Life Insurance Surplus Models Expand to Three Technical Bases

A life insurance policy can look steady on paper for years, then surprise insurers when old assumptions no longer match reality. This review shows why the math behind premiums and reserves has to keep pace with contracts that can run for decades. The paper organizes Markov models of surplus, built on a counting process following Norberg (1991), and sorts them by one and two technical bases using a clear notation. It then extends the framework to three technical bases to handle different valuation approaches. That extension forces two changes: the authors expand the idea of a technical basis to include non-contractual cashflows recognized in the Thiele equation, and they add new mainly systematic terms to surplus. The paper also treats some cashflows as dynamic, or quasi-contractual, to match real practice. Two examples are given: the paid-up valuation principle and reversionary bonus on participating contracts. The result is a map of how actuarial models connect theory with practice, especially in settings drawn largely from UK experience.

A life policy can last for decades. The price may start in one era. The reserve check may happen in another. That gap is the problem at the heart of this review. A technical basis is the set of assumptions used to price and value a policy. The old Northampton method uses the same basis for both premiums and reserves. That sounds tidy. It also feels safe. But long-lived contracts age. When the basis ages too, the accounting can start to look silly. Then surplus can appear in the wrong place, or vanish before it should. UK life business has lived with that tension for a long time. This is where the surprise begins. The safest-looking rule can become the least honest one.

Why one basis is not always enough

The review sorts the field into one-base and two-base models first. A one-base model uses one assumption set for the whole job. A two-base model splits the pricing basis from the reserve basis. The jump to three bases is the main move. It helps when valuation rules do not line up with the contract's own cash flows. To make that work, the review widens 'technical basis' to include cash flows outside the contract. Those flows still matter in the Thiele equation, which tracks how reserves change over time. The result is new, mostly systematic terms in surplus. The same cash flows can then be treated as quasi-contractual. That means they act like part of the deal in valuation. That covers paid-up valuation and reversionary bonus on participating contracts, which share profits with policyholders.

How the model keeps track of policy change

The method starts with a counting process. That is a running tally of events as they happen. It follows Norberg's 1991 setup. The policy moves between states, like active, paid up, or ended. A Markov model means the next move depends on the current state, not the full past. The review uses that idea to sort surplus models by technical basis. It then adds a clear notation. That notation shows which cash flows belong to pricing. It also shows which belong to reserves. And it flags flows outside the contract. They still shape surplus. The payoff is a map that links theory with practice.

Why this changes surplus talk

Surplus is the money left after reality meets assumptions. That sounds simple. It is not. If a policy uses one old basis forever, surplus can be distorted by stale rules. If it uses several bases, the model can match different ways of valuing the same contract. That matters for paid-up policies, where premiums stop but value remains. It also matters for reversionary bonus, where profits get added to participating contracts. The review's larger point is practical. Some cash flows are not written into the contract. Still, they behave like part of the contract once valuation begins. Treating them as quasi-contractual keeps the accounting honest.

The sharp question left open

The surprising move is to treat some outside cash flows as quasi-contractual. That sounds odd at first. It becomes useful fast. It lets paid-up valuation and reversionary bonus sit inside one map. That makes one thing possible. A single surplus map can fit both rules. It does this without treating every cash flow as contractual. Old assumptions can survive only where they still make sense. The next test is other long-lived UK-style participating contracts. If it can fit them, the model gives a cleaner way to compare valuation rules. If it cannot, the break point will be plain.