The warranty provision nobody recalculated, a full year after a product recall

This is a composite, illustrative scenario, not an account of a named client. The company's warranty provision had been calculated at a flat 1.5% of product sales value for three consecutive years, a rate set once and simply rolled forward each year without recalculation. The middle of those three years included a product recall on one appliance line, after which actual warranty claims paid that year ran at roughly 6% of sales for that line — four times the rate still being used to provide for future claims the following year, with no adjustment made to reflect the higher claims experience the recall had just demonstrated.

Why a flat rate survived a year that should have broken it

The finance team treated the 1.5% rate as a fixed policy rather than an estimate that had to be revisited each year against actual experience. The recall itself was expensed as it happened, through separate recall cost lines, which meant the warranty provision calculation — a mechanical multiplication of current-year sales by the historical rate — never ran into the recall data directly. Nobody connected the two numbers because they sat in different sections of the working papers.

What the provision reasonableness test found

1. 01Compared the provision movement to actual claims paid, product line by product line, for each of the three yearsFor every line except the recalled one, actual claims tracked close to the 1.5% rate. For the recalled line, actual claims had run at roughly 6% in the recall year and remained elevated at around 3% the following year as residual claims continued coming in.
2. 02Asked why the rate used for the provision had not been updated for that lineThere was no formal basis — it was simply the same rate applied company-wide, with no line-level differentiation at all.
3. 03Recalculated the provision for the recalled product line using its own actual claims experienceThis increased the required provision for that line materially above what the flat rate produced.
4. 04Checked whether the recall's root cause had been fixed or was still a live exposureManagement confirmed a design fix had been implemented from a specific production batch onward, which supported using the post-fix claims rate going forward rather than assuming the elevated rate would persist indefinitely.

What went into the file

• The claims-versus-provision comparison by product line across three years, isolating the recalled line from the rest.
• The recalculated provision for that line and management's revised estimate going forward, with the design-fix evidence supporting a tapering rate rather than a permanently elevated one.
• A note on why the generic company-wide rate was an inappropriate basis once one product line had a materially different claims profile.