What Is CECL (Current Expected Credit Loss)?
CECL, the current expected credit loss methodology, is the US GAAP standard for measuring credit losses on loans and other financial assets carried at amortized cost. Introduced by the Financial Accounting Standards Board in Accounting Standards Update 2016-13 and codified in ASC Topic 326, it requires institutions to estimate the credit losses they expect over the full life of an asset, rather than waiting until a loss has been incurred.
CECL is an accounting standard, but in practice it is also a data project. Estimating lifetime losses depends on years of loan-level history, current conditions, and forecasts, all of which have to be complete, consistent, and traceable for auditors and examiners.
CECL replaced the incurred loss model with an estimate of expected lifetime credit losses, based on historical loss experience, current conditions, and reasonable and supportable forecasts. It took effect for SEC filers that are not eligible to be smaller reporting companies in fiscal years beginning after December 15, 2019, and for all other entities in fiscal years beginning after December 15, 2022. CECL does not prescribe a single method, many methods rely on loan-level data, and US banking agencies note that, depending on the method, institutions may need to capture additional data and keep it longer. Governing that data, with defined critical data elements, lineage, and controls, is what makes the allowance defensible.
What Is CECL?
Under CECL, an institution records an allowance for credit losses when it originates or acquires a financial asset, equal to the credit losses it expects over the asset's contractual life. The estimate draws on three kinds of information, according to the interagency guidance from the Federal Reserve, FDIC, NCUA, and OCC:
- Past events, including historical credit loss experience on assets with similar risk characteristics.
- Current conditions, such as the present state of the portfolio and the economy.
- Reasonable and supportable forecasts of conditions that affect collectability. Beyond the period the institution can reasonably forecast, it reverts to historical loss experience.
CECL applies to financial assets measured at amortized cost, such as loans and held-to-maturity debt securities, as well as to certain off-balance-sheet credit exposures. It is not limited to banks: any entity reporting under US GAAP with such assets applies it, although banks and credit unions feel its impact most.
The standard does not specify a single method. Institutions may use any reasonable approach, and common choices include loss-rate, vintage, discounted cash flow, probability of default and loss given default, and remaining-life methods. The choice of method shapes which data the institution needs.
CECL vs the Incurred Loss Model
| Aspect | Incurred loss model | CECL |
|---|---|---|
| When losses are recognized | When a loss is probable and has been incurred | At origination or acquisition, for expected lifetime losses |
| Time horizon | Losses already incurred at the reporting date | The full contractual life of the asset |
| Information used | Mainly past events and current conditions | Past events, current conditions, and reasonable and supportable forecasts |
| Data needs | Often pool-level loss history | Historical data over the life of the assets, often at loan level |
| Typical effect on allowance | Lower, recognized later | Often higher, recognized earlier |
FASB introduced CECL after the financial crisis, when users of financial statements were concerned that the incurred loss model recognized credit losses too little and too late.
Who Must Comply and When
CECL has applied to all entities reporting under US GAAP since 2023. The original effective dates in ASU 2016-13 were staggered, and ASU 2019-10 later deferred them for most institutions:
| Entity type | Effective for fiscal years beginning after |
|---|---|
| SEC filers that are not eligible to be smaller reporting companies | December 15, 2019 |
| All other entities, including smaller reporting companies, private companies, and credit unions | December 15, 2022 |
For a calendar-year bank in the second group, that meant adopting CECL on January 1, 2023. Banks report the resulting allowance for credit losses in their financial statements and in regulatory reports such as the Call Report.
CECL Data Requirements
The interagency guidance is explicit that data is a central challenge. It says that, depending on the estimation method, institutions "may need to capture additional data and retain data longer than they have in the past on loans that have been paid off or charged off," and encourages them to discuss historical loss data with their lending, credit risk, and IT teams and with their core loan service providers.
The exact fields depend on the method. Pool-level methods need less detail, but many CECL processes rely on loan-level data such as:
- Loan characteristics: product type, origination date, maturity date, original amount, interest rate, and payment terms.
- Risk attributes: credit score, risk rating, loan-to-value, collateral type, industry, and geography, used to segment the portfolio.
- Performance history: balances over time, delinquency status, modifications and renewals, charge-off and recovery amounts and dates, and payoffs.
- Economic inputs: the forecast variables used in the model, such as unemployment or property prices, and the reversion assumptions.
This data usually lives in several places: the core banking system, loan servicing and collateral systems, loan files, and external economic forecasts. Bringing it together consistently, for loans that may have closed years ago, is where most CECL effort goes.
Data Lineage and Controls for CECL
The allowance for credit losses is a significant estimate in a bank's financial statements, so it sits inside the scope of internal control over financial reporting: under SOX for public companies, and under FDICIA's audit and reporting rules for insured banks above its asset thresholds. The model behind it is also subject to model risk management. Both require the same thing from the data: evidence of where it came from and how it was checked.
- Lineage. Document how each input flows from source systems into the CECL data layer and the model, at the level of individual columns where it matters, so a change in a source field can be traced to its effect on the allowance.
- Completeness and accuracy controls. Reconcile loan balances to the general ledger, check for missing loans and fields, and validate that charge-off and recovery history is complete.
- Definitions. Agree on what each key field means, such as when a loan counts as charged off or modified, and apply the same definition across systems and periods.
- Ownership. Name a business owner for each critical input, so data issues have someone accountable for resolving them.
- Retention. Keep the history needed for the chosen method, including closed and charged-off loans, and document where it is stored.
Common CECL Data Challenges
- Missing history. Older loans were often archived only at pool level or not at all, and system migrations broke the history.
- Inconsistent definitions. Different systems record default, modification, or charge-off in different ways.
- Vendor data. Core and servicing providers hold much of the history, and extracting it consistently takes coordination.
- Manual adjustments. Qualitative adjustments and spreadsheet steps are hard to trace unless they are documented as part of the data flow.
- Change over time. New products, acquisitions, and system changes alter the data each year, and the model has to keep up.
How Dawiso Supports It
Dawiso does not calculate the allowance. It provides the governance layer that makes the data behind it understandable and traceable.
- Shared definitions. The business glossary records terms and business indicators, such as charge-off or outstanding balance, with definitions, calculations, and named Owners and Stewards.
- Lineage from source to report. Data lineage shows flows between systems, tables, and columns, supporting impact analysis when a source field or transformation changes.
- A catalog of the inputs. The data catalog documents the tables, reports, and other data assets that make up the CECL process, so auditors and new team members can find them.
- Quality expectations on governed data. When CECL inputs are published as data products, quality checks and SLAs can be recorded on the product and its contracts.
For the wider financial services context, see Dawiso for financial services and consistent financial reporting.
Conclusion
CECL changed when and how institutions recognize credit losses, and with it the data they need. Expected lifetime losses have to be estimated from consistent historical data, often at loan level, together with current conditions and documented forecasts. The accounting standard leaves the method open, but auditors, examiners, and model validators all ask the same questions about the data: where it came from, what it means, who owns it, and how it was checked. Treating CECL inputs as critical data elements, with lineage and controls, is what turns a model output into a defensible number.
Sources
- Federal Reserve, FDIC, NCUA, and OCC - Frequently Asked Questions on the New Accounting Standard on Financial Instruments - Credit Losses (interagency FAQs, including data needs and effective dates)
- Federal Reserve - SR 19-8: Frequently Asked Questions on the Current Expected Credit Losses Methodology (April 2019, revised July 2020; notes the ASU 2019-10 deferral)
- Financial Accounting Standards Board - Accounting Standards Update 2016-13, Financial Instruments - Credit Losses (Topic 326), and ASU 2019-10
See it in action
Interactive Data Lineage
Visualizing how data moves, transforms, and connects across systems, applications, and reports.