Pension Provisions as State-Dependent Cash Flows

Introduction

A pension plan provision is not just a benefit amount. It specifies who may receive a payment, what event makes it payable, and when it is paid. For example, a retirement pension may be payable only if a member reaches a stated retirement date while eligible. A death benefit may instead be payable if death occurs before that date. A survivor pension may continue after the member's death, but to a different recipient and perhaps at a different amount.

Valuation begins by translating this language into cash flows attached to member states and triggering events. The key is to keep three questions separate: Is the person eligible? Which event has occurred? At what time is the payment made? This tutorial develops that representation and uses it to calculate expected present values. All plans and data below are hypothetical.

Definition: A state-dependent cash flow is a payment whose amount, recipient, or existence depends on the member's status at a specified time. A trigger is the event that makes a provision payable, such as retirement, death, or disability. Eligibility is a condition that must be satisfied for the provision to apply; eligibility alone need not cause a payment.

Representing Provisions as Conditional Payments

For a simple annual model, let the possible states at the end of year $t$ include active, retired and alive, disabled, and dead. The plan's terms determine which states or transitions produce payments. For instance, a retirement pension might be payable when an eligible active member retires. A death-in-service lump sum might be payable when an active member dies. A survivor pension might be payable to a spouse in years after the member's death.

At each payment date, list the possible outcomes and the payment in each outcome. If the outcomes are mutually exclusive, the expected payment is the sum of each outcome's probability multiplied by its payment. A payment that is not due in a particular outcome is recorded as zero. For payment time $t$, let $C_t$ denote the random payment and let $i$ be the effective annual interest rate. Under a deterministic interest assumption, the expected present value at time 0 is:

\operatorname{EPV}=\sum_{t=1}^{n} v^t E[C_t], \qquad v=\frac{1}{1+i}.

This formula assumes that cash flows are valued at the stated annual dates, that $i$ is constant and effective per year, and that each payment amount and its probability are defined consistently. If a payment occurs immediately at an event time within a year, discounting it as if it occurred at year-end is not exact unless the model explicitly places it at year-end or supplies a suitable within-year timing assumption.

Modeling rule: Write a benefit as a payment amount together with its eligibility condition, trigger, recipient, and payment date. Then identify the probability of the complete condition being met by that date. Do not treat a trigger probability as an eligibility probability unless the problem says they are the same.

For a provision paying $b$ at time $t$ only if event $G$ has occurred and eligibility condition $H$ is satisfied, the expected payment is $bP(G\cap H)$. If the event probability is conditional on eligibility, use the product rule $P(G\cap H)=P(H)P(G\mid H)$. The conditioning matters: multiplying or adding probabilities from different conditional populations can produce an incorrect payment expectation.

Why the Present-Value Formula Works

The valuation formula follows from linearity of expectation. For a finite set of payment dates, the present value is the random sum $\sum_t v^t C_t$. Therefore:

E\left[\sum_{t=1}^{n}v^tC_t\right] =\sum_{t=1}^{n}E[v^tC_t] =\sum_{t=1}^{n}v^tE[C_t].

The last equality holds because $v^t$ is fixed under the deterministic interest assumption. No independence between payments at different dates is required for this expectation identity. Dependence still matters when finding each $E[C_t]$: for example, the probability of a survivor payment in year 3 depends on what happened to the member before then.

State the payment date.
Fix when the amount is due, such as at the end of each year.

List the outcomes at that date.
Use mutually exclusive states or triggering events and include a zero payment where no benefit is due.

Apply eligibility and trigger conditions.
Calculate the probability of the full condition required for each payment.

Calculate and discount expected payments.
Multiply each payment by its probability, then discount from its own payment date.

Worked Valuations

Worked Example: Competing Events at the End of Year 1

A hypothetical active member is subject to three mutually exclusive year-end outcomes: death with probability 0.02, disability with probability 0.08, and remaining active with probability 0.90. The plan pays a $10,000 death benefit if death occurs and a $1,500 benefit if disability occurs. No payment is made for remaining active. Each benefit is paid at the end of year 1. The effective annual interest rate is 4%. Find the expected present value.

Method. First calculate the payment expectation by outcome. The year-1 expected payment is:

E[C_1]=0.02(10{,}000)+0.08(1{,}500)+0.90(0) =200+120 =320.

The payment is at the end of year 1, so it is discounted once:

\operatorname{EPV}=\frac{320}{1.04} =307.6923\ldots =\$307.69.

Check. The benefit is paid only in the death or disability outcome, whose combined probability is 0.10. The conditional average payment across those two outcomes is $320/0.10=\$3{,}200$, so the unconditional expected payment is $0.10(\$3{,}200)=\$320$, as calculated. The expected present value is $307.69 to the nearest cent.

Common mistake avoided. Adding the two benefit amounts and multiplying by the combined event probability would give $0.10(10{,}000+1{,}500)=\$1{,}150$. That treats both benefits as payable in the same outcome. Because death and disability are mutually exclusive here, each amount must instead be weighted by its own probability.

Worked Example: A Pension and a Later Survivor Payment

In a hypothetical plan, 60% of a group of members are eligible and retire at the end of year 1. Each retiree is alive at the end of year 2 and then receives a $1,900 pension payment. Between the ends of years 2 and 3, each retiree has a 0.10 probability of dying. If the retiree is alive at the end of year 3, the member receives $1,900; if the member has died, the spouse receives $1,000. Assume the spouse survives through year 3. Payments are made at the ends of years 2 and 3. The effective annual interest rate is 4%. Find the expected present value per original group member.

Method. At the end of year 2, the payment occurs only for the 0.60 who retired. Thus:

E[C_2]=0.60(1{,}900)=1{,}140.

At the end of year 3, first condition on having retired. The member is alive with probability 0.90 and dead with probability 0.10. The expected year-3 payment for a retiree is:

$$ 0.90(1{,}900)+0.10(1{,}000) =1{,}710+100 =1{,}810. $$

Only 0.60 of the original group retired, so the unconditional expected year-3 payment is $E[C_3]=0.60(1{,}810)=\$1{,}086$. Discount the payments from their respective dates:

\begin{aligned} \operatorname{EPV} &=\frac{1{,}140}{(1.04)^2}+\frac{1{,}086}{(1.04)^3}\\ &=1{,}053.9941+965.4500\\ &=2{,}019.4441. \end{aligned}

The expected present value is $2,019.44 per original group member, to the nearest cent. The calculation does not assume a year-3 pension for someone who did not retire; it includes the spouse payment only for a retiree who died during the stated interval.

Common mistake avoided. The 0.10 death probability is conditional on being a retiree who reached year 2 alive. It is not the death probability for the entire original group. The year-3 expectation must include the 0.60 retirement probability as well as the conditional death and survival probabilities.

Worked Example: Following Survivor Payments Across Two Years

A hypothetical plan has a 0.40 probability that a member retires at the end of year 1. If retired, the member has a 0.25 annual probability of death in each subsequent year, independently from year to year. At the ends of years 2 and 3, the plan pays $1,200 to a retired member who is alive or $600 to the spouse if the member has died. Assume the spouse survives and is eligible for both scheduled payments after the member's death. The effective annual interest rate is 5%. Find the expected present value per original member.

Method. At the end of year 2, conditional on retirement, the member is alive with probability 0.75 and dead with probability 0.25. The unconditional expected payment is:

E[C_2]=0.40\left[0.75(1{,}200)+0.25(600)\right] =0.40(900+150) =420.

At the end of year 3, conditional on retirement, the member survives both years with probability $0.75^2=0.5625$. The probability of death by the end of year 3 is $1-0.5625=0.4375$. The survivor receives the payment in the death state, so:

\begin{aligned} E[C_3] &=0.40\left[0.5625(1{,}200)+0.4375(600)\right]\\ &=0.40(675+262.50)\\ &=375. \end{aligned}

The expected present value is:

\begin{aligned} \operatorname{EPV} &=\frac{420}{(1.05)^2}+\frac{375}{(1.05)^3}\\ &=380.9524+323.9391\\ &=704.8915. \end{aligned}

The expected present value is $704.89 per original member, to the nearest cent. As a check on the year-3 probabilities, the alive and dead probabilities conditional on retirement sum to $0.5625+0.4375=1$.

Common mistake avoided. Applying a single-year survival probability of 0.75 at year 3 ignores the first post-retirement year. The member must survive both years to receive the member payment at the end of year 3; otherwise the spouse payment applies under the stated plan terms.

Timing, Eligibility, and Common Traps

A sound valuation keeps the benefit definition separate from the event probabilities. Eligibility can be established before a trigger occurs, can depend on the member's status at the trigger date, or can fail if an earlier event occurs. The probability used in the expected payment must match the exact wording of the provision. If a death benefit applies only to an eligible active member, for example, the relevant event is not simply death: it is death while active and eligible.

Provision wording	Payment condition to model	Common error
Retirement pension payable on eligible retirement	Eligibility and retirement both hold by the payment date	Paying the pension merely because the member was eligible earlier
Death-in-service benefit	Death occurs while the member is still in the covered active status	Including deaths after retirement if the provision excludes them
Survivor pension after member death	The member has died and the survivor meets the stated conditions	Stopping all payments at member death despite a survivor provision
Payment at a specified year-end	Discount the expected payment from that year-end	Discounting one year too many or too few

For mutually exclusive states, add probability-weighted payments. For sequential conditions, use conditional probabilities or joint probabilities. For example, if a probability $r$ of retirement is followed by conditional survival probability $s$, then the probability of both retirement and survival is $rs$, not $r+s$. A survivor payment needs its own amount and payment rule; it is not automatically equal to the member's pension or automatically zero.

The expected present value is an average over possible outcomes, not a promise that a particular member will receive that average. Its usefulness is that it places contingent provisions on a common financial basis while retaining their distinct timing and triggers. In more detailed pension valuations, the same representation can be combined with richer eligibility rules and decrement probabilities; the initial task remains an accurate statement of who receives what, and when.

Key takeaway: For each payment date, determine the eligible population, identify mutually exclusive states or triggering events, assign the correct payment in each outcome, and then discount the expected payment from that date. This structure prevents benefit amounts, probabilities, and timing from being mixed together.

Check Your Understanding

For each problem, state the relevant conditional event before calculating. Amounts are hypothetical and payments occur at the specified year-ends.

An active group has year-1 probabilities 0.01 of death, 0.04 of disability, and 0.95 of remaining active. The respective death and disability benefits are $20,000 and $2,000; no payment is made for remaining active. At an effective annual interest rate of 3%, find the expected present value. (Answer: $271.84)
At the end of year 2, 50% of a group become eligible for a fixed $2,400 pension payment at the end of year 3. Assume all eligible members receive it and no other payments are made. At an effective annual interest rate of 4%, find the expected present value per original member. (Answer: $1,066.80)
A member retires at the end of year 1 with probability 0.40. At the end of year 2, a retiree alive receives $1,500 and a spouse receives $750 if the member has died; conditional on retirement, the member's probability of surviving to year 2 is 0.90. At an effective annual interest rate of 5%, find the expected present value of the year-2 payment per original member. (Answer: $517.01)
A member retires at the end of year 1 with probability 0.25. Conditional on retirement, the member's annual death probability is 0.20 in each of the next two years, independently. At the ends of years 2 and 3, the member receives $1,000 if alive and the spouse receives $400 if the member has died. The spouse survives. At an effective annual interest rate of 4%, find the expected present value per original member. (Answer: $377.65)

What You'll Learn