Discounting Duration Profits to Time Zero

Why the Emergence Time Matters

A profit amount is not fully specified for valuation until its timing is known. A dollar of profit emerging at the end of the first policy year and a dollar emerging at the end of the second year have different time-zero values, even when their nominal amounts are equal. The discount period is measured from time zero to the stated profit emergence time.

This tutorial focuses on discounting each expected duration profit separately. It distinguishes an expected profit conditional on a policy being in force from an expected amount per policy originally issued, and makes the timing convention explicit for both annual and fractional durations. This treatment is aligned with the SOA Exam ALTAM (Advanced Long-Term Actuarial Mathematics) syllabus.

Definition: Let $t$ be the time, measured in years from issue, at which a profit emerges. At effective annual interest rate $i$, using the equivalent fractional-year convention described below, the time-zero value of a profit amount $G_t$ is $v^tG_t$, where $v=1/(1+i)$. If $G_t$ is random, its time-zero expected contribution is $v^tE[G_t]$.

Assigning the Correct Duration

Under the annual timing convention used here, the profit for policy year $k$ emerges at the end of that policy year, at time $t=k$. In particular, the profit for the first policy year is discounted for one year, not zero years. A profit emerging halfway through the first policy year is at $t=0.5$; halfway through the second policy year it is at $t=1.5$.

The duration is the actual elapsed time from issue to emergence, not merely the policy-year label. When cash flows within a policy year have already been aggregated into an annual profit under a specified timing convention, discount that annual profit at its assigned emergence time. Do not discount its component cash flows again.

For an integer duration $k$, the discount factor is $v^k=(1+i)^{-k}$. For a fractional duration $t$, this tutorial assumes equivalent fractional-year discounting: accumulation over $t$ years is $(1+i)^t$, and discounting is its reciprocal. This is equivalent to using the constant force $\delta=\ln(1+i)$ over the fractional interval. An effective annual rate by itself specifies annual accumulation; if a problem specifies a different within-year convention, use that convention instead.

Timing rule: First identify the emergence time $t$ from issue. Then discount by $v^t$. For end-of-year profits in policy years 1, 2, and 3, the factors are $v$, $v^2$, and $v^3$, respectively.

Worked Example: First-Year and Second-Year Profits

A hypothetical policy has an effective annual interest rate of 5%. The expected profit for policy year 1, conditional on being in force at the start of that year, is $200. All policies are in force at issue. At the start of policy year 2, the probability of being in force is 0.8, and the conditional expected profit for that year is $150. Find each year's expected profit per policy originally issued and its time-zero value. Round dollar amounts to the nearest cent.

Method. Convert any conditional profit to an expected amount per originally issued policy using the relevant in-force probability, then discount that amount at the end-of-year duration.

For policy year 1, all policies are in force at its start, so the expected profit per original policy is $1.00(200)=\$200$. The profit emerges at $t=1$, so its time-zero value is

200v^1=\frac{200}{1.05}=190.476190\ldots

This is $190.48 to the nearest cent.

For policy year 2, the expected profit per original policy is $0.8(150)=\$120$. It emerges at $t=2$, so its time-zero value is

120v^2=\frac{120}{(1.05)^2} =\frac{120}{1.1025} =108.843537\ldots

This is $108.84 to the nearest cent. As a check using unrounded values, accumulating the two contributions back to their emergence times gives $190.476190\ldots(1.05)=200$ and $108.843537\ldots(1.1025)=120$.

Common mistake avoided. Discounting the conditional $150 directly gives $150/1.1025=\$136.05$, but that amount is conditional on the policy being in force. The expected profit per original policy is only $120. Applying the in-force probability after discounting would give the same result if done exactly once; ignoring it or applying it twice would not.

Conditional Profits and Expected Profits per Original Policy

A profit may be stated conditionally, for example, “expected profit during year $k$, given that the policy is in force at the start of the year.” That is not automatically the expected amount for each policy originally issued. If profit is zero for a policy that is not in force at the relevant time, the unconditional expected profit per original policy is the conditional expected profit multiplied by the probability of being in force.

More generally, let $I_t$ indicate that the policy is in force at the conditioning time, and suppose there is no duration-$t$ profit for policies for which $I_t=0$. If the conditional expected profit is $m_t=E[G_t\mid I_t=1]$, then

$$ E[G_t]=P(I_t=1)m_t. $$

The time-zero expected contribution is therefore $v^tP(I_t=1)m_t$. This is the appropriate form when the profit is stated conditionally but the valuation is per policy originally issued. If the amount is already an unconditional expected profit per original policy, do not multiply by the in-force probability again.

Worked Example: Discounting Unconditional Annual Profits

A hypothetical profit projection gives unconditional expected profits per original policy of $500 at the end of year 1 and $420 at the end of year 2. The effective annual interest rate is 5%. Find each time-zero contribution, rounded to the nearest cent.

Method. These amounts are already unconditional, so discount each at its stated duration without applying another probability weight.

The year-1 contribution is

500v^1=\frac{500}{1.05} =476.190476\ldots =\$476.19\text{ to the nearest cent}.

The year-2 contribution is

420v^2=\frac{420}{(1.05)^2} =\frac{420}{1.1025} =380.952380\ldots =\$380.95\text{ to the nearest cent}.

Using the unrounded contributions, accumulation back to the emergence dates checks the arithmetic: $476.190476\ldots(1.05)=500$ and $380.952380\ldots(1.1025)=420$.

Common mistake avoided. The second-year discount factor is $v^2$, not $v$. Using $420/1.05$ would value the profit as if it emerged after one year, producing $400.00 instead of $380.95.

Fractional Durations

Some profit components emerge between policy anniversaries. When the emergence time is fractional, retain that fraction in the exponent. For example, with effective annual rate $i$, the discount factor for a profit at time $1.5$ is $(1+i)^{-1.5}$, not $(1+i)^{-1}$ or $(1+i)^{-2}$. This convention is particularly useful when the problem explicitly assigns a midyear emergence time.

Worked Example: A Midyear Profit in the Second Policy Year

A hypothetical insurer uses an effective annual interest rate of 21% and the equivalent fractional-year convention. At the start of policy year 2, the probability that a policy is in force is 0.8. Conditional on being in force at that time, the expected profit emerging halfway through year 2, at $t=1.5$, is $332.75. Find the time-zero expected contribution per policy originally issued, rounded to the nearest cent.

Method. First convert the conditional profit to an unconditional expected amount per original policy. Then discount for the full 1.5 years.

The expected profit per original policy at $t=1.5$ is

0.8(332.75)=\$266.20.

The fractional-year accumulation factor is

(1.21)^{1.5}=1.21\sqrt{1.21} =1.21(1.10)=1.331.

Thus the time-zero contribution is

\frac{266.20}{1.331}=\$200.00.

The forward check is $200.00(1.331)=266.20$, the expected amount at time 1.5. The conditional amount is also recovered by dividing $266.20 by 0.8, giving $332.75.

Common mistake avoided. Discounting by only one year would ignore the half-year after the first policy anniversary. Discounting by two years would treat the profit as if it emerged at the end of year 2. The stated emergence time is $t=1.5$, so the exponent must be 1.5.

Why Discounting an Expected Profit Works

For a fixed emergence time $t$, the discount factor $v^t$ is deterministic under the assumed interest rate. Therefore it can be taken outside the expectation. This gives a direct justification for discounting the expected profit amount rather than separately discounting every possible profit outcome.

Represent the profit at its emergence time.
Let $G_t$ be the random profit amount paid or received at the fixed time $t$.

Apply the time-zero discount factor.
The random time-zero amount is $v^tG_t$, because the discount factor depends on the fixed time and interest assumption, not on the profit outcome.

Take expectations.
By linearity of expectation and because $v^t$ is constant, $E[v^tG_t]=v^tE[G_t]$. Thus the expected amount is discounted at the same duration-specific factor.

This argument assumes a fixed emergence time and a deterministic interest rate. If the timing itself is random, or if discount factors depend on random future investment returns, the discount factor may not be a fixed constant that can be taken outside the expectation. Those cases require the timing and dependence assumptions to be handled explicitly.

Common Timing and Weighting Traps

Starting the count at zero for year 1. Under end-of-year emergence, the first policy-year profit occurs at $t=1$, so it is discounted by $v$, not by 1.
Using the policy-year number as an exponent without checking the timing. Year $k$ profit is discounted by $v^k$ only when it emerges at the end of year $k$. A midyear or other within-year emergence has a different time.
Confusing conditional and unconditional amounts. Multiply a conditional expected profit by the relevant in-force probability when converting to an amount per original policy. Do not apply that probability again if the stated profit is already unconditional.
Rounding too early. Keep the unrounded discounted amounts through intermediate calculations, then round the final dollar result to cents. Early rounding can affect later checks or sums.
Assuming a fractional convention without checking the problem statement. The exponent $t$ follows the stated timing convention. If a problem specifies a different within-year accumulation rule, use it rather than silently imposing equivalent fractional-year discounting.

The key discipline is to label each expected profit with both its basis and its timing: conditional or unconditional, per in-force policy or per original policy, and emerging at which exact time. Once these are clear, the appropriate discount factor follows directly.

Check Your Understanding

For each problem, use the stated emergence time and distinguish conditional from unconditional profits where needed. Round dollar answers to the nearest cent.

A hypothetical policy has an effective annual interest rate of 6%. Its unconditional expected profit per original policy is $318 at the end of year 1. Find its time-zero contribution. (Answer: $300.00)
At an effective annual rate of 5%, an unconditional expected profit of $250 emerges at the end of year 2. Find its time-zero contribution. (Answer: $226.76)
At an effective annual rate of 21%, using equivalent fractional-year discounting, an unconditional expected profit of $110 emerges halfway through the first policy year. Find its time-zero contribution. (Answer: $100.00)
At an effective annual rate of 5%, the conditional expected profit at the end of year 3 is $180 for a policy in force at the start of that year. The probability of being in force then is 0.7, and policies not in force have no profit at that duration. Find the time-zero contribution per original policy. (Answer: $108.84)
At an effective annual rate of 21%, a conditional expected profit of $532.40 emerges at $t=1.5$, and the probability of being in force at the relevant conditioning time is 0.5. Assume no profit at that duration when not in force. Find the time-zero contribution per original policy. (Answer: $200.00)

What You'll Learn