Insurance, Risk and the Underwriting–Claims Value Chain

Introduction

Insurance is a financial and social mechanism for managing uncertain losses. Individuals, firms, and public institutions face risks such as death, illness, fire, theft, liability, crop failure, cyberattack, and natural catastrophe. Insurance does not eliminate these events. Instead, it transfers specified financial consequences from an individual policyholder to an insurer in exchange for a premium.

The insurer can accept this transfer because it combines many exposure units into a portfolio. Although the loss of any one policyholder is uncertain, the aggregate results for a sufficiently large and reasonably homogeneous portfolio can become more predictable. This principle supports risk pooling, pricing, capital allocation, underwriting, and claims management.

The insurance operation is best understood as a value chain:

1. identifying and segmenting risks;
2. designing products and policy conditions;
3. selecting and pricing risks through underwriting;
4. issuing policies and collecting premiums;
5. monitoring exposure and preventing losses;
6. receiving, investigating, and settling claims; and
7. learning from claims experience to improve future underwriting.

Underwriting and claims are therefore not separate administrative functions. Underwriting determines which risks are accepted and on what terms; claims management determines whether reported events fall within those terms and what amount is payable. The quality of each function affects the other.

Learning Objectives

By the end of this chapter, a student should be able to:

• distinguish risk, uncertainty, peril, hazard, loss, and exposure;
• explain risk pooling and the law of large numbers;
• identify the principal conditions of insurability;
• explain indemnity and calculate a basic claim payment correctly;
• analyse adverse selection and moral hazard;
• describe the underwriting–claims value chain;
• evaluate ethical, regulatory, and technological issues in insurance operations.

1. Risk and Its Insurance Meaning

Risk is the possibility of an uncertain outcome, especially an outcome involving financial loss. In insurance, risk normally has both a probability dimension and a consequence dimension. A low-probability event with catastrophic consequences may be more important than a frequent event with a small cost.

Important terms include:

• Peril: the direct cause of loss, such as fire, collision, illness, or theft.
• Hazard: a condition that increases the probability or severity of loss.
• Physical hazard: a material condition, such as defective wiring.
• Moral hazard: increased loss potential arising from dishonest or irresponsible behaviour.
• Morale hazard: carelessness because the person feels protected by insurance.
• Exposure: a person, asset, activity, or liability subject to possible loss.
• Loss: an adverse financial consequence.
• Frequency: how often losses occur.
• Severity: the size of a loss when it occurs.

A useful conceptual relationship is that expected loss depends on both frequency and severity. In a simple discrete setting:

Here, pᵢ is the probability of outcome i, Lᵢ is its monetary loss, and the summation covers all relevant outcomes. If there is a 2% probability of a loss of $10,000 and no other loss, expected loss is $200. Expected loss is not a prediction that the policyholder will actually lose $200; it is a probability-weighted average across comparable cases.

Insurance generally covers pure risks, where the possible outcomes are loss or no loss. Speculative risks, such as business investment, may produce either gain or loss and are ordinarily not insured in the same way, although some associated pure risks may be covered.

2. Risk Pooling and the Law of Large Numbers

Risk pooling

Risk pooling combines many policyholders who face broadly similar classes of risk. Each contributes a premium to a common fund. Claims are paid from that fund, together with investment income and capital where necessary. The pool does not require every policyholder to suffer a loss at the same time.

Pooling works best when:

• exposures are sufficiently numerous;
• risks are reasonably homogeneous or can be segmented accurately;
• losses are not perfectly correlated;
• policy conditions are clear and enforceable; and
• the insurer has adequate capital and reinsurance.

Pooling is not the same as eliminating risk. A catastrophe affecting a whole region can produce correlated losses and weaken diversification. Reinsurance, catastrophe modelling, geographic diversification, exclusions, and capital buffers address this limitation.

Law of large numbers

Suppose independent, identically distributed loss amounts are represented by X₁, X₂, …, Xₙ, each with expected value μ and variance σ². The portfolio-average loss is:

Under the law of large numbers, X̄ₙ tends toward μ as n increases, subject to the assumptions of the model. If the risks are independent and have finite variance, the standard deviation of the average is:

Thus, increasing n reduces the relative fluctuation of the average loss. This does not mean that the insurer’s total claims become smaller. Total expected claims generally increase with the number of policies; rather, average claims become more predictable.

For independent Bernoulli losses, where each exposure has probability p of a claim, the expected number of claims is np and the variance is np(1 − p). The standard deviation is √[np(1 − p)]. Independence is an important condition. If losses are correlated, diversification is weaker and the formula cannot be applied without modification.

Numerical illustration

Assume each of 10,000 buildings has a 1% annual probability of a fire, and each fire produces a fixed insured loss of $50,000. The expected number of fires is:

Expected aggregate loss is:

The standard deviation of the number of fires, under an independence assumption, is:

The corresponding standard deviation of aggregate claims is approximately:

The actual result can be higher or lower. A regional wildfire could violate independence and produce a much larger deviation.

3. Insurability

A risk is insurable when an insurer can accept it, price it, and manage its consequences within legal, technical, and capital constraints. No universal list applies to every product, but the following conditions are central.

Fortuitous and uncertain loss

The event should be uncertain from the policyholder’s perspective and not deliberately caused. Insurance cannot function if a person can purchase cover after a loss has already occurred or intentionally create the loss.

Definable loss

The occurrence, timing, cause, and financial consequence should be capable of reasonable verification. Claims documentation, expert assessment, medical records, police reports, and financial records help establish these facts.

Measurable or estimable probability

The insurer needs credible information about frequency and severity. Historical data, engineering studies, mortality tables, medical evidence, climate models, and expert judgement may be used. New or rapidly changing risks are more difficult to price.

Sufficiently large and diversified portfolio

A single unusual loss may be difficult to predict, but a portfolio can be modelled more reliably. Geographic, industry, age, asset, and peril concentration must be monitored.

Economically viable premium

The premium must be acceptable to the customer and sufficient, over time, to support expected claims, expenses, commission, reinsurance, capital costs, and profit or surplus requirements. A simplified premium identity is:

This is a pricing framework rather than a universal legal formula. Actual pricing also reflects taxes, regulation, competition, policy conditions, and investment assumptions.

Limited catastrophic correlation or effective risk transfer

If a single event can cause losses to almost all policyholders simultaneously, ordinary pooling may be inadequate. Reinsurance, catastrophe bonds, deductibles, exclusions, limits, and government-backed arrangements may be required.

Legal and social acceptability

The subject matter and contract must comply with insurance law, consumer protection, sanctions rules, anti-money-laundering requirements, privacy law, and applicable discrimination restrictions.

4. Indemnity and the Claim-Payment Mechanism

The principle of indemnity aims to place the insured, after a covered loss, approximately in the same financial position as immediately before the loss. It prevents insurance from becoming a source of profit from damage. Indemnity is subject to policy wording, valuation basis, deductibles, limits, exclusions, underinsurance provisions, salvage, subrogation, and applicable law.

Basic formula and limit convention

To remove a common ambiguity, this chapter uses the following convention:

• L = covered loss measured under the policy’s valuation basis, in monetary units;
• U = policy limit, defined as the maximum covered loss amount considered before applying the deductible;
• D = deductible, in monetary units;
• P = insurer’s payment, excluding any interest or separately recoverable expense.

Under this convention:

The zero floor is essential: a deductible cannot create a negative claim payment. The policy limit is applied first to the covered loss, and the deductible is then subtracted. Some policies instead define the limit as the maximum insurer payment after the deductible. Such wording is a different convention and must be identified explicitly before calculation.

Worked example

A covered loss is $18,000, the policy limit under the pre-deductible convention is $15,000, and the deductible is $2,000.

If the covered loss is $1,200 and the deductible is $2,000, with a limit of $15,000:

The policyholder bears the loss because it does not exceed the deductible.

Limits of indemnity

Indemnity does not mean that every economic consequence is recoverable. Policies may use actual cash value, replacement cost, agreed value, market value, reinstatement value, or another basis. A replacement-cost policy may require repair or replacement before the full amount is payable. Liability policies may include defence costs within or outside the limit. Business interruption policies may use a defined indemnity period and documented loss of gross profit rather than physical repair cost.

Subrogation allows an insurer that has paid a claim to pursue a legally responsible third party, subject to applicable law. Contribution addresses situations where more than one policy covers the same interest and loss. These doctrines support indemnity by reducing double recovery.

5. Adverse Selection and Moral Hazard

Adverse selection

Adverse selection occurs when people with higher-than-average expected losses are more likely to seek or retain insurance, while lower-risk people withdraw or decline to purchase it. It arises from information asymmetry: applicants may know more about their risk than the insurer.

Examples include a person with a concealed medical condition seeking unusually high health cover or a property owner in a flood-prone location buying insurance only when a storm is imminent. If premiums are based on average risk but high-risk customers dominate, premiums rise. Higher-risk customers may then remain while lower-risk customers leave, producing a destabilising cycle.

Insurers respond through underwriting questions, inspections, waiting periods, exclusions, deductibles, risk classification, experience rating, policy limits, mandatory participation, and credible data. These controls must comply with law and should not amount to unfair discrimination.

Moral hazard

Moral hazard occurs when insurance changes behaviour in a way that increases the probability or severity of loss. A dishonest policyholder may deliberately cause a loss or exaggerate a claim. More commonly, reduced care may arise because the financial consequences are partly transferred to the insurer.

Controls include deductibles, coinsurance, policy conditions, warranties, audits, loss-prevention requirements, fraud analytics, claims investigation, salvage, subrogation, and sanctions for misrepresentation. Controls must be proportionate. Excessive friction can unfairly delay legitimate claims and damage trust.

Adverse selection concerns who enters or remains in the pool; moral hazard concerns behaviour after protection is obtained. The two problems can occur together but are conceptually distinct.

6. Underwriting in the Insurance Value Chain

Underwriting is the systematic assessment, classification, selection, pricing, and monitoring of risk. Its objectives are to accept desirable business, decline or modify unacceptable business, achieve an appropriate portfolio mix, and ensure that policy terms reflect exposure.

A typical underwriting workflow is:

1. Submission: receive the application, proposal, financial information, prior loss history, and supporting evidence.
2. Verification: check identity, insurable interest, ownership, exposure characteristics, and completeness.
3. Risk analysis: assess perils, hazards, frequency, severity, aggregation, and legal environment.
4. Classification: assign the risk to a rating segment or refer it for specialist review.
5. Terms: determine premium, deductible, limit, exclusions, conditions, warranties, and coverage extensions.
6. Authority decision: accept, decline, or refer according to delegated authority.
7. Documentation: issue a clear policy and retain an auditable record.
8. Monitoring: review changes, endorsements, renewals, loss experience, and accumulation.

Underwriting is not merely a sales approval function. Poor selection can produce adverse claims experience; overly restrictive selection can reduce access and weaken portfolio diversity. Underwriters therefore balance technical adequacy, customer needs, competition, regulation, and capital.

7. Claims Management

Claims management converts contractual promises into operational outcomes. A sound process is prompt, fair, consistent, evidence-based, and proportionate to the claim’s complexity.

A typical lifecycle includes:

• first notification of loss;
• coverage and policy-status verification;
• triage by urgency, complexity, fraud indicators, and severity;
• investigation of cause, circumstances, quantum, and liability;
• use of adjusters, engineers, medical experts, or legal advisers where appropriate;
• calculation of the payable amount;
• communication of decisions and reasons;
• payment, repair, replacement, settlement, or denial; and
• recovery through salvage, contribution, or subrogation.

Claims data is strategically important. It reveals emerging perils, inadequate wording, pricing errors, fraud patterns, repair inflation, service bottlenecks, and distribution problems. Claims feedback should flow back to product design and underwriting rather than being treated only as an expense.

Claims decisions must distinguish between a valid claim and an attractive narrative. At the same time, investigation should not presume dishonesty. The insurer should apply the policy wording, evidential standards, regulatory time limits, complaints procedures, and principles of fair treatment.

8. The Underwriting–Claims Feedback Loop

Underwriting and claims form a continuous loop. Underwriting creates the risk pool and defines the contract. Claims experience tests whether risk selection, pricing, wording, and controls were effective. The resulting information improves future underwriting.

For example, repeated water-damage claims may reveal that a building class was underpriced, that a policy exclusion is unclear, or that a prevention service would be economical. A sudden rise in suspicious claims may indicate organised fraud, but it may also reflect a change in reporting behaviour or a legitimate external event. Analysis must therefore combine quantitative indicators with investigation and professional judgement.

Useful portfolio measures include claim frequency, average severity, loss ratio, expense ratio, and combined ratio. In simplified form:

These ratios require consistent definitions and accounting bases. A combined ratio below 100% traditionally indicates underwriting profit before investment income, but interpretation depends on reserving quality, development of claims, reinsurance, and accounting rules.

9. Technology, Data, and Artificial Intelligence

Insurers increasingly use automated underwriting, telematics, satellite imagery, optical character recognition, predictive models, and artificial intelligence to estimate risk and triage claims. These tools can improve speed, consistency, fraud detection, and loss prevention.

However, technology creates governance obligations. Data may be incomplete, historically biased, unlawfully obtained, or unsuitable for a particular decision. Proxy variables can reproduce prohibited discrimination even when sensitive attributes are excluded. Automated claims decisions may be difficult for customers to understand or challenge.

Good practice includes documented purpose limitation, data quality testing, explainability proportionate to risk, human review for material adverse decisions, model validation, cybersecurity, privacy safeguards, monitoring for drift, and clear accountability. AI should support professional judgement rather than remove responsibility from the insurer.

10. Ethical, Legal, and Regulatory Considerations

Insurance decisions affect access to healthcare, housing, transport, credit-related protection, and business continuity. Ethical underwriting therefore requires more than technical profitability. Core expectations include good faith, fair treatment, transparency, confidentiality, competent investigation, timely communication, and consistent application of policy terms.

Important governance controls include segregation of duties, delegated authority limits, claim-reserving controls, audit trails, complaints handling, conflict-of-interest procedures, fraud controls, sanctions screening, and regulatory reporting. The exact legal rules vary by jurisdiction, but the professional principle is stable: decisions must be lawful, evidence-based, documented, and reviewable.

A denial is not automatically unfair, and payment is not automatically correct. The relevant question is whether the decision follows the contract, evidence, law, and fair claims-handling standards.

Applied Case: Illustrative Commercial Property Claim

Background

A small manufacturer insures its factory under a commercial property policy. The policy has a $1,000,000 limit, defined in this case as the maximum covered loss before the deductible, and a $25,000 deductible. The policy covers fire but excludes deliberate acts by the insured. It also requires reasonable fire-prevention measures.

Problem

A fire damages machinery and stock. The insured reports a total loss of $1,300,000. The insurer must determine coverage, the payable amount, and whether underwriting or claims controls require review.

Analysis

An engineer confirms that an electrical fault caused the fire. Inspection records show that required maintenance was performed. There is no reliable evidence of deliberate action. The covered loss, after applying the policy valuation rules and excluding non-covered items, is assessed at $920,000.

Using the stated limit convention:

The original reported amount does not determine payment; the policy valuation and verified evidence do. The insurer should document the cause assessment, valuation, deductible, communications, and any salvage or recovery rights.

Outcome and learning points

The claim is paid at $895,000, subject to the policy’s other conditions. The claim is then coded for electrical-cause analysis. If similar losses are emerging, underwriting may revise inspection requirements, pricing, risk engineering, or wording. The case demonstrates that claims management both enforces the contract and generates information for portfolio improvement.

Challenges and Limitations

Risk pooling is weakened by catastrophe correlation, climate change, inflation, cyber accumulation, and rapid technological change. Historical data may not represent future conditions. Pricing models can create opacity or unfair segmentation. Fraud controls can inconvenience honest customers. Underwriting restrictions can reduce availability, while insufficient restrictions can threaten solvency.

No single formula resolves these tensions. Professional insurance management combines actuarial analysis, legal interpretation, engineering evidence, customer communication, capital management, and ethical judgement.

Chapter Summary

Insurance transfers specified financial consequences of uncertain events to an insurer in exchange for a premium. Risk pooling makes aggregate outcomes more predictable, especially when exposures are numerous, reasonably homogeneous, and not strongly correlated. The law of large numbers explains why average outcomes tend toward expected values, but it does not eliminate catastrophe or model risk.

Insurable risks are generally fortuitous, definable, measurable, economically viable, legally acceptable, and suitable for portfolio diversification or risk transfer. Indemnity seeks to restore the insured financially rather than create a profit. Under the explicitly stated pre-deductible limit convention, the basic claim formula is P = max(0, min(L, U) − D). The zero floor prevents a negative payment.

Adverse selection results from information asymmetry before or at entry; moral hazard concerns behaviour after insurance is obtained. Underwriting selects, classifies, prices, and monitors risk. Claims management verifies coverage, investigates loss, calculates payment, and delivers the contractual promise. Claims data should feed back into underwriting, product design, pricing, prevention, and governance. Technology can improve performance, but privacy, fairness, explainability, cybersecurity, and human accountability remain essential.

Key Terms

Risk: The possibility of an uncertain adverse outcome.

Peril: The direct cause of a loss.

Hazard: A condition that increases loss probability or severity.

Exposure: A person, asset, activity, or liability subject to risk.

Frequency: The rate or number of losses occurring in a period.

Severity: The monetary size of a loss.

Risk pooling: Combining many exposures so aggregate outcomes become more predictable.

Law of large numbers: A statistical principle that sample averages tend toward expected values as observations increase, under stated conditions.

Insurability: The practical and legal suitability of a risk for insurance.

Indemnity: Restoration of the insured approximately to the pre-loss financial position.

Deductible: The portion of a covered loss borne by the insured.

Policy limit: The maximum covered loss or payment specified by the policy; its convention must be identified.

Adverse selection: Disproportionate participation by higher-risk individuals due to information asymmetry.

Moral hazard: Behaviour that increases loss likelihood or severity after insurance protection is obtained.

Underwriting: Risk selection, classification, pricing, and acceptance.

Claims management: The investigation, assessment, decision, and settlement of reported losses.

Subrogation: Recovery by an insurer from a responsible third party after paying a claim.

Loss ratio: Incurred claims divided by earned premium.

Combined ratio: Loss ratio plus expense ratio.

Risk pooling: The aggregation of exposures to share uncertain losses.

Review Questions

Short Answer Questions

1. Distinguish a peril from a hazard and give one example of each.
2. Why does the law of large numbers matter to insurers?
3. State two conditions that support insurability.
4. What is the difference between adverse selection and moral hazard?
5. What is the purpose of a deductible?

Long Answer Questions

1. Explain how risk pooling operates and discuss the conditions under which it becomes unreliable.
2. Explain indemnity, including the roles of limits, deductibles, valuation bases, subrogation, and contribution.
3. Describe the underwriting process from submission to renewal monitoring.
4. Discuss the claims lifecycle and explain why claims data is valuable to underwriting.
5. Evaluate the benefits and governance risks of using artificial intelligence in underwriting and claims.

Analytical / Application Questions

1. A policy uses the pre-deductible limit convention. A covered loss is $72,000, the limit is $60,000, and the deductible is $7,500. Calculate the payment and explain each step.
2. An insurer observes rising claim frequency after lowering deductibles. Analyse two possible behavioural explanations and recommend controls, while considering customer fairness.
3. A portfolio contains geographically concentrated properties exposed to one flood peril. Explain why ordinary pooling assumptions may fail and propose underwriting, reinsurance, and claims-management responses.

Multiple Choice Questions with Answers

Question 1

Which statement best describes risk pooling?

A. It guarantees that no policyholder will suffer a loss.

B. It combines many exposures so aggregate outcomes become more predictable.

C. It transfers all insurance risk to the government.

D. It removes the need for underwriting.

Correct Answer: B. It combines many exposures so aggregate outcomes become more predictable.

Question 2

Under the stated pre-deductible limit convention, which formula gives the basic claim payment?

A. P = min(L, U) + D

B. P = max(0, min(L, U) − D)

C. P = L − U − D without a floor

D. P = max(0, L + U − D)

Correct Answer: B. P = max(0, min(L, U) − D)

Question 3

A covered loss is $1,200, the limit is $10,000, and the deductible is $2,000. What is the payment under the chapter’s convention?

A. $0

B. $800

C. $1,200

D. $2,000

Correct Answer: A. $0

Question 4

What is a necessary assumption for the simple standard-deviation relationship σ/√n for an average portfolio loss?

A. All losses must be identical in amount.

B. The risks must be perfectly correlated.

C. The observations should satisfy suitable independence and finite-variance conditions.

D. The insurer must have no deductible.

Correct Answer: C. The observations should satisfy suitable independence and finite-variance conditions.

Question 5

Adverse selection primarily occurs because:

A. Insurers always pay claims late.

B. Applicants may possess more information about their risk than insurers.

C. Claims adjusters inspect losses after payment.

D. Reinsurance eliminates diversification.

Correct Answer: B. Applicants may possess more information about their risk than insurers.

Question 6

Which example most directly illustrates moral hazard?

A. A customer hides a known flood exposure before buying cover.

B. An insurer changes its logo.

C. An insured becomes less careful because insurance absorbs much of the loss cost.

D. An underwriter verifies a building inspection.

Correct Answer: C. An insured becomes less careful because insurance absorbs much of the loss cost.

Question 7

In the simplified premium framework, which item is normally included in gross premium?

A. Expected claims cost

B. Only the policyholder’s deductible

C. Only investment losses

D. The insured’s unrelated tax bill

Correct Answer: A. Expected claims cost

Question 8

Which activity is principally part of claims management?

A. Assessing reported loss circumstances and determining coverage and payment

B. Choosing the insurer’s corporate name

C. Eliminating all uncertainty from the economy

D. Guaranteeing that every submitted risk is accepted

Correct Answer: A. Assessing reported loss circumstances and determining coverage and payment

Question 9

A covered loss is $18,000, the pre-deductible policy limit is $15,000, and the deductible is $2,000. What is payable?

A. $13,000

B. $15,000

C. $16,000

D. $18,000

Correct Answer: A. $13,000

Question 10

Why can catastrophe risk weaken ordinary insurance pooling?

A. Catastrophe losses are always below deductibles.

B. The losses may be highly correlated across many insureds.

C. Catastrophes cannot be investigated.

D. Pooling works only for liability insurance.

Correct Answer: B. The losses may be highly correlated across many insureds.