Meridian Health: a flat 12% PMI increase on a deteriorating, ageing book

The 12% is not derived from the experience, so it cannot be confirmed as presented.

Technical indication. 94/78 − 1 = 20.5% on current premium would restore the target loss ratio, before allowing for the 11.4% episode-cost trend, 3.8% volume growth, 0.6-year annual age drift, or the fact that the age curve is four years out of date. On the aggregate, 12% is under-indicated: the loss ratio would only fall to 94/1.12 = 83.9%, above target. Yet some bands will be over-indicated.

Why a flat increase is the wrong instrument.

• The drift is not uniform. Older bands carry both the cost growth and the worst lapse deterioration, so a single factor on a stale age curve entrenches the cross-subsidy from young to old.
• Band boundaries create cliff edges: a member crossing from 60-64 to 65-69 sees the flat factor plus the stale relativity in one renewal, which is hard to justify and hard to explain.
• The competitor's modular product attacks the entry point. A 12% rise on the cheapest plan damages acquisition and encourages selective lapse of the young and healthy - the premium-drift spiral.
• Fair value obligations require evidence that the increase is proportionate to cost, plus consideration of forbearance for vulnerable customers and clear communication.

Alternatives to model: rebalanced age and benefit relativities; higher excesses or co-payments on the loss-making bands; a restructured entry product; increases differentiated by duration and channel; staging across renewal dates rather than one April event.

Monitoring: monthly loss ratio and lapse by band, lapse elasticity estimates, new-business mix and retention offers, so the committee can see whether the retained book's morbidity worsens.

A sliding-scale ceding commission is not a flat proportional share. As the ceded loss ratio rises, the commission falls, so the insurer's net cost of the same 30% cession rises precisely when experience is bad; the reverse applies when experience improves. So the treaty already cushions, or amplifies, the effect of a mispriced increase.

Why this matters for the repricing decision.

• Part of any improvement in loss ratio following the increase flows back through a higher ceding commission and through profit commission, so the net retained benefit of a given rate increase is smaller than a naive 70% share.
• The experience rating has a lag. A bad 2024 will still be influencing treaty terms at the next renewal, so the decision is not a one-year one.
• The treaty can distort behaviour if it rewards ceded premium volume without regard to quality of the underlying risk.

What I would model and negotiate. Project the treaty cash flows under three scenarios (flat 12%, rebalanced repricing, no change) on a net-of-reinsurance basis, not gross. Then check and, if necessary, renegotiate: the loss ratio definition (does it include expenses and large claims?); whether there is a floor or cap on the commission; whether the reinsurer must consent to changes in the rating basis; the profit commission threshold; and whether a smaller quota share plus a catastrophe stop-loss gives cheaper protection for the tail than a 30% share with a punitive sliding scale. I would also confirm the treaty does not penalise the insurer for the legacy age curve being corrected.

Lapse is not random. The members most likely to leave after a large increase are the young, the healthy and the price-sensitive, so the retained book's expected claim cost per member rises even if nothing else changes. The spiral is self-reinforcing: higher lapse worsens morbidity, forcing a further increase, which drives further selective lapse.

Indicators to track by age band, product and duration:

• lapse and non-renewal rates, separated into voluntary lapse and lapse following a retention offer;
• lapse among members with no claims in the prior 24 months versus claimants (the selectivity measure);
• direct debit cancellation and failed-payment rates in the first three months after renewal;
• quote-to-bind and re-quote rates on the modular entry product;
• complaints and ombudsman referrals about the increase;
• movement of members down to excess-increased or downgraded plans rather than out of the book.

Triggers I would propose. If 55+ lapse exceeds 12% in the first quarter after renewal, pause further increases in that band and reassess. If the no-claim lapse rate exceeds the claimant lapse rate by more than 3 percentage points, recalculate the retained book's expected loss ratio immediately rather than waiting for the annual review. If downgrade rates exceed 8%, treat it as a product design problem, not a price problem.

Governance. Report the indicators monthly to the pricing committee with a pre-agreed escalation route, and document at the outset that the 12% was approved on a stated set of assumptions about lapse elasticity.