Guide

Fed bank stress tests explained

Harbor Credit Union’s ALM team extended a wholesale funding line to a mid-size regional bank in early 2025 after the institution “passed” the Federal Reserve’s annual stress test with a projected common equity tier 1 (CET1) ratio still above the regulatory minimum in the severely adverse scenario. Six months later, unrealized securities losses and deposit beta pressure pushed the bank’s credit spreads wider anyway — the line repriced 95 basis points and Harbor ate unexpected carry. Post-mortem: traders treated a regulatory pass as a forecast of stable market funding when stress tests are a standardized, lagging, model-driven exercise with explicit limitations on trading-book coverage and no promise about near-term spread paths.

After the refactor, Harbor reads CCAR and DFAST disclosures for capital distribution constraints and tail-loss magnitudes, then layers market-implied funding stress indicators on top. Counterparty limit breaches fell from 4 to 0 over two review cycles; wholesale rollover forecast error dropped from 18 to 5 basis points. This guide covers DFAST supervisory mechanics, CCAR capital planning review, scenario design, CET1 and the stress capital buffer (SCB), dividend and buyback gates, the Harbor Credit Union workflow, a technique decision table versus generic portfolio stress tests, pitfalls, and a production checklist.

What Fed bank stress tests are

Large U.S. bank holding companies (“large banks,” generally those with $100 billion or more in assets, with tiering for smaller but still significant firms) must demonstrate they can absorb severe macro and market shocks without failing or cutting critical lending. The Federal Reserve runs two related programs, usually announced together each June:

DFAST (Dodd-Frank Act Stress Test) — a supervisory exercise. The Fed publishes standardized scenarios (baseline, severely adverse, and sometimes exploratory) and projects each firm’s losses, revenues, expenses, and capital ratios using both Fed models and firm-submitted data. Results determine whether minimum capital requirements rise via the stress capital buffer.
CCAR (Comprehensive Capital Analysis and Review) — a qualitative and quantitative review of capital planning. Large banks submit their own stress scenarios, capital actions (dividends, buybacks), and internal models. The Fed can object to capital plans even when DFAST math looks adequate if governance, data, or planning process fails scrutiny.

Passing does not mean “no risk.” It means that under the Fed’s specified hypothetical paths for unemployment, GDP, rates, and asset prices, projected capital stays above regulatory floors in the model. Real shocks differ in timing, correlation, and instruments stressed.

Scenarios: baseline, severely adverse, and exploratory

Baseline

A moderate macro path consistent with consensus near-term forecasts. Useful for comparing firm projections to a common reference, not for tail-risk sizing.

Severely adverse

The headline scenario: deep recession, sharp equity decline, widened spreads, and rate paths designed to stress net interest income and securities portfolios simultaneously. The Fed publishes scenario variables quarterly in advance; banks run projections over a nine-quarter horizon. Key design choices that readers often miss:

Instantaneous vs path. Some market variables jump in quarter one; others evolve gradually. Funding cost spikes may lag loan losses — affecting projected CET1 trough timing.
Global market shock (GMS). A separate component for the largest trading firms, stressing market risk factors over a short horizon. Regional banks with small trading books see less GMS impact.
Counterparty default module. Models assumed default of largest counterparties for systemic trading firms; linkage to real-world bilateral exposure is indirect.

Exploratory (when published)

Additional scenarios testing structural risks — e.g., persistent inflation with higher rates, or commercial real estate concentration. Exploratory results typically do not set capital buffers but signal supervisory focus areas for the next cycle.

Capital ratios, SCB, and what “pass” means

Stress tests feed into capital requirements, not just disclosure. The ratio most cited in headlines is CET1 — core equity as a percent of risk-weighted assets. After DFAST, the Fed assigns a stress capital buffer (SCB) equal to the decline in projected CET1 under the severely adverse scenario (subject to floors and caps), replacing much of the old static capital conservation buffer for large banks.

A firm “passes” DFAST if projected CET1 in the severely adverse path stays above regulatory minimums (including the new SCB) throughout the horizon. CCAR adds a second gate: the Fed may object to planned capital distributions if cumulative dividends and buybacks would erode capital below stressed minima or if the planning process is deficient.

Compare stressed capital paths to bank reserves and liquidity coverage separately — stress tests are solvency-centric; March 2023 showed deposit runs can stress funding before modeled loan losses fully materialize. Also read the Fed balance sheet context: QT and reserve levels influence deposit pricing betas that stress models may update only annually.

CCAR vs DFAST: practical differences

Dimension	DFAST	CCAR
Primary question	Does standardized stress breach capital mins?	Is the capital plan credible and safe to distribute?
Scenarios	Fed-published baseline and severely adverse	Firm-designed scenarios plus Fed scenarios
Outputs	SCB, public aggregate results	Objection or non-objection to capital plan
Qualitative review	Limited	Governance, controls, model risk management
Market read-through	Capital buffer changes, loss severity	Dividend/buyback capacity, supervisory tone

Equity investors often focus on CCAR day buyback announcements; credit analysts should weight DFAST loss rates by asset class (commercial real estate, credit card, wholesale) and the implied SCB change year over year.

Calendar, disclosure, and market transmission

The typical annual cycle:

February — Fed releases scenario templates for the upcoming cycle.
April — banks submit capital plans and data.
June — Fed announces DFAST results, SCBs, and CCAR objections/non-objections; banks may announce revised distributions.

Transmission channels to watch:

Higher SCB → retained earnings need rises → lower dividend payout ratios or slower buybacks.
Large projected loan losses in a sector → supervisory scrutiny on concentration limits; can widen subordinated debt spreads even after a pass.
CCAR objection → immediate distribution freeze until plan resubmission; rare but high impact for preferred and AT1 holders.
Exploratory scenario pain → forward indicator of next year’s severely adverse design tweaks.

Harbor Credit Union now tags each counterparty with last disclosed SCB delta, projected CRE loss rate in severely adverse, and three-month change in five-year CDS — stress pass alone no longer clears limit increases.

Harbor Credit Union refactor (worked example)

After the 2025 funding-line miss, Harbor rebuilt its bank counterparty framework:

Primary screen: market spreads (CDS, sub debt, TED-style wholesale proxies) updated weekly.
Regulatory overlay: DFAST projected CET1 trough, SCB change vs prior year, and CCAR distribution headroom as scenario bounds.
Limit sizing: exposure capped at 40% of the lower of stressed min capital headroom (implied) and internal rating ceiling.
June event protocol: no limit increases for 30 days post-release; reassess after Q2 10-Q reconciles modeled vs realized NIM.
Early warning: flag any bank where exploratory CRE losses exceed severely adverse by more than 20% (model divergence signal).

Outcomes (internal benchmarks, H1 2026): wholesale rollover forecast error 18 bp → 5 bp; zero surprise line repricings; counterparty committee review time down 30%. Regulatory stress results remain essential — as slow-moving capital constraints, not real-time spread forecasts.

Technique decision table

Your goal	Use Fed stress tests	Prefer instead
Size bank counterparty limits	SCB and CET1 trough as ceiling	Market CDS and funding spreads alone
Forecast near-term dividend cuts	CCAR objection history and payout ratios	DFAST loss rates only
Estimate CRE tail risk at a bank	Severely adverse CRE loss disclosure	Baseline scenario
Trade CCAR release day	Buyback surprise vs consensus	Assuming all passes are equal
Stress your own investment portfolio	Inspiration for macro paths	Bank DFAST numbers directly
Detect liquidity run risk	Supplement with LCR and deposit data	DFAST pass as funding safety
Compare global banks	U.S. DFAST is not comparable to EBA tests	Harmonized metrics (CET1, TLAC) with local caveats

Common pitfalls

Pass equals safe credit — models use lagged data; market stress can front-run supervisory cycles.
Ignoring SCB changes — a pass with a higher buffer is tighter capital policy than last year.
Baseline confusion — baseline paths are not tail scenarios; do not size hedges from them.
Trading-book blind spot — regional banks with securities AOCI losses may face market pressure DFAST projects slowly.
CCAR vs DFAST conflation — distribution objections are CCAR; capital minimums are DFAST.
Static scenario extrapolation — next year’s severely adverse CRE shock may be sharper after exploratory results.
Deposit beta optimism — firm-submitted assumptions can flatter NIM under stress; compare across peers.

Production checklist

Bookmark Fed stress test hub and annual scenario publication dates.
Archive each bank’s DFAST CET1 trough, SCB, and loss rates by asset class.
Log CCAR objection/non-objection and planned buyback changes on release day.
Plot SCB delta year over year for top ten counterparties.
Overlay market CDS changes 90 days before and after June results.
Compare severely adverse CRE losses to firm 10-K concentration disclosures.
Read exploratory scenario sections for next-cycle design hints.
Reconcile DFAST NIM assumptions with realized deposit betas quarterly.
Stress liquidity separately using LCR and uninsured deposit shares.
Flag firms where internal stress losses exceed Fed severely adverse.
Review qualitative CCAR deficiencies in Fed press releases.
Update counterparty limits only after market and regulatory screens align.

Key takeaways

DFAST is a standardized supervisory stress test that sets capital buffers; CCAR reviews whether capital plans and distributions are safe.
Passing means modeled CET1 stays above floors in Fed scenarios — not that market funding spreads will stay tight.
The stress capital buffer translates projected losses into higher ongoing capital requirements.
June release day moves buyback expectations; credit analysts should focus on loss severity and SCB changes.
Harbor Credit Union cut wholesale forecast error from 18 to 5 bp by pairing stress disclosures with live funding stress indicators.