Why a Capital Model Is Not the Whole ORSA
A capital model can estimate the capital needed for specified risks under defined assumptions. That estimate is useful, but it cannot by itself answer every question an insurer faces about its solvency. The insurer must also consider whether the model captures its actual risk profile, how risks could change under its plans, and whether its resources and management actions would remain effective in adverse conditions.
An Own Risk and Solvency Assessment (ORSA) is the insurer’s own assessment of its risks and solvency needs in relation to its risk profile and business strategy. A standalone capital model is a tool that may inform that assessment. The distinction matters: a calculated amount is an output conditional on a model’s scope and assumptions, not a complete conclusion about the insurer’s ability to meet obligations or pursue its plans.
“Standalone” describes the model’s role here: it produces an estimate within its defined boundaries. It does not mean the model is necessarily unsophisticated or detached from the insurer’s business. Even a detailed model cannot settle issues it does not represent, such as whether a critical control will work, whether a planned response is feasible, or whether a strategic change makes historical assumptions less relevant.
What a Model Output Does—and Does Not—Say
A capital result should be read with its scope attached. Relevant questions include which risks are included, what horizon and measure are used, which assumptions drive the result, and what data limitations remain. A result may be suitable for comparing modeled exposures under a consistent basis while still being incomplete as an assessment of the insurer’s broader position.
For example, a model may quantify losses from specified catastrophe, market, or underwriting risks but not fully reflect a new reliance on one claims administrator. The model result does not prove that the service-provider risk is immaterial. It says only that the result was generated from the modeled inputs and methods. ORSA analysis must consider whether an unmodeled or weakly modeled risk could affect operations, capital, liquidity, or the execution of the business plan.
The same caution applies to management actions. A model may show that a planned action reduces a loss, but management must consider whether the action can be taken in time, whether it is legally and operationally available, and whether it could have side effects. A numerical benefit is not automatically a credible plan.
Why Model Scope Limits What the Output Can Establish
The basic logic is straightforward. Let \(M\) denote a model’s output, \(I\) its included inputs, and \(A\) its assumptions and methods. Then the output can be represented as \(M=g(I,A)\). If a material risk is outside the model’s scope, changing that risk while leaving \(I\) and \(A\) unchanged will not change \(M\). Therefore, the model output alone cannot establish that the omitted risk has no effect on the insurer’s overall solvency assessment.
This is not a claim that every omitted risk requires a capital add-on. It is a reason to investigate material omissions and explain how they affect the assessment. Depending on the evidence, the appropriate response could be a scenario analysis, a change in controls, a revised plan, a targeted model enhancement, or a documented judgment that the risk does not alter the conclusion materially. The response should follow from the risk analysis rather than from a desire to force every concern into one number.
A useful discipline is to distinguish three things: the model’s reported output, additional evidence or stresses considered in ORSA, and the management conclusion drawn from them. Keeping these separate prevents an informal stress from being presented as though it were a calibrated capital requirement.
Worked Applications
Worked Example: A Model Omits a Critical Service-Provider Exposure
A hypothetical insurer’s standalone model reports required capital of $120 million. The modeled risk amounts before diversification are $72 million for claims risk, $38 million for asset risk, and $30 million for catastrophe risk; the model applies a $20 million diversification credit. Available capital is $150 million. A qualitative review identifies dependence on one claims administrator, and a hypothetical operational stress estimates a $25 million loss if a prolonged interruption occurs. Determine what the model says, what the stress adds, and what should not be inferred.
Method. First reproduce the model result, then compare available capital with that result. Treat the operational stress as a separate sensitivity, not as an automatically calibrated addition to required capital.
The modeled amount is:
Modeled capital headroom is:
The separate operational stress is $25 million. If, purely as a simple sensitivity, that loss reduces the same available-capital measure dollar for dollar and is not already reflected elsewhere, stressed available capital would be \(150-25=125\) million dollars. Relative to the modeled $120 million requirement, that leaves \(125-120=5\) million dollars of headroom.
The calculation is a warning signal for further assessment, not a proof that the insurer’s ORSA capital need is exactly $145 million, or that the operational event has a particular probability. The stress may have a different basis from the model, may overlap with modeled losses, and may affect operations or cash flows in ways not represented by the simple subtraction. Management should examine the provider’s substitutability, recovery arrangements, and the plausibility of the stress. The common mistake is to add $25 million mechanically to the model’s $120 million and label $145 million “the ORSA requirement.” That treats an uncalibrated sensitivity as a like-for-like capital measure and may double count risk.
Check. The model result is $120 million, the modeled headroom is $30 million, and the illustrative sensitivity leaves $5 million of headroom. These figures are arithmetically consistent; the last figure is not a new calibrated capital requirement.
Worked Example: Growth Changes the Forward-Looking Assessment
A hypothetical insurer has current available capital of $240 million and a standalone model requirement of $200 million. Its plan is to expand a specialty product line. Under a stated planning scenario, the expansion increases modeled capital need by $32 million and is expected to reduce available capital by $12 million because of launch costs and an initial operating loss. Calculate current and projected headroom and the projected capital-to-requirement ratio. What does the result contribute to ORSA?
Method. Calculate the current position first, then update both the requirement and available capital using the scenario assumptions. Do not adjust only one side of the comparison.
Current headroom is:
Projected required capital is \(200+32=232\) million dollars. Projected available capital is \(240-12=228\) million dollars. Thus projected headroom is:
The projected capital-to-requirement ratio is:
Rounded to three decimal places, the scenario indicates a $4 million shortfall against the projected model requirement and a ratio of 0.983. The current $40 million headroom does not resolve the forward-looking issue: the plan changes both the risk profile and the available resources.
The result does not, by itself, dictate whether the insurer should abandon the expansion. ORSA analysis should consider the scenario’s assumptions, timing, uncertainty, and whether credible actions could address the shortfall. Possible decisions might include phasing the expansion, securing additional resources, or changing the plan. The common mistake is to report only the current $40 million headroom as evidence that the growth plan is affordable. It ignores the stated future changes.
Check. Projected resources are $228 million against a $232 million requirement, so the difference is a $4 million shortfall; \(228/232\) rounds to 0.983.
Worked Example: Capital Adequacy Does Not Settle Liquidity
A hypothetical insurer reports available capital of $145 million and a standalone model requirement of $100 million. Separately, it expects $90 million of assets to be available for near-term cash needs and $65 million of cash obligations within 30 days. A catastrophe stress creates an additional $30 million of cash outflow by day 30, while $18 million of recoveries are expected only by day 90. Under these simplified assumptions, calculate the day-30 cash position and the day-90 position after the recoveries arrive. Explain why the capital model result does not settle the issue.
Method. Compare cash available with cash outflows by date. Do not treat a later recovery as though it were already available on day 30.
The standalone model reports capital headroom of \(145-100=45\) million dollars. This is a capital comparison. For the stated cash-flow stress, the day-30 position is:
The hypothetical day-30 cash position is a $5 million shortfall. After the $18 million recovery arrives by day 90, and assuming no other cash flows, the position becomes:
The insurer can therefore show positive capital headroom in the model and still face a timing gap in this simplified liquidity stress. The figures are not interchangeable: capital adequacy and cash availability answer different questions. ORSA should consider the timing and reliability of the recoveries, potential sources of cash, and whether the stress assumptions are reasonable. The common mistake is to subtract the $30 million outflow from the $45 million capital headroom and conclude that liquidity is covered. That calculation mixes measures and ignores the cash-flow dates.
Check. Day 30 is short by $5 million; after the stated $18 million recovery, the day-90 position is positive $13 million. The capital model’s $45 million headroom does not remove the day-30 cash gap.
Using Judgment Without Replacing Analysis with Opinion
Management judgment is necessary because not every material issue can be reduced to a reliable model estimate. Judgment should nevertheless be disciplined and traceable. It should identify the evidence considered, the uncertainty or limitation involved, the decision affected, and why the chosen response is proportionate. Where a numerical stress is used, its assumptions and relationship to the model should be clear.
A useful sequence is to keep the model output visible, assess the risk or limitation that may not be captured, and then explain the implications for the insurer’s plans and solvency conclusion. The purpose is not to create a second, competing “true” capital number. It is to make sure the conclusion reflects the insurer’s own circumstances rather than only the boundaries of a calculation.
State the modeled risks, assumptions, horizon, and limitations relevant to the decision.
Consider qualitative exposures, changes in strategy, operational dependencies, and whether proposed responses are feasible.
Use scenarios, sensitivities, and other evidence where appropriate, while avoiding unsupported or double-counted numerical adjustments.
State what the evidence supports, what uncertainty remains, and how that uncertainty affects the insurer’s assessment or plans.
A useful exam distinction is that the model answers a bounded quantitative question; ORSA considers whether that answer, together with other relevant evidence and judgment, is adequate for the insurer’s own assessment. This is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus.
Common Traps
- Treating model output as a complete risk inventory. A risk absent from the model can still be material to the insurer’s assessment.
- Turning every stress into a capital add-on. A stress may be useful for testing resilience without being calibrated to the same basis as the model requirement.
- Ignoring the business plan. A current capital comparison does not establish that future growth, acquisitions, or other planned changes are supportable.
- Assuming a modeled management action will occur. An action should be considered in light of its timing, authority, feasibility, and potential side effects.
- Confusing cash and capital. A favorable capital comparison does not establish that funds will be available at the time obligations fall due.
Check Your Understanding
For each problem, distinguish the model calculation from the broader ORSA consideration.
- A hypothetical insurer’s model reports a $74 million requirement. Its model includes underwriting and market risk but excludes a newly identified dependence on one data-hosting provider. A reasonable operational stress is $11 million, but no event probability is estimated. What is the model result, and why is it not automatically correct to report an $85 million ORSA requirement? (Answer: The model result is $74 million; adding $11 million mechanically is unjustified because the stress is not established as a like-for-like calibrated requirement and could overlap with modeled effects.)
- A company has projected available capital of $156 million and a projected model requirement of $164 million after a planned product launch. Calculate projected headroom. Name one type of management response ORSA could assess. (Answer: \(156-164=-8\), a shortfall of $8 million; ORSA could assess phasing the launch or obtaining additional resources.)
- A model shows $28 million of capital headroom. In a separate hypothetical stress, cash available in 30 days is $42 million, obligations are $35 million, and a stressed outflow is $12 million. Calculate stressed day-30 cash headroom and explain whether the model’s capital headroom resolves it. (Answer: \(42-35-12=-5\), a $5 million cash shortfall; the $28 million model headroom does not by itself resolve a cash-timing gap.)
- A hypothetical model requirement is $90 million. Available capital is $110 million today. Under a strategic scenario, the requirement rises by $15 million and available capital falls by $8 million. Calculate the projected headroom and state the key interpretive point. (Answer: Projected requirement is $105 million and projected available capital is $102 million, giving \(102-105=-3\), a $3 million shortfall; the current position alone does not establish that the future plan is supportable.)
- A model indicates sufficient capital after assuming a recovery action that would generate $20 million, but the action depends on a contract renewal that has not been secured. What should the ORSA discussion address? (Answer: It should assess the action’s availability and timing, the uncertainty created by the unsecured renewal, and the effect on the conclusion if the action is not available; the modeled benefit should not be treated as certain without support.)