Trade receivable ECL is often described as simple because many entities apply a lifetime expected loss approach without tracking the general three-stage model. The measurement work can still be demanding. Customer mix, ageing, disputes, write-offs, recoveries and economic conditions rarely behave uniformly.
There is no single method that fits every receivables book. The right approach depends on portfolio size, homogeneity, data history, credit terms and the significance of individual balances.
Quick comparison
| Method | Best suited to | Data need | Main strength | Main limitation |
|---|---|---|---|---|
| Provision matrix | Homogeneous ageing-based portfolios | Medium | Intuitive and scalable | Can hide differences within buckets |
| Historical loss rate | Stable short-cycle portfolios | Low–medium | Simple and transparent | Can be too backward-looking |
| Roll-rate analysis | Portfolios with meaningful delinquency migration | Medium–high | Models deterioration paths | Needs stable state definitions |
| Vintage analysis | High-volume books affected by origination cohort | High | Reveals seasoning and cohort quality | Young vintages need projection |
| PD–LGD–EAD | Credit-managed or longer-term receivables | High | Granular risk decomposition | Can be over-engineered |
| Discounted cash shortfall | Large or distressed individual balances | Case-specific | Reflects specific recovery facts | Judgement-intensive |
| Hybrid approach | Mixed portfolios | Variable | Fits method to exposure | Requires clear method boundaries |
1. Provision matrix
A provision matrix applies expected loss rates to ageing buckets or other risk groups. For example, current balances may receive a lower rate than balances 31–60, 61–90 or more than 90 days past due.
The matrix should not begin and end with ageing. Segment by characteristics that drive collection outcomes—customer type, geography, product, channel, security, credit insurance or other relevant factors. Historical rates then need adjustment for current and reasonable, supportable forward-looking conditions.
Best when: the portfolio is large, reasonably homogeneous and ageing is predictive.
Watch for: arbitrary buckets, small-sample volatility and blanket forward-looking uplifts.
2. Historical loss-rate method
This method starts with observed credit losses over an appropriate period and applies an adjusted rate to a defined exposure base. It can be suitable for short-cycle books with stable terms and few delinquency stages.
The calculation must define loss consistently. Write-offs alone may omit recoveries, delayed write-off practices or disputed balances. The historical window should capture relevant experience without allowing obsolete periods to dominate.
Best when: exposure behaviour and customer mix are stable.
Watch for: zero-loss history being treated as proof of zero expected loss.
3. Roll-rate analysis
Roll-rate methods estimate how receivables migrate from one ageing or risk state to another—for example, current to 1–30 days past due, then to deeper delinquency, default, cure or write-off.
Multiplying transition probabilities through the recovery path can estimate the chance that each starting bucket ultimately becomes a loss. This captures deterioration better than a static rate when migration behaviour is informative.
Best when: enough monthly or quarterly history exists and states are defined consistently.
Watch for: changes in collection policy, billing systems or bucket definitions that break comparability.
4. Vintage analysis
Vintage analysis groups receivables by origination period—such as invoice month or customer onboarding cohort—and tracks cumulative loss as the cohort seasons.
It can reveal that recent cohorts perform differently because of changes in sales strategy, underwriting, customer mix or economic conditions. It is particularly useful where loss emerges in a repeatable pattern after origination.
Best when: transaction volumes are high and cohort performance matters.
Watch for: projecting young vintages using mature-vintage patterns without considering changed conditions.
5. PD–LGD–EAD method
Some receivables behave more like credit exposures than short-term invoices. A PD–LGD–EAD approach can be appropriate for longer credit terms, formal customer ratings, significant concentrations or portfolios already managed through credit-risk models.
PD estimates default likelihood, LGD estimates the loss after recoveries and EAD estimates exposure at default. Lifetime horizons and forward-looking information should reflect the contractual and behavioural life of the receivable.
Best when: data and risk differentiation justify component modelling.
Watch for: false precision and importing lending assumptions that do not fit commercial receivables.
6. Discounted cash-shortfall assessment
Large, disputed, restructured or distressed balances may be assessed individually. Estimate probability-weighted cash recoveries, their timing and relevant costs, then discount the shortfall using the applicable effective interest rate logic.
Scenarios might include full settlement under revised terms, partial recovery, negotiated settlement, insurance recovery or insolvency. Support should come from correspondence, legal status, payment behaviour, collateral or other case evidence.
Best when: a balance is individually material or pooling would hide its facts.
Watch for: optimistic recovery timing and double counting individual balances in the collective matrix.
7. Hybrid approach
Many entities need more than one method. A practical framework may use provision matrices for the core portfolio, individual assessment for material distressed balances and a separate method for insured, related-party or long-term exposures.
The framework should define clear boundaries, prevent omissions or duplication and reconcile every exposure to one measurement route. Hybrid does not mean ad hoc; it means deliberately matching method to risk.
Best when: the receivables population is economically diverse.
Watch for: moving exposures between methods to obtain a preferred result.
Worked provision-matrix illustration
| Ageing bucket | Balance | Adjusted lifetime loss rate | ECL |
|---|---|---|---|
| Current | ₹5,00,00,000 | 0.40% | ₹2,00,000 |
| 1–30 days past due | ₹1,50,00,000 | 1.50% | ₹2,25,000 |
| 31–60 days past due | ₹60,00,000 | 5.00% | ₹3,00,000 |
| 61–90 days past due | ₹30,00,000 | 15.00% | ₹4,50,000 |
| More than 90 days | ₹20,00,000 | 45.00% | ₹9,00,000 |
| Total | ₹7,60,00,000 | ₹20,75,000 |
The rates above are illustrative, not benchmarks. An entity should derive rates from its own relevant experience, segmentation and forward-looking assessment.
Five decisions to document whichever method is used
- Which exposures are in scope and which measurement route applies?
- Which default, loss and write-off definitions are used?
- How are historical observations selected and adjusted?
- How are current and forward-looking conditions incorporated?
- How are the result, changes and management judgement reviewed?
Frequently asked questions
Does the simplified approach mean no forward-looking adjustment is needed?
No. Simplification relates primarily to recognising lifetime ECL without tracking stage changes for qualifying assets. Measurement still uses relevant historical, current and reasonable, supportable forward-looking information.
Can every customer use the same provision matrix?
Only if their loss behaviour is sufficiently similar. Material differences in customer type, geography, product, security, payment terms or other risk characteristics may require segmentation.
How should credit insurance be treated?
Consider coverage terms, exclusions, counterparty strength, claim timing and whether the arrangement is integral to the contractual terms for ECL measurement. Do not replace analysis with the policy's headline coverage percentage.
Choose a method that fits the portfolio
Explore ECL for Trade Receivables or review the site's detailed guide to ECL for trade receivables and contract assets.
Technical references
- IFRS Foundation, IFRS 9 Financial Instruments
- IFRS Foundation, IFRS 9 educational session