Testing a warranty provision against historical claim experience means rebuilding the claim rate and average claim cost directly from the entity's actual claims data for a sufficient historical period, applying that rebuilt rate to current-period sales still within the warranty window, and comparing the result to the provision the entity has recognised — then separately checking whether the prior year's provision, tested the same way a year ago, was adequate against what was actually claimed.
Why the claim rate itself is the thing to test
A warranty provision computed correctly as a percentage of sales is still wrong if the percentage itself has not been revisited. Product changes, quality issues in a specific batch, and changes in the warranty terms offered all move the right rate, and a rate inherited from several years ago stops being evidence of anything.
The testing sequence
- 01Obtain the raw claims data for a sufficient historical period, not the client's summarised rateClaim date, product or batch, cost of the claim, and the original sale date and warranty term. Enough history to see a trend, not just a single year's average.
- 02Segment claims by product line or batch where claim experience genuinely differsA single blended rate across products with materially different failure patterns understates the provision for the higher-risk line and overstates it for the lower-risk one.
- 03Independently recompute the claim rate and average cost per claim from the raw dataBuild the rate from first principles rather than reviewing the client's computation, so a change in methodology or a selective date range used by the client is actually caught.
- 04Identify sales still within the warranty window at the reporting dateApply the recomputed rate and average cost to this population, segmented consistently with the way the rate itself was segmented.
- 05Adjust for known changes that the pure historical average would not captureA product recall, a known batch defect, or a change in warranty terms offered this year should be reflected as a specific adjustment, documented separately from the base statistical rate.
- 06Compare the current provision to the recomputed estimate and quantify any differenceA provision materially below the recomputed estimate, carried for more than one period, is a pattern worth raising on its own, independent of whether this year's number alone is material.
- 07Test the prior year's provision against actual claims paid during the current yearThis is the single best piece of evidence about whether the estimation approach is reliable — a provision that consistently understates or overstates actual claims points to a flawed model, not bad luck.
| Input | Source |
|---|---|
| Claim rate | Recomputed from raw historical claims data, segmented by product where experience differs |
| Average cost per claim | Recomputed from actual claim cost records, not an assumed unit cost |
| Population within the warranty window | Sales register, filtered to the warranty term still open at reporting date |
| Known specific adjustments | Documented separately — recalls, known defects, changed warranty terms |
| Prior year back-test | Actual claims paid this year compared to last year's provision |
Common practice, not a requirement
Running the prior-year back-test before looking at the current year's provision at all is not a requirement, but it is the fastest way to tell whether this year's number deserves scrutiny or whether the model has reliably tracked actual outcomes for several years running.
How we think about this
Rebuilding the claim rate from raw data and applying it to the current sales population is a computation that should be reperformed in full, not reviewed against the client's own summary. Deciding how much weight a specific known event — a recall, a batch defect — deserves against the pure statistical average is the judgement call, and it belongs to the auditor, not the model.
This is practical guidance written for audit teams, not an authoritative statement of any standard or statute. Standards, forms and timelines change. Check the current text issued by the ICAI and the relevant authority before relying on anything here in an audit file.
Related resources
- Audit documentationInd AS 37: provisions, contingent liabilities and contingent assets — the three-way splitThree categories, three completely different treatments — recognise it, disclose it, or say nothing. Getting the category wrong is the whole misstatement.
- Audit workflow & automationSA 540: auditing accounting estimates, where judgment replaces a verifiable numberAn estimate can't be vouched to a source document the way an invoice can. SA 540 is about what replaces vouching when there's no single right answer to check against.
- Revenue, receivables & cut-offAuditing sales returns and credit notesWhy post-year-end credit notes are a classic overstatement signal, how to match them to the original invoice, and the GST reporting implications auditors should check.
- Inventory & stockAuditing standard costing variances in manufacturingMaterial, labour and overhead variance analysis under standard costing, and what each variance tells an auditor about inventory valuation risk.
Free tools, no signup
Each one reads a file entirely in your browser — or, for two of them, just answers a few questions — and nothing is ever uploaded.
- Purchase register checkDuplicate payments, round-sum amounts, weekend-dated invoices, incomplete records.
- GSTR-2B reconcilerMatches your purchase register against GSTR-2B — missing invoices either side, amount mismatches.
- Bank reconciliation checkerMatches your cash book against a bank statement — what's outstanding on each side.
- Fixed asset register checkDuplicate asset codes, missing location or useful life, negative net book value.
- Payroll red-flag scannerBank accounts shared across employees, and records missing an ID, PAN or salary.
- CARO applicability checkerAnswer a few questions about the company and see which CARO 2020 clauses likely apply.
- MSME payment delay checkerMSME-flagged invoices paid — or still unpaid — beyond the 45-day Section 22 limit.
- Related-party transaction screenerTransactions above a threshold you set with no approval reference, grouped by counterparty.
- Depreciation useful life checkerCompares a useful life in use against the Schedule II, Part C indicative figure.
- Materiality calculatorOverall and performance materiality against PBT, revenue, assets or equity, with reasoning shown.
- Sample size calculatorStatistical sample size for a test of controls or a substantive test of detail.