Why Capital Comparisons Need a Common Basis
Two insurer capital assessments can report different dollar amounts or ratios without disagreeing about the same thing. One result may apply to a consolidated group and another to a legal entity; one may use a broader definition of available capital; or the assessments may cover different dates, risks, or diversification assumptions. A numerical comparison made before checking these features can produce a confident but invalid conclusion.
The appropriate task is not to force different models into identical calculations. It is to identify which differences are attributable to the model and which arise because the results describe different objects. Comparisons should be presented as like-for-like only when the entity, capital definitions, measurement basis, risk coverage, and treatment of diversification are sufficiently aligned.
A capital comparison is a comparison of assessments, not just numbers. Before interpreting a difference, specify what each result measures. In particular, distinguish available capital from required capital: a change in either changes the ratio, but the two changes have different meanings.
A Six-Part Comparability Check
Check whether each result covers the same legal insurer, insurance group, or other consolidation perimeter. Group resources are not automatically available to every subsidiary.
Identify what counts in \(C\), including any restrictions or deductions, and what is included in \(K\). Do not assume two measures called “capital” use the same definition.
Record the valuation date and the relevant risk horizon. An amount calculated at one date or over one horizon cannot be treated as a same-basis result without an appropriate bridge or recalculation.
List the included risks and whether amounts are gross or net of risk transfer. A model that omits a risk category does not measure the same total requirement as one that includes it.
Establish whether charges are summed, aggregated using a dependence assumption, or subject to a limit on diversification credit. Different aggregation rules can produce different totals from identical component charges.
Compare each result with the threshold or interpretation rule belonging to that assessment. Equal ratios, or equal distances from a threshold, need not have equal meaning across different scales.
A useful comparison records these six items beside each result. If a basis difference cannot be quantified reliably, state that limitation rather than presenting a false precision. The earlier material on capital ratios and diversification is useful background, but the central discipline here is to keep each adjustment tied to the definition it changes.
Worked Comparisons
Worked Example: Group and Legal-Entity Results on Different Capital Definitions
Two hypothetical assessments are being discussed as if they describe the same insurer. Assessment A reports consolidated group available capital of $1,200 million and required capital of $800 million. Within that group, the operating legal entity has available capital of $780 million and required capital of $600 million. Assessment B reports the operating legal entity with available capital of $720 million and required capital of $540 million. B’s numerator deducts $60 million of assets that A includes; B’s denominator is $60 million lower because it gives an additional $60 million of diversification credit. Reconcile the comparison to the stated operating-entity basis used by A.
Method. Calculate the reported ratios first, then isolate the entity and definition differences. The comparison basis is the operating entity with A’s stated definitions.
Assessment A’s consolidated ratio is \(1{,}200/800=1.500\), or 150.0%. Assessment B’s operating-entity ratio is \(720/540=1.3333\), or 133.3% to one decimal place. These reported figures differ by 16.7 percentage points, but they do not compare the same entity.
For A’s operating entity, the ratio is \(780/600=1.300\), or 130.0%. To express B on that stated basis, add back the $60 million numerator deduction and remove the $60 million additional diversification credit from its denominator:
The aligned B ratio is \(780/600=1.300\), or 130.0%, matching A’s operating-entity ratio. The reconciliation checks because both aligned amounts now equal A’s operating-entity amounts. The group result remains a group result and must not be substituted for the legal-entity result.
Common mistake. Comparing 150.0% with 133.3% and concluding that A gives the operating insurer a stronger assessment mixes a group result with a legal-entity result and leaves two definition differences unadjusted. Compare the same entity on the same capital basis instead.
Worked Example: Separating Measurement-Date and Exposure Changes
A hypothetical insurer’s June 30 assessment reports \(C=\$840\) million and \(K=\$560\) million. Its December 31 assessment reports \(C=\$900\) million and \(K=\$600\) million. The two assessments use the same entity and definitions. A documented bridge shows that, between the dates, retained earnings increased available capital by $120 million, a dividend reduced it by $60 million, and new business increased required capital by $40 million. Compare the reported ratios and identify the effect of the exposure change on the December result.
Method. Compute each date’s ratio, verify the capital bridge, and then calculate the December ratio on the June exposure basis. The latter is a diagnostic calculation, not a replacement for the December assessment.
At June 30, \(R=840/560=1.500\), or 150.0%. The available-capital bridge gives \(840+120-60=900\) million at December 31. The required-capital bridge gives \(560+40=600\) million.
At December 31, the reported ratio is \(900/600=1.500\), also 150.0%. The equal ratios do not mean the underlying amounts or exposures were unchanged: required capital increased by $40 million, while available capital increased by a net $60 million.
To isolate the effect of new business on the December denominator, hold December available capital fixed and remove the documented $40 million requirement increase. The resulting ratio on the June exposure basis is:
Thus, on this specified diagnostic basis, the December exposure increase reduces the ratio from 160.7% to the reported 150.0%. The calculation checks against the bridge: the only denominator change removed is the stated $40 million. It does not imply that capital should be assessed at June exposures at December; it shows why measurement dates and portfolio changes matter.
Common mistake. Calling the two assessments “identical” because both ratios equal 150.0% overlooks their different capital amounts and exposure bases. Conversely, comparing $560 million with $600 million as though the $40 million difference alone showed deterioration ignores the changed portfolio and available capital.
Worked Example: Risk Coverage and Diversification
Two hypothetical models use the same component charges, in millions: property 400, credit 300, and operational risk 100. Model P includes all three risks and aggregates them using the stated formula with pairwise correlation \(0.25\) for every pair. Model Q includes property and credit only and adds those charges without diversification. Calculate both reported amounts, then show why their difference cannot be attributed solely to aggregation.
Method. Apply Model P’s specified correlation aggregation formula, then calculate Model Q’s included-risk total. For three charges \(K_1,K_2,K_3\) with the same pairwise correlation \(\rho\), the model’s aggregate is:
This expression follows by expanding the variance of a sum: the variance is the sum of the component variances plus twice each pairwise covariance. Under the model’s specified correlation structure, each covariance contribution is represented by the corresponding correlation-weighted cross-product. Taking the square root gives the aggregate measure. The formula applies here because the model explicitly specifies this aggregation rule and common pairwise correlation; it is not a universal rule for every rating model.
For Model P, the sum of squared charges is \(400^2+300^2+100^2=160{,}000+90{,}000+10{,}000=260{,}000\). The pairwise products sum to \(400(300)+400(100)+300(100)=120{,}000+40{,}000+30{,}000=190{,}000\). The correlation term is \(2(0.25)(190{,}000)=95{,}000\). Therefore:
Model Q reports \(400+300=\$700\) million. Its result excludes the $100 million operational charge as well as using a different aggregation rule. As a diagnostic, adding the omitted operational component to Q’s un-diversified total gives \(700+100=\$800\) million. That figure and Model P’s $595.8 million still differ because the aggregation rules differ, but Q’s actual $700 million also differs in risk coverage.
The arithmetic checks: Model P includes all three components in the formula; Model Q’s reported amount contains only property and credit. Common mistake. Treating \(700-595.8=\$104.2\) million as “the diversification benefit” is wrong because the models do not cover the same risks. Align risk coverage before attributing any difference to aggregation.
Worked Example: Interpreting Headroom Against Different Thresholds
Two hypothetical assessments each report a capital ratio of 1.40. Assessment M uses a stated reference threshold of 1.25, while Assessment N uses a stated reference threshold of 1.10. Calculate each ratio’s relative headroom above its own threshold, defined for this comparison as \(R/T-1\), where \(T\) is that assessment’s threshold. Explain what the calculation does and does not establish.
Method. Divide each reported ratio by its own threshold and subtract 1. For M, \(1.40/1.25=1.12\), so relative headroom is \(1.12-1=0.12\), or 12.0%. For N, \(1.40/1.10=1.272727\ldots\), so relative headroom is \(0.272727\ldots\), or 27.3% to one decimal place.
Both calculations use the supplied ratios and thresholds; the arithmetic checks directly. They show the ratios’ positions relative to their respective reference thresholds under the stated definition. They do not establish that the two assessments use equivalent thresholds, rating categories, or interpretations. A threshold is meaningful within its own assessment framework.
Common mistake. Equal reported ratios do not imply equal headroom when thresholds differ. Nor does the larger relative headroom automatically mean a stronger rating across agencies; that inference would require comparable scales and interpretations, not just this arithmetic.
How to Report a Defensible Comparison
A concise, defensible comparison distinguishes the observed difference from its possible causes. State the reported values and then identify whether the results match on each relevant basis. If a numerical adjustment is made, show its amount, direction, and justification. Do not treat an adjustment as exact when the available information supports only a qualitative diagnosis.
| Comparison issue | Question to resolve | Risk if ignored |
|---|---|---|
| Entity | Do both results cover the same insurer or group? | Group resources may be mistaken for entity-level support. |
| Capital definition | Are available and required capital defined consistently? | A ratio difference may reflect numerator or denominator adjustments. |
| Date and horizon | Do the results use the same measurement date and risk horizon? | Changes in capital, exposure, or assumptions may be misread as model differences. |
| Risk coverage | Are the same risks included, on a gross or net basis? | One result may omit a charge or recognize a risk-transfer benefit absent in the other. |
| Diversification | Are the aggregation rules and limits comparable? | Different totals may be incorrectly attributed to different component risks. |
| Threshold scale | Is each result interpreted against its own stated threshold? | Equal ratios or headroom may be assigned an unwarranted common meaning. |
The central exam skill is diagnosis before interpretation. A capital result is not fully described by a dollar amount or ratio alone. A comparison that clearly names its basis—and acknowledges what cannot be reconciled—supports better risk analysis and is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus.
Check Your Understanding
For each problem, check both the arithmetic and whether the stated comparison is on a common basis.
- A group assessment reports available capital of $960 million and required capital of $640 million. Its operating legal entity has $540 million available capital and $450 million required capital. Calculate both ratios and state why the group ratio cannot stand in for the legal-entity ratio. (Answer: Group 150.0%; legal entity 120.0%; they cover different entities.)
- At one date an insurer has \(C=\$720\) million and \(K=\$480\) million. At a later date, its available capital is $780 million and its required capital is $520 million. Calculate both ratios and the later required-capital increase. (Answer: Both ratios are 150.0%; required capital increased by $40 million.)
- Three component charges are 180, 240, and 100, with pairwise correlation 0.50 under the common-correlation aggregation formula used in this tutorial. Calculate the aggregate to one decimal place. (Answer: \(\sqrt{185{,}200}\approx \$430.3\) million.)
- An assessment reports \(R=1.32\) against its own threshold \(T=1.20\). Using relative headroom \(R/T-1\), calculate the headroom as a percentage. (Answer: 10.0%.)
- A model reports $310 million net of reinsurance and excludes an operational charge of $45 million. For a stated comparison basis that includes the operational charge and removes a $20 million reinsurance credit, calculate the adjusted amount, assuming the stated adjustments are additive and no other changes apply. (Answer: \(310+45+20=\$375\) million.)