Why Pension State Diagrams Matter
A pension member’s status can change many times: employment may end, a deferred right may become payable, disability benefits may begin, or a pensioner may die. A state diagram records these possible changes in a consistent way. Its main purpose is not to calculate a pension amount; it is to make clear which status applies at a given time and what event changes it.
This matters because the same event can have different consequences depending on the member’s current status. Death while active, for example, is not the same transition as death after retirement. The diagram makes the starting status, triggering event, and resulting status explicit. The pension rules determine which transitions are allowed; the diagram represents those rules rather than replacing them.
Define the States Before Drawing Arrows
For a basic pension membership model, use \(A\) for active, \(V\) for vested deferred, \(D\) for disabled, \(R\) for retired, and \(X\) for deceased. These labels describe benefit status, not just employment or health. For example, “retired” means that a retirement pension has commenced, not merely that a person has stopped working.
\(A\), active: The member is alive, accruing service or otherwise treated as active under the plan, and is not classified as disabled, deferred, or retired.
\(V\), vested deferred: The member is alive, has a vested pension right that is deferred, and has not commenced a retirement pension. The member is not classified as active or disabled.
\(D\), disabled: The member is alive and is receiving, or has been approved for, the plan’s disability benefit status before a retirement pension has commenced. The plan must specify the relevant disability status.
\(R\), retired: The member is alive and a retirement pension has commenced.
\(X\), deceased: The member has died. This state takes precedence over all other states and is absorbing in this basic model: after death, the member does not return to \(A\), \(V\), \(D\), or \(R\).
The precedence rule is essential. A person who was retired and then dies is no longer classified as \(R\) at a time after death; the applicable current state is \(X\). Likewise, \(A\), \(V\), \(D\), and \(R\) apply only while the member is alive. This makes the states mutually exclusive. A member cannot simultaneously be classified as active and deferred, or as retired and deceased, under these definitions.
A plan may have an unvested leaver or another status that does not fit these five categories. Do not force that person into \(V\): an unvested leaver does not have a vested deferred pension right by definition. Add a separate state if that status is within the model’s scope, or state that the diagram covers only members who reach one of its defined states.
Constructing the Basic Diagram
For a basic model, the following transitions are common. Each arrow needs both a triggering event and a resulting state. The list is illustrative: a plan may permit fewer transitions or require additional states.
The “eligible re-employment” arrow is conditional. Some plans reinstate prior service; others preserve the deferred right while treating new service separately. If the member can simultaneously have active service and a separate deferred pension right, a single-state model may not retain enough information. One solution is to use a combined state, such as “active with a preserved deferred benefit”; another is to keep the active state and record the deferred right as an additional flag. The chosen representation must match the plan rules.
Similarly, the \(D \to R\) arrow represents a plan rule that converts disability status to retirement status when a retirement pension commences. It is not automatic for every plan. If a member recovers and resumes active service, \(D \to A\) may be appropriate only when the plan’s rules and the model’s scope treat that member as active again.
Why Transitions Must Be Labeled
A line between two states is not enough. The triggering event explains why the change occurs, and the resulting state confirms how the member is classified afterward. “Termination” alone may be ambiguous: termination may lead to \(V\) if the member is vested, or to an out-of-scope unvested-leaver state if not. The arrow label should therefore capture the condition that matters, such as “termination after vesting.”
A useful construction method is to take each defined state in turn, list the events that can occur while the member occupies it, and determine the state after each event. Then check that every arrow is permitted by the stated plan rules and that every included state has a clear definition.
Specify which members and benefit statuses the diagram represents. Decide whether unvested leavers and combined statuses are included.
Write a precise test for each state, including whether the member must be alive and whether a pension must have commenced.
Use events stated by the plan, such as vested termination, approved disability, retirement commencement, recovery, or death.
For each event, determine the post-event state. Apply the deceased-state precedence rule and any stated eligibility conditions.
Confirm that every arrow is justified, that no member is assigned to two states at once, and that the states preserve the information needed for the model’s purpose.
Justifying the Classification Rule
The mutually exclusive classification follows from the state definitions and the precedence rule, rather than from the drawing itself. First check whether the member is deceased. If so, assign \(X\), regardless of the member’s earlier status. If the member is alive, check whether a retirement pension has commenced; if so, assign \(R\). Otherwise, check disability status, then vested deferred status, then active status, using the definitions and any plan-specific eligibility rules.
This ordered check produces at most one classification: once a condition is met, the later cases do not apply. It also shows why definitions matter. If “retired” were defined only as “a retirement pension has commenced,” without saying the member must be alive or applying death precedence, a deceased pensioner would appear to meet both \(R\) and \(X\). The rule resolves that overlap by assigning \(X\) at times after death.
Worked Example: Tracing Re-employment and Disability
Method: Start with the member’s defined state, then apply each event in chronological order using the hypothetical plan rules. A member is hired at age 35, becomes vested after five years, and is entitled to reinstatement of prior service after eligible re-employment. At age 42, after seven years of active service, the member terminates employment. The member is rehired at age 44, enters the plan’s disability status at age 46, recovers and returns to active service at age 47, commences a retirement pension at age 60, and dies at age 72. Classify the member after each event.
At age 35: The member is alive and active, so the initial state is \(A\).
At age 42: The member has seven years of service, which exceeds the five-year vesting requirement. The triggering event is vested termination, so \(A \to V\). The member’s pension is deferred rather than in payment.
At age 44: The plan permits reinstatement of prior service on this re-employment. The member returns to active status, so \(V \to A\). This classification uses the stated reinstatement rule; without it, the deferred right might need to remain recorded separately.
At age 46: The member enters the plan’s disability status before any retirement pension has commenced, so \(A \to D\).
At age 47: The member recovers and resumes active service. Under the stated plan assumption, \(D \to A\).
At age 60: The retirement pension commences, so \(A \to R\). The relevant event is commencement, not merely reaching age 60.
At age 72: The member dies. The transition is \(R \to X\). At a time after death the member is classified as deceased, not retired.
The full path is \(A \to V \to A \to D \to A \to R \to X\). The common mistake this method avoids is treating termination, age attainment, or disability as sufficient by itself to determine status without checking the plan’s vesting, commencement, and recovery rules.
Worked Example: Death and Retirement on the Same Calendar Date
Method: Use the effective order of events, then apply the state definitions at the time being classified. A hypothetical vested deferred member is scheduled to begin a pension on 1 July. In Case 1, the plan treats commencement as effective at 00:00 on 1 July if the member is alive then; the member dies at 15:00 that day. In Case 2, commencement is effective only at 23:59 and only if the member is alive then; the member dies at 10:00. Identify the state immediately after each death.
Case 1: At 00:00 the member is alive and the pension commences. The transition is \(V \to R\). At 15:00 the member dies, giving \(R \to X\). Immediately after death the state is \(X\), not \(R\).
Case 2: The member dies at 10:00, before the scheduled commencement time. The member is still in \(V\) immediately before death, so the transition is \(V \to X\). The condition for commencement at 23:59 is not met because the member is not alive then.
In both cases the state after death is \(X\), but the transition path differs: \(V \to R \to X\) in Case 1 and \(V \to X\) in Case 2. The common mistake is to classify a member as retired merely because the scheduled commencement date has arrived. The effective time and the plan’s condition must be checked.
Worked Example: Checking State Counts in a Cohort
Method: Apply the transition counts in stages and verify that the total number of members is conserved. A hypothetical cohort of 12 active members experiences these events: 3 have vested termination, 2 enter disability status, 1 dies, 4 commence retirement pensions, and 2 remain active. Later, among the 3 deferred members, 1 retires, 1 dies, and 1 is rehired into active status. Of the 2 disabled members, 1 recovers into active status and 1 dies. Of the 4 retired members, 1 dies. Find the final number in each state.
After the first set of events: The cohort has 3 in \(V\), 2 in \(D\), 1 in \(X\), 4 in \(R\), and 2 in \(A\). The check is \(3+2+1+4+2=12\).
Update the deferred members: The 3 in \(V\) become 1 in \(R\), 1 in \(X\), and 1 in \(A\). There are now 4 retired members in total before accounting for deaths in retirement: the original 4 plus the 1 newly retired, for 5.
Update the disabled members: The 2 in \(D\) become 1 in \(A\) and 1 in \(X\). Active membership is now \(2+1+1=4\): the 2 who remained active, the 1 rehired member, and the 1 who recovered.
Update the retired members: Of the 5 in \(R\), 1 dies and 4 remain retired. The deceased count is \(1+1+1+1=4\): the initial death, the deferred-member death, the disabled-member death, and the retired-member death.
Final check: There are 4 in \(A\), 0 in \(V\), 0 in \(D\), 4 in \(R\), and 4 in \(X\). The total is \(4+0+0+4+4=12\), matching the initial cohort. The common mistake avoided is to count a deceased retiree in both \(R\) and \(X\); after death, that member contributes only to \(X\).
Common Traps and Model Limitations
- Overlapping definitions: Define \(R\) as alive with a pension commenced, and make \(X\) take precedence after death. Do not leave it unclear whether a deceased pensioner remains classified as retired.
- Unlabeled arrows: An arrow from \(A\) to \(V\) should say which kind of termination leads to a vested deferred right. Eligibility conditions belong in the label or accompanying plan assumptions.
- Confusing employment with benefit status: A member who stops working is not necessarily \(V\); vesting and the right to a deferred pension must be established.
- Assuming every route is universal: Re-employment, disability recovery, and disability-to-retirement conversion depend on plan provisions. Do not add those arrows without stating the relevant rule.
- Forcing combined statuses into one node: If an active member can retain a separate deferred right, a single label may lose information. Add a combined state or an additional status variable when the model requires both facts.
Check Your Understanding
For each problem, state the resulting state or transition under the assumptions given.
- A member is active and alive. The member terminates after satisfying the plan’s vesting requirement, but does not commence a pension. What transition occurs, and what is the resulting state? (Answer: \(A \to V\); vested deferred.)
- A deferred member is alive on the date a pension is scheduled to commence, but the plan requires the member to be alive at the stated commencement time. The member dies before that time. What is the state immediately after death? (Answer: \(X\); the transition is \(V \to X\).)
- A disabled member recovers and resumes active service under a plan that explicitly restores active status. What transition applies? (Answer: \(D \to A\).)
- A retired member dies. Give the transition and explain why the member is not counted in two states after death. (Answer: \(R \to X\); \(X\) takes precedence, so the member is classified only as deceased after death.)
- A plan permits a rehired member to accrue new service while preserving a separate deferred pension right. Why might the five-state model be inadequate, and give one suitable adjustment. (Answer: One state cannot represent both statuses without losing information; use a combined active-with-deferred state or an additional deferred-right flag.)