What you walk away with: surrender value and nonforfeiture
By now you know the deep secret of a level-premium policy: in the early years the customer pays *more* than the true cost of their cover, and that overpayment is stored up as the policy reserve. So a fair question follows immediately. If a policyholder paid those early overcharges for ten years and then decides to quit, do they just lose everything they prepaid? Morally they should not, and the law in most places agrees. The money they get back when they walk away is the cash surrender value, and the legal protections that guarantee it are called nonforfeiture rights.
Here is the honest catch, though: the surrender value is *not* simply the reserve. It is the reserve minus a surrender charge. The reason is the new-business strain you already met — the insurer spent heavily up front on commission and underwriting (the deferred acquisition costs) and expected to recover those costs slowly over many years of premiums. A policyholder who quits early walks out before the insurer has earned that money back. The surrender charge lets the insurer recoup its unamortized acquisition costs so that the remaining policyholders, and the company, are not left holding the loss. Surrender charges are heaviest in the first years and fade to zero after roughly a decade.
Nonforfeiture is broader than a cash refund. Instead of taking the money, a policyholder can usually convert the surrender value into *reduced paid-up* insurance — a smaller amount of fully prepaid cover with no further premiums — or into *extended term* insurance, keeping the full death benefit but only for as long as the surrender value can fund it. All three options spend the same pot of money in different shapes. The single sentence to carry away: the surrender value is the policyholder's equity in the contract, and nonforfeiture law makes sure that equity cannot simply be confiscated.
The first-year hole: why a fresh policy looks broke
Recall the clean benchmark from earlier in this rung: the net premium reserve pretends the customer pays exactly the expense-free net premium and ignores running costs entirely. That makes it tidy to compute — but it quietly tells a lie about year one. In reality the insurer pays a fat commission and underwriting bill in the first year, far more than the small expense margin built into the first year's gross premium. So in the first year the company spends cash it has not yet been reimbursed for.
Now look at what the net premium reserve demands. It says: at the end of year one, you must already be *holding* a reserve. But the insurer just spent its first premium on the agent and the medical exam — it has nothing left to set aside. The clean reserve insists the company put up money it does not have, on top of money it has already spent. This double squeeze on a young, growing book of business is exactly what actuaries call new-business strain: writing more policies, paradoxically, drains capital before it later replenishes it.
This is not a flaw in the arithmetic; the net premium reserve is doing exactly what it was defined to do. It is a mismatch between a *deliberately expense-free* reserve and the *expense-heavy* cash flows of real life. A regulator who forces a young insurer to hold the full net premium reserve from day one would punish growth and could even push a perfectly sound, fast-growing company toward insolvency on paper. So the profession built a release valve — a reserve that bends to acknowledge the first-year hole.
Modified reserves: full preliminary term and Zillmer
Modified reserves are a clever accounting move that eases new-business strain *without* changing a single promise to the policyholder. The trick is to *re-imagine* the premium pattern used to compute the reserve. Instead of pretending the customer pays one level net premium every year, the modified method pretends they pay a *smaller* premium in year one — just enough to cover that year's pure death cost and leave nothing over — and a slightly *larger* level premium in every later year to make up the difference. The total present value is unchanged, so the equivalence principle still holds; only the *timing* of the assumed funding is shifted.
The most generous version is full preliminary term (FPT). It treats the entire first year as if the policy were one-year term insurance: the first-year valuation net premium is set to the cost of one year's death cover, so the *reserve at the end of year one is zero*. That exactly frees up room for the heavy first-year acquisition expense. From year two on, the policy is valued as if it were a fresh whole-life issued at age x+1, with a single higher level premium. FPT effectively lets the insurer borrow the first-year expense from the future and repay it through the slightly larger renewal premiums.
Zillmer's method is the dial-able cousin of FPT. Rather than wiping out the *entire* first-year reserve, Zillmer shifts only a chosen amount — the *Zillmer rate*, often expressed as a percentage of the sum assured — from the first year into the renewals. FPT is just the special case where you shift the maximum the regulator allows. Many supervisory regimes cap the Zillmer rate precisely so insurers cannot relieve so much strain that the early reserve becomes dangerously thin. Think of FPT and Zillmer as the same release valve, with Zillmer offering a turnable knob and FPT being that knob opened all the way.
Does it actually make money? Profit testing
All the machinery so far answers 'what must we charge?' and 'what must we hold?' — but neither directly answers the question that keeps a finance director awake: *will this product actually earn a profit, and when?* That is the job of profit testing. The idea is wonderfully concrete. Take one representative policy, lay out a full set of best-estimate expense assumptions, mortality rates, lapse rates, and an investment return, then march forward year by year and write down the *actual cash the shareholder gains or loses* at the end of each year, after paying claims and expenses and after putting up (or releasing) the required reserve.
The string of those yearly profits, one number per policy year, is called the profit signature. For almost every traditional protection product it has the same unmistakable shape: a deep negative dip in year one — that is the new-business strain again, the acquisition cost the insurer has not yet recovered — followed by a long tail of small positive profits as the policy matures and the early outlay is slowly paid back. The signature is the financial fingerprint of the product. Reading it tells you not just *whether* a product is profitable but *how long* you must wait and *how much capital* you must sink up front to write it.
Profit signature for one policy (shareholder cash at year-end, per unit)
Year: 1 2 3 4 5 ... 20
Profit: -38.0 +6.5 +6.8 +7.0 +7.2 ... +9.1
^^^^
new-business strain (heavy first-year expense, reserve set up)
Discount the whole stream at the risk discount rate r = 10%:
NPV = -38.0/1.10 + 6.5/1.10^2 + 6.8/1.10^3 + ... + 9.1/1.10^20
= +11.4 (> 0, so the product adds value at a 10% hurdle)
Payback: the cumulative profit first turns positive around year 7.Actuaries boil the signature down into a few headline numbers. Discount the whole stream at the company's risk discount rate (its required return) and you get the *net present value of profit*; divide that by the present value of premiums and you get the *profit margin*. The discount rate that makes the NPV exactly zero is the product's *internal rate of return*. And the year in which cumulative profit first turns positive is the *payback period* — how long the company's capital is at risk. A product can be perfectly priced by the equivalence principle yet still be rejected here, if its return is too slow or too small for the capital it ties up.
Tying pricing to reporting — and an honest word on limits
Step back and see how the whole rung now connects. Pricing chose the gross premium so the deal breaks even on the chosen assumptions. Reserving decided how much the company must hold as the years pass — and modified reserves smoothed the brutal first year. Profit testing then ran the *same* assumptions forward to ask whether the shareholder is actually rewarded for the capital they risk. The reserve is not a sideshow to profit: because the reserve is money the insurer must lock up, *the reserving basis directly shapes the profit signature*. A heavier reserve deepens the early strain and slows payback; a lighter (modified) reserve releases capital sooner and lets profit emerge faster. Pricing, reserving, and reporting are one loop, not three departments.
Now the honesty this field demands. A profit test is only as good as its assumptions, and it is built on a *single representative policy* multiplied up — real portfolios are messier, with a spread of ages, sizes, and behaviours. Worse, the lapse assumption is treacherous: profit signatures often *depend on* policyholders lapsing at a certain rate (lapse-supported pricing), so a product that looks profitable can quietly turn into a loser if too few — or too many — people surrender. And every figure rests on a guessed investment return and mortality table that the future is free to ignore. The profit signature is a careful projection under one scenario, *not* a prophecy.
That is why prudent actuaries never read a single profit signature in isolation. They *sensitivity-test* it — rerun with mortality 10% worse, lapses doubled, investment returns a point lower — and watch how fast the comfortable NPV evaporates. A product whose profitability survives only under perfectly benign assumptions is not a profitable product; it is a bet. The discipline you have been building all rung — be explicit about every assumption, and be honest about how fragile your answer is to it — is precisely what turns a calculation into something a company, and a regulator, can trust.