CCAR and CECL: Differences, Stress Test Feed, and 2026 Changes

CCAR and CECL do different jobs. The Comprehensive Capital Analysis and Review is a Federal Reserve stress test that checks whether a large bank holds enough capital to keep lending through a hypothetical severe recession. The Current Expected Credit Loss standard is an accounting rule that requires banks to reserve today for losses they expect over the full remaining life of every loan on the books. They intersect because the reserve a bank calculates under CECL becomes an input to the stress test, and CECL’s forward-looking mechanics tend to push projected losses earlier and higher in a downturn scenario, which can enlarge the capital cushion the Fed requires the bank to hold.

What CCAR Measures

CCAR is the Federal Reserve’s annual assessment of capital adequacy at the largest bank holding companies, savings and loan holding companies, and intermediate holding companies of foreign banking organizations. It is codified in the capital plan rule at 12 CFR 225.8, and it applies to institutions with $100 billion or more in total consolidated assets.1eCFR. 12 CFR 225.8 – Capital Planning and Stress Capital Buffer Requirement Covered banks submit capital plans each April projecting revenues, losses, reserves, and capital levels over at least nine consecutive quarters.

The Fed designs hypothetical macroeconomic scenarios containing specific paths for unemployment, GDP, interest rates, and asset prices. The severely adverse scenario is the one that drives capital requirements: it models a deep recession and measures how far a bank’s common equity tier 1 (CET1) ratio would fall.2Board of Governors of the Federal Reserve System. 2026 Stress Test Scenarios CCAR once had a qualitative component that let the Fed reject a plan on governance grounds, but that objection was eliminated for most firms in 2019, leaving CCAR overwhelmingly a numbers exercise.3Board of Governors of the Federal Reserve System. Federal Reserve Board Announces It Will Limit the Use of the Qualitative Objection in CCAR

The primary output is the Stress Capital Buffer (SCB), a firm-specific requirement that sits on top of the 4.5% CET1 minimum. To calculate it, the Fed takes the bank’s starting CET1 ratio, subtracts the lowest projected CET1 ratio at any point during the stress horizon, and adds four quarters of planned common stock dividends scaled by risk-weighted assets at the trough. The result, or 2.5%, whichever is greater, becomes the SCB.1eCFR. 12 CFR 225.8 – Capital Planning and Stress Capital Buffer Requirement A bank whose actual CET1 ratio slips into the SCB zone faces automatic restrictions on dividends and share buybacks.

What CECL Measures

CECL, codified as Accounting Standards Codification Topic 326, was issued by the Financial Accounting Standards Board in 2016 and replaced the older “incurred loss” model.4National Credit Union Administration. CECL Accounting Standards The old rule delayed recognition until a loss was probable and estimable, which tended to bunch write-downs into downturns. CECL requires banks to estimate expected credit losses over the full remaining life of each financial asset from the moment it is originated or acquired.

The estimate draws on three categories of information: historical loss experience, current economic conditions, and reasonable and supportable forecasts of the future. Beyond the forecast horizon, the bank must revert to historical loss rates without adjusting them for expected conditions.5FASB. FASB Staff Q&A – Topic 326, No. 2 The output is the Allowance for Credit Losses (ACL), a balance sheet reserve that offsets the amortized cost of the loan portfolio. Each reporting period, the bank compares the current estimate to the existing allowance and books the difference as credit loss expense in net income.6FASB. ASU 2025-08 – Financial Instruments – Credit Losses (Topic 326) Because that expense runs through the income statement, it reduces retained earnings and, in turn, regulatory capital.

CECL is principles-based. Banks may use discounted cash flow methods, loss-rate methods, roll-rate methods, probability-of-default models, aging schedules, or combinations of these, and they can use different approaches for different asset classes. Two banks with similar loan books can produce meaningfully different allowances based on model choice and forecast assumptions.

Large SEC-reporting banks adopted CECL for fiscal years beginning after December 15, 2019.7Board of Governors of the Federal Reserve System. Frequently Asked Questions on the New Accounting Standard on Financial Instruments – Credit Losses Because CECL generally produces a larger day-one allowance than the incurred loss model, regulators allowed institutions to phase in the capital impact over up to five years, with a full offset for the first two years followed by a step-down at 75%, 50%, and 25%.8Federal Register. Regulatory Capital Rule – Revised Transition of the Current Expected Credit Losses Methodology That transition has now ended for the largest banks, so CECL’s full capital impact is embedded in their reported numbers.

The Core Differences

The most fundamental gap is purpose. CCAR asks whether a bank can survive a hypothetical catastrophe. CECL asks how much a bank expects to lose on loans it has already made. One is a supervisory resilience test; the other is an accounting measurement embedded in the financial statements.

The authorities differ. CCAR is a Federal Reserve exercise grounded in banking supervision.9Board of Governors of the Federal Reserve System. Comprehensive Capital and Analysis Review and Dodd-Frank Act Stress Tests – Questions and Answers CECL is a FASB standard that applies to every entity reporting under U.S. GAAP, not just banks.

The horizons differ. CCAR projects capital over at least nine consecutive quarters using scenarios the Fed designs. CECL looks across the full remaining contractual life of each loan, which can run 30 years for a residential mortgage. CCAR uses the Fed’s severely adverse scenario as its benchmark. CECL relies on management’s own reasonable and supportable forecast, then reverts to unadjusted historical loss rates once that window closes.5FASB. FASB Staff Q&A – Topic 326, No. 2

The frequency differs. CCAR is an annual cycle with a single submission and a single set of results. CECL runs continuously; the allowance is remeasured every reporting period.

How CECL Feeds Into the Stress Test

The Allowance for Credit Losses that a bank calculates under CECL is a starting input to the CCAR stress projection. When the Fed models a severely adverse scenario, it projects how that allowance would evolve as the hypothetical economy deteriorates. Because CECL requires lifetime loss estimates that respond to forecasted conditions, the allowance climbs sharply as the outlook darkens across the entire loan book at once, rather than easing up as individual loans reach the threshold of probable loss.

The result is a front-loading effect. Under the old incurred loss model, provisions built gradually because a loss had to be virtually certain before recognition. Under CECL, the bank must increase its allowance the moment the forward-looking forecast worsens, even before defaults materialize. In a stress test, provision expense surges earlier and harder, dragging projected net income down and eroding the CET1 ratio faster than the older accounting would have.

That flows straight into the Stress Capital Buffer. The SCB is driven by the gap between the starting CET1 ratio and the lowest projected CET1 ratio during the stress horizon. A sharper, earlier decline widens that gap, producing a larger SCB. The effect is most pronounced at banks with heavy exposure to long-duration assets like mortgages, where lifetime loss estimates are especially sensitive to economic assumptions.

One consequence is that CCAR results have become more volatile. A bank’s SCB can move meaningfully year to year based on scenario severity, changes in loan mix, and how CECL’s forward-looking mechanics interact with the Fed’s hypothetical downturn. An accounting standard designed for transparency and a supervisory test designed for resilience now feed each other in ways neither framework fully anticipated.

What Is Changing in 2026 and 2027

For the 2026 cycle, the Federal Reserve voted to freeze existing SCB requirements rather than update them with new stress test results. The freeze runs until 2027, when the Fed expects to have incorporated public feedback on its supervisory models before recalculating.10Board of Governors of the Federal Reserve System. Federal Reserve Board Finalizes Hypothetical Scenarios for Its 2026 Supervisory Stress Test

The freeze is part of a broader transparency proposal published in late 2025. The Fed proposed to publish comprehensive stress test model documentation each year, invite public comment on material model changes before implementing them, and respond to substantive feedback, a shift from the historical practice of keeping model details confidential to prevent gaming. The proposal also floats averaging a bank’s results across two consecutive years before setting the SCB, which would dampen the year-to-year volatility that CECL has helped produce. To fit a public comment period on scenarios, the stress test jump-off date would shift from December 31 to September 30, with proposed scenarios published by October 15 of the prior year.11Federal Register. Enhanced Transparency and Public Accountability of the Supervisory Stress Test Models and Scenarios The changes are still at proposal stage, and no new SCB math is expected before the 2027 cycle.