CECL Training for Accountants: Build ASC 326 Credit-Loss Estimates From Receivables to Review-Ready Allowances

By Vincent Howard, CPA | Managing Partner, Howard, Howard and Hodges | SkillAbility for Accounting Firms

Last updated: September 1, 2026  |  34-minute read

What Is CECL Training for Accountants?

CECL training develops an accountant’s ability to estimate lifetime expected credit losses for in-scope financial assets using relevant historical experience, current conditions, reasonable-and-supportable forecasts, and supportable assumptions.

The CECL model is fundamentally different from waiting for a loss to become probable.

For financial assets measured at amortized cost that fall within ASC 326-20, expected credit losses are generally recognized through an allowance at origination or acquisition and updated each reporting period.

For a typical commercial accounting team, that can affect:

• trade accounts receivable,
• contract assets,
• notes receivable,
• loans held for investment,
• held-to-maturity debt securities,
• net investments in leases,
• certain reinsurance recoverables,
• certain off-balance-sheet credit exposures.

But the estimate does not have to look like a bank’s probability-of-default model.

ASC 326 does not prescribe one required method.

A nonfinancial company with short-term trade receivables may use an aging schedule or loss-rate approach if that method reasonably reflects expected credit losses for the portfolio.

This article connects naturally to Revenue Recognition Training for Staff Accountants, Statement of Cash Flows Training for Staff Accountants, Business Combination Accounting Training for Staff Accountants, Fair Value Accounting Training for Staff Accountants, and Workpaper Review Checklist.

Why CECL Is a Judgment-Development Topic

I have practiced public accounting since 1990, founded my accounting firm in 1993, and helped grow Howard, Howard and Hodges from three people to approximately 50 staff. Since 2020, I have built SkillAbility around a recurring staff-development problem: accountants are often given last year’s allowance workbook and taught to update the balances—not to challenge the estimate.

That is especially risky under CECL.

The spreadsheet can multiply receivables by loss percentages.

It cannot decide whether:

• the receivable population is complete,
• risk segments still make sense,
• one troubled customer should leave the collective pool,
• historical loss data still represents the current portfolio,
• a recession forecast is already embedded in current delinquency data,
• a qualitative factor duplicates a quantitative forecast adjustment,
• a post-balance-sheet cash collection should be considered,
• collateral actually supports a practical expedient,
• an unfunded commitment requires a liability,
• or a debt security belongs under the AFS impairment model instead of CECL.

“Allowance = 2.1% of AR, same as last quarter” is not a conclusion.

“The allowance reflects this population, these risk characteristics, this loss history, these current conditions, these forecast assumptions, this reversion method, and these specifically identified exposures” is.

What Is Current in CECL in 2026?

KPMG’s latest comprehensive Credit Impairment Handbook is dated July 2026. It includes ASC 326-20, ASC 326-30, trade receivables, off-balance-sheet exposures, purchased financial assets, business combinations, write-offs, recoveries, forecasts, and disclosures. The 2026 edition incorporates both ASU 2025-05 and ASU 2025-08.

Deloitte’s current CECL Roadmap remains its June 2025 edition and reflects ASC 326 after CECL became effective for all entities for fiscal years beginning after December 15, 2023.

Chart: Where CECL Judgment Concentrates

SkillAbility training heat map—not a FASB ranking. Actual risk depends on asset type, contractual term, data quality, portfolio concentration, forecast horizon, collateral, credit enhancements, acquisition status, and entity-specific policies.

The ALLOWANCE READY Framework

A — Scope the Asset Before You Build an Allowance

CECL does not apply to every receivable-like balance or every debt security.

Common ASC 326-20 CECL assets

• Trade accounts receivable
• current and noncurrent contract assets
• notes receivable
• held-to-maturity debt securities
• loans held for investment
• net investment in leases
• certain reinsurance receivables
• certain financial guarantees and off-balance-sheet credit exposures

Common items requiring another model

• Available-for-sale debt securities: ASC 326-30 impairment model
• Financial assets measured at fair value through earnings: credit risk is reflected in fair value rather than a separate CECL allowance
• Equity securities: generally outside the ASC 326-20 CECL model
• Operating lease receivables: specialized lease guidance can apply depending on the fact pattern
• Loans held for sale: measured under applicable lower-of-cost-or-fair-value guidance rather than CECL as held-for-investment loans

Staff should build a scope matrix before building an allowance.

L — Contractual Term Defines the Lifetime You Are Estimating

CECL requires an estimate over the contractual term of the financial asset, subject to the standard’s guidance on expected prepayments and extensions.

Expected prepayments matter

Expected prepayments can shorten the period over which the entity is exposed to credit loss.

For example, a five-year note receivable may have an expected weighted-average life shorter than five years if meaningful prepayments are expected.

Extensions and renewals are not automatically included

Staff should not extend the CECL horizon simply because management expects to renew a customer note or credit arrangement.

Expected extensions, renewals, and modifications are generally excluded unless the borrower has an option to extend the contractual term that is not unconditionally cancellable by the entity or other specific guidance applies.

Short-term trade receivables can still require lifetime expected loss

“It turns over in 45 days” does not mean “no CECL.”

It means the lifetime is short—which often makes a simpler method appropriate.

L — Pool Assets With Similar Risk Characteristics, Then Pull Out Exceptions

ASC 326 requires financial assets to be evaluated collectively when similar risk characteristics exist.

Possible segmentation factors include:

• Customer type
• industry
• geography
• credit rating
• collateral type
• loan size
• aging bucket
• payment terms
• product or service line
• origination vintage
• risk grade

Trade receivable example

A $12 million receivable population includes:

• $7M large commercial customers with 30-day terms
• $3M small business customers with 45-day terms
• $1M healthcare customers subject to reimbursement delays
• $1M international distributors with higher historical loss rates

One blanket loss rate may hide meaningful differences in expected credit risk.

Individual assessment

An asset can leave a collective pool when it no longer shares similar risk characteristics.

Examples:

• customer bankruptcy,
• severe delinquency,
• material dispute,
• known fraud,
• borrower-specific restructuring,
• collateral-dependent collection.

The allowance model should therefore have both:

Segmentation should change when risk changes

If the company enters a new market, changes underwriting standards, experiences a major customer concentration, or shifts payment terms, the historical segments may no longer be appropriate.

CECL Methods Accountants Should Understand

ASC 326 does not mandate one method.

The method should be appropriate for the asset and reasonably estimate lifetime expected losses.

Aging schedule

Often practical for trade receivables.

Example:

• Current: 0.4%
• 1–30 days past due: 1.0%
• 31–60: 4.0%
• 61–90: 12.0%
• 90+: 35.0%

Those percentages still need to reflect current and forecast conditions—not merely last year’s matrix.

Loss-rate method

Historical net credit losses are compared with an exposure base such as:

• average receivables,
• originations,
• sales,
• outstanding balances.

Roll-rate method

Estimates how balances migrate through delinquency states toward default/write-off.

Probability of default / loss given default

Common in lending portfolios:

More sophisticated does not automatically mean more appropriate.

Discounted cash flow

Estimates the present value difference between contractual cash flows and cash flows expected to be collected.

DCF can be useful for individually evaluated loans or portfolios where timing of cash shortfalls is important.

Remaining life / weighted-average remaining maturity methods

Other methods can be appropriate when they reasonably estimate expected losses and align with the asset population and available data.

O — Historical Loss Data Is the Base, Not the Final Answer

Historical credit-loss experience generally provides a starting point for CECL.

Staff should understand where the loss data came from.

Data questions

• How many years are included?
• Does the period contain a recession?
• Does it contain unusually benign credit conditions?
• Were credit policies consistent?
• Were write-offs recorded consistently?
• Were recoveries included?
• Were acquired portfolios mixed into originated portfolios?
• Did customer terms change?
• Did the company enter/exit markets?
• Are the historical segments comparable to today’s assets?

Example: historical rate that no longer fits

A company has five-year net write-off experience of 0.60%.

But during those five years:

• 80% of customers were investment-grade national companies.
• The current portfolio now contains 45% small businesses.
• payment terms moved from 30 days to 60 days.
• customer concentration increased materially.

The 0.60% loss rate is evidence.

It is not automatically the expected-loss estimate.

External loss data

When internal data is limited, external information can be relevant.

But staff should document comparability:

• industry,
• credit quality,
• geography,
• term,
• underwriting,
• economic period.

W + N — Current Conditions, Reasonable-and-Supportable Forecasts, and Reversion

CECL requires entities to consider past events, current conditions, and reasonable-and-supportable forecasts relevant to collectibility.

Potential forecast variables

• Unemployment
• GDP / economic growth
• interest rates
• industry defaults
• commodity prices
• housing values
• customer bankruptcy trends
• reimbursement conditions
• company-specific credit trends

The relevant variable depends on the exposure.

Reasonable and supportable does not mean forecast forever

An entity does not need to forecast economic conditions over the entire contractual life when it cannot do so reasonably and supportably.

Beyond that forecast period, the model reverts to historical loss information using a reasonable method.

Common reversion approaches

• Immediate reversion
• straight-line reversion over a defined period
• other systematic/reasonable reversion methods

Example

A loan portfolio has a five-year expected life.

Management can reasonably support a two-year economic forecast.

The model could:

• use forecast-adjusted losses in Years 1–2,
• revert to historical experience during Years 3–4,
• use fully historical experience in Year 5.

The workpaper should explain why two years is supportable and why the reversion method is reasonable.

A — Use Qualitative Adjustments Without Double Counting Risk

Qualitative adjustments are often necessary when historical data and the core quantitative model do not fully capture current expected credit losses.

They are also one of the easiest places to introduce management bias.

Common qualitative factors

• Changes in lending or credit policies
• Changes in customer financial condition
• Changes in delinquency trends
• Changes in portfolio concentrations
• Changes in collateral values
• Changes in economic conditions
• Changes in staffing or collection effectiveness
• Changes in competitive/legal/regulatory environment
• Changes in model/data limitations

Double-counting example

Assume the base model already adjusts loss rates upward for forecasted unemployment.

Management then applies a separate +20 basis point qualitative factor labeled “economic recession / unemployment risk.”

The same risk may be counted twice.

Build a Q-factor bridge

Q factors should move when facts move

A qualitative adjustment should not remain +25 bps for four years merely because that number is “in the model.”

Staff should ask:

• What changed?
• What evidence supports the direction?
• How does the magnitude relate to the risk?
• What would make us reduce or remove the factor?

Trade Receivables: The 2026 ASU 2025-05 Opportunity

ASU 2025-05 is especially relevant to CPA firms serving private companies because it targets the complexity of estimating expected credit losses on current accounts receivable and current contract assets arising from ASC 606 revenue transactions.

Practical expedient — available to all entities

An eligible entity may elect a practical expedient allowing it to assume that current conditions as of the balance-sheet date do not change for the remaining life of eligible current accounts receivable and current contract assets.

For short-duration receivables, that can reduce the need to develop a separate forward-looking macroeconomic forecast when the practical expedient is appropriate and elected.

Additional policy election — entities other than PBEs

Entities other than public business entities can also elect to consider cash collection activity after the balance-sheet date when estimating expected credit losses on the eligible current receivable/contract-asset population.

This policy election is available only if the practical expedient is also elected.

Example

At December 31, a private company has:

• $4.0M current trade receivables
• $600K aged over 60 days
• $180K specifically concerning balance

By February 15—before financial statements are issued—the company collects:

• $3.6M of the total AR
• $420K of the >60-day bucket
• only $10K of the $180K concerning balance

If the company properly elected the ASU 2025-05 practical expedient and eligible non-PBE cash-collection policy election, subsequent collection activity can be relevant evidence in the year-end CECL estimate for those eligible receivables.

That does not mean every subsequent receipt automatically changes the allowance.

Staff still need to determine:

• which balances are eligible,
• whether the election is in place,
• whether the collection evidence relates to collectibility at the balance-sheet date,
• whether remaining balances need additional expected-loss analysis.

C — Collateral-Dependent Assets and Zero-Loss Conclusions Need Evidence

Collateral-dependent financial assets

When a borrower is experiencing financial difficulty and repayment is expected to be provided substantially through operation or sale of collateral, specialized collateral-dependent guidance can affect how expected credit losses are measured.

Staff should document:

• borrower financial difficulty,
• collateral type,
• current fair value,
• costs to sell where applicable,
• seniority and liens,
• expected source of repayment,
• timing of realization.

Coordinate material collateral fair values with Fair Value Accounting Training for Staff Accountants.

Zero allowance can be appropriate—but not because “this customer always pays”

ASC 326 can result in a zero expected-credit-loss allowance when historical loss experience adjusted for current/forecast conditions supports an expectation of nonpayment of zero.

Examples may include certain high-quality short-term assets, depending on facts.

But the conclusion should be based on evidence such as:

• historical nonpayment experience,
• credit quality,
• guarantees or credit enhancements within the applicable accounting model,
• current conditions,
• expected lifetime.

E — Write-Offs, Recoveries, and Subsequent Credit Events

Write-offs

Financial assets are written off when they are deemed uncollectible.

A write-off reduces:

• the asset’s amortized cost basis, and
• the allowance for credit losses.

It should not create a second loss if the expected loss was already appropriately provided through the allowance.

Example

A $100,000 receivable has a $90,000 individual allowance.

The entity concludes only $10,000 remains collectible and writes off $90,000.

The gross receivable falls to $10,000 and the related allowance is reduced by $90,000.

Recoveries

Expected recoveries of amounts previously written off can be included in the measurement of expected credit losses subject to ASC 326 limitations, and actual recoveries affect the allowance/credit-loss accounting when received or expected as applicable.

Staff should distinguish:

• recoveries on prior write-offs,
• ordinary collections on still-recorded balances,
• credit insurance proceeds,
• collateral proceeds.

Subsequent events

A customer bankruptcy shortly after year-end can provide important evidence.

Staff should assess whether the event:

• provides additional evidence about conditions existing at the balance-sheet date, or
• reflects a new condition arising after the balance-sheet date.

That analysis is separate from the ASU 2025-05 subsequent-cash-collection election.

E — Off-Balance-Sheet Credit Exposures Can Require a Liability

CECL can apply to certain off-balance-sheet credit exposures not accounted for as insurance.

Examples can include:

• loan commitments,
• lines of credit,
• certain financial guarantees.

The expected credit-loss liability generally reflects:

• expected funding over the contractual period the entity is exposed to credit risk, and
• expected credit losses on the amounts expected to be funded.

If an entity can unconditionally cancel a commitment, exposure beyond amounts already funded may be excluded under applicable guidance.

Example

A company has a $5M committed line to a counterparty.

At year-end:

• $2M is funded,
• $3M remains unfunded,
• management expects $1M of the unfunded amount to be drawn,
• expected loss rate on funded/drawn exposure is 3%.

Illustratively:

The exact method depends on facts, but the training point is simple: the balance sheet can contain both an allowance against funded assets and a separate liability for qualifying unfunded exposure.

AFS Debt Securities Are Not Measured Under the CECL Model

Available-for-sale debt securities are within ASC 326—but not ASC 326-20 CECL.

ASC 326-30 uses a separate impairment model.

That model focuses on whether a decline in fair value below amortized cost is attributable to credit, subject to the standard’s allowance mechanics and limitations.

Staff should not apply:

• trade-receivable aging rates,
• loan lifetime CECL percentages,
• the ASC 326-20 pooling model

directly to an AFS security.

Worked Example 1: Trade Receivable Aging Allowance

Assume a commercial company has the following year-end trade AR:

But one $100,000 balance in the 90+ bucket belongs to a customer that filed bankruptcy and is expected to pay only $15,000.

Staff removes that balance from the collective calculation.

If the 90+ pool excluding that customer becomes $100,000:

Individual customer:

Revised total allowance:

The more important lesson is not the math.

It is that segmentation and individual assessment changed the estimate materially.

Worked Example 2: Historical Loss + Forecast + Q-Factor Bridge

Assume a $20M receivable pool has a historical lifetime loss rate of 0.70%.

Current/forecast evidence:

• Industry default outlook worsened.
• The quantitative forecast model increases expected loss by 0.20%.
• Customer concentration rose materially, not captured in the model: +0.10% qualitative adjustment.
• Collection staffing improved, partially offsetting risk: −0.05% qualitative adjustment.

Staff then asks the critical review question:

Does the +0.10% concentration adjustment capture risk that is genuinely absent from the 0.20% quantitative forecast adjustment?

If yes, support it.

If not, remove the double count.

A Quarter-End CECL Close Workflow

Build one controlled allowance population

Every CECL workpaper should trace back to a controlled population that includes:

• Customer / borrower
• asset type
• invoice or loan identifier
• origination date
• maturity / due date
• outstanding balance
• aging / delinquency
• risk segment
• credit rating or risk grade where relevant
• collateral
• specific-watch-list status
• write-off / recovery status
• expected funding for commitments where applicable

This prevents the allowance workbook, disclosure schedule, and GL from becoming three different populations.

ASC 326 Self-Review Checklist Before Manager Review

• Did I identify every material financial asset subject to a credit-loss model?
• Did I distinguish ASC 326-20 CECL assets from ASC 326-30 AFS securities?
• Did I identify assets measured at fair value through earnings that do not need a separate CECL allowance?
• Did I identify loans held for sale separately from held-for-investment loans?
• Did I confirm the receivable/loan population ties to the GL?
• Did I reconcile the population to subledger detail?
• Did I identify acquired assets and whether special purchased-asset guidance applies?
• Did I identify the contractual term for each material pool?
• Did I consider expected prepayments?
• Did I avoid including unsupported expected renewals/extensions?
• Did I identify contractual extension options controlled by the borrower where applicable?
• Did I group assets with similar risk characteristics?
• Did I document why each segment is appropriate?
• Did I review whether segmentation needs to change from prior periods?
• Did I identify customers/borrowers no longer sharing pool risk characteristics?
• Did I remove individually evaluated exposures from collective calculations where appropriate?
• Did I identify bankruptcies, disputes, fraud, restructurings, and severe delinquencies?
• Did I document the method used for each portfolio?
• Did I confirm the method estimates lifetime expected credit losses?
• If using an aging method, did I support aging buckets and loss rates?
• If using a loss-rate method, did I define the exposure base consistently?
• If using roll rates, did I validate migration data?
• If using PD/LGD, did I validate probability and loss severity assumptions?
• If using DCF, did I validate expected cash flows and discount-rate mechanics?
• Did I identify the historical data period used?
• Did I reconcile historical write-offs to source accounting records?
• Did I include recoveries consistently in historical loss information?
• Did I review changes in credit policy during the historical period?
• Did I review changes in customer mix during the historical period?
• Did I identify whether historical data includes unusual recessionary or benign periods?
• Did I evaluate external data comparability when internal data was insufficient?
• Did I identify relevant current-condition changes?
• Did I identify relevant reasonable-and-supportable forecast variables?
• Did I support the length of the reasonable-and-supportable forecast period?
• Did I avoid forecasting beyond the period management can reasonably support?
• Did I document the reversion method?
• Did I document the reversion period?
• Did I ensure reversion is systematic and reasonable?
• Did I identify qualitative risks not captured quantitatively?
• Did I avoid double counting risks already embedded in forecast adjustments?
• Did I support the direction of every Q factor?
• Did I support the magnitude of every Q factor?
• Did I identify the evidence that would cause a Q factor to increase, decrease, or disappear?
• Did I compare Q factors with prior periods and explain changes?
• For eligible current ASC 606 receivables, did I identify whether the ASU 2025-05 practical expedient was elected?
• If the entity is not a PBE, did I identify whether the related subsequent-cash-collection policy election was made?
• Did I confirm that cash-collection evidence was used only for eligible assets under the entity’s policy?
• Did I distinguish subsequent collection evidence from unrelated subsequent events?
• Did I identify collateral-dependent financial assets?
• Did I support borrower financial difficulty where required?
• Did I obtain current collateral fair value where applicable?
• Did I consider costs to sell where required?
• Did I evaluate senior liens and legal access to collateral?
• Did I support zero-loss conclusions with actual evidence?
• Did I avoid assuming zero loss merely because a counterparty is large or government-related?
• Did I identify expected recoveries on prior write-offs where applicable?
• Did I ensure expected recoveries do not create an inappropriate negative allowance beyond applicable limits?
• Did I write off balances when deemed uncollectible?
• Did I reduce both gross asset and allowance appropriately for write-offs?
• Did I avoid recording a duplicate credit-loss expense when writing off a fully reserved balance?
• Did I identify subsequent bankruptcies or credit events before issuance?
• Did I determine whether subsequent events provide evidence about year-end conditions?
• Did I identify off-balance-sheet commitments within ASC 326?
• Did I estimate expected funding of commitments where required?
• Did I identify unconditionally cancellable commitments separately?
• Did I calculate a separate liability for qualifying off-balance-sheet credit exposure?
• Did I reconcile funded and unfunded exposure without duplication?
• Did I identify AFS debt securities separately?
• Did I avoid applying CECL lifetime-loss methods directly to AFS debt securities?
• Did I prepare the allowance rollforward?
• Did beginning allowance tie to prior-period financial statements?
• Did current-period provision expense tie to the income statement?
• Did write-offs and recoveries tie to source schedules?
• Did ending allowance tie to the balance sheet?
• Did I reconcile model output to the journal entry?
• Did I prepare required qualitative disclosure of methods and assumptions?
• Did I prepare required quantitative disclosure support by portfolio class where applicable?
• Did I document changes in methodology from prior periods?
• Did I document changes in segmentation from prior periods?
• Did I document changes in forecast variables from prior periods?
• Did I explain material allowance movement rather than only quantify it?
• Can another accountant reproduce the allowance from the population, loss history, forecast, Q factors, exceptions, write-offs, and rollforward?

100-Point CECL Readiness Scorecard

Suggested readiness bands

• 90–100: Ready to own defined recurring CECL workstreams with normal manager/technical review.
• 82–89: Generally review-ready; targeted coaching remains in forecast, Q factors, or individual/collateral analysis.
• 72–81: Controlled ownership with manager checkpoints before forecast and overlay conclusions are finalized.
• Below 72: Continue structured ASC 326 practice.

Override the numerical score for fabricated Q-factor support, intentionally biased forecasts, suppressed write-offs, manipulated segmentation, unsupported zero-loss conclusions, or material allowance plugs.

A 30/60/90-Day CECL Training Plan

Days 1–30: Build the receivable foundation

Require staff to:

• tie AR to the GL,
• build aging buckets,
• calculate historical net loss rates,
• identify specific troubled accounts,
• process write-offs,
• reconcile the allowance rollforward.

Days 31–60: Make them defend the estimate

Add:

• changing customer concentrations,
• a deteriorating economic forecast,
• improving collections,
• ASU 2025-05 elections,
• post-year-end cash collections,
• Q-factor double-counting traps.

Days 61–90: Add exceptions

Add:

• collateral-dependent loan,
• unfunded commitment,
• HTM security,
• AFS security,
• zero-loss conclusion,
• acquired loan,
• recovery on prior write-off.

Use Scenario-Based Training for Accountants so staff practice estimation and escalation before the live quarter-end close.

15 Realistic CECL Training Scenarios

1. Same allowance percentage for six years

The portfolio, economy, and customer mix changed, but the allowance remains 1.5%. Staff rebuilds the evidence rather than inheriting the percentage.

2. One customer distorts the 90+ bucket

A bankrupt customer represents half of the aging bucket. Staff removes the exposure for individual assessment instead of contaminating the collective pool.

3. Recession captured twice

The quantitative forecast already includes unemployment deterioration, but management adds a separate recession Q factor. Staff identifies double counting.

4. Historical losses came from a different customer population

The company shifted from enterprise to small-business customers. Staff challenges comparability of the historical base.

5. No allowance because receivables turn in 30 days

Staff explains that short life simplifies the estimate but does not eliminate lifetime expected-loss recognition.

6. ASU 2025-05 used without an election

The team uses subsequent cash collections to reduce year-end allowance, but the nonpublic entity never adopted the required policy election.

7. Cash collection after year-end is unrelated evidence

A customer pays only because a new investor recapitalized the business after year-end. Staff distinguishes the event from ordinary balance-sheet-date collectibility evidence.

8. Government receivable given zero loss automatically

Staff documents actual credit history, terms, and relevant current/forecast evidence before concluding zero allowance.

9. Collateral value is stale

A loan is collateral dependent, but the appraisal is two years old. Staff obtains current evidence and evaluates selling costs and liens.

10. A fully reserved receivable is written off through expense again

Staff corrects the write-off to reduce the gross receivable and allowance rather than record duplicate loss.

11. Recoveries are mixed into normal collections

Historical loss rates are understated because recovered write-offs were treated as current-period AR collections without clear tracking.

12. Unfunded line of credit ignored

The company has a funded allowance but no liability for expected credit loss on a qualifying expected draw.

13. AFS security put in CECL pool

Staff separates the security into the ASC 326-30 impairment model.

14. Purchased loan ASU applied too early

Staff treats ASU 2025-08 as mandatory in 2026. Reviewer confirms the 2027 effective date unless early adopted.

15. Q factor never changes

A +25 bps overlay remains unchanged despite improving delinquencies and credit conditions. Staff requires fresh evidence.

What CPA Firms Should Measure

Connect these measures to your Staff Accountant Competency Checklist, Accounting Employee Development Plan, and Accountants Shifting From Preparers to Reviewers.

Common CECL Training Mistakes

Mistake 1: Treat the prior-period allowance percentage as the model

The estimate is rolled forward without revisiting population, loss history, forecasts, or qualitative factors.

Mistake 2: Mix CECL and AFS debt-security accounting

Different ASC 326 models get combined into one allowance workbook.

Mistake 3: Segment only by aging when other risk characteristics matter

Customer type, geography, collateral, or concentration risk is ignored because the aging report is easier to obtain.

Mistake 4: Use historical losses without asking whether history is comparable

Credit policy, customer mix, product mix, or economic environment changed materially.

Mistake 5: Add forecasts and Q factors for the same risk

Expected credit losses become conservatively overstated through duplicated overlays.

Mistake 6: Treat Q factors as permanent percentages

Adjustments survive because nobody defined what evidence would remove them.

Mistake 7: Ignore individually risky customers because they are “already in the aging”

A material bankruptcy or dispute is diluted by a collective percentage.

Mistake 8: Assume post-year-end cash collections can always reduce the allowance

The entity’s policy election, asset eligibility, and evidence about year-end conditions are not evaluated.

Mistake 9: Write off a reserved balance through expense again

The same credit loss is recorded twice.

Mistake 10: Build disclosures after the model is finalized

The narrative of methods, assumptions, risk segments, and movement does not reconcile to the actual allowance process.

How SkillAbility Builds CECL Capability

BASE — Receivable and allowance execution

• ASC 326 scope
• AR population tie-out
• aging methods
• historical loss rates
• write-offs and recoveries
• allowance rollforward

MAPS — Estimation judgment

• Segmentation
• individual assessment
• forecast variables
• reversion
• Q-factor evidence
• ASU 2025-05 elections
• collateral-dependent exposures

SUMMIT — Reviewer and model-governance readiness

• Review CECL methodology
• challenge forecast assumptions
• challenge management overlays
• approve segmentation changes
• review purchased-financial-asset issues
• review AFS vs. CECL model selection
• review disclosures
• coach staff without rebuilding the allowance

Frequently Asked Questions About CECL Training

What is CECL?

CECL is the current expected credit losses model in ASC 326-20. It generally requires entities to recognize an allowance for lifetime expected credit losses on financial assets measured at amortized cost that are within the model.

Does CECL apply to trade accounts receivable?

Yes. Trade accounts receivable are commonly within ASC 326-20, including short-term receivables. A simple aging or loss-rate method may be appropriate when it reasonably estimates lifetime expected losses.

Does CECL require a specific model?

No. ASC 326 does not prescribe one required method. An entity can use an aging schedule, loss-rate, roll-rate, PD/LGD, DCF, or another reasonable method depending on the asset and available data.

What information is used in a CECL estimate?

CECL considers relevant historical loss experience, current conditions, and reasonable-and-supportable forecasts, followed by reversion to historical information when forecasts cannot be supported over the full contractual life.

What does lifetime expected credit loss mean?

It means expected credit losses over the asset’s contractual life, adjusted for expected prepayments and applicable guidance on extension or renewal options.

How should receivables be segmented?

Financial assets are evaluated collectively when they share similar risk characteristics. Segments can consider customer type, industry, geography, delinquency, credit rating, collateral, term, vintage, or other relevant risk factors.

When should a receivable be evaluated individually?

An exposure may require individual assessment when it no longer shares the risk characteristics of its collective pool, such as a material bankruptcy, dispute, restructuring, fraud issue, or collateral-dependent collection.

What is a CECL qualitative adjustment?

A qualitative adjustment addresses expected credit risk not adequately captured in historical or quantitative model inputs. It should be supported by evidence and should not duplicate risk already incorporated elsewhere in the model.

What is a reasonable-and-supportable forecast period?

It is the period over which management can reasonably support forward-looking assumptions relevant to expected credit losses. ASC 326 does not require forecasting economic conditions over the asset’s full life when that cannot be supported.

What is reversion under CECL?

Reversion is the process of returning forecast-adjusted expected-loss estimates to historical loss information after the reasonable-and-supportable forecast period. The method should be systematic and reasonable.

What changed for trade receivables in 2026?

ASU 2025-05 became effective for fiscal years beginning after December 15, 2025. It provides a practical expedient for eligible current ASC 606 receivables and contract assets and an additional policy election for entities other than PBEs.

What is the ASU 2025-05 practical expedient?

It permits an entity to assume that current conditions as of the balance-sheet date do not change for the remaining life of eligible current accounts receivable and current contract assets arising from ASC 606 revenue transactions.

Can a private company consider cash collected after year-end?

Under ASU 2025-05, an entity other than a public business entity can elect to consider subsequent cash collection activity for eligible current receivables and contract assets, but only if it also elects the related practical expedient.

When are receivables written off?

Financial assets are written off when deemed uncollectible. The write-off reduces both the amortized-cost basis and the related allowance rather than creating a second loss when the exposure was already reserved.

Can CECL result in a zero allowance?

Yes, if the entity’s historical experience adjusted for current and forecast conditions supports an expectation of zero nonpayment. The conclusion must be documented rather than assumed.

What is a collateral-dependent financial asset?

It is an asset for which the borrower is experiencing financial difficulty and repayment is expected to be provided substantially through operation or sale of collateral, triggering specialized ASC 326 measurement guidance.

Does CECL apply to unfunded commitments?

CECL can apply to certain off-balance-sheet credit exposures. The liability generally reflects expected funding and expected credit losses during the contractual period of exposure, subject to guidance for unconditionally cancellable commitments.

Does CECL apply to available-for-sale debt securities?

No. AFS debt securities are within ASC 326 but follow the separate ASC 326-30 impairment model rather than the CECL model in ASC 326-20.

What is ASU 2025-08?

ASU 2025-08 changes accounting for certain purchased seasoned loans by expanding gross-up accounting. It is effective for annual periods beginning after December 15, 2026, with early adoption permitted.

How do you know when an accountant is review-ready for CECL?

A review-ready accountant can scope the assets, control the population, segment risk, validate loss history, support forecasts and reversion, avoid Q-factor duplication, handle individual and collateral exposures, account for write-offs and unfunded commitments, reconcile the allowance, and explain the disclosures.

Current Research and Authority Resources

• KPMG — Credit Impairment Handbook, July 2026
• Deloitte — Current Expected Credit Losses Roadmap
• FASB — ASU 2025-05: Measurement of Credit Losses for Accounts Receivable and Contract Assets
• FASB — ASU 2025-08: Purchased Loans
• Google Search Central — Optimizing for Generative AI Features
• Google Search Central — Generative AI Performance Reports

ASC 326 can intersect with ASC 606 revenue/contract assets, ASC 842 lease receivables, ASC 805 business combinations, ASC 820 fair value, ASC 320 debt securities, ASC 460 guarantees, ASC 310 receivables, ASC 230 cash flows, income taxes, regulatory reporting, and industry-specific guidance. Verify current authoritative literature and entity-specific facts for live estimates.

The Bottom Line

CECL training should not produce accountants who can update an allowance percentage.

It should produce accountants who can defend the estimate.

Scope the asset before selecting the method.

Control the population and contractual term.

Segment assets by actual risk characteristics.

Use historical losses as evidence—not autopilot.

Incorporate current conditions and reasonable forecasts.

Document reversion.

Use Q factors only for risk that is not already captured.

Apply ASU 2025-05 only to eligible assets under the entity’s elected policy.

Pull troubled assets out of collective pools when appropriate.

Recognize write-offs, recoveries, collateral, and unfunded exposure correctly.

Keep AFS securities out of the CECL model.

Reconcile the allowance, expense, rollforward, and disclosure.

That is ALLOWANCE READY.

Protect Knowledge. Develop People. Scale the Firm.

To staff who can explain why the allowance changed before review has to rebuild it,

Vincent Howard, CPA
Managing Partner, Howard, Howard and Hodges
SkillAbility for Accounting Firms

© 2026 SkillAbility for Accounting Firms. This article provides general educational information and does not replace client-specific U.S. GAAP, audit, tax, lending, legal, valuation, regulatory, SEC, or other professional advice.