Insurance Financial Analysis and ModellingIntroduction and contextInsurance companies receive premiums today in exchange for uncertain obligations that may arise in the future. Their financial performance therefore cannot be assessed solely from current revenue and profit. Analysts must examine underwriting quality, claim development, reserve adequacy, capital strength, reinsurance dependence, investment returns, liquidity and the interaction between these elements. Insurance analysis differs between general insurance and life insurance. General insurance contracts commonly cover property, motor, liability, health and other risks over relatively short periods. Life insurance contracts may extend for decades and combine mortality, longevity, morbidity, lapse, expense and investment risks. Nevertheless, both sectors require disciplined measurement of expected cash flows, uncertainty and capital adequacy. This chapter develops a practical framework for analysing insurance financial statements and building models for underwriting results, claims, reserves, cash flows and solvency. Learning objectivesBy the end of this chapter, a student should be able to: - Explain the structure of an insurer’s financial statements.
- Calculate and interpret premium growth, loss, expense and combined ratios.
- Distinguish reported, incurred, paid and ultimate claims.
- Explain claims reserving and apply basic development methods.
- Analyse reinsurance and investment performance.
- Construct a simplified underwriting and solvency model.
- Evaluate model assumptions, governance, ethics and regulatory constraints.
1. The economics of insuranceThe insurer’s business model has four linked stages: - Risk selection and pricing: estimating expected claims, expenses, capital costs and profit requirements.
- Risk pooling: combining many exposures so that aggregate results are more predictable than individual outcomes.
- Claims settlement: paying valid claims and managing fraud, litigation and service quality.
- Asset and capital management: investing premium cash flows while maintaining sufficient liquidity and solvency.
An insurer’s profit may be represented conceptually as: Profit = Underwriting result + Net investment result + Other income − Other expenses − Tax The underwriting result relates to insurance operations. The investment result reflects income and gains on assets supporting both shareholder capital and insurance liabilities. The two components must be analysed separately because strong investment returns can conceal weak pricing or claims performance. Main financial statements| Statement | Main analytical purpose |
|---|
| Statement of financial position | Measures assets, liabilities, capital and liquidity | | Income statement | Reports premiums, claims, expenses, investment income and profit | | Statement of cash flows | Shows actual operating, investing and financing cash movements | | Notes and regulatory returns | Explain reserves, risk exposures, reinsurance and capital requirements |
Important insurance liabilities include the unearned premium reserve, reported claim reserve, incurred-but-not-reported reserve, life policy reserve, premium deficiency reserve and, where applicable, risk margins or contractual service margins under the relevant accounting framework. 2. Premium analysisPremium analysis begins with distinguishing the following measures: - Written premium: premium recorded when the policy is issued or renewed.
- Earned premium: the portion relating to insurance coverage provided during the reporting period.
- Gross premium: premium before reinsurance ceded.
- Net premium: premium retained after reinsurance ceded, subject to the accounting basis used.
- New business premium: premium from newly issued policies.
- Renewal premium: premium from continuing policies.
For a short-term policy, premium is generally earned over the period of risk. Under a simple time-apportionment assumption: Earned premiumt = Written premiumt + Opening unearned premiumt − Closing unearned premiumt The equation assumes that premiums are recognised consistently with the passage of risk. It is not appropriate where risk is highly seasonal or benefits are delivered in a materially non-linear pattern. Premium growth is calculated as: Premium growth rate = (Premiumt − Premiumt−1) ÷ Premiumt−1 × 100% Growth should be decomposed into price change, exposure growth, renewal retention, new business, product mix, currency movement and acquisitions. Rapid growth may indicate successful expansion, but it may also reflect underpricing, weak underwriting controls or aggressive sales incentives. 3. Claims and underwriting ratiosLoss ratioThe basic loss ratio is: Loss ratio = Incurred claims ÷ Earned premium × 100% “Incurred claims” normally includes claims paid during the period plus the change in outstanding claim reserves, and may include claims-handling expenses depending on the reporting convention. The analyst must identify the definition before comparing companies. A more explicit formulation is: Incurred claims = Claims paid + Closing outstanding claims − Opening outstanding claims This assumes that claim reserves are measured on a consistent basis and excludes or separately identifies recoveries, salvage and subrogation. Expense ratioExpense ratio = Underwriting expenses ÷ Earned premium × 100% Underwriting expenses may include acquisition commissions, administration, policy issuance, personnel and claims-handling costs. Expense classification differs across accounting systems, so published ratios require careful reconciliation. Combined ratioCombined ratio = Loss ratio + Expense ratio A combined ratio below 100% indicates an underwriting profit before investment income and certain other items. A ratio above 100% indicates an underwriting loss under the specified definitions. The ratio is not a complete measure of insurer profitability because it excludes investment income, taxes, capital costs and potentially catastrophe volatility. Retention and reinsurance ratiosRetention ratio = Net written premium ÷ Gross written premium × 100% A high retention ratio may improve premium scale and reduce reinsurance cost, but it also increases exposure to severe losses. A low retention ratio may reduce volatility while creating counterparty, pricing and dependence risks. Worked example: underwriting performanceAn insurer reports earned premium of CU 80 million, claims paid of CU 42 million, opening outstanding claims of CU 18 million and closing outstanding claims of CU 25 million. Underwriting expenses are CU 20 million. First, incurred claims are: CU 42m + CU 25m − CU 18m = CU 49m. Therefore: - Loss ratio = CU 49m ÷ CU 80m = 61.25%.
- Expense ratio = CU 20m ÷ CU 80m = 25.00%.
- Combined ratio = 61.25% + 25.00% = 86.25%.
The insurer generated an underwriting margin of approximately 13.75% of earned premium before other adjustments. However, the conclusion depends on reserve adequacy: if the closing claim reserve is understated, the reported loss ratio is artificially low. 4. Claims development and reservingWhy reserves are necessaryAt the reporting date, many claims have not been paid. Some have been reported but not settled, while others have occurred without being reported. The insurer must estimate the ultimate cost of these claims. Key categories are: - Case reserves: estimates attached to individually reported claims.
- IBNR reserves: claims incurred but not reported, including development on known claims where the terminology is used broadly.
- Outstanding claims reserve: case reserves plus IBNR, often after considering recoveries.
- Ultimate claims: the final expected cost after all future development.
- Loss-adjustment expense reserve: expected costs of investigating and settling claims.
A reserve is an estimate, not cash held in a separate account. The insurer invests assets to support the liability, but assets remain part of the general investment portfolio unless legally segregated. Claims trianglesA claims triangle arranges accident-year or underwriting-year claims by development period. For example, cumulative paid claims may appear as follows: | Accident year | 12 months | 24 months | 36 months |
|---|
| 2022 | 40 | 58 | 63 | | 2023 | 44 | 66 | — | | 2024 | 51 | — | — |
Values may be in CU millions. The diagonal incompleteness reflects that recent years have not fully developed. Chain-ladder methodThe basic chain-ladder method assumes that historical development patterns provide a reasonable guide to future development. A cumulative development factor from age j to age j+1 is: fj = Σ Ci,j+1 ÷ Σ Ci,j where Cᵢ,ⱼ is cumulative claims for accident year i at development age j, and the summation includes selected years with observations at both ages. The ultimate claim estimate is: Ultimatei = Ci,j × LDFj where LDFⱼ is the selected age-to-ultimate loss development factor. The reserve is: Reservei = Ultimatei − Ci,j The method assumes reasonably stable claims processes, consistent case reserving, comparable exposure and no major changes in inflation, settlement speed, policy terms or legal environment. These assumptions often fail during inflationary shocks, catastrophes, rapid growth or claims-management changes. Bornhuetter–Ferguson approachThe Bornhuetter–Ferguson method combines observed claims with an a priori expected loss estimate. In simplified form: Ultimate = Reported-to-date claims + Expected ultimate claims × (1 − Percentage reported) The expected ultimate claims are often calculated as expected loss ratio × earned premium. This method is less sensitive than chain-ladder to random early claims experience, making it useful for immature years, but it depends heavily on the quality of the expected loss ratio. Reserve diagnosticsAnalysts should investigate: - reserve releases and strengthening by accident year;
- changes in average claim severity and frequency;
- paid-to-incurred ratios;
- development beyond selected tail periods;
- discounting and inflation assumptions;
- catastrophe and latent claims;
- adequacy of case reserves;
- reserve risk margins and sensitivity ranges.
Persistent reserve releases may indicate conservative reserving, favourable development or prior under-reserving followed by corrective action. They should never be interpreted without examining the underlying years and assumptions. 5. Life insurance analysisLife insurers analyse a longer-duration cash-flow structure. Major drivers include mortality, longevity, morbidity, lapse, expenses, premium persistency, new business strain and investment yields. For a simplified life policy, expected present value is calculated by discounting probability-weighted cash flows: PV = Σt=1T E(CFt) ÷ (1 + i)t where E(CFₜ) is the expected net cash flow at time t, i is the annual effective discount rate, and T is the projection horizon. The formula assumes a deterministic discount rate and expected cash flows; real valuation models may use stochastic economic scenarios, risk-neutral methods or market-consistent discounting. A simplified expected mortality claim for a group is: Expected claimst = Number exposedt × Mortality ratet × Average benefitt The model must distinguish lives exposed for a full year from partial-year exposure. Mortality rates, benefit amounts, expenses and lapses may vary by age, sex where legally permissible, policy duration, underwriting class and product. Important life metrics include: - Persistency or lapse rate: proportion of policies terminating other than by maturity or insured event.
- New business margin: expected value created by new policies relative to premiums or required capital.
- Solvency coverage ratio: available eligible capital divided by required capital.
- Embedded value or appraisal value: a valuation framework that requires careful treatment of assumptions and is not identical to accounting equity.
6. Reinsurance analysisReinsurance transfers specified risks from a cedant to a reinsurer. Quota-share reinsurance transfers a fixed percentage of premiums and claims. Excess-of-loss reinsurance responds when losses exceed an attachment point, subject to a limit. Facultative reinsurance covers individually negotiated risks, while treaty reinsurance covers a portfolio under a standing agreement. A simplified net claims relationship is: Net claims = Gross claims − Reinsurance recoveries The actual economic result must also include ceded premiums, reinstatement premiums, commissions, collateral and the probability of reinsurer default. Reinsurance analysis should address: - Risk transfer: Is the arrangement genuinely transferring insurance risk?
- Cost: Does the protection justify the ceded premium and expenses?
- Protection quality: Is the reinsurer financially strong and appropriately diversified?
- Concentration: Could one reinsurer failure create material loss?
- Basis risk: Do contract exclusions or limits leave important exposures uncovered?
- Liquidity: How quickly can recoveries be collected after a catastrophe?
Reinsurance can reduce earnings volatility and capital requirements, but it does not eliminate the underlying risk. It may also create collateral, contract-interpretation and counterparty risks. 7. Investment performance and asset–liability managementInsurers invest premium receipts until claims and benefits are paid. Investment analysis must therefore consider both return and liability compatibility. A simple portfolio return is: Investment return = (Investment income + Realised gains + Unrealised gains − Investment expenses) ÷ Average invested assets The numerator and denominator must be measured on a consistent accounting basis. For example, a total return measure should not be compared directly with an income-only measure. Key considerations include: - duration matching between assets and liabilities;
- credit quality and default risk;
- liquidity for expected claims;
- currency matching;
- interest-rate sensitivity;
- equity and property volatility;
- concentration limits;
- inflation protection;
- regulatory admissibility of assets.
A high investment yield may result from taking excessive credit or duration risk. Conversely, a low yield may be appropriate for a conservative portfolio supporting short-tail claims. Investment results should be evaluated after considering capital requirements and liability risk, not in isolation. 8. Building an insurance financial modelA practical model should be modular and auditable. A suitable architecture is: - Input module: exposure, premium, pricing, claims, expense, reinsurance, investment and capital assumptions.
- Underwriting module: written and earned premium, claims, expenses and ratios.
- Reserve module: claim triangles, development factors and ultimate losses.
- Investment module: asset balances, income, market-value changes and reinvestment.
- Cash-flow module: premium receipts, claim payments, expenses, reinsurance and investment flows.
- Capital module: eligible assets, liabilities, required capital and stress tests.
- Output module: dashboards, ratios, scenarios and exception reports.
Simplified annual projectionSuppose an insurer begins with CU 100 million of cash and investments. During the year it receives CU 120 million of premiums, pays CU 75 million of claims and CU 25 million of expenses, receives CU 8 million of reinsurance recoveries, earns CU 6 million of investment income and pays CU 2 million of tax and other costs. Ending liquid resources are: CU 100m + CU 120m − CU 75m − CU 25m + CU 8m + CU 6m − CU 2m = CU 132 million. This is a cash-flow result, not necessarily accounting profit. Premiums may include amounts not yet earned, claim payments may relate to prior years and investment income may differ from total return. Solvency modellingA simplified solvency coverage ratio is: Solvency coverage ratio = Eligible own funds ÷ Solvency capital requirement × 100% For illustration, if eligible own funds are CU 180 million and the capital requirement is CU 120 million: CU 180m ÷ CU 120m × 100% = 150%. The precise regulatory calculation depends on jurisdiction and may include market, underwriting, credit, operational and diversification components. A ratio above a regulatory minimum does not guarantee safety: stress losses may be underestimated, assets may be illiquid, or liabilities may be exposed to model error. Useful scenarios include: | Scenario | Typical shock | Main question |
|---|
| Claims inflation | Higher severity and reserve development | Can reserves and capital absorb the increase? | | Catastrophe | Large correlated loss | Is reinsurance sufficient and collectible? | | Interest-rate fall | Lower reinvestment yield and possible liability increase | Is duration risk controlled? | | Equity decline | Market-value reduction | Does capital remain above requirements? | | Lapse shock | Higher policy termination | Is liquidity and profitability affected? | | Reinsurer default | Recovery impairment | Can the insurer pay claims without recoveries? |
Scenario analysis changes selected assumptions one at a time or in coherent combinations. Stress testing should not be confused with forecasting: its purpose is to evaluate resilience under adverse conditions. 9. Illustrative case: Northbridge General InsuranceThis is a fictional illustrative case. BackgroundNorthbridge writes motor and commercial property insurance. Its earned premium increased from CU 200 million to CU 250 million. Management reports that growth resulted from a 12% increase in exposure, a 5% average price increase and a new distribution partnership. ProblemThe reported combined ratio improved from 101% to 96%, but the board is concerned that reserves may be optimistic and that catastrophe reinsurance is expensive. AnalysisThe 25% premium growth is materially higher than the stated exposure and price effects considered separately. The difference may reflect product mix, timing or aggressive recognition. The analyst should reconcile written and earned premium and examine renewal retention. A claims review shows that current-year reported claims are low, but the claims triangle indicates that recent accident years develop more slowly than historical years. Applying an unadjusted chain-ladder method may therefore understate ultimate claims. Management should compare chain-ladder, Bornhuetter–Ferguson and independent actuarial selections. The reinsurance programme has a low attachment point, reducing catastrophe volatility but increasing ceded premium. The programme is economically valuable if the company’s capital charge for catastrophe risk exceeds the cost of protection. However, the reinsurer concentration is high, so credit quality and collateral provisions require review. Investment income increased because the company extended asset duration. This improves yield but creates greater market-value sensitivity if interest rates rise. Motor claims are paid relatively quickly, whereas commercial property claims may develop over several years; the asset portfolio should reflect these different liability profiles. Approach and outcomeNorthbridge builds a three-year model with separate motor and property segments. It tests: - a 10% increase in claims severity;
- a two-year delay in property settlement;
- a 20% reduction in reinsurance recoveries;
- a 200-basis-point interest-rate increase;
- a combined catastrophe and reinsurer-default scenario.
The base-case combined ratio remains 96%, but the adverse combined scenario rises to 112%. Solvency coverage falls from 155% to 118%, still above the internal warning level of 110% but close enough to require management action. The company increases catastrophe retention only modestly, diversifies reinsurers, strengthens property reserves and shortens the duration of part of its asset portfolio. Learning points- Reported ratios are not reliable without reserve-development analysis.
- Growth must be decomposed into exposure, price and mix.
- Reinsurance decisions involve both expected cost and capital protection.
- Asset strategy must reflect claim timing and liability duration.
- Scenario analysis reveals vulnerabilities hidden by base-case profitability.
10. Governance, ethics and limitationsInsurance models influence pricing, claims decisions, capital allocation and customer outcomes. Governance should include documented assumptions, version control, independent validation, access controls, audit trails and clear model ownership. Ethical risks include unfair discrimination in pricing, opaque use of personal data, claims delay, selective reserve strengthening, misleading performance measures and incentives that reward premium growth without adequate risk control. Predictive models should be tested for discrimination, stability, explainability and data leakage. Human review remains important when automated decisions affect vulnerable policyholders. Important limitations include: - parameter uncertainty, especially for long-tail claims;
- structural breaks caused by inflation, law or climate change;
- dependence on historical data that may not represent future risks;
- correlation between risks during crises;
- accounting differences between insurers;
- false precision created by complex spreadsheets or algorithms;
- model risk from coding errors and inappropriate aggregation.
Best practice is to use independent challenge, sensitivity analysis, back-testing, stress testing and documented expert judgment. Climate risk, cyber risk, pandemic risk and social inflation increasingly require forward-looking analysis rather than simple extrapolation. Contemporary developmentsModern insurers increasingly use cloud-based data platforms, automated claims triage, telematics, satellite data, machine learning and real-time dashboards. These technologies can improve speed and segmentation, but they also increase cyber, privacy, third-party and explainability risks. Artificial intelligence should support—not replace—actuarial oversight, underwriting accountability and fair claims treatment. Regulatory reporting is also becoming more risk-sensitive and market-consistent in many jurisdictions. Students should therefore distinguish accounting profit, economic value, regulatory capital and management performance measures. No single metric captures the complete financial condition of an insurer. Chapter SummaryInsurance financial analysis integrates underwriting, reserving, reinsurance, investments, cash flows and capital. Premium growth must be decomposed into exposure, price, retention, product mix and timing. The loss ratio measures claims relative to earned premium; the expense ratio measures underwriting cost; and the combined ratio evaluates underwriting profitability before investment income. Claims reserving requires distinguishing paid, reported, incurred, IBNR and ultimate claims. Chain-ladder and Bornhuetter–Ferguson methods are useful but depend on assumptions about development patterns and expected loss ratios. Reinsurance reduces volatility but introduces cost, counterparty and basis risks. Investment performance must be assessed alongside liquidity, duration, credit quality and liability characteristics. A sound financial model is modular, transparent and scenario-based. Solvency analysis compares eligible own funds with required capital, while stress testing examines resilience to claims inflation, catastrophes, market movements, lapse changes and reinsurer default. Ethical governance, independent validation and careful interpretation are essential because model outputs affect both institutional stability and policyholder welfare. Key TermsPremium: Amount charged for insurance coverage. Written premium: Premium recorded when a policy is issued or renewed. Earned premium: Premium relating to coverage provided during a period. Incurred claims: Claims paid plus the change in outstanding claim reserves. Loss ratio: Incurred claims divided by earned premium. Expense ratio: Underwriting expenses divided by earned premium. Combined ratio: Loss ratio plus expense ratio. Case reserve: Estimate for an individual reported claim. IBNR: Claims incurred but not reported, including relevant late development. Ultimate claims: Final expected cost of claims after all development. Claims triangle: Claims data arranged by origin year and development period. Chain-ladder: Reserving method based on historical development factors. Bornhuetter–Ferguson: Reserving method combining observed claims with expected losses. Reinsurance: Transfer of insurance risk to another insurer. Retention: Portion of risk or premium retained by the primary insurer. Solvency coverage ratio: Eligible own funds divided by required capital. Asset–liability management: Coordinating asset characteristics with liability cash flows. Stress test: Evaluation of financial resilience under adverse assumptions. Model risk: Risk of loss caused by model error, misuse or inappropriate assumptions. Review QuestionsShort Answer Questions- Why can an insurer report an underwriting profit while having weak overall profitability?
- Distinguish written premium from earned premium.
- What does a combined ratio above 100% indicate?
- Explain the purpose of an IBNR reserve.
- Identify two risks created by dependence on reinsurance.
Long Answer Questions- Explain how premium growth should be decomposed and analysed by an insurance company.
- Compare the chain-ladder and Bornhuetter–Ferguson reserving methods, including their assumptions and limitations.
- Discuss why investment performance must be analysed together with insurance liabilities.
- Explain the construction and governance of a modular insurance financial model.
- Discuss ethical and regulatory risks associated with artificial intelligence in insurance.
Analytical / Application Questions- An insurer has earned premium of CU 150 million, incurred claims of CU ninety million and underwriting expenses of CU 45 million. Calculate the loss ratio, expense ratio and combined ratio, and interpret the result.
- A general insurer’s premium grows by 30%, but exposure grows by only 10% and average price by 8%. Identify possible explanations and propose additional data required for analysis.
- An insurer’s solvency coverage ratio is 145% in the base case and 112% after a catastrophe and reinsurer-default stress. Recommend actions management should consider and explain the limitations of the ratio.
Multiple Choice Questions with AnswersQuestion 1Which measure relates premium to the period of insurance coverage actually provided? A. Written premium B. Earned premium C. Ceded premium D. Gross premium Correct Answer: B. Earned premium Question 2An insurer has incurred claims of CU 60 million and earned premium of CU 100 million. What is its loss ratio? A. 40% B. 60% C. 100% D. 160% Correct Answer: B. 60% Question 3If the loss ratio is 65% and the expense ratio is 28%, what is the combined ratio? A. 37% B. 65% C. 93% D. 108% Correct Answer: C. 93% Question 4Which reserve relates primarily to claims that have occurred but have not yet been reported? A. Unearned premium reserve B. IBNR reserve C. Capital reserve D. Investment reserve Correct Answer: B. IBNR reserve Question 5What is a central assumption of the basic chain-ladder method? A. Future development patterns resemble selected historical patterns B. All claims are paid immediately C. Investment returns are always zero D. Premiums never change Correct Answer: A. Future development patterns resemble selected historical patterns Question 6An insurer has eligible own funds of CU 240 million and a capital requirement of CU 160 million. What is its solvency coverage ratio? A. 66.7% B. 100% C. 150% D. 400% Correct Answer: C. 150% Question 7Which reinsurance arrangement typically responds to losses above an attachment point, subject to a limit? A. Quota share B. Excess of loss C. Coinsurance only D. Premium financing Correct Answer: B. Excess of loss Question 8Why may a high investment yield be misleading as a measure of insurer performance? A. It may reflect excessive credit or duration risk B. It proves that reserves are adequate C. It excludes all asset risk by definition D. It guarantees strong liquidity Correct Answer: A. It may reflect excessive credit or duration risk Question 9Which item is most directly a cash-flow measure rather than an accounting profit measure? A. Ending liquid resources after receipts and payments B. Combined ratio C. Loss development factor D. Solvency capital requirement Correct Answer: A. Ending liquid resources after receipts and payments Question 10What is a principal purpose of stress testing an insurer? A. To guarantee the forecast outcome B. To eliminate all model assumptions C. To evaluate resilience under adverse conditions D. To replace regulatory reporting Correct Answer: C. To evaluate resilience under adverse conditions
|