Reserve Differences From Benefit Basis Changes

Why Isolate the Benefit Basis?

A prospective reserve is the value, at a valuation date, of future outgo less future income, conditional on the contract being in force at that date. If only the benefits used in the valuation change, the reserve difference can be found by comparing those future benefit values directly. Premiums, expenses, interest, and mortality do not need to be recalculated if their assumptions and projected cash flows are unchanged.

The benefit change may affect the amount paid, the time it is paid, or the conditions under which it is paid. These are distinct effects. A benefit paid later has a different present value even if its amount and triggering event are unchanged; a waiting period can remove some claims while leaving others eligible.

Definition: Let $V_t^{\text{old}}$ and $V_t^{\text{new}}$ be prospective reserves at duration $t$ under old and new valuation benefit bases. If premiums, expenses, interest, decrements, and all other projected cash flows are identical, define the benefit-basis reserve difference as $\Delta V_t=V_t^{\text{new}}-V_t^{\text{old}}$. A positive difference means the new basis produces a higher reserve; a negative difference means it produces a lower reserve.

This comparison concerns the benefits included in the valuation. A change to a valuation benefit basis need not mean that the contractual benefit paid to a policyholder has changed. In either interpretation, the calculation below isolates the benefit effect only; if premiums or other cash flows also change, their differences must be included separately.

Annual Benefit Changes: Amount and Timing

Consider a life aged $x$ at the valuation date and a contract with $n$ remaining policy years. Suppose a death benefit $B_k$ is paid at the end of the year of death if death occurs in future year $k$. Under the usual annual model, the conditional probability of death in year $k$ is ${}_{k-1}p_x q_{x+k-1}$, and the discount factor to the valuation date is $v^k$, where $v=1/(1+i)$.

V_t^{\text{death}}=\sum_{k=1}^{n} {}_{k-1}p_x q_{x+k-1} B_k v^k

If both bases pay at the end of the same year for the same death event, subtracting the old reserve from the new reserve gives:

\Delta V_t^{\text{death}} =\sum_{k=1}^{n}{}_{k-1}p_x q_{x+k-1} \left(B_k^{\text{new}}-B_k^{\text{old}}\right)v^k

This formula assumes the same mortality probabilities and the same payment date under both bases. If payment dates differ, discount each basis’s benefit to its own payment date before subtracting. More generally, for mutually exclusive claim events $j$, let $P_j$ be the probability of event $j$ conditional on being in force at the valuation date, and let $B_j$ and $\tau_j$ be the benefit amount and payment time under a basis. Then that basis’s benefit value is $\sum_j P_jB_jv^{\tau_j}$. The event probability is conditional on the valuation date; do not multiply by the probability of surviving from issue to that date again.

Why the Difference Formula Holds

The result follows directly from the prospective-reserve definition. Write future benefits as $B$, premiums as $P$, and expenses as $E$, with each symbol denoting the present value of the corresponding future cash flows at the valuation date. Under each basis, the reserve is benefits plus expenses less premiums. If only benefits differ, subtracting the two reserve equations cancels the identical expense and premium values.

\begin{aligned} V_t^{\text{new}}&=B^{\text{new}}+E-P,\\ V_t^{\text{old}}&=B^{\text{old}}+E-P,\\ V_t^{\text{new}}-V_t^{\text{old}} &=(B^{\text{new}}+E-P)-(B^{\text{old}}+E-P)\\ &=B^{\text{new}}-B^{\text{old}}. \end{aligned}

The benefit values are sums of event probabilities multiplied by discounted payments. Thus, for benefits paid on the same events and dates, the difference is the probability-weighted present value of the benefit increments. If a condition or payment time changes, the new and old benefit values must instead be constructed separately, event by event.

Fix the valuation date and condition.
Work conditional on the contract being in force at that date, using only future probabilities from that point onward.

List the benefit events.
For each future event, state whether it is covered under each basis, the benefit amount, and the payment time.

Value each basis consistently.
Use the same interest and decrement assumptions for both, and discount each payment from its actual payment date.

Subtract new minus old.
Check the sign against the direction of the benefit change and round only the final reserve difference.

Worked Calculations

Worked Example: Increasing a Benefit in the Second Year

A hypothetical two-year term contract is valued at issue. The death benefit is increased by $8,000 only for death in year 2; all other benefits are unchanged. Assume independent annual mortality probabilities with $q=0.08$ and $p=0.92$ in each year, and an effective annual interest rate $i=0.05$. Find the reserve difference, new minus old.

Method. The benefit increment applies only to death in year 2. For that event, the policyholder must survive year 1 and die in year 2, so its probability is $0.92(0.08)=0.0736$. The increment is paid at the end of year 2.

The present value of the increase is therefore:

\Delta V_0 =8{,}000(0.0736)(1.05)^{-2} =\frac{588.80}{1.1025} =534.0589569\ldots

Rounded to the nearest cent, the reserve difference is +$534.06. The new reserve is $534.06 higher. This calculation avoids applying the year-2 mortality probability without first allowing for survival through year 1.

Worked Example: Deferring Payment After Death

A hypothetical one-year death benefit of $12,000 is payable at the end of the year of death under the old basis. Under the new basis, the same benefit for a death during year 1 is paid one year later, at the end of year 2. At issue, assume $q=0.05$ and $i=0.08$. Find the reserve difference, new minus old.

Method. The event is death in year 1 under both bases, so its probability is 0.05 in each calculation. The old payment is discounted one year; the new payment is discounted two years:

\begin{aligned} B^{\text{old}}&=12{,}000(0.05)(1.08)^{-1} =\frac{600}{1.08}=555.5555556\ldots,\\ B^{\text{new}}&=12{,}000(0.05)(1.08)^{-2} =\frac{600}{1.1664}=514.4032922\ldots,\\ \Delta V_0&=514.4032922\ldots-555.5555556\ldots\\ &=-41.1522634\ldots \end{aligned}

Rounded to the nearest cent, the reserve difference is −$41.15. The later payment lowers the reserve even though the benefit amount and probability of death are unchanged. The common error is to discount both payments for one year, which incorrectly gives a zero difference.

Worked Example: Adding a Waiting Period and Changing the Later Benefit

A hypothetical two-year term contract pays $10,000 at the end of the year of death for death in either year under the old basis. Under the new basis, death in year 1 is not covered, while death in year 2 is covered for $14,000. Assume $q=0.10$, $p=0.90$ in each year, and $i=0.05$. Find the reserve difference at issue.

Method. First value each basis. Death in year 1 has probability 0.10; death in year 2 has probability $0.90(0.10)=0.09$. The old basis pays in both years, while the new basis pays only in year 2.

\begin{aligned} B^{\text{old}} &=\frac{10{,}000(0.10)}{1.05} +\frac{10{,}000(0.09)}{1.05^2}\\ &=952.3809524\ldots+816.3265306\ldots\\ &=1{,}768.7074830\ldots,\\ B^{\text{new}} &=\frac{14{,}000(0.09)}{1.05^2} =\frac{1{,}260}{1.1025} =1{,}142.8571429\ldots,\\ \Delta V_0 &=1{,}142.8571429\ldots-1{,}768.7074830\ldots\\ &=-625.8503401\ldots \end{aligned}

Rounded to the nearest cent, the reserve difference is −$625.85. The larger year-2 benefit adds value, but removing the year-1 benefit has a greater negative effect. The separate valuation makes both effects visible and avoids treating the waiting period as a mere change in amount.

Payment Conditions and Practical Checks

A payment condition can be represented by specifying which future events qualify under each basis. Examples include a waiting period, survival to a specified duration, or a benefit payable only for a stated cause of decrement. For each event, assign a zero benefit when the condition is not met. If the condition changes the probability of an event itself, rather than merely whether a payment is made for that event, use the appropriate event probabilities under each basis. Do not assume the probabilities cancel unless the decrement assumptions are identical.

For a reserve calculated at duration $t$, use probabilities conditional on the policy being in force at duration $t$. For example, a benefit payable in the second year after the valuation date is weighted by the probability of surviving the first future year and then experiencing the covered event in the second. It is not weighted by survival from issue unless the reserve is being calculated at issue.

Common exam trap: Subtracting benefit amounts before checking whether their payment dates and eligibility conditions match. A benefit can have the same amount but a different present value when paid later; conversely, an increase in amount can be outweighed by excluding an earlier payment. Write the event, probability, amount, and payment time for each basis before calculating.

A final sign check is useful. If a benefit amount increases on the same events and at the same dates, with all else fixed, the reserve difference must be nonnegative. If the only change is deferring payments, the difference must be nonpositive for a positive interest rate. When both favorable and unfavorable changes occur, as in the waiting-period example, calculate the net difference rather than inferring its sign from only one component.

Key result: When only valuation benefits change, the reserve difference is the new prospective benefit value minus the old prospective benefit value. Value both streams using their own amounts, contingencies, and payment dates; identical premiums, expenses, and other assumptions cancel.

This method is aligned with the SOA Exam ALTAM (Advanced Long-Term Actuarial Mathematics) syllabus. It is especially useful when a problem changes one benefit feature while holding the remaining valuation basis fixed: isolating the affected future cash flows reduces unnecessary recalculation and makes the sign and source of the reserve change easier to audit.

Check Your Understanding

For each problem, calculate the reserve difference as new minus old and round the final monetary answer to the nearest cent.

A benefit is increased by $5,000 for death in year 1 only. At issue, $q=0.06$ and $i=0.04$. Find the reserve difference. (Answer: +$288.46)
A two-year contract has a $9,000 benefit for death in year 2. The new basis reduces that benefit by $2,000. Assume $q=0.07$, $p=0.93$ in each year, and $i=0.03$. Find the reserve difference. (Answer: −$122.73)
A $25,000 benefit for death during year 1 is paid at the end of year 1 under the old basis and at the end of year 3 under the new basis. Assume $q=0.04$ and $i=0.05$. Find the reserve difference. (Answer: −$88.54)
A two-year contract pays $8,000 for death in either year under the old basis. Under the new basis, year-1 death is excluded and the year-2 benefit is $10,000. Assume annual $q=0.12$, $p=0.88$, and $i=0.04$. Find the reserve difference. (Answer: −$727.81)
At duration 1, a benefit is increased by $3,000 only for death in the second future year. The conditional death probabilities for the first and second future years are 0.05 and 0.04, respectively; assume $i=0.05$. Find the reserve difference at duration 1. (Answer: +$103.40)

What You'll Learn