Why Feasibility Must Be Tested
A management action may appear valuable in a stressed projection yet be unavailable when needed. An insurer might plan to sell assets, change underwriting terms, obtain reinsurance, or raise capital. A projection can estimate the effect of such an action, but that estimate does not establish that the insurer has authority to act, can complete the action in time, or can find willing counterparties. Feasibility testing asks whether the action can be carried out under the conditions of the scenario being assessed.
The test is specific to the stress and the proposed action. A transaction that is feasible in ordinary conditions may not be feasible during a market disruption, when counterparties are constrained or transaction prices have changed. Similarly, a plan that is legally permitted may fail if implementation systems, staff, or approvals are not available before the action is needed. A defensible ORSA assessment therefore distinguishes the action's modeled effect from the evidence that supports its execution.
A Structured Feasibility Test
A practical assessment starts by stating the action precisely: what will be done, by whom, in what amount or volume, by what date, and under which scenario conditions. Vague statements such as “raise capital if necessary” cannot be tested until the required approvals, investor process, amount, and deadline are identified. The assessment should then examine each constraint that could prevent, delay, or limit execution.
Identify the legal, contractual, governance, and regulatory permissions required. Check whether the decision-maker has authority in the relevant legal entity and whether any consent, notice, or restriction applies.
Map the time needed for approvals, preparation, implementation, and settlement against the date by which the action must be effective. Include required notice periods and dependencies, not just the final execution step.
Determine the volume the insurer can process and the amount counterparties or markets can absorb under the stressed conditions. Use the binding constraints, rather than assuming the planned amount can be completed in full.
Identify which policyholders, employees, investors, counterparties, or other stakeholders must respond or cooperate. Distinguish a documented commitment from an expectation of participation.
Identify actions that must precede or enable other actions, and document the evidence supporting each key assumption. Record conditions that would trigger reassessment or a fallback plan.
The constraints can be expressed as limits on the amount of an action that can be executed. Suppose a proposed action has a unit of execution, such as one policy change or one dollar of asset sales. Let \(c_1,c_2,\ldots,c_k\) be the maximum amounts permitted by each necessary constraint, all measured in the same unit. An executable amount \(n\) must satisfy every constraint:
It follows that \(n\) cannot exceed the tightest constraint, so the maximum supportable amount is \(\min(c_1,c_2,\ldots,c_k)\). This result follows directly: if \(c_j\) is the smallest limit, every feasible \(n\) must be at most \(c_j\); if the other stated constraints are the only limits, an amount up to \(c_j\) satisfies them all. In practice, the calculation is only as reliable as the constraints and evidence included. A missing approval or an unmodeled dependency can make the apparent maximum too high.
Worked Feasibility Assessments
Worked Example: Testing a Policy Migration Under a Short Deadline
A hypothetical insurer proposes migrating 12,000 eligible policies to a revised product after a severe claims stress. The migration is expected to provide $250 of benefit per policy successfully migrated; these figures are hypothetical planning assumptions, not market data. The scenario provides a three-month implementation window. Applicable rules require 60 days of notice before a policy can be migrated. The operations team can process at most 4,000 policies per month after the notice period, and the insurer assumes that 80% of policyholders offered the migration will accept. Assume the insurer has confirmed legal authority to make the offer, that the 60-day notice period equals two months for this planning exercise, and that there are no other capacity constraints. Test whether the modeled 12,000-policy action is feasible.
Method. First determine how much implementation time remains after the mandatory notice period. Three months are available, and two are consumed by notice, leaving \(3-2=1\) month for processing.
The operational capacity during that one month is \(4{,}000 \times 1=4{,}000\) offers. This is below the 12,000 policies in the plan. Applying the stated 80% acceptance assumption gives \(4{,}000 \times 0.80=3{,}200\) completed migrations.
The supportable benefit is \(3{,}200 \times \$250=\$800{,}000\). By comparison, the modeled benefit for all 12,000 migrations would be \(12{,}000 \times \$250=\$3{,}000{,}000\). The action is not feasible at the modeled scale within the window under the stated assumptions; the supported amount is 3,200 completed migrations and $800,000 of benefit.
Common mistake. Ignoring the notice requirement would use all three months of operational capacity: \(3 \times 4{,}000 \times 0.80=9{,}600\) migrations, implying $2,400,000 of benefit. That calculation is wrong because processing cannot begin until the two-month notice period has elapsed. The corrected calculation uses only the remaining one month. The check is that 3,200 is both below the 4,000 offers that can be processed and below the 12,000-policy target.
Worked Example: Testing an Asset Sale Against Market Depth
A hypothetical insurer plans to generate at least $25 million of net proceeds from selling securities within four business days of a stress trigger. It holds $50 million of the relevant securities, of which $10 million is pledged and unavailable for sale. Under the stressed market assumption, buyers can absorb at most $12 million of gross sales per business day. Settlement requires two business days after a sale, and each sale incurs a 2% reduction from gross sale value to net proceeds. Assume the insurer has the required internal and legal authority, can execute sales on each business day, and needs the proceeds settled by the end of day four. Can it meet the $25 million target?
Method. Separate the inventory, daily market capacity, settlement timing, and transaction reduction. The unpledged inventory is \(\$50\text{ million}-\$10\text{ million}=\$40\text{ million}\). Inventory is not the binding limit here, because the market can absorb at most $12 million gross per sale day.
To have settlement by the end of day four, the latest sale can occur by the end of day two: it then settles two business days later. Thus, at most two days of sales can settle by the deadline. The maximum gross sales are \(2 \times \$12\text{ million}=\$24\text{ million}\), which is below the $40 million of unpledged inventory.
Net proceeds are 98% of gross sales, so the maximum timely net proceeds are \(0.98 \times \$24\text{ million}=\$23.52\text{ million}\). The shortfall from the target is \(\$25.00\text{ million}-\$23.52\text{ million}=\$1.48\text{ million}\). The insurer cannot support the target under these assumptions.
Common mistake. Treating four days as four sale days would produce \(4 \times \$12\text{ million}=\$48\text{ million}\) of gross capacity and $47.04 million net. That is not timely capacity: sales made on days three and four settle after the deadline. The corrected test includes only two sale days. It also applies the 2% transaction reduction to gross sales, not as an extra amount added to the target.
Worked Example: Testing a Dependent Capital-Raising Package
A hypothetical insurer proposes raising $18 million through a capital placement, with settled funds required by the end of day 39 after a stress trigger. The board decision takes eight days. A data pack needed for the regulatory submission takes ten days to prepare and can be developed in parallel with board consideration. The regulatory review takes 15 days after both the board decision and data pack are complete. Investor documentation and execution then take ten days after regulatory approval, followed by five days for settlement. Assume these durations are fixed for this exercise and that no step can be shortened. Investors have provided binding subscriptions totaling $14 million, and no additional subscriptions are available in the scenario. Test the package.
Method. Map the dependencies before adding durations. Board consideration and data-pack preparation run in parallel, so both are complete after the longer of eight and ten days: \(\max(8,10)=10\) days. The regulatory review then ends on day \(10+15=25\). Investor execution ends on day \(25+10=35\), and settlement ends on day \(35+5=40\).
The settled funds arrive on day 40, one day after the day-39 deadline. The package therefore fails the timing test. It also fails the amount test: binding subscriptions support $14 million, which is \(\$18\text{ million}-\$14\text{ million}=\$4\text{ million}\) below the target. Both constraints must be satisfied, so the proposed $18 million action is infeasible as specified.
Common mistake. Adding the board and data-pack durations as if they were sequential gives \(8+10+15+10+5=48\) days. That incorrectly treats parallel work as serial. The correct critical-path calculation uses \(\max(8,10)\), giving 40 days. Although 40 days is shorter than 48, it still misses the 39-day deadline. A second error would be to treat the $18 million target as available merely because it is the planned amount; the binding subscription evidence supports only $14 million.
Evidence, Dependencies, and Limits of the Conclusion
A feasibility conclusion should be traceable to evidence appropriate to the constraint. Authority may be supported by legal analysis, governance documents, or documented regulatory requirements. Operational capacity may require staffing plans, system testing, and implementation schedules. Market access may require evidence of counterparties, transaction capacity, or binding commitments under conditions consistent with the stress. Stakeholder response assumptions should distinguish actual commitments from indications of interest or historical experience that may not persist under stress.
Dependencies need particular care. If actions are serial, their required times accumulate. If work can proceed in parallel, the relevant elapsed time is governed by the longest path to the next dependent step. An action may also depend on another action in a substantive, rather than purely chronological, way: for example, a planned transaction may require information produced by a separate operational change. The assessment should identify that relationship rather than counting each action's timing independently.
Feasibility is not a binary label for an entire plan when only part of an action can be completed. Where a constraint caps execution, the conclusion should state the supported amount or volume, the binding constraint, and the assumptions on which that limit depends. If the action is contingent on approvals or market conditions that are not secured, describe it as conditional and identify the evidence needed before relying on it. Do not treat a possible action as certain merely because it is included in a projection.
Common Feasibility Traps
- Confusing authority with execution. Legal permission does not establish that the action can be completed operationally or by the required date.
- Using the full planning horizon as implementation time. Notice periods, approvals, settlement, and preparation consume time before an action takes effect.
- Assuming normal-condition capacity in a stress. Market depth, staff availability, and stakeholder willingness may all be lower under the scenario being tested.
- Counting nonbinding interest as committed capacity. Expressions of interest do not necessarily establish a reliable amount available when the action is needed.
- Ignoring dependencies or double-counting parallel work. Map which steps require completion of earlier steps, and which genuinely can proceed at the same time.
- Reporting only “feasible” or “not feasible.” State the amount supportable, the constraint that limits it, the deadline, and any unresolved conditions.
Check Your Understanding
For each hypothetical case, identify the binding feasibility constraint and calculate the supported amount where the data allow it.
- A product change requires 30 days of customer notice. The insurer has a 50-day window and can process 600 accounts per day after notice. If 75% of offered changes are accepted and each completed change is expected to provide $40, state the maximum completed changes and benefit. (Answer: 12,000 offers, 9,000 completed changes, and $360,000 of benefit.)
- An insurer must receive $15 million net from an asset sale by day five. It can sell $5 million gross per day, but settlement takes three business days and transaction costs reduce gross proceeds by 4%. Assume sales can be made each business day. State the maximum net amount settled by the deadline and whether the target is met. (Answer: Two sale days are available; maximum net proceeds are \(2 \times \$5\text{ million} \times 0.96=\$9.60\text{ million}\), so the target is not met.)
- A capital transaction requires five days for board approval and seven days for data preparation, which can proceed in parallel. Regulatory review then takes 12 days, execution takes eight days, and settlement takes four days. How many days are required from the trigger to settled funds? (Answer: \(\max(5,7)+12+8+4=31\) days.)
- A hypothetical reinsurance action is planned to transfer $30 million of exposure. Legal authority permits the transaction, operational staff can complete documentation for $24 million, and counterparties under the stress scenario have offered binding capacity for $20 million. What amount is supportable before considering any other constraints? (Answer: $20 million, limited by binding counterparty capacity.)
- A proposed $10 million placement must settle within 25 days. Approvals and documentation take 10 days in sequence; investor execution then takes eight days and settlement takes five days. Binding investor subscriptions are $8 million. State whether the full proposed action is feasible and identify each failed constraint. (Answer: No. The timeline is \(10+8+5=23\) days and meets the deadline, but binding subscriptions support only $8 million, $2 million below the target.)