A reserve review matters because claims can continue to develop long after an insurer records them. A claim that looks settled today may still involve new costs, while an open claim may close for less than expected. We use reserve adequacy analysis to check whether booked amounts can cover future claim obligations.
Ultimate claims means the total expected cost of covered claims after all claims have fully developed and settled. Estimates depend on claim maturity, data quality, and professional judgment. A young book of claims needs different care than a mature one with steady patterns.
We can make the review useful by keeping its scope clear and its assumptions visible. The process starts with claims data and ends with a record of the choices, tests, and actions behind the estimate. This also helps insurer financial reporting explain why a reserve changed or stayed level.
The most useful assessment makes its assumptions visible, not just its final number.
Key Takeaways
- Compare carried reserves with a matching estimate of ultimate claims.
- Use reconciled data and methods suited to the claims being reviewed.
- Check whether old claim patterns still fit current operations and costs.
- Treat a ratio as a signal to investigate, not a final verdict.
- Record assumptions, uncertainty, and follow-up actions.
What Does Reserve Adequacy Measure?
A reserve is adequate when the amount held is sufficient for the claim obligations included in the assessment, based on selected estimates and assumptions. Reserve adequacy compares two different amounts: the projected obligation and the amount booked for unpaid claims.
Unpaid claim obligations can include reported claims and incurred-but-not-reported claims, often called IBNR. The projected obligation estimates what those claims may cost after development. The booked reserve, by contrast, is the amount already recorded in the insurer’s accounts.
A favorable or adverse difference is a reason to investigate, not proof by itself. A favorable difference means the selected estimate falls below the booked amount. An adverse difference means the estimate is higher, which may point to a shortfall under the chosen assumptions.
Compare the booked balance with a projected obligation that covers the same claims, losses, and expenses.
The meaning of a result depends on consistent definitions. One estimate may include loss adjustment expenses, while another may include losses alone. A reserve balance may also cover a different set of lines or claim periods.
That mismatch can make a sound balance appear weak, or make a thin balance appear sufficient. We should also note the valuation date, because claim payments and new information change the picture over time.
Before interpreting any difference, confirm that the booked amount and projected obligation cover the same claims and expenses. A reserve assessment is only as clear as the terms used to define its inputs.
Which Insurance Obligations Belong in the Review?
A P&C review generally tests unpaid losses and loss adjustment expenses, while a life asset adequacy review focuses on whether supporting assets and cash flows can meet policy obligations. These are related but distinct questions about an insurer’s ability to meet future commitments.
Reserve reviews support a focused view of claim liabilities. A broader insurance solvency risk assessment can also examine capital, market, credit, liquidity, and operational risks.
P&C loss and loss adjustment expense reserves
For property/casualty insurers, the review usually covers unpaid losses and loss adjustment expenses. Those expenses include costs tied to handling and settling claims. The scope may also include other reserves when an applicable opinion or legal rule requires them.
ASOP No. 36 applies to qualifying written opinions on P&C loss, loss adjustment expense, or other reserves. It covers opinions prepared to meet NAIC Property/Casualty Annual Statement Instructions, applicable law, or a representation of compliance. The scope of ASOP No. 36 helps define which actuarial work falls under the standard.
Casualty insurance reserve adequacy analysis can involve long reporting delays and claim costs that change over time. A review should make clear which business, accident periods, expenses, and valuation date it covers.
Life insurance asset adequacy
Life asset adequacy testing asks whether assets and expected cash flows can support policy liabilities. The American Academy of Actuaries’ Asset Adequacy Analysis Practice Note was updated in September 2024 for companies filing the NAIC Life, Accident & Health/Fraternal statutory annual statement.
The practice note describes cash flow testing and gross premium valuation as asset adequacy methods. It is not an ASOP and does not bind actuaries. A life practice note and a P&C loss-reserve opinion address related solvency concerns, but they are not interchangeable.
Keep the liability being tested explicit, especially when a company writes both P&C and life business. Naming the scope at the start prevents a claim-reserve result from being mistaken for a test of all policy obligations.
What Data Should You Gather Before Projecting Claims?
Start with a reconciled history of paid and incurred losses, then add the claim, exposure, and reserve detail needed to explain how those losses changed. Comparable records help us build a useful view of claims development rather than simply collect more files.
Gather five types of information before choosing a projection method. Each type answers a different question about claim costs, business volume, or the records behind the balance.
- Paid and incurred losses by accident, policy, or report period.
- Claim counts, open-claim counts, and details about claim status.
- Current carried reserves by line of business.
- Exposure and premium measures tied to the claims being reviewed.
- Operational and accounting history that explains changes in records.
Paid losses show cash already spent. Incurred losses combine paid amounts with case reserves, the amounts assigned to open claims. Claim counts add another view: a rising loss total may reflect more claims, higher costs per claim, or both.
Exposure and premium measures help us understand how much business generated the claims. Operational history can show when a new claims system or accounting rule changed the way records were stored or reported.
P&C annual statements include Schedule P loss-development information. Reserve analyses also commonly reconcile source records with annual-statement or audited totals. A difference may point to a timing issue, a coding problem, or a mismatch in scope.
Missing records can distort projections, as can data that use different loss or expense definitions. We should keep the source, date, and meaning of each total clear enough for another reviewer to trace.
Investigate an unreconciled total before using it to calibrate a projection. A clean-looking triangle cannot fix records that do not agree with the company’s reported balance.
How Do Development Patterns Produce Ultimate Claims Estimates?
A loss development factor projects how a cumulative claim amount may grow from its current development age to its ultimate value. A loss triangle arranges claim experience by origin period and development period, so we can compare similar stages. The pattern is evidence, not a promise.
A development factor takes a current cumulative amount toward an estimate of ultimate claims. For example, reported claims multiplied by a loss development factor can produce an ultimate claims estimate. The selected factor reflects observed development and judgment about whether that pattern still fits.
The following 2024 NAIC annual-statement dashboard figures show how amounts changed across development ages. The incurred-loss amount shown for accident year 2017 is $373.4 billion at both listed points.
| Metric | Accident year | Earlier point and amount | Later point and amount |
|---|---|---|---|
| Incurred losses | 2017 | Development year 2 (2019), $373.4B | Development year 7 (2024), $373.4B |
| Paid losses | 2020 | Development year 0 (2020), $185.7B | Development year 4 (2024), $353.5B |
| Carried reserves | 2019 | Development year 0 (2019), $213.3B | Development year 4 (2023), $42.17B |
| Case reserves | 2021 | Development year 0 (2021), $96.59B | Development year 3 (2024), $28.89B |
For accident year 2020, the reserve-to-paid ratio was 1.21 at development year 0. Its paid-to-incurred ratio was 88.9% at development year 4 in 2024.
These measures show that paid losses can rise as claims settle, while carried and case reserves can fall. They do not mean every later period will follow the same path. The 2024 dashboard development figures describe observed experience, not a forecast for every insurer.
Claim handling, business mix, or settlement timing can change, breaking the link between past and future development. Check that the periods compared have consistent definitions before treating their development as a repeatable pattern.
How Do You Calculate the Reserve Adequacy Ratio?
Divide current reserves by the projected ultimate claims for the same block of business, then multiply by 100% to calculate the reserve adequacy ratio. The calculation is simple arithmetic, but its result depends on matched inputs.
Reserve Adequacy Ratio (%) = (Current Reserves / Ultimate Claims Estimate) x 100%. If current reserves are lower than the selected ultimate claims estimate, the ratio is below 100%. If the two amounts match, the ratio is 100%.
For a symbolic example, call current reserves R and the ultimate claims estimate U. When R is less than U, R divided by U produces a value below one, so the ratio falls below 100%.
When R equals U, the division equals one and the ratio is 100%. The formula does not change if the estimate includes loss adjustment expenses, but both amounts must include them on the same basis.
Compare the same line, valuation date, loss basis, and expense scope. A ratio based on one year of claims cannot be compared fairly with reserves covering several years.
We should not mix gross and net amounts, either. Reinsurance can change the amount an insurer expects to pay, so the reserve balance and ultimate estimate must use a matching treatment.
Label the valuation date and included loss expenses beside every ratio you report. Those details tell a reader what the percentage actually compares.
How to Conduct Insurance Reserve Adequacy Analysis
A sound review follows a repeatable sequence, but the final estimate still depends on whether the methods fit the company’s data and claims. These five steps help us keep the work organized without turning actuarial judgment into a mechanical exercise.
Use more than one suitable projection method, and explain why each method fits. A method may be unsuitable when its data assumptions do not match the business being reviewed.
- Define the block of business and the valuation date.
- Reconcile and validate the experience used in the review.
- Project ultimate claims with more than one suitable method.
- Compare the selected estimate or range with carried reserves.
- Document assumptions, sensitivities, and actions to consider.
Incurred-loss, paid-loss, ultimate-claim, claim-count, and loss-ratio approaches are distinct evidence streams. Each can show a different part of the claim story. The best fit depends on the data available and how claims develop in that line.
A paid view relies on amounts already paid, while an incurred view also uses case reserves. Claim counts can help separate changes in frequency from changes in cost. An ultimate-claim view carries the projection through to expected final costs.
The sequence gives the team a shared way to discuss evidence. It does not remove the need to judge data quality, method fit, or the weight assigned to each result.
A broad risk review can also consider the insurer’s capital and other exposures, but it does not replace claim reserve testing. Keep the purpose of each analysis clear so decision-makers do not treat different measures as interchangeable.
Record why a method was accepted or rejected, not only the estimate it produced. That record lets a later reviewer understand the choices behind the result.
Which Changes Can Distort Historical Claims Data?
A new claims system or case-reserving rule can create an apparent change in loss development even when underlying claim costs have not changed. Process changes matter because a triangle reflects how the company records and handles claims, not just the cost of claims.
Check whether a break in the data came from the business or from the way the business was recorded. Underwriting, claims handling, data processing, and accounting changes can all alter observed patterns.
Changes in claims operations and accounting
A faster reporting process can shift claims into earlier periods. New settlement priorities may change how quickly claims close. A different case-reserve practice may raise or lower incurred losses without changing paid amounts.
Policy limits and deductibles can alter the insurer’s share of a claim. System definitions can also change what counts as open, closed, paid, or incurred. Even small coding changes can make older and newer records hard to compare.
Choosing periods and segments that remain comparable
Accident-year, policy-year, and report-year views place claims into different time groups. A change in reporting delay may stand out more clearly in a report-year view, while policy-year results can show a shift tied to the coverage period.
Use a focused view to test whether a shift affects one group or the whole portfolio. Look for these signals:
- A sudden shift in open-claim counts.
- A change in case reserves per open claim.
- A break in the mix of limits, territories, or claim types.
More homogeneous segments can clarify a change, but excessive segmentation can leave too little credible experience. Compare the affected data with a less distorted view before discarding the full history.
How Should You Adjust for Frequency, Severity, and Inflation?
Historical losses need adjustment when claim frequency, average severity, or the cost of settling claims has shifted since the experience period. Frequency and severity describe different parts of claim cost, so we should examine them separately.
Frequency measures claim counts relative to an exposure base, while severity measures loss cost per claim. A rise in frequency means more claims per unit of exposure. A rise in severity means each claim costs more on average.
Economic inflation can raise repair, medical, legal, or other claim costs. Claim-cost trends may also shift for reasons beyond broad inflation. Those changes affect how well older losses represent the expected cost of future claims.
Frequency and severity can move in different directions. For example, fewer claims do not necessarily mean lower total losses if the remaining claims become more costly. The distinction between claim frequency and severity helps us avoid treating one measure as a stand-in for the other.
Use this four-part adjustment checklist:
- Identify the trend measure used for the claim type.
- Select a period that reflects current conditions.
- Apply the adjustment consistently to relevant losses.
- Test the result under alternative assumptions.
A common mistake is applying one trend rate to every line or claim type without checking the experience. Auto repair costs and liability claim costs may not move in the same way, so a single rate can hide meaningful differences.
Show the unadjusted estimate beside the trend-adjusted estimate so reviewers can see how much the assumption changes the result. The comparison makes the effect of the adjustment plain.
Which Reserving Methods Should You Compare?
Comparing methods helps reveal whether a result depends on one particular assumption or data view. Different methods draw on different evidence, so a comparison can expose a result that changes sharply when one assumption changes.
No method is best for every line of business. We should give more weight to approaches whose evidence matches the maturity and stability of the portfolio.
Development-based projections
Chain ladder uses observed development patterns to project later claim costs. It can be applied to paid losses or incurred losses. The paid view uses cash already spent, while the incurred view also reflects case reserves on open claims.
These views can differ when case-reserve practices change or claims settle slowly. A mature, consistent history can support development-based projections. Shifts in reporting or settlement patterns can weaken their value.
Expected-loss and hybrid projections
Bornhuetter-Ferguson combines expected losses with observed development. It can help when experience is immature or volatile because it does not rely only on early claim development.
Cape Cod blends observed experience with an expected-loss view using exposure and development-adjusted experience. Both approaches use an expected-loss starting point, so that assumption needs support from the business context and data.
Other evidence can include claim counts, average claim severity, and loss ratios. The table shows the main trade-offs.
| Method | Primary evidence | Useful situation | Key limitation |
|---|---|---|---|
| Chain ladder | Observed development | Mature, consistent history | Sensitive to changing patterns |
| Bornhuetter-Ferguson | Expected losses plus observed development | Immature or volatile experience | Depends on expected-loss assumption |
| Cape Cod | Exposure and development-adjusted experience | Blending observed and expected loss views | Depends on exposure and trend assumptions |
Give more weight to methods whose assumptions fit the maturity and stability of the portfolio. A method that looks precise can still mislead when its core assumptions do not fit.
How Do You Test Estimates and Select a Reasonable Range?
A selected estimate is more credible when its historical performance and sensitivity to assumptions have been examined. Testing the estimate means checking how it performed and how much it shifts under reasonable changes.
A range reflects uncertainty and judgment, not a statistical guarantee. It helps show how alternative assumptions affect the view of unpaid claims, rather than suggesting that every possible outcome falls inside fixed bounds.
Back-testing compares prior projections with later observed outcomes. Large gaps may show that an assumption needs review, or that the portfolio changed after the original estimate.
Check how well each method fits the observed data. Seasonal or cyclical effects can also matter, especially when recent experience covers only part of a cycle.
Test material assumptions, including trends, expected losses, and development patterns. Then note the strengths and weaknesses of each method, so a reviewer can see which evidence carries the most weight.
Before relying heavily on incurred development, check whether case-reserve adequacy shifted over time. If case reserves changed while claim costs did not, incurred patterns may give a misleading signal.
ASOP No. 36 was revised by the Actuarial Standards Board in March 2024. The revision expanded guidance on risk of material adverse deviation, including added disclosure requirements.
Preserve the rationale for the selected point or range alongside the sensitivity results. A clear record explains why the estimate was chosen despite the uncertainty.
When Should You Segment the Portfolio?
Segment the portfolio when business mix or claim behavior differs enough that one blended estimate would hide a meaningful pattern. Separate views can show which lines or claim groups drive the total result.
Use segments when they improve comparison without leaving too little credible experience. A company-wide total can blend together lines with very different claim costs and development patterns.
Auto, liability, and property experience may develop at different speeds and respond to different cost pressures. Reviewing them separately can reveal a pattern that disappears in a combined estimate.
Claim type or loss size can also isolate a concentrated exposure. A large-loss layer may need its own review when a few severe claims dominate the overall result.
For example, a liability portfolio may have many small claims and a handful of very large claims. If those few losses drive most of the current estimate, a separate large-loss view can show whether the broad pattern applies to that layer.
Useful dimensions can include state, subline, territory, size of risk, and size of loss. The right split depends on whether the categories reflect real differences in exposure or claim behavior.
Too many narrow segments can leave each group with too few claims to support a stable projection. We can then mistake random movement for a lasting change.
Keep a segment separate only when its data remains credible enough to support a useful projection. A split that cannot support a sound estimate adds detail without adding clarity.
What Do Reserve Development and Industry Results Signal?
An industry can report stronger underwriting results while a particular line still experiences adverse development in prior-year reserves. Reserve development occurs when later estimates change for claims from earlier periods.
Adverse development means prior-period claim estimates increased; favorable development means they decreased. Either result can affect current financial measures, even though the claims began in earlier years.
U.S. commercial auto insurers had $2.0 billion of adverse reserve development in 2025. That line-specific result can coexist with broader improvement across the P&C industry.
The U.S. P&C industry net combined ratio fell from 96.6 in 2024 to 92.9 in 2025. Net underwriting gain rose from $23 billion to about $63 billion over those years.
| Metric | 2024 | 2025 |
|---|---|---|
| Net combined ratio | 96.6 | 92.9 |
| Net underwriting gain | $23 billion | About $63 billion |
| Catastrophe losses as combined-ratio points | 8.8 | 7.6 |
Catastrophe losses accounted for 8.8 combined-ratio points in 2024 and 7.6 points in 2025. Those figures help place the change in underwriting results in context.
The 2024 and 2025 industry results can prompt questions about trends, but they cannot establish whether a particular insurer’s reserves are adequate. Company results depend on its own claims, mix, and assumptions.
Treat industry measures as context, then test the company’s own claims experience and assumptions. A strong market result does not erase adverse development in a specific book.
What Should a Reserve Adequacy Report Include?
A useful report lets a reviewer trace each conclusion from reconciled data through estimates to a decision. A decision-ready report ties evidence to owners, so open questions do not vanish after the analysis ends.
Show the data, estimate comparison, uncertainty, and proposed action in a format leaders can use. A reusable insurance reserve adequacy analysis template can keep those items consistent across review periods.
A decision-ready results table
For each line, show current reserve level, ultimate claims estimate, reserve adequacy ratio, and any proposed adjustment. The supporting table can also name the evidence, decision, and owner for each key task.
| Report element | Evidence shown | Decision it supports | Owner |
|---|---|---|---|
| Data reconciliation | Source totals and exceptions | Whether inputs are usable | Finance |
| Estimate comparison | Carried reserves and selected ultimate estimate by line | Whether a shortfall or redundancy is indicated | Actuarial |
| Sensitivity | Alternative assumptions and range | How uncertainty affects the conclusion | Actuarial |
| Action log | Proposed adjustment and approval status | Who will decide and when | Management |
Communication, actions, and ownership
Explain the main result in plain terms, then show the assumptions that matter most. A short account of the data limits can help leaders separate a real change from a weak comparison.
When the report and findings are submitted to the company, they must be communicated to the board or audit committee. Direct presentation is not required by the NAIC.
Keep a record of questions, decisions, and later changes to the action plan. Assign an owner and due date to every proposed follow-up action.
Which U.S. Actuarial and NAIC Requirements Apply?
The applicable rules depend first on the insurer’s domicile state and on whether the work is a qualifying written actuarial opinion. State rules matter because NAIC requirements do not override differing laws in the insurer’s state.
Check the domicile state’s adopted requirements before applying a general NAIC rule. ASOP No. 36 covers qualifying written opinions on P&C loss, loss adjustment expense, or other reserves.
The NAIC catalog includes the 2026 Annual Statement Instructions for Property/Casualty annual statements and required supplemental information. Under NAIC requirements described for 2026, carried reserves need not equal the Appointed Actuary’s estimate.
An opinion is not “reasonable” if carried reserves fall outside the actuary’s range. The Appointed Actuary may give an explicit range or decide whether carried reserves fall within an implicit range.
Use this five-item compliance check:
- Confirm the domicile state’s adopted rules.
- Identify whether the opinion falls within ASOP No. 36.
- Document the actuary’s range or implicit-range conclusion.
- Check whether carried reserves fall within that range.
- Verify the applicable annual-statement instructions.
NAIC requirements are not authoritative unless adopted by the insurer’s state of domicile. Differing state laws and regulations take precedence.
The NAIC does not require a quarterly Statement of Opinion or an actuarial estimate supporting reserves in quarterly statements. It also does not require the same Appointed Actuary to be reappointed each year, though appointment is required upon initial appointment.
Check current domicile-state requirements alongside the applicable annual-statement instructions before relying on a general NAIC rule. A rule that applies in one state may not apply in another.
How Can You Monitor Reserve Adequacy Between Reviews?
A reserve estimate should be revisited when new claim experience materially departs from the pattern or assumptions used to select it. Monitoring is a trigger for investigation, not an automatic order to book an adjustment.
Watch paid and incurred emergence, case-reserve movement, claim frequency and severity, large losses, and claims operations. A shift in any of these areas may call for a closer look at the estimate.
Compare emerging paid and incurred losses with the selected estimates. Review large claims as they develop, since a small number of severe losses can change a line’s results.
Watch for changes in how claims are reported, reserved, or settled. A steady company-wide total can conceal deterioration in one line and improvement in another.
When a signal appears, identify what changed and whether the estimate still fits the observed experience. A temporary swing may call for more review, while a lasting shift may support a new projection.
Use monitoring to open an investigation, not to make an automatic reserve adjustment.
Set a review trigger around material departures from the selected estimate, and record who investigates each one. Named ownership helps the company respond before the next full review.
Conclusion
A reserve adequacy conclusion is strongest when the supporting evidence is traceable and its uncertainty is visible. Comparable data, several suitable projections, and transparent judgment help us explain what the estimate means.
The next step can be small and concrete: choose one line of business and reconcile its current carried reserves to the underlying claim data before calculating a ratio. That first check can reveal whether the basic inputs support a meaningful comparison.
Start with one line of business, reconcile its figures, and document the assumptions behind its estimate.