State-Indexed Cash-Flow Classification

Why the Cash-Flow Trigger Matters

A contract may promise a premium while coverage is active, a lump sum when disability begins, or a payment on an anniversary if the insured is then receiving long-term care. These amounts may all be called “cash flows,” but they are governed by different contractual triggers. The trigger determines when a payment is due and what facts must hold for it to be due.

Getting the trigger right is essential before valuing the contract. A state-contingent payment depends on the state occupied at a specified time or over a period. A transition payment depends on a change from one state to another. A fixed-time payment is tied to a stated date, possibly with a state condition. A termination payment is triggered by the contract ending. These categories describe contractual rules; they are not four different methods of discounting or valuation.

Definition: Let $J(t)$ denote the state occupied at time $t$. A state-triggered cash flow is due because the contract is in a specified state. A transition-triggered cash flow is due because a specified change of state occurs. A fixed-time cash flow is due at a specified time, with any contractual eligibility condition checked at that time. A termination cash flow is due because coverage ends under a specified termination cause, such as death, surrender, or maturity.

For clarity, the examples use the insurer’s perspective: benefits and expenses paid by the insurer are positive outflows, and premiums received are negative outflows. A contract’s cash-flow rule must identify the amount, the trigger, and any conditions on the state or event. Naming only the amount is not enough.

Representing Each Trigger Precisely

A state-triggered rule asks which state is occupied. For example, “pay 400 dollars for each scheduled monthly payment date on which the insured is in the care state” is based on $J(t)=C$ at each stated date. A transition-triggered rule instead asks whether a particular jump occurred: “pay 2,000 dollars when the insured first moves from active to disabled.” Being disabled at a later time does not, by itself, satisfy that transition trigger.

A fixed-time rule names a date, such as a policy anniversary. It may be unconditional, or it may require a state condition at that date. “Pay 1,000 dollars at the second anniversary” and “pay 1,000 dollars at the second anniversary if the insured is in the care state” are different rules. A termination rule names both the ending event and the payment made because of it. For instance, a surrender value is triggered by surrender, not simply by being in an active state.

\text{State at }t:\quad b_j(t)\mathbf{1}\{J(t)=j\}, \qquad \text{transition }i\to j:\quad b_{ij}(t)\text{ when that jump occurs.}

Here $\mathbf{1}\{\cdot\}$ equals 1 when the stated condition is true and 0 otherwise. For transitions, let $N_{ij}(t)$ count jumps from state $i$ to state $j$ up to time $t$. A payment triggered by a specified jump can be represented by an amount multiplied by the increment in that count at the jump time. This notation makes the contractual distinction explicit: a state rule tests occupancy, while a transition rule tests an event.

Worked Example: Separate State and Transition Triggers

A hypothetical policy has states active $A$, disabled $D$, and long-term care $C$. It receives a 70-dollar premium at each of two specified dates, $t=0.5$ and $t=1$, if the insured is active on that date. It pays 2,000 dollars on the first $A\to D$ transition and 3,000 dollars on the first $D\to C$ transition. It also pays 1,000 dollars at $t=1$ if the insured is in state $C$. The path is $A$ until $t=0.3$, $D$ from $t=0.3$ until $t=0.7$, and $C$ thereafter. Classify the rules and calculate the cash flows on this path.

Assumptions and parameters. This is a deterministic path, with no probability distribution or mortality model applied. The effective annual interest rate is 0%, so amounts are counted without discounting. The premium is 70 dollars per eligible date; the transition benefits are 2,000 and 3,000 dollars; the fixed-time benefit is 1,000 dollars.

Step 1: Classify the premium. The premium depends on the state at each named date, so it is state-triggered at fixed payment dates. At $t=0.5$, the state is $D$, so no premium is received. At $t=1$, the state is $C$, so no premium is received. Total premiums received are $0+0=\$0$.

Step 2: Classify and check the transition benefits. The move $A\to D$ occurs at $t=0.3$, so the 2,000-dollar transition benefit is due. The move $D\to C$ occurs at $t=0.7$, so the 3,000-dollar transition benefit is due. These amounts are not triggered merely by being in $D$ or $C$; each requires the specified change of state.

Step 3: Check the fixed-time benefit. The payment date is $t=1$, and $J(1)=C$. The condition is satisfied, so the 1,000-dollar benefit is due. It is fixed-time and state-conditioned, rather than a transition benefit.

Step 4: Total the signed cash flows. Benefits paid are $2{,}000+3{,}000+1{,}000=\$6{,}000$. Premium receipts are zero, so the net outflow is $\$6{,}000-\$0=\$6{,}000$.

Check and method. Each rule was tested against its own trigger: the two premium dates, the two specified jumps, and the state at the anniversary. The common mistake avoided is treating the fixed-time benefit as payable merely because a transition into $C$ occurred; its stated test is instead the state at $t=1$.

Why Indicator Rules Work

The indicator representation follows directly from the contract’s conditional wording. If an amount $b$ is payable exactly when condition $E$ holds, its pathwise cash flow is $b\mathbf{1}\{E\}$. When $E$ is true, the indicator is 1 and the amount is $b$; when $E$ is false, the indicator is 0 and no amount is paid. This is a bookkeeping identity, not a probability approximation.

C_k=b_k\mathbf{1}\{J(t_k)=j\} \quad\Longrightarrow\quad \sum_k C_k =\sum_k b_k\mathbf{1}\{J(t_k)=j\}.

For example, a benefit of 500 dollars at each of three specified dates when the insured is in state $C$ produces one term for each date. A date with $J(t_k)\ne C$ contributes zero; a date with $J(t_k)=C$ contributes 500 dollars. By contrast, for a payment on an $i\to j$ transition, the condition is about a jump, not occupancy. Counting occupancy in state $j$ cannot determine how many qualifying entries occurred: the policy may enter $j$, leave it, and return.

Termination can also be written as a conditional rule. If $T$ is the contract’s termination time and $R$ is the termination cause, a payment schedule may specify an amount $g(R,T)$. The cause matters: surrender, death, and maturity need not produce the same amount. A termination payment is therefore not interchangeable with an ordinary benefit payable while coverage continues.

Worked Example: Checking Every Fixed Premium Date

A hypothetical income policy charges a 120-dollar premium at each quarter-year date from $0.25$ through $2.00$, inclusive, provided the insured is active $A$ on that date. The contract pays 1,500 dollars on the first $A\to D$ transition and 1,800 dollars on the first entry into long-term care $C$. The state path is $A$ on $[0,0.4)$, $D$ on $[0.4,0.9)$, $C$ on $[0.9,1.4)$, $D$ on $[1.4,1.8)$, and $A$ from $1.8$ through $2.0$. Determine the premium total and the net undiscounted outflow.

Assumptions and parameters. The state path is given deterministically; no probability distribution or mortality model is used. The effective annual interest rate is 0%, and all amounts are in dollars. The premium is 120 dollars per eligible date; the two transition benefits are 1,500 and 1,800 dollars.

Step 1: List the due dates. The eight dates are $0.25, 0.50, 0.75, 1.00, 1.25, 1.50, 1.75,$ and $2.00$.

Step 2: Determine the state at each date. Applying the stated intervals gives $A,D,D,C,C,D,D,A$, respectively. In particular, $1.75$ lies in $[1.4,1.8)$, so the state then is $D$, not $A$. The active dates are $0.25$ and $2.00$, exactly two dates.

Step 3: Calculate premiums. Premiums received are $2\times \$120=\$240$.

Step 4: Calculate transition benefits. The first $A\to D$ transition is at $t=0.4$, so the 1,500-dollar benefit is due. Entry into $C$ occurs at $t=0.9$, so the 1,800-dollar benefit is due. Total benefits are $\$1{,}500+\$1{,}800=\$3{,}300$.

Step 5: Find net outflow and check. With benefits positive and premiums as receipts offsetting outflow, the net undiscounted outflow is $\$3{,}300-\$240=\$3{,}060$. The count of two premium dates agrees with the state list: only its first and last entries are $A$.

Method and common mistake. Classify each premium as a state test at a named date, then evaluate the path at every one of those dates. The common error is to misread the state at $1.75$ as active; it is disabled because $1.4\leq1.75<1.8$. That error would add an unjustified 120-dollar premium.

Termination Is a Distinct Contractual Trigger

A contract can end for different reasons, and each reason can have a different settlement. The word “termination” identifies the general trigger—coverage ends—but the rule must still identify the cause and the amount. It is also important to distinguish the termination amount from payments governed by the state immediately before termination. For instance, a care benefit may be payable because the insured is in $C$, while a surrender value is payable because the policyholder surrenders and the contract ends.

A termination event may be represented by a transition to an absorbing terminated state $X$, but the contractual classification should retain the cause. A move to $X$ following death is not necessarily equivalent to a move to $X$ following surrender. If the contract specifies different amounts, the cash-flow rule must distinguish those events. A fixed maturity date can likewise be both a specified time and a termination trigger; its contractual consequence is determined by the maturity provision, not by labelling alone.

Worked Example: Distinguishing Termination Causes

A hypothetical policy has the following termination provisions: surrender produces a payment of 8,400 dollars and an associated termination expense of 250 dollars; death produces a benefit of 60,000 dollars and an associated termination expense of 500 dollars; maturity produces a payment of 15,000 dollars with no termination expense. Compare the net outflow under each separate termination scenario.

Assumptions and parameters. Each scenario is evaluated separately and deterministically; no probability distribution or mortality assumption is used to assign likelihoods. The effective annual interest rate is 0%, so there is no discounting. The listed payments and expenses are the complete termination amounts for this comparison.

Step 1: Classify the surrender case. The 8,400-dollar amount is triggered by surrender, and the 250-dollar expense is also triggered by that termination. Net outflow is $\$8{,}400+\$250=\$8{,}650$.

Step 2: Classify the death case. Death terminates the policy and triggers the 60,000-dollar benefit. The associated expense adds 500 dollars, giving net outflow $\$60{,}000+\$500=\$60{,}500$.

Step 3: Classify the maturity case. Maturity is the specified termination event. The payment is 15,000 dollars, and there is no termination expense, so net outflow is $\$15{,}000+\$0=\$15{,}000$.

Check and method. The three results are tied to three separate causes: surrender, death, and maturity. They should not be combined or treated as alternative labels for one common payment. The common mistake avoided is applying the surrender value to every policy termination, even when the contract specifies a different death or maturity amount.

Classification Checks and Common Traps

When translating contract wording into a model, first identify the trigger phrase, then state the required condition precisely. “While disabled” refers to state occupancy; “on becoming disabled” refers to a transition; “on each anniversary” specifies a date; “on surrender” specifies a termination cause. A phrase may contain more than one condition, as in “on the anniversary if disabled,” so both the date and the state test must be preserved.

Confusing occupancy with entry. “Pay while in $C$” is not the same as “pay on entry into $C$.” Remaining in $C$ can satisfy the first rule without creating another entry event.
Ignoring a stated date. A payment conditional on the state at an anniversary must be checked at the anniversary, not at a nearby transition or at the start of the year.
Assuming termination causes are interchangeable. Use the specific death, surrender, or maturity provision that applies.
Forgetting the sign convention. When calculating net outflow from the insurer’s perspective, benefits and expenses increase outflow, while premium receipts reduce it.
Counting eligible dates from an approximate path description. Test every specified date against the state intervals, including interval endpoints as defined by the contract or model.

Key takeaway: A cash-flow rule is complete only when it states the amount, trigger, and all eligibility conditions. State rules test $J(t)$; transition rules test a specified jump; fixed-time rules test a named date and any conditions there; termination rules test the ending event and its cause. This classification is aligned with the SOA Exam ALTAM (Advanced Long-Term Actuarial Mathematics) syllabus.

Check Your Understanding

For each problem, classify the trigger and state the resulting cash flow or rule.

A hypothetical policy pays 600 dollars at time $t=1$ only if the insured is disabled at that time. The path is active until $t=0.8$, then disabled through $t=1.2$. Is the payment due, and what is its classification? (Answer: Yes; 600 dollars, fixed-time and state-conditioned.)
A care policy pays 2,400 dollars on the first transition $D\to C$. The path is $D\to C\to D\to C$. How many payments are due under the “first transition” wording? (Answer: One payment of 2,400 dollars.)
A hypothetical contract receives a 95-dollar premium at each of $t=0.5,1.0,1.5$ if the state is active. The states at those dates are $A,D,A$. What is the total premium received? (Answer: $2\times \$95=\$190$.)
At surrender, a policy pays a value of 7,250 dollars and incurs a 175-dollar termination expense. Under the insurer-outflow convention, what is the net outflow, and what triggers it? (Answer: $\$7{,}250+\$175=\$7{,}425$, triggered by surrender termination.)
A benefit is described as “1,100 dollars at the policy anniversary if the insured entered state $C$ during the preceding policy year.” At the anniversary the insured is in $D$, but did enter $C$ earlier in that year. Is the benefit due, and how should its trigger be classified? (Answer: Yes, 1,100 dollars; it is fixed-time and conditioned on a transition having occurred during the stated period, not on occupancy of $C$ at the anniversary.)

What You'll Learn