CECL / credit loss case study

CECL (ASC 326) Implementation Case Study for AR

A $180M B2B services company replaced its decade-old incurred-loss bad-debt reserve methodology with an ASC 326 CECL forward-looking expected credit loss model — segmented by customer tier, calibrated against historical loss experience, and tested for forecast sensitivity. Audit-clean on first cycle.

Client profile: Composite case study based on a $180M revenue B2B services company with $32M average AR balance, 1,200 active customers, mix of enterprise and mid-market customer base. NetSuite GL with detailed AR aging history. Annual external audit; private with PE-backed sponsor.

See measured results Talk through your situation

Company context Before — what was actually broken What Ledger Summit implemented CECL mechanics for AR — what the engine handles Implementation timeline Measured results

Company context

ASC 326 (CECL — Current Expected Credit Loss) replaced the historical incurred-loss model for credit losses on financial assets, including trade receivables. Effective for non-public, non-PBE entities for fiscal years beginning after Dec 15, 2022. Pre-CECL: reserves recognized when loss is probable and estimable. Post-CECL: reserves recognized at origination based on lifetime expected loss, with forward-looking macroeconomic factors considered.

For B2B trade receivables (typical 30–60 day payment terms), CECL methodology is simpler than for banks (lifetime is short, historical loss experience is good predictor). But it still requires: customer segmentation, historical loss rate analysis, qualitative adjustments for forward-looking conditions, and documented methodology.

$180M revenue B2B services company
$32M average AR balance
1,200 active customers
Mix: enterprise (top 20 = 45% of revenue), mid-market, SMB tail
Historical write-off rate: ~0.4% / year
Industry: professional services + software; cyclical exposure
NetSuite GL with 5+ years of AR aging history

Before — what was actually broken

The pre-CECL reserve was a flat 1.2% of AR — set 8 years ago when annual write-offs were higher, never refreshed. Auditor accepted it as immaterial year over year. CECL adoption forced a methodology refresh; first-year application produced a small cumulative-effect adjustment to opening retained earnings.

Flat 1.2% reserve methodology, 8 years old
No customer segmentation in reserve calc
No forward-looking adjustment
Historical write-off analysis ad hoc
Documentation thin

What Ledger Summit implemented

A four-track CECL implementation: (1) customer segmentation by risk profile; (2) historical loss rate calibration per segment; (3) forward-looking qualitative adjustment with documented basis; (4) ongoing reserve model with quarterly review.

Customer segmentation: enterprise (top 20), mid-market (next 100), SMB (long tail), distressed (named list)
Historical loss rate per segment using 5-year rolling history with seasonality adjustment
Forward-looking qualitative adjustment: 25 bps overlay reflecting economic forecast (recession risk, industry conditions)
Specific identification for distressed customers (bypass general reserve)
Quarterly reserve refresh with named reviewer sign-off
CECL methodology memo: documentation of approach, segmentation, calibration, adjustments
Disclosure pack: rollforward, methodology, sensitivity to economic assumption changes

CECL mechanics for AR — what the engine handles

Topic	What it requires
Lifetime expected credit loss	For trade AR, lifetime is short (30-60 days); historical loss experience is reasonable predictor
Customer segmentation	Group customers with similar credit risk characteristics; segments need historical data depth
Historical calibration	5-year rolling window typical; longer if cyclical history reveals pattern
Forward-looking adjustment	Qualitative overlay reflecting economic forecast; documented and refreshed quarterly
Specific identification	Customers in distress (bankruptcy filing, restructuring, named workout) bypass general reserve; specific reserve at NRV
Practical expedient	Aging-based methodology acceptable if reasonable proxy for expected loss; documented
Provisioning	Initial recognition at origination; reversed when actually written off
Recoveries	Recognized when received; reduce write-offs in same period
Disclosure	Rollforward (beginning balance + provision + writeoffs + recoveries = ending balance); methodology; sensitivity
CECL adoption	Modified retrospective: opening retained earnings adjustment for cumulative effect

Implementation timeline

Weeks 1–2: 5-year AR aging data analysis; historical write-off pattern by customer / industry / segment
Weeks 3–4: Customer segmentation defined; loss rate calibration per segment; forward-looking adjustment methodology
Weeks 5–6: Reserve model built; quarterly refresh process designed; auditor walkthrough scheduled
Week 7: First quarter of CECL reserve calculated; cumulative-effect adjustment quantified
Week 8: Disclosure pack drafted; auditor walkthrough completed

Measured results

Metric	Before	After	Delta
Reserve methodology	Flat 1.2%, 8 years old	Segmented + forward-looking	—
Reserve % of AR	1.2%	1.8% post-CECL	+60 bps
Cumulative-effect adjustment	—	~$190K to opening RE	—
Audit findings — credit losses	None previously, but methodology untested	None	—
Quarterly reserve refresh	Annual at best	Quarterly	—
Documentation depth	Thin	Methodology + segmentation + adjustment	—

Alternatives considered

Option	Time	Cost band	Strengths	Weaknesses
Big-4 advisory	3 months	$180K–$320K	Brand	Cost; over-scoped
CECL software (Moody's, S&P)	2 months	$80K–$140K / yr	Purpose-built	For banks; over-scoped for trade AR
Internal-only	—	$0	No vendor	Methodology defensibility weak
Ledger Summit + auditor concurrence (selected)	8 weeks	$60K–$110K	Right-sized; documented	Maintenance ongoing

When this approach fits

$50M+ revenue with material AR balance
B2B with concentrated customer base
Annual external audit
IPO or M&A diligence pressure
Existing reserve methodology aged or undocumented
NetSuite, Sage Intacct, or similar GL

Lessons learned

Segment customers, don't aggregate. Single rate for all customers loses the signal CECL is trying to capture.
Document the forward-looking adjustment. Qualitative overlay needs a defensible basis (economic forecast, industry conditions); 25 bps without rationale is a finding.
Specific identification for distress. Customers in bankruptcy or workout get specific reserve; don't average them into the general pool.
Quarterly refresh is enough. Monthly is overkill; annual is too infrequent for a forward-looking model.
Auditor walkthrough early. Methodology buy-in before quarter-end calculations saves rework.

Frequently asked questions

Why did CECL replace incurred-loss?

Pre-CECL underreserved during expansionary periods (no probable loss yet) and required catch-up during downturns. CECL is forward-looking; reserves accrue at origination based on lifetime expected loss.

Does CECL apply to all financial assets?

Trade AR, loans, debt securities, lease receivables, financial guarantees — most financial assets at amortized cost. Doesn't apply to: AFS debt securities (CECL-light), non-financial assets, intangibles.

How is trade AR CECL different from bank loan CECL?

Trade AR is short-duration, simple. Banks have multi-year exposure with complex modeling. For trade AR: aging-based, segmented, simpler than bank CECL.

Can we use the practical expedient?

Aging-based methodology is acceptable if it's a reasonable proxy for expected loss. Most B2B trade AR portfolios qualify. Documented as policy.

What's a forward-looking adjustment?

Qualitative overlay reflecting economic forecast — recession risk, industry conditions, customer concentration. Documented basis and refreshed quarterly.

How does specific identification work?

Customers in distress (bankruptcy, workout) get specific reserve based on expected NRV from receivables. They're excluded from general segment to avoid double-counting.

What disclosure does CECL require?

Rollforward (beginning + provision − writeoffs + recoveries = ending), methodology description, sensitivity analysis to key assumptions, vintage analysis (industry-specific).

What was the cumulative-effect adjustment?

Modified retrospective adoption: difference between pre-CECL and post-CECL reserve at adoption date. Posted to opening retained earnings net of tax.

Does CECL affect tax provision?

Tax impact through deferred tax asset (CECL increases reserve = tax-deductible difference). Reversal pattern follows reserve releases (writeoffs).

What about M&A diligence?

Acquired AR is fair-valued at acquisition (ASC 805); going-forward, CECL applies. Acquired specific reserves carry through.

CECL methodology aged or undocumented?

A 30-minute call walks your AR portfolio and tells you whether the current reserve is defensible.

Book a free call