Employers' Liability Reserve Adequacy After a Causation Ruling

The £420m net reserve is more likely to be understated than adequate.

• Legal change. The Court of Appeal ruling widens the causation test for industrial disease. That can affect both the number of claims and the average cost. The 15% Q4 increase in reported claims is a warning sign, but it is not yet clean evidence: some may be a backlog after the ruling, some may be genuine new claimants, and some may be re-opens of previously closed claims. The actuary should split the movement by cause, accident year and disease type before extrapolating.
• Case reserves. With 2,000 open claims and a long duration, case reserves are slow to react. The £240m case estimate may not reflect the new legal test, especially for claims where liability was previously denied. A claim-by-claim review of the largest and most uncertain cases is needed.
• IBNR. £180m of IBNR on £80m earned premium is plausible for long-tail EL/PL, but only three years of post-transfer data makes the chain ladder unstable. The portfolio transfer doubled exposure, so development patterns before the transfer are not representative. Bornhuetter-Ferguson or frequency-severity methods with external benchmarks should be used alongside the chain ladder, not instead of judgement.
• Discounting. A 3.5% discount rate over nine years reduces the nominal reserve materially. If claims inflation for industrial disease is running above 3.5%, the real value of the reserve is being eroded. The actuary should show the reserve gross and net of discount, and stress a lower discount rate or higher inflation.
• Margin. The CFO's proposal to release £20m assumes the central estimate is robust. Given the legal uncertainty and data discontinuity, the central estimate itself may need to increase. Releasing margin before the review would be premature.

Recommendation: do not sign off £420m as adequate without a legal review, claim-level review, and a range of methods. Present a central estimate with a wider uncertainty range and hold a margin for the legal change. The dividend decision should wait.

Reinsurance credit must be assessed claim by claim and in aggregate.

• Per-claim XL. The £5m xs £2m cover only responds above £2m per claim. Most industrial disease claims may be below that, so the XL protects against the severe tail, not the frequency. The actuary should estimate the number of claims expected to breach £2m and the resulting recoveries, allowing for the widened causation test increasing both frequency and severity.
• Aggregate stop-loss. The £30m xs £50m aggregate cover is the more important protection if the legal change produces many medium-sized claims. The key issue is whether the new claims are one 'event' or separate occurrences. If the reinsurer treats them as separate, the stop-loss may not attach in the way the insurer expects; if they aggregate, the reinsurer may argue the deductible or event definition is not met. The wording and any hours or aggregation clause must be reviewed by legal counsel.
• Recoverability and timing. Even if the cover responds, the insurer must consider counterparty credit risk, the reinsurer's dispute, and the timing of recoveries. A disputed recovery should not be recognised at full value; a haircut or capital charge is appropriate until resolved.
• Net position. The £420m is a net figure. The actuary should show gross reserves, recoveries and net reserves separately, and stress a scenario where reinsurance recoveries are delayed or reduced. The board needs to see that the net reserve is not artificially supported by uncertain recoveries.

Recommendation: recognise reinsurance recoveries only to the extent they are probable and measurable, disclose the aggregation dispute, and hold additional margin for recoverability. The reserve review should include gross and net views.

Recommended reserve basis and communication

• Basis: set reserves at a central estimate that reflects the legal change, plus an explicit risk margin for the uncertainty in the legal outcome, data discontinuity and reinsurance recoverability. Do not describe the result as 'prudent' without explaining what confidence level it represents. A 75–80% sufficiency target may be defensible for a long-tail EL portfolio, but it must be documented and consistent with the board's risk appetite.
• Methods: use at least three approaches - chain ladder, Bornhuetter-Ferguson with external loss ratios, and frequency-severity with a legal scenario - and reconcile them. Give the board a range, not a single number.
• Regulatory communication: tell the regulator what has changed, what is uncertain, and what actions are being taken. Do not present the central estimate as settled while the legal position is still developing. A clear plan for a further review, with dates and triggers, is more credible than an optimistic point estimate.
• Board communication: separate the accounting question (what to hold) from the dividend question (what to distribute). Explain that releasing £20m now would reduce the buffer before the legal effect is known, and that a dividend funded by reserve release is effectively a distribution of uncertain capital.
• Documentation: record the legal advice, the claim review, the methods, the judgements and any dissent. If the board proceeds with the release, the actuary should ensure the minutes reflect the advice given.