Selecting Stress Tests for an ORSA

Why Test Selection Starts With a Decision

An ORSA stress-test program is not a catalogue of every adverse event that could be imagined. Its purpose is to examine whether the insurer can withstand material risks and to provide information relevant to decisions such as reinsurance purchasing, risk limits, asset allocation, and business-plan approval. A test is useful when its risk, time horizon, and outcomes connect clearly to one of those needs.

Selection begins with the risk inventory and the insurer’s solvency assessment: identify material exposures, consider where the business plan could change them, and ask which uncertainties could alter a decision or the conclusion about capital adequacy. A test that produces a large numerical movement but addresses no unresolved decision may be less useful than a smaller, well-targeted test that clarifies a significant exposure.

Definition: A standalone sensitivity changes one specified risk driver, or one risk in isolation, to examine the insurer’s exposure to that risk. A broader scenario describes a coherent set of adverse conditions that may affect several risks or business assumptions together. Both are stress tests; they answer different questions.

A standalone sensitivity is useful for isolating exposure. For instance, a catastrophe sensitivity can help assess whether a proposed underwriting limit or reinsurance structure leaves an unacceptable concentration. A broader scenario can test a question that isolation cannot answer, such as whether a severe event occurring alongside impaired reinsurance recoveries would create a solvency problem. The selection decision is not “sensitivity or scenario” in general; it is which form, or combination of forms, best addresses the stated need.

Compare Solvency Outcomes Consistently

For the examples below, define the solvency coverage ratio as available capital divided by internal capital need. A stress can reduce available capital, increase capital need, or do both. Comparing only the change in available capital can therefore misstate its effect on solvency.

R=\frac{A}{C}

Here, $A$ is available capital, $C$ is internal capital need, and $R$ is the coverage ratio. If a stress changes these amounts by $\Delta A$ and $\Delta C$, then the stressed ratio is $(A+\Delta A)/(C+\Delta C)$. The changes are signed: a reduction in available capital is a negative $\Delta A$, while an increase in capital need is a positive $\Delta C$.

This ratio is an organizing measure for comparing outcomes, not a complete description of risk. A selection decision also needs the business context, the assumptions behind the test, and the decision the result is intended to inform. A single ratio does not describe liquidity, timing, or the feasibility of responding to a stress.

Why Both Sides of the Ratio Matter

The effect of a stress on coverage follows directly by comparing stressed and baseline ratios. Provided $C$ and $C+\Delta C$ are positive:

\frac{A+\Delta A}{C+\Delta C}-\frac{A}{C} = \frac{C\Delta A-A\Delta C}{C(C+\Delta C)}

To verify the result, put the two ratios over the common denominator $C(C+\Delta C)$. The numerator becomes $C(A+\Delta A)-A(C+\Delta C)$, which simplifies to $C\Delta A-A\Delta C$. Thus, a loss of available capital ($\Delta A<0$) reduces the ratio, and an increase in capital need ($\Delta C>0$) also reduces it. This is why selecting and comparing tests requires both effects, not just the loss amount.

Worked Example: Choosing Targeted Sensitivities for Separate Decisions

A hypothetical insurer has current available capital of $240 million and internal capital need of $200 million. Its risk committee must review catastrophe reinsurance, interest-rate hedging, and operational resilience. A preliminary assessment supplies the following hypothetical stress impacts: catastrophe, $\Delta A=-$36$ million and $\Delta C=+$8$ million; interest rates, $\Delta A=-$18$ million and $\Delta C=-$6$ million; and operational disruption, $\Delta A=-$12$ million and $\Delta C=+$2$ million. Calculate the baseline and stressed ratios, then explain which tests should be included for the decisions described.

Method. Calculate the baseline ratio and, for each standalone sensitivity, calculate the changed available capital and changed capital need before dividing. The impacts are supplied stress outputs; no probability is assigned to them.

Baseline coverage is $240/200=1.2000$, or 120.00%.

Under the catastrophe sensitivity, available capital is $240-36=$204$ million and capital need is $200+8=$208$ million. Coverage is $204/208=0.980769$, or 98.08%.

Under the interest-rate sensitivity, available capital is $240-18=$222$ million and capital need is $200-6=$194$ million. Coverage is $222/194=1.144330$, or 114.43%.

Under the operational sensitivity, available capital is $240-12=$228$ million and capital need is $200+2=$202$ million. Coverage is $228/202=1.128713$, or 112.87%.

Selection rationale. The catastrophe test is directly relevant to the reinsurance review, and the operational test is relevant to the resilience review. The interest-rate test is relevant to the hedging decision even though its ratio remains above 100%; a test need not produce the worst outcome to inform a decision. The committee should therefore retain all three targeted tests if those are the decisions under review. The ratios show different effects, but do not on their own establish that any action is required.

Check. The catastrophe stress reduces available capital by $36 million and raises capital need by $8 million, so the calculation uses $204 million over $208 million. This produces a ratio below 100%, consistent with the stressed capital need exceeding available capital.

Common mistake avoided. Using $204/200=1.0200$ for the catastrophe stress applies the available-capital loss but ignores the increase in capital need. The correct denominator is $208 million, so the stressed ratio is 0.980769, not 1.0200.

When a Broader Scenario Is Needed

A standalone test is designed to isolate a risk. It is not a substitute for a broader scenario when the decision depends on risks occurring together, shared assumptions changing at once, or interactions across the business. Conversely, a combined scenario may reveal a vulnerability without showing which individual risk is driving it. An effective test set can use both forms: targeted tests to isolate exposures and broader scenarios to examine combined resilience.

A scenario should be selected because the combined conditions address a material question, not merely because several adverse assumptions can be placed together. The scenario’s rationale should identify what makes the conditions relevant to the insurer’s risk profile and which decisions its results could inform. The design and calibration of a particular scenario require separate consideration.

Worked Example: Selecting a Combined Test for Reinsurance Dependence

A hypothetical insurer projects $310 million of available capital and $260 million of internal capital need at a planning date. The reinsurance committee is reviewing its dependence on recoverables. The assessment provides two standalone outputs: a catastrophe stress of $\Delta A=-$42$ million and $\Delta C=+$10$ million, and a reinsurer failure stress of $\Delta A=-$18$ million and $\Delta C=+$6$ million. A separately modeled combined scenario produces $\Delta A=-$68$ million and $\Delta C=+$22$ million. Calculate the ratios and explain why the combined test adds information.

Method. First calculate the baseline, then apply each set of impacts to both components of the ratio. Treat the combined scenario’s supplied impacts as its own model output; do not assume they equal the sum of the standalone impacts.

Baseline coverage is $310/260=1.192308$, or 119.23%.

For the catastrophe sensitivity, available capital is $310-42=$268$ million and capital need is $260+10=$270$ million. Coverage is $268/270=0.992593$, or 99.26%.

For the reinsurer failure sensitivity, available capital is $310-18=$292$ million and capital need is $260+6=$266$ million. Coverage is $292/266=1.097744$, or 109.77%.

For the combined scenario, available capital is $310-68=$242$ million and capital need is $260+22=$282$ million. Coverage is $242/282=0.858156$, or 85.82%.

Selection rationale. The catastrophe sensitivity isolates underwriting exposure, and the reinsurer failure sensitivity isolates counterparty dependence. The combined scenario is also warranted because the committee’s question concerns reliance on reinsurance under adverse conditions, rather than each risk considered separately. Its result is below both standalone ratios and makes the combined vulnerability visible.

The standalone impacts would sum to a $60 million reduction in available capital and a $16 million increase in capital need. The combined scenario instead has a $68 million reduction and a $22 million increase. The additional $8 million reduction and $6 million increase are part of the supplied combined output; they are not an arithmetic result that can be inferred from the standalone tests.

Check. The combined result uses $242/282$, which is approximately 0.8582. The denominator is above the numerator, so coverage is below 100%, as the result indicates.

Common mistake avoided. Treating the combined scenario as just the sum of standalone impacts would give $250/276=0.905797$. That is not the specified combined result and would miss the additional effects represented by its model output.

Build a Test Set With Complementary Coverage

A useful selection process makes the purpose and coverage of each test explicit. It asks whether the test addresses a material exposure, whether its form fits the question, and whether its result could inform a decision or change a conclusion. It also checks whether existing monitoring already covers the risk, so the ORSA does not add tests that are repetitive without adding insight.

Start with the decision.
State what the assessment needs to inform, such as a reinsurance renewal, a risk limit, or confidence in projected capital adequacy.

Identify the relevant exposures.
Use the risk inventory and planning assessment to identify which material risks or changing assumptions bear on that decision.

Choose a suitable test form.
Use a standalone sensitivity when isolating one exposure answers the question. Use a broader scenario when the question depends on multiple conditions or their interaction.

Check coverage and duplication.
Consider whether existing analysis already examines the risk and whether the proposed test adds a distinct decision-relevant view.

Record the rationale and limits.
Explain why the test was selected, what it addresses, and what important exposures or questions it does not address.

For example, an insurer may already review equity-market sensitivities routinely in its investment monitoring. That does not make equity risk immaterial; it means the ORSA selection should consider whether repeating the same analysis adds decision value. If a business-plan change creates a new equity concentration, the test set may need to change. Selection should follow the current risk profile and decision needs, not a fixed list.

Worked Example: Prioritizing Tests Within a Limited Review

A hypothetical insurer has projected available capital of $150 million and internal capital need of $125 million at its two-year planning point. The board has asked about catastrophe reinsurance, cyber controls, and overall resilience. The assessment provides these standalone impacts: catastrophe, $\Delta A=-$24$ million and $\Delta C=+$5$ million; cyber disruption, $\Delta A=-$10$ million and $\Delta C=+$3$ million; and an equity decline, $\Delta A=-$16$ million and $\Delta C=-$2$ million. A broader adverse scenario relevant to overall resilience has $\Delta A=-$42$ million and $\Delta C=+$14$ million. Routine investment monitoring already reports the equity sensitivity. Select an ORSA test set that addresses the board’s questions and calculate the relevant ratios.

Method. Calculate each ratio from the projected baseline of $150/125=1.2000$, or 120.00%. Then select tests by decision relevance and added coverage, not by ranking the standalone ratio movements alone.

The catastrophe sensitivity gives available capital of $150-24=$126$ million and capital need of $125+5=$130$ million. Its ratio is $126/130=0.969231$, or 96.92%.

The cyber sensitivity gives available capital of $150-10=$140$ million and capital need of $125+3=$128$ million. Its ratio is $140/128=1.093750$, or 109.38%.

The equity sensitivity gives available capital of $150-16=$134$ million and capital need of $125-2=$123$ million. Its ratio is $134/123=1.089431$, or 108.94%.

The broader scenario gives available capital of $150-42=$108$ million and capital need of $125+14=$139$ million. Its ratio is $108/139=0.776978$, or 77.70%.

Selection rationale. Include the catastrophe sensitivity for the reinsurance decision and the cyber sensitivity for the control decision. Include the broader scenario for the board’s overall-resilience question, which neither isolated test answers. The equity sensitivity remains relevant, but the facts state that routine investment monitoring already reports it; the ORSA can reference that analysis rather than repeat it without a distinct purpose. If the risk profile or board’s question changes, that selection should be reconsidered.

Check. The broader scenario ratio is $108/139$, approximately 0.7770. It is lower than either selected standalone ratio and reflects the scenario’s stated combined impacts.

Common mistake avoided. Selecting only the test with the lowest ratio would choose the broader scenario but leave the reinsurance and cyber-control decisions without risk-specific evidence. A broad result does not identify or replace every targeted analysis needed for separate decisions.

Common Selection Traps

Choosing tests by severity alone. A severe-looking result is not automatically the most useful test. Link selection to a material exposure and a decision need.
Using a standalone test to answer a combined-risk question. Isolated sensitivities help attribute exposure; they do not establish the outcome under simultaneous conditions.
Using a broad scenario as a substitute for isolation. A combined result may not show which exposure drives the outcome. Add targeted tests when a decision requires that distinction.
Ignoring changes to capital need. A stress can affect both available capital and internal capital need. Calculate both before comparing coverage.
Repeating analysis without added value. Existing monitoring may already cover an exposure. Confirm whether an ORSA test adds a distinct perspective or decision use before including it.
Assuming the chosen set is exhaustive. A selected set is a reasoned coverage choice, not proof that untested risks cannot matter. State material gaps and revisit selection when the risk profile changes.

Key takeaway: Select ORSA stress tests to match material risks with decision needs. Use standalone sensitivities to isolate exposures, broader scenarios to examine combined conditions, and both where the assessment needs both forms of evidence.

Check Your Understanding

For each problem, use the stated stress impacts and explain the selection point where asked. Amounts are hypothetical and in millions of dollars.

A hypothetical insurer has available capital of $400 million and internal capital need of $320 million. A standalone stress reduces available capital by $30 million and raises capital need by $10 million. Calculate the baseline and stressed coverage ratios. (Answer: Baseline 125.00%; stressed $370/330=1.1212$, or 112.12%.)
A projected position has available capital of $225 million and internal capital need of $180 million. A rate sensitivity changes these by $\Delta A=-$12$ million and $\Delta C=-$8$ million. Calculate the stressed ratio and state whether the test can still be decision-relevant if coverage increases. (Answer: $213/172=1.2384$, or 123.84%; yes, it can inform a hedging decision even though coverage increases.)
A hypothetical insurer has available capital of $280 million and internal capital need of $240 million. A catastrophe sensitivity has impacts $(- $30,\ +$8)$, a counterparty sensitivity has impacts $(- $14,\ +$4)$, and the supplied combined scenario has impacts $(- $52,\ +$18)$, where each pair is $(\Delta A,\Delta C)$. Calculate the combined ratio and explain why the standalone tests do not replace it. (Answer: $228/258=0.8837$, or 88.37%; the combined test assesses the stated joint conditions, which the isolated tests do not.)
A board asks whether a new distribution channel creates operational exposure and whether the insurer remains resilient when that exposure occurs alongside a market decline. Which test forms are appropriate, and why? (Answer: A standalone operational sensitivity helps isolate the channel exposure; a broader scenario is also appropriate to examine the combined operational and market conditions.)
A projected insurer has available capital of $180 million and internal capital need of $150 million. A selected sensitivity reduces available capital by $20 million and increases capital need by $5 million. Calculate stressed coverage, then identify one reason the result alone may not answer a broader resilience question. (Answer: $160/155=1.0323$, or 103.23%; it does not by itself show the effect of other risks or assumptions changing together.)

What You'll Learn