Stress Testing in Finance: Scenarios, Capital Buffer, and Rules

Financial stress testing is a regulatory exercise in which the Federal Reserve simulates a severe economic crisis and measures whether each large bank would still have enough capital to keep operating and lending through it. In the 2025 cycle, 22 banks absorbed $549 billion in projected losses under the Fed’s severely adverse scenario and their combined capital cushion held above the required minimum.¹ For the 2026 cycle, 32 firms are on the participant list.²

Why the Test Exists

Before 2008, bank supervision worked mostly by looking backward at recent financial statements. The financial crisis exposed the flaw: banks that looked healthy on paper were destroyed by losses no one had planned for. Stress testing reverses the lens. Instead of asking how the bank did last year, it asks whether the bank could survive if a deep recession, a market crash, and a wave of loan defaults arrived at the same time.

The point isn’t to predict a downturn. It’s to make sure that if one comes, the largest banks can keep lending to households and businesses without breaching capital minimums or needing a rescue.

Sensitivity Analysis Versus Scenario Analysis

Two analytical approaches sit under the “stress test” umbrella. Sensitivity analysis isolates one variable. An analyst might model a 2% jump in interest rates while everything else stays constant, which reveals how exposed a particular portfolio is to that single risk.

Scenario analysis moves everything at once. Real recessions don’t arrive as isolated shocks; unemployment rises, equity markets fall, home values drop, and corporate defaults climb together. The Fed’s supervisory stress test is a scenario analysis, built around a coherent macroeconomic story severe enough to expose hidden weaknesses.

What the Scenarios Look Like

Every year the Fed publishes at least two hypothetical environments: a baseline that tracks consensus forecasts, and a severely adverse scenario that simulates a deep recession. The statute requires at least those two.³ The severely adverse scenario does the heavy lifting. For 2026, its parameters include:

Unemployment rising 5.5 percentage points, from 4.5% to a peak of 10% in the third quarter of 2027
GDP declining 4.8% from the fourth quarter of 2025 to a mid-2027 trough
Equity prices falling roughly 54% over the first three quarters
House prices dropping 29%
Commercial real estate declining 40%
The VIX peaking at 72% in the second quarter of 2026

The Fed calibrates these numbers under a published policy framework so that severity stays meaningful regardless of where the economy sits at the start of the cycle.⁴ The 2025 version assumed an even sharper GDP decline of 7.8% and a 50% equity drop.⁵

Banks with large trading operations face two extra layers on top of the macro scenario: a global market shock applied to their trading positions and a simulated default of their largest counterparty. Projections run for 13 quarters, though the capital ratios that determine outcomes are measured through the first nine.⁶

Which Banks Are Tested

Any U.S. bank holding company or intermediate holding company with average total consolidated assets of $100 billion or more is subject to the supervisory stress test. Once a firm crosses that line it stays in unless its assets fall below the threshold for four consecutive quarters.⁷

Frequency depends on size and complexity. Global systemically important banks and Category II and III firms are tested every year. Category IV firms are tested every other year, in even-numbered years.⁷ The 2026 participant list of 32 firms includes all eight U.S. G-SIBs along with intermediate holding companies of foreign banks operating here, such as Barclays US, DB USA, HSBC North America, and UBS Americas.²

Firms holding more than $250 billion in total consolidated assets must also run their own internal stress tests and submit the results to their primary regulator under the statute.³

How the Fed Actually Runs the Test

A common misconception is that banks calculate their own results and submit them. They don’t. Banks send in detailed portfolio data, and the Federal Reserve runs the projections through its own independently developed models. Those models are calibrated on industry-wide data and applied uniformly, so differences across firms come from differences in portfolios, not from differences in modeling assumptions.⁸

The Fed generally does not tailor projections firm by firm or apply firm-specific overlays. Uniformity is the design principle: one yardstick for everyone, so banks can’t game favorable assumptions into their own numbers.⁸ In the limited cases where supervisory data can’t support an independent estimate, the Fed may fall back on firm-provided figures or third-party data.

Passing or Failing: Capital Ratios

The main metric is the Common Equity Tier 1 (CET1) capital ratio. CET1 is a bank’s highest-quality capital, mostly common shares and retained earnings that are permanently available to absorb losses.⁹ Under U.S. rules that follow Basel III, every bank must hold at least 4.5% CET1 against its risk-weighted assets at all times.¹⁰

The Tier 1 leverage ratio provides a cross-check by dividing Tier 1 capital by average total assets without any risk weighting. That catches situations where a portfolio looks low-risk on paper but concentrates exposure in practice.

During the test, the Fed projects each bank’s CET1 ratio at its lowest point over the nine-quarter window. In 2025, the 22 participating banks entered with an aggregate CET1 ratio of 13.4% and dropped to a projected minimum of 11.6% under the severely adverse scenario, well above the 4.5% floor.¹ Breaching the minimum would signal a serious shortfall and trigger regulatory consequences.

The Stress Capital Buffer

Before 2020, the Fed used stress test results to approve or reject each bank’s plans for dividends and stock buybacks by issuing an “objection” or “non-objection.” That pass/fail system was replaced by the stress capital buffer, which converts projected losses directly into an ongoing capital requirement.

The buffer is calculated by taking the projected decline in a bank’s CET1 ratio under the severely adverse scenario and adding four quarters of planned dividends. The result has a floor of 2.5%. It stacks on top of the 4.5% CET1 minimum, so every bank must maintain at least a 7.0% CET1 ratio, and banks with larger projected losses carry a higher buffer.¹¹

Distribution limits are now automatic. A bank that falls below its combined minimum-plus-buffer level faces graduated restrictions on dividends and discretionary bonuses, and a bank in the bottom quartile of its buffer cannot make any capital distributions without prior Board approval.¹² The mechanism replaced the once-a-year cliff where a bank either got approval or didn’t.

The Legal Framework

Supervisory stress testing sits in Section 165(i) of the Dodd-Frank Act, codified at 12 U.S.C. § 5365(i), which directs the Federal Reserve to run annual analyses of whether covered institutions hold enough capital to absorb losses under adverse conditions.³ The implementing regulation is 12 CFR Part 252, Subpart E. It defines who is tested, on what cycle, and what the Fed must publish, and it sets two hard dates: scenarios must be communicated to firms by February 15 each year, and a summary of results must be publicly disclosed by June 30.⁷

These supervisory tests operate alongside the Comprehensive Capital Analysis and Review (CCAR), the broader annual exercise assessing capital positions and planning practices under the capital plan rule at 12 CFR 225.8.¹³

Where the Framework Is Heading

The rules aren’t static. In November 2025, the Fed proposed opening its stress test models and scenario design methodology to public comment for the first time, a shift for a process that has historically operated as a closed system. The comment period closed in February 2026.¹⁴

A separate proposal would average the current year’s stress test results with the prior year’s when calculating each bank’s stress capital buffer, smoothing the year-over-year swings in capital requirements that banks have long complained about.¹¹ Whether either proposal takes its current form is still open, but the direction points toward a more transparent and more predictable system.

1
Board of Governors of the Federal Reserve System. Results for Banks Under the Severely Adverse Scenario
2
Board of Governors of the Federal Reserve System. 2026 Stress Test Scenarios
3
Office of the Law Revision Counsel. 12 USC 5365 – Enhanced Supervision and Prudential Standards
4
Board of Governors of the Federal Reserve System. 2026 Stress Test Scenarios
5
Board of Governors of the Federal Reserve System. 2025 Stress Test Scenarios
6
Federal Register. Enhanced Transparency and Public Accountability of the Supervisory Stress Test Models and Scenarios
7
eCFR. 12 CFR Part 252 Subpart E – Supervisory Stress Test Requirements
8
Board of Governors of the Federal Reserve System. Approach to Supervisory Model Development and Validation
9
Bank for International Settlements. Definition of Capital in Basel III – Executive Summary
10
Board of Governors of the Federal Reserve System. Annual Large Bank Capital Requirements
11
Federal Register. Modifications to the Capital Plan Rule and Stress Capital Buffer Requirement
12
Board of Governors of the Federal Reserve System. Amendments to the Regulatory Capital, Capital Plan, and Stress Buffer Requirements
13
Board of Governors of the Federal Reserve System. Comprehensive Capital Analysis and Review and Dodd-Frank Act Stress Tests: Questions and Answers
14
Federal Register. Enhanced Transparency and Public Accountability of the Supervisory Stress Test Models and Scenarios