarter before the start of the NBER
recession date, to its peak, a value within the range implied by the
guide (see Table 14, first column).
[[Page 51903]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.053
Figure 3 illustrates the comparison of the guide-implied peak range
of the VIX (delineated by the solid lines) against the peaks in past
stress test scenarios (shown as dots), accompanied by the respective
jump-off points from the data (dashed line). Two key results emerge.
First, past peaks chosen by the Board in previous severely adverse
scenarios are mostly within the bounds that would have been stipulated
by the VIX guide. In the two instances where the Board would have been
more constrained, one episode was higher than the upper bound and the
other lower than the lower bound. Thus, the guide is likely to result,
on average, in similar stress test severity as before this revision to
the policy statement while having the benefit of each individual year's
scenario being more transparent and predictable. Second, the
flexibility in the proposed guide to have a minimum increase of 10
percentage points provides adequate severity during stressful times,
even beyond the upper end of the range for typical jump-off quarters.
For instance, in 2020Q1, when the COVID-19 pandemic unfolded and the
VIX jumped, reaching the historical maximum of the VIX, the prescribed
peak would have been higher than 75 percent. Given the severity of the
underlying conditions in 2020Q1, the peak would be determined by the
jump-off point and the 10-percentage-point increment, resulting in a
peak of 93 percent.
---------------------------------------------------------------------------
\179\ Sources: Chicago Board Options Exchange via Bloomberg
Finance L.P. and Federal Reserve staff estimates.
---------------------------------------------------------------------------
In its formulation of the annual scenarios, the Board's
considerations would include the overall level of cyclical systemic
risk, the current level of the VIX as a contemporaneous indicator of
uncertainty and financial stress, and the performance of equity prices
within the past 12 months as a forward-looking indicator of economic
and financing conditions to determine the appropriate increase in the
VIX in the scenario. As discussed in Section IX.F of this Supplementary
Information, the Board expects to calibrate the increment in the VIX
based on its views of the status of cyclical systemic risk.
Specifically, the Board would be more likely to set the VIX peak value
at the higher end of the range if the Board expects that cyclical
systemic risks are high (as it would be after a sustained long
expansion), and alternatively would be more likely to set the peak
value to the lower end of the range if cyclical systemic risks are low
(as it would be in the earlier stages of a recovery), provided doing so
remained consistent with the goal of ensuring that firms were properly
capitalized to withstand severe economic and financial conditions. This
may result in a scenario that is more intense than normal if the Board
expects that cyclical systemic risks were to be increasing in a period
of sustained robust expansion. Conversely, it would also allow the
Board to specify a scenario that is less intense than normal in an
environment where systemic risks appeared subdued, such as in the early
stages of a recovery. This choice would consider that the scenario does
not add unduly to remaining stress, thereby exacerbating the initial
adverse shock. The lower end of the increase range could also be
appropriate when underlying market uncertainty and financial stress
start to recede and higher-than-usual credit losses stemming from
previously elevated vulnerabilities were either already realized--or
are in the process of being realized--and thus removed from firms'
balance sheets.\180\
---------------------------------------------------------------------------
\180\ Evidence of market uncertainty and financial stress
receding could include strong stock market performance or positive
economic news related to GDP, unemployment or nonfarm payroll.
Evidence that credit losses are being realized could include
elevated charge-offs on loans and leases or loan-loss provisions in
excess of gross charge-offs.
---------------------------------------------------------------------------
Alternative Peak Guide Option
The Board considered an alternative in which the VIX would increase
to 75 percent or by at least 10 percentage points from the jump-off
value, whichever results in a higher level. In this alternative peak
option, the VIX would be set at a level of 75 percent in typical future
scenarios. This prescriptive implementation would follow the principle
of conservatism by always moving the VIX close to its historical
maximum. It would also have the benefit of increasing the
predictability of the guide. However, when the VIX at the jump-off
value is elevated but has been declining or is
[[Page 51904]]
projected to decline and firm balance sheets are recovering, this
alternative would remove the Board's discretion to choose a lower peak
for the VIX. A lower but still constant value for the VIX guide in a
typical scenario, for instance with a lower bound of 65 percent
(corresponding to the average value across past scenarios) might not
provide sufficient resilience in normal times, as the actual peaks of
the VIX in the 2007-2009 financial crisis and during the COVID-19
pandemic exceeded 80 percent. Although a lower anchor could be coupled
with a higher minimum increment value--for example, 20 percentage
points--such a large increment in already stressful times removes the
Board's discretion to choose a less severe VIX peak. Although the
proposed and alternative guides are discussed, and the Board views a
more restrictive alternative guide as potentially reasonable, the Board
believes the alternative guide is inferior to the proposed guide, given
the variation in peak levels of the VIX the Board has found appropriate
in past stress test scenarios. Nonetheless, the purpose of the
alternative guide discussion is to invite comment on a reasonable
alternative considered by the Board and to transparently lay out the
Board's present arguments for choosing the proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Peak Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the peak and its timing signify the magnitude and timing
of the most severe point in this trajectory. The guide stipulates that
the peak level in the scenario would be reached in quarter 2, which is
consistent with past severely adverse scenarios and historical
observations. The peak was reached in quarter 2 in both the 2007-2009
financial crisis and in the COVID-19 pandemic (see Table 14).\181\
Averaging across all four financial stress episodes used to calibrate
the guide yields a peak in quarter 3. As historical maximum values of
the VIX were reached in the 2007-2009 financial crisis and during the
COVID-19 pandemic, and the scenario narrative specifies that the event
is triggered by a financial crisis similar to events in the fall of
2008, the Board considers the peak timing in quarter 2 more appropriate
for both the proposed and the alternative guide.
---------------------------------------------------------------------------
\181\ The Board determined that the timing of the start of the
stress period should sometimes differ from the start date of the
recession determined by the NBER. For potentially fast-moving
variables (such as the VIX), the Board times the onset of the stress
period during the 2007-2009 financial crisis based on the Lehman
Brothers bankruptcy on September 15, 2008. This event is widely
considered to be the most significant of the events that roiled
financial markets during the 2007-2009 financial crisis episode. As
stress test data operate at quarterly frequency, the Board's timing
of this event for determining the subsequent timing of the peak VIX
is in 2008Q3. The focus on the Lehman Brothers bankruptcy as the
triggering event is more consistent with the stress test scenario
narrative in which a financial shock sets the stress test scenario
dynamics in motion than the NBER recession date.
---------------------------------------------------------------------------
Trajectory to Peak Value
To reach the peak value, the guide prescribes that the highest
share, 60 to 80 percent, of the VIX increase occurs in the first
quarter of the scenario. Such frontloading of the increase is broadly
consistent with empirical evidence and with the behavior of the other
fast-moving variables (such as equity prices) in the scenario.
Additionally, the academic literature considers the VIX (and other
measures of uncertainty) a contemporaneous stress indicator that can
respond to shocks on impact and stresses the importance of
contemporaneous feedback between uncertainty and financial
conditions.\182\ For instance, 100 percent of the increase in the VIX
occurred in the first quarter of the 1990Q3-1991Q1 recession. During
the 2007-2009 financial crisis, nearly 40 percent of the increase in
the VIX occurred in the first quarter.\183\ In specifying a target for
the proportion of increase to be realized in the first quarter, the
Board would follow the same approach that it would use to assess
appropriate severity for the peak value. In particular, during economic
booms, the Board might formulate a scenario with greater frontloading
of the VIX increases, as the scenarios with greater frontloading would
contribute to higher severity. In the case of an economy that is
characterized by moderate or slowing economic growth, the Board would
likely stipulate the middle of the range of the VIX increases. Whereas
in economic downturns or at the beginning of a recovery, the Board
would expect to formulate a scenario with less frontloading of the VIX
increases.
---------------------------------------------------------------------------
\182\ The importance of contemporaneous feedback between
uncertainty and financial conditions is discussed, for example, in
S. Gilchrist, J. Sim, & E. Zakrajsek, Uncertainty, Financial
Frictions, and Investment Dynamics, NBER Working Paper (2014), and
D. Caldara et al., The Macroeconomic Impact of Financial and
Uncertainty Shocks, 88 European Econ. Rev. 1166 (2016) (``Caldara
(2016)'').
\183\ See Table 14.
---------------------------------------------------------------------------
5-Year Treasury Yield
The stress test scenario sets out trajectories for several
variables, including the 5-year Treasury yield, which is measured using
the quarterly average of the yield on 5-year U.S. Treasury notes.\184\
---------------------------------------------------------------------------
\184\ This series is constructed by Federal Reserve staff based
on the Svensson smoothed term structure model. L. Svensson,
Estimating Forward Interest Rates with the Extended Nelson-Siegel
Method, 3 Sveriges Riksbank Q. Rev. 13 -26 (1995).
---------------------------------------------------------------------------
Because banks generally engage in maturity transformation by
borrowing short-term (i.e., deposits) to fund longer-term assets,
fluctuations in interest rates can affect their financial health in
various ways.\185\ The 5-year Treasury yield is an important benchmark
rate for credit markets and is, thus, directly related to the
profitability of banks' investments in loans and securities as well as
their trading activities. For example, a decline in longer-term
Treasury yields that exceeds the decline in short-term yields (known as
a flattening of the yield curve) tends to compress firms' net interest
margins and can therefore reduce their profitability. At the same time,
the decline in such yields tends to increase the market value of firms'
investments in long-term fixed-rate bonds, some which is reflected in
various measures of capital at firms.\186\ Incorporating the 5-year
Treasury yield into the supervisory stress test helps to ensure that
firms are prepared for a wide range of market conditions, including
periods with a sudden decline in a credit market benchmark rate. This
helps maintain the overall stability and resilience of the financial
system.
---------------------------------------------------------------------------
\185\ See W. English, S. Van den Heuvel, & E. Zakrajsek,
Interest Rate Risk and Bank Equity Valuations, 98 Journal of
Monetary Economics 80-97 (2018).
\186\ The change in the fair value of securities held for sale
is reflected in common equity for all firms and in common equity
tier 1 for firms subject to Category I and Category II standards, as
well as firms that opt into that treatment. See 12 CFR part 252.
---------------------------------------------------------------------------
The Board uses a quarterly average of the 5-year Treasury yield in
the stress test scenarios. Quarterly averages smooth out excessive (and
potentially irrelevant) volatility that is present at daily or even
monthly frequencies. Using quarterly averages strikes a balance between
being sensitive enough to capture market trends and stable enough to
avoid overreaction to market noise. Relatedly, the 5-year yield
reflects long-term expectations of overall economic conditions.
Therefore, removing short-term volatility from this measure via
quarterly averaging is likely to, more-often-than-not, result in a
[[Page 51905]]
better representation of macroeconomic conditions.
In determining the appropriate level of scenario severity, the
Board adheres to scenario design principles discussed in Section IX.F
of this Supplementary Information. While doing so, the Board also
strives to avoid introducing additional sources of procyclicality into
the financial system. In the context of the 5-year yield, these
principles are applied in calibrating three key aspects of the guide:
the trough value, the timing of the trough value, and the trajectory to
trough. This approach ensures that the 5-year yield guide aligns with
the established stress testing literature while mitigating potential
systemic risks for the financial system.
The rest of this section is organized as follows. First, Table 15
presents an overview of the 5-year Treasury yield guide components,
followed by the guide description of the trough component. The next
subsection provides the data- and scenario-based rationale for the
calibration of the trough component. A discussion of an alternative
trough option follows in the next subsection, comparing the
implementation and caveats to the proposed guide option. Finally,
additional guide parameters (trough value timing and trajectory to the
trough) and the rationale for their calibration are discussed.
[GRAPHIC] [TIFF OMITTED] TP18NO25.054
a. Trough Value Component of the Proposed Guide
Under the proposed guide, the 5-year Treasury yield decreases from
its starting value by 1.5 to 3.5 percentage points. The Board expects
to determine the size of the scenario's decline based on relevant
banking, macroeconomic, or other conditions in the economy or financial
markets.\187\ Additionally, the size of the decline will likely be
informed by (a) the behavior of short-term interest rates in the
macroeconomic model for stress testing that the Board has developed
specifically to aid in communicating the stress test scenario to the
public,\188\ (b) estimates of the likely term premiums in a period of
economic weakness consistent with the scenario narrative, and (c) risks
that are apparent in relevant indicators of economic and financial
conditions.\189\ However, the guide restricts the 5-year Treasury yield
from falling below a lower bound of 0.3 percent or a decline of 0.3
percentage points from the jump-off level, whichever is lower.
---------------------------------------------------------------------------
\187\ Depending on the level of short-term interest rates, in
some scenarios, the short-term rate could reach its trough slower
than the 5-year and 10-year yields. In those cases, the scenario
would include the inversion of the yield curve in the first few
scenario quarters. Such behavior is in line with past scenarios as
well as behavior of interest rates preceding past stress episodes,
like the 2001Q1-2001Q4 recession, the 2007-2009 financial crisis and
the COVID-19 pandemic.
\188\ See https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
\189\ In the Board's macroeconomic model for the stress test,
the path of the 5-year Treasury yield is determined as the sum of
the expected federal funds rate implied by the scenario and the
paths of the term premiums.
---------------------------------------------------------------------------
Data- and Scenario-Based Rationale for the Proposed Trough Value
In the recession approach chosen by the Board, risk-free long-term
interest rates fall because reduced economic activity and inflation
result in an easing of monetary policy. As noted above, declining
interest rates can have both positive and negative implications for
firms' capital levels, depending on the firm's business model and the
specific composition of its assets and liabilities at the start of the
stress test.
In line with these guiding principles as well as those emphasized
by the stress testing literature discussed in Section IX.F of this
Supplementary Information, the Board considers the behavior of the 5-
year Treasury yield during four financial stress episodes since the
mid-1980s, including the 2007-2009 financial crisis, to calibrate the
guide (Table 16).\190\ The average decline in the 5-year Treasury yield
[[Page 51906]]
during those financial stress episodes has been around 2.7 percentage
points, ranging from 2.1 to 3.5 percentage points.\191\ Notably, the
percentage-point decline in the 5-year yield across these recessions is
consistent even though the level of the yield at the start of the
period has varied considerably.
---------------------------------------------------------------------------
\190\ In contrast with the calibration of other scenario
variable guides, the Board considers the behavior of the 5-year
Treasury yield during four financial stress episodes only after the
mid-1980s. These financial stress episodes include NBER recessions
in 1990Q3-1991Q1, 2001Q1-2001Q4, 2008Q3-2009Q2 (Lehman Brothers
bankruptcy as a forcing event), and 2019Q4-2020Q2. For the purposes
of calibrating representative yield behavior during stress episodes,
the Board chose to focus on the recessions since the mid-1980s, as
the period after the mid-1980s is characterized by a major monetary
policy regime shift and stabilization in the interest rate
environment. The mid-1980s marked the end of the ``Great
Inflation,'' an era that began in the mid-1960s and was
characterized by persistently high inflation and accommodative
monetary policy. In response, monetary policy underwent a major
regime shift in the early 1980s. This regime shift began the era of
``Great Moderation'' marked by low and stable inflation and reduced
macroeconomic volatility. See, e.g., R. Clarida, J. Gali, & M.
Gertler, Monetary Policy Rules and Macroeconomic Stability: Evidence
and Some Theory, 115 Q. J. of Econ. 147-80 (2000); Federal Reserve
History, Federal Reserve Bank of St. Louis, ``Great Inflation,''
https://www.federalreservehistory.org/essays/great-inflation;
Federal Reserve History, Federal Reserve Bank of St. Louis, ``Great
Moderation,'' https://www.federalreservehistory.org/essays/great-moderation.
\191\ The average decline during all the NBER recessions
starting from the 1969Q4-1970Q4 recession is 2.5 percentage points,
which is close to the average since the mid-1980s, but the range of
declines is wider.
[GRAPHIC] [TIFF OMITTED] TP18NO25.055
The evidence from the historical stress episodes along with the
principle of conservatism and the goal of avoiding the addition of
sources of procyclicality suggest that a decline of 1.5 to 3.5
percentage points in the 5-year Treasury yield would be reasonable. The
lower end of the range (i.e., 1.5 percentage points) is somewhat below
the historical average decline in the yield during financial stress
episodes and in previous severely adverse scenarios (Table 14), leaving
room to adjust the decline--and thus severity of the scenario--relative
to the historical average. The 5-year yield declined by 2.1 percentage
points during the 1990Q3-1991Q1 recession.
---------------------------------------------------------------------------
\192\ Quarterly average of the yield on 5-year U.S. Treasury
notes, constructed by Federal Reserve staff based on the Svensson
smoothed term structure model. See Svensson (1995), supra note 184.
---------------------------------------------------------------------------
The higher end of the range for the decline (i.e., 3.5 percentage
points) is driven by observations in the data as well as the guiding
principles: first, the largest decline in the 5-year yield during NBER
recessions since the mid-1980s has been 3.5 percentage points. This
decline took place during the 2001Q1-2001Q4 recession. However, outside
recessions, there are episodes displaying more sizable drops in the 5-
year yield over the horizon of 13 quarters (matching the scenario
horizon), the declines ranging from 2.6 to 6.1 percentage points. In
particular, the episode spanning 1984Q2-1987Q2 had a drop of 6.1
percentage points, the episode spanning 1990Q3-1993Q3 had a drop of 3.5
percentage points, and the episode spanning 1999Q4-2002Q4 had a drop of
3.1 percentage points. These observations suggest that a decline of 3.5
percentage points in the 5-year yield is coherent with past
experiences. Second, allowing the 5-year yield to potentially fall more
than what has been observed, on average, during past recessions speaks
to the guiding principle that adequate severity might sometimes require
a scenario that is somewhat beyond typical historical experiences.
The guide also imposes a 0.3 percent lower bound for the value of
the 5-year Treasury yield. The Board opted for this near zero, albeit
positive, lower bound for a few reasons. First, the lower bound is
intended to limit the extent that an annual scenario may unduly
disincentivize bank lending when the economy is entering or recovering
from a severe downturn. Second, this choice increases the
predictability of the 5-year Treasury yield path in the scenario.
Third, the lower bound is calibrated to be in line with the historical
episodes. The 5-year Treasury yield declined to similar levels during
the COVID-19 pandemic, reaching 0.3 percent in 2020Q3. Finally, the
guide imposes a decline of 0.3 percentage points in the yield when the
jump-off value of the 5-year yield is close to or below its historical
minimums at the scenario jump-off. In particular, this element binds
when the yield is below 0.6 percent at the jump-off. This element
further increases transparency on the yield trough level in the
scenarios in various potential interest rate environments outside
historical experiences.
To illustrate how the trough levels of scenarios consistent with
this guide would compare to the past stress test scenarios, consider
the history of the 5-year yield and its scenario values at the trough
over the period in which the Board has been conducting annual stress
tests, from 2014 to 2025 (Figure 4). The past stress test scenario
troughs are depicted as dots, whereas the range that is spanned by the
proposed guide is indicated by the solid lines, incorporating the lower
bound. The dashed line depicts the quarterly average of the 5-year
Treasury yield observed in the data. This period contains both low- and
high-interest rate environments: The quarterly average of the 5-year
Treasury yield over that period was 2.1 percent, ranging between 0.3
and 4.5 percent at the jump-off quarter.
[[Page 51907]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.056
During the periods in which the 5-year yield was below 1.7 percent,
such as most quarters between 2013 and 2016, and 2020 to 2021, the
guide would prescribe the lower bound for the 5-year yield. In a
higher-rate environment, however, a severe drop in the 5-year yield
would not necessarily imply a yield close to zero, which the guide
takes into an account. Between 2017Q1 and 2019Q2, the interest rate
environment was such that the decline in the yield within the range of
1.5 to 3.5 percentage points would have provided the Board with the
discretion to choose trough levels in the range of 0.3 and 1.4 percent.
After 2022Q1, the proposed guide would have constrained the Board at
times to a choice of yield levels significantly greater than the 0.3
percent lower bound.
---------------------------------------------------------------------------
\193\ Quarterly average of the yield on 5-year U.S. Treasury
notes, constructed by Federal Reserve staff based on the Svensson
smoothed term structure model. See Svensson (1995), supra note 184.
---------------------------------------------------------------------------
Figure 4 also illustrates that the troughs implied by the proposed
guide are quantitatively close to, but not the same as those featured
in the past stress test scenarios during the low-interest-rate
environment from 2013 to 2022 (the dots are located closer to the
binding lower bound in most years). In several of those years, the
Board chose a level for the trough that was modestly above the level
that would have been prescribed by the guide and in one case the Board
chose a level below the guide. However, with interest rates having
risen to moderate levels between 2023 and 2025, the guide would have
required the Board to choose a higher trough in 2024 and 2025 than it
did.
The Board considers these deviations from past scenarios to be an
acceptable consequence of adopting the guide, given its goal of
increasing predictability and transparency in the stress test. On the
one hand, a more flexible guide, which would encompass a higher share
of the past scenario troughs both in the lower and higher interest rate
environments, would call for a wider range in the variable component of
the guide. While a wider range would increase scenario flexibility, it
would come at the expense of predictability. The proposed range strikes
a balance between providing an adequate amount of flexibility to allow
for adjusting scenario severity based on economic and financial
conditions and keeping scenarios predictable. On the other hand,
keeping the range as is, one could also consider shifting the range up
or down to better enclose the past scenario troughs. If the range was
shifted down (e.g., 1.0 to 3.0 percent), the guide would better
encompass the past scenario troughs during the low-interest-rate
periods, but the opposite would be true for the post-COVID-19 periods.
If the range was shifted up (e.g., 2.0 to 4.0 percent), the guide would
better encompass the past scenario troughs between 2023 and 2025, but
the lower bound would bind for a larger number of scenario troughs
between 2013 and 2021. Thus, shifting the range would not meaningfully
change how well the guide aligns with the past scenario troughs.
Lastly, as the deviations from past scenarios would have been in both
directions, the Board expects that the proposed guide will be broadly
consistent with maintaining an average level of severity of stress
tests going forward that is similar to what it has been under the
Scenario Design Policy Statement.
In setting the 5-year Treasury yield trough value, the Board could
consider the overall level of cyclical systemic risk, and the current
level of the 5-year Treasury yield as a benchmark measure of overall
economic and financing conditions. As discussed earlier in Section IX.F
of this Supplementary Information, the Board expects to calibrate the
increment in the 5-year yield in consideration of observable cyclical
systemic risk. The Board would also consider how declines in Treasury
yields, which decrease net income but increase the market value of
firms' long-term securities holdings, interact with the current
vulnerabilities in the banking sector. In general, a decline in long-
term interest rates may have a positive or negative effect on the
severity of the scenario for a given firm
[[Page 51908]]
depending on the firm's exposure to interest rate risk, which may vary
from year to year depending on the firm's portfolio. In reaching its
determination to set this guide in an annual scenario, the Board will
consider how the choice would promote stress test credibility and the
resilience of the financial system to even worse outcomes. If the Board
observes that cyclical systemic risks were increasing in a period of
sustained robust expansion, the Board might choose a scenario that is
more severe than normal. The choice would also depend on firms'
exposure to interest rate risk. Conversely, the Board could specify a
scenario that is less intense than normal in an environment where
systemic risks appeared subdued, such as in the early stages of a
recovery, provided that doing so remained consistent with the goal of
ensuring that firms were properly capitalized to withstand severe
economic and financial conditions. A less severe scenario could also be
appropriate when underlying market uncertainty and financial stress
start to recede and higher-than-usual credit losses were either already
realized--or are in the process of being realized--and thus removed
from firms' balance sheets. The choice would consider that the scenario
does not add unduly to remaining stress, thereby exacerbating the
initial adverse shock, and it would be particularly appropriate if the
Board judges that firms are already taking steps to reduce their
risk.\194\
---------------------------------------------------------------------------
\194\ Evidence of market uncertainty and financial stress
receding could include strong stock market performance or positive
economic news related to GDP, inflation, unemployment or nonfarm
payroll. Evidence that credit losses are being realized could
include elevated charge-offs on loans and leases or loan-loss
provisions in excess of gross charge-offs.
---------------------------------------------------------------------------
Alternative Trough Guide Option
As an alternative, the Board also considered a guide in which the
5-year Treasury yield would decline by 2.5 percentage points regardless
of the jump-off conditions, with the lower bound still applying. Under
this alternative, the decline of 2.5 percentage points is chosen based
on the same observations shown in Table 16. In particular, 2.5
percentage points is close to the average decline in the 5-year
Treasury yield observed during the financial stress episodes (2.7
percentage points). The choice of a single value in the middle of the
range proposed in the proposed version of this guide reflects the
offsetting effects of interest rates on net interest margin and fair
value of securities.
The Board considered this alternative because of its goal of
increasing transparency and predictability of the stress test, while
maintaining sufficient severity. However, the Board recognizes that
this alternative guide would not avoid adding sources of procyclicality
as effectively as the proposed guide. In particular, it would reduce
the Board's flexibility during periods of moderate or high interest
rates to test the resilience of firms' net income to a sharper decline
in interest rates. However, as noted above, declines in yields have
offsetting effects on firms' regulatory capital in the stress test
because they decrease net income but increase the market value of their
long-term securities holdings. Thus, a more flexible guide could have
more-balanced effects on the stress capital buffer calculated from the
stress test results.
While the alternative troughs fall within the range determined by
the proposed guide, these trough levels can be significantly higher or
lower than the values chosen by the Board in prior severely adverse
scenarios. These deviations could impair the Board's ability to ensure
that the stress test severity fully considers the risks that are
apparent in relevant indicators of economic and financial conditions,
particularly those related to the Treasury term premium, when
determining the trough value. The Board views the alternative guide as
reasonable. As compared to the proposed guide, the alternative guide
would provide firms and the public with increased predictability
regarding the trough value to be set for 5-year Treasury yields.
However, this increased predictability under the alternative guide
comes at the expense of the added flexibility inherent in the proposed
guide to set the trough based on risks that are apparent in relevant
indicators of economic and financial conditions and to avoid adding
sources of procyclicality in the proposed guide. The purpose of the
alternative guide discussion is to invite comment on a reasonable
alternative considered by the Board and to transparently lay out the
Board's present arguments for choosing the proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Trough Value Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the trough and its timing signify the magnitude and timing
of the most severe point in this trajectory. The proposed guide
suggests that the 5-year Treasury yield would reach the trough between
quarters 1 and 4 of the scenario. This timing is chosen such that the
trough is consistent with the scenario narrative: the severely adverse
scenario is triggered by a sizeable financial shock combined with a
pronounced increase in unemployment and decrease in inflation. In
response to these developments, both short- and long-term interest
rates typically would fall sharply. The timing of the trough is also
broadly consistent with the historical data (Table 16). Averaging
across the four financial stress episodes, the trough is placed in the
fifth quarter, but the trough occurred earlier during the two most
recent recessions.\195\ The 5-year yield reached its trough in quarter
3 during the 2007-2009 financial crisis and in quarter 4 during the
COVID-19 pandemic. In the past stress test scenarios, the trough was
also reached in quarter 3, on average. In setting this part of the
guide in an annual scenario, the Board expects to consider the same
indicators and other factors described above for the choice of the
trough in the 5-year rate, so as best to promote stress test
credibility and the resilience of the financial system to even worse
outcomes.
---------------------------------------------------------------------------
\195\ These four episodes include 1990Q3-1991Q1, 2001Q1-2001Q4,
2008Q3-2009Q2 (Lehman Brothers bankruptcy as a forcing event), and
2019Q4-2020Q2 recessions.
---------------------------------------------------------------------------
Trajectory to Trough Value
The proposed guide stipulates that the largest share of the decline
in the 5-year Treasury yield would be realized in quarter 1. A rapid,
frontloaded decline of the 5-year yield to its trough would be
consistent with the scenario narrative and the implied dynamics of the
other variables, mainly the large increase in unemployment and
resulting declines in inflation and output. In response to these
developments, both short- and long-term interest rates would fall
sharply, consistent with the Board's macroeconomic model for stress
testing, and specifically the expectational component of the 5-year
Treasury yield, which accounts for the future expected realizations of
the macro variables that determine the policy rate rule.\196\
---------------------------------------------------------------------------
\196\ Existing studies suggest that it is beneficial to
frontload interest rate cuts in response to shocks. See, e.g., R.
Caballero & A. Simsek, A Note on Temporary Supply Shocks with
Aggregate Demand Inertia, 5 Am. Econ. Rev.: Insights 241-58 (2023);
R. Caballero & A. Simsek, Monetary Policy and Asset Price
Overshooting: a Rationale for the Wall/Main Street Disconnect, 79 J.
of Fin. 1719-53 (2024).
---------------------------------------------------------------------------
To determine the specific path of the 5-year Treasury yield for a
given trough timing, the Board considered a simple formula that can map
the trough value timing to a share of decline in quarter
[[Page 51909]]
1. To do so, the Board considered lower and upper bound of trough
timing described in the previous section. If the trough timing is
quarter 1 (e.g., lower bound of the range), then the formula should
yield 100 percent of the decline occurring in the first quarter. For
trough timing of quarter 4, the Board took example of COVID-19
pandemic. During the COVID-19 pandemic, the 5-year yield reached its
trough in quarter 4, and nearly 50 percent of the decline in the 5-year
yield was realized during the first quarter. Using these reference
points, the Board concluded that the following simple formula could
determine the approximate share of the decline realized in quarter 1
---------------------------------------------------------------------------
as:
100%-15% * (Trough value timing-1).
This simple formula stipulates that when the scenario trough is
realized in quarter 4, about 55 percent of the decline would be
realized in quarter 1:
100%-15% * (4-1) = 55%
This result is broadly in line--if not exactly in line--with the
data from the COVID-19 pandemic. The guide also stipulates that, after
the initial decline realized in quarter 1, the yield declines to its
trough at smoothly decreasing percent reductions.
10-Year Treasury Yield
The stress test scenarios set out trajectories for several
variables, including the 10-year Treasury yield, which is measured
using the quarterly average of the yield on 10-year U.S. Treasury
notes.\197\ Because banks generally engage in maturity transformation
by borrowing short-term (i.e., deposits) to fund longer-term assets,
fluctuations in interest rates can affect their financial health in
various ways. The 10-year Treasury yield is an important benchmark rate
for credit markets and is, thus, directly related to the profitability
of firms' investments in loans and securities as well as their trading
activities. For example, a decline in longer-term Treasury yields that
exceeds the decline in short-term yields (known as a flattening of the
yield curve) tends to compress firms' net interest margins and can
therefore reduce their profitability. At the same time, the decline in
such yields tends to increase the value of firms' investments in long-
term fixed-rate bonds, some of which is reflected in various measures
of capital at firms.\198\ Incorporating the 10-year Treasury yield into
the supervisory stress test helps to ensure that firms are prepared for
a wide range of market conditions, including periods with a sudden
decline in a credit market benchmark rate. This helps maintain the
overall stability and resilience of the financial system.
---------------------------------------------------------------------------
\197\ This series is constructed by the Board based on the
Svensson smoothed term structure model. See Svensson (1995), supra
note 184.
\198\ The change in the fair value of securities held for sale
is reflected in common equity for all firms and in common equity
tier 1 for firms subject to Category I and Category II standards, as
well as firms that opt into that treatment. See 12 CFR part 252.
---------------------------------------------------------------------------
The Board uses a quarterly average of the 10-year Treasury yield in
the stress test scenarios. Quarterly averages smooth out excessive (and
potentially irrelevant) volatility that is present at daily or even
monthly frequencies. Using quarterly averages strikes a balance between
being sensitive enough to capture market trends and stable enough to
avoid overreaction to market noise. Relatedly, the 10-year yield
reflects long-term expectations of overall economic conditions.
Therefore, removing short-term volatility from this measure via
quarterly averaging is likely to, more-often-than-not, result in a
better representation of current macroeconomic conditions.
In determining the appropriate level of scenario severity, the
Board adheres to scenario design principles discussed in the earlier
Section IX.F of this Supplementary Information. While doing so, the
Board also strives to avoid introducing additional sources of
procyclicality into the financial system. In the context of the 10-year
yield, these principles are applied in calibrating three key aspects of
the guide: the trough value, the timing of the trough value, and the
trajectory to trough. This approach helps ensure that the 10-year yield
guide aligns with the established stress testing literature while
mitigating potential systemic risks for the financial system.
The rest of this section is organized as follows. First, Table 17
presents an overview of the 10-year Treasury yield guide components,
followed by the guide description of the trough component. The next
subsection provides the data- and scenario-based rationale for the
calibration of the trough component. A discussion of an alternative
trough option follows in the next subsection, comparing the
implementation and caveats to the proposed guide option. Finally,
additional guide parameters (trough value timing and trajectory to the
peak) and the rationale for their calibration are discussed.
[GRAPHIC] [TIFF OMITTED] TP18NO25.057
[[Page 51910]]
a. Trough Value Component of the Proposed Guide
The 10-year Treasury yield decreases from its starting value by
between 1.0 to 3.0 percentage points. The Board will determine the size
of an annual scenario's decline based on a number of factors, including
relevant banking, macroeconomic, or other conditions in the economy or
financial markets.\199\ Additionally, the size of the decline will be
informed by (a) the behavior of short-term interest rates in the
macroeconomic model for stress testing that the Board has developed
specifically to aid in communicating the stress test scenario to the
public,\200\ (b) estimates of the likely term premiums in period of
economic weakness consistent with the scenario narrative, and (c) risks
that are apparent in relevant indicators of economic and financial
conditions.\201\ However, the guide restricts the 10-year Treasury
yield from falling below a lower bound of 0.5 percent or a decline of
0.3 percentage points from the jump-off level, whichever is lower.
---------------------------------------------------------------------------
\199\ Depending on the level of short-term interest rates, in
some scenarios, the short-term rate could reach its trough slower
than the 5-year and 10-year yields. In those cases, the scenario
would include the inversion of the yield curve in the first few
scenario quarters. Such behavior is in line with past scenarios as
well as behavior of interest rates in past stress episodes, like the
2001Q1-2001Q4 recession, the 2007-2009 financial crisis and the
COVID-19 pandemic.
\200\ See https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
\201\ In the macroeconomic model for stress testing, the path of
the 10-year Treasury yield is determined as the sum of the expected
federal funds rate implied by the scenario and the paths of the term
premiums.
---------------------------------------------------------------------------
Data- and Scenario-Based Rationale for the Proposed Trough Value
In the recession approach chosen by the Board, risk-free long-term
interest rates fall because reduced economic activity and inflation
result in an easing of monetary policy. As noted above, declining
interest rates can have both positive and negative implications for
firms' capital levels, depending on the firm's business model and the
specific composition of its assets and liabilities at the start of the
stress test.
In line with these guiding principles as well as those emphasized
by the stress testing literature discussed in Section IX.F of this
Supplementary Information, the Board considers the behavior of the 10-
year Treasury yield during four financial stress episodes since the
mid-1980s, including the 2007-2009 financial crisis, to calibrate the
guide (Table 18).\202\ The average decline in the 10-year Treasury
yield during those financial stress episodes has been around 1.9
percentage points, ranging from 1.3 to 2.4 percentage points.\203\
Notably, the percentage-point decline in the 10-year yield across these
recessions is similar even though the level of the yield at the start
of the period has varied considerably.
---------------------------------------------------------------------------
\202\ In contrast with the calibration of other scenario
variable guides, the Board considers the behavior of the 10-year
Treasury yield during four financial stress episodes only after the
mid-1980s. These financial stress episodes include NBER recessions
in 1990Q3-1991Q1, 2001Q1-2001Q4, 2008Q3-2009Q2 (Lehman Brothers
bankruptcy as a forcing event), and 2019Q4-2020Q2. For the purposes
of calibrating representative yield behavior during stress episodes,
the Board chose to focus on the recessions since the mid-1980s, as
the period after the mid-1980s is characterized by a major monetary
policy regime shift and stabilization in the interest rate
environment. The mid-1980s marked the end of the ``Great
Inflation,'' an era that began in the mid-1960s and was
characterized by persistently high inflation and accommodative
monetary policy. In response, monetary policy underwent a major
regime shift in the early 1980s. This regime shift began the era of
``Great Moderation'' marked by low and stable inflation and reduced
macroeconomic volatility. See supra note 190.
\203\ The average decline during all the NBER recessions
starting from the 1973Q4-1975Q1 recession--the first NBER recession
for which the 10-year Treasury yield data is available--is also 1.9
percentage points, but the range of declines is wider.
[GRAPHIC] [TIFF OMITTED] TP18NO25.058
The evidence from the historical stress episodes along with the
principle of conservatism and the goal of avoiding the addition of
sources of procyclicality suggest that a decline of 1.0 to 3.0
percentage points in the 10-year Treasury yield would be reasonable.
The lower end of the range (i.e., 1.0 percentage points) is somewhat
below the historical average decline in the yield during financial
stress episodes and in previous severely adverse scenarios (Table 18),
leaving room to adjust the decline--and thus severity of the scenario--
relative to the historical average. The 10-year yield declined by 1.3
percentage points during the 1990Q3-1991Q1 recession.
---------------------------------------------------------------------------
\204\ Source: Quarterly average of the yield on 10-year U.S.
Treasury notes, constructed by Federal Reserve staff based on the
Svensson smoothed term structure model. See Svensson (1995), supra
note 184.
---------------------------------------------------------------------------
The higher end of the range for the decline (i.e., 3.0 percentage
points) is driven by observations in the data as well as the guiding
principles: first, the largest decline in the 10-year yield during NBER
recessions since the mid-
[[Page 51911]]
1980s has been 2.4 percentage points. This decline took place during
both the 2001Q1-2001Q4 recession and the COVID-19 pandemic. However,
outside recessions, there are episodes displaying more sizeable drops
in the 10-year yield over the horizon of 13 quarters (matching the
scenario horizon), the declines ranging from 2.2 to 5.8 percentage
points. In particular, the episode spanning 1984Q2-1987Q2 had a drop of
5.8 percentage points, the episode spanning 1990Q3-1993Q3 had a drop of
3.0 percentage points, the episode spanning 1999Q4-2002Q4 had a drop of
2.2 percentage points, and the episode spanning 2018Q4-2021Q4 had a
drop of 2.4 percentage points. These observations suggest that a
decline of 3.0 percentage points in the 10-year yield is coherent with
past experiences. Second, allowing the 10-year yield to potentially
fall more than what has been observed during past recessions, on
average, speaks to the guiding principle that adequate severity should
be somewhat beyond historical experiences.
The guide also imposes a 0.5 percent lower bound for the value of
the 10-year Treasury yield. The Board opted for this near zero, albeit
positive, lower bound for a few reasons. First, the lower bound is
intended to limit the extent that an annual scenario may unduly
disincentivize bank lending when the economy is entering or recovering
from a severe downturn. Second, this choice increases the
predictability of the 10-year Treasury yield path in the scenario.
Third, the lower bound is in line with the historical episodes. The 10-
year Treasury yield declined to similar levels during the COVID-19
pandemic, reaching 0.6 percent in 2020Q3, but it has never fallen below
that level. Finally, the guide imposes a decline of 0.3 percentage
points in the yield when the jump-off value of the 10-year yield is
close to or below its historical minimums at the scenario jump-off. In
particular, this element binds when the yield is below 0.8 percent at
the jump-off. This element further increases transparency on the yield
trough level in the scenarios in various potential interest rate
environments outside historical experiences.
To illustrate how the trough levels of scenarios consistent with
this guide would compare to the past stress test scenarios, consider
the history of the 10-year yield and its scenario values at the trough
over the period in which the Board has been conducting annual stress
tests, from 2014 to 2025 (Figure 5). The past stress test scenario
troughs are depicted as dots, whereas the range that is spanned by the
proposed guide is indicated by the solid lines, which incorporate the
lower bound. The dashed line depicts the quarterly average of the 10-
year Treasury yield observed in the data. This period contains both
low- and high-interest rate environments, and the quarterly average of
the 10-year Treasury yield (depicted as a dashed line) has been 2.5
percent, with a range between 0.6 and 4.5 percent at the jump-off
quarter.
[GRAPHIC] [TIFF OMITTED] TP18NO25.059
For periods when the 10-year yield is below 1.5 percent, such as
the period surrounding the COVID-19 pandemic, the guide would prescribe
the lower bound for the 10-year yield. In other periods between 2013-
2025, the 10-year yield has been high enough such that the lower bound
of the guide is not strictly binding after applying the minimum amount
of decline in the guide. In a higher-rate environment, a severe drop in
the 10-year yield would not necessarily imply a yield close to zero.
Figure 5 illustrates that the range of troughs consistent with the
proposed guide usually includes the values
[[Page 51912]]
featured in the past stress test scenarios during the low-interest-rate
environment from 2013 to 2022 (the dots are located within the guide-
prescribed range, or close to the binding lower bound in most years).
In three of those years, the Board chose a level for the trough that
was above the maximum level that would have been allowed by the guide
and in two cases the Board chose a level modestly below the minimum
level consistent with guide. With interest rates having risen to
moderate levels between 2023 and 2025, the guide would have required
the Board to choose a slightly higher trough in 2023 and 2025 and a
notably higher trough in 2024 than the Board chose in those scenarios.
---------------------------------------------------------------------------
\205\ Source: Quarterly average of the yield on 10-year U.S.
Treasury notes, constructed by Federal Reserve staff based on the
Svensson smoothed term structure model. See Svensson (1995), supra
note 184.
---------------------------------------------------------------------------
The Board considers these deviations from past scenarios to be an
acceptable consequence of adopting the guide, given its goal of
increasing predictability and transparency in the stress test. As the
deviations from past scenarios would have been in both directions, the
Board expects that the proposed guide will be broadly consistent with
maintaining an average level of severity of stress tests going forward
that is similar to what it has been under the previous scenario design
framework.
In setting the 10-year Treasury yield trough value, the Board could
consider the overall level of cyclical systemic risk, and the current
level of the 10-year Treasury yield as a benchmark measure of overall
economic and financing conditions. As discussed in earlier Section IX.F
of this Supplementary Information, the Board expects to calibrate the
increment in the 10-year yield in consideration of observable cyclical
systemic risk. The Board would also consider how declines in Treasury
yields, which decrease net income but increase the market value of
firms' long-term securities holdings, interact with the current
vulnerabilities in the banking sector. In general, a decline in long-
term interest rates may have a positive or negative effect on the
severity of the scenario for a given firm depending on the firm's
exposure to interest rate risk, which may vary from year to year
depending on the firm's portfolio. In reaching its determination to set
this guide in an annual scenario, the Board will consider how the
choice would promote stress test credibility and the resilience of the
financial system to even worse outcomes.
If the Board observes that cyclical systemic risks were increasing
in a period of sustained robust expansion, the Board might choose a
scenario that is more intense than normal. The choice would also depend
on firms' exposure to interest rate risk. Conversely, the Board could
specify a scenario that is less intense than normal in an environment
where systemic risks appeared subdued, such as in the early stages of a
recovery, provided that doing so remained consistent with the goal of
ensuring that firms were properly capitalized to withstand severe
economic and financial conditions. A less severe scenario could also be
appropriate when underlying market uncertainty and financial stress
start to recede and higher-than-usual credit losses were either already
realized--or are in the process of being realized--and thus removed
from firms' balance sheets. The choice would consider that the scenario
does not add unduly to remaining stress, thereby exacerbating the
initial adverse shock, and it would be particularly appropriate if the
Board judges that firms are already taking steps to reduce their risk--
for instance, by potentially restricting lending to otherwise qualified
borrowers.\206\
---------------------------------------------------------------------------
\206\ Evidence of market uncertainty and financial stress
receding could include strong stock market performance or positive
economic news related to GDP, inflation, unemployment, or nonfarm
payroll. Evidence that credit losses are being realized could
include elevated charge-offs on loans and leases or loan-loss
provisions in excess of gross charge-offs.
---------------------------------------------------------------------------
Alternative Trough Guide Option
As an alternative, the Board considered a guide in which the 10-
year Treasury yield would decline by 2.0 percentage points regardless
of the jump-off conditions. The lower bound would still bind. The
decline of 2.0 percentage points is chosen based on the same
observations shown in Table 18. In particular, 2.0 percentage points is
close to the average decline in the 10-year Treasury yield observed
during the financial stress episodes (1.9 percentage points) and the
average decline in previous severely adverse scenarios (1.7 percentage
points). The choice of a single value in the middle of the range
proposed in the more flexible version of this guide balances the
offsetting effects of interest rates on net interest margin and fair
value of securities.
The Board considered this alternative because of its goal of
increasing transparency and predictability of the stress test, while
maintaining sufficient severity. The Board recognizes that this
alternative guide would not avoid adding sources of procyclicality as
effectively as the proposed guide. In particular, it would reduce the
Board's flexibility during periods of moderate or high interest rates
to test the resilience of firms' net income to a sharper decline in
interest rates. However, as noted above, declines in yields have
offsetting effects on firms' regulatory capital in the stress test
because they decrease net income but increase the market value of their
long-term securities holdings. Thus, a more flexible guide would allow
the Board to balance its assessment of these two vulnerabilities in the
stress test scenario.
While the alternative troughs fall within the range determined by
the proposed guide, these trough levels can be significantly higher or
lower than the values chosen by the Board in prior severely adverse
scenarios. These deviations could impair the ability of the Board to
ensure the stress test severity that fully considers the risks that are
apparent in relevant indicators of economic and financial conditions,
particularly those related to inflation and inflation expectations,
when determining the trough value. The Board views the alternative
guide as reasonable. Compared to the proposed guide, the alternative
guide would provide firms and the public with increased predictability
regarding the trough value to be set for 10-year Treasury yields.
However, this increased predictability under the alternative guide
comes at the expense of the added flexibility inherent in the proposed
guide to set the trough based on risks that are apparent in relevant
indicators of economic and financial conditions and to avoid adding
sources of procyclicality in the proposed guide. The purpose of the
alternative guide discussion is to invite comment on a reasonable
alternative considered by the Board and to transparently lay out the
Board's present arguments for choosing the proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Trough Value Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the trough and its timing signify the magnitude and timing
of the most severe point in this trajectory. The proposed guide
suggests that the 10-year Treasury yield would reach the trough in
quarters 1 to 4 of the scenario. This timing is chosen such that the
trough is consistent with the scenario narrative: the severely adverse
scenario is triggered by a sizeable financial shock combined with a
pronounced increase in unemployment and decrease in inflation. In
response to these developments, both short- and long-term interest
rates typically would fall sharply. The timing of the trough is also
broadly consistent with the historical
[[Page 51913]]
data (Table 18). Averaging across the four financial stress episodes,
the trough is placed in the sixth quarter, but the trough occurred
earlier during the two most recent recessions.\207\ The 10-year yield
reached its trough in quarter 3 during the 2007-2009 financial crisis
and in quarter 4 during the COVID-19 pandemic. In the past stress test
scenarios, the trough was reached in quarter 1, on average. In setting
this part of the guide in an annual scenario, the Board will consider
the same indicators and other factors described above for the choice of
the trough in the 10-year rate, so as best to promote stress test
credibility and the resilience of the financial system to even worse
outcomes.
---------------------------------------------------------------------------
\207\ These four episodes include 1990Q3-1991Q1, 2001Q1-2001Q4,
2008Q3-2009Q2 (Lehman Brothers bankruptcy as a forcing event), and
2019Q4-2020Q2 recessions.
---------------------------------------------------------------------------
Trajectory to Trough Value
The proposed guide stipulates that the largest share of the decline
in the 10-year Treasury yield would be realized in quarter 1. A rapid,
frontloaded decline of the 10-year yield to its trough would be
consistent with the scenario narrative and the implied dynamics of the
other variables, mainly a large rise in unemployment and resulting
declines in inflation and output. In response to these developments,
both short- and long-term interest rates would fall sharply, consistent
with the Board's macroeconomic model for stress testing, because the
expectational component of the 10-year Treasury yield accounts for the
future expected realizations of the macro variables that determine the
policy rate rule.\208\
---------------------------------------------------------------------------
\208\ See supra note 196.
---------------------------------------------------------------------------
To determine the specific path of the 10-year Treasury yield for a
given trough timing, the Board considered a simple formula that can map
the trough value timing to a share of decline in quarter 1. To do so,
the Board considered lower and upper bound of trough timing described
in the previous section. If the trough timing is quarter 1 (e.g., lower
bound of the range), then the formula should yield 100 percent of the
decline occurring in the first quarter. For trough timing of quarter 4,
the Board took example of COVID-19 pandemic. During the COVID-19
pandemic, the 10-year yield reached its trough in quarter 4, and 52
percent of the decline in the 10-year yield was realized during the
first quarter. Using these reference points, the Board concluded that
the following simple formula could set the approximate share of the
decline realized in quarter 1 as:
100%-15% * (Trough value timing-1).
This simple formula stipulates that when the scenario trough is
realized in Quarter 4, about 55 percent of the decline would be
realized in Quarter 1:
100%-15% * (4-1) = 55%.
This is broadly in line--if not exactly in line--with the data from
the COVID-19 pandemic. The guide also stipulates that, after the
initial decline realized in quarter 1, the yield declines to its trough
at smoothly decreasing percent reductions.
BBB Yield
The stress test scenarios set out the trajectory of the BBB
corporate spread, measured by the quarterly average of ICE BofA U.S.
Corporate 7-10 Year Yield-to-Maturity Index relative to the 10-year
Treasury yield.\209\ The BBB corporate spread represents the
performance of corporate debt rated as investment grade by a major
ratings agency.
---------------------------------------------------------------------------
\209\ The source for the BBB corporate spread series is ICE
BofAML U.S. Corporate 7-10 Year Yield-to-Maturity Index, ICE Data
Indices, LLC, (C4A4 series). The 10-year yield is computed as the
quarterly average of the yield on 10-year U.S. Treasury notes,
constructed by Federal Reserve staff based on the Svensson smoothed
term structure model. See Svensson (1995), supra note 184.
---------------------------------------------------------------------------
Although firms subject to the supervisory stress test do not hold
substantial volumes of BBB corporate bonds on their balance sheets,
they make business loans to large- and middle-market firms and hold
other types of business debt on their balance sheets, e.g., commercial
and industrial (C&I) loans and collateralized loan obligations (CLOs).
Corporate bond spreads and CLO spreads tend to move together in times
of financial stress and high uncertainty. C&I loans to large- and
middle-market firms, some of whom are also issuers of corporate bonds,
account for 65 percent of total C&I loans. Because of these
similarities with bond-issuing firms, changes in business conditions
that underlie changes in spreads on BBB corporate bonds would affect
these borrowers as well (and hence the balance sheets of the stress
tested firms). In fact, empirical research finds that bank borrowers
are more sensitive to macroeconomic and financial shocks than publicly-
traded borrowers due to their relatively more-restricted access to
funding resources. Hence, in the context of the severely adverse
scenario, the Board views BBB corporate bond spreads as a measure
representing conditions in the business sector more generally.
Instead of a higher frequency, such as daily, for which the
underlying data is available, the Board uses a quarterly average of the
BBB spreads in the stress test scenario for several reasons. First, BBB
bonds face liquidity issues and their prices can be quite volatile at
higher frequencies for reasons unrelated to underlying business
conditions.\210\ Using quarterly averages strikes a balance between
being sensitive enough to capture market trends and stable enough to
avoid overreaction to high-frequency volatility. Relatedly, as noted
above, in the context of stress testing, the BBB spreads provide a good
representation of business borrowing and underlying economic
confidence. Therefore, removing short-term noise from this measure via
quarterly averaging results in a more reasonable representation of
underlying business borrowing conditions.
---------------------------------------------------------------------------
\210\ There is empirical support for excessive volatility in
corporate bonds and find that it has little relation to firm
fundamentals. J. Bao, & J. Pan, Bond Illiquidity and Excess
Volatility, 26 Review of Financial Studies 3068-3103 (2013). In a
working version of the paper, the authors stress that such excessive
volatility is pervasive at higher frequencies, being the strongest
at daily and weekly horizons and staying significant at monthly
horizons. J. Bao & J. Pan, Excess Volatility of Corporate Bonds
(2008), https://haas.berkeley.edu/wp-content/uploads/bond_vol.pdf.
---------------------------------------------------------------------------
In determining the appropriate level of scenario severity, the
Board adheres to scenario design principles discussed in the earlier
Section IX.F of this Supplementary Information. While doing so, the
Board also strives to avoid introducing additional sources of
procyclicality into the financial system. In the context of the BBB
spreads, these principles are applied in calibrating three key aspects
of the guide: the peak value, the timing of the peak value, and the
trajectory to peak. This approach helps ensure that the BBB spread
guide aligns with the established stress testing literature while
mitigating potential systemic risks for the financial system.
The rest of this section is organized as follows. First, an
overview of the BBB spread guide components is given in Table 19, which
is followed by the description of the component of the guide that
determines the peak of the spread. The next subsection provides the
data- and scenario-based rationale for the calibration of the peak
component. Next is a discussion of an alternative calibration for the
peak component, comparing the implementation and caveats to the
proposed guide option. Finally, additional guide parameters (peak value
timing and trajectory to the peak) and the rationale for their
calibration are discussed.
[[Page 51914]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.060
a. Peak Value Component of the Guide
The BBB corporate bond yield is expected to move such that its
spread relative to the 10-year Treasury yield would either increase
from its initial level by 100 basis points or to a level ranging
between 500 and 600 basis points, whichever results in a higher level.
Data- and Scenario-Based Rationale for the Peak Value
In line with the guiding principles emphasized by the stress
testing literature and discussed in Section IX.F of this Supplementary
Information, the Board references the behavior of the BBB spreads
during financial stress episodes, including the 2007-2009 financial
crisis, to calibrate the guide for BBB spreads in the supervisory
stress test scenarios. The higher end of the range for the peak level
(i.e., 600 basis points) corresponds to the quarterly-average peak
value observed during the 2007-2009 financial crisis (Table 19).
Additionally, weekly averages of the BBB spread peaked at 688 basis
points over the same period.\211\ The lower end of the range for the
peak level (500 basis points) is motivated by the data as well. A level
of 500 basis points also constitutes a severe BBB spread value from a
statistical point of view.\212\ At daily frequency, the BBB spread
reached values of around 500 basis points several times during the
2007-2009 financial crisis and during the COVID-19 pandemic the BBB
spread reached about 450 basis points.\213\
---------------------------------------------------------------------------
\211\ Weekly average calculated using ICE BofA U.S. Corporate 7-
10 Year Yield-to-Maturity Index (ICE Data Indices, LLC) and the
yield on 10-year U.S. Treasury notes, constructed by Federal Reserve
staff based on the Svensson smoothed term structure model. See
Svensson (1995), supra note.
\212\ For instance, in the weekly data from December 1988
through February 2025, 500 basis points and 600 basis points
represent the top (i.e., the riskiest) percentiles of the BBB spread
historical distribution: 98.5 and the 99.3, respectively.
\213\ The daily frequency BBB spread peak during the COVID-19
pandemic measured about 450 basis points, before declining after
unprecedented government support programs were announced in March of
2020. Stress tests are designed to assess firms in the absence of
such government support. During the 2007-2009 financial crisis, the
weeks in which spreads exceeded 500 basis points preceded the weeks
with even higher BBB spread values.
---------------------------------------------------------------------------
These examples illustrate that the value of 500 basis points
represents a severe point in the historical distribution of the BBB
spread observed during crisis events, which could be followed by
further worsening of conditions. Even if peaks of 500 to 600 basis
points have been rather short-lived, they could potentially trigger
events that would cause inadequately capitalized firms to fire-sale
their assets--a risk the Board seeks to reduce through the use of
stress testing. Moreover, setting a floor for the BBB spread at 500
basis points recognizes that, not only are cyclical systemic risks
likely to build up at financial intermediaries during robust
expansions, but that these risks are also easily obscured by a buoyant
environment.
To ensure sufficient severity in the event that the BBB spread at
the start of a stress test cycle is around or higher than the peak
levels attained in the history (e.g., above 500 basis points), the
Board contemplates a minimum increment of 100 basis points, in line
with the principle that adequate severity requires a guide to be able
to go somewhat beyond historical experiences when initial conditions
warrant.\214\ The minimum increment of 100 basis points ensures
adequate scenario severity, maintaining the credibility of the stress
test while at the same time constraining the peak from becoming unduly
contractionary and deviating too far from historically observed
levels.\215\ Applying a larger value of a minimum increment (e.g., 200
basis points) could result in a peak level that is unjustifiably
distant from historical peaks and might not allow the Board to reflect
near-term changes, such as emerging signs of financial stabilization,
resulting in inappropriately high scenario severity at a time when the
economy and financial markets are already stressed.
---------------------------------------------------------------------------
\214\ See Schuermann (2014), supra note 99.
\215\ If a future financial distress event causes the BBB spread
to rise beyond the current peak of about 600 basis points, the Board
may consider an update of the peak range to reflect that new
empirical evidence in subsequent future tests.
---------------------------------------------------------------------------
BILLING CODE 6210-01-P
[[Page 51915]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.061
To illustrate how the Board would use this guide to formulate the
scenarios, and how the implied peak levels of the guides compare to the
past stress test scenarios, consider the jump-off values in 2014-2025
cycles, the period during which the Board has been conducting stress
tests (Figure 6). The past stress test scenario peaks are depicted as
dots in the figure, whereas the proposed guide is indicated as a range
by the solid lines. This period contains both stressful times (the
COVID-19 pandemic) as well as the slow recovery after 2009 and some
periods of very low unemployment and robust growth. It is therefore
quite representative in capturing a variety of jump-off values. In this
time frame, the quarterly average of the BBB spread (depicted as a
dashed line) has been between about 100 and 265 basis points at the
jump-off quarter, while the average of those quarterly jump-off values
was about 170 basis points. Going to 500 or 600 basis points from such
jump-off values represented a substantial increase in the spread,
possibly more than 400 basis points. This is a plausible increase when
markets become strained or bad economic news pervades. For instance,
during the 2007-2009 financial crisis, the difference between the
average BBB spread during 2007Q3 and the BBB spread at the peak that
episode (2008Q4) amounted to 431 basis points (Table 20).
---------------------------------------------------------------------------
\216\ Federal Reserve staff calculated BBB spread using the U.S.
BBB corporate yield, computed using the quarterly average of ICE
BofAML U.S. Corporate 7-10 Year Yield-to-Maturity Index (ICE Data
Indices, LLC, C4A4 series), and the quarterly average of the yield
on 10-year U.S. Treasury notes, constructed by Federal Reserve staff
based on the Svensson smoothed term structure model. See Svensson
(1995), supra note 184.
---------------------------------------------------------------------------
Figure 6 also illustrates that the peak range of 500 to 600 basis
points implied by the proposed guide brackets the peak values of the
BBB spread used in the past stress test scenarios (the dots are located
within or on the borders outlined by the solid lines throughout the
time span of the figure). In other words, the proposed guide is
consistent with the Board's past stress test practices in determining
the peak.
[[Page 51916]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.062
BILLING CODE 6210-01-C
In its formulation of the annual scenarios, the Board could
consider the overall level of cyclical systemic risk, or the current
level of the BBB spreads as a contemporaneous indicator of uncertainty
and financial stress. As discussed in Section IX.F of this
Supplementary Information, the Board expects to calibrate the increment
in the BBB spreads based on its views of the status of cyclical
systemic risk. Specifically, the Board would be more likely to set the
BBB spreads peak value at the higher end of the range if the Board
expects that cyclical systemic risks are high (as it would be after a
sustained long expansion), and alternatively would be more likely to
set the peak value to the lower end of the range if cyclical systemic
risks are low (as it would be in the earlier stages of a recovery),
provided doing so remained consistent with the goal of ensuring that
firms were properly capitalized to withstand severe economic and
financial conditions. This might result in a scenario that is more
severe than normal if the Board expects that cyclical systemic risks
were increasing in a period of sustained robust expansion.
---------------------------------------------------------------------------
\217\ Calculated using data from ICE Data Indices, LLC; the
quarterly average of the yield on 10-year U.S. Treasury notes,
constructed by Federal Reserve staff based on the Svensson smoothed
term structure model; and Federal Reserve staff estimates. See
Svensson (1995), supra note.
---------------------------------------------------------------------------
Conversely, it would also allow the Board to specify a scenario
that is less severe than normal in an environment where systemic risks
appeared subdued, such as in the early stages of a recovery. The lower
end of the increase range could also be appropriate when underlying
market uncertainty and financial stress start to recede and higher-
than-usual credit losses stemming from previous elevated levels of the
BBB spreads were either already realized--or are in the process of
being realized--and thus removed from firms' balance sheets. This
choice would consider that the scenario does not add unduly to
remaining stress, thereby exacerbating the initial adverse shock, and
it would be particularly appropriate if the Board judges that firms are
already taking steps to reduce their risk; for instance, by potentially
restricting lending to otherwise qualified borrowers.\218\
---------------------------------------------------------------------------
\218\ Evidence of market uncertainty and financial stress
receding could include decreased defaults in public bond markets,
strong stock market performance or positive economic news related to
GDP, unemployment or nonfarm payroll. Evidence that credit losses
are being realized could include elevated charge-offs on loans and
leases or loan-loss provisions in excess of gross charge-offs.
---------------------------------------------------------------------------
Alternative Peak Guide Option
As an alternative, the Board also considered a guide that would
choose a peak level as a maximum between 600 basis points and an
increase from the jump-off value by 100 basis points. The justification
for considering this peak calibration is as follows. Unlike the
proposed guide, the alternative allows for less discretion and
therefore would provide more certainty to firms and to market
participants about the severity of the stress test each year. However,
the Board considered the importance of ensuring that the chosen
calibration would be sufficiently severe, because, as noted above,
insufficiently severe stress tests can undermine the credibility of the
results. Therefore, to attain adequate scenario severity for this
option, the Board considered the peak calibration level of 600 basis
points--the value corresponding to the BBB spread peak observed during
the 2007-2009 financial crisis and the upper bound of the range
considered for the proposed guide. This alternative guide could be less
desirable as it is less flexible and may end up being too severe,
especially during economic downturns, when the proposed guide would
offer the flexibility to choose a lower peak from the range that could
avoid adding sources of procyclicality in the results.
Consider the application of the alternative guide in 2013-2024
against the peaks of past scenarios and the proposed guide. Given the
initial conditions in this time period, the alternative guide would
prescribe the 600 basis points for the peak value in all quarters of
the considered time span. Compared with the past stress tests,
[[Page 51917]]
such prescriptions are often more severe, resulting in the peaks that
can be as much as 100 basis points higher than those of the past stress
tests.
Although the proposed and the alternative guides are both
discussed, and the Board views the alternative guide as reasonable, it
was judged to be inferior to the proposed guide as discussed in this
section. The purpose of the alternative guide discussion is to invite
comment on a reasonable alternative considered by the Board and to
transparently lay out the Board's present arguments for choosing the
proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Peak Value Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the peak and its timing signify the magnitude and timing
of the most severe point in this trajectory. The guide stipulates that
the peak level in the scenario would be reached in quarter 3 or quarter
4, which is consistent with historical observations. In post-war
recessions, the BBB spread reached its peak in quarter 4 (on average),
whereas the 2007-2009 financial crisis yields a peak in quarter 2 (see
Table 20). The empirical literature that studies responses of corporate
spreads to shocks (e.g., unexpected increases in uncertainty or
financial riskiness) often documents a delayed peak. For instance, the
response of the corporate spread to an uncertainty shock can peak after
month 6 (into quarter 3) in the U.S. data.\219\ In the past stress test
scenarios, the peak was also reached in quarter 4, on average.
---------------------------------------------------------------------------
\219\ See Caldara (2016), supra note . The delayed peak feature
is particularly prominent for the Jurado et al. (2015) measure of
uncertainty--a widely accepted measure in this literature. K. Jurado
et al., Measuring Uncertainty, 105 Am. Econ. Rev. 1177-1216 (2015).
---------------------------------------------------------------------------
The Board expects that the timing of the start of the stress period
should sometimes differ from the start date of the recession determined
by the NBER. For potentially fast-moving variables (such as BBB spread,
equity prices or VIX), the Board times the onset of the stress period
during the 2007-2009 Financial Crisis based on the Lehman Brothers
bankruptcy on September 15, 2008. This event is widely considered to be
the most significant of the events that roiled financial markets during
the 2007-2009 Financial Crisis.\220\ As stress test data operate at
quarterly frequency, the Board's timing of this event for purposes of
dating the peak of the BBB corporate spread is in 2008Q3. Indeed, the
BBB corporate spread remained largely flat between 2008Q1 and 2008Q2,
rising somewhat in 2008Q3 (because the Lehman Brothers bankruptcy
occurred close to the end of the quarter, it had little effect on the
quarterly average) before increasing sharply to the observed maximum in
2008Q4.\221\ Therefore, the focus on the Lehman Bankruptcy as the
triggering event is more consistent with the stress test scenario
narrative in which a financial shock sets the stress test scenario
dynamics in motion than the NBER recession date.
---------------------------------------------------------------------------
\220\ See, e.g., R. Wiggins et al., The Lehman Brothers
Bankruptcy A: Overview, 1 Journal of Financial Crises 39-62 (2019).
\221\ Demonstrated by the calculation of the BBB spread over
time using the U.S. BBB corporate yield, computed using the
quarterly average of ICE BofAML U.S. Corporate 7-10 Year Yield-to-
Maturity Index (ICE Data Indices, LLC, C4A4 series), and the
quarterly average of the yield on 10-year U.S. Treasury notes,
constructed by Federal Reserve staff based on the Svensson smoothed
term structure model. See Svensson (1995), supra note 184.
---------------------------------------------------------------------------
Trajectory to Peak Value
To reach the peak spread value, the guide prescribes that the
highest share of the spread increase (about 60 to 80 percent) occurs in
the first quarter of the scenario. Such frontloading of the spread
increase is consistent with the historical evidence and academic
literature.\222\ For instance, in the 2007-2009 financial crisis, the
largest increase in the spread (about 67 percent of the jump-off. A
very similar result emerges when considering the Enron/Dotcom stress
episode and 1990 bond market stress episode.\223\ On average (across
all three bond market stress episodes), about 66 percent of the
increase to the peak in the spread was realized in a single quarter
after the onset of the stress episode. After quarter one and until the
peak is reached, the guide stipulates a smooth trajectory with half of
the remaining adjustment made in quarter two and with the rest of the
adjustment made either in quarter three (when the peak occurs in
quarter three) or equally distributed between quarter three and four
(when the peak occurs in quarter four). As an example, if the increase
share in the first quarter was around 60 percent, then the adjustment
in quarter two would be about 20 percent with the remaining 20 percent
in quarter three (if the peak is in quarter three) or with the
remaining distributed approximately 10 percent each in quarter three
and four (if the peak is in quarter four). This simple adjustment rule
mimics a hump-shaped response of the corporate spread to shocks, a
feature well-documented in the empirical literature.\224\
---------------------------------------------------------------------------
\222\ In the academic literature, spreads are well-known to be
contemporaneous indicators that move the most at the onset of a
stress event or crisis. For instance, Krishnamurthy (2025) documents
rapid changes in spreads at the onset of financial crises, whereas
Bernanke (2005) classify spreads and stock market prices as ``fast-
moving'' variables that respond to shocks on impact. A.
Krishnamurthy & T. Muir, How Credit Cycles across a Financial
Crisis, 80 J. of Fin. 1339-78 (2025) (``Krishnamurthy (2025)''); B.
Bernanke et al., Measuring the Effects of Monetary Policy: A Factor-
Augmented Vector Autoregressive (FAVAR) Approach, 120 Q.J. of Econ.
387-422 (2005). Caldara (2016), supra note 182, provides empirical
evidence of such behavior of spreads in response to financial shocks
and uncertainty shocks.
\223\ For a more detailed discussion of the Enron/Dotcom
episode, see D. Romer, Preventing the Next Catastrophe: Where Do We
Stand? (Conference paper). Rethinking Macro Policy II: First Steps
and Early Lessons Conference (2013); M. Bordo & J. Haubrich, Deep
Recessions, Fast Recoveries, and Financial Crises: Evidence from the
American Record, 55 Econ. Inquiry 527-41 (2017). The 1990 bond
market stress episode is discussed, for example, in M. Wolfson,
Financial Crises: Understanding the Postwar U.S. Experience (1994).
\224\ Some of the recent examples include D. Caldara & E.
Herbst, Monetary Policy, Real Activity, and Credit Spreads, 11 Am.
Econ. J. 157-92 (2019) and Caldara (2016), supra note 182.
---------------------------------------------------------------------------
Mortgage Rate
The stress test scenarios sets out trajectories for several
variables, including the mortgage spread as proxied by the quarterly
average of weekly series for the interest rate of a conventional,
conforming, 30-year fixed-rate mortgage, obtained from the Freddie Mac
Primary Mortgage Market Survey relative to the 10-year Treasury
yield.\225\ For purposes of this guide, mortgage spread refers to the
difference, in basis points, between mortgage and Treasury rates
defined above.
---------------------------------------------------------------------------
\225\ The 10-year Treasury yield is calculated using the
quarterly average of the yield on 10-year Treasury notes by the
Federal Reserve Board based on the Svensson smoothed term structure
model. See Svensson (1995), supra note 184.
---------------------------------------------------------------------------
In the supervisory stress test, the mortgage spread can act as both
(i) an indicator of stress for certain important assets under the
scenarios and (ii) a source of stress for firms subject to the
supervisory stress test with substantial exposure to assets that are
tied to mortgage spreads, such as mortgage loan portfolio or mortgage-
backed securities, which are reported by firms on FR Y-14M, Schedule A
(First Lien) and FR Y-14Q, Schedule B (Securities). Firms subject to
the supervisory stress test typically have substantial exposure to the
assets referenced in the mortgage spread, and as a result, by
incorporating the mortgage spread into scenarios, stress tests help
ensure that firms are prepared for a wide range of market conditions,
including periods of
[[Page 51918]]
elevated mortgage spreads, in part reflecting financial shocks and any
associated economic downturn. This helps maintain the overall stability
and resilience of the financial system.
The Board uses a quarterly average of the mortgage rate spread in
the stress test scenarios. Quarterly averages smooth out excessive (and
potentially irrelevant) volatility that is present at weekly or monthly
frequencies. Using quarterly averages strikes a balance between being
sensitive enough to capture market trends and stable enough to avoid
overreaction to market volatility that is not representative of
underlying trends in housing markets or the broader economy.
In determining the appropriate level of scenario severity, the
Board adheres to scenario design principles discussed in Section IX.F
of this Supplementary Information. While doing so, the Board also
strives to avoid introducing additional sources of procyclicality into
the financial system. In the context of the mortgage spread, these
principles are applied in calibrating three key aspects of the guide:
the trough value, the timing of the trough value, and the trajectory to
trough. This approach helps ensure that the mortgage spread guide
aligns with the established stress testing literature while mitigating
potential systemic risks for the financial system.
The rest of this section is organized as follows. First, Table 21
provides an overview of the mortgage rate guide components, which is
followed by a description of the peak component for the guide. The next
subsection provides the data- and scenario-based rationale for the
calibration of the peak component. A discussion of an alternative peak
option follows in the next section, comparing the implementation and
caveats to the proposed guide option. Finally, additional guide
parameters (trough value timing and trajectory to the peak) and the
rationale for their calibration are discussed.
BILLING CODE 6210-01-P
[GRAPHIC] [TIFF OMITTED] TP18NO25.063
a. Peak Value Component of the Guide
The mortgage rate is expected to move such that its spread relative
to the 10-year Treasury yield would increase from its jump-off level
(i.e., the value of the variable in the quarter before the start of the
scenario) to a range determined by that level plus 70 basis points to
160 basis points, with a lower bound of 300 basis points.
Data- and Scenario-Based Rationale for the Peak Value
In line with the guiding principles emphasized by the stress
testing literature and discussed in Section IX.F of this Supplementary
Information, the Board uses the behavior of the mortgage spreads during
financial stress episodes, including the 2007-2009 financial crisis, to
calibrate the guide for the mortgage spread in the supervisory stress
test scenarios. In particular, the Board considers the behavior of the
mortgage spread in three severe recessions, including the 2007-2009
financial crisis, to calibrate the guide for mortgage spreads in the
supervisory stress test scenarios. In particular, the calibration of
the lower bound of 300 basis points in the guide is based on evidence
from historical stress episodes along with the principle of
conservatism. The average peak value for the mortgage spread observed
in severe recessions has been 278 basis points (Table 22), ranging from
225 to 380 basis points.\226\ In the 2007-2009 financial crisis, the
peak mortgage spread measured about 249 basis points at a weekly
frequency.\227\ The calibration of the lower bound of 300 basis
points--a value that is slightly above the historical average during
severe recessions--speaks to the guiding principle that adequate
severity should be somewhat beyond historical experiences. In addition,
setting a floor for the mortgage spreads at 300 basis points recognizes
the fact that, not only do cyclical systemic risks build up at
financial intermediaries during robust expansions, but that these risks
are also easily obscured by a buoyant environment.
---------------------------------------------------------------------------
\226\ A similar average peak value of 260 points is obtained
from averaging across episodes of housing market stress, which
include the 1973 recession along with the previously defined housing
recessions (1980Q2-1985Q2, 1989Q4-1997Q1, 2005Q4-2012Q1). See 12 CFR
252, Appendix A.
\227\ The spread measure at weekly frequency is obtained as an
average over daily values starting from Thursday of the previous
week and ending on Wednesday of the next week. Accordingly, the
value of approximately 249 basis points was reached in the calendar
week ending on December 21, 2008. A close value of 248 basis points
was reached in the calendar week ending on August 31, 2008.
---------------------------------------------------------------------------
[[Page 51919]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.064
The historical \228\ maximum value of the mortgage spread occurred
in the 1980-1985 episode--in a high-inflation environment with high
unemployment as well--and measured 404 basis points, based on quarterly
averages of the spread. Weekly averages of the spread during this
episode would result in a higher peak level of 541 basis points, which
was reached in the calendar week ending on April 20, 1980 (Figure 7).
Between 2022 and 2024, inflation accelerated, and the mortgage spread
rose above the 2007-2009 peak, reaching a quarterly-frequency maximum
of 284 basis points in 2023 Q2 (304 basis points at a weekly frequency,
in the calendar week ending on May 28, 2023) despite a strong economy
and well-functioning mortgage markets. Hence, guide calibration of the
mortgage spread should account for conditions in the housing market,
including interest rate volatility, and the phase of the business cycle
as described above, as well as the level of inflation and inflation
expectations at the jump-off quarter to elucidate their effect on
firms' balance sheets.
---------------------------------------------------------------------------
\228\ Quarterly average of weekly series for the interest rate
of a conventional, conforming, 30-year fixed-rate mortgage is
obtained from the Primary Mortgage Market Survey of the Federal Home
Loan Mortgage Corporation. Quarterly average of the yield on 10-year
Treasury notes is constructed by the Federal Reserve Board based on
the Svensson smoothed term structure model. See Svensson (1995),
supra note 184. Data also derived from Federal Reserve staff
calculations.
---------------------------------------------------------------------------
[[Page 51920]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.065
In its formulation \229\ of the annual scenarios, the Board could
consider the overall level of cyclical systemic risk, or the current
level of the mortgage spreads as a contemporaneous indicator of
uncertainty and financial stress. As discussed in Section IX.F of this
Supplementary Information, the Board expects to calibrate the increment
in the mortgage spreads based on its views of the status of cyclical
systemic risk. Specifically, the Board would be more likely to set the
mortgage spread peak value at the higher end of the range if the Board
expects that cyclical systemic risks are high (as it would be after a
sustained long expansion), and alternatively would be more likely to
set the peak value to the lower end of the range if cyclical systemic
risks are low (as it would be in the earlier stages of a recovery),
provided doing so remained consistent with the goal of ensuring that
firms were properly capitalized to withstand severe economic and
financial conditions. This might result in a scenario that is more
intense than normal if the Board expects that cyclical systemic risks
were increasing in a period of sustained robust expansion. Conversely,
it would also allow the Board to specify a scenario that is less
intense than normal in an environment where systemic risks appeared
subdued, such as in the early stages of a recovery. The lower end of
the range could also be appropriate when underlying market uncertainty
and financial stress start to recede and higher-than-usual credit
losses stemming from previously elevated levels of mortgage spreads
were either already realized or are in the process of being realized,
and thus removed from firms' balance sheets. This choice would consider
that the scenario does not add unduly to remaining stress, thereby
exacerbating the initial adverse shock, and it would be particularly
appropriate if the Board judges that firms are already taking steps to
reduce their risk--for instance, by potentially restricting lending to
otherwise qualified borrowers.\230\
---------------------------------------------------------------------------
\229\ Data derived from Primary Mortgage Market Survey of the
Federal Home Loan Mortgage Corporation, weekly and quarterly average
of the yield on 10-year Treasury notes, constructed by Federal
Reserve staff based on the Svensson smoothed term structure model.
See Svensson (1995), supra note 184. Data also derived from Federal
Reserve staff estimates.
\230\ Evidence of market uncertainty and financial stress
receding could include stronger lending growth, an easing of lending
standards, strong stock market performance or positive economic news
related to GDP, unemployment, or nonfarm payroll. Evidence that
credit losses are being realized could include elevated charge-offs
on loans and leases or loan-loss provisions in excess of gross
charge-offs.
---------------------------------------------------------------------------
Consider the application of the range component of the guide (70 to
160 basis points from the jump-off value) illustrated in Figure 8
(solid lines) for 2014-2025 stress test cycles. This time period is
illustrative as it contains various stages of the business and
financial cycle (normalization after the 2007-2009 financial crisis,
the COVID-19 pandemic and the normalization after it in a higher-
inflation environment). Accordingly, the initial conditions in this
period are quite representative. While the lower bound of the range
(300 basis points) was explained above, the application of the upper
part of the range results in values from 300 basis points to 440 basis
points, with the higher values achieved in 2022-2024, a period of
higher inflation. Per the discussion above, these values, while being
severe, do not deviate too far from historically observed values. And
consistently with historical experiences, these values also reflect the
inflation environment.
[[Page 51921]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.066
BILLING CODE 6210-01-C
To illustrate \231\ the implications of the guide, the Board
applies it to recent historical data and compares the implied peak
prescriptions with the corresponding peaks from past stress test
scenarios (Figure 8). From 2013Q1 until 2019Q4, the stress test peak
values, depicted by the dots, were quantitatively close to the upper
end of the range of the proposed guide, depicted by the solid lines
(Figure 8). From the onset of the COVID-19 pandemic in 2020 through
2021, the stress test peaks were mostly within the bands of the
proposed guide, while in 2022-2024 the stress test peaks were located
at or very close to the lower end of the range suggested by the
proposed guide. Summing up, comparison of the guide-implied peaks with
the past stress test peak values shows that the guide is broadly
consistent with past scenario values. The range of the guide should
allow the Board to account for risks that are apparent in relevant
indicators of economic and financial conditions and constrain the peak
to historically plausible bounds during normal periods, while adjusting
to future periods in which spreads may move toward record levels.
---------------------------------------------------------------------------
\231\ Data derived from Primary Mortgage Market Survey of the
Federal Home Loan Mortgage Corporation and Federal Reserve staff
estimates.
---------------------------------------------------------------------------
Alternative Peak Guide Option
As an alternative, the Board also considered a guide that would
choose a peak level as a maximum between 300 basis points and an
increase from the jump-off value by 100 basis points. A comparison of
the alternative and the proposed guides in 2013-2024 illustrates
episodes when the alternative guide deviates from the proposed guide.
As the alternative guide has a flat increment regardless of the
underlying conditions, it would propose systematically lower peak
values in the pre-pandemic period and systematically higher values in
the post-pandemic period, when compared with the stress test peaks.
Additionally, following this alternative guide would not allow the
Board to respond to risks not already reflected in the current level of
the mortgage spread.
Although the proposed and the alternative guides are both
discussed, and the Board views the alternative guide as reasonable, the
benefits to the public from increased predictability in the alternative
guide are considered to be outweighed by the added flexibility to
reflect risks related to mortgage finance that are apparent in relevant
indicators of economic and financial conditions or to avoid adding
additional sources of procyclicality in the proposed guide. The purpose
of the alternative guide discussion is to invite comment on a
reasonable alternative considered by the Board and to transparently lay
out the Board's present decision making in choosing the proposed guide.
b. Additional Guide Parameters and Rationale Behind Them
Peak Value Timing
In general, the entire 13-quarter trajectory of stress test
variables is important as it ultimately affects implied firm losses.
The value of the peak and its timing signify the magnitude and timing
of the most severe point in this trajectory. The proposed guide
stipulates that the peak level in the scenario would be reached in
quarters 3 to 4, which is consistent with historical observations and
past severely adverse scenarios (Table 22). The proposed guide sets a
range of peak timings between 3 and 4 quarters, whereas the alternative
guide eliminates this flexibility and stipulates a peak in quarter 3.
Keeping the magnitude of the peak constant, a more delayed peak timing
generally results in less severity of the overall path, as a less
abrupt worsening in conditions and credit quality gives firms and
mortgage borrowers more time to adjust to the shock. In contrast, an
earlier peak timing would increase the scenario severity.
For the proposed guide, a range in the timing (quarter 3 or quarter
4) is used as an additional lever (together with the peak magnitude
range) to avoid adding
[[Page 51922]]
sources of procyclicality in the stress test. An earlier peak timing
would increase the scenario severity. The factors that the Board would
consider in setting the timing of the peak are the same as those
discussed above influencing the level of the peak.
Trajectory to Peak Value
To reach the peak spread value, the guide prescribes that the
highest share of the spread increase (50 to 70 percent) occurs in the
first quarter of the scenario. After quarter one and until the peak is
reached, the guide stipulates a smooth trajectory. Such frontloading of
the spread increase is consistent with the historical evidence and
academic literature.\232\ Averaging across all financial stress
episodes, the share of the mortgage spread increase that occurs in the
first quarter after the onset of the stress is about 60 percent; in
other words, 60 percent of the distance from the jump-off point to the
peak is covered in the first quarter. This number is quantitatively
similar to the past stress test scenarios in 2013-2025, where on
average the corresponding share measures 64 percent. Averaging across
severe historical episodes in the data yields a share of 73 percent. At
the same time, there are severe episodes with a somewhat smaller
increase in the share occurring in the first quarter. For instance, the
severe episode surrounding the 1981 recession measured 47 percent of
the mortgage spread increase in the first quarter. Hence, both the
guide calibration (over 50 percent) as well as the average obtained
across the mortgage spread paths in 2013-2025 stress test scenarios (64
percent) lie within historically plausible bounds.
---------------------------------------------------------------------------
\232\ See supra note 222.
---------------------------------------------------------------------------
International Variables
As described in the Scenario Design Policy Statement, a scenario
that targets all specific risk factor groups includes judgement on the
projected paths of selected international variables. Recessions that
occur simultaneously across countries are an important source of stress
to the balance sheets of firms with notable international exposures but
are not a typical feature of the international economy even when the
United States is in recession. As a result, simply adopting the typical
path of international variables in a severe U.S. recession would likely
underestimate the risks stemming from the international economy.
Consequently, an approach that relies on both judgement and insights
from economic models informs the path of international variables. As
part of the review of the scenario design framework, the Board has
developed simple quantitative guides for the proposed paths of the
international variables used in the severely adverse scenario of the
supervisory stress test. Consistent with the Scenario Design Policy
Statement, the international component of the stress test scenarios
contains the path for real GDP, consumer price inflation, and the
nominal exchange rate for four country blocs: the euro area, the U.K.,
Japan, and Developing Asia.\233\ These economies capture the majority
of the foreign exposure of U.S. banks.
---------------------------------------------------------------------------
\233\ For the purpose of the supervisory stress tests, the Board
defines Developing Asia as China, India, Hong Kong, South Korea, and
Taiwan. Aggregate variables for this bloc (GDP, inflation, and the
nominal exchange rate) are obtained by weighting country-specific
variables by their relative share of the total nominal GDP
(expressed in U.S. dollars).
---------------------------------------------------------------------------
The following guides apply to each international variable:
A peak/trough value, which represents the extreme value
(either peak or trough, depending on the variable) that is typically
reached in the severely adverse scenario. For all variables the peak/
trough is reached after 4 quarters.
An end value for the last period in the scenario, that is
13 quarters after initial impact.
A scenario path, which describes the path of international
variables from the jump-off value to the peak/trough value and then to
the end value.
A scenario range, which specifies by how much each
variable can deviate from the scenario path to adapt to relevant
changes in banking, macroeconomic, or other conditions.
a. Overview of Approach
In designing the paths of the international variables in the
severely adverse scenario, the Board opted to follow a prescriptive
approach that is informed by the experience of the 2007-2009 financial
crisis. Given its global repercussions, the 2007-2009 financial crisis
is a useful benchmark for the economic effects of a large global
financial shock.
To generate the proposed paths of GDP and inflation in the four
economic regions for the severely adverse scenario, the Board first
computed the distance between the realized outcomes of GDP and
inflation during the 2007-2009 financial crisis from the baseline
forecasts prior to the onset of the 2007-2009 financial crisis. These
baseline forecasts were derived from publicly available forecasts from
the Blue Chip Economic Indicators and the IMF World Economic Outlook
(WEO).\234\ The Blue Chip and WEO forecasts provide values for year-
over-year real GDP growth and inflation. To forecast quarterly GDP
growth rates and inflation rates, quarterly values are first derived
from the annual growth rates using linear interpolation; then a
Hodrick-Prescott filter is used to smooth the path of GDP and inflation
across the forecast period.\235\ Based on this procedure, the Board
specifies guides for the values of the variables of interest. These
values are usually reached in the scenario, but the Board reserves the
right to depart from these values within specified ranges.
---------------------------------------------------------------------------
\234\ The Blue Chip data provide forecasts over a two-year
horizon and are updated at a monthly frequency. The WEO data provide
forecasts over a six-year horizon, which are updated biannually in
April and September/October each year. To produce the baseline
scenario, the Blue Chip forecasts are used for the first two years,
whereas the WEO forecasts are used for the remaining years.
\235\ The Hodrick-Prescott filter is an empirical tool that can
be employed to remove the cyclical component of a time series data.
This technique was developed by Whittaker (1923) and popularized in
economics by Hodrick and Prescott (1997). See E. Whittaker, ``On a
New Method of Graduation.'' Proceedings of the Edinburgh
Mathematical Association. 41: 63-75 (1923), https://doi.org/10.1017%2FS0013091500077853; R. Hodrick & E. Prescott, Postwar U.S.
Business Cycles: An Empirical Investigation, 29 J. of Money, Credit
& Banking 1-16 (Feb. 1997), https://doi.org/10.2307/2953682.
---------------------------------------------------------------------------
The data for the euro area, the U.K., and Japan were aggregated to
obtain identical guides for GDP and inflation for these Advanced
Foreign Economies (AFEs). The Board favored identical guides for these
regions to prevent possible credit allocation incentives that could
arise if guides differed systematically between the AFEs. However,
identical guides do not imply that the actual severely adverse scenario
features identical paths for the euro area, the U.K., and Japan. The
scenario paths of the three regions can vary with the given ranges.
The key elements of the international guides derived from this
procedure are summarized in Table 23; Figure 9 shows the behavior of
the variables of interest during the 2007-2009 financial crisis from
which the guides are derived. Detailed explanations and alternative
considerations are provided thereafter. For GDP, the deviation is
computed as percentage change from the baseline real GDP level. For
inflation, the deviation is computed as percentage point difference
from the baseline path of inflation. For exchange rates, the guide is
expressed in terms of percent deviation from the jump-off point.
[[Page 51923]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.067
Table 23 also provides ranges for each variable to allow for
flexibility: This flexibility enables judgment to be exercised to
capture the possibility that the foreign economies might react
differently to financial stress, either because of changes in the
global macroeconomic landscape or in country-specific vulnerabilities.
The prescriptive approach for the international variables in the
severely adverse scenario yields guides that are qualitatively and
quantitatively reasonable based on the Board's judgment and broadly
accepted models of international economies.\236\ The Board opted for
the more prescriptive approach because the advantages of increased
transparency and simplicity outweighed the disadvantage of less
flexibility.
---------------------------------------------------------------------------
\236\ See, e.g., M. Adrian et al., A quantitative model for the
integrated policy framework. IMF WP/20/122 (2020).
---------------------------------------------------------------------------
b. GDP
Trough Value
The magnitude of the prescribed economic downturn in the specified
foreign economies is informed by the deterioration in foreign economic
activity which occurred between 2008Q1 and 2009Q1. In particular, the
trough value for GDP is obtained by considering the deviation of the
real GDP level in 2009Q1 from a baseline path derived from the April
2008 IMF WEO forecast.\237\ This approach implies that, four quarters
after jump-off, the GDP levels in the euro area, the U.K., and Japan
are 7.5 percent below the baseline scenario.\238\ In Developing Asia,
GDP growth declines until GDP reaches a level 3 percent below baseline.
These values are in line with the behavior of real GDP reported in the
top panel of Figure 9.
---------------------------------------------------------------------------
\237\ The April 2008 WEO provides forecasts for annual GDP
growth and for annual inflation between 2008 and 2013. Blue Chip
forecast for international variables are not available until 2009.
The baseline for quarterly GDP growth rates and inflation, over the
period 2008Q2 until 2011Q2, is generated using the same procedure
employed to create the baseline scenario: first, quarterly values
for the GDP level are obtained by linearly interpolating the annual
growth rates available in the WEO forecast, and then a Hodrick-
Prescott filter is used to smooth the GDP level path across the
forecast period.
\238\ This value is in line with the average deviation from
baseline across these advanced economies in 2009Q1, when weighting
the deviations from baseline by the nominal GDP (in U.S. dollars) in
each country bloc in 2007.
---------------------------------------------------------------------------
End Value
In the euro area, the U.K., and Japan, the level of GDP at the end
of the severely adverse scenario deviates from the corresponding value
in the baseline (13 quarters after initial impact) by the same
magnitude as the trough value of GDP deviates from the corresponding
value in the baseline (4 quarters after initial impact). This
assumption implies that, in line with the experience of the 2007-2009
global financial crisis, after reaching the trough, GDP in the AFEs
grows at the same rate as in the (pre-crisis) baseline. The guide
proposes that GDP recovers more quickly in developing Asia, returning
to the GDP baseline in levels at the end of the scenario, in line with
the evidence from 2011Q2. These GDP paths are consistent with Figure 9
and with empirical evidence which suggests that advanced economies
suffer very persistent output losses following a financial crisis,
while developing economies experience less severe slowdowns.\239\
---------------------------------------------------------------------------
\239\ See, e.g., V. Cerra & S. Saxena, Growth Dynamics: The Myth
of Economic Recovery, 98 American Economic Review 439-57 (2008);
[Ograve]. Jord[agrave] et al., When credit bites back, 45 J. of
Money, Credit & Banking 3-28 (2013); M. Laeven & M. Valencia,
Systemic Banking Crises Revisited, International Monetary Fund, WP/
18/206 (2018).
---------------------------------------------------------------------------
Path
Real GDP reaches the reference trough four quarters after the jump-
off date and then gradually converges to the end value of the scenario.
After reaching the trough, the AFEs experience a similar GDP growth
rate in the scenario as in the baseline, whereas Developing Asia grows
faster in the scenario to catch up with the level of GDP in the
baseline. The path of GDP is created with a two-step procedure similar
to the one used to generate the baseline scenario. First, the series is
linearly interpolated between the jump-off value and the
[[Page 51924]]
trough value and from the trough value to the end value. Then, a
Hodrick-Prescott filter is used to smooth the GDP path over the
duration of the scenario. This approach generates a smooth path for GDP
consistent with business cycle dynamics.
Range
The path described above captures the GDP dynamics during the 2007-
2009 financial crisis. In determining the magnitude of the
international shock in the severely adverse scenario, the Board would
consider several factors, including the current economic performance of
foreign economies, the risks posed by country-specific vulnerabilities,
and the scope for countercyclical policy actions in each country bloc.
For example, in periods of sub-par foreign economic performance, the
Board would likely reduce the magnitude of the shock, whereas when
foreign growth is particularly strong, the magnitude of the shock would
be increased. In addition, the allocation of shocks across blocs can be
altered to highlight relevant country-specific risks. This strategy is
implemented by increasing or decreasing the severity of the shock, as
measured by the deviation of GDP from baseline at the scenario trough,
by at most 2.5 percent.\240\ As a result, at the trough, real GDP can
fall between 5 and 10 percent below the baseline in advanced foreign
economies, bracketing the real GDP outcomes of the three AFEs in the
2007-2009 financial crisis. For Developing Asia, real GDP can fall
between 0.5 and 5.5 percent below the baseline. These adjustments are
performed while keeping the overall stress on foreign economies, as
measured by the average GDP deviation from baseline at the trough,
within a range of 4 to 9 percent.\241\
---------------------------------------------------------------------------
\240\ This value is in line with the average standard deviation
of four-quarter GDP growth across the four country blocs, computed
over the pre-COVID-19 historical sample. When adjusting the
reference peak/trough value, the reference end value is adjusted
proportionally, to keep the ratio with the trough value constant.
\241\ Total effect on the foreign economies is computed
weighting the deviations from baseline in each country bloc by the
bloc's nominal GDP (in U.S. dollars) in the year preceding the jump-
off date. The range of -4 to -9 percent is centered around -6.5
percent--that is, the average deviation from baseline across the
foreign economies in 2009Q1.
---------------------------------------------------------------------------
[[Page 51925]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.068
c. Inflation
Trough Value
Inflation is assumed to decline below the baseline scenario for the
first four quarters of the simulation, consistent with the demand-
driven decline in GDP growth over the same period. This behavior is
broadly in line with the historical evidence between 2008Q1 and 2009Q1
for the four country blocs. The maximum decline in inflation is
calibrated to reflect the difference between the realized rate of
inflation and the one derived from the April 2008 IMF WEO forecast for
2009Q1 (middle panel of Figure 9Figure). This method provides that,
four quarters after the jump-off date, inflation is below baseline by
about 3 percentage points in the euro area, the UK, and in Japan, and
by 5 percentage points in Developing Asia.\242\
---------------------------------------------------------------------------
\242\ The decline in inflation in the Euro area, UK, and Japan
is in line with the average deviation from baseline across these
advanced economies in 2009Q1, when weighting the deviations from
baseline by the nominal GDP (in U.S. dollars) in each country bloc
in 2007.
---------------------------------------------------------------------------
End Value
After reaching the trough, inflation gradually returns to baseline
by the 13th quarter of the simulation. This inflation path is
consistent with the evidence from the 2007-2009 financial crisis, when
inflation returned to, or even surpassed, the WEO baseline by 2011. In
[[Page 51926]]
addition, academic research suggests that financial crises typically do
not have large or persistent effects on inflation.\243\
---------------------------------------------------------------------------
\243\ See, e.g., M. Schularick & A. Taylor, Credit Booms Gone
Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870-
2008, 102 Am. Econ. Rev. 1029-61 (2012); S. Gilchrist et al.,
Inflation Dynamics during the Financial Crisis, 107 Am. Econ. Rev.
785-823 (2017).
---------------------------------------------------------------------------
Path
The path for inflation is obtained by using the same strategy
employed for GDP, which combines linear interpolation and a Hodrick-
Prescott filter.
Range
If the path of GDP is different from the reference path, the path
for inflation will be adjusted to preserve the same ratio between the
deviation of GDP and the deviation of inflation from baseline under the
reference path--the values of these ratios are 2.5 for advanced foreign
economies and 0.6 for Developing Asia.\244\ As a result, inflation
declines between 2 percentage points and 4 percentage points below
baseline in the advanced foreign economies, and between 0.8 percentage
points and 9 percentage points below baseline in Developing Asia.
---------------------------------------------------------------------------
\244\ This value is equal to the ratio between the deviation of
real GDP from baseline and the deviation of inflation from baseline
at the trough in the international guides, that is the ratio between
7.5% and 3% for advanced foreign economies and the ratio between 3
percent and 5 percent for Developing Asia.
---------------------------------------------------------------------------
d. Exchange Rates
Trough/Peak Value
The Board assumes that over the first four quarters of the
simulation the U.S. Dollar experiences a 15 percent appreciation
against the Euro and the British Pound. This appreciation is in line
with the change in the Nominal Advanced Foreign Economies U.S. Dollar
Index between 2008Q1 and 2009Q1 (bottom left panel of Figure 9).\245\
Over the same period, the U.S. Dollar appreciates by 15 percent also
against the exchange rate for Developing Asia, in line with the
fluctuation in the Nominal Emerging Market Economies U.S. Dollar Index
between 2008Q1 and 2009Q1 (see bottom right panel of Figure 9).\246\
Consistent with the evidence between 2008Q1 and 2009Q1, the U.S. Dollar
experiences a mild 1 percent depreciation against the Japanese Yen,
which is typically considered a safe-haven currency, a currency which
retains its value during times of global economic stress.\247\
---------------------------------------------------------------------------
\245\ Nominal Advanced Foreign Economies U.S. Dollar Index
[DTWEXAFEGS], https://fred.stlouisfed.org/series/DTWEXAFEGS.
\246\ Nominal Emerging Market Economies U.S. Dollar Index
[DTWEXEMEGS], https://fred.stlouisfed.org/series/DTWEXEMEGS.
\247\ See, e.g., M. Botman, et al., The Curious Case of the Yen
as a Safe Haven Currency: A Forensic Analysis, International
Monetary Fund, WP/13/228 (2013). The Yen/USD exchange rate went from
approximately 99.9 at the end of 2008Q1 to 99.15 at the end of
2009Q1, a decline of about one percent.
---------------------------------------------------------------------------
End Value
Exchange rates gradually reach their peak/trough and then revert
back to their jump-off values by the end of the scenario horizon.
Path
The path for the exchange rate is obtained by using the same
strategy employed for GDP and inflation, which combines linear
interpolation and a Hodrick-Prescott filter.
Range
For exchange rates, which are highly volatile and only weakly
linked to macroeconomic fundamentals, the Board can adjust the maximum
fluctuation of each of the four foreign currencies within a range of
plus or minus 10 percent from the reference peak/trough value.\248\ For
each country bloc, the magnitude of the depreciation is adjusted
depending on the realized change in the exchange rate in the year
preceding the jump-off date. For example, if over the past year the
dollar has already appreciated by 5 percent against the euro, the Board
would lower the appreciation rate in the scenario from 15 percent to 10
percent.
---------------------------------------------------------------------------
\248\ This value is in line with the average standard deviation
of 4-quarter changes in the exchange rates of the four country
blocs, computed over the pre-COVID-19 historical sample.
---------------------------------------------------------------------------
e. Alternative Considerations
The Board considered a range of different approaches to derive the
guides for the international variables in the severely adverse
scenario. First, distinct instead of common guides for GDP and
inflation for each of the AFEs were explored. Following the methodology
explained in Section IX.F of this Supplementary Information, the trough
values for GDP during the 2007-2009 financial crisis were 6 percent
below baseline in both the euro area and the U.K., and 10 percent below
baseline in Japan. For inflation, the trough values in the 2007-2009
financial crisis were 3 percent below baseline in the euro area, 2
percent below baseline in the U.K., and 4.5 percent below baseline in
Japan. The Board decided against using region-specific guides for the
AFEs, as systematic differences in the guides across regions could
affect credit allocations. However, the issued guides still allow for
region-specific paths in the severely adverse scenario within the
specified ranges to reflect region-specific circumstances when
desirable. The Board may also use the specified ranges to raise or
lower the sensitivity of all regions at the same time in the severely
adverse scenario.
Second, the Board examined other global or regional economic
downturns of significance in addition to the 2007-2009 financial crisis
to refine its guides for the international variables of the adverse
scenario. There are only a few global recessions in recent history but
none of them--besides the 2007-2009 crisis--were driven by financial
factors. The COVID-19 recession of 2020 led to a sharper contraction in
global economic activity than the 2007-2009 financial crisis but did
not result in persistent financial stress. One distinct feature of the
2007-2009 financial crisis was the resilience of Developing Asia where
GDP dropped by only 3 percent relative to baseline. Only a decade prior
to the 2007-2009 financial crisis, several countries in Developing Asia
had experienced negative doubled-digit GDP growth rates as part of the
1997 Asian Financial Crisis. Based on the experience of the 1997 Asian
Financial Crisis, the Board considered a lower trough value for
Developing Asia. However, the Board decided against this approach for
several reasons. China was generally unaffected by the Asian Financial
Crisis and maintained its high GDP growth rate throughout the crisis,
significantly increasing the regional GDP growth rate during this
period despite the sharp declines experienced elsewhere in the region.
In addition, the countries that were most affected by the Asian
Financial Crisis changed to more robust economic policies--switching
from fixed/managed exchange rates to flexible inflation targeting and
from policies that implied large current account deficits to policies
that led to surpluses. Finally, the Board looked to the euro area debt
crisis for additional guidance. But since this crisis directly followed
the 2007-2009 financial crisis no additional insight for the design of
the severely adverse scenario was obtained that was not already
embedded in the analysis of the 2007-2009 financial crisis. In the
interest of transparency and simplicity, the Board decided to adopt the
findings derived solely from the 2007-2009 financial crisis but added
flexibility by allowing ranges for variables.
[[Page 51927]]
H. Global Market Shock
Design of the Global Market Shock
The global market shock component comprises a large set of
financial risk factors and associated hypothetical shocks to those risk
factors. The Board considers emerging and ongoing areas of financial
market vulnerabilities in the development of the global market shock
component, informed by financial stability reports, supervisory
information, and internal and external assessments of potential sources
of distress such as geopolitical, economic, and financial market
events. Financial risk factor shocks are calibrated based on assumed
time horizons that reflect several scenario design considerations. The
Board also considers liquidity characteristics of the different asset
classes that constitute certain risk factors. These liquidity horizons
approximate the variation in speed at which banks could reasonably
close out, or effectively hedge, the associated risk exposures in the
event of market stress.
The chosen risk factors of the global market shock scenario are
important to specifying how a stress scenario unfolds across financial
markets and capturing salient risks within the banking system. These
include, but are not limited to:
Public equity returns from key advanced economies and from
developing and emerging market economies, along with selected points
along term structures of equity option-implied volatilities;
Exchange rates of foreign currencies, along with selected
points along term structures of foreign exchange option-implied
volatilities;
Government yields at selected maturities (e.g., 10-year
U.S. Treasuries), swap rates, and other types of interest rates for key
advanced economies and from developing and emerging market economies;
Implied volatilities on interest rate options for selected
maturities and expiration dates, which are key inputs to the pricing of
interest rate derivatives;
Futures prices at various expiration dates for commodity
products such as energy, oil, metals, and agricultural products; and
Credit spreads or prices for selected credit-sensitive
products, including corporate bonds, credit default swaps (CDS),
securitized products, sovereign debt, and municipal bonds.
The global market shock is typically applied to positions held by
the firms on a given as-of date, reflecting a hypothetical
instantaneous ``shock'' to a large number of risk factors that
determine the mark-to-market values of trading positions. Additionally,
the global market shock in a given annual severely adverse scenario is
a standardized set of market shocks that apply to all of the firms with
significant trading activities. The selection of a single date, and a
single global market shock, has tended to enhance the operational
consistency and simplicity of the annual supervisory stress test, while
managing burden on reporting firms.
The Board is considering enhancements to the design of the global
market shock in the annual stress test to improve the stress test's
ability to capture the impact of severe economic stress in financial
markets. Alternative approaches to the global market shock could
include employing instantaneous shock events across multiple as-of
dates, rather than the current approach of selecting a single date for
an instantaneous shock event. Another approach could involve an annual
stress test that features multiple global market shock components on a
single as-of date, which would allow the Board to compare a given
firm's losses across a variety of types of shocks for a given set of
trading position. The set of losses generated by such multiple market
scenarios could be aggregated using a simple average, an average of the
two worst outcomes, or another technique. These alternatives could
enhance the dynamism of the annual stress tests and improve the Board's
ability to evaluate the impact of severe economic stress on trading
positions in a given annual stress test. However, these changes could
also increase the complexity of the tests, and affect their
predictability from year to year.
Question 44: What changes could the Board implement to improve the
general design of the global market shock? What, if any, alternative
approaches should the Board consider? For instance, should the Board
consider adjusting the global market shock so that shock events occur
on multiple dates within the as-of date window? Should the Board
consider testing more than one global market shock component in a given
annual stress test or on a particular date? If so, how should the Board
assess whether the current design, or alternative approaches,
contribute to outcomes that are overly volatile or insufficiently
representative of risks? If the Board should adopt these alternative
approaches, what information should the Board provide to the public
about how it will implement these alternatives, and should that
information be published as part of a revised Scenario Design Policy
Statement, codified as part of Regulation YY, the annual scenario
disclosure, or some other means?
Question 45: If the Board did adjust the global market shock to
consider multiple dates within the as-of date window or more than one
global market shock component in a given annual stress test or on a
particular date, what method should the Board use to aggregate these
values to calculate a firm's trading and counterparty losses in the
stress test and why? For example, should the Board consider averaging
the two instances of highest trading and counterparty losses? What
would be the advantages and disadvantages of these aggregation methods?
Question 46: The global market shock component and the largest
counterparty default component of the severely adverse scenario are
both based on the global market shock. Should the Board consider
removing one or both of these components from the severely adverse
scenario? If so, what alternative approaches should the Board consider
to account for trading and counterparty losses in the supervisory
stress test? For example, should trading and counterparty losses be
considered as part of the macroeconomic scenario as opposed to the
global market shock? What would be the advantages and disadvantages of
retaining these components or replacing them with alternative
approaches?
Question 47: Should the Board continue to include a global market
shock component in the severely adverse scenario? What would be the
advantages and disadvantages of including a market shock component in
the severely adverse scenario? If the Board determines to remove the
market shock component, are there additional changes that the Board
should implement that would mitigate any disadvantages from this
change?
Question 48: The global market shock component currently applies to
firms subject to Category I, II, and III standards that have aggregate
trading assets and liabilities of $50 billion or more, or trading
assets and liabilities equal to or greater than 10 percent of total
consolidated assets. What are the advantages and disadvantages of
applying the global market shock component to this group of firms?
Should this component apply to a different set or subset of firms? If
so, how should the Board determine which set or subset of firms should
be subject to the global market shock component?
Shock Values
The Board generates shock values for all exposures in the global
market shock template. Shock values represent the
[[Page 51928]]
magnitudes of changes to the financial risk factors and reflect the
severity of market stress that these risk factors experience in the
scenario. Table 24 provides an overview of the proposed shock
definitions by asset class.
[GRAPHIC] [TIFF OMITTED] TP18NO25.069
[[Page 51929]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.070
Liquidity Horizons
Financial risk factor shocks are calibrated based on assumed time
horizons that reflect several scenario design considerations. The
horizons are generally longer than the typical times needed to
liquidate exposures under normal conditions because they are designed
to capture the unpredictable liquidity conditions that prevail in times
of stress. The Board is proposing to add descriptions of the liquidity
horizons in the Scenario Design Policy Statement.
As discussed below, the Board is proposing horizons that are
intended to maintain consistency with the timeline for attributing
losses stemming from these risk factors. Specifically, losses
associated with the global market shock component are recognized in the
first quarter of the projection horizon, which indicates that these
shocks occur within a three-month period and thus implies a three-month
upper bound for calibrating the shocks.
The Board is proposing to amend its Scenario Design Policy
Statement to use shock liquidity horizons that are broadly consistent
with the proposed standards in the Basel Committee on Banking
Supervision's Fundamental Review of the Trading Book (FRTB).\249\ The
risk factors in the FRTB are similar to the ones in global market
shock. The horizons in the FRTB were determined by the Basel Committee
in consultations with the financial industry and represent the general
consensus of a broad range of regulation authorities and the industry.
Therefore, they are a reasonable benchmark for defining the shock
horizons used in the global market shock. The Board departed from the
FRTB slightly by specifying the same liquidity horizon to all risk
factors in the same asset class. This choice was consistent with the
Board's stress test principle of simplicity and facilitated a more
straightforward modeling framework for the global market shock.
---------------------------------------------------------------------------
\249\ Basel Committee on Banking Supervision, ``Calculation of
RWA for market risk,'' in The Basel Framework 675-970, https://www.bis.org/baselframework/BaselFramework.pdf. See also 88 FR 64028,
64138 (Sep. 18, 2023).
---------------------------------------------------------------------------
The liquidity horizons used in the global market shock component
are not perfectly matched with the FRTB liquidity horizons due to
granularity differences between the FRTB standards and the global
market shock template. The FRTB specifies horizons at a more granular
level, often using different horizons within each asset class. For
example, the FRTB specifies sovereign risk factor horizons by credit
rating. In contrast, the global market shock template specifies
sovereign shocks by country to capture country-specific risks not
reflected by credit ratings. Moreover, the Board uses the same
liquidity horizon for all risk factors within each asset class, whereas
the FRTB allows for different horizons within asset classes. Given
these differences, the global market shock scenario aims at aligning
with the horizons specified by the FRTB by using a weighted average of
the FRTB horizons within each asset class. The weights are determined
using aggregate firm exposures over past submission quarters. For
example, FRTB horizons for equity risk factors vary between 10 and 60
business days, and the global market shock horizon for this asset class
would be four weeks. Because the Board imposes an upper bound on global
market shock horizons of one quarter, there are cases where the range
of FRTB horizons would be longer than the global market shock horizon.
For example, FRTB horizons for corporate credit risk factors vary
between 60 and 120 business days, but the Board uses a horizon of three
months for corporate credit. See Table 25.
[[Page 51930]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.071
Question 49: What are the advantages and disadvantages of the
Board's proposed liquidity horizons? What, if any, additional or
alternative liquidity time horizons should the Board consider?
Global Market Shock Simplification
As discussed in Section II.B of this Supplementary Information, the
global market shock specifies hypothetical shocks to a standard set of
risk factors. Currently, the global market shock discloses more than
20,000 risk factors that reflect sudden market distress and heightened
uncertainty. Statistical models are used to generate a subset of risk
factors out of these 20,000 risk factors with the remaining ones
generated by simple mapping. However, this latter category includes
many risk factors that are often not material (for example, certain
commodity shocks). These low-materiality exposures do not necessarily
enhance the risk capture of the global market shock component.
To address these issues and simplify the global market shock
component, the Board is proposing to substantially reduce the number of
disclosed risk factors. Specifically, this would reduce the number of
disclosed risk factors to approximately 2,300 shocks, determined based
on their relevance for developing a global market shock scenario
narrative, the materiality of the risk factor, data quality, and
consistency across asset classes.
Under this approach, the Board would also review consistency across
asset classes. In this regard, where possible, the Board would generate
shocks to the same set of countries, regions, and tenor points across
different asset classes. Such consistency would simplify shock
comparison across different asset classes and improve public
understanding of the global market shock component. Additionally, the
Board is proposing to remove the inclusion of shocks to the values of
private equity positions in section 3.2(b)(viii) of the Scenario Design
Policy Statement, because private equity exposures are now stressed
using the severely adverse macroeconomic scenario.
Question 50: What are the advantages and disadvantages of
simplifying the global market shock's specification of risk factor
shocks? What are the advantages and disadvantages of removing shocks
related to the value of private equity positions from the global market
shock component?
X. Economic Analysis
Introduction
In December 2024, the Board announced that it would seek public
[[Page 51931]]
comment on significant changes to improve the transparency of its
supervisory stress test and to reduce the volatility of resulting
capital buffer requirements.\250\ As discussed in Section II.E of this
Supplementary Information, this proposal would improve the transparency
and public accountability surrounding the stress test models and
scenarios, as well as make certain changes to their underlying
methodologies, which could provide meaningful benefits to the public as
discussed below. This section provides economic analysis of the
enhanced disclosure of the supervisory stress test framework.
---------------------------------------------------------------------------
\250\ See Board, Press Release (Dec. 23, 2024), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20241223a.htm.
In February 2025, the Board stated that it would begin the public
comment process on comprehensive changes to the supervisory stress
test in 2025. See Board, Press Release (Feb. 5, 2025), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20250205a.htm.
---------------------------------------------------------------------------
The Board's supervisory stress test has historically operated with
some disclosure regarding the stress test models and scenarios used,
with an increase in public information provided beginning in 2019, as
discussed in Section II.B of this Supplementary Information. The
comprehensive model documentation that the Board is publishing on its
website, as well as the proposed enhanced disclosure process for the
models and scenarios, provides several benefits, including improved
credibility of the stress test, improvement in feedback regarding the
modeling process, better informed investors, and improved market
discipline. However, the enhanced disclosure comes with costs as well,
including reduced model dynamism, and increased systemic reliance on a
single model, that is, ``model monoculture.''
Baseline
The economic analysis uses the current stress testing framework,
including the current disclosure regime, as the baseline. Throughout
the analysis, the Board assesses the economic impact of the proposal by
comparing outcomes under the proposal to the outcomes estimated under
the baseline.
Proposed Policy Changes
With this proposal, the Board is providing a comprehensive
description of the modeling framework used to conduct the supervisory
stress test: the equations, variables and parameters of each model used
to estimate the projections that, when aggregated, produce the results
of the supervisory stress test. This proposal would also codify an
enhanced disclosure process under which the Board would annually
publish the stress test models, invite public comment on any material
changes to the models, and seek comment on the annual stress test
scenarios. This represents a significant increase in disclosure
relative to present, as current stress test disclosures are more
limited, for example, current disclosures cover the structure of the
stress testing model framework and key variables, along with
hypothetical portfolio loss rates for select corporate and retail loss
models.
In addition, this proposal would change the stress test jump-off
date and the GMS as-of date, as described in Sections VI.A and VI.B of
this Supplementary Information. These changes would adjust the stress
testing schedule to accommodate the public comment process and mitigate
risks that the enhanced disclosure provided under this proposal would
undermine the goals of supervisory stress testing.
Section VIII.A of this Supplementary Information summarizes
proposed changes to the stress testing models from the 2025 to the 2026
supervisory stress test, which would inform the Board's determination
of firms' stress capital buffer requirements.\251\ Section VIII.B
provides an analysis of the potential effects of these proposed model
changes.
---------------------------------------------------------------------------
\251\ For a more detailed discussion of the proposed model
changes, see https://www.federalreserve.gov/supervisionreg/dfa-stress-tests-2026.htm.
---------------------------------------------------------------------------
Finally, Sections V and IX of this Supplementary Information
describe proposed changes to the Board's Stress Testing Policy
Statement and Scenario Design Policy Statement. The proposed changes to
the Board's Stress Testing Policy Statement and Scenario Design Policy
Statement are intended to express the Board's expectations for how the
Board conducts the annual supervisory stress test and designs annual
scenarios for the annual supervisory stress test. These changes provide
additional transparency, public accountability, and predictability
without creating binding legal obligations or economic impact.
Analysis of Benefits and Costs of Enhanced Model Disclosure
Benefits
a. Improved Credibility of the Stress Test
The supervisory stress test has material safety and soundness
benefits and these benefits are likely more sustainable when the
Board's stress testing program operates with high levels of
accountability and credibility. Disclosing comprehensive model
documentation to the public ensures that all institutions and
stakeholders have equal access to the supervisory methodology, which
could improve accountability in supervisory decision-making, promote
fairness, and reinforce trust in the stress testing process. Publicly
disclosing the stress test models and scenarios also enhances trust in
the stress testing process,\252\ as stakeholders may be able to better
assess the soundness of models and their alignment with best
practices.\253\ As a result, firms may understand better where there
are discrepancies between their own internal stress testing models and
the supervisory stress testing models, and consequently they may be
better positioned to communicate specific concerns with supervisors.
With greater transparency and public accountability, stakeholders may
be more confident that the supervisory stress test results do not
reflect the desires of firms or supervisors to obtain a specific
outcome.\254\ While the Board has previously released enhanced
disclosures of the stress test models, such as portfolio-level average
loss rates, macro-to-loss linkages, and risk drivers, the comprehensive
model documentation disclosed in connection with this proposal better
illustrates how supervisors incorporate model refinements and emerging
risks, which could further improve credibility over time.
---------------------------------------------------------------------------
\252\ For an overview of studies on the impact of government
transparency, which generally suggest a mixed-to-positive impact on
trust, see M. Cucciniello et al., 25 Years of Transparency Research:
Evidence and Future Directions, 77 Public Admin. Rev. 32-44 (2016).
\253\ See I. Goldstein & Y. Leitner, ``Stress test disclosure:
theory, practice, and new perspectives,'' Handbook of Financial
Stress Testing 208-223 (2022).
\254\ See I. Goldstein & H. Sapra, Should Banks' Stress Test
Results Be Disclosed? An Analysis of the Costs and Benefits, 8
Foundations and Trends in Finance 1-54 (2013); F. Niepmann & V.
Stebunovs, Modeling our stress away, 158 Journal of Banking &
Finance 107042 (2024). When regulators are more constrained in their
ability to make the models more or less severe, this could alleviate
inefficient strategic interactions between supervisors and banks,
referred to as ``policy traps.'' J. Shapiro & J. Zeng, Stress
Testing and Bank Lending, 37 Rev. of Fin. Studies 1265-1314 (2024).
---------------------------------------------------------------------------
In addition, as described in Section VI.B of this Supplementary
Information, this proposal would extend the date selection range of GMS
as-of date from five months (between October 1 of the previous year and
March 1 of a given year) to a full year (between October 1 of two years
prior to a given stress test cycle to October 1 of the year prior to a
given stress test cycle). Thus, the GMS
[[Page 51932]]
could be applied to market risk positions held by the firms on any
selected date within the full year instead of the current five months.
This change could reduce firm's risk gaming activities such as ``window
dressing'' for firms subject to the GMS. Therefore, the resulting
improved risk capture would further enhance the credibility of the
stress test results.
b. Improved Model Feedback
The Board's supervisory stress test models consist of equations,
parameters, and assumptions that translate hypothetical macroeconomic
shocks under designed stress scenarios into loss estimates across asset
classes, income streams, and capital ratios. Despite their complexity,
the supervisory stress test models and stress scenarios, like all
theoretical models, remain simplified representations of reality. As
such, they benefit from feedback and refinement. Public disclosure of
models and scenarios should provide academics, industry professionals,
and the broader risk community with the information to provide more
effective feedback.\255\ For example, in past supervisory stress
testing cycles, stakeholders have raised concerns about loss rates on
certain asset classes. Over time, such feedback could help to refine
and improve the models and scenarios as they could be updated to
mitigate concerns, as appropriate. In this sense, the proposal's
enhanced disclosure could facilitate stakeholders' feedback, ultimately
leading to better modelling performance and further enhancing the
credibility of the supervisory stress testing process.\256\
---------------------------------------------------------------------------
\255\ See I. Goldstein & Y. Leitner, ``Stress test disclosure:
theory, practice, and new perspectives,'' Handbook of Financial
Stress Testing 208-223 (2022); B. Hirtle, ``Structural and Cyclical
Macroprudential Objectives in Supervisory Stress Testing,'' Remarks
at ``The Effects of Post-Crisis Banking Reforms'' conference (Jun.
22, 2018).
\256\ As an example of feedback on the Pre-provision Net Revenue
Model under the current disclosure regime, see M. Xiao, ``What
Goldman's appeal victory means for Fed stress tests,'' Risk.net
(Oct. 30, 2024), https://www.risk.net/risk-management/7960102/what-goldmans-appeal-victory-means-for-fed-stress-tests.
---------------------------------------------------------------------------
c. Improved Ability To Evaluate Business Plans
Comprehensive disclosure of the stress test models also may help
firms better understand how supervisors assess losses under severely
stressed hypothetical scenarios. This may allow firms to more
accurately predict their required capital ratios, reducing capital
planning uncertainty \257\ and possibly increasing firms' willingness
to lend.\258\ Reduced capital requirement uncertainty could help firms
better plan their future business decisions.
---------------------------------------------------------------------------
\257\ See G. Gallardo et al., Stress testing convergence, 9 J.
of Risk Mgmt. in Fin. Institutions 32-45 (2016); B. Hirtle,
``Structural and Cyclical Macroprudential Objectives in Supervisory
Stress Testing,'' Remarks at ``The Effects of Post-Crisis Banking
Reforms'' conference (Jun. 22, 2018).
\258\ For evidence on the impact of regulatory uncertainty on
lending, see S. Gissler et al., Lending on hold: regulatory
uncertainty and bank lending standards, 81 J. of Monetary Econ. 89-
101 (2016).
---------------------------------------------------------------------------
d. Better Informed Investors and Improved Market Discipline
Research suggests that investors use stress test results to assess
firms' resilience. Indeed, disclosures of results from the stress test
tend to affect firms' stock prices and CDS spreads.\259\ Through such
financial market signals, investors may help discipline firms' risk
taking.\260\ This ``market discipline'' may constrain risk taking and
incentivize firms to strengthen capital positions.\261\ The
comprehensive disclosure of the supervisory stress testing models may
allow investors to make better informed decisions, potentially
improving the effectiveness of market discipline.
---------------------------------------------------------------------------
\259\ See C. Sahin et al., Banking stress test effects on
returns and risks, 117 J. of Banking & Fin. 105843 (2020); L.
Guerrieri & M. Modugno, The information content of stress test
announcements, 160 J. of Banking & Fin. 107087 (2024); M. Flannery
et al., Evaluating the information in the federal reserve stress
tests, 29 J. of Fin. Intermediation 1-18 (2017); G. Petrella & A.
Resti, Supervisors as information producers: Do stress tests reduce
bank opaqueness?, 37 J. of Banking & Fin. 5406-20 (2013); D. Morgan
et al., The Information Value of the Stress Test, 46 J. of Money,
Credit & Banking 1479-1500 (2014); C. Alves et al., Do stress tests
matter? A study on the impact of the disclosure of stress test
results on European financial stocks and CDS markets, 47 Applied
Economics 1213-29 (2015); O. Georgescu et al., Do stress tests
matter? European Central Bank Working Paper 2054 (2017), https://www.ecb.europa.eu/pub/pdf/scpwps/ecb.wp2054.sv.pdf; L. Ahnert et
al., Regulatory stress testing and bank performance, 26 European
Fin. Mgmt 1449-88 (2020); L. Gu, K. Wang., & J. Wu, ``The asset
market effects of bank stress-test disclosures,'' in Stress Testing
(2nd Edition): Approaches, Methods and Applications (2019).
\260\ See supra note 33.
\261\ For evidence on the impact of stress test disclosure on
bank risk-taking, see supra note . However, the impact on risk-
taking is attributed more to supervisory scrutiny than disclosure in
other research. See C. Kok et al., The disciplining effect of
supervisory scrutiny in the EU-wide stress test, 53 J. of Fin.
Intermediation 101015 (2023).
---------------------------------------------------------------------------
Costs
a. Reduced Dynamism
As discussed above, models are necessarily a simplified version of
reality. As forecasting methodologies evolve or conditions in the
economy and the financial system change, the existing models may no
longer adequately capture risks. For this reason, an effective stress
test must be able to adapt. Under this proposal, material changes to
the stress testing models would be published for comment, and the Board
would be required to respond to such comment, before implementing the
material model changes in the supervisory stress test. This process
would increase the resources needed to develop, propose, and implement
material model changes, particularly to the extent that any changes are
complex, present many alternatives, or affect firms' ability to
distribute capital. As a result, the use of new models or model changes
that explore risks that are less established may pose a high resource
burden under the proposed enhanced disclosure regime, potentially
limiting the supervisory stress test to simpler, less controversial,
and more familiar approaches.\262\ Tests of new risk dimensions or
emerging threats may take significantly more time to implement. With
less dynamism, the supervisory stress test may fail to capture new
risks and could produce an increasingly stale view of how firms would
be likely to perform under stressed conditions.\263\ In addition, as
described in Section VI.A of this Supplementary Information, this
proposal would change the jump-off date of the supervisory and company-
run stress tests from December 31 to September 30, to allow the Board
to publish the annually disclosed stress test information for comment
after the jump-off date of the stress test and to prevent firms from
adjusting their exposures based on the published information. As a
result, the tested balance sheets would be older by one quarter, which
would add additional staleness to the stress test and stress test
results, because firm balance sheets as well as economic conditions
could change substantially within a quarter.
---------------------------------------------------------------------------
\262\ See M. Flannery, Transparency and model evolution in
stress testing, SSRN Working Paper (2019), http://dx.doi.org/10.2139/ssrn.3431679. Even the current approach to stress testing
may not allow for the optimal level of dynamism or macroprudential
considerations. See D. Tarullo, Reconsidering the regulatory uses of
stress testing, Hutchins Center Working Paper 92 (2024), https://www.brookings.edu/wp-content/uploads/2024/05/WP92_Tarullo-stress-testing.pdf; W. Bassett & D. Rappoport, ``Enhancing stress tests by
adding macroprudential elements,'' in Handbook of Financial Stress
Testing 455-83 (2022).
\263\ For an example of the reduced utility of a stale stress
model, see W. Frame et al., The failure of supervisory stress
testing: Fannie Mae, Freddie Mac, and OFHEO, Federal Reserve Bank of
Boston Working Paper 15-4 (2015), https://www.bostonfed.org/-/media/Documents/Workingpapers/PDF/wp1504.pdf.
---------------------------------------------------------------------------
b. Reduced Risk Sensitivity and Overreliance on a Single Model
Framework
Supervisory stress testing results are important inputs to the
capital requirements associated with firms'
[[Page 51933]]
banking activities. With comprehensive model disclosure likely reducing
the uncertainty of supervisory stress test results, firms' estimates of
future regulatory capital requirements could rely more on the Board's
stress test models and less on their own internal stress testing models
or internal risk management tools, both of which may be less useful
than before for managing regulatory capital.\264\ To the extent that
firms' own internal stress testing models or risk management tools
provide additional information about risk, the expected capital
requirements could become less risk-sensitive as a result and it may
reduce firms' incentives to independently measure and manage their
vulnerabilities.
---------------------------------------------------------------------------
\264\ T. Schuermann, ``The Fed's Stress Tests Add Risk to the
Financial System,'' W.S.J. (Mar. 19, 2013), https://www.wsj.com/articles/SB10001424127887324532004578362543899602754?gaa_at=eafs&gaa_n=ASWzDAgXgiqB0fwSIwZXAJZF5iLfwSHPFItS1v9pIwVWyP1FFRG2TyjbJ153&gaa_ts=68e66a22&gaa_sig=QXBddH1PbBwcemmdRad58NRIsIlftxSu-CxAv7UOygRlCujSJqcMQF1rlakd0GGI4045knXKHn-H06BNwTBP-Q%3D%3D.
---------------------------------------------------------------------------
Disclosure could also enable firms to more easily optimize their
exposures to minimize capital requirements in the supervisory stress
test, which could allow vulnerabilities to build up where risks are not
well or fully accounted for by standardized supervisory models.
Reliance on the supervisory stress testing models could extend
further if disclosure results in firms increasing the similarity of
their own stress models to the stress test models.\265\ Increased
reliance of all stress tested firms on a single model, known as ``model
monoculture,'' or delaying material model changes while risks build up
in areas that are treated benignly in the stress test would pose risks,
as firms may face a greater incentive to shift business activities
towards these areas to reduce their capital requirements.\266\ The
resulting convergence of risk taking could increase the vulnerability
of the banking system, particularly to those risks that are under-
reflected by the supervisory stress testing models.\267\
---------------------------------------------------------------------------
\265\ Of course, as noted above, there is benefit to these
changes to the extent that they are adopted to improve the ability
of firms' models to capture risk.
\266\ Relatedly, banks may have a stronger incentive to
temporarily curtail those risk exposures treated adversely by the
stress testing models, i.e., to ``window dress.'' See P. Alexander,
``How banks game stress tests: the `shocking' truth,'' Risk.net
(Sep. 30, 2019), https://www.risk.net/regulation/6989811/how-banks-game-stress-tests-the-shocking-truth; M. Cornett et al., An
Examination of Bank Behavior around Federal Reserve Stress Tests, 41
Journal of Financial Intermediation 100789 (2020).
\267\ See Y. Leitner & B. Williams, Model Secrecy and Stress
Tests, 78 J. of Fin. 1055-95 (2023); K. Rhee & K. Dogra, Stress
Tests and Model Monoculture, 152 J. of Fin. Econ. 103760 (2024); B.
Hirtle, ``Structural and Cyclical Macroprudential Objectives in
Supervisory Stress Testing,'' Remarks at ``The Effects of Post-
Crisis Banking Reforms'' conference (Jun. 22, 2018), https://www.newyorkfed.org/newsevents/speeches/2018/hir180622; Flannery,
M.J., 2019. Transparency and Model Evolution in Stress Testing.
SSRN, Working Paper, http://dx.doi.org/10.2139/ssrn.3431679; B.
Bernanke, ``Stress testing banks: what have we learned?'' Remarks at
``Maintaining Financial Stability: Holding a Tiger by the Tail''
conference (Apr. 8, 2013), https://www.bis.org/review/r130409c.pdf;
I. Goldstein & Y. Leitner, ``Stress test disclosure: theory,
practice, and new perspectives,'' Handbook of Financial Stress
Testing 208-223 (2022); F. Br[auml]uning & J. Fillat, Stress Testing
Effects on Portfolio Similarities Among Large US Banks, Federal
Reserve Bank of Boston Policy Perspectives, Paper 19-1 (2019),
https://www.bostonfed.org/-/media/Documents/Workingpapers/PDF/2019/cpp1901.pdf.
---------------------------------------------------------------------------
Conclusion
As discussed in the introduction to Section X.D of this
Supplementary Information, the Board's supervisory stress test has
historically operated with partial disclosure regarding the stress test
models used. The comprehensive model documentation published in
connection with this proposal, as well as the proposed enhanced
disclosure process, provides several benefits that outweigh the costs
of the proposal.
Taken together, the Board assessed that the benefits of the
proposal justify the costs.
Question 51: What, if any, additional material costs or benefits
should the Board consider, in addition to those discussed in the
proposal?
Question 52: What alternatives that achieve the objectives of the
proposal should the Board evaluate? Please provide specific suggestions
and rationales for any proposed alternatives, including how they might
address potential unintended consequences or better achieve the
proposal's goals.
XI. Administrative Law Matters
A. Paperwork Reduction Act Analysis
In accordance with the requirements of the Paperwork Reduction Act
(PRA) of 1995 (44 U.S.C. 3501-3521), the Board may not conduct or
sponsor, and the respondent is not required to respond to, an
information collection unless it displays a currently valid Office of
Management and Budget (OMB) control number. The Board reviewed the
information collections related to the proposed rule under the
authority delegated to the Board by OMB.
The proposed rule would not create any information collections
subject to the PRA; however, the Board is proposing to revise the FR Y-
14A/Q/M to reduce regulatory reporting burden by retiring items and
removing supporting documentation requirements that are no longer
needed to conduct the supervisory stress test. Additionally, the Board
is proposing to collect additional information to support the proposed
supervisory stress test models.
The Board invites public comment on the following information
collection:
(a) Whether the collection of information is necessary for the
proper performance of the Board's functions, including whether the
information has practical utility;
(b) The accuracy of the Board's estimate of the burden of the
proposed information collection, including the validity of the
methodology and assumptions used;
(c) Ways to enhance the quality, utility, and clarity of the
information to be collected;
(d) Ways to minimize the burden of the information collection on
respondents, including through the use of automated collection
techniques or other forms of information technology; and
(e) Estimates of capital or start-up costs and costs of operation,
maintenance, and purchase of services to provide information.
Proposal Under OMB Delegated Authority To Extend for Three Years, With
Revision, the Following Information Collection
Collection title: Capital Assessments and Stress Testing Reports.
Collection identifier: FR Y-14A/Q/M.
OMB control number: 7100-0341.
General description of collection: This family of information
collections is composed of the following three reports:
The annual FR Y-14A collects quantitative projections of
balance sheet, income, losses, and capital across a range of
macroeconomic scenarios and qualitative information on methodologies
used to develop internal projections of capital across scenarios.\268\
---------------------------------------------------------------------------
\268\ In certain circumstances, a firm may be required to re-
submit its capital plan. See 12 CFR 225.8(e)(4); 12 CFR
238.170(e)(4). Firms that must re-submit their capital plan
generally also must provide a revised FR Y-14A in connection with
their resubmission.
---------------------------------------------------------------------------
The quarterly FR Y-14Q collects granular data on various
asset classes, including loans, securities, trading assets, and pre-
provision net revenue for the reporting period.
The monthly FR Y-14M is comprised of three retail
portfolio- and loan-level schedules, and one detailed address-matching
schedule to supplement two of the portfolio- and loan-level schedules.
[[Page 51934]]
The data collected through the FR Y-14A/Q/M provide the Board with
the information needed to help ensure that large firms have strong,
firm-wide risk measurement and management processes supporting their
internal assessments of capital adequacy and that their capital
resources are sufficient, given their business focus, activities, and
resulting risk exposures. The data within the reports are used in
connection with setting firms' stress capital buffer requirements. The
data are also used to support other Board supervisory efforts aimed at
enhancing the continued viability of large firms, including continuous
monitoring of firms' planning and management of liquidity and funding
resources, as well as regular assessments of credit risk, market risk,
and operational risk, and associated risk management practices.
Information gathered in this collection is also used in the supervision
and regulation of respondent financial institutions. Respondent firms
are currently required to complete and submit up to 17 filings each
year: one annual FR Y-14A filing, four quarterly FR Y-14Q filings, and
12 monthly FR Y-14M filings.\269\ Compliance with the information
collection is mandatory.
---------------------------------------------------------------------------
\269\ Holding companies that do not meet the materiality
thresholds described in the instructions for the FR Y-14M are not
required to file that report. This results in some holding companies
submitting fewer than 17 filings each year.
---------------------------------------------------------------------------
Current Actions: The proposal would modify the FR Y-14A/Q/M to
remove supporting documentation requirements, schedules, and data items
that are no longer needed to conduct the supervisory stress test. The
proposal would also make other revisions necessary to facilitate the
stress test modeling decisions. All proposed revisions would be
effective for the September 30, 2026, report date.
Supporting Documentation
a. FR Y-14A
The FR Y-14A collects detailed data on firms' quantitative
projections of assets, liabilities, income, losses, and capital across
a range of macroeconomic scenarios. Firms are also required to provide
qualitative information on the methodologies used to develop their
projections and any other analysis that supports or contributes to
these projections. This qualitative supporting documentation helps
supervisors assess the accuracy and comprehensiveness of the
projections included in firms' FR Y-14A submissions. This information
was previously critical to assess the data systems and modeling
methodologies that firms used to report the FR Y-14A. However, as these
systems and frameworks have matured, much of the supporting
documentation has become outdated or is not needed by supervisors to
make such assessments. To ensure that the FR Y-14A requirements do not
capture information that is no longer needed and to reduce reporting
burden, the Board is proposing to remove Appendix A ``Supporting
Documentation'' from the FR Y-14A. However, supervisors may request
similar information to what is currently required from Appendix A from
firms through supervisory channels, as deemed appropriate and on a
targeted basis, in support of the annual capital plan review. Firms
would only be expected to provide information that supervisors request
each cycle. The proposed removal of the FR Y-14A supporting
documentation reporting requirement would not impact any other capital
planning expectations.
b. FR Y-14Q
FR Y-14Q, Schedule L (Counterparty) collects data on firms'
counterparty credit risk, including derivative and securities financing
transaction exposures. Applicable firms are required to report two
versions of Schedule L: an ``unstressed'' version under the actual
economic conditions on the reported date, and a ``stressed'' version
under the hypothetical stress scenarios used in the supervisory and
company-run stress tests. To support firms' estimates of credit
valuation adjustment and counterparty losses under the stress
scenarios, the FR Y-14Q requires that firms provide detailed
descriptions of the methodologies used to generate values for the
``stressed'' version. As for the FR Y-14A, this information was
previously important in understanding firms' counterparty submissions
but is no longer required for supervisors to assess Schedule L data.
However, the Board has identified supporting information that is
relevant to understanding a firm's estimated credit valuation
adjustment and largest counterparty default losses. Therefore, to
streamline Schedule L and reduce reporting burden, the Board is
proposing replacing the existing Schedule L supporting documentation
with this more limited set of questions. These questions would concern
excluded counterparties, estimation assumptions, drivers of changes in
credit valuation adjustment, and other related topics.
Similarly, qualitative information is needed to assess firms'
trading mark-to-market projections under the global market shock. As a
firm's projections are directly connected to the exposures reported on
FR Y-14Q, Schedule F (Trading), the Board is proposing to introduce
supporting documentation for Schedule F that includes five questions
related to a firm's trading projections and Schedule F submissions.
Together with the Schedule L supporting documentation, this would
ensure that supervisors have the necessary information to assess a
firm's projections under the global market shock.
Home Equity Data Collection
FR Y-14M, Schedule B.1 (Home Equity Loan-Level Table) collects
loan-level data on firms' HELOCs. These data are used in support of
stress test modeling and monitoring of firms' home equity portfolios.
The Board has identified several items on Schedule B.1 that are not
needed to assess a home equity loan or HELOC's risk characteristics or
are captured elsewhere on Schedule B.1. Therefore, to maintain
appropriate risk coverage and reduce reporting burden, the Board is
proposing to retire the following fields from Schedule B.1.
Item 18 (Number of Units)
Item 31 (ARM Periodic Rate Cap)
Item 32 (ARM Periodic Rate Floor)
Item 38 (Bankruptcy Flag)
Item 48 (Foreclosure Referral Date)
Item 51 (Pre-Payment Penalty Term)
Item 58 (Interest Rate Frozen)
Item 59 (Principal Deferred)
Item 62 (First Mortgage Serviced in House)
Item 72 (Term Modification)
Item 73 (Principal Write-Down)
Item 74 (Line Re-Age)
Item 75 (Loan Extension)
Item 86 (Accrual Status)
Item 87 (Foreclosure Suspended)
Item 88 (Property Valuation Method at Origination)
Item 92 (Third Party Sale Flag)
Item 107 (Entity Type)
Collection of Mailing Address Information
FR Y-14M, Schedule C (Address Matching) collects address
information on each loan reported on FR Y-14M, Schedule A (First Lien)
or Schedule B (Home Equity). This collection includes both property and
mailing address data used in support of the supervisory stress test
models. However, the Board has determined that the mailing address
items are no longer needed for stress testing or supervisory purposes.
Therefore, the Board is proposing to remove item 6 (Mailing Stress
Address), item 7 (Mailing City), item 8 (Mailing State), and item 9
(Mailing Zip Code) from Schedule C.
[[Page 51935]]
Unpaid Principal Balance
FR Y-14M, Schedule B.1 item 95 (Unpaid Principal Balance (Net))
collects information on the current net unpaid principal balance of a
home equity line of credit. The instructions provide a definition for
calculating net unpaid principal balance and note that this value
should equal the book value on regulatory filings. However, reporting
of unpaid principal balance can vary across regulatory reporting,
including by considering loan premiums, which item 95 does not include.
To address this inconsistency, the Board is proposing to remove this
language from the instructions for item 95.
Private Equity
Beginning with the 2025 supervisory stress test, the Board
calculated losses on private equity exposures under the macroeconomic
scenario over a nine-quarter projection horizon as opposed to under the
GMS, which would have considered the impact only in the first quarter
of the projection horizon. As described in the Board's 2025 Supervisory
Stress Test Methodology disclosure,\270\ the new treatment better
aligns with the characteristics of private equity exposures, which are
principally long-term investments that are managed as banking book
positions. To better capture private equity data in a manner that
aligns with this new treatment, the Board is proposing several
revisions to FR Y-14Q, Schedule F (Trading).
---------------------------------------------------------------------------
\270\ See Board, 2025 Stress Test Scenarios (Feb. 2025), https://www.federalreserve.gov/publications/files/2025-stress-test-scenarios-20250205.pdf.
---------------------------------------------------------------------------
First, the Board is proposing to move the fourth quarter as-of date
for reported private equity exposures to December 31 of a given year,
as opposed to the as-of date of the GMS. Schedule F is reported on a
quarterly basis. However, to gather data necessary to subject firms to
the GMS, firms are required to report Schedule F as of the GMS as-of
date and not as of December 31 for the fourth quarter submission.
Therefore, the Board is proposing to require private equity exposures
to be reported as of December 31, as private equity exposures are no
longer stressed under the GMS.
Second, the Board is proposing to revise Schedule F such that
private equity carry values are reported net of embedded goodwill or
investments in the capital of unconsolidated financial institutions
that are deducted from common equity tier 1 capital. The Board's
capital rule provides that certain amounts of goodwill and investments
in the capital of unconsolidated financial institutions be deducted
from CET1 capital,\271\ and the carry value of private equity exposures
reported on Schedule F can be affected by these deducted amounts.
---------------------------------------------------------------------------
\271\ See 12 CFR 217.22.
---------------------------------------------------------------------------
Firms subject to Category I through III standards are required to
report these deduction items on FR Y-14A, Schedule A.1.d (Capital). To
ensure that deductions are not double-counted when calculating trading
and counterparty losses, firms may report an adjusted starting value
for these items to reflect the impact of the global market shock.
However, as currently reported, a portion of these amounts may be
attributable to private equity. Therefore, the Board is proposing
revising Schedule F to require firms to exclude the amounts
attributable to private equity from the carry value. This revision
would ensure that losses are not assigned to balances that have been
deducted from capital when calculating private equity losses.
Third, the Board is proposing to require hedges on private equity
exposures to be separately reported on Schedule F. Hedges on private
equity exposures are currently reported on Schedule F but are not
segmented from other hedges on trading exposures. Given that private
equity exposures are no longer stressed as part of the GMS, the Board
is proposing to require hedges on private equity exposures to be
reported separately so that they can be considered as part of the
macroeconomic scenario.
Lastly, the Board is proposing to implement a new materiality
threshold for the reporting of Schedule F.24 (Private Equity).
Currently, Schedule F.24 is reported only by firms subject to Category
I through III standards with substantial trading operations, which is
defined as having, on average for four quarters, aggregate trading
assets and liabilities of $50 billion or more, or aggregate trading
assets and liabilities equal to 10 percent or more of total
consolidated assets. However, private equity exposures are primarily
banking book positions for which the FR Y-14 uses a separate reporting
threshold. For firms subject to Category IV standards, material
portfolios for banking book positions are defined as those with asset
balances greater than $5 billion or with asset balances greater than
ten percent of tier 1 capital on average for the four quarters
preceding the reporting period. For firms subject to Category I through
III standards, material portfolios for banking book positions are
defined as those with asset balances greater than $5 billion or asset
balances greater than five percent of tier 1 capital on average for the
four quarters preceding the reporting period.
To align the materiality threshold for private equity with other
banking book schedules, the Board is proposing to revise the FR Y-14Q
instructions to apply the materiality threshold to Schedule F.24 that
is currently applied to the banking book schedules. Additionally, since
a firm subject to Category IV standards could have its private equity
losses contribute to its supervisory stress test results, the Board
also proposes to require a firm subject to Category IV standards to
submit Schedule F.24 if it meets the proposed materiality threshold.
Similarly, consistent with reporting expectations for other banking
book positions, the Board is proposing to update FR Y-14Q, Schedule K
(Supplemental) such that firms report the carrying value of funded and
unfunded private equity exposures that do not meet the materiality
threshold for Schedule F.24 reporting. These revisions would ensure
consistent reporting and treatment of private equity in the supervisory
stress test.
Additionally, the Board is proposing a revision to FR Y-14A,
Schedule A.4 (Trading) which captures trading profit and loss
projections under the global market shock. As private equity shocks are
no longer included in the global market shock, items related to private
equity are no longer needed to capture trading profit and loss
projections. Therefore, the Board is proposing to remove item 15
(``Private Equity''), item 15A (``Private Equity: Funded''), item 15B
(``Private Equity: Unfunded''), item 15C (``Private Equity: Other'')
from Schedule A.4.
Other Hedges
Currently, the FR Y-14Q captures certain types of hedges, including
hedges on accrual loans and loans held under the fair value option and
certain designated accounting hedges on securities, but is not
comprehensive, which limits the ability of the supervisory stress test
to account for these positions. For example, FR Y-14Q, Schedule B
(Securities) does not provide sufficient information to independently
revalue the hedging instrument. Additionally, interest rate risk hedges
that are used to mitigate risk on instruments other than securities
from changes in interest rates are not captured by the FR Y-14Q.
Schedule B was designed to capture basic information on traditional
hedges on securities and does not consistently and comprehensively
capture portfolio layer
[[Page 51936]]
method or interest rate risk hedges for valuation purposes.
Separately, fair value option hedges are positions that are used to
hedge loan assets that are held-for-sale or held under fair value
option accounting, and do not meet the definition of trading assets or
liabilities. This includes synthetic securitizations, a form of loss
mitigation in which a firm partially transfers credit risk on specific
portfolios to outside investors through credit derivatives or
guarantees. Fair value option hedges are currently reported as a
separate instance of Schedule F. However, Schedule F is subject to a
materiality threshold, so fair value option hedges are not reported
comprehensively by all relevant firms on the FR Y-14Q.
To improve the risk capture of the supervisory stress test by
incorporating the effects of additional hedges, the Board is proposing
to revise FR Y-14Q, Schedule B.2 (Investment Securities with Designated
Accounting Hedges) to capture all qualified accounting hedges,
including portfolio layer method and all designated accounting hedges.
Additionally, the Board is proposing to implement FR Y-14Q, Schedule
B.3 to more comprehensively map hedging relationships. Similarly, the
Board is proposing to revise Schedule F to capture data on hedges from
any firms with reportable hedges.
Question 53: Would the new fields proposed in FR Y-14Q, Schedule
B.2 or B.3 prove burdensome to report for firms?
Question 54: Do the new fields proposed in FR Y-14Q, Schedule B.2
provide sufficient information to independently model the value of the
hedging instrument?
Question 55: Should changes be made to the fields or definitions
proposed in FR Y-14Q, Schedule B.2 to better account for more esoteric
derivatives such as swaptions, cap, or floors?
Exchange Traded Funds
Exchange traded funds (ETFs) are investment funds comprised of
exposures to multiple underlying assets, such as commodities, equities,
or currencies. Currently, Schedule F instructs firms to decompose
certain ETF exposures based on the fund's underlying assets. However,
the instructions also provide that all other ETFs should be reported in
the equity worksheets. This ambiguity may lead to classifying non-
equity ETFs in the equity worksheets.
All ETFs should be reported based on the underlying asset holdings
and associated risk factors. For example, ETFs for which rates or
credit exposures are the underlying holdings should be reported on the
corresponding worksheet. To provide clarity and ensure consistent
reporting, the Board is proposing to clarify the Schedule F
instructions such that all ETFs are reported in the worksheet that
corresponds to the underlying asset class and risk exposures.
Credit Card Revenue and Loss Share Agreements
Revenue and loss sharing agreements (RLSAs) are partnership
agreements firms have with private entities to share a portion of
profits, revenues, and/or losses generated by a specified asset. As
discussed in the Credit Cards Model description, the Board accounts for
private RLSAs when projecting credit card losses in the supervisory
stress test. Currently, the Board's adjustment accounts for a specific
case where a firm accounts for loss sharing payments by reducing
provisions. However, as agreement terms and reporting practices vary,
the current adjustment may not fully or consistently address
differential RLSA treatment across firms. Therefore, the Board is
considering additional enhancements to the current RLSA adjustment to
more comprehensively capture RLSAs in the supervisory stress test.
Specifically, the Board is considering one modeling approach that would
account for RLSAs at the portfolio level and a second that would
account for RLSAs at the agreement level. To facilitate the portfolio
level enhancement, the Board is proposing to collect portfolio level
details on FR Y-14M, Schedule D (Credit Card) of individual revenue
components (e.g., interest income, interest expense, noninterest
income, and noninterest expense), charge-offs and recoveries, and
provision build. Additionally, the amount of each that is subject to
partner sharing agreements and where the partner shares portions of
each are reported on the FR Y-9C, as well as the shared amounts of net
profit, net revenue, and net charge offs. To facilitate the agreement
level enhancement, the Board is proposing to collect the same
information at the agreement level, as well as effective share rates
and contractual share rates of the individual revenue, loss, and
provision components. For both approaches, the Board is proposing to
expand Schedule D.1, item 70 (``Loss Sharing'') to collect information
on the type of RLSA. If either the portfolio level or agreement level
enhancement is adopted, the Board would only adopt the corresponding FR
Y-14 revisions. If the Board does not adopt either enhancement to the
RLSA adjustment, then neither set of revisions would be implemented. If
either RLSA modeling enhancement is adopted, the corresponding FR Y-14
revision would represent an increase in estimated FR Y-14 burden hours
of approximately 2,500 hours if adopted.
Stress Test Date Changes
a. FR Y-14A Jump-Off Date
The FR Y-14A collects data on firms' projections of balance sheet
asset and liabilities, income, losses, and capital across a range of
hypothetical scenarios. These projections span a nine-quarter horizon
beginning with the first quarter of the year in which the report is
filed. This means that the jump-off date for the FR Y-14A is December
31 of the previous year, consistent with the supervisory stress test.
However, as discussed in Section VI.A of this Supplementary
Information, the Board is proposing to shift the jump-off date of the
stress test to September 30 so that the scenarios are released for
comment after the finalization of firms' balance sheets.
Consistent with this proposed jump-off date change, the Board is
proposing to modify the FR Y-14A to use a September 30 jump-off date.
These revisions would include updating the instructions to note that
the projection horizon begins in the fourth quarter of the year
preceding the reporting year, and noting that firms should report
actual capital actions in the first and second quarters of the
projection horizon, as they occur before the due date. The FR Y-14A and
capital plans would still be due April 5.
b. Global Market Shock as-of Date Submissions
As discussed in Section VI.B of this Supplementary Information, the
Board is proposing to expand the as-of date range for the global market
shock to be between October 1 of two years prior to a given stress test
cycle to October 1 of the year prior to a given stress test cycle. To
facilitate this proposed change, the Board is proposing several changes
to the FR Y-14A and FR Y-14Q.
On the FR Y-14A, the Board is proposing to update Schedule A.4
(Trading) and Schedule A.5 (Counterparty) such that the as-of date for
these schedules may fall between October 1 of two years prior to a
given stress test cycle to October 1 of the year prior to a given
stress test cycle. These schedules would still be due on April 5 of the
following year.
[[Page 51937]]
Currently, the fourth quarter submissions of FR Y-14Q, Schedule F
(Trading) and Schedule L (Counterparty) are submitted as of the global
market shock as-of date instead of quarter end. However, under the
proposal, the as-of date for the global market shock could fall in a
quarter other than the fourth quarter. Therefore, Board is proposing to
modify the submission cadence for Schedule F (Trading) and Schedule L
(Counterparty) such that, for whichever quarter contains the global
market shock as-of date, Schedule F and Schedule L would be submitted
as of that date, as opposed to quarter end. Submissions for all other
quarters would be submitted as-of quarter end.
Question 56: What, if any, other FR Y-14 revisions are needed to
facilitate the proposed changes to the stress test jump-off date and
global market shock as-of date?
Frequency: Annually, quarterly, and monthly.
Respondents: Holding companies with $100 billion or more in total
consolidated assets, as based on (1) the average of the firm's total
consolidated assets in the four most recent quarters as reported
quarterly on the firm's Consolidated Financial Statements for Holding
Companies (FR Y-9C; OMB No. 7100-0128) or (2) the average of the firm's
total consolidated assets in the most recent consecutive quarters as
reported quarterly on the firm's FR Y-9Cs, if the firm has not filed an
FR Y-9C for each of the most recent four quarters.
Total estimated number of respondents: 35.
Estimated change in burden:
FR Y-14A: -4,235 hours.
FR Y-14Q: -700 hours.
FR Y-14M: +792 hours.
Total estimated change in burden: -4,143.
Total estimated annual burden hours: 757,696.
B. Regulatory Flexibility Act Analysis
The Board is providing an initial regulatory flexibility analysis
with respect to this proposed rule. The Regulatory Flexibility Act
(RFA) \272\ requires an agency to consider whether the rules it
proposes will have a significant economic impact on a substantial
number of small entities.\273\ In connection with a proposed rule, the
RFA requires an agency to prepare and invite public comment on an
initial regulatory flexibility analysis describing the impact of the
rule on small entities, unless the agency certifies that the proposed
rule, if promulgated, will not have a significant economic impact on a
substantial number of small entities. An initial regulatory flexibility
analysis must contain (1) a description of the reasons why action by
the agency is being considered; (2) a succinct statement of the
objectives of, and legal basis for, the proposed rule; (3) a
description of, and, where feasible, an estimate of the number of small
entities to which the proposed rule will apply; (4) a description of
the projected reporting, recordkeeping, and other compliance
requirements of the proposed rule, including an estimate of the classes
of small entities that will be subject to the requirement and the type
of professional skills necessary for preparation of the report or
record; (5) an identification, to the extent practicable, of all
relevant Federal rules which may duplicate, overlap with, or conflict
with the proposed rule; and (6) a description of any significant
alternatives to the proposed rule which accomplish the stated
objectives of applicable statutes and minimize any significant economic
impact of the proposed rule on small entities.
---------------------------------------------------------------------------
\272\ 5 U.S.C. 601 et seq.
\273\ Under regulations issued by the U.S. Small Business
Administration (SBA), a small entity includes a depository
institution, bank holding company, or savings and loan holding
company with total assets of $850 million or less. 13 CFR 121.201.
Consistent with the SBA's General Principles of Affiliation, the
Board includes the assets of all domestic and foreign affiliates
toward the applicable size threshold when determining whether to
classify a particular entity as a small entity. 13 CFR 121.103. As
of December 31, 2024, there were approximately 2,364 small bank
holding companies, approximately 85 small savings and loan holding
companies, and approximately 451 small state member banks.
---------------------------------------------------------------------------
The Board has considered the potential impact of the proposed rule
on small entities in accordance with the RFA. Based on its analysis and
for the reasons stated below, the Board believes that this proposed
rule will not have a significant economic impact on a substantial
number of small entities. Nevertheless, the Board is publishing and
inviting comment on this initial regulatory flexibility analysis. In
connection with this proposal, the Board also proposes to make changes
to the Board's reporting forms.
As discussed in detail above, under the proposal, the Board is
inviting public comment on the models used to conduct the Board's
supervisory stress test, changes to those models to be implemented in
the 2026 stress test, and proposed changes to enhance the transparency
and public accountability of the Board's stress testing framework. The
proposal would amend the Policy Statement on the Scenario Design
Framework for Stress Testing, including to implement guides for
additional scenario variables, and the Stress Testing Policy Statement.
The proposal would also codify an enhanced disclosure process under
which the Board would annually publish comprehensive documentation on
the stress test models, invite public comment on any material changes
that the Board seeks to make to those models, and annually publish the
stress test scenarios for comment. Lastly, the proposal would make
changes to the FR Y-14A/Q/M to remove items that are no longer needed
to conduct the supervisory stress test and to collect additional data
to support the stress test models and improve risk capture.
As discussed above, several statutory authorities, including the
International Lending Supervision Act of 1983,\274\ section 5(b) of the
Bank Holding Company Act,\275\ the International Banking Act,\276\
section 10(g) of the Home Owners' Loan Act,\277\ and section 165 of the
Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank
Act) \278\ (as amended by section 401 of the Economic Growth,
Regulatory Relief, and Consumer Protection Act \279\), provide
authority for the Board's stress testing and stress capital buffer
framework, including this proposed rule.
---------------------------------------------------------------------------
\274\ See 12 U.S.C. 3902(1); 3907(a); 3909(a).
\275\ 12 U.S.C. 1844(b).
\276\ See 12 U.S.C. 3106.
\277\ See 12 U.S.C. 1467a(g)(1).
\278\ Dodd-Frank Act, supra note 2.
\279\ Economic Growth, Regulatory Relief, and Consumer
Protection Act, supra note 3.
---------------------------------------------------------------------------
The International Lending Supervision Act of 1983 provides the
Board with broad discretionary authority to set minimum capital levels
for state member banks and affiliates of insured depository
institutions, including holding companies, supervised by the
Board.\280\ Under section 5(b) of the Bank Holding Company Act, the
Board may issue such regulations and orders relating to capital
requirements of bank holding companies as may be necessary for the
Board to carry out the purposes of the Bank Holding Company Act.\281\
Foreign banking organizations with a U.S. subsidiary bank, branch, or
agency are made subject by the International Banking Act to the
provisions of the Bank Holding Company Act in the same manner as bank
holding companies; \282\ therefore, the Board is also authorized under
section 5(b) of the Bank Holding Company Act to impose these
[[Page 51938]]
requirements on those foreign banking organizations.
---------------------------------------------------------------------------
\280\ See 12 U.S.C. 3902(1); 3907(a); 3909(a).
\281\ 12 U.S.C. 1844(b).
\282\ See 12 U.S.C. 3106.
---------------------------------------------------------------------------
Similarly, with regard to savings and loan holding companies,
section 10(g) of the Home Owners' Loan Act authorizes the Board to
issue such regulations and orders relating to capital requirements as
the Board deems necessary and appropriate to carry out the purposes of
the Home Owners' Loan Act.\283\ Moreover, section 165 of the Dodd-Frank
Act, as amended by section 401 of the Economic Growth, Regulatory
Relief, and Consumer Protection Act, requires the Board to establish
risk-based capital requirements for large bank holding companies and
nonbank financial companies supervised by the Board.\284\ Additionally,
section 165(i)(1) of the Dodd-Frank Act, as amended by section 401 of
the Economic Growth, Regulatory Relief, and Consumer Protection Act,
requires the Board to conduct an annual supervisory stress test of
these large firms.\285\
---------------------------------------------------------------------------
\283\ See 12 U.S.C. 1467a(g)(1).
\284\ See 12 U.S.C. 5365(b)(1)(A)(i).
\285\ See 12 U.S.C. 5365(i)(1).
---------------------------------------------------------------------------
The proposed rule would apply to bank holding companies, U.S.
intermediate holding companies of foreign banking organizations, and
savings and loan holding companies, each with at least $100 billion in
total consolidated assets, as well as state member banks with more than
$250 billion in total consolidated assets, certain nonbank financial
companies supervised by the Board, and any other bank holding company
or covered savings and loan holding company domiciled in the United
States that is made subject to the capital plan rule by order of the
Board.\286\ The proposed rule would not apply to any small entities.
Further, although the Board does not project there to be a direct
impact to reporting, recordkeeping, or other compliance requirements as
a result of the proposed rule, the Board also is proposing to revise
the FR Y-14A/Q/M (Capital Assessments and Stress Testing) reports to
remove items that are no longer needed to conduct the supervisory
stress test and to collect data that would improve the calculation of
the stress capital buffer requirement. These reports are submitted by
firms subject to the Board's capital plan rule requirements; thus, the
changes would not impact small entities. In addition, the Board is
aware of no other Federal rules that duplicate, overlap, or conflict
with the proposed changes to the capital and stress testing rules.
Accordingly, the Board believes that the proposed rule will not have a
significant economic impact on a substantial number of small banking
organizations supervised by the Board and, therefore, believes that
there are no significant alternatives to the proposed rule that would
reduce the economic impact on small banking organizations supervised by
the Board.
---------------------------------------------------------------------------
\286\ There currently are no entities with less than $100
billion in total consolidated assets subject to the capital plan
rule or to the stress test rules.
---------------------------------------------------------------------------
The Board welcomes comment on all aspects of its analysis.
C. Plain Language
Section 722 of the Gramm-Leach-Bliley Act (Pub. L. 106-102, 113
Stat. 1338, 1471, 12 U.S.C. 4809) requires the federal banking agencies
to use plain language in all proposed and final rules published after
January 1, 2000. The Board has sought to present the notice of proposed
rulemaking in a simple and straightforward manner and invites comment
on the use of plain language. For example:
Is the material organized to suit your needs? If not, how
could the Board present the proposed rule more clearly?
Are the requirements in the proposed rule clearly stated?
If not, how could the proposed rule be more clearly stated?
Does the proposal contain technical language or jargon
that is not clear? If so, which language requires clarification?
Would a different format (grouping and order of sections,
use of headings, paragraphing) make the proposed rule easier to
understand? If so, what changes would achieve that?
Is this section format adequate? If not, which of the
sections should be changed and how?
What other changes can the Board incorporate to make the
proposed rule easier to understand?
D. Providing Accountability Through Transparency Act of 2023
The Providing Accountability Through Transparency Act of 2023 (12
U.S.C. 553(b)(4)) requires that a notice of proposed rulemaking include
the internet address of a summary of not more than 100 words in length
of the proposed rule, in plain language, that shall be posted on the
internet website under section 206(d) of the E-Government Act of 2002
(44 U.S.C. 3501 note).
The proposal and such a summary can be found at https://www.regulations.gov and https://www.federalreserve.gov/supervisionreg/reglisting.htm.
List of Subjects
12 CFR Part 225
Administrative practice and procedure, Banks, Banking, Federal
Reserve System, Holding companies, Reporting and recordkeeping
requirements, Securities.
12 CFR Part 238
Administrative practice and procedure, Banks, Banking, Federal
Reserve System, Holding companies, Reporting and recordkeeping
requirements, Securities.
12 CFR Part 252
Administrative practice and procedure, Banks, Banking, Capital
planning, Federal Reserve System, Holding companies, Reporting and
recordkeeping requirements, Securities, Stress testing.
Authority and Issuance
For the reasons stated in the preamble, the Board of Governors of
the Federal Reserve System proposes to amend 12 CFR chapter II as
follows:
PART 225--BANK HOLDING COMPANIES AND CHANGE IN BANK CONTROL
(REGULATION Y)
0
1. The authority citation for part 225 continues to read as follows:
Authority: 12 U.S.C. 1817(j)(13), 1818, 1828(o), 1831i, 1831p-
1, 1843(c)(8), 1844(b), 1972(1), 3106, 3108, 3310, 3331-3351, 3906,
3907, and 3909; 15 U.S.C. 1681s, 1681w, 6801 and 6805.
Subpart A--General Provisions
0
2. In Sec. 225.8:
0
a. Revise paragraph (d)(16).
0
b. Remove the text ``final,'' and add in its place the text ``third,''
in subparagraph (f)(2)(i)(A).
0
c. In paragraphs (f)(2)(i)(C)(1), (f)(4), (h)(2)(ii)(A),
(h)(2)(ii)(A)(1), (h)(2)(ii)(A)(2), (h)(2)(ii)(B), (h)(2)(ii)(B)(1),
and (h)(2)(ii)(B)(2), remove the text ``fourth through seventh'',
wherever it appears and add in its place the text ``fifth through
eighth''.
0
d. In paragraph (k)(2), remove the text ``fourth'' and replace with the
text ``fifth.''
The revisions read as follows:
Sec. 225.8 Capital Planning and stress capital buffer requirement.
* * * * *
(d) * * *
(16) Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter two quarters preceding the quarter
in which
[[Page 51939]]
the bank holding company submits its capital plan, over which the
relevant projections extend.
* * * * *
(f) * * *
(2) * * *
(i) * * *
(A) The ratio of a bank holding company's common equity tier 1
capital to risk-weighted assets, as calculated under 12 CFR part 217,
subpart D, as of the third quarter of the previous capital plan cycle,
unless otherwise determined by the Board; minus
* * * * *
(C) * * *
(1) The sum of the bank holding company's planned common stock
dividends (expressed as a dollar amount) for each of the fifth through
eighth quarters of the planning horizon
* * * * *
(4) Adjustment of stress capital buffer requirement. In each
calendar year in which the Board does not calculate a Category IV bank
holding company's stress capital buffer requirement pursuant to
paragraph (f)(1) of this section, the Board will adjust the Category IV
bank holding company's stress capital buffer requirement to be equal to
the result of the calculation set forth in paragraph (f)(2) of this
section, using the same values that were used to calculate the stress
capital buffer requirement most recently provided to the bank holding
company, except that the value used in paragraph (f)(2)(i)(C)(1) of
this section will be equal to the bank holding company's planned common
stock dividends (expressed as a dollar amount) for each of the fifth
through eighth quarters of the planning horizon as set forth in the
capital plan submitted by the bank holding company in the calendar year
in which the Board adjusts the bank holding company's stress capital
buffer requirement.
* * * * *
(h) * * *
(2) * * *
(ii) * * *
(A) Determine whether the planned capital distributions for the
fifth through eighth quarters of the planning horizon under the
Internal baseline scenario would be consistent with effective capital
distribution limitations assuming the stress capital buffer requirement
provided by the Board under paragraph (h)(1) or (i)(5) of this section,
as applicable, in place of any stress capital buffer requirement in
effect; and
(1) If the planned capital distributions for the fifth through
eighth quarters of the planning horizon under the Internal baseline
scenario would not be consistent with effective capital distribution
limitations assuming the stress capital buffer requirement provided by
the Board under paragraph (h)(1) or (i)(5) of this section, as
applicable, in place of any stress capital buffer requirement in
effect, the bank holding company must adjust its planned capital
distributions such that its planned capital distributions would be
consistent with effective capital distribution limitations assuming the
stress capital buffer requirement provided by the Board under paragraph
(h)(1) or (i)(5) of this section, as applicable, in place of any stress
capital buffer requirement in effect; or
(2) If the planned capital distributions for the fifth through
eighth quarters of the planning horizon under the Internal baseline
scenario would be consistent with effective capital distribution
limitations assuming the stress capital buffer requirement provided by
the Board under paragraph (h)(1) or (i)(5) of this section, as
applicable, in place of any stress capital buffer requirement in
effect, the bank holding company may adjust its planned capital
distributions. A bank holding company may not adjust its planned
capital distributions to be inconsistent with the effective capital
distribution limitations assuming the stress capital buffer requirement
provided by the Board under paragraph (h)(1) or (i)(5) of this section,
as applicable; and
(B) Notify the Board of any adjustments made to planned capital
distributions for the fifth through eighth quarters of the planning
horizon under the Internal baseline scenario.
* * * * *
(k) * * *
(2) The dollar amount of the capital distribution will exceed the
dollar amount of the bank holding company's final planned capital
distributions, as measured on an aggregate basis beginning in the fifth
quarter of the planning horizon through the quarter at issue.
* * * * *
PART 238--SAVINGS AND LOAN HOLDING COMPANIES (REGULATION LL)
0
3. The authority citation for part 238 continues to read as follows:
Authority: 5 U.S.C. 552, 559; 12 U.S.C. 1462, 1462a, 1463,
1464, 1467, 1467a, 1468, 5365; 1813, 1817, 1829e, 1831i, 1972; 15
U.S.C. 78l.
Subpart O--Supervisory Stress Test Requirements for Covered Savings
and Loan Holding Companies
0
4. In Sec. 238.130:
0
a. Add definitions of Material model change, Model change, and Models.
0
b. Revise definition of Planning horizon.
0
5. In Sec. 238.132:
0
a. Revise paragraph (b).
0
b. Add subsection (e).
The revisions read as follows:
Sec. 238.130 Definitions.
* * * * *
Material model change means a model change that could have, in the
Board's estimation, an impact on the post-stress CET1 regulatory
capital ratio of any covered company, or on the average post-stress
CET1 capital ratios of all covered companies required to participate in
the upcoming stress test cycle, including covered companies under 12
CFR part 252, subpart E, based on the prior year's severely adverse
scenario and prior year's input data, equal to (i) a change of 20 basis
points or more in the projected CET1 ratio of any such covered company;
or (ii) a change of 10 basis points or more in the average of the
absolute change to the values of the projected CET1 ratios of such
covered companies.
Model change means the introduction of a new model or a conceptual
change to an existing model.
Models means the analytical techniques that the Board determines
are appropriate for use in the supervisory stress test.
* * * * *
Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter prior to the start of the stress
test cycle, over which the relevant projections extend.
* * * * *
Sec. 238.132 Analysis conducted by the Board.
* * * * *
(b) Economic and financial scenarios related to the Board's
analysis. The Board will conduct its analysis using a minimum of two
different scenarios, including a baseline scenario and a severely
adverse scenario. The Board will disclose proposed scenarios by October
15 of the calendar year one year prior to the year in which the stress
test is performed, and will provide for at least a 30-day period for
public input. The Board will notify covered companies of the final
scenarios that the Board will apply to conduct the analysis for each
stress test cycle to which the covered company is subject by no later
[[Page 51940]]
than February 15 of that year, except with respect to trading
components of the scenarios and any additional scenarios that the Board
will apply to conduct the analysis, which will be communicated by no
later than March 1 of that year. The data used in such trading
components of the scenarios must be as of a date selected by the Board
that is no earlier than October 1 of the calendar year two years prior
to the year in which the stress test is performed and that precedes
October 1 of the calendar year one year prior to the year in which the
stress test is performed. Unless otherwise determined by the Board, the
as-of date for such trading or other components of the scenarios will
be communicated by the Board by October 15 of the calendar year prior
to the year in which the stress test is performed.
* * * * *
(e) Disclosure of models and material model changes--
(1) Annual disclosure. The Board will publicly disclose the models
that the Board used to conduct the analysis for the stress test by May
15 of the calendar year in which the stress test was performed pursuant
to Sec. 238.132.
(2) Material model changes from previous stress test cycle. The
Board will disclose and invite public input on any material model
changes before implementing them in the stress test.
(3) Response to public input on material model changes. The Board
will consider and respond to substantive public input on any material
model changes before implementing such material model changes in the
stress test.
* * * * *
Subpart P--Company-Run Stress Test Requirements for Savings and
Loan Holding Companies
0
7. In Sec. 238.141:
0
a. Revise the definition of Planning horizon.
0
8. In Sec. 238.143:
0
a. Revise subparagraph (b)(2)(i).
0
b. Revise subparagraph (b)(4)(i).
The revisions read as follows:
Sec. 238.141 Definitions.
* * * * *
Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter prior to the start of the stress
test cycle, over which the relevant projections extend.
* * * * *
Sec. 238.143 Stress test.
* * * * *
(b) * * *
(2) * * *
(i) The Board may require a covered company with significant
trading activity, as determined by the Board and specified in the
Capital Assessments and Stress Testing report (FR Y-14), to include a
trading and counterparty component in its severely adverse scenario in
the stress test required by this section. The data used in this
component must be as of a date that is no earlier than October 1 of the
calendar year two years prior to the year in which the stress test is
performed and that precedes October 1 of the calendar year one year
prior to the year in which the stress test is performed pursuant to
this section. Unless otherwise determined by the Board, the as-of date
of such component will be communicated to the company by October 15 of
the calendar year one year prior to the year in which the stress test
is performed and a final description of the component will be
communicated to the company by no later than March 1 of the calendar
year in which the stress test is performed pursuant to this section.
* * * * *
(4) * * *
(i) Notification of additional component. If the Board requires a
covered company to include one or more additional components in its
severely adverse scenario under paragraph (b)(2) of this section or to
use one or more additional scenarios under paragraph (b)(3) of this
section, the Board will notify the company in writing and include a
discussion of the basis for its determination. The Board will provide
such notification no later than September 30 of the preceding calendar
year. The notification will include a general description of the
additional component(s) or additional scenario(s) and the basis for
requiring the company to include the additional component(s) or
additional scenario(s).
* * * * *
Subpart S--Capital Planning and Stress Capital Buffer Requirement
0
9. In Sec. 238.170:
0
a. Revise paragraph (d)(14).
b. Remove the text ``final,'' and add in its place the text
``third,'' in subparagraph (f)(2)(i)(A).
0
c. In paragraphs (f)(2)(i)(C)(1), (f)(4), (h)(2)(ii)(A),
(h)(2)(ii)(A)(1), (h)(2)(ii)(A)(2), (h)(2)(ii)(B), (h)(2)(ii)(B)(1),
and (h)(2)(ii)(B)(2), remove the text ``fourth through seventh'',
wherever it appears and add in its place the text ``fifth through
eighth''.
0
d. In paragraph (k)(2), remove the text ``fourth'' and replace with the
text ``fifth.''
The revisions read as follows:
* * * * *
(d) * * *
(14) Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter two quarters preceding the quarter
in which the covered savings and loan holding company submits its
capital plan, over which the relevant projections extend.
* * * * *
(f) * * *
(2) * * *
(i) * * *
(A) The ratio of a covered savings and loan holding company's
common equity tier 1 capital to risk-weighted assets, as calculated
under 12 CFR part 217, subpart D, as of the third quarter of the
previous capital plan cycle, unless otherwise determined by the Board;
minus
* * * * *
(C) * * *
(1) The sum of the covered savings and loan holding company's
planned common stock dividends (expressed as a dollar amount) for each
of the fifth through eighth quarters of the planning horizon; to
* * * * *
(4) Adjustment of stress capital buffer requirement. In each
calendar year in which the Board does not calculate a Category IV
savings and loan holding company's stress capital buffer requirement
pursuant to paragraph (f)(1) of this section, the Board will adjust the
Category IV savings and loan holding company's stress capital buffer
requirement to be equal to the result of the calculation set forth in
paragraph (f)(2) of this section, using the same values that were used
to calculate the stress capital buffer requirement most recently
provided to the covered savings and loan holding company, except that
the value used in paragraph (f)(2)(i)(C)(1) of the calculation will be
equal to the covered savings and loan holding company's planned common
stock dividends (expressed as a dollar amount) for each of the fifth
through eighth quarters of the planning horizon as set forth in the
capital plan submitted by the covered savings and loan holding company
in the calendar year in which the Board adjusts the covered savings and
loan holding company's stress capital buffer requirement.
* * * * *
(h) * * *
(2) * * *
(ii) * * *
[[Page 51941]]
(A) Determine whether the planned capital distributions for the
fifth through eighth quarters of the planning horizon under the
Internal baseline scenario would be consistent with effective capital
distribution limitations assuming the stress capital buffer requirement
provided by the Board under paragraph (h)(1) or (i)(5) of this section,
as applicable, in place of any stress capital buffer requirement in
effect; and
(1) If the planned capital distributions for the fifth through
eighth quarters of the planning horizon under the Internal baseline
scenario would not be consistent with effective capital distribution
limitations assuming the stress capital buffer requirement provided by
the Board under paragraph (h)(1) or (i)(5) of this section, as
applicable, in place of any stress capital buffer requirement in
effect, the covered savings and loan holding company must adjust its
planned capital distributions such that its planned capital
distributions would be consistent with effective capital distribution
limitations assuming the stress capital buffer requirement provided by
the Board under paragraph (h)(1) or (i)(5) of this section, as
applicable, in place of any stress capital buffer requirement in
effect; or
(2) If the planned capital distributions for the fifth through
eighth quarters of the planning horizon under the Internal baseline
scenario would be consistent with effective capital distribution
limitations assuming the stress capital buffer requirement provided by
the Board under paragraph (h)(1) or (i)(5) of this section, as
applicable, in place of any stress capital buffer requirement in
effect, the covered savings and loan holding company may adjust its
planned capital distributions. A covered savings and loan holding
company may not adjust its planned capital distributions to be
inconsistent with the effective capital distribution limitations
assuming the stress capital buffer requirement provided by the Board
under paragraph (h)(1) or (i)(5) of this section, as applicable; and
(B) Notify the Board of any adjustments made to planned capital
distributions for the fifth through eighth quarters of the planning
horizon under the Internal baseline scenario.
* * * * *
(k) * * *
(2) The dollar amount of the capital distribution will exceed the
dollar amount of the covered savings and loan holding company's final
planned capital distributions, as measured on an aggregate basis
beginning in the fifth quarter of the planning horizon through the
quarter at issue.
* * * * *
PART 252--ENHANCED PRUDENTIAL STANDARDS (REGULATION YY)
0
10. The authority citation for part 252 continues to read as follows:
Authority: 12 U.S.C. 321-338a, 481-486, 1467a, 1818, 1828,
1831n, 1831o, 1831p-1, 1831w, 1835, 1844(b), 1844(c), 3101 et seq.,
3101 note, 3904, 3906-3909, 4808, 5361, 5362, 5365, 5366, 5367,
5368, 5371.
Subpart B--Company-Run Stress Test Requirements for State Member
Banks With Total Consolidated Assets Over $250 Billion
0
11. In Sec. 252.12:
0
a. Revise the definition of Planning Horizon.
0
12. In Sec. 252.14:
0
a. Revise subparagraph (b)(2)(i).
0
b. Revise subparagraph (b)(4)(i).
The revisions read as follows:
Sec. 252.12 Definitions.
* * * * *
Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter prior to the start of the stress
test cycle, over which the relevant projections extend.
* * * * *
Sec. 252.14 Stress test.
* * * * *
(b) * * *
(2) * * *
(i) The Board may require a state member bank with significant
trading activity, as determined by the Board and specified in the
Capital Assessments and Stress Testing report (FR Y-14), to include a
trading and counterparty component in its severely adverse scenario in
the stress test required by this section. The Board may also require a
state member bank that is subject to 12 CFR part 217, subpart F or that
is a subsidiary of a bank holding company that is subject to section
Sec. 252.54(b)(2)(i) to include a trading and counterparty component
in the state member bank's severely adverse scenario in the stress test
required by this section. The data used in this component must be as of
a date that is no earlier than October 1 of the calendar year two years
prior to the year in which the stress test is performed and that
precedes October 1 of the calendar year one year prior to the year in
which the stress test is performed. Unless otherwise determined by the
Board, the as-of date for such component will be communicated to the
company by October 15 of the calendar year one year prior to the year
in which the stress test is performed and a final description of the
component will be communicated to the company by no later than March 1
of the calendar year in which the stress test is performed pursuant to
this section.
* * * * *
(4) * * *
(i) Notification of additional component or scenario. If the Board
requires a state member bank to include one or more additional
components in its severely adverse scenario under paragraph (b)(2) of
this section or to use one or more additional scenarios under paragraph
(b)(3) of this section, the Board will notify the company in writing by
September 30 of the preceding calendar year and include a discussion of
the basis for its determination.
* * * * *
Subpart E--Supervisory Stress Test Requirements for Certain U.S.
Banking Organizations With $100 Billion or More in Total
Consolidated Assets and Nonbank Financial Companies Supervised by
the Board
0
13. In Sec. 252.42:
0
a. Add definitions of Material model change, Model change, and Models.
0
b. Revise the definition of Planning Horizon.
0
14. In Sec. 252.44:
0
a. Revise paragraph (b).
0
b. Add subsection (e).
The revisions read as follows:
Sec. 252.42 Definitions.
* * * * *
Material model change means a model change that could have, in the
Board's estimation, an impact on the post-stress CET1 regulatory
capital ratio of any covered company, or on the average post-stress
CET1 capital ratios of all covered companies required to participate in
the upcoming stress test cycle, including covered companies under 12
CFR part 238, subpart O, based on the prior year's severely adverse
scenario and prior year's input data, equal to (i) a change of 20 basis
points or more in the projected CET1 ratio of any such covered company;
or (ii) a change of 10 basis points or more in the average of the
absolute change to the values of the projected CET1 ratios of such
covered companies.
Model change means the introduction of a new model or a conceptual
change to an existing model.
Models means the analytical techniques that the Board determines
[[Page 51942]]
are appropriate for use in the supervisory stress test.
* * * * *
Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter prior to the start of the stress
test cycle, over which the relevant projections extend.
* * * * *
Sec. 252.44 Analysis conducted by the Board.
* * * * *
(b) Economic and financial scenarios related to the Board's
analysis. The Board will conduct its analysis using a minimum of two
different scenarios, including a baseline scenario and a severely
adverse scenario. The Board will disclose proposed scenarios by October
15 of the calendar year one year prior to the year in which the stress
test is performed, and will provide for at least a 30-day period for
public input. The Board will notify covered companies of the final
scenarios that the Board will apply to conduct the analysis for each
stress test cycle to which the covered company is subject by no later
than February 15 of that year, except with respect to trading or any
other components of the scenarios and any additional scenarios that the
Board will apply to conduct the analysis, which will be communicated by
no later than March 1 of that year. The data used in such trading or
other components of the scenarios must be as-of a date selected by the
Board that is no earlier than October 1 of the calendar year two years
prior to the year in which the stress test is performed and that
precedes October 1 of the calendar year one year prior to the year in
which the stress test is performed. Unless otherwise determined by the
Board, the as-of date for such trading or other components of the
scenarios will be communicated by the Board by October 15 of the
calendar year prior to the year in which the stress test is performed.
* * * * *
(e) Disclosure of models and material model changes--
(1) Annual disclosure. The Board will publicly disclose the models
that the Board used to conduct the analysis for the stress test by May
15 of the calendar year in which the stress test was conducted pursuant
to Sec. 252.44.
(2) Material model changes from previous stress test cycle. The
Board will disclose and invite public input on any material model
changes before implementing such material model changes in the stress
test.
(3) Response to public input on material model changes. The Board
will consider and respond to substantive public input on any material
model changes before implementing such material model changes in the
stress test.
* * * * *
Subpart F--Company-Run Stress Test Requirements for Certain U.S.
Bank Holding Companies and Nonbank Financial Companies Supervised
by the Board
0
16. In Sec. 252.52:
0
a. Revise the definition of Planning horizon.
0
17. In Sec. 252.54:
0
a. Revise subparagraph (b)(2)(i).
0
b. Revise subparagraph (b)(4)(i).
The revisions read as follows:
Sec. 252.52 Definitions.
* * * * *
Planning horizon means the period of at least nine consecutive
quarters, beginning with the quarter prior to the start of the stress
test cycle, over which the relevant projections extend.
* * * * *
Sec. 252.54 Stress test.
* * * * *
(b) * * *
(2) * * *
(i) The Board may require a covered company with significant
trading activity to include a trading and counterparty component in its
severely adverse scenario in the stress test required by this section.
The data used in this component must be as of a date selected by the
Board that is no earlier than October 1 of the calendar year two years
prior to the year in which the stress test is performed that precedes
October 1 of the calendar year one year prior to the year in which the
stress test is performed pursuant to this section. Unless otherwise
determined by the Board, the as-of date for such component will be
communicated to the company by October 15 of the calendar year one year
prior to the year in which the stress test is performed and a final
description of the component will be communicated to the company by no
later than March 1 of the calendar year in which the stress test is
performed pursuant to this section. A covered company has significant
trading activity if it has:
(A) Aggregate trading assets and liabilities of $50 billion or
more, or aggregate trading assets and liabilities equal to 10 percent
or more of total consolidated assets;
(B) Is not a Category IV bank holding company.
* * * * *
(4) * * *
(i) Notification of additional component. If the Board requires a
covered company to include one or more additional components in its
severely adverse scenarios under paragraph (b)(2) of this section or to
use one or more additional scenarios under paragraph (b)(3) of this
section, the Board will notify the company in writing. The Board will
provide such notification no later than September 30 of the preceding
calendar year. The notification will include a general description of
the additional component(s) or additional scenario(s) and the basis for
requiring the company to include the additional component(s) or
additional scenario(s).
* * * * *
Appendix A to Part 252--Policy Statement on the Scenario Design
Framework for Stress Testing
0
18. Appendix A to part 252 is revised to read as follows:
1. Background
(a) The Board has imposed stress testing requirements through
its regulations (stress test rules) implementing section 165(i) of
the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-
Frank Act or Act), section 10(g) of the Home Owners' Loan Act, and
section 401(e) of the Economic Growth, Regulatory Relief, and
Consumer Protection Act, and through its capital plan rule (12 CFR
225.8). Under the stress test rules, the Board conducts a
supervisory stress test of each bank holding company with total
consolidated assets of $100 billion or more, intermediate holding
company of a foreign banking organization with total consolidated
assets of $100 billion or more, and nonbank financial company that
the Financial Stability Oversight Council has designated for
supervision by the Board (together, covered companies).\287\ In
addition, under the stress test rules, certain firms are also
subject to company-run stress test requirements.\288\ The Board will
provide two different sets of conditions (each set, a scenario),
including baseline and severely adverse scenario for both
supervisory and company-run stress tests (macroeconomic
scenarios).\289\
---------------------------------------------------------------------------
\287\ 12 U.S.C. 5365(i)(1); 12 CFR part 252, subpart E.
\288\ 12 U.S.C. 5365(i)(2); 12 CFR part 252, subparts B and F.
\289\ The stress test rules define scenarios as those sets of
conditions that affect the United States economy or the financial
condition of a company that the Board determines are appropriate for
use in stress tests, including, but not limited to, baseline and
severely adverse scenarios. The stress test rules define baseline
scenario as a set of conditions that affect the United States
economy or the financial condition of a company and that reflect the
consensus views of the economic and financial outlook. The stress
test rules define severely adverse scenario as a set of conditions
that affect the U.S. economy or the financial condition of a company
and that overall are significantly more severe than those associated
with the baseline scenario and may include trading or other
additional components.
---------------------------------------------------------------------------
[[Page 51943]]
(b) The stress test rules provide that the Board will notify
covered companies by no later than February 15 of each year of the
scenarios it will use to conduct its supervisory stress tests and
provide, also by no later than February 15, covered companies and
other financial companies subject to the final rules the set of
scenarios they must use to conduct their company-run stress tests.
Under the stress test rules, the Board may require certain companies
to use additional components in the severely adverse scenario or
additional scenarios. For example, the Board expects to require
large banking organizations with significant trading activities to
include a trading and counterparty component (market shock,
described in the following sections) in their severely adverse
scenario. The Board will provide any additional components or
scenarios by no later than March 1 of each year.\290\ The Board
expects that the scenarios it will require the companies to use will
be the same as those the Board will use to conduct its supervisory
stress tests (together, stress test scenarios).
---------------------------------------------------------------------------
\290\ 12 CFR 252.14(b); 12 CFR 252.44(b); 12 CFR 252.54(b).
---------------------------------------------------------------------------
(c) In addition, Sec. 225.8 of the Board's Regulation Y
(capital plan rule) requires covered companies to submit annual
capital plans, including stress test results, to the Board in order
to allow the Board to assess whether they have robust, forward-
looking capital planning processes and have sufficient capital to
continue operations throughout times of economic and financial
stress.\291\
---------------------------------------------------------------------------
\291\ See 12 CFR 225.8.
---------------------------------------------------------------------------
(d) Stress tests required under the stress test rules and under
the capital plan rule require the Board and financial companies to
calculate pro-forma capital levels--rather than ``current'' or
actual levels--over a specified planning horizon under baseline and
stressful scenarios. This approach integrates key lessons of the
2007-2009 financial crisis and subsequent stress events into the
Board's supervisory framework. During the financial crisis, investor
and counterparty confidence in the capitalization of financial
companies eroded rapidly in the face of changes in the current and
expected economic and financial conditions, and this loss in market
confidence imperiled companies' ability to access funding, continue
operations, serve as a credit intermediary, and meet obligations to
creditors and counterparties. Importantly, such a loss in confidence
occurred even when a financial institution's capital ratios were in
excess of regulatory minimums. This is because the institution's
capital ratios were perceived as lagging indicators of its financial
condition, particularly when conditions were changing.
(e) The stress tests required under the stress test rules and
capital plan rule are a valuable supervisory tool that provide a
forward-looking assessment of large financial companies' capital
adequacy under hypothetical economic and financial market
conditions. Currently, these stress tests primarily focus on credit
risk, operational risk, and market risk--that is, risk of mark-to-
market losses associated with companies' trading and counterparty
positions--and not on other types of risk, such as liquidity risk.
Pressures stemming from these sources are considered in separate
supervisory exercises. No single supervisory tool, including the
stress tests, can provide an assessment of a company's ability to
withstand every potential source of risk.
(f) Selecting appropriate scenarios is an especially significant
consideration for stress tests required under the capital plan rule,
which ties the review of a company's performance under stress
scenarios to its ability to make capital distributions. More severe
scenarios, all other things being equal, generally translate into
larger projected declines in banks' capital. Thus, a company would
need more capital today to meet its minimum capital requirements in
more stressful scenarios and have the ability to continue making
capital distributions, such as common dividend payments. This
translation is far from mechanical, however; it will depend on
factors that are specific to a given company, such as underwriting
standards and the company's business model, which would also greatly
affect projected revenue, losses, and capital.
2. Overview and Scope
(a) This policy statement provides more detail on the
characteristics of the stress test scenarios and explains the
considerations and procedures that underlie the approach for
formulating these scenarios. The considerations and procedures
described in this policy statement apply to the Board's stress
testing framework, including to the stress tests required under 12
CFR part 252, subparts B, E, and F as well as the Board's capital
plan rule (12 CFR 225.8).\292\
---------------------------------------------------------------------------
\292\ 12 CFR 252.14(a); 12 CFR 252.44(a); 12 CFR 252.54(a).
---------------------------------------------------------------------------
(b) Although the Board does not envision that the broad approach
used to develop scenarios will change from year to year, the stress
test scenarios will reflect changes in the outlook for economic and
financial conditions and changes to specific risks or
vulnerabilities that the Board, in consultation with the other
federal banking agencies, determines should be considered in the
annual stress tests. The stress test scenarios should not be
regarded as forecasts; rather, they are hypothetical paths of
economic variables that will be used to assess the strength and
resilience of the companies' capital in various economic and
financial environments.
(c) The remainder of this policy statement is organized as
follows. Section 3 provides a broad description of the baseline and
severely adverse scenarios and describes the relationship between
the macroeconomic scenario and the market shock component of the
severely adverse scenario applicable to companies with significant
trading activity. This section also describes the types of variables
that the Board expects to include in the macroeconomic scenarios and
the market shock component. Section 4 describes the Board's approach
for developing the macroeconomic scenarios, and section 5 describes
the approach for the market shocks. Section 6 provides a timeline
for the formulation and publication of the macroeconomic assumptions
and market shocks.
3. Content of the Stress Test Scenarios
(a) The Board will publish two different scenarios, including
baseline and severely adverse conditions, for use in stress tests
required in the stress test rules.\293\ In general, the Board
anticipates that it will not issue additional scenarios. Specific
circumstances or vulnerabilities that in any given year the Board
may determine require particular vigilance to help ensure the
resilience of the banking sector may be captured in the severely
adverse scenario, and are expected to be explained through the
comment process in those stress test cycles.
---------------------------------------------------------------------------
\293\ 12 CFR 252.14(b); 12 CFR 252.44(b); 12 CFR 252.54(b).
---------------------------------------------------------------------------
(b) While the Board generally expects to use the same scenarios
for all companies subject to the final rule, it may require a subset
of companies--depending on a company's financial condition, size,
complexity, risk profile, scope of operations, or activities, or
risks to the U.S. economy--to include additional scenario components
or additional scenarios that are designed to capture different
effects of adverse events on revenue, losses, and capital. One
example of such components is the market shock that applies only to
companies with significant trading activity. Additional components
or scenarios may also include other stress factors that may not
necessarily be directly correlated to macroeconomic or financial
assumptions but nevertheless can materially affect companies' risks,
such as the unexpected default of a major counterparty.
(c) Early in each stress testing cycle, the Board plans to
publish the macroeconomic scenarios along with a brief narrative
summary that provides a description of the economic situation
underlying the scenario and explains how the scenarios have changed
relative to the previous year. In addition, to assist companies in
projecting the paths of additional variables in a manner consistent
with the scenario, the narrative will provide descriptions of the
general path of some additional variables. These descriptions will
be general--that is, they will describe developments for broad
classes of variables rather than for specific variables--and will
specify the intensity and direction of variable changes but not
numeric magnitudes. These descriptions should provide guidance that
will be useful to companies in specifying the paths of the
additional variables for their company-run stress tests. Note that
in practice it will not be possible for the narrative to include
descriptions of all the additional variables that companies may need
for their company-run stress tests. In cases where scenarios are
designed to reflect particular risks and vulnerabilities, the
narrative will also explain the underlying motivation for these
features of the scenario. The Board also plans to release a
description of the market shock components.
[[Page 51944]]
3.1 Macroeconomic Scenarios
(a) The macroeconomic scenarios will consist of the future paths
of a set of economic and financial variables.\294\ The economic and
financial variables included in the scenarios will likely comprise
those included in the ``2014 Supervisory Scenarios for Annual Stress
Tests Required under the Dodd-Frank Act Stress Testing Rules and the
Capital Plan Rule'' (2013 supervisory scenarios). The domestic U.S.
variables provided for in the 2013 supervisory scenarios included:
---------------------------------------------------------------------------
\294\ The future path of a variable refers to its specification
over a given time period. For example, the path of unemployment can
be described in percentage terms on a quarterly basis over the
stress testing time horizon.
---------------------------------------------------------------------------
(1) Six measures of economic activity and prices: Real and
nominal gross domestic product (GDP) growth, the unemployment rate
of the civilian non-institutional population aged 16 and over, real
and nominal disposable personal income growth, and the Consumer
Price Index (CPI) inflation rate;
(2) Four measures of developments in equity and property
markets: The Core Logic National House Price Index, the National
Council for Real Estate Investment Fiduciaries Commercial Real
Estate Price Index, the Dow Jones Total Stock Market Index, and the
Chicago Board Options Exchange Market Volatility Index; and
(3) Six measures of interest rates: The rate on the 3-month
Treasury bill, the yield on the 5-year Treasury bond, the yield on
the 10-year Treasury bond, the yield on a 10-year BBB corporate
security, the prime rate, and the interest rate associated with a
conforming, conventional, fixed-rate, 30-year mortgage.
(b) The international variables provided for in the 2014
supervisory scenarios included, for the euro area, the United
Kingdom, developing Asia, and Japan:
(1) Percent change in real GDP;
(2) Percent change in the CPI or local equivalent; and
(3) The U.S./foreign currency exchange rate.\295\
---------------------------------------------------------------------------
\295\ The Board may increase the range of countries or regions
included in future scenarios, as appropriate.
---------------------------------------------------------------------------
(c) The economic variables included in the scenarios influence
key items affecting financial companies' net income, including pre-
provision net revenue and credit losses on loans and securities.
Moreover, these variables exhibit fairly typical trends in adverse
economic climates that can have unfavorable implications for
companies' net income and, thus, capital positions.
(d) The economic variables included in the scenario may change
over time. For example, the Board may add variables to a scenario if
the international footprint of companies that are subject to the
stress testing rules changed notably over time such that the
variables already included in the scenario no longer sufficiently
capture the material risks of these companies. Alternatively,
historical relationships between macroeconomic variables could
change over time such that one variable (e.g., disposable personal
income growth) that previously provided a good proxy for another
(e.g., light vehicle sales) in modeling companies' pre-provision net
revenue or credit losses ceases to do so, resulting in the need to
create a separate path, or alternative proxy, for the other
variable. However, recognizing the amount of work required for
companies to incorporate the scenario variables into their stress
testing models, the Board expects to eliminate variables from the
scenarios only in rare instances.
(e) The Board expects that the company may not use all of the
variables provided in the scenario, if those variables are not
appropriate to the company's line of business, or may add additional
variables, as appropriate. The Board expects the companies to ensure
that the paths of such additional variables are consistent with the
scenarios the Board provided. For example, the companies may use, as
part of their internal stress test models, local-level variables,
such as state-level unemployment rates or city-level house prices.
While the Board does not plan to include local-level macro variables
in the stress test scenarios it provides, it expects the companies
to evaluate the paths of local-level macro variables as needed for
their internal models, and ensure internal consistency between these
variables and their aggregate, macro-economic counterparts. The
Board will provide the macroeconomic scenario component of the
stress test scenarios for a period that spans a minimum of 13
quarters. The scenario horizon reflects the supervisory stress test
approach that the Board plans to use. Under the stress test rules,
the Board will assess the effect of different scenarios on the
consolidated capital of each company over a forward-looking planning
horizon of at least nine quarters.
3.2 Market Shock Component
(a) The market shock component of the severely adverse scenario
will only apply to companies with significant trading activity and
their subsidiaries.\296\ The component consists of large moves in
market prices and rates that would be expected to generate losses.
Market shocks differ from macroeconomic scenarios in several ways,
both in their design and application. For instance, market shocks
that might typically be observed over an extended period (e.g., 3
months) are assumed to affect the market value of the companies'
trading assets and liabilities immediately. In addition, under the
stress test rules, the as-of date for market shocks will differ from
the quarter-end, and the Board will provide the as-of date for
market shocks no later than February 1 of each year. Finally, as
described in section 4, the market shock includes a much larger set
of risk factors than the set of economic and financial variables
included in macroeconomic scenarios. Broadly, these risk factors
include shocks to financial market variables that affect asset
prices, such as a credit spread or the yield on a bond, and, in some
cases, the value of the position itself (e.g., the market value of
securitized positions).
---------------------------------------------------------------------------
\296\ Currently, companies with significant trading activity
include any bank holding company or intermediate holding company
that (1) has aggregate trading assets and liabilities of $50 billion
or more, or aggregate trading assets and liabilities equal to 10
percent or more of total consolidated assets, and (2) is not a
Category IV firm. The Board may also subject a state member bank
subsidiary of any such bank holding company to the market shock
component. The set of companies subject to the market shock
component could change over time as the size, scope, and complexity
of financial company's trading activities evolve.
---------------------------------------------------------------------------
(b) The Board envisions that the market shocks will include
shocks to a broad range of risk factors that are similar in
granularity to those risk factors that trading companies use
internally to produce profit and loss estimates, under stressful
market scenarios, for all asset classes that are considered trading
assets, including public equities, credit, interest rates, foreign
exchange rates, and commodities. Examples of risk factors include,
but are not limited to:
(1) Public equity indices to which companies with significant
trading activity may have exposure, along with term structures of
implied volatilities;
(2) Cross-currency foreign exchange rates of selected
currencies, along term structures of implied volatilities;
(3) Term structures of government rates (e.g., U.S. Treasuries),
interbank rates (e.g., swap rates) and potentially other key rates
(e.g., commercial paper) for developed markets and for developing
and emerging market nations to which companies may have exposure;
(4) Term structures of implied volatilities that are key inputs
to the pricing of interest rate derivatives;
(5) Term structures of futures prices for energy products
including crude oil (differentiated by country of origin), natural
gas, and power;
(6) Term structures of futures prices for metals and
agricultural commodities; and
(7) Credit spreads or instrument prices for credit-sensitive
product segments including: corporate bonds, credit default swaps,
and collateralized debt obligations by risk; non-agency residential
mortgage-backed securities and commercial mortgage-backed securities
by risk and vintage; sovereign debt; and, municipal bonds.
4. Approach for Formulating the Macroeconomic Assumptions for Scenarios
(a) This section describes the Board's approach for formulating
macroeconomic assumptions for each scenario. The methodologies for
formulating this part of each scenario differ by scenario, so these
methodologies for the baseline and severely adverse scenarios are
described separately in each of the following subsections.
(b) In general, the baseline scenario will reflect the most
recently available consensus views of the macroeconomic outlook
expressed by professional forecasters, government agencies, and
other public-sector organizations as of the beginning of the stress-
test cycle. The severely adverse scenario will consist of a set of
economic and financial conditions that reflect the conditions of
post-war U.S. recessions.
(c) Each of these scenarios is described further in sections
below as follows: Baseline (subsection 4.1) and severely adverse
(subsection 4.2)
[[Page 51945]]
4.1 Approach for Formulating Macroeconomic Assumptions in the
Baseline Scenario
(a) The stress test rules define the baseline scenario as a set
of conditions that affect the U.S. economy or the financial
condition of a banking organization, and that reflect the consensus
views of the economic and financial outlook. Projections under a
baseline scenario are used to evaluate how companies would perform
in more likely economic and financial conditions. The baseline
serves also as a point of comparison to the severely adverse
scenario, giving some sense of how much of the company's capital
decline could be ascribed to the scenario as opposed to the
company's capital adequacy under expected conditions.
(b) The baseline scenario will be developed around a
macroeconomic projection that captures the prevailing views of
private-sector forecasters (e.g., Blue Chip Consensus Forecasts and
the Survey of Professional Forecasters), government agencies, and
other public-sector organizations (e.g., the International Monetary
Fund and the Organization for Economic Co-operation and Development)
near the beginning of the annual stress-test cycle. The baseline
scenario is designed to represent a consensus expectation of certain
economic variables over the time period of the tests and it is not
the Board's internal forecast for those economic variables. For
example, the baseline path of short-term interest rates is
constructed from consensus forecasts and may differ from that
implied by the Federal Open Market Committee's Summary of Economic
Projections.
(c) For some scenario variables--such as U.S. real GDP growth,
the unemployment rate, and the consumer price index--there will be
many different forecasts available to project the paths of these
variables in the baseline scenario. For others, a more limited
number of forecasts will be available. If available forecasts
diverge notably, the baseline scenario will reflect an assessment of
the forecast that is deemed to be most plausible. The Board also
considers the output of a macroeconomic model, for which the Board
will maintain a description separately on the Board's website,
developed by Board staff for use in constructing the values of some
of the variables in the scenarios for the annual stress test. In
setting the paths of variables in the baseline scenario, particular
care will be taken to ensure that, together, the paths present a
coherent and plausible outlook for the U.S. and global economy,
given the economic climate in which they are formulated. However,
the macroeconomic model was designed to meet the specific needs of
the stress testing program, and the resulting baseline scenarios are
not Federal Reserve forecasts.
4.2 Approach for Formulating the Macroeconomic Assumptions in the
Severely Adverse Scenario
The stress test rules define a severely adverse scenario as a
set of conditions that affect the U.S. economy or the financial
condition of a financial company and that overall are significantly
more severe than those associated with the baseline scenario. The
financial company will be required to publicly disclose a summary of
the results of its stress test under the severely adverse scenario,
and the Board intends to publicly disclose the results of its
analysis of the financial company under the severely adverse
scenario.
4.2.1 General Approach: The Recession Approach
(a) The Board intends to use a recession approach to develop the
severely adverse scenario. In the recession approach, the Board will
specify the future paths of variables to reflect conditions that
characterize post-war U.S. recessions, generating either a typical
or specific recreation of a post-war U.S. recession. The Board chose
this approach because it has observed that the conditions that
typically occur in recessions--such as increasing unemployment,
declining asset prices, and contracting loan demand--can put
significant stress on companies' balance sheets. This stress can
occur through a variety of channels, including higher loss
provisions due to increased delinquencies and defaults; losses on
trading positions through sharp moves in market prices; and lower
bank income through reduced loan originations. For these reasons,
the Board believes that the paths of economic and financial
variables in the severely adverse scenario should, at a minimum,
resemble the paths of those variables observed during a recession.
(b) This approach requires consideration of the type of
recession to feature. All post-war U.S. recessions have not been
identical: Some recessions have been associated with very elevated
interest rates, some have been associated with sizable asset price
declines, and some have been relatively more global. Recessions that
are caused by or exacerbated by a financial crisis often are deeper
and more protracted than other recessions. The Board therefore
believes that the severely adverse scenario should be triggered by a
sudden and substantial increase in risk aversion and uncertainty
that causes sharp declines in risky financial asset prices, lower
interest rates on safe assets, and a rise in volatility big enough
to disrupt functioning in some markets. Although markets resume
normal functioning within a few months, the rise in uncertainty and
decline in wealth causes businesses to take nearly simultaneous
steps to reduce employment and investment and households to reduce
spending. Negative feedback effects between contracting economic
activity and financial markets' response lead to a deep and
prolonged decline in overall economic activity, inflation, and asset
prices followed by a shallow recovery.
(c) Indeed, the most common features of recessions are increases
in the unemployment rate and contractions in aggregate incomes and
economic activity. For this and the following reasons, the Board
intends to use a rise in the unemployment rate as the primary basis
for calibrating the severity of the severely adverse scenario.
First, the unemployment rate is likely the most representative
single summary indicator of adverse economic conditions. Second, in
comparison to GDP, labor market data have traditionally featured
more prominently than GDP in the set of indicators that the National
Bureau of Economic Research reviews to inform its recession
dates.\297\ Third and finally, the growth rate of potential output
can cause the size of the decline in GDP to vary between recessions.
While changes in the unemployment rate can also vary over time due
to demographic factors, this seems to have more limited implications
over time relative to changes in potential output growth. The
unemployment rate used in the severely adverse scenario will reflect
an unemployment rate that has been observed in severe post-war U.S.
recessions, measuring severity by the absolute level of and relative
increase in the unemployment rate.\298\
---------------------------------------------------------------------------
\297\ More recently, a monthly measure of GDP has been added to
the list of indicators.
\298\ Even though all recessions feature increases in the
unemployment rate and contractions in incomes and economic activity,
the size of this change has varied over post-war U.S. recessions.
Table 1 documents the variability in the depth of post-war U.S.
recessions. Some recessions--labeled mild in Table 1--have been
relatively modest, with GDP edging down just slightly and the
unemployment rate moving up about a percentage point. Other
recessions--labeled severe in Table 1--have been much harsher, with
GDP dropping 3.75 percent and the unemployment rate moving up a
total of about 4 percentage points.
---------------------------------------------------------------------------
(d) The Board believes that the severely adverse scenario should
also reflect a housing recession. The house prices path set in the
severely adverse scenario will reflect developments that have been
observed in post-war U.S. housing recessions, measuring severity by
the absolute level of and relative decrease in the house prices.
(e) As described below, the Board has developed guides for
several additional variables including equity prices, the Chicago
Board Options Exchange's Volatility Index (VIX), BBB spread,
mortgage rate spread, commercial real estate prices, and 5-year and
10-year Treasury yields. The international variables (GDP,
inflation, and exchange rates) are also subject to guides.
(f) The Board will specify the paths of those other
macroeconomic and financial market variables based on their behavior
during previous recessions or other periods of financial stress, as
well as informed assessments of how that behavior co-moved with the
paths of unemployment, income, house prices, and activity during
periods of macrofinancial stress. Some of these other variables,
however, have taken divergent paths in previous recessions (e.g.,
foreign GDP). The analysis that the Board conducted to develop the
guides informed its judgment in selecting the appropriate ranges for
the peak or trough, the timing of that peak or trough, and ending
values, as well as the trajectory of these variables between the
starting and ending values. In general, the path for these variables
also will be based on their underlying structure at the time that
the scenario is designed (e.g., economic or financial-system
vulnerabilities in other countries).
(g) The Board considered alternative methods for scenario design
of the severely adverse scenario, including a probabilistic
approach. The probabilistic approach constructs a baseline forecast
from a large-
[[Page 51946]]
scale macroeconomic model and identifies a scenario that would have
a specific probabilistic likelihood given the baseline forecast. The
Board believes that, at this time, the recession approach is better
suited for developing the severely adverse scenario than a
probabilistic approach because it guarantees a recession of some
specified severity. In contrast, the probabilistic approach requires
the choice of an extreme tail outcome--relative to baseline--to
characterize the severely adverse scenario (e.g., a 5 percent or a 1
percent tail outcome). In practice, this choice is difficult as
adverse economic outcomes are typically thought of in terms of how
variables evolve in an absolute sense rather than how far away they
lie in the probability space away from the baseline. In this sense,
a scenario featuring a recession may be somewhat clearer and more
straightforward to communicate. Finally, the probabilistic approach
relies on estimates of uncertainty around the baseline scenario and
such estimates are in practice model-dependent.
4.2.2 Setting Variables in the Severely Adverse Scenario
(a) Generally, the variables in the severely adverse scenario
will be specified to be consistent with their expected behavior in
severe recessions or periods of market stress. The approach for
specifying the paths of these variables in the scenario will reflect
the Board's assessment of:
(1) how economic models suggest that these variables should
evolve given the path of the unemployment rate,
(2) how these variables have typically evolved in past U.S.
recessions or other relevant periods of significant stress in
particular asset markets, and
(3) other relevant factors, including the current state of the
economy; the level of vulnerabilities in the financial system; and
consensus estimates of long-run equilibrium values of potential GDP,
interest rates, and inflation expectations.
(b) For certain variables subject to guides that provide a range
or potential values (BBB spread, VIX, commercial real estate prices,
and mortgage rate), the Board expects that it could be appropriate
to set the paths for these variables at similar levels of severity.
In making this determination, the Board would consider the expected
severity of the unemployment rate and house prices variables and the
prevailing macroeconomic and financial conditions described in the
baseline scenario.
(c) The expected trajectories for the variables related to
unemployment, long-term bond yields and spreads, asset prices, and
volatility will be informed by quantitative guides. These guides
provide plausible ranges within which the Board expects to choose
the level of the peak or trough that each of these variables will
reach in the scenario, the timing of that peak or trough, the value
of the variable at the end of the scenario, and the trajectory of
the variable between the starting and ending value. The Board's
choices within those ranges will be informed by the factors listed
in section (a), above.
(d) Economic models--such as medium-scale macroeconomic models--
should be able to generate plausible paths consistent with the
unemployment rate for a number of scenario variables, such as real
GDP growth, CPI inflation, and short-term interest rates, which have
relatively stable (direct or indirect) relationships with the
unemployment rate (e.g., Okun's Law, the Phillips Curve, and
interest rate feedback rules). The Board has developed a model
specifically structured and calibrated to the needs of the stress
testing program to inform the trajectories of these variables (as
well as disposable personal income, or DPI), a description of which
will be maintained on the Board's website. The output of this model
is not a forecast of the Federal Reserve.
(e) In addition, judgment is necessary in projecting the path of
a scenario's international variables. Recessions that occur
simultaneously across countries are an important source of stress to
the balance sheets of companies with notable international exposures
but are not a typical feature of the international economy even when
the U.S. is in recession. As a result, simply adopting the typical
path of international variables in a severe U.S. recession would
likely underestimate the risks stemming from the international
economy. Consequently, an approach that uses both judgment and
economic models informs the path of international variables.
(f) The Board expects that the variables described in this
section 4.2.2 will be specified in the annual scenarios in the
severely adverse scenario to be consistent with the guides for each
variable below. In designing these guides and setting the values for
the variables in the severely adverse scenario, the Board will
consider the following scenario design principles:
(1) Severity: The scenarios should be sufficiently severe.
Severity is an important component in ensuring that covered
companies are adequately capitalized against a hypothetical severe
recession and in maintaining the public credibility of stress tests.
In determining the adequate level of severity for these guides, the
principle of severity requires that, at times, variable paths may
exceed levels observed in the historical data. Since no single
scenario can account for all potential contingencies, the scenario
must be sufficiently severe to ensure that banks will be resilient
to a range of alternative and plausible scenarios that could
generate net losses that are of similar magnitudes, even if such
scenarios would have different characteristics from the single
annual scenario. In establishing a sufficiently severe scenario, the
Board considers the potential unintended effects of the scenario on
the operations of firms subject to the stress tests.
(2) Credibility: The scenarios should be credible. Credible
stress tests maintain the confidence of the public and financial
markets that the stress tests are sufficiently severe to ensure that
the firms are properly capitalized to withstand severe economic and
financial conditions.
(3) Avoiding adding procyclicality: The scenarios should avoid
adding sources of procyclicality. If stress tests are relatively
more severe in already stressed conditions, then this severity could
add undue stress to the financial system, reducing financial
intermediation with negative implications for the macroeconomy. The
stress tests should balance the need for an adequately severe
scenario without magnifying existing procyclical tendencies in the
financial system.
(4) Flexibility: While the Board's scenario design framework
promotes transparency and predictability, fixed guides often would
fail to achieve at least one of the Board's goals of severity,
credibility, and not adding to procyclicality, as well as the
principles established in the Board's Stress Testing Policy
Statement.\299\ As a result, the Board has designed guides in this
section that generally establish ranges of historically observed
values that can be selected for a given severely adverse scenario,
while also enabling the Board to consider unexpected shocks that may
have implications for the economy and the financial stability of the
United States, and therefore, firms' future financial condition.
Further, flexibility is important to enable the Board to implement
reasonable technical adjustments to the values and trajectories of
the variables, consistent with these scenario design principles.
---------------------------------------------------------------------------
\299\ 12 CFR 252, Appendix B.
---------------------------------------------------------------------------
(g) The guides described in this section set out paths for each
variable over the 13 quarters in the severely adverse scenario. The
stress test requires projections of 13 quarters' worth of losses to
determine capital ratios at the end of 9 quarters of the scenario,
because loss provisions in quarter 9 are affected by bank
performance in quarters 10 to 13. To describe these paths, most
guides adopt a simple framework involving the following 4
parameters:
(1) the jump-off value: the value of the variable in the quarter
preceding the scenario. The jump-off value will be set to reflect
the conditions at the time that the scenario is designed.
(2) the peak or trough value: the paths in each guide specify
that each variable in the scenario will either increase or decrease
from its jump-off value. If the variable increases, it will reach a
maximum or peak value during the scenario. If it decreases, it will
reach a minimum or trough value during the scenario.
(3) the timing of the peak or trough: the quarter of the
scenario in which the variable path reaches its peak or trough.
(4) the trajectory from jump-off to peak or trough: the values
between the jump-off and peak or trough will be determined with a
roughly linear interpolation, a nonlinear function, or by specifying
the proportion of the change from jump-off to peak or trough that
will obtain in each of the intervening quarters.
(h) The severely adverse scenario will also set out end values
and trajectories to end values. The end value is expected to
generally be consistent with the historical values of a given
variable within a 10 to 15 quarter window after the beginning of
either a recession or other identified stress event. The trajectory
from peak or trough to end value is expected to generally be
determined by a roughly linear interpolation. The trajectory from
the peak or trough to the end
[[Page 51947]]
value generally will be smooth for variables determined by guides
and follow the model path for modeled variables.
4.2.2.1 Setting the Unemployment Rate Under the Severely Adverse
Scenario
(a) The Board anticipates that the severely adverse scenario
will feature an unemployment rate peak value that increases between
3 to 5 percentage points from its jump-off value. However, if a 3 to
5 percentage point increase in the unemployment rate does not raise
the level of the unemployment rate to at least 10 percent, the path
of the unemployment rate in most cases will be specified so as to
raise the unemployment rate to at least 10 percent.
(b) The Board anticipates that the unemployment rate peak value
will occur between quarters 6 and 8 after the jump-off point for the
scenario. The trajectory to peak value is expected to experience
high initial changes with smaller subsequent changes quarter to
quarter.
4.2.2.2 Setting House Prices in the Severely Adverse Scenario
(a) In specifying the path for nominal house prices, the Board
will consider the ratio of the nominal house price index (HPI) to
nominal per capita DPI. The Board anticipates that the severely
adverse scenario will feature an HPI-DPI ratio that falls by at
least 25 percent, or enough to bring the ratio down to the trough
reached in the first quarter of 2012 after the 2007-2009 financial
crisis.
(b) The trough is expected to occur between quarter 8 and
quarter 10 after the jump-off quarter. The trajectory to trough
value is expected to experience 20 percent of the decline realized
in quarter 1 and another 20 percent of the decline in quarter 2 (40
percent in total), with a roughly linear trajectory to trough
thereafter.
4.2.2.3 Setting the BBB Spread for the Severely Adverse Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a BBB corporate spread value, defined as the difference
between the yield on BBB corporate bonds and the 10-year Treasury
yield, that increases to the higher of (1) between a spread level of
500 to 600 basis points, or (2) a total increase of about 100 basis
points from the jump-off value.
(b) The Board anticipates that the BBB spread peak value will
occur between quarters 3 and 4 after the jump-off point for the
scenario. The trajectory to peak value is expected to experience the
highest share of the increase in quarters 1 and 2, with between 60
and 80 percent of the increase in quarter 1, followed by a smooth
trajectory to peak thereafter.
4.2.2.4 Setting the Mortgage Rate for the Severely Adverse Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a mortgage rate spread value, relative to the 10-year
Treasury yield, that increases between 70 to 160 basis points from
its initial level. The initial level will be set based on the
conditions at the time that the scenario is designed. However, if a
70 to 160 basis point increase in the mortgage rate spread does not
raise the level of the mortgage rate spread to at least 280 basis
points, the path of the mortgage rate spread in most cases will be
specified so as to raise the mortgage rate spread to at least 280
basis points.
(b) The Board anticipates that the mortgage rate spread peak
value will occur between quarters 3 and 4 after the jump-off point
for the scenario. The trajectory to peak value is expected to
experience between 50 and 70 percent of the increase realized in
quarter 1, with a smooth trajectory to peak thereafter.
4.2.2.5 Setting the VIX for the Severely Adverse Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a VIX peak value that will increase to a level between
65 and 75 percent or by at least 10 percentage points from the jump-
off value, whichever results in a higher level.
(b) The Board anticipates that the VIX peak value will occur in
quarter 2 after the jump-off point for the scenario. The trajectory
to peak value is expected to experience the largest share of the
increase, of 60 to 80 percent, in quarter 1.
4.2.2.6 Setting Equity Prices for the Severely Adverse Scenario
(a) The Board anticipates that the severely adverse scenario
will feature an equity price value that falls by around 50 percent
plus or minus up to 10 percent, depending on the performance of
equity prices over the 12-month period prior to the jump-off value.
When equity prices have risen over the past 12 months, equity prices
will fall to a trough level below the jump-off value of 50 percent
plus one half of the percentage increase in equity prices up to a
maximum of 10 percent. When equity prices have decreased over the
past 12 months, equity prices will fall to a trough level below the
jump-off value of 50 percent minus one half of the percentage
decrease in equity prices, up to a maximum of 10 percent. Thus, the
equity prices reach a trough level of between 40 and 60 percent
below the jump-off value.
(b) The Board anticipates that the equity price trough value
will occur in quarter 3 or 4 after the jump-off point for the
scenario. The trajectory to trough value is expected to experience
the highest share of the decrease, 60 to 70 percent, in quarter 1,
with 10 to 20 percent of the decline occurring in quarter 2 and the
remaining decline realized about equally in the remaining quarter(s)
to the trough value.
4.2.2.7 Setting CRE Prices for the Severely Adverse Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a CRE price value that falls between 30 and 45 percent
from its jump-off value.
(b) The Board anticipates that the CRE trough value will occur
between 8 and 10 quarters after the jump-off value for the scenario.
The trajectory to trough value is expected to be roughly linear.
4.2.2.8 Setting the 5-Year Treasury Yield for the Severely Adverse
Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a 5-year Treasury yield value that falls between 1.5
and 3.5 percentage points from its jump-off value, subject to a
lower bound of 0.3 percent, or a decline of 0.25 percent from the
jump-off level, whichever is lower.
(b) The Board anticipates that the 5-year Treasury yield trough
value will occur between 1 and 4 quarters after the jump-off value
for the scenario. The trajectory to trough value is expected to
experience the highest share of the decrease in quarter 1, depending
on the quarter that the trough value will occur, such that the share
of the decrease in quarter 1 will be between 55 percent and 100
percent. If the trough value is set to occur in quarters 2, 3, or 4,
the yield decline trajectory following quarter 1 will decrease
smoothly to the trough quarter.
4.2.2.9 Setting the 10-Year Treasury Yield for the Severely Adverse
Scenario
(a) The Board anticipates that the severely adverse scenario
will feature a 10-year Treasury yield value that falls between 1 and
3 percentage points from its jump-off value, subject to a lower
bound of 0.5 percent, or a decline of 0.25 percent from the jump-off
level, whichever is lower.
(b) The Board anticipates that the 10-year Treasury yield trough
value will occur between 1 and 4 quarters after the jump-off value
for the scenario. The trajectory to trough value is expected to
experience the highest share of the decrease in quarter 1, depending
on the quarter that the trough value will occur, such that the share
of the decrease in quarter 1 will be between 55 percent and 100
percent. If the trough value is set as quarters 2, 3, or 4, the
yield decline trajectory following quarter 1 will decrease smoothly
to the trough quarter.
4.2.2.10 Setting the Calibration of International Variables
(a) The Board expects to calibrate values for certain
international variables in the euro area, the United Kingdom, Japan,
and Developing Asia.
(b) For the euro area, the Board expects in general to specify
that GDP will decline by 7.5 percent from the baseline value to its
trough in the scenario, and reach an end value of 7.5 percent below
the baseline value. However, the Board may choose a value for the
decline in GDP between 5 and 10 percent. The Board expects to
specify that euro area inflation will decline by 3 percentage points
from the baseline scenario to its trough, and reach an end value of
0 percentage points below the baseline value. However, the Board may
choose a value for the decline in inflation between 2 and 4
percentage points. The Board expects to specify that the U.S. dollar
will appreciate against the euro by approximately 15 percent from
its jump-off value at its peak and then revert back to the jump-off
value by the end of the scenario. However, the Board may choose a
value for U.S. dollar appreciation between 5 and 25 percent(c).
For the United Kingdom, the Board expects in general to specify
that GDP will decline by 7.5 percent from the baseline value to its
trough in the scenario, and reach an end value of 7.5 percent below
the baseline value. However, the Board may choose a value for the
decline in GDP between 5 and 10 percent. The Board expects to
specify that inflation will decline by 3 percentage points from the
[[Page 51948]]
baseline value to its trough, and reach an end value of 0 percentage
points below the baseline value. However, the Board may choose a
value for the decline in inflation between 2 and 4 percentage
points. The Board expects to specify that the U.S. dollar will
appreciate against the Great Britain Pound by 15 percent from its
jump-off value at its peak and then revert back to the jump-off
value by the end of the scenario. However, the Board may choose a
value for U.S. dollar appreciation between 5 and 25 percent.
(d) For Japan, the Board expects in general to specify that GDP
will decline by 7.5 percent from the baseline value to its trough in
the scenario, and reach an end value of 7.5 percent below the
baseline value. However, the Board may choose a value for the
decline in GDP between 5 and 10 percent. The Board expects to
specify that inflation will decline by 3 percentage points from the
baseline value to its trough, and reach an end value of 0 percentage
points below the baseline value. However, the Board may choose a
value for the decline in inflation between 2 and 4 percentage
points. The Board expects to specify that U.S. dollar will
depreciate against the Japanese yen by 1 percent from its jump-off
value at its peak and then revert back to the jump-off value by the
end of the scenario. However, the Board may choose a value for
change in value of the U.S. dollar against the Japanese yen ranging
from a 9 percent depreciation to an 11 percent appreciation.
(e) For Developing Asia, the Board expects in general to specify
that GDP will decline by 3 percent from the baseline value to its
trough, and reach an end value of 0 percent below the baseline
value. However, the Board may choose a value for the decline in GDP
between 0.5 and 5.5 percent. The Board expects to specify that
inflation will decline by 5 percentage points from the baseline
value to its trough, and reach an end value of 0 percentage points
below the baseline value. However, the Board may choose a value for
the decline in inflation between 0.8 and 9 percentage points. The
Board expects to specify that the U.S. dollar will appreciate
against the currencies in Developing Asia by 15 percent from its
jump-off value at its peak and then revert back to the jump-off
value by the end of the scenario. However, the Board may choose a
value for the appreciation of the U.S. dollar between 5 and 25
percent.
4.2.3 Adding Salient Risks to the Severely Adverse Scenario
(a) The severely adverse scenario will be developed to reflect
specific risks to the economic and financial outlook that are
especially salient but that would feature minimally in the scenario
if the Board were to use only approaches that looked to past
recessions or relied on historical relationships between variables.
(b) There are some important instances in which it will be
appropriate to augment the recession approach with salient risks.
For example, if an asset price were especially elevated and thus
potentially vulnerable to an abrupt and potentially destabilizing
decline, it would be appropriate to include such a decline in the
scenario even if such a large drop were not typical in a severe
recession. Likewise, if economic developments abroad were
particularly unfavorable, assuming a weakening in international
conditions larger than what typically occurs in severe U.S.
recessions would likely also be appropriate.
(c) Clearly, while the recession component of the severely
adverse scenario is within some predictable range, the salient risk
aspect of the scenario is far less so, and therefore, needs an
annual assessment. Each year, the Board will identify the risks to
the financial system and the domestic and international economic
outlooks that appear more elevated than usual, using its internal
analysis and supervisory information and in consultation with the
Federal Deposit Insurance Corporation (FDIC) and the Office of the
Comptroller of the Currency (OCC). Using the same information, the
Board will then calibrate the paths of the macroeconomic and
financial variables in the scenario to reflect these risks.
(d) The Board will factor in particular risks to the domestic
and international macroeconomic outlook identified by its
economists, bank supervisors, and financial market experts and make
appropriate adjustments to the paths of specific economic variables.
These adjustments will not be reflected in the general severity of
the recession and, thus, all macroeconomic variables; rather, the
adjustments will apply to a subset of variables to reflect co-
movements in these variables that are historically less typical. The
Board plans to discuss the motivation for the adjustments that it
makes to variables to highlight systemic risks in the narrative
describing the scenarios, which will be released for public comment
and subsequently adjusted, if necessary, in response to those
comments.\300\
---------------------------------------------------------------------------
\300\ The means of effecting an adjustment to the severely
adverse scenario to address salient systemic risks differs from the
means used to adjust variables within the ranges specified by the
guides or the paths suggested by the macroeconomic model. For
example, in adjusting the scenario for an increased unemployment
rate, the Board would modify all variables such that the future
paths of the variables would be similar to how these variables have
moved historically in response to a change in the unemployment rate.
In contrast, to address salient risks, the Board may only modify a
small number of variables in the scenario and, as such, their future
paths in the scenario would be somewhat more atypical, but not
implausible, given existing risks.
---------------------------------------------------------------------------
5. Approach for Formulating the Market Shock Component
(a) This section discusses the approach the Board proposes to
adopt for developing the market shock component of the severely
adverse scenario appropriate for companies with significant trading
activities. The design and specification of the market shock
component differs from that of the macroeconomic scenarios because
profits and losses from trading are measured in mark-to-market
terms, while revenues and losses from traditional banking are
generally measured using the accrual method. As noted above, another
critical difference is the time-evolution of the market shock
component. The market shock component consists of a sudden ``shock''
to a large number of risk factors that determine the mark-to-market
value of trading positions, while the macroeconomic scenarios supply
a projected path of economic variables that affect traditional
banking activities over the entire planning period.
(b) The development of the market shock component that are
detailed in this section are as follows: baseline (subsection 5.1)
and severely adverse (subsection 5.2).
5.1 Approach for Formulating the Market Shock Component Under the
Baseline Scenario
Market shocks are large, previously unanticipated moves in asset
prices and rates. Under the baseline scenario, asset prices should,
broadly speaking, reflect consensus opinions about the future
evolution of the economy. Sudden price movements, as envisioned in
the market shock, should not occur along the baseline path. As a
result, the market shock will not be included in the baseline
scenario.
5.2 Approach for Formulating the Market Shock Component Under the
Severely Adverse Scenario
This section addresses possible approaches to designing the
market shock component in the severely adverse scenario, including
important considerations for scenario design, possible approaches to
designing scenarios, and a development strategy for implementing the
preferred approach.
5.2.1 Design Considerations for Market Shocks
(a) The general market practice for stressing a trading
portfolio is to specify market shocks either in terms of changes to
observable, broad financial market indicators and risk factors or
directly as changes to the mark-to-market values of financial
instruments.
(b) While the number of market shocks used in companies' pricing
and stress-testing models typically exceeds that provided in the
Board's scenarios, the number of market shocks in the Board's
scenarios allows for the consistency and comparability of market
losses across companies. However, the benefit from specifying a
large set of market shocks is at least partly offset by the
potential difficulty in creating shocks that are coherent and
internally consistent, particularly as the framework for developing
market shocks deviates from historical events. The Board's process
for generating the scenario market shocks has developed over time to
rely less on models and has expanded its use of simpler methods,
such as multipliers and mappings to modeled risk factors.
(c) Also, importantly, the ultimate losses associated with a
given market shock will depend on a company's trading positions,
which can make it difficult to rank order, ex ante, the severity of
the scenarios. In certain instances, market shocks that include
large market moves may not be particularly stressful for a given
company. Aligning the market shock with the macroeconomic scenario
for consistency may result in certain companies actually benefiting
from risk factor moves of larger magnitude in the market scenario if
the companies are hedging against salient risks to other parts of
their
[[Page 51949]]
business. Thus, the severity of market shocks must be calibrated to
take into account how a complex set of risks, such as directional
risks and basis risks, interacts with each other, given the
companies' trading positions at the time of stress. For instance, a
large depreciation in a foreign currency would benefit companies
with net short positions in the currency while hurting those with
net long positions. In addition, longer maturity positions may move
differently from shorter maturity positions, adding further
complexity.
(d) The sudden nature of market shocks and the early recognition
of mark-to-market losses add another element to the design of market
shocks, and to determining the appropriate severity of shocks. The
design of the market shocks must factor in appropriate assumptions
around the period of time during which market events will unfold and
any associated market responses.
(e) The design of market shocks includes calibration of shock
magnitudes based on assumed time horizons that reflect several
scenario design considerations. One consideration is the liquidity
characteristics of different asset classes. More specifically, the
calibration horizons reflect the variation in speed at which banks
could reasonably close out, or effectively hedge, the associated
risk exposures in the event of a market stress. The horizons are
generally longer than the typical times needed to liquidate
exposures under normal conditions because they are designed to
capture the unpredictable liquidity conditions that prevail in times
of stress. Another consideration is maintaining consistency between
the assumed time horizons used to calibrate market shocks and the
timeline for attributing the losses stemming from them.
Specifically, losses associated with the global market shock
component are attributed in one quarter of the stress test horizon,
which implies an upper bound of three months for calibrating the
shocks.
(f) Given these considerations, shock liquidity horizons are
chosen to be broadly consistent with the proposed standards in the
Fundamental Review of the Trading Book (FRTB). The horizons in the
FRTB are specified based on recommendations from consultations with
the financial industry and its regulators. The horizons in the FRTB
are therefore considered a reasonable benchmark for defining the
shock horizons used in the global market shock. The liquidity
horizons used in the market shock scenarios are not expected to be
perfectly matched with the FRTB liquidity horizons due to
granularity differences between the FRTB standards and the global
market shock template. The FRTB specifies horizons at a more
granular level, often using different horizons within each asset
class, whereas the Board uses the same liquidity horizon for all
market shocks within each asset class. Given these differences, the
global market shock scenario aims to align with the horizons
specified by the FRTB by using a weighted average of the FRTB
horizons within each asset class. The weights are determined using
aggregate firm exposures. For example, FRTB horizons for equity risk
factors vary between 10 and 60 business days, and the global market
shock horizon for this asset class is assumed to be 4 weeks. Because
the Board imposes an upper bound on global market shock horizons of
one quarter, there are cases where in which the range of FRTB
horizons is longer than the global market shock horizon. For
example, FRTB horizons for corporate credit market shocks vary
between 60 and 120 business days, but the Board uses a horizon of 3
months for corporate credit.
5.2.2 Approaches to Market Shock Design
(a) As an additional component of the severely adverse scenario,
the Board plans to use a standardized set of market shocks that
apply to all companies with significant trading activity. The market
shocks could be based on a single historical episode, hypothetical
(but plausible) events, or some combination of historical episodes,
with or without the addition of-hypothetical events (hybrid
approach). Depending on the type of hypothetical events, a scenario
based on such events may result in changes in risk factors that were
not previously observed.
(b) For the market shock component in the severely adverse
scenario, the Board plans to use the hybrid approach to develop
shocks. The hybrid approach allows the Board to maintain certain
core elements of consistency in market shocks each year while
providing flexibility to add hypothetical elements based on market
conditions at the time of the stress tests. In addition, this
approach will help ensure internal consistency in the scenario
because of its basis in historical episodes; however, combining the
historical episode and hypothetical events may require some
adjustments to ensure mutual consistency of the joint moves. In
general, the hybrid approach provides considerable flexibility in
developing scenarios that are relevant each year, and by introducing
variations in the scenario, the approach will also reduce the
ability of companies with significant trading activity to modify or
shift their portfolios to minimize expected losses in the severely
adverse market shock.
(c) The Board has considered a number of alternative approaches
for the design of market shocks. For example, the Board explored an
option of providing tailored market shocks for each trading company,
using information on the companies' portfolios gathered through
ongoing supervision, or other means. By specifically targeting known
or potential vulnerabilities in a company's trading position, the
tailored approach would be useful in assessing each company's
capital adequacy as it relates to the company's idiosyncratic risk.
However, the Board does not believe this approach to be well-suited
for the stress tests required by regulation. Consistency and
comparability are key features of annual supervisory stress tests
and annual company-run stress tests required in the stress test
rules. It would be difficult to use the information on the
companies' portfolios to design a common set of shocks that are
universally stressful for all covered companies. As a result, this
approach would be better suited to more customized, tailored stress
tests that are part of the company's internal capital planning
process or to other supervisory efforts outside of the stress tests
conducted under the capital rule and the stress test rules.
5.2.3 Development of the Market Shock
(a) Consistent with the approach described above, the market
shock component for the severely adverse scenario will incorporate
key elements of market developments during historical periods of
stress, and may include other price and rate movements in certain
markets that the Board deems to be plausible, though such movements
may not have been observed historically.
(b) The Board will identify potential market stress scenarios,
based on multiple sources of information, including financial
stability reports, supervisory information, and internal and
external assessments of market risks and potential flash points. The
hypothetical elements could originate from major geopolitical,
economic, or financial market events with potentially significant
impacts on market risk factors. The severity of these hypothetical
moves will likely be guided by similar historical events,
assumptions embedded in the companies' internal stress tests or
market participants, and other available information.
(c) Once broad market scenarios are agreed upon, the
implications for key risk factor groups will be defined. For
example, a scenario involving the failure of a large, interconnected
globally active financial institution could begin with a sharp
increase in credit default swap spreads and a precipitous decline in
asset prices across multiple markets, as investors become more risk
averse and market liquidity evaporates. These broad market movements
will be extrapolated to the granular level for all risk factors by
examining transmission channels and the historical relationships
between variables, though in some cases, the movement in particular
risk factors may be amplified based on theoretical relationships,
market observations, or the saliency to company trading books. If
there is a disagreement between the risk factor movements in the
historical event used in the scenario and the hypothetical event,
the Board will reconcile the differences by assessing a priori
expectations based on financial and economic theory and the
importance of the risk factors to the trading positions of the
firms.
6. Consistency Between the Macroeconomic Scenarios and the Market Shock
(a) As discussed earlier, the market shock comprises a set of
movements in a large number of risk factors that are realized in the
first quarter of the stress test horizon. Among the risk factors
specified in the market shock are several variables also specified
in the macroeconomic scenarios, such as short- and long-maturity
interest rates on Treasury and corporate debt, the level and
volatility of U.S. stock prices, and exchange rates.
(b) The market shock component is an add-on to the macroeconomic
scenarios that reflects abrupt market disruptions. As a result, the
market shock component may not always be directionally consistent
with the macroeconomic scenario. Because the market shock is
designed, in part, to mimic the effects of a sudden market
dislocation, while the macroeconomic scenarios are designed to
provide a description of the evolution of the real economy over two
or more years,
[[Page 51950]]
assumed economic conditions can move in significantly different
ways. In effect, the market shock can simulate a market panic,
during which financial asset prices move rapidly in unexpected
directions, and the macroeconomic assumptions can simulate the
severe recession that follows. Indeed, the pattern of a financial
crisis, characterized by a short period of wild swings in asset
prices followed by a prolonged period of moribund activity, and a
subsequent severe recession is familiar and plausible.
(c) As discussed in section 4.2.3, the Board may feature a
particularly salient risk in the macroeconomic assumptions for the
severely adverse scenario, such as a fall in an elevated asset
price. In such instances, the Board may also seek to reflect same
risk in the market shock. For example, if the macroeconomic scenario
were to feature a substantial decline in house prices, it may be
plausible for the market shock to feature a significant decline in
market values of any securities that are closely tied to the housing
sector or residential mortgages.
7. Timeline for Scenario Publication
(a) The Board will provide a final description of the
macroeconomic scenarios by no later than February 15. During the
period immediately preceding the publication of the scenarios, the
Board will collect and consider information from academics,
professional forecasters, international organizations, domestic and
foreign supervisors, and other private-sector analysts that
regularly conduct stress tests based on U.S. and global economic and
financial scenarios, including analysts at the firms. In addition,
the Board will consult with the FDIC and the OCC on setting the
guides in the scenarios. The Board expects to conduct this process
each year and disclose the developed scenarios for public comment.
The Board will update the scenarios, based on the public comments
and incoming macroeconomic data releases and other information.
(b) The Board expects to provide a broad overview of the market
shock component along with the macroeconomic scenarios. The Board
will publish the market shock templates by no later than March 1 of
each year, and intends to publish the market shock earlier in the
stress test and capital plan cycles to allow companies more time to
conduct their stress tests.
BILLING CODE 6210-01-P
[[Page 51951]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.072
[[Page 51952]]
[GRAPHIC] [TIFF OMITTED] TP18NO25.074
BILLING CODE 6210-01-C
* * * * *
Appendix B to Part 252--Stress Testing Policy Statement
0
19. To amend appendix B to part 252:
0
a. Add paragraph (a)(iv) to section 2.1;
0
b. Revise section 2.2;
0
c. Remove the text ``and non-public information about'' from section
3.1;
0
d. Revise paragraph (c) of section 3.2.
The revisions read as follows:
* * * * *
2.1 Soundness in Model Design
(a) During development, the Federal Reserve
(i) subjects supervisory models to extensive review of model
theory and logic and general conceptual soundness;
(ii) examines and evaluates justifications for modeling
assumptions;
(iii) tests models to establish the accuracy and stability of
the estimates and forecasts that they produce; and
(iv) invites, evaluates, and responds to substantive public
input on material model changes.
* * * * *
2.2. Disclosure of Information Related to the Supervisory Stress
Test
(a) In general, the Board does not disclose information related
to the supervisory stress test to covered companies if that
information is not also publicly disclosed. However, the Board will
generally provide additional information directly to a covered
company about such covered company's supervisory stress test
results, provided that the Board will only do so if it provides the
same type of information to all other covered companies
participating in the same stress test cycle.
(b) The Board has increased the breadth of its public disclosure
since the inception of the supervisory stress test to include
comprehensive descriptions of the supervisory stress models, changes
to those models, and, for each supervisory stress test cycle, more
information about model changes and key risk drivers, in addition to
more detail on different components of projected net revenues and
losses. Increasing public disclosure can help the public understand
and interpret the results of the supervisory stress test,
particularly with respect to the condition and capital adequacy of
participating firms. Providing additional information about the
supervisory stress test allows the public to make an evaluation of
the quality of the Board's assessment. This policy also promotes
consistent and equitable treatment of covered companies by ensuring
that institutions do not have access to information about the
supervisory stress test that is not also accessible to other covered
companies, corresponding to the principle of consistency and
comparability.
* * * * *
3.1. Structural Independence
* * * * *
(b) In addition, the Model Validation Council, a council of
external academic experts, provides independent advice on the
Federal Reserve's process to assess models used in the supervisory
stress test. In biannual meetings with Federal Reserve officials,
members of the council discuss selective supervisory models, after
being provided with detailed model documentation for those models.
The documentation and discussions enable the council to assess the
effectiveness of the models used in the supervisory stress tests and
of the overarching model validation program.
* * * * *
3.2. Technical Competence of Validation Staff
* * * * *
(c) The model validation program covers three main areas of
validation:
(1) Conceptual soundness;
(2) ongoing monitoring; and
(3) outcomes analysis.
[[Page 51953]]
Validation staff evaluates all aspects of model development,
implementation, and use, including but not limited to theory,
design, methodology, input data, testing, performance, documentation
standards, implementation controls (including access and change
controls), and code verification.
By order of the Board of Governors of the Federal Reserve
System.
Benjamin W. McDonough,
Deputy Secretary of the Board.
[FR Doc. 2025-20211 Filed 11-17-25; 8:45 am]
BILLING CODE 6210-01-P