The price was deliberately cautious — so what happens to the slack?
By now you know how a life premium is built. The gross premium starts from the bare risk and then loads on expenses and a margin for safety. And every piece of it rests on *assumptions*: an assumed mortality table, an assumed investment return, an assumed level of expenses. A careful insurer does not assume the *expected* future — it assumes a deliberately prudent one: a little more death than it truly expects, a little less investment income, a little more cost. That cushion is what keeps the company solvent when a bad year arrives.
But prudence has a flip side. If you price for a future darker than the one that actually arrives — and most years, it does arrive kinder — the policy will, by design, throw off a steady stream of surplus: money that came in beyond what the cautious assumptions said was needed. Here is the central choice that splits the whole life market in two. *Who gets to keep that surplus?* In a non-participating policy, it belongs to the company (and its shareholders); the policyholder paid a fixed price and that is the end of it. In a participating policy — also called with-profits — the policyholder is invited to *share* in the surplus their own prudent pricing helped create.
Three rivers of surplus: mortality, interest, expense
Where does the surplus actually come from? Beautifully, it comes from exactly the three assumptions the premium was built on. Each is a place where *real experience can beat the assumption* — and each gap, summed over the whole in-force book, is a named source of surplus. Actuaries call this [[mortality-interest-expense-in-product-design|source-of-surplus analysis]], and the three classic rivers are the mortality gain, the interest gain, and the expense gain.
The mortality gain appears when fewer policyholders die than the assumed table predicted. Suppose the table said 100 deaths this year on a block of policies, but only 90 lives were lost. The death benefits for those 10 "extra survivors" were already funded inside the reserves — and now they need not be paid out this year. That released money is mortality surplus. (Be careful with sign: for a *life-annuity* book it is the opposite — *more* deaths than expected releases reserve and creates the gain. Always ask which way the benefit points.)
The interest gain is usually the biggest river. The reserves are not idle cash sitting in a vault — they are *invested*, and the premium assumed they would earn some cautious rate, say 3%. If the assets actually earned 4.5%, the extra 1.5% on a large pile of reserves is pure surplus. Finally, the expense gain appears when the real cost of running the policies — commissions, admin, the items behind the expense assumptions — comes in below what was loaded into the premium. Three assumptions, three places reality can do better than feared, three rivers of surplus.
ONE YEAR OF SURPLUS ON A BLOCK OF PAR POLICIES (rough sketch) source assumption actual surplus this year -------- -------------- -------------- ----------------- interest earn 3.0% earned 4.5% +1.5% x reserves mortality 100 deaths 90 deaths +10 released death benefits expense loaded 120/policy spent 95/policy +25/policy total divisible surplus = interest gain + mortality gain + expense gain The gain is the GAP between assumed and actual, summed over the in-force book. If experience is WORSE than assumed, a river runs negative (a strain, not a gain).
Declaring the dividend — the actuary's annual judgement
The surplus is real, but turning it into policyholder dividends is an act of careful judgement, not a mechanical payout. Each year (or each valuation cycle) the actuary measures the surplus that emerged, decides how much is *durable* enough to give back versus how much should stay as a buffer, and then sets a dividend scale — a schedule of how much each policy receives, varying by issue age, plan, duration, and size. A guiding instinct is the contribution principle: distribute the divisible surplus roughly in proportion to how much each policy *contributed* to creating it. Fairness here means giving back to the policies whose own prudent pricing generated the slack.
Crucially, a declared dividend is rarely guaranteed in advance — and that asymmetry is the whole point. The guaranteed part of a par policy (the sum assured, the guaranteed cash value) is priced on the *prudent* assumptions, so it is rock-solid. The dividend sits on top and flexes with experience: rich in good years, thin or zero in bad ones. This is exactly what lets the insurer offer a participating contract without betting the firm. The policyholder accepts some upside *uncertainty* in exchange for sharing in the good times — a very different bargain from the fixed, known price of a non-par term policy.
How the money reaches the policyholder
Once a dividend is declared, the policyholder usually chooses *how* to take it — and the menu reveals how cleverly these contracts are wired into the rest of the policy. The classic options:
- Take it in cash — the dividend is simply paid out, the most transparent option.
- Reduce the next premium — the dividend is netted off the bill, so you pay less out of pocket.
- Leave it to accumulate at interest — the insurer holds it like a savings account attached to the policy.
- Buy paid-up additions — spend the dividend as a single premium on a tiny extra slice of fully-paid insurance, quietly growing both the death benefit and the cash value.
Outside North America the same idea wears a slightly different costume: the bonus. A traditional UK-style with-profits policy adds a reversionary bonus each year — a permanent increase to the guaranteed sum assured, which once added cannot be taken away — and then a one-off terminal bonus at death or maturity that hands over the share of surplus held back along the way. Whether you call it a dividend bought as a paid-up addition or a reversionary bonus, the deep mechanism is identical: surplus is recycled into more guaranteed benefit, and the contract's value ratchets up over time. The terminal bonus, by contrast, is deliberately *not* locked in — it is the insurer's release valve, smoothing returns and absorbing the final reckoning of experience.
Honest caveats — and how par sits among its cousins
Participating policies invite a few honest misconceptions, worth naming plainly. First, a dividend is *not investment income you earned* — it is the return of surplus from prudent pricing, and a chunk of any genuine investment gain is exactly the interest source we met above. Second, an illustrated dividend is a *projection*, not a promise; illustrations have a long history of being shown too optimistically, and a fall in the dividend scale is the most common complaint against these products. Third, the smoothing that makes with-profits feel safe also makes it *opaque*: it is genuinely hard to see what your money is really earning, because the insurer is deliberately holding surplus back in good years to top up bad ones.
It helps to place par among the family you have been meeting. A non-par whole-life policy fixes the price and keeps all surplus for shareholders — simple, transparent, no upside for you. A participating whole-life policy shares surplus back through dividends or bonuses, smoothing the ride. [[universal-life|Universal life]], by contrast, *un-bundles* the very same three levers and shows them to you directly: an explicit interest credit, explicit cost-of-insurance charges, explicit expense deductions in a transparent account. In a sense, universal life is the par mechanism turned inside out — the same mortality, interest, and expense rivers, but flowing in plain view rather than pooled and smoothed behind a dividend. Understanding the three sources of surplus is therefore the master key to the whole product shelf.