Two engines that have been running separately
Up to now this ladder has built two machines, and kept them in separate rooms. The interest rung taught you to move money through time: a dollar promised in n years is worth less than a dollar today, so you discount it back using the present value you computed over and over. The probability and mortality rungs taught you a different skill: when an outcome is uncertain, you weight each possibility by how likely it is and add them up — that is what an expectation *is*. Both engines are reliable. They have just never been switched on at the same time.
Life insurance forces them together, because a life-contingent promise has *both* problems at once. Consider the cleanest possible contract: pay $100,000 at the end of the year in which a certain 40-year-old dies. When will that dollar leave the insurer? Nobody knows — it depends on the person's future lifetime, which is a random variable. And whenever it leaves, it must be discounted, because a payment ten years out is cheaper to promise than one next year. So the value of this promise is uncertain in its *timing* and discounted by its *distance*. Neither engine alone can price it.
The whole of life contingencies is the act of running both engines on the same cash flow. Discount each possible payment for its timing; weight each by the probability it actually happens; then add everything up. That single combined number is the keystone of this entire rung — and it has a name.
The actuarial present value, in plain words
The actuarial present value — APV, also called the *expected present value* — of a future payment is its present value, discounted for time, multiplied by the probability that the payment is actually made. If a payment is certain, that probability is 1 and the APV is just the ordinary present value you already know. The moment the payment becomes uncertain, you shave it down by its probability. APV is therefore not a new idea bolted on top of the old ones; it is the old present value with a probability factor riding alongside the discount factor.
Where does the probability come from? Straight out of the mortality machinery from the last rung. The chance that our 40-year-old is still alive after t years is read from the survival function (or equivalently from the life table); the chance they die in a particular future year is the difference between two survival probabilities. So a benefit that pays only on death gets weighted by death probabilities, and a benefit that pays only on survival gets weighted by survival probabilities. The cash flow tells you *which* probabilities to reach for; the survival model hands them over.
A tiny worked example: a 3-year pure endowment
Let's make APV concrete with the simplest life-contingent contract there is: a pure endowment. It pays $1,000 in exactly 3 years' time, but *only if* the policyholder is still alive then; if they die first, it pays nothing at all. There is a single payment at a single, fixed date — so timing is not random here — but whether it happens *is*. Suppose the effective annual interest rate is 5%, and from the life table the probability our person survives all 3 years is 0.94. Watch the two engines combine.
Contract: pay 1,000 in 3 years IF alive, else 0
i = 5% -> 3-year discount factor v^3 = 1/1.05^3 = 0.863838
3-year survival probability 3_p_40 = 0.94 (from the life table)
Step 1 ordinary present value (if it were certain):
1,000 * 0.863838 = 863.84
Step 2 weight by probability it actually pays:
863.84 * 0.94 = 812.01 <- actuarial present value
Contrast:
certain 3-yr payment of 1,000 PV = 863.84
life-contingent same payment APV = 812.01 (lower, by the 0.94)
expected dollars paid, ignoring time = 1,000 * 0.94 = 940.00 (too high)Look hard at the two contrasts at the bottom, because they pin down what APV really is. It is *lower* than the plain present value of $863.84 — the survival probability of 0.94 shaved off the chance the money is never paid. And it is *lower* than the naive expected payout of $940 — the discount factor shaved off the time value. Strip away either engine and you get a wrong, higher number. The APV of $812.01 is the only figure that respects both truths at once: the payment is uncertain, and it is in the future.
The equivalence principle: setting a fair premium
Now flip the contract around. The insurer's *benefit* — the money it might one day pay out — is a life-contingent cash flow, so it has an APV. But the policyholder's *premiums* are also a life-contingent cash flow: a regular payment that stops the moment the person dies, because you do not pay premiums on a policy that has already paid out. Premiums, too, have an actuarial present value. We now have two APVs facing each other across the contract: what the insurer expects to pay, and what the policyholder expects to pay.
The equivalence principle is the rule that decides the premium: set it so that, at the moment the policy is issued, the actuarial present value of the premiums *equals* the actuarial present value of the benefits. Neither side starts with an advantage; in *expectation*, and after discounting, the two promises are worth exactly the same. You already met this idea in disguise on the interest rung — it is the equation of value balancing what flows in against what flows out — but now both sides carry survival and death probabilities, not just discount factors.
- Write the APV of the benefit — every possible payout, each discounted for its timing and weighted by the probability it occurs.
- Write the APV of the premiums for one dollar of annual premium — the stream of premium payments, each discounted and weighted by the probability the policyholder is still alive (and so still paying) at that time.
- Set the two equal and solve: premium = (APV of benefit) / (APV of one-dollar premium stream). The number you get is the net (benefit) premium.
The premium that comes out of this division is the net premium (often the *net level premium* when it is the same every year). "Net" is the honest word: it covers the benefits and nothing else — no salaries, no commissions, no profit margin. Those real-world loadings get added later when you move from the net premium to the gross premium an actual customer pays. The equivalence principle gives you the bedrock fair price; commerce builds on top of it.
What the principle quietly assumes
The equivalence principle is powerful precisely because it is honest about *expectation* — and honesty means owning what it leaves out. "Equal in expectation" does not mean equal for any one policyholder. The person who dies in year two got vastly more than they paid; the person who lives to ninety paid far more than they ever received. The balance only holds *on average across the whole pool*, which is why insurance only works as the risk-pooling exercise you met at the very start of this ladder. APV is a statement about the crowd, never a promise about you.
One more honest point, because it will save you confusion in the reserves guides ahead. The equivalence principle balances the two sides *only at issue*, at time zero. The instant the policy is in force, that balance starts to drift: the benefit gets closer and more likely, while fewer premiums remain to be collected. The growing gap between the APV of future benefits and the APV of future premiums is exactly what a policy reserve must hold — but that is the next idea, and it stands squarely on this one.
What you carry forward
You now hold the keystone of life contingencies. The actuarial present value is nothing more exotic than your old present value with a probability factor riding alongside the discount factor — discount for timing, shave for likelihood, sum it all up. The equivalence principle takes that one tool and points it both ways: value the benefit, value the premium, set them equal at issue, and a fair net premium falls out. Everything else in this rung — whole life, term, endowment insurances, the premiums that fund them, the reserves that back them — is this single move applied to richer and richer cash flows.
If you remember just one sentence from this guide, make it this: an actuarial present value is an honest expected value of money over time, and the equivalence principle is the demand that two such honest values balance. Hold that idea steady and the notation cannot frighten you — when the next guides introduce the compact symbols of life-contingent valuation, you will recognise each one as a tidy package for the discount-weight-sum loop you have now done by hand. The symbols are the shorthand; the meaning is yours.