How to test actuarial assumptions behind the leave encashment liability

Testing the actuarial assumptions behind a leave encashment or compensated absences liability means independently benchmarking the discount rate, the attrition or withdrawal rate, and the salary growth rate the actuary used, against external market data and the entity's own experience, rather than accepting the actuary's certificate at face value because a qualified professional signed it. This procedure sits downstream of the general leave encashment audit procedure, which confirms the valuation exists and is correctly recorded; this one tests whether the valuation's own inputs are reasonable.

Why the assumptions, not the model, carry the real risk

The actuarial method for valuing compensated absences — typically the projected unit credit method — is standard and rarely the source of error. The liability is highly sensitive to three specific assumptions, and each one has an external benchmark the auditor can use independently of the actuary's own certificate: the discount rate against government bond yields, the attrition rate against the entity's own historical employee turnover, and the salary growth rate against the entity's own recent increment history and sector trend.

The testing sequence

1. 01Obtain the actuarial valuation report in full, not just the summary liability figureThe full report discloses each assumption used, the sensitivity of the liability to each one, and the employee data the valuation was run on.
2. 02Benchmark the discount rate against market yields on government bonds of a matching currency and termThe discount rate should reflect the yield on government bonds (or, where a deep market exists, high-quality corporate bonds) with a term approximating the expected timing of the benefit payments. A discount rate materially different from the prevailing yield curve as at the valuation date, with no documented reason, should be challenged specifically.
3. 03Compare the attrition rate assumption to the entity's own actual employee turnover over the past several yearsRequest the entity's headcount and separation data and compute the actual historical attrition rate by relevant employee category or grade, since attrition genuinely differs between, for example, a junior sales workforce and a senior management cadre, and a single blended rate across very different populations understates the sensitivity of the result.
4. 04Compare the salary growth assumption to the entity's own recent increment history and approved compensation plansObtain the actual average salary increase given over the past two to three years and the budgeted increase for the coming year, and test whether the long-term salary growth assumption used in the valuation is consistent with that pattern rather than an arbitrary round number.
5. 05Test the employee census data supplied to the actuary for completeness and accuracyAgree a sample of the census — date of birth, date of joining, salary, accumulated leave balance — to the entity's own payroll and HR records, since an actuarial valuation run on an incomplete or stale census produces a precise-looking liability built on wrong inputs.
6. 06Obtain and test the sensitivity analysis disclosed in the valuation reportA one-percentage-point change in the discount rate and in the salary growth rate should each be shown with its impact on the liability; use this to assess whether the overall assumption set, taken together, produces a result at the conservative or aggressive end of what the sensitivity range would support.
7. 07Compare this year's assumption set to last year's and investigate any unexplained changeA change in the discount rate consistent with the movement in government bond yields is expected and needs no particular challenge; a change in the attrition or salary growth assumption with no corresponding change in the entity's actual experience, made in a direction that reduces the liability, deserves a specific question to management and the actuary.