Life, Health and Personal Lines Underwriting

Introduction and Context

Underwriting is the disciplined process by which an insurer evaluates risk, decides whether to accept it, determines the terms of acceptance, and allocates an appropriate premium. In life, health and personal lines insurance, underwriting is especially important because the insured risks are closely connected with human behaviour, medical condition, occupation, property use and financial circumstances. Poor underwriting may produce adverse selection, excessive claims, unfair discrimination, inadequate premiums or regulatory criticism.

The underwriter’s task is not simply to identify “good” and “bad” risks. It is to distinguish relevant differences in expected loss while applying the principles of fairness, consistency, legality, confidentiality and commercial sustainability. Underwriting decisions commonly result in one of five outcomes: standard acceptance, acceptance with a loading, acceptance subject to an exclusion or higher deductible, postponement pending further information, or declinature.

This chapter examines mortality, morbidity, health, motor, homeowners and personal accident underwriting. It also considers medical and lifestyle information, risk classification, exclusions, loadings, automated underwriting and governance.

Learning Objectives

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

1. Explain the purpose and principles of life, health and personal lines underwriting.
2. Distinguish mortality risk from morbidity, frequency, severity and behavioural risks.
3. Analyse medical, occupational, lifestyle, vehicle, property and liability factors.
4. Explain standard classifications, loadings, exclusions, deductibles, postponements and declinatures.
5. Apply basic expected-loss and premium reasoning to underwriting decisions.
6. Evaluate the advantages and limitations of automated and AI-supported underwriting.
7. Recommend fair, evidence-based and compliant underwriting actions in practical cases.

1. The Underwriting Function

Insurance depends on pooling exposures that are sufficiently similar for aggregate losses to become reasonably predictable. Underwriting protects the pool by controlling which risks enter it and on what terms. The fundamental relationship is:

Probability is the likelihood that an insured event will occur during the relevant period; severity is the monetary consequence if it occurs. The relationship is a simplified expected-value model. It assumes that the probability and financial consequence can be estimated and that the event definition is sufficiently clear. It does not by itself determine the final premium, because expenses, capital costs, reinsurance, commission, taxes, profit margin and uncertainty must also be considered.

For a portfolio with several mutually exclusive loss outcomes, the expected loss is:

where E(L) is expected loss, pᵢ is the probability of outcome i, and Lᵢ is the loss associated with outcome i. Probabilities are dimensionless and losses are expressed in currency, so expected loss is also expressed in currency per policy period.

A simplified risk premium may therefore be represented as:

The risk margin compensates for uncertainty, volatility and capital requirements. A commercial premium additionally includes acquisition and administration expenses, claims-handling expenses, reinsurance costs, taxes and profit allowance. Actual pricing models are more sophisticated and may use credibility theory, generalized linear models, survival analysis, catastrophe models and capital-based pricing.

Adverse Selection and Moral Hazard

Adverse selection occurs when individuals with higher-than-average risk are more likely to seek insurance or select higher benefits, while lower-risk individuals withdraw because they consider the premium excessive. Underwriting, waiting periods, deductibles, benefit limits and legally permitted risk classification help control this problem.

Moral hazard arises when insurance changes behaviour after coverage is obtained. A motor policyholder may drive less cautiously, or a health-insurance member may use services more intensively when the cost is largely transferred to the insurer. Moral hazard is controlled through deductibles, co-payments, exclusions, utilization review, fraud controls and communication of policy conditions.

2. A General Underwriting Framework

A professional underwriting process usually follows these stages:

1. Information collection: proposal forms, declarations, medical evidence, inspection reports, claims history, driving records, credit or financial information where lawful, and external databases.
2. Verification: checking identity, ownership, prior losses, licences, medical disclosures and inconsistencies.
3. Risk analysis: identifying hazards, estimating frequency and severity, and comparing the case with underwriting rules.
4. Classification: assigning the risk to a homogeneous class such as preferred, standard, substandard or unacceptable.
5. Terms selection: choosing premium adjustments, exclusions, deductibles, limits, waiting periods or special conditions.
6. Decision and documentation: recording the rationale, authority level, evidence and customer communication.
7. Monitoring: reviewing claims experience, persistency, complaints, appeals and portfolio results.

Underwriting should distinguish a hazard, which increases the likelihood or severity of loss, from a peril, which is the cause of loss. In life insurance, smoking may be a behavioural or physical hazard and death may be the insured event. In motor insurance, poor vehicle maintenance is a physical hazard and collision is a peril. A moral hazard involves dishonesty or intentional conduct; a morale hazard involves carelessness created by the existence of insurance.

3. Life Insurance Underwriting

3.1 Mortality Risk

Mortality underwriting estimates the likelihood and timing of death. The central demographic quantity is the one-year mortality rate for a life aged x:

where dₓ is the number of deaths between exact ages x and x + 1, and lₓ is the number of lives alive at exact age x at the beginning of the interval. Both quantities refer to the same population or life table. Thus, qₓ is a probability or rate for the specified interval, not a currency amount.

The corresponding survival probability is:

For a one-year term life benefit of sum insured S, a highly simplified expected death benefit is:

This ignores discounting, expenses, policy lapses, timing of death, underwriting selection and policy conditions. It is useful for illustrating the logic of mortality risk, not for producing a complete premium.

3.2 Evidence Used in Life Underwriting

Important factors include:

• age and sex, subject to applicable law and permitted rating practice;
• current and historical medical conditions;
• height, weight and body-mass indicators;
• smoking, alcohol and substance use;
• occupation, travel, residence and hazardous hobbies;
• family medical history;
• previous insurance applications, exclusions and claims;
• income, existing cover and financial need;
• policy term, benefit amount and product type.

The sum insured must be economically justified. Financial underwriting reduces overinsurance, speculative applications and possible incentives for intentional loss. It may involve income replacement, debt protection, education funding, business-continuity needs and succession planning.

3.3 Medical and Lifestyle Factors

Medical underwriting considers diagnosis, severity, duration, treatment, control, complications, recurrence and prognosis. The same condition may receive different decisions depending on whether it is stable, recently diagnosed, poorly controlled or associated with organ damage. For example, well-controlled hypertension may attract a modest loading, whereas uncontrolled hypertension combined with renal disease may require a larger loading, exclusion where appropriate, postponement or declinature.

Smoking generally increases mortality and may affect premium classification. The insurer must define smoking status consistently, specify the relevant look-back period and verify declarations through lawful evidence. Alcohol misuse, recreational drug use, dangerous sports, aviation, mountaineering and work at height may influence both probability and severity of death.

3.4 Life Risk Classifications

A loading increases the premium to reflect higher expected risk. An exclusion removes a specified cause or condition from coverage. A rating-up may treat the applicant as older than their actual age for pricing purposes. These approaches are not interchangeable: a loading may preserve broad coverage at a higher price, while an exclusion transfers a defined risk back to the applicant.

4. Health and Medical Expense Underwriting

Health underwriting concerns morbidity: the likelihood of illness, injury, treatment, disability or utilization of health services. Unlike mortality, morbidity may involve repeated events, variable duration and substantial behavioural and provider-driven influences.

Health risks may be evaluated through:

• age, sex and geographic location where legally permitted;
• pre-existing conditions and previous treatment;
• hospital admissions, chronic medication and surgery;
• occupational exposure and functional capacity;
• family history and preventive-care indicators;
• claims and utilization patterns;
• benefit design, provider network and waiting periods.

The underwriter must separate a diagnosed condition from an unverified suspicion and distinguish an exclusion of a pre-existing condition from an exclusion of all future illnesses. Clear wording is essential. A waiting period delays eligibility for certain benefits; a benefit limit caps the insurer’s liability; a deductible makes the insured bear an initial amount; a co-payment shares a stated proportion or amount of each eligible cost.

Frequency and Severity in Health Insurance

Health claims are often analysed using frequency and severity:

This expression is a portfolio approximation. Frequency means expected claims per policy or member during a period; severity means average cost per claim. It is most useful when the two quantities are defined consistently and their dependence is not ignored. For a single policyholder, a more complete model may require separate treatment of outpatient, inpatient, pharmacy and chronic-care claims.

Health underwriting must account for provider behaviour, medical inflation, technology, benefit design and anti-selection at renewal. Claims data can indicate utilization but should not be treated as a complete explanation of health status. Privacy, medical confidentiality, consent and data minimization are fundamental.

5. Motor Underwriting

Motor underwriting combines actuarial, behavioural, vehicle and environmental factors. Common rating variables include:

• driver age, experience and licence history;
• prior accidents, convictions and claims;
• vehicle make, model, age, value, repair cost and theft frequency;
• annual mileage and purpose of use;
• geographic location and garaging arrangements;
• security devices and telematics;
• cover selected, deductible and liability limit;
• named drivers and household driving patterns.

The underwriter evaluates both claim frequency and claim severity. A high-performance vehicle may have greater collision frequency, while an expensive vehicle may generate greater repair severity. Liability exposure may be particularly important because bodily injury claims can be large and long-tailed.

The no-claims discount rewards claim-free experience but must be designed carefully. A claim-free period may reflect genuine lower risk, but it can also be affected by underreporting, policy switching and protection features. Telematics may measure speed, braking, time of travel and mileage, but its use requires transparency, data security and safeguards against biased interpretation.

Important motor exclusions may concern unlicensed driving, intoxication, racing, unauthorized use, deliberate damage and use outside the declared purpose. Exclusions must comply with insurance law and should not be used deceptively to defeat reasonable expectations.

6. Homeowners and Personal Property Underwriting

Homeowners underwriting evaluates the building, contents, liability and surrounding environment. Relevant factors include:

• construction type, age and condition of the building;
• roof, electrical, plumbing and heating systems;
• occupancy, vacancy and tenancy arrangements;
• location, flood, storm, earthquake, wildfire and crime exposure;
• security, alarms, sprinklers and protective devices;
• rebuilding cost and contents valuation;
• prior claims and maintenance practices;
• business use, pools, trampolines, pets and other liability hazards.

A central distinction is between market value and replacement cost. Market value reflects what property may sell for; replacement cost estimates the cost of repairing or rebuilding with comparable materials, subject to policy conditions. Underinsurance occurs when the insured amount is below the value required by the policy’s valuation basis.

Some policies apply an average or coinsurance condition. In simplified form, where a policy requires insurance equal to a specified proportion of value:

This formula applies only where the contract contains such a proportional condition, the loss is within the clause’s scope, and the amount required and amount insured are measured on the same basis. For example, if a policyholder insures property for $240,000 when the required amount is $300,000 and suffers a covered loss of $60,000, the proportional payment is $48,000 before any deductible:

$60,000 × ($240,000 ÷ $300,000) = $48,000.

The formula is not universal. Some policies provide full replacement subject to specified conditions, while others use agreed value, first-loss limits or sublimits.

Personal lines also include personal liability, valuables, travel, household employees and personal accident. Valuables may require scheduled items, appraisals, security conditions and separate limits because their loss severity and fraud exposure differ from ordinary contents.

7. Personal Accident Underwriting

Personal accident insurance responds to accidental death, permanent disablement, temporary disablement or medical expenses, depending on the contract. Underwriting focuses on occupation, travel, sports, transport use, age, prior injuries and the definition of accident.

The distinction between accident and illness is crucial. A benefit may depend on an unforeseen, external and visible event, or on a specified injury resulting directly and independently from accidental means. Ambiguous definitions produce disputes. A policy may exclude self-inflicted injury, intoxication, criminal activity, war, professional sport or hazardous occupations, subject to local law and policy wording.

Occupation classification should consider physical demands, work environment, height, machinery, travel and protective controls. A manual worker, office employee and professional diver may have materially different accident exposure. Fair underwriting uses evidence relevant to the insured peril rather than broad assumptions about social groups.

8. Exclusions, Loadings and Decision Quality

An underwriter should select the least restrictive term that appropriately reflects the evidence. A loading is generally suitable when the elevated risk is measurable across the covered causes. An exclusion may be appropriate when one identifiable condition creates disproportionate exposure but other risks remain acceptable. A deductible is useful where small or frequent losses create administrative cost or moral hazard. A limit or sublimit controls severity. Postponement is appropriate when temporary circumstances or insufficient evidence prevent reliable classification.

The decision should be:

• evidence-based: supported by relevant and current information;
• consistent: comparable cases receive comparable treatment;
• proportionate: the response matches the additional risk;
• communicated: terms and reasons are explained clearly where required;
• reviewable: an appeal or reassessment pathway exists;
• documented: evidence, authority and rationale are recorded.

Underwriters must avoid prohibited discrimination. Some variables may be statistically associated with risk but still be restricted or unlawful. Legal requirements differ by jurisdiction, so insurers must follow applicable insurance, equality, consumer-protection, privacy, medical-information and anti-money-laundering rules.

9. Automated Underwriting and Artificial Intelligence

Automated underwriting uses rules engines, application programming interfaces, predictive models and external data to produce immediate decisions or route cases for human review. A simple workflow is: data capture → validation → eligibility rules → risk score → decision or referral → audit record.

Automation offers speed, consistency, lower processing cost and improved customer experience. It can also support triage: straightforward standard risks are processed automatically while complex or high-value cases are referred to specialists.

However, models may reproduce historical discrimination, use inaccurate proxy variables, fail on unusual cases, or become unreliable when population behaviour changes. Data quality, model drift, cyber risk, explainability and accountability are major concerns. A statistically predictive variable is not automatically an ethically or legally acceptable underwriting variable.

Good governance includes:

1. documented purpose and lawful data basis;
2. data minimization and consent where required;
3. validation on relevant and representative populations;
4. testing for disparate impact and unfair outcomes;
5. human oversight and meaningful referral mechanisms;
6. adverse-decision explanation and correction procedures;
7. access controls, retention limits and cybersecurity;
8. independent model validation and periodic monitoring.

The underwriter remains responsible for the decision even when software provides a recommendation. Automation should support professional judgment, not eliminate accountability.

10. Illustrative Case Study: Integrated Personal Risk Assessment

Background. A fictional applicant, Amina, aged 42, seeks life insurance of $500,000, comprehensive motor insurance and homeowners cover. She is a non-smoker with controlled hypertension, works as a civil engineer, drives 18,000 kilometres annually, owns a six-year-old vehicle and lives in a region exposed to seasonal flooding.

Problem. The insurer must determine whether the risks can be accepted on standard terms. The application reveals one prior motor claim, a partially finished basement and a requested building sum insured below an external rebuilding-cost estimate.

Analysis. Amina’s non-smoking status is favourable, while age and hypertension require medical assessment. The underwriter requests recent blood-pressure readings, medication details and relevant physician evidence rather than assuming that the diagnosis alone determines the result. Her occupation includes occasional site visits, but the evidence does not indicate unusually hazardous duties.

The motor claim is investigated for cause, driver responsibility, repair severity and recency. Annual mileage increases exposure relative to a low-mileage driver, but the vehicle’s safety features and secure parking may reduce severity or theft risk.

For homeowners cover, the flood exposure is a location hazard. The basement may increase water-damage severity. The building sum insured is compared with the insurer’s replacement-cost basis, not market value. The applicant is advised that inadequate insurance may trigger a proportional settlement condition where the policy contains one.

Approach and solution. The life application is accepted subject to standard medical evidence, with a possible modest loading only if the evidence shows materially elevated mortality. Motor cover is accepted with a premium reflecting driving history and mileage, together with the ordinary deductible. Homeowners cover is accepted subject to flood terms, an appropriate flood deductible or sublimit, confirmation of mitigation measures and correction of the building sum insured.

Outcome. The insurer accepts the overall relationship without declining the applicant merely because several moderate factors exist. Each term is linked to a specific hazard, documented and explained. The decision demonstrates that underwriting should assess the total risk while avoiding unsupported aggregation of unrelated concerns.

Learning points. First, a diagnosis is not a complete mortality assessment. Second, a prior claim must be interpreted rather than mechanically penalized. Third, location and valuation are central to property underwriting. Finally, differentiated terms can be fairer and more commercially sound than automatic acceptance or rejection.

11. Challenges, Limitations and Contemporary Developments

Underwriting information is incomplete, self-reported and sometimes outdated. Medical conditions evolve, property risks change, vehicles become connected, and climate-related perils may alter the historical relationship between location and loss. Catastrophe exposure can make local claims experience unstable, requiring reinsurance and portfolio-level modelling rather than individual underwriting alone.

Climate change increases the importance of flood, wildfire, storm and heat-related risks in personal lines. Connected devices and telematics provide more frequent data but raise surveillance and consent concerns. Genetic information, wearable health data and social-media data are especially sensitive and may be restricted or inappropriate even when technically predictive.

Underwriters should combine quantitative evidence with professional review, recognize uncertainty, and avoid false precision. A model output such as a score is not a causal explanation. Portfolio monitoring should examine loss ratio, claim frequency, claim severity, retention, complaints, referral rates, turnaround time and outcomes across relevant segments. A common insurance performance measure is:

Both numerator and denominator must be defined consistently for the same period and portfolio. The ratio may be expressed as a percentage. It can indicate underwriting performance, but it is affected by reserving, catastrophe timing, premium adequacy and accounting choices; it should not be interpreted in isolation.

Best-Practice Checklist

Before finalizing a decision, the underwriter should ask:

• Is the information relevant, sufficient and lawfully obtained?
• Have frequency, severity and uncertainty been considered?
• Is the classification consistent with underwriting guidelines?
• Is a loading, exclusion, limit or deductible more suitable than declinature?
• Are terms clear to the customer?
• Has potential unfair discrimination been tested?
• Is the decision supported by evidence and within delegated authority?
• Will claims, complaints and portfolio results be monitored?

Chapter Summary

Life, health and personal lines underwriting converts individual information into an insurance decision that protects both the customer and the risk pool. Life underwriting focuses primarily on mortality and financial need, while health underwriting examines morbidity, utilization and treatment costs. Motor underwriting combines driver, vehicle, usage and environmental variables. Homeowners underwriting considers construction, occupancy, location, valuation, catastrophe exposure and liability. Personal accident underwriting emphasizes occupation, activities and the contractual definition of accidental injury.

Expected loss provides a foundation for reasoning, but a complete premium also reflects uncertainty, expenses, capital, reinsurance and profit. Loadings, exclusions, deductibles, limits, waiting periods, postponements and declinatures are distinct tools and should be selected proportionately. Ethical and regulatory obligations require privacy, consistency, explainability and avoidance of prohibited discrimination. Automated underwriting can improve efficiency, but it requires data governance, fairness testing, human oversight and continuing validation. Effective underwriting is therefore both a technical and professional activity: it combines evidence, judgement, documentation and accountability.

Key Terms

Adverse selection: The tendency for higher-risk people to seek or retain more insurance than lower-risk people.

Automated underwriting: Use of rules, data and software to make or support underwriting decisions.

Deductible: The amount of a covered loss borne by the insured before the insurer pays.

Exclusion: A contractual provision removing a specified risk, cause or condition from coverage.

Frequency: The number or expected number of claims during a defined period.

Hazard: A circumstance that increases the probability or severity of loss.

Loading: An additional premium charged for an elevated risk.

Moral hazard: Increased risk arising from intentional or dishonest behaviour.

Morale hazard: Increased carelessness because insurance exists.

Morbidity: The incidence or likelihood of illness, injury, disability or health-service utilization.

Mortality: The incidence or likelihood of death in a defined population.

Postponement: Deferral of an underwriting decision until information or circumstances improve.

Premium: The amount paid for insurance coverage.

Replacement cost: The cost of repairing or rebuilding property on the policy’s specified valuation basis.

Severity: The size or financial impact of an individual loss.

Standard risk: A risk expected to perform close to the insurer’s average accepted risk.

Substandard risk: An accepted risk with materially higher expected loss than standard.

Underinsurance: Insuring property for less than the amount required under the policy’s valuation conditions.

Review Questions

5 Short Answer Questions

1. What is the difference between mortality and morbidity?
2. Define adverse selection and give one underwriting response.
3. Distinguish a loading from an exclusion.
4. Identify four factors used in motor underwriting.
5. Why is replacement cost relevant to homeowners insurance?

5 Long Answer Questions

1. Explain the stages of a professional underwriting process and the purpose of each stage.
2. Discuss the medical and lifestyle factors considered in life insurance underwriting.
3. Explain how frequency and severity affect health and motor insurance decisions.
4. Compare standard acceptance, postponement, substandard acceptance and declinature.
5. Evaluate the benefits and risks of automated underwriting and AI-supported decisions.

3 Analytical / Application Questions

1. An applicant has controlled diabetes, no recent hospitalization and a physically demanding occupation. Analyse the evidence an underwriter should request and recommend possible terms, explaining why.
2. A homeowner insures a building for $180,000 when the policy requires $240,000 and suffers a covered loss of $30,000. Assuming a proportional condition applies and no deductible, calculate the claim payment and explain the limitation of your calculation.
3. A motor insurer’s algorithm charges substantially higher premiums to applicants from one postal area. Design an investigation to determine whether the result reflects legitimate risk, a proxy for prohibited discrimination, poor data quality or model drift.

Multiple Choice Questions with Answers

Question 1

Which relationship best represents simplified expected loss?

A. Probability of loss × severity of loss

B. Premium ÷ probability of loss

C. Severity of loss − probability of loss

D. Number of policies × commission rate

Correct Answer: A. Probability of loss × severity of loss

Question 2

In the life-table expression qₓ = dₓ ÷ lₓ, what does dₓ represent?

A. Lives alive at exact age x

B. Deaths between exact ages x and x + 1

C. Annual premium at age x

D. Survival probability at age x

Correct Answer: B. Deaths between exact ages x and x + 1

Question 3

Which underwriting action removes a specified cause or condition from coverage?

A. Loading

B. Deductible

C. Exclusion

D. Rating-up

Correct Answer: C. Exclusion

Question 4

Which factor is most directly associated with motor claim frequency?

A. Driver accident history

B. Building roof material

C. Life-policy beneficiary

D. Hospital waiting period

Correct Answer: A. Driver accident history

Question 5

A property is insured for $240,000, the required amount is $300,000, and the covered loss is $60,000. Under a proportional condition, what is the payment before deductible?

A. $36,000

B. $48,000

C. $60,000

D. $75,000

Correct Answer: B. $48,000

Question 6

What is a principal concern with using historical claims data in an automated underwriting model?

A. It always eliminates adverse selection

B. It may reproduce historical bias

C. It cannot be stored electronically

D. It guarantees causal explanations

Correct Answer: B. It may reproduce historical bias

Question 7

Which characteristic most clearly describes a postponed application?

A. The risk is automatically accepted at standard terms

B. The risk is outside all possible underwriting appetite

C. A decision is deferred pending information or changed circumstances

D. The policy has no deductible

Correct Answer: C. A decision is deferred pending information or changed circumstances

Question 8

Which homeowners valuation concept estimates the cost to repair or rebuild rather than the selling price?

A. Market value

B. Replacement cost

C. Surrender value

D. Book value

Correct Answer: B. Replacement cost

Question 9

Which control is primarily designed to reduce the insured’s incentive to submit very small claims?

A. Deductible

B. Beneficiary designation

C. Medical questionnaire

D. Mortality table

Correct Answer: A. Deductible

Question 10

Which principle is most appropriate when selecting an underwriting term?

A. Use the most restrictive term in every uncertain case

B. Ignore evidence that complicates automation

C. Apply a proportionate response linked to the relevant hazard

D. Treat all applicants as identical regardless of risk evidence

Correct Answer: C. Apply a proportionate response linked to the relevant hazard