Deriving the Annual Participating Reserve Recursion

Why an Annual Reserve Recursion Is Useful

A prospective reserve expresses the value at a given time of future policy cash flows. The annual reserve recursion gives a second way to relate reserves at successive policy anniversaries: it balances the funds available during the year against the expected value of the outgo at year-end. For a participating policy, the outgo may include a death benefit with declared additions, a cash dividend on survival, and the reserve needed for a surviving policy.

The recursion is only valid when its timing and valuation assumptions are explicit. In particular, a premium received at the start of the year is not treated like a benefit paid at the end of the year, and a survival payment is weighted by the probability of surviving the year. The derivation below uses one stated annual convention throughout.

Definition: Let $V_t$ be the reserve immediately before the premium at time $t$, conditional on the policy being in force. The premium $P_t$ is received at time $t$, and an expense $E_t$ is incurred at that time. Let $i$ be the effective annual interest rate on the valuation basis. Write $q_t$ for the probability of death during the year, conditional on being in force at time $t$, and $p_t=1-q_t$ for the probability of surviving to time $t+1$. The death payment $B_t$, any cash dividend $D_t$ paid on survival, and the reserve $V_{t+1}$ for a surviving policy are all valued at time $t+1$. The reserve $V_{t+1}$ is immediately before any premium due at that time.

This is an annual discrete model: it places death benefits at the end of the policy year, even though actual claims may be paid at other times. It assumes the same valuation interest rate applies throughout the year, and that death and survival are the only outcomes. Lapses, partial-year claim timing, and stochastic investment returns require additional terms or a different timing model. The amounts $B_t$ and $D_t$ are defined by the policy cash flows: for example, $B_t$ can include a declared addition payable on death, while $D_t$ is a cash amount paid on survival.

Deriving the Recursion

At time $t$, the reserve and the net premium received after the time-$t$ expense provide $V_t+P_t-E_t$ to support the policy. Under the valuation basis, these funds accumulate for one year at rate $i$. At time $t+1$, the policy has one of two outcomes. If death occurs, the insurer pays $B_t$. If the policy survives, the insurer pays $D_t$ and holds $V_{t+1}$ for the future obligations on that policy.

Equating the accumulated funds to the expected year-end outgo gives the annual reserve recursion:

(V_t+P_t-E_t)(1+i) = q_t B_t+p_t(D_t+V_{t+1}).

The probabilities apply to the entire outgo associated with each outcome. The death payment is weighted by $q_t$; both the survival dividend and the next-year reserve are weighted by $p_t$. The equation is an equivalence of values conditional on the policy being in force at time $t$, not a claim that the same amount is paid in every individual policy experience.

Identify time-$t$ cash flows.
Start with $V_t$, add the premium $P_t$, and subtract the expense $E_t$, using the stated convention that the premium and expense occur at time $t$.

Accumulate to the year-end.
Multiply the time-$t$ funds by $1+i$. This step uses the effective annual valuation rate and the assumption that the funds are invested for a full year.

Write the outcome-specific outgo.
For death, use $B_t$. For survival, use $D_t+V_{t+1}$. Weight these amounts by $q_t$ and $p_t$, respectively.

Equate and solve.
Set accumulated funds equal to expected year-end outgo. Rearrange for the reserve that the question asks for, retaining the same timing convention in every term.

The equation can be rearranged in either direction. Solving for the current reserve gives:

V_t = \frac{q_t B_t+p_t(D_t+V_{t+1})}{1+i} -P_t+E_t.

Alternatively, solving for the next-year reserve gives:

V_{t+1} = \frac{(V_t+P_t-E_t)(1+i)-q_tB_t}{p_t} -D_t, \qquad p_t>0.

The first form discounts expected year-end outgo to time $t$, then allows for the time-$t$ premium and expense. The second form isolates the survival outgo and divides by $p_t$, because $V_{t+1}$ is payable as a reserve only in the survival outcome. If $p_t=0$, the second form cannot determine $V_{t+1}$: there are no surviving policies in this one-year model to which that reserve applies.

Worked Applications

Worked Example: Solving for the Next-Year Reserve

A hypothetical participating policy has reserve $V_t=\$79$, an annual premium of $\$30$, and a time-$t$ expense of $\$5$. The valuation interest rate is $i=4\%$. The probability of death during the year is $q_t=0.05$, so $p_t=0.95$. The end-of-year death payment is $\$400$, and a cash dividend of $\$8$ is paid at year-end on survival. Find $V_{t+1}$.

Method. Apply the accumulated-funds form of the recursion. The time-$t$ funds after the premium and expense are $(79+30-5)=\$104$. Accumulating for one year gives $104(1.04)=\$108.16$.

The probability-weighted death outgo is $0.05(400)=\$20$. On survival, the outgo before probability weighting is $8+V_{t+1}$, so expected survival outgo is $0.95(8+V_{t+1})$. Equating the two sides:

108.16=20+0.95(8+V_{t+1}) =27.60+0.95V_{t+1}.

Therefore, $0.95V_{t+1}=108.16-27.60=80.56$, and $V_{t+1}=80.56/0.95=\$84.80$, to the nearest cent.

Check. The survival outgo is $8+84.80=\$92.80$. Expected year-end outgo is $0.05(400)+0.95(92.80)=20+88.16=\$108.16$, equal to the accumulated funds. This check also confirms that the dividend is included once, in the survival outgo.

Worked Example: Solving for the Current Reserve

A hypothetical participating policy has an annual premium of $\$60$, an expense of $\$10$ at the start of the year, and a valuation interest rate of $5\%$. The death probability is $q_t=0.04$, so the survival probability is $p_t=0.96$. The year-end death payment is $\$900$. On survival, a $\$20$ cash dividend is paid and the next-year reserve is $\$70$. Find $V_t$, and state the survival outgo before probability weighting.

Method. First combine the survival cash flows: the outgo conditional on survival is $20+70=\$90$. It is not yet probability-weighted. The expected death outgo is $0.04(900)=\$36$, and expected survival outgo is $0.96(90)=\$86.40$. Thus expected year-end outgo is $36+86.40=\$122.40$.

The reserve equation is $(V_t+60-10)(1.05)=122.40$, or $(V_t+50)(1.05)=122.40$. Dividing by $1.05$ gives $V_t+50=122.40/1.05=116.571428\ldots$, so $V_t=\$66.57$, rounded to the nearest cent. The survival outgo before probability weighting is $\$90$.

Check. Using the unrounded reserve, accumulated funds are $(66.571428\ldots+50)(1.05)=\$122.40$, matching expected year-end outgo.

Worked Example: Including a Declared Addition in the Death Payment

A hypothetical participating policy has $V_t=\$128.10$, an annual premium of $\$120$, and a time-$t$ expense of $\$20$. The valuation rate is $5\%$, $q_t=0.05$, and $p_t=0.95$. The death payment is $\$800$, including a declared addition of $\$40$. On survival, a cash dividend of $\$10$ is paid, and $V_{t+1}=\$200$. Check whether the reserve recursion balances, allowing for rounding of the stated reserve to cents.

Method. The time-$t$ funds after premium and expense are $128.10+120-20=\$228.10$. At the end of the year, these accumulate to $228.10(1.05)=\$239.505$.

The death outgo is the stated total of $\$800$, which already includes the declared addition. Its expected value is $0.05(800)=\$40$. Conditional on survival, the outgo is $10+200=\$210$, so its expected value is $0.95(210)=\$199.50$. Total expected year-end outgo is $40+199.50=\$239.50$.

Check. The two sides are $\$239.505$ and $\$239.50$, a difference of half a cent attributable to using a reserve rounded to cents. The unrounded reserve from the recursion is found from $(V_t+100)(1.05)=239.50$, giving $V_t=239.50/1.05-100=\$128.095238\ldots$, which rounds to $\$128.10$. The addition is not added again to the $\$800$ death payment. Any future liability for the declared addition is reflected in $V_{t+1}$ if the valuation basis requires it; it is not an extra current-year survival payment unless the policy pays one.

Timing, Participating Benefits, and Common Traps

The recursion depends on where each cash flow falls in the year. Here, premiums and expenses occur at time $t$, while the death payment, survival dividend, and $V_{t+1}$ are all at time $t+1$. If a problem instead places an expense or premium at year-end, it cannot simply be included in the time-$t$ amount $V_t+P_t-E_t$; it must be placed at its stated time and accumulated or discounted accordingly. Similarly, a death benefit paid at the moment of death generally cannot be treated as a year-end payment without an appropriate adjustment.

For participating benefits, separate the outcome amounts before applying probabilities. A cash dividend paid only on survival belongs in $D_t$. A declared addition payable on death belongs in $B_t$, as part of the total death payment. The next-year reserve is a separate survival outgo. This separation helps prevent counting a bonus both as part of the benefit and as an additional payment.

A frequent exam error is to use the survival outgo without multiplying by the probability of survival. In the second worked example, the incorrect calculation $0.04(900)+(20+70)=\$126$ treats the $\$90$ survival outgo as though it occurred with certainty. It would produce the incorrect reserve $126/1.05-60+10=\$70.00$. The correction is to weight that outgo by $0.96$: $0.04(900)+0.96(20+70)=\$122.40$. Hence $V_t=122.40/1.05-60+10=\$66.57$.

A useful final check is to substitute the proposed reserve back into the original, unrearranged equation. Confirm that premiums and expenses are placed at the correct time, that each conditional outgo is weighted by its corresponding probability, and that any declared addition is included exactly once. This check is especially valuable when solving for $V_{t+1}$, because the rearranged equation divides by $p_t$.

Key takeaway: Under the stated annual timing convention, accumulated current funds equal probability-weighted year-end outgo: $(V_t+P_t-E_t)(1+i)=q_tB_t+p_t(D_t+V_{t+1})$. The formula is useful only when the timing, valuation basis, and treatment of participating benefits are consistent on both sides.

Check Your Understanding

For each problem, use the stated annual timing convention: premiums and expenses are at the start of the year; benefits, dividends, and next-year reserves are at year-end. Round monetary answers to the nearest cent.

A hypothetical policy has $V_t=\$50$, premium $\$20$, expense $\$4$, $i=3\%$, $q_t=0.10$, death payment $\$300$, and survival dividend $\$5$. Find $V_{t+1}$. (Answer: $\$37.20$)
A hypothetical policy has premium $\$40$, expense $\$6$, $i=4\%$, $q_t=0.02$, death payment $\$1{,}000$, survival dividend $\$12$, and $V_{t+1}=\$150$. Find $V_t$. (Answer: $\$137.88$)
A hypothetical policy has $V_t=\$120$, premium $\$50$, expense $\$8$, $i=2\%$, $q_t=0.03$, death payment $\$600$, and survival dividend $\$10$. Find $V_{t+1}$. (Answer: $\$141.79$)
A hypothetical policy has premium $\$80$, expense $\$10$, $i=4\%$, $q_t=0.05$, death payment $\$500$, survival dividend $\$6$, and $V_{t+1}=\$100$. Find $V_t$. (Answer: $\$50.87$)
A hypothetical policy has $V_t=\$90$, premium $\$25$, expense $\$5$, $i=3\%$, $q_t=0.10$, death payment $\$250$, survival dividend $\$4$, and $V_{t+1}=\$80$. Calculate accumulated funds and expected year-end outgo, then state the expected surplus of accumulated funds over outgo. (Answer: accumulated funds $\$113.30$; expected outgo $\$100.60$; surplus $\$12.70$)

What You'll Learn