Actuarial Analysis Informs ORSA; It Does Not Own the Risk
An ORSA needs technically credible estimates of how risks may affect an insurer’s solvency, but technical analysis is only one input to the assessment. Actuaries can estimate liabilities, test assumptions, project financial outcomes, and explain uncertainty. Those contributions help management and the board understand the insurer’s risk profile and options. They do not, by themselves, make the actuary accountable for accepting the risk, setting risk appetite, or choosing a business response.
The distinction is practical, not merely organizational. A reserve analysis may indicate that adverse claims development could reduce projected capital. The actuarial contribution is to describe the estimate, its assumptions, its sensitivity, and its limitations. Management must decide what action to take, such as changing underwriting, purchasing reinsurance, or accepting the exposure within the insurer’s strategy and risk appetite. The board retains its oversight role.
This distinction is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus. It does not imply that actuaries have no responsibility for the quality of their own work. An actuary remains responsible for communicating methods, assumptions, material limitations, and conclusions clearly and professionally. That professional responsibility is different from owning the business risk being analyzed.
What Actuaries Can Contribute
The specific contribution depends on the question, the actuary’s role, and the information available. In an ORSA, useful actuarial work may include:
- Estimating insurance liabilities and explaining how experience, assumptions, and methods affect those estimates.
- Projecting premiums, claims, expenses, and other insurance-related cash flows under the business plan or specified scenarios.
- Testing sensitivities and stresses, then explaining which assumptions or exposures drive the results.
- Assessing how reinsurance terms or other risk-transfer arrangements affect retained claims and projected outcomes.
- Identifying uncertainty, data limitations, model limitations, and circumstances in which a result may not be reliable.
- Interpreting what a result means for the insurer’s financial position, while making clear which conclusions are technical and which require management judgment.
An actuarial estimate is not automatically a complete ORSA conclusion. A result has meaning only in context: what was measured, over what period, under which assumptions, and for which entity or group? The actuary can help answer those questions and explain implications. Management remains responsible for considering the full business context, including strategic objectives, operational constraints, and risk appetite.
Why Assumptions and Interpretation Matter
A useful actuarial analysis does more than report a number. It shows how the number was constructed and what could cause it to change. For example, a projected claims amount may depend on exposure, claim frequency, and average severity. Under a simple model with homogeneous exposure and compatible definitions of frequency and severity, expected claims are:
The relationship follows by first estimating the expected number of claims as exposure multiplied by expected frequency per unit, and then multiplying by expected cost per claim. This decomposition helps identify whether a change in projected claims arises from more exposure, a higher expected frequency, or greater severity. It is a modeling relationship, not a guarantee that the components are independent or stable. If the portfolio is heterogeneous, definitions differ, or frequency and severity interact, a more suitable model or segmentation may be needed.
Interpretation is also important when a scenario changes more than one assumption. If severity rises while frequency falls, the net effect is not found by adding the percentage changes: the factors multiply. An actuary should state the calculation and explain whether it is a sensitivity, a forecast, or a stress. Those labels matter because a stress scenario need not be the most likely outcome, and a sensitivity does not establish what management should do.
Worked Example: Interpreting a Reserve Sensitivity
A hypothetical insurer reports an estimated unpaid claims liability of $120.00 million. For a one-year sensitivity, the actuary holds exposure and all other assumptions constant, reduces expected claim frequency by 2%, and increases expected claim severity by 5%. Determine the revised estimate and explain what the result does and does not establish.
Method and assumptions. Treat the changes as multiplicative factors applied to the baseline estimate, consistent with a simplified frequency-severity model. This is a deterministic sensitivity, not a probability distribution or a forecast. Interest and mortality assumptions do not apply to this non-life claims sensitivity.
The frequency factor is \(1-0.02=0.98\), and the severity factor is \(1+0.05=1.05\). The combined factor is:
The revised liability is:
The increase is \(\$123.48-\$120.00=\$3.48\) million. The arithmetic checks: the combined assumptions increase the baseline by 2.9%, and \(0.029 \times \$120.00=\$3.48\) million.
Interpretation. The actuary can report that this specified combination of assumptions increases the estimated liability by $3.48 million, subject to the model’s simplifying assumptions. The result does not show that this combination is likely, nor does it determine whether the insurer should change underwriting, buy reinsurance, or accept the exposure. Those are management decisions informed by the broader ORSA.
Common mistake avoided. Adding the assumption changes gives \(+5\%-2\%=+3\%\), or an estimated increase of $3.60 million. That is wrong because frequency and severity factors multiply. Their product is 1.029, not 1.03.
From Technical Result to Management Decision
Actuarial work can help decision-makers compare outcomes, but a calculation alone does not select a course of action. The following sequence keeps the technical contribution and the ownership boundary visible:
Specify the risk, entity, time horizon, baseline, and scenario being analyzed.
Describe the data, assumptions, methods, and material limitations; test whether the result is internally consistent.
Explain its effect on relevant financial outcomes and identify uncertainty or important drivers.
Explain how plausible actions could change the analysis, without presenting a technical estimate as a decision already made.
Those responsible for governance determine whether to accept, mitigate, transfer, or otherwise manage the risk, consistent with their authority and responsibilities.
This sequence does not prevent an actuary from expressing a professional view or raising a concern. It requires clarity about the status of that view. “Under these assumptions, the projected liability increases by this amount” is a technical conclusion. “The insurer should accept this exposure” is a management decision, even if actuarial analysis informs it.
Worked Example: Assessing a Catastrophe Scenario and Reinsurance
A hypothetical insurer estimates a gross catastrophe loss of $42 million in a specified stress scenario. Its treaty has a $10 million retention and a $25 million limit on recoveries. Assume the treaty responds as described, and apply an 8% haircut to the calculated recovery to reflect a modeled collection uncertainty. Starting available capital is $90 million. For this simplified illustration, treat the net loss as reducing available capital dollar for dollar and hold the capital requirement constant at $62 million. Determine the modeled net loss, post-stress capital, and remaining capital above the requirement.
Method and assumptions. First calculate the contractual recovery as the lesser of the amount above the retention and the treaty limit. Then apply the stated haircut. This is a deterministic stress calculation; no probability distribution, interest rate, or mortality assumption is used. The capital bridge is deliberately simplified and excludes other balance-sheet effects.
The gross loss above the retention is \(\$42-\$10=\$32\) million. The contractual recovery is capped at $25 million:
The 8% collection haircut is \(0.08 \times \$25=\$2\) million. The modeled collectible recovery is therefore \(\$25-\$2=\$23\) million. The net loss is \(\$42-\$23=\$19\) million.
Under the stated dollar-for-dollar assumption, post-stress available capital is \(\$90-\$19=\$71\) million. Capital above the unchanged requirement is \(\$71-\$62=\$9\) million. The check is consistent: the initial amount above the requirement was \(\$90-\$62=\$28\) million, and a $19 million loss leaves $9 million.
Interpretation. The actuarial contribution is the calculated effect of this scenario under the specified treaty and simplified capital assumptions. It can support discussion of reinsurance structure, counterparty exposure, and solvency resilience. It does not decide whether to purchase additional coverage or accept the remaining exposure; that decision belongs to management within the insurer’s governance arrangements.
Common mistake avoided. Subtracting the full $25 million treaty limit from the gross loss gives a $17 million net loss. That ignores the stated collection haircut. The modeled recovery is $23 million, so the consistent net loss is $19 million.
Communicating Limits, Options, and Uncertainty
A technically correct result can still mislead if its limitations are hidden. Actuarial communication should make clear whether figures are estimates or actual outcomes, which assumptions are fixed, and how sensitive the conclusion is to those assumptions. If data are incomplete or the model omits a material feature, the analysis should say so. This does not make the work less useful; it helps decision-makers understand how much weight to place on it.
The distinction between analysis and ownership can be especially important when results are uncomfortable. An actuary may identify that a proposed plan relies on assumptions that appear optimistic, or that a stress creates a narrow capital margin. The actuary should explain the evidence and uncertainty, and raise material concerns through appropriate channels. Management must then consider the finding and decide on action. The actuary’s responsibility to provide sound analysis is not a substitute for management’s responsibility to manage the risk.
Worked Example: Interpreting a Growth Projection
A hypothetical insurer projects one-year results for a line of business. The baseline has $100 million of premium, a 65% claims ratio, $25 million of expenses, and investment income equal to 3% of premium. A proposed growth plan increases premium volume by 12%; the claims ratio is assumed to rise to 68%, expenses are projected at $28 million, and investment income remains 3% of premium. Calculate the baseline and growth-plan result, then state an appropriate actuarial interpretation.
Method and assumptions. Use the simplified result formula: premium less claims and expenses, plus investment income. Claims are premium multiplied by the stated claims ratio. The calculation is deterministic over one year; no probability distribution, interest rate beyond the stated investment-income assumption, or mortality assumption is used.
For the baseline, claims are \(0.65 \times \$100=\$65\) million, and investment income is \(0.03 \times \$100=\$3\) million. The result is:
Under the growth plan, premium is \(\$100 \times 1.12=\$112\) million. Claims are \(0.68 \times \$112=\$76.16\) million, and investment income is \(0.03 \times \$112=\$3.36\) million. The projected result is:
The projected result decreases by \(\$13.00-\$11.20=\$1.80\) million, even though premium grows by 12%. The result as a percentage of premium falls from \(\$13/\$100=13.00\%\) to \(\$11.20/\$112=10.00\%\). The checks show that both ratios follow directly from the calculated results.
Interpretation. The actuary can identify that, under these assumptions, the larger volume does not produce a larger projected dollar result and that the projected margin is lower. The analysis should also identify the claims-ratio and expense assumptions as important drivers. It does not establish that the growth plan should be rejected: management must weigh the projection against strategy, risk appetite, uncertainty, and other relevant considerations.
Common mistake avoided. Assuming that 12% more premium means 12% more profit would project \(\$13 \times 1.12=\$14.56\) million. That scaling assumes unchanged economics. Here, the claims ratio and expenses change, so the result must be recomputed from the stated components.
Common Governance Traps
- Treating the technical specialist as the risk owner. Expertise in measurement does not make the actuary accountable for the business activity that generates the risk.
- Confusing a scenario with a prediction. A specified stress shows an outcome under stated conditions; it does not, by itself, give the likelihood of that outcome.
- Reporting a number without its basis. A liability or capital impact should be accompanied by the scope, assumptions, and material limitations needed to interpret it.
- Turning analysis into an implied decision. A result can inform a recommendation, but the responsible decision-maker must own the choice and its rationale.
- Omitting professional challenge. Preserving management ownership does not mean withholding relevant analysis or concerns. The actuarial contribution includes communicating material limitations and disagreements appropriately.
A sound ORSA uses actuarial work as evidence for informed judgment, not as a replacement for it. The actuary makes the technical basis visible and interpretable. Management and the board retain responsibility for risk ownership, governance, and decisions within their respective roles.
Check Your Understanding
For each situation, distinguish the actuarial contribution from the management decision. State the key calculation where requested.
- A hypothetical insurer has a $50 million baseline liability. A deterministic sensitivity increases expected frequency by 4% and decreases expected severity by 3%, with all else held constant. What is the revised estimate, and does it determine whether management should reduce the business? (Answer: \(1.04 \times 0.97=1.0088\), so the revised estimate is $50.44 million; no, the calculation informs but does not make the management decision.)
- A modeled event produces a $30 million gross loss. Reinsurance has a $5 million retention and a $15 million limit, and the recovery is assumed fully collectible. What is the net loss? Who decides whether to buy more coverage? (Answer: Recovery is \(\min(30-5,15)=\$15\) million, so net loss is $15 million; management decides whether to buy more coverage.)
- A projection relies on a claims ratio that is based on limited experience. Name two matters the actuary should communicate and identify who decides whether to proceed with the plan. (Answer: Communicate the data limitation and the sensitivity or uncertainty associated with the claims-ratio assumption; management owns the decision whether to proceed.)
- A growth scenario has $80 million of premium, a 70% claims ratio, $18 million of expenses, and investment income of 2% of premium. Calculate the simplified result. Does a positive result alone establish that the risk is within appetite? (Answer: Claims are $56 million and investment income is $1.60 million, so the result is \(80-56-18+1.60=\$7.60\) million; no, management must assess the result in the broader ORSA and against risk appetite.)
- An actuary calculates that a specified stress leaves $6 million of capital above a stated requirement, but the model excludes a material operational disruption. What should the actuarial communication make clear, and who determines the response? (Answer: Clearly disclose the model’s exclusion and explain that the result may not capture the disruption’s effect; management and the board determine the appropriate response within their responsibilities.)