Underwriting and Risk Selection

Introduction

Underwriting is the disciplined process by which an insurer evaluates an applicant, estimates the probability and financial consequences of loss, decides whether to accept the risk, and determines appropriate terms, conditions, limits and price. Risk selection is the related portfolio activity of choosing risks that are consistent with the insurer’s underwriting policy, capital position, reinsurance arrangements and strategic objectives.

Underwriting is central to both life and general insurance. In life insurance, the insurer assesses mortality, morbidity, longevity, occupation, financial justification and lifestyle. In general insurance, the assessment may concern property damage, liability, motor accidents, business interruption, marine loss, natural catastrophes or other perils. Although the evidence differs, the underlying question is similar: what is the expected loss, how uncertain is it, and can the risk be accepted on terms that preserve fairness and financial sustainability?

A sound underwriting decision is neither simply an acceptance nor a rejection. It may involve standard acceptance, an increased premium, an excess or deductible, reduced limits, exclusions, warranties, special conditions, postponement pending further evidence, or referral to a senior underwriter.

Learning Objectives

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

1. Explain the purpose, principles and stages of underwriting.
2. Distinguish life insurance risk factors from general insurance risk factors.
3. Evaluate proposal forms, inspections, medical evidence and financial information.
4. Apply classification, risk scoring and underwriting authority frameworks.
5. Explain policy conditions, exclusions, warranties, deductibles and risk-improvement requirements.
6. Analyse portfolio, accumulation, catastrophe and reinsurance considerations.
7. Recognise ethical, regulatory, data-protection and anti-discrimination requirements.
8. Make and justify an underwriting recommendation in a practical case.

1. The Nature and Purpose of Underwriting

Insurance operates through risk pooling. Many policyholders transfer uncertain financial consequences to an insurer in return for a premium. The insurer can sustain this arrangement only when it estimates risks systematically and prices or controls them adequately. Underwriting therefore performs five functions:

• Selection: deciding which risks fit the insurer’s appetite.
• Classification: grouping risks with broadly similar expected loss characteristics.
• Pricing support: supplying information for premium determination.
• Risk control: applying conditions, deductibles, limits and loss-prevention requirements.
• Portfolio management: ensuring that individual decisions remain compatible with aggregate exposure and capital.

Underwriting is different from claims assessment. Underwriting occurs before or at inception and concerns anticipated risk. Claims assessment occurs after an event and determines whether the event is covered and what amount is payable. However, underwriting quality strongly influences claims frequency, severity and disputes.

Fundamental underwriting principles

Several insurance principles guide underwriting:

Utmost good faith. Applicants and insurers are expected to disclose and represent material facts honestly, subject to the applicable law. A material fact is one that would influence a prudent insurer’s decision to accept, reject or alter the risk.

Insurable interest. The applicant must have a legally recognised relationship with the subject matter or life insured. This prevents insurance from becoming a speculative wager.

Indemnity. Most general insurance contracts aim to restore the insured, so far as money can do so, to the financial position immediately before the loss. Life insurance generally pays an agreed benefit rather than measuring actual financial loss.

Proximate cause. Claims analysis identifies the dominant effective cause of loss, subject to policy wording and applicable law. Underwriters must therefore understand how covered and excluded perils interact.

Subrogation and contribution. After paying an indemnity claim, an insurer may acquire rights against a responsible third party. Where more than one policy covers the same interest and loss, contribution rules may apply. These principles affect both policy design and risk selection.

2. The Underwriting Process

Underwriting is best understood as a sequence of connected decisions rather than a single checklist.

1. Submission and preliminary screening: The insurer identifies the applicant, requested cover, period, limits, location and apparent eligibility.
2. Completeness and authenticity review: Missing answers, inconsistent statements, unusual documentation or possible misrepresentation are investigated.
3. Hazard identification: The underwriter identifies physical, moral, morale, legal, financial, operational and environmental hazards.
4. Evidence collection: Depending on the class of business, this may include inspections, medical examinations, financial statements, valuations, driving records, engineering reports, loss histories and third-party data.
5. Risk analysis and classification: The risk is compared with underwriting guidelines, rating factors and relevant experience.
6. Decision and terms: The insurer accepts, declines, postpones or refers the risk and specifies premium, limits, deductibles, exclusions and conditions.
7. Documentation and communication: The rationale, evidence and authority level are recorded. Terms are communicated clearly before contract formation where required.
8. Monitoring and renewal: Claims, changes in exposure, inspection findings and emerging information are reviewed at renewal.

Figure: The underwriting cycle links individual risk assessment with continuing portfolio feedback.

3. Information Sources and Evidence Quality

Proposal forms

The proposal form establishes the facts on which underwriting is based. It normally records the applicant’s identity, requested cover, previous insurance, claims history, occupation or business activity, construction, location, health, lifestyle, financial interest and desired limits. Questions should be clear, relevant and proportionate. Ambiguous or overly broad questions weaken the evidential value of the answers and may create disputes.

Underwriters should test information for completeness, consistency and plausibility. For example, a commercial proposal stating that a warehouse has no combustible stock should be reconciled with its business description and inventory records. In life insurance, an applicant’s stated occupation, income, requested sum assured and lifestyle should be considered together.

Inspections and physical evidence

Property and engineering inspections can identify construction quality, electrical hazards, fire protection, housekeeping, security, flood exposure, maintenance standards and business continuity arrangements. A motor inspection may verify vehicle condition, use, modifications and security devices. Inspection reports should distinguish observed facts from the inspector’s opinions and should state the date and scope of inspection.

A survey does not eliminate uncertainty. Conditions may change, hidden defects may exist, and inspection quality varies. Material recommendations should therefore be assigned responsibility and completion dates, with follow-up where appropriate.

Medical and lifestyle evidence in life insurance

Life underwriting may use medical questionnaires, attending physician reports, medical examinations, laboratory results, prescription records where legally permitted, and information on smoking, alcohol, hazardous pursuits and occupational exposure. The underwriter’s task is not to diagnose the applicant but to estimate how disclosed characteristics affect mortality or morbidity relative to the relevant underwriting class.

Evidence must be collected with informed consent and handled confidentially. Medical information should be accessed only by authorised personnel and used for legitimate insurance purposes.

Financial evidence

Financial underwriting is important where the benefit could create over-insurance, moral hazard or an inappropriate economic incentive. Evidence may include income statements, tax records, audited accounts, debt schedules, asset valuations, loan agreements, budgets and business plans. In life insurance, the relationship between income, liabilities, dependants, existing cover and proposed benefit should be reasonable. In commercial insurance, financial statements can indicate solvency, continuity risk, fraud risk and the ability to maintain safety controls.

External data and verification

Credit information, motor records, sanctions screening, catastrophe maps, property databases, fraud indicators and industry statistics may support underwriting. Such data must be accurate, relevant, lawfully obtained and subject to correction procedures. Data should support professional judgement, not replace it blindly.

4. Hazard Analysis and Risk Classification

A hazard is a condition that increases the probability or severity of loss. A peril is the immediate cause of loss, such as fire, theft, illness or collision.

• Physical hazard: a tangible condition, such as defective wiring or poor construction.
• Moral hazard: intentional dishonesty or conduct designed to obtain an improper insurance benefit.
• Morale hazard: careless behaviour arising because the person is insured, such as neglecting ordinary precautions.
• Legal or liability hazard: a legal environment, contract or activity that increases potential liability.
• Financial hazard: weak financial condition, excessive leverage or inadequate maintenance capacity.
• Operational hazard: failures in processes, people, systems or controls.

Classification groups risks according to characteristics associated with expected loss. Examples include age and health class in life insurance; construction, occupancy, protection and exposure in property insurance; driver profile and vehicle use in motor insurance; and industry, turnover, payroll and claims history in liability insurance.

A classification system should be:

• actuarially relevant;
• sufficiently homogeneous within classes;
• practical to administer;
• legally permissible;
• understandable enough to support fair treatment; and
• periodically validated against actual experience.

Classification does not imply that every member of a class has identical risk. It is a method of making uncertainty manageable while avoiding arbitrary distinctions.

5. Risk Scoring and Quantitative Decision Support

Risk scoring assigns points or probabilities to relevant characteristics. A simple additive score may be written as:

Here, S is the score, xᵢ is the coded value of risk characteristic i, wᵢ is its assigned weight, and n is the number of characteristics. The score is dimensionless unless a documented scale gives it units. The model assumes that the characteristics are defined consistently and that the weights have been developed and validated using appropriate data.

A score is not itself a premium. It may trigger a decision band, such as standard acceptance, referral, additional evidence or decline. Underwriters must examine cases near a threshold and cases where the score conflicts with strong qualitative evidence.

For expected loss, a basic relationship is:

where E[L] is expected loss, pⱼ is the probability of loss outcome j, sⱼ is the associated monetary severity, and the sum covers mutually exclusive outcomes. If frequency and average severity are modelled separately, then under the assumptions of a stable exposure period and a suitable frequency-severity decomposition:

where E[N] is expected claim count and E[X] is expected severity per claim. This is an expectation, not a prediction of the exact loss for an individual policy.

A simplified indicated premium can be represented as:

where P is premium, E[L] expected claims cost, E[C] expected claim-related and administrative expenses, E[R] risk-transfer or reinsurance cost, and E[Π] target contribution for profit, capital and contingencies. The formula is conceptual: taxes, commissions, investment income, policy fees and regulatory pricing rules may require additional terms. It is valid only when all components are measured on the same exposure basis and for the same policy period.

Worked example: expected loss and underwriting terms

Suppose a small commercial property has the following estimated annual outcomes:

The expected annual loss is:

E[L] = (0.94 × 0) + (0.05 × 20,000) + (0.01 × 100,000) = 2,000

Assume expected expenses are $700, reinsurance and risk-transfer cost is $250, and the target contribution is $550. The indicated premium is therefore $3,500. This result does not mean that the insurer will collect exactly the loss amount or that the property is automatically acceptable. The underwriter must still examine accumulation, loss control, policy limits, deductibles, data quality and authority requirements.

6. Life Insurance Underwriting

Life underwriting commonly evaluates the following dimensions:

Age and mortality basis. Mortality generally varies with age, but age alone is not sufficient. The insurer uses an approved mortality basis and may apply medical or lifestyle adjustments.

Medical history. Existing conditions, treatment, prognosis, recurrence risk and functional impact are relevant. A disclosed condition should be assessed in context rather than treated as an automatic decline.

Lifestyle. Tobacco use, alcohol consumption, recreational drugs, travel and hazardous activities can materially affect risk.

Occupation and environment. Mining, aviation, offshore work, explosives, armed security and other occupations may involve elevated mortality or disability exposure.

Financial justification. The amount and purpose of cover should be consistent with income, liabilities, dependants, estate planning or business needs.

Policy duration and benefit structure. Term assurance, whole-life cover, endowment-related products and disability benefits have different risk profiles. Guaranteed benefits and investment-linked features may introduce additional product and conduct considerations.

Life decisions may include standard rates, table-rated extra premiums, exclusions, benefit limits, postponement or decline. The rationale should be recorded and communicated with sensitivity. Medical underwriting must not become a substitute for clinical care, and personal information must not be disclosed beyond legitimate purposes.

7. General Insurance Underwriting

General insurance underwriting is class-specific, but several recurring dimensions apply.

Property: construction, occupancy, protection, exposure, location, valuation, maintenance, fire load, security, catastrophe hazards and business continuity.

Motor: driver age and experience, vehicle type, use, annual mileage, territory, modifications, claims record, security and repair cost.

Liability: nature of operations, products or services, contractual assumptions, safety management, turnover, payroll, jurisdiction, past claims and potential severity.

Business interruption: dependence on key premises, suppliers, utilities, customers and technology; recovery time; documentation of gross profit; and adequacy of indemnity periods.

Agricultural, marine and specialty risks: weather, voyage, cargo, biological, geopolitical, valuation and accumulation factors may require specialist expertise.

For general insurance, the underwriter must distinguish between frequency and severity controls. A deductible may reduce small-claim frequency and administrative cost, while a limit or exclusion controls the insurer’s maximum exposure. A risk with low frequency but extreme severity may require catastrophe modelling, reinsurance or restricted capacity rather than merely a higher premium.

8. Policy Conditions and Risk Control

Policy terms convert underwriting analysis into contractual controls.

• Deductible or excess: the portion borne by the insured before the insurer pays, subject to wording. It can reduce small claims and encourage loss prevention.
• Limit of indemnity or sum insured: the maximum payable for a specified coverage, event or period.
• Exclusion: a stated circumstance, peril, property or loss type outside coverage.
• Warranty or condition: a contractual requirement concerning conduct, maintenance, disclosure or risk management. Legal effect varies by jurisdiction and wording.
• Endorsement: an amendment that adds, removes or changes coverage.
• Subjectivity: a condition that must be satisfied before binding or continuation, such as receiving an acceptable inspection report.
• Coinsurance or average condition: a mechanism that may reduce recovery where the insured amount is inadequate relative to the value or exposure, depending on the policy and law.

Conditions must be specific, proportionate, measurable and capable of compliance. An insurer should not impose a condition that it cannot reasonably verify or that is unrelated to the risk. Wording should identify the consequence of non-compliance and comply with consumer-protection rules.

9. Portfolio and Reinsurance Considerations

An individual risk may be acceptable but unsuitable for the portfolio. Portfolio underwriting considers:

• concentration by geography, peril, industry, age, occupation or distribution channel;
• catastrophe accumulation, such as many properties exposed to the same flood zone;
• correlation between risks, where one event can generate many claims;
• capacity, capital and solvency requirements;
• volatility and tail risk;
• reinsurance terms, retentions, exclusions and reporting obligations; and
• diversification across classes and territories.

A simple portfolio loss estimate is the sum of individual losses, but aggregate uncertainty depends on dependence among those losses. If claims are highly correlated, diversification benefits are smaller than under an independence assumption. Underwriters must therefore disclose material accumulations and avoid relying on average historical experience when exposures or climate conditions are changing.

Reinsurance may be proportional, sharing premiums and losses by an agreed percentage, or non-proportional, responding above a retention. Reinsurance does not remove the insurer’s responsibility to underwrite correctly: exclusions, warranties, data quality and reporting requirements remain important.

10. Illustrative Case: Commercial Warehouse Submission

Background

A fictional insurer receives a proposal for a three-year-old warehouse storing packaged consumer goods. The applicant requests property damage and business interruption cover. The location is near a river, and the requested limit is $2 million.

Problem

The proposal states that the premises have automatic sprinklers, 24-hour security and no prior losses. A preliminary data check identifies a flood map showing elevated exposure. The inspection finds that sprinklers cover the main storage area but not a recently added extension. The applicant’s declared values also appear lower than the replacement-cost estimate.

Analysis

The underwriter identifies flood as an external hazard, incomplete sprinkler protection as a physical hazard, possible underinsurance as a valuation problem, and inconsistent loss information as a verification issue. The proposed business interruption limit is also unsupported because the applicant has not supplied financial statements or an indemnity-period calculation.

The risk should not be accepted on the original terms. The underwriter requests:

1. a current replacement-cost valuation;
2. clarification of prior flood and water-damage incidents;
3. completion or isolation of the sprinkler deficiency;
4. a flood-risk mitigation plan and evidence of drainage maintenance;
5. financial statements supporting the business interruption calculation; and
6. confirmation of stock storage height and emergency procedures.

Approach and solution

Subject to satisfactory evidence, the insurer may offer cover with a flood sublimit or separate deductible, a higher all-perils deductible, an endorsement requiring completion of sprinkler works, a documented valuation basis and a business interruption limit based on verified financial information. If the flood accumulation at that location exceeds portfolio tolerance, the insurer may decline the flood component or use reinsurance capacity.

Outcome

The applicant provides evidence that prior water damage was minor and unrelated to river flooding, completes the sprinkler extension and supplies a professional valuation. The insurer accepts the property risk with specified conditions, a flood deductible, revised limits and a 12-month inspection follow-up. The decision is recorded and reported to the accumulation-control function.

Learning points

The case demonstrates that underwriting is iterative. A proposal form is a starting point, not conclusive evidence. A risk can become acceptable through risk improvement, better information and appropriate contractual controls. Portfolio exposure may still override an otherwise sound individual decision.

11. Ethics, Governance and Regulation

Underwriting decisions affect access to protection and can materially influence households and businesses. Ethical underwriting requires consistency, transparency, confidentiality and proportionality.

Important governance practices include:

• documented underwriting guidelines and delegated authority;
• segregation of duties for unusual or high-value decisions;
• audit trails showing evidence and reasoning;
• controls against fraud, bribery, money laundering and sanctions violations;
• complaint and appeal mechanisms;
• monitoring for unfair discrimination or disparate outcomes;
• secure handling of medical, financial and identity data; and
• periodic review of models, rules and exceptions.

Legal requirements differ across jurisdictions, but commonly address disclosure, privacy, consumer fairness, licensing, solvency, product governance, unfair contract terms and anti-discrimination. Factors such as health, disability, sex, race, religion or genetic information may be restricted or subject to special rules. Even where a factor is legally permitted, its use should have a defensible actuarial and social rationale.

An underwriter should avoid both careless acceptance and unjustified conservatism. A decline should be based on relevant evidence and policy, not stereotypes or convenience. Automated decisions should provide appropriate human oversight, explainability and a route for correction.

12. Technology, AI and Contemporary Developments

Insurance technology supports automated data capture, rules engines, geospatial analysis, telematics, remote inspections, electronic medical evidence, catastrophe models and claims feedback. Artificial intelligence can identify patterns, classify documents and assist triage. These tools may improve speed and consistency, but they introduce risks of biased training data, proxy discrimination, privacy intrusion, cyberattack, model drift and opaque decisions.

Best practice is to treat automated underwriting as decision support unless law and governance arrangements clearly permit otherwise. Insurers should validate models on relevant populations, monitor error rates, test for unfair impacts, control access to data, document model versions and maintain human escalation for exceptional cases. Climate change, cyber risk, supply-chain disruption, inflation, medical advances and changing mobility patterns also challenge historical assumptions. Underwriting should therefore combine historical data with scenario analysis, expert judgement and periodic recalibration.

Best-Practice Checklist for Underwriters

Before finalising a decision, the underwriter should ask:

1. Is the submission complete, internally consistent and verified where material?
2. What are the main perils, hazards, frequency drivers and severity drivers?
3. Is the risk within the insurer’s class definition and appetite?
4. Are limits, values, benefits and deductibles appropriate?
5. Are exclusions and conditions clear, relevant and enforceable?
6. Does the decision comply with law, authority and reinsurance requirements?
7. Could this risk create an unacceptable accumulation?
8. Is the evidence sufficient to justify the decision to a reviewer or customer?
9. Has personal and confidential information been handled properly?
10. What future information, inspection or renewal review is required?

Chapter Summary

Underwriting evaluates individual risks and converts evidence into decisions about acceptance, classification, price, limits and conditions. Risk selection extends this analysis to the insurer’s portfolio, capital, reinsurance and accumulation profile. Effective underwriting uses proposal information, inspections, medical evidence, financial documents, external data and professional judgement. Risk factors include physical, moral, morale, legal, financial, operational and environmental hazards. Classification and scoring can improve consistency, but scores are decision-support tools rather than substitutes for judgement.

Life underwriting focuses particularly on mortality, morbidity, medical history, lifestyle, occupation and financial justification. General insurance underwriting varies by class but commonly examines construction, occupancy, protection, exposure, valuation, liability, claims history and catastrophe potential. Deductibles, limits, exclusions, warranties and risk-improvement requirements translate analysis into contractual controls. Ethical and regulatory duties require fairness, privacy, accountability, model governance and careful treatment of vulnerable applicants. Finally, underwriting quality depends on continuous monitoring: claims and portfolio experience should inform revised rules, models and risk appetite.

Key Terms

Underwriting: Evaluation and acceptance of insurance risks on specified terms.

Risk selection: Choosing risks consistent with underwriting policy and portfolio capacity.

Peril: The immediate cause of a possible loss.

Hazard: A condition that increases loss probability or severity.

Physical hazard: A tangible condition that increases risk.

Moral hazard: Intentional dishonesty or conduct intended to obtain an improper benefit.

Morale hazard: Carelessness encouraged by the existence of insurance.

Proposal form: An application containing information used for underwriting.

Material fact: Information that could influence an insurer’s decision.

Risk classification: Grouping risks with similar expected characteristics.

Risk score: A numerical indicator derived from selected risk characteristics.

Expected loss: The probability-weighted average financial loss.

Deductible: The amount or portion borne by the insured before payment, subject to policy terms.

Limit of indemnity: The maximum amount payable under specified coverage.

Exclusion: A circumstance or loss not covered by the policy.

Warranty: A contractual requirement whose legal effect depends on wording and applicable law.

Accumulation: Concentration of exposures vulnerable to one event or related events.

Reinsurance: Insurance purchased by an insurer to transfer part of its risk.

Moral risk: The possibility that incentives or dishonesty affect loss behaviour.

Portfolio risk: Aggregate uncertainty arising from the insurer’s collection of policies.

Review Questions

Short Answer Questions

1. What is the difference between underwriting and claims assessment?
2. Define physical hazard, moral hazard and morale hazard.
3. Why is financial underwriting relevant in life insurance?
4. What purposes can a deductible serve?
5. Why can an individually acceptable risk be unsuitable for a portfolio?

Long Answer Questions

1. Explain the stages of the underwriting process and the evidence used at each stage.
2. Compare life insurance underwriting with property or liability underwriting.
3. Discuss how classification and risk scoring can improve underwriting consistency and where they may fail.
4. Explain how policy conditions, exclusions, limits and warranties control insurance risk.
5. Analyse the ethical, legal and governance requirements of modern underwriting.

Analytical / Application Questions

1. A factory has excellent fire protection but is located in a severe flood zone and has inadequate business interruption records. Recommend the additional evidence and terms an underwriter should consider.
2. An insurer’s motor portfolio shows acceptable average claims but rapidly increasing losses in one city. Analyse possible explanations and propose underwriting or portfolio actions.
3. A life applicant requests a large sum assured, reports modest income and provides incomplete medical information. Construct a reasoned underwriting response, including evidence, financial justification and possible outcomes.

Multiple Choice Questions with Answers

Question 1

Which activity is the primary purpose of underwriting?

A. Settling claims after a loss

B. Evaluating and selecting risks before or at policy inception

C. Investing all premium income

D. Auditing an insurer’s annual accounts

Correct Answer: B. Evaluating and selecting risks before or at policy inception

Question 2

Which is an example of a physical hazard in property insurance?

A. Intentional exaggeration of a claim

B. Carelessness because insurance exists

C. Defective electrical wiring

D. A disputed court interpretation

Correct Answer: C. Defective electrical wiring

Question 3

Using the figures in the chapter’s warehouse example, what is the expected annual loss?

A. $1,500

B. $2,000

C. $2,500

D. $3,500

Correct Answer: B. $2,000

Question 4

What does a deductible generally require?

A. The insurer pays every loss in full

B. The insured bears a specified portion before the insurer pays, subject to policy terms

C. The policy has no coverage limit

D. The insurer transfers all risk to a reinsurer

Correct Answer: B. The insured bears a specified portion before the insurer pays, subject to policy terms

Question 5

Which evidence is especially relevant to financial underwriting in life insurance?

A. Building sprinkler spacing only

B. Driving mileage only

C. Income, liabilities, dependants and existing cover

D. The colour of the applicant’s vehicle

Correct Answer: C. Income, liabilities, dependants and existing cover

Question 6

What is the main portfolio concern when many properties are exposed to the same flood?

A. Diversification benefit increases without limit

B. Accumulation and correlated losses

C. Medical misrepresentation

D. Reduced policy documentation

Correct Answer: B. Accumulation and correlated losses

Question 7

In the formula E[L] = E[N] × E[X], what does E[N] represent?

A. Expected claim count

B. Expected claim severity

C. Expected premium

D. Expected reinsurance commission

Correct Answer: A. Expected claim count

Question 8

Which response is most appropriate when a material proposal answer is incomplete?

A. Automatically accept without review

B. Ignore the omission if the premium is high

C. Seek clarification or additional evidence before deciding

D. Delete the proposal

Correct Answer: C. Seek clarification or additional evidence before deciding

Question 9

Why should automated underwriting models receive continuing validation?

A. Models never change after implementation

B. Validation can identify error, bias, drift and changing risk patterns

C. Validation eliminates the need for data protection

D. Validation guarantees every claim will be profitable

Correct Answer: B. Validation can identify error, bias, drift and changing risk patterns

Question 10

Which statement best describes a policy exclusion?

A. A stated circumstance or loss outside the coverage

B. A guaranteed claim payment

C. A method of calculating mortality only

D. A form of portfolio diversification

Correct Answer: A. A stated circumstance or loss outside the coverage