Termination Benefits and Exit-State Payments

Why the Payment Trigger Matters

A reserve for a policy in a covered state may include regular benefits payable while coverage continues and a lump sum payable when the policy leaves that state. These payments have different triggers: an occupancy benefit depends on time spent in the state, while an exit benefit depends on a transition. Valuing both correctly requires identifying the cash flow at the time it is due and the state immediately after a transition.

The same distinction applies whether an exit ends the contract or moves the policyholder into another covered state. If coverage continues in a destination state, the prospective reserve after the transition is also part of the value immediately before the transition. A lump sum paid at exit is not a substitute for that destination-state reserve.

Definition: An occupancy payment is payable because the policy occupies a specified state, often as a continuous rate per unit of time. An exit-state payment is a lump sum payable when a specified transition out of that state occurs. For a transition from state $i$ to state $j$, write $b_{ij}(t)$ for the payment at the transition time $t$. If the contract continues in state $j$, its reserve $V_j(t)$ must also be included in the value just before the transition.

Prospective Value of Occupancy and Exit Payments

Consider a continuous-time model with states, transition intensities, and a deterministic force of interest $\delta$. Let $c_i(s)$ be the rate of occupancy payments while the policy is in state $i$ at time $s$. Let $\mu_{ij}(s)$ be the transition intensity from $i$ to $j$, and let $b_{ij}(s)$ be the lump sum paid if that transition occurs. A transition intensity is a rate: over a short interval of length $h$, the conditional probability of transition from $i$ to $j$ is approximately $\mu_{ij}(s)h$.

For a finite horizon $T$, the prospective reserve at time $t$, given current state $i$, is the conditional expected present value of future payments. It includes continuous occupancy payments, lump sums at transitions, and any terminal payment specified at $T$. In general, the expected value of a transition payment depends on both the probability of occupying its starting state and the transition intensity out of that state.

V_i(t)=E\left[ \int_t^T e^{-\delta(s-t)}c_{X_s}(s)\,ds +\sum_{\tau\in(t,T]} e^{-\delta(\tau-t)}b_{X_{\tau-},X_\tau}(\tau) +e^{-\delta(T-t)}G_{X_T} \;\middle|\; X_t=i \right].

Here $X_s$ is the state at time $s$, $\tau$ ranges over transition times, and $G_{X_T}$ is any terminal payment associated with the state at the horizon. Payments that do not apply to a particular state or transition are set to zero. If a transition is followed by continued coverage, its resulting future payments are included through the later part of the expectation; equivalently, the destination reserve is part of the transition's continuation value.

Key result: In a one-state covered period with constant exit force $\lambda$, constant occupancy payment rate $c$, constant exit payment $B$, no payment after exit, and horizon $n$, the value at the start is $$ \int_0^n e^{-(\delta+\lambda)s}(c+\lambda B)\,ds =(c+\lambda B)\frac{1-e^{-(\delta+\lambda)n}}{\delta+\lambda}, \qquad \delta+\lambda\ne 0. $$ The term $c$ is paid per unit of time while coverage continues. The term $\lambda B$ is the expected exit-payment rate conditional on still being covered. This expression assumes constant forces and no additional payments or reserve after exit.

Why the Exit Term Has a Transition Intensity

Suppose the policy begins in a covered state and can only leave it through one exit, with constant force $\lambda$. The probability of still being covered at time $s$ is $e^{-\lambda s}$. Therefore the expected discounted occupancy payment over a small interval of length $ds$ around $s$ is $e^{-\delta s}e^{-\lambda s}c\,ds$.

Conditional on still being covered at time $s$, the probability of exiting in the next $ds$ is approximately $\lambda ds$. The expected discounted exit payment during that interval is therefore approximately $e^{-\delta s}e^{-\lambda s}\lambda B\,ds$. Adding the two flows and integrating from 0 to $n$ gives

\int_0^n e^{-(\delta+\lambda)s}c\,ds+ \int_0^n e^{-(\delta+\lambda)s}\lambda B\,ds =(c+\lambda B)\frac{1-e^{-(\delta+\lambda)n}}{\delta+\lambda}.

This derivation explains why the exit amount is multiplied by the transition force in a continuous-time valuation. The amount $B$ is paid only when an exit occurs; it is not a continuous benefit rate. If the policy remains covered beyond $n$, or if an exit leads to another covered state, the relevant survival, transition, terminal, or destination-reserve terms must be included instead.

Worked Example: Continuous Occupancy Benefit and Exit Lump Sum

A hypothetical contract begins in a covered state and remains there until exit. The exit force is constant at $\lambda=0.20$ per year. It pays continuously at $\$90$ per year while covered and pays $\$1{,}000$ immediately upon exit. The valuation force of interest is $\delta=0.03$ per year, and the horizon is two years. Assume no payment after exit and no other terminal benefit. Find the value at the start.

Method. Use the constant-force prospective value formula. The probability of remaining covered to time $s$ is $e^{-0.20s}$, and discounting contributes $e^{-0.03s}$. Thus the combined factor is $e^{-0.23s}$. The expected payment rate while covered, including the exit payment, is $90+0.20(1{,}000)=290$.

V(0)=\int_0^2 e^{-0.23s}\bigl(90+0.20(1{,}000)\bigr)\,ds =290\frac{1-e^{-0.46}}{0.23}.

Using $e^{-0.46}\approx0.6312836455$, the integral factor is \[ \frac{1-0.6312836455}{0.23}\approx1.603114585. \] The occupancy-payment value is $90(1.603114585)\approx\$144.2803$. The exit-payment value is $200(1.603114585)\approx\$320.6229$. Their total is approximately $\$464.9032$, or $\$464.90$ to the nearest cent.

Check. Adding the separately valued components gives $\$144.2803+\$320.6229=\$464.9032$, agreeing with the combined integral. The calculation avoids treating the $\$1{,}000$ lump sum as though it were payable continuously: its expected rate while covered is $\lambda B=\$200$ per year.

Annual Models: Payments, Transitions, and Reserves

In an annual model, list the possible outcomes over the year and attach the correct cash flow to each. An occupancy benefit may be payable only if the policy remains in a specified state, while a transition payment is payable only if the corresponding transition occurs. If an outcome leads to an ongoing state, include that state's reserve at year-end. With an effective annual interest rate $i$, a premium paid immediately at the valuation time is not discounted; subtract it after discounting the expected year-end value.

V_i(0)=\frac{E[\text{year-end payments and reserves}\mid X_0=i]}{1+i} -\text{premium paid immediately at time }0.

Worked Example: A Transition Benefit and a Benefit for Remaining in State

A hypothetical one-year contract starts in state $L$. At year-end, it remains in $L$ with probability $0.70$, moves to covered state $A$ with probability $0.20$, or terminates with probability $0.10$. The $L$-to-$A$ transition pays $\$1{,}000$, and termination pays $\$5{,}000$. A $\$600$ occupancy benefit is paid only if the policy remains in $L$ for the year. The year-end reserves are $V_L(1)=\$4{,}000$ and $V_A(1)=\$2{,}000$; after termination, the reserve is zero. The effective annual interest rate is $i=0.05$, and a premium of $\$1{,}500$ is paid immediately at time 0. Find the reserve before that premium.

Method. For each outcome, add the cash flow triggered by that outcome to its year-end reserve. The three conditional amounts are:

Remain in $L$: $\$600+\$4{,}000=\$4{,}600$.
Move from $L$ to $A$: $\$1{,}000+\$2{,}000=\$3{,}000$.
Terminate: $\$5{,}000+\$0=\$5{,}000$.

The expected year-end amount is $0.70(4{,}600)+0.20(3{,}000)+0.10(5{,}000)=3{,}220+600+500=\$4{,}320$. Its value at time 0 is $\$4{,}320/1.05=\$4{,}114.2857$. Subtract the immediately payable premium: $\$4{,}114.2857-\$1{,}500=\$2{,}614.2857$. The reserve before the premium is therefore $\$2{,}614.29$ to the nearest cent.

Check. Reweighting the three outcomes gives $0.70(4{,}600)+0.20(3{,}000)+0.10(5{,}000)=\$4{,}320$, as used above. This method avoids adding the occupancy benefit to transition outcomes: the $\$600$ is paid only on remaining in $L$.

Competing Exit Types

A covered state may have more than one possible exit, with a different force and payment for each. Under constant forces, the probability of still occupying the covered state at time $s$ depends on the sum of the exit forces. The expected payment rate from the exits is the sum of each exit force multiplied by its own payment. This allows a separate value to be assigned to each exit type without confusing its payment with an occupancy benefit.

Worked Example: Valuing Two Different Exit Payments

A hypothetical contract is in covered state $C$ for at most three years. It exits to state $R$ with constant force $0.12$ per year or terminates with constant force $0.08$ per year. Either exit ends payments under this contract, so there is no destination reserve. While in $C$, it pays continuously at $\$120$ per year. It pays $\$400$ on exit to $R$ and $\$2{,}500$ on termination. The valuation force is $\delta=0.04$ per year. Find the value at the start.

Method. The total exit force is $0.12+0.08=0.20$, so the probability of remaining in $C$ at time $s$ is $e^{-0.20s}$. Including interest, the discount-and-occupancy factor is $e^{-0.24s}$. The payment rate while in $C$, including the expected transition payments per unit time, is $120+0.12(400)+0.08(2{,}500)=120+48+200=\$368$.

V(0)=368\int_0^3 e^{-0.24s}\,ds =368\frac{1-e^{-0.72}}{0.24}.

Using $e^{-0.72}\approx0.4867522560$, the integral factor is $(1-0.4867522560)/0.24\approx2.138532267$. The occupancy-payment value is $120(2.138532267)\approx\$256.6239$. The value of exit payments to $R$ is $48(2.138532267)\approx\$102.6495$, and the termination-payment value is $200(2.138532267)\approx\$427.7065$. Adding them gives $\$786.9799$, or $\$786.98$ to the nearest cent.

Check. The three payment rates sum to $120+48+200=368$, and the separate values sum to approximately $\$786.9799$, agreeing with the combined calculation. The result assumes both exits end the contract; if the policy continued after moving to $R$, the reserve in $R$ would need to be included.

Common Mistake: Applying a Benefit to the Wrong Outcome

In the annual example, a tempting but incorrect calculation is to omit the $\$600$ occupancy benefit because it is not an exit payment. That gives an expected year-end amount of $0.70(4{,}000)+0.20(1{,}000+2{,}000)+0.10(5{,}000)=\$3{,}900$, and a reserve before premium of $\$3{,}900/1.05-\$1{,}500=\$2{,}214.29$. This is too low because the benefit is payable when the policy remains in $L$; it is a real contract payment even though no transition occurs.

The correction is to attach every payment to its actual trigger. Add $\$600$ only to the remain-in-$L$ outcome, add the applicable lump sum to each exit outcome, and include the reserve in any continuing destination state. In the example, this restores the expected year-end amount to $\$4{,}320$ and the reserve before premium to $\$2{,}614.29$.

Key takeaway: At each possible payment time, identify the state occupied, whether a transition occurred, the payment triggered by that event, and any reserve remaining after it. Discount the resulting cash flows using their actual timing. A payment for continued occupancy and a lump sum on exit are separate cash flows, even when both are valued from the same starting state.

Check Your Understanding

For each problem, state the payment timing carefully before calculating.

A contract remains in one covered state until exit. The constant exit force is $0.10$ per year, the continuous occupancy benefit is $\$50$ per year, and the exit payment is $\$300$. The valuation force is $\delta=0.02$, and the horizon is two years. There are no payments after exit. Find the value at the start. (Answer: $\$142.25$)
A one-year model starts in state $X$. It remains in $X$ with probability $0.60$, moves to $Y$ with probability $0.30$, and terminates with probability $0.10$. A $\$200$ benefit is paid only on remaining in $X$; the $X$-to-$Y$ transition pays $\$500$, and termination pays $\$1{,}000$. Year-end reserves are $\$700$ in $X$, $\$900$ in $Y$, and zero after termination. With $i=0.04$, find the reserve before an immediate $\$100$ premium. (Answer: $\$919.23$)
A contract has one covered state with two exits, at constant forces $0.10$ and $0.05$ per year. Their respective exit payments are $\$200$ and $\$1{,}000$. The continuous occupancy benefit is $\$80$ per year, the valuation force is $\delta=0.03$, and the horizon is two years. Both exits end the contract. Find the value at the start. (Answer: $\$251.94$)
In an annual model, explain why an occupancy benefit payable only on remaining in state $L$ should be included in the remain-in-$L$ outcome, rather than in every outcome or omitted from the reserve calculation. (Answer: It is triggered by continued occupancy, so include it only on the remain-in-$L$ outcome; it is not payable on exits.)
A covered state has a transition to another covered state, with a lump sum payable at the transition. What additional value must be included at the transition when calculating the reserve just before it? (Answer: Include the reserve in the destination state, in addition to the transition lump sum.)

What You'll Learn