Underwriting Information and Risk Assessment Methods

Introduction and Context

Insurance underwriting is the disciplined process of deciding whether a risk should be accepted, on what terms, and at what price. The underwriter’s decision depends not only on the hazard itself but also on the quality, relevance, timeliness, and lawful use of information about the applicant, property, activity, and expected loss environment.

Underwriting information may come from proposal forms, inspections, financial statements, medical evidence, claims histories, geospatial databases, credit indicators, public records, industry databases, and specialist reports. Each source answers different questions. A proposal form may describe the applicant’s declared circumstances; an inspection may test whether those declarations are supported by physical evidence; financial statements may indicate solvency and incentives; and loss histories may reveal frequency, severity, and risk controls.

The central professional challenge is to transform imperfect information into a fair and defensible risk assessment. Underwriters must avoid two opposite errors:

• Underassessment, in which important hazards are missed and premiums or terms are inadequate.
• Overassessment, in which irrelevant, inaccurate, discriminatory, or excessive information produces unjustified refusal or restrictive terms.

Modern underwriting therefore combines technical analysis with professional judgement, governance, data protection, explainability, and continuous review.

Learning Objectives

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

1. Explain the purpose and limitations of major underwriting information sources.
2. Distinguish hazard, exposure, vulnerability, loss experience, and financial capacity.
3. Apply structured methods for assessing data quality, relevance, bias, and privacy.
4. Interpret inspections, financial statements, medical evidence, claims histories, geospatial data, and credit indicators.
5. Use basic frequency, severity, expected-loss, and risk-scoring concepts in underwriting.
6. Identify ethical, regulatory, and operational controls for defensible underwriting decisions.
7. Evaluate the use of external data and artificial intelligence in insurance underwriting.

1. The Information Logic of Underwriting

A risk assessment normally follows a sequence:

1. Identify the exposure: What person, property, liability, activity, vehicle, or portfolio is being insured?
2. Identify hazards: What conditions increase the probability or severity of loss?
3. Assess controls: What measures reduce the likelihood or impact of loss?
4. Estimate expected outcomes: How frequently might losses occur, and how severe might they be?
5. Assess uncertainty: How complete, reliable, and representative is the evidence?
6. Make and document a decision: Accept, decline, modify, refer, monitor, or price the risk.

A useful distinction is between a peril, a hazard, and an exposure. Fire is a peril; defective electrical wiring is a physical hazard; the insured building and its contents are the exposure. Poor maintenance, dishonesty, or reckless behaviour may be moral or behavioural hazards. Legal or economic conditions can create morale hazard, where the existence of insurance affects incentives.

Underwriters should also distinguish risk from uncertainty. Risk generally refers to situations in which possible outcomes and their approximate probabilities can be estimated. Uncertainty refers to limited knowledge about outcomes or probabilities. A new technology, incomplete medical evidence, or an unprecedented climate event may involve substantial uncertainty even when the underlying hazard is understood.

Basic loss measures

For a group of similar risks, the expected number of claims can be represented as:

where E[N] is the expected number of claims, n is the number of exposure units, and λ is the expected claim frequency per exposure unit. Frequency must be measured over a defined period, such as claims per vehicle-year or claims per policy-year.

If S denotes the cost of an individual claim, the expected aggregate loss is:

where E[L] is expected aggregate loss and E[S] is expected severity, assuming the frequency and severity relationship is modelled appropriately. The expression is commonly used as a practical frequency–severity approximation. It should not be treated as universally exact when claim frequency and severity are dependent or when policy limits, deductibles, inflation, and development are ignored.

For example, if a portfolio has 2,000 exposure units, an annual claim frequency of 0.04, and average claim severity of CU 8,000:

This is an expected loss estimate, not a prediction that exactly 80 claims will occur. Random variation, catastrophic events, reporting delays, fraud, inflation, and changes in exposure may cause actual losses to differ substantially.

A simplified technical premium may be expressed as:

This is a conceptual identity for underwriting analysis, not a complete regulatory or accounting formula. Taxes, commissions, reinsurance costs, investment income, capital requirements, and target profit may also be relevant.

2. Proposal Forms and Applicant Declarations

The proposal form is the initial structured representation of the risk. It normally collects information about:

• The proposer and intended insured parties.
• The subject matter and use of the property or activity.
• Previous insurance and claims.
• Existing controls, maintenance, security, and safety procedures.
• Financial, medical, occupational, or operational circumstances.
• Requested limits, deductibles, extensions, and policy period.

A well-designed proposal form uses clear questions, defined terms, logical sequencing, and proportionate information requirements. Questions should be relevant to underwriting decisions. Ambiguous questions create inconsistent answers and increase disputes about misrepresentation or non-disclosure.

Underwriters should assess:

• Completeness: Are all mandatory fields answered?
• Consistency: Do answers conflict with one another or with external evidence?
• Plausibility: Are the answers reasonable in context?
• Materiality: Could the information change acceptance, pricing, limits, or terms?
• Verifiability: Can the information be confirmed independently?
• Timeliness: Does it describe the current risk rather than a previous situation?

The principle of materiality is central. Information is material when it could reasonably influence the insurer’s decision or terms. Materiality should not be confused with curiosity: information may be interesting but irrelevant to the insured peril.

A declaration is not automatically reliable merely because it has been signed electronically or physically. The underwriter should consider the applicant’s ability to understand the questions, the possibility of translation problems, incentives to understate risk, and whether the form was completed by the applicant, an intermediary, or another person.

3. Inspections and Physical Risk Surveys

An inspection provides direct or specialist evidence about the risk. It may be conducted before acceptance, during renewal, after a material change, or following a loss. Property inspections may examine construction, occupancy, housekeeping, electrical systems, fire protection, security, machinery, storage, access, and neighbouring exposures.

A professional inspection should state:

• Scope and date of inspection.
• Areas inspected and areas not accessible.
• Methods and instruments used.
• Observed conditions, photographs, measurements, and documents reviewed.
• Identified hazards and their potential consequences.
• Existing controls and their operating condition.
• Recommended improvements and deadlines.
• The inspector’s qualifications and limitations.

An inspection report is evidence, not an absolute truth. A survey may miss concealed defects, may be affected by weather or temporary conditions, and may reflect the inspector’s professional judgement. Underwriters should distinguish observation from inference. “Unprotected combustible storage was observed within two metres of a heat source” is an observation. “The premises has a high fire risk” is an evaluative conclusion requiring supporting reasoning.

Risk improvements should be proportionate and measurable. “Improve security” is weak; “install monitored intrusion detection covering the rear loading area within 60 days” is more actionable. Conditions precedent, warranties, recommendations, and advisory comments must be clearly distinguished because their legal effects may differ by jurisdiction and policy wording.

4. Financial Statements and Financial Capacity

Financial information is important in commercial, credit, trade credit, surety, directors’ and officers’ liability, business interruption, and some life and health underwriting contexts. It can indicate the applicant’s ability to maintain assets, fund risk controls, absorb deductibles, and continue as a going concern.

Key documents include the statement of financial position, income statement, cash-flow statement, budgets, management accounts, audit reports, debt schedules, and ageing reports.

Important indicators

These ratios are interpretive tools rather than automatic decision rules. Their meaning depends on industry, seasonality, accounting policies, asset quality, and the reliability of the accounts. A high current ratio may reflect obsolete inventory; a low ratio may be normal in a business with rapid cash sales. A profitable company may still fail because cash is tied up in receivables.

Underwriters should examine whether financial statements are audited, the auditor’s opinion, related-party transactions, contingent liabilities, unusual revenue recognition, rapid changes in receivables, and large uninsured exposures. They should compare reported values with proposal answers, inspection findings, tax or regulatory filings where lawful, and prior periods.

Financial distress can create both direct and indirect underwriting concerns. It may reduce maintenance expenditure, increase incentives for fraud, or make the applicant unable to fund deductibles and risk improvements. However, financial difficulty should not be treated as proof of dishonesty. It is a factor requiring proportionate investigation.

5. Medical Evidence and Life or Health Underwriting

Medical evidence may include application declarations, attending physician statements, examination reports, laboratory results, prescription histories, occupational information, and medical records. Its use must be limited to lawful, relevant purposes and handled with heightened confidentiality because health information is sensitive personal data.

Underwriters should consider:

• Diagnosis and date of onset.
• Treatment, medication, and compliance.
• Prognosis and functional impact.
• Recurrence or progression risk.
• Occupational and lifestyle context.
• Whether evidence is current and clinically reliable.
• Whether the evidence supports the specific underwriting decision.

A single medical measurement rarely determines the whole risk. Context, trends, age, treatment, and relevant clinical guidelines matter. Underwriting should not rely on unsupported stereotypes, such as assuming that a diagnosis necessarily predicts incapacity or mortality in the same way for every individual.

Medical evidence should be obtained with informed consent, secure transmission, access controls, retention limits, and procedures for correction of inaccurate records. Decisions should be explainable at an appropriate level without disclosing confidential information improperly.

6. Loss Histories and Claims Experience

Loss history is often one of the most useful sources of underwriting information because it records actual experience. A loss run should ideally show policy periods, exposure basis, date of loss, cause, paid amount, outstanding reserve, recovery, expenses, status, and whether the claim is open or closed.

Underwriters should analyse both frequency and severity. One large loss may indicate a low-frequency catastrophe, while repeated small losses may reveal an operational control failure. Claims should be adjusted for:

• Changes in exposure size and business activity.
• Inflation and medical or repair-cost trends.
• Changes in deductibles and policy limits.
• Claims handling and reserve development.
• Acquisitions, disposals, or changes in location.
• Changes in risk controls.
• Exceptional or one-off events.

A simple historical frequency estimate is:

The exposure denominator must be appropriate. Claims per vehicle-year, claims per employee-year, and claims per CU 1 million of sales measure different phenomena. A frequency estimate based on only a few exposure units may be unstable. Underwriters should therefore avoid treating a small number of observations as a precise estimate of future risk.

Loss histories also require careful interpretation. A low number of reported claims may indicate excellent controls, high deductibles, under-reporting, or an immature reporting system. A high number may reflect improved reporting rather than worsening risk. Claims should be linked to root causes and corrective actions, not merely counted.

7. Geospatial Information and Location-Based Risk

Geospatial underwriting uses location, distance, terrain, land use, infrastructure, climate, crime, flood, wildfire, storm, and supply-chain information. It is particularly important for property, agriculture, motor, marine, catastrophe, and business interruption insurance.

Relevant geospatial variables may include:

• Geographic coordinates and elevation.
• Distance to rivers, coasts, fault lines, or industrial facilities.
• Historical flood, wildfire, wind, or earthquake zones.
• Building density and neighbouring exposures.
• Road access, emergency response distance, and fire-station availability.
• Soil, drainage, vegetation, and topography.
• Utility and telecommunications dependencies.

Geospatial analysis must distinguish location-level risk from individual-level assumptions. A postcode may correlate with income, ethnicity, or other protected characteristics. Using a location variable as a proxy for a protected characteristic can create indirect discrimination even when the protected characteristic is not explicitly collected.

Data quality concerns include inaccurate coordinates, outdated hazard maps, inconsistent geographic boundaries, different spatial resolutions, and models that do not reflect local mitigation. A catastrophe map is not a guarantee that a particular property will flood or escape flooding. It provides evidence about a hazard environment.

A sound geospatial workflow verifies the property address, identifies the data source and version, records spatial resolution, tests sensitivity to boundary changes, and allows expert review where the result is surprising or consequential.

8. Credit Indicators and External Data

Credit indicators may include payment history, insolvency records, commercial credit ratings, public filings, trade references, and credit-based scores. They can be relevant to credit insurance, surety, commercial underwriting, premium financing, and, where legally permitted, other lines of business.

Credit information should be assessed for:

• Accuracy and date of the record.
• Whether the score is predictive for the specific insurance outcome.
• The applicant’s opportunity to correct errors.
• The legal basis and notice requirements.
• Potential disparate impact on protected or vulnerable groups.
• Whether the information is being used beyond its original purpose.

External data should never be treated as automatically superior to applicant-supplied information. It may be incomplete, duplicated, stale, incorrectly matched, or generated from opaque methods. Identity matching is especially important: a similar name or address does not prove that a record belongs to the applicant.

The principle of purpose limitation requires that data collected for one legitimate purpose not be casually reused for an unrelated purpose. Data minimisation requires collecting and retaining only what is reasonably necessary. These principles support both privacy and better underwriting discipline.

9. Assessing Data Quality, Relevance, Bias, and Privacy

A practical information-governance framework evaluates four dimensions.

Data quality

Important dimensions include:

• Accuracy: Does the data correctly represent the subject?
• Completeness: Are important fields missing?
• Consistency: Do records agree across sources?
• Timeliness: Is the information current enough for the decision?
• Validity: Does it conform to defined formats and meanings?
• Uniqueness: Are duplicate records present?
• Traceability: Can the source and transformation history be established?

A data-quality exception should not simply be ignored. It should lead to verification, a conservative assumption, referral to a senior underwriter, or a decision not to use the variable.

Relevance and predictive value

A variable should have a rational connection to the insured peril, expected loss, claims cost, or ability to perform contractual obligations. Historical correlation alone is not enough. The relationship may be accidental, unstable, or caused by a third factor. A variable that predicts one product’s claims may have little relevance to another product.

Bias and fairness

Bias may arise from unrepresentative training data, historical discrimination, measurement error, proxy variables, selective reporting, or unequal access to data. A model can be statistically accurate overall while producing materially worse outcomes for a subgroup.

Useful monitoring questions include:

• Are error rates materially different across relevant groups?
• Does the variable act as a proxy for a protected characteristic?
• Are applicants treated differently for reasons unrelated to insured risk?
• Is there a legitimate business justification for the variable?
• Can the decision be explained and challenged?

Fairness is not achieved by blindly removing sensitive variables. In some contexts, sensitive variables are needed to test disparate impact or improve risk adjustment. Their collection and use must be legally justified and tightly controlled.

Privacy and confidentiality

Core controls include lawful collection, informed notice, consent where required, access restrictions, encryption, secure sharing, retention schedules, breach response, vendor oversight, and deletion or anonymisation when information is no longer needed. Sensitive information should be visible only to personnel with a legitimate business need.

10. Structured Underwriting and Decision Governance

A defensible underwriting file should allow an independent reviewer to understand:

1. What information was available at the decision date.
2. Which information was considered material.
3. How contradictions were investigated.
4. Which rules, guidelines, or models were applied.
5. What judgement the underwriter exercised.
6. Why the final terms were proportionate.
7. What conditions, referrals, or monitoring requirements were imposed.

A common decision structure is:

• Accept: Evidence supports ordinary terms.
• Accept with modification: Apply a deductible, exclusion, limit, warranty, premium adjustment, or risk-improvement condition.
• Refer: Escalate because of complexity, uncertainty, authority limits, or potential adverse impact.
• Decline: Do not insure because the risk falls outside appetite, capacity, legality, or technical viability.
• Post-bind review: Verify outstanding information and monitor material changes.

Underwriters should separate facts, assumptions, interpretations, and decisions in their records. This improves auditability and reduces hindsight bias after a claim occurs.

11. Worked Application: Commercial Property Renewal

Background

An insurer is considering renewal of a small manufacturing plant. The applicant reports no major changes, two minor water-damage claims, and installation of improved fire alarms. The requested limit is CU 12 million.

Evidence collected

• Proposal form: production increased by 20% and operating hours expanded.
• Inspection: combustible packaging is stored near a heat-producing process; the alarm system is installed but monitoring is not yet connected.
• Financial statements: revenue increased, but receivables aged over 90 days also increased significantly.
• Loss history: two water losses of CU 25,000 and CU 32,000 in the previous year.
• Geospatial data: the plant is outside the high-risk flood zone but lies in an area with increasing surface-water exposure.
• External data: local emergency response time has lengthened because of road construction.

Analysis

The proposal form’s statement that there were “no major changes” is incomplete in context. Production and operating hours have increased, changing the exposure. The inspection identifies a serious fire-control gap: equipment exists, but monitoring is not operational. The two water claims are individually modest but may indicate a recurring maintenance problem. The geospatial evidence does not establish that flooding will occur, but it increases the importance of drainage and business-continuity planning. Financial information suggests possible cash-flow pressure, which may affect maintenance and the ability to fund a deductible.

Approach and solution

The underwriter should:

1. Verify the cause and corrective action for each water loss.
2. Require monitored alarm connection before or shortly after inception, subject to lawful and clearly drafted policy conditions.
3. Obtain updated values and business-interruption information reflecting increased production.
4. Inspect drainage, roof condition, and water shut-off procedures.
5. Consider a deductible or modified terms based on the confirmed risk, not merely on the applicant’s financial position.
6. Refer the case if the requested limit exceeds authority or if catastrophe accumulation is material.
7. Document the evidence, assumptions, and reasons for each term.

Illustrative expected-loss calculation

Suppose comparable exposure data suggest an annual frequency of 0.06 claims per plant-year and an average adjusted severity of CU 180,000. For one plant:

This CU 10,800 is not the final premium. It excludes expenses, catastrophe uncertainty, inflation, reinsurance, capital costs, commissions, taxes, and profit. Because the plant’s operating hours, storage practice, and emergency response have changed, the comparable frequency and severity may not be directly transferable. The underwriter should therefore treat the figure as a starting estimate and apply professional review.

Outcome

The renewal is offered with updated values, a higher water deductible, a documented fire-monitoring condition, and a follow-up inspection. The applicant is given a clear explanation of the required improvements and a review date.

Learning points

The case demonstrates that no single data source is decisive. Proposal forms reveal declared circumstances; inspections test physical conditions; financial statements provide context; loss history reveals experience; geospatial data identifies environmental exposure; and external information highlights infrastructure dependence. Effective underwriting integrates these sources while distinguishing evidence from assumption.

12. Technology, Automation, and Artificial Intelligence

Automated underwriting may use rules engines, predictive models, document extraction, image analysis, geospatial models, and machine learning. These tools can improve speed, consistency, triage, and detection of missing information. They do not eliminate the need for governance or professional judgement.

Key controls include:

• Documented purpose and approved data sources.
• Validation before deployment and periodic revalidation.
• Testing for accuracy, drift, disparate impact, and robustness.
• Human review for high-impact, unusual, or contested decisions.
• Version control and change management.
• Explainable decision reasons appropriate to the audience.
• Cybersecurity and vendor oversight.
• A process for correction, appeal, and reconsideration.

Automation bias occurs when staff accept a system output without sufficient challenge. Conversely, excessive manual override can undermine consistency and make the model impossible to evaluate. Override reasons should be recorded and analysed.

Emerging data sources include connected devices, satellite imagery, wearable devices, telematics, building sensors, and real-time weather information. These may improve risk measurement but raise concerns about surveillance, consent, unequal access, data security, and continuous repricing. Insurers should ask whether the additional information produces a proportionate underwriting benefit and whether customers understand its use.

13. Professional, Ethical, and Regulatory Responsibilities

Underwriting decisions should be lawful, fair, consistent, proportionate, and financially sound. Requirements vary by jurisdiction, but common responsibilities include:

• Treating customers fairly.
• Complying with insurance, contract, privacy, anti-discrimination, and consumer-protection law.
• Maintaining adequate records and audit trails.
• Protecting confidential and sensitive information.
• Avoiding misleading reasons for acceptance, pricing, or refusal.
• Managing conflicts of interest and intermediary incentives.
• Applying sanctions, anti-money-laundering, and fraud controls lawfully.
• Ensuring delegated authority is monitored.

There is an ethical distinction between risk-based differentiation and unfair discrimination. Insurance requires differentiation because risk characteristics affect expected loss. The professional obligation is to ensure that differentiation is relevant, evidence-based, transparent where required, and not an unjustified proxy for protected characteristics.

Challenges and Limitations

Underwriting information is never perfect. Major limitations include:

• Small samples and random loss variation.
• Incomplete or inconsistent historical records.
• Changes in technology, climate, law, behaviour, and exposure.
• Correlated risks and accumulation across policies.
• Model error and dependence on historical patterns.
• Fraud, deliberate misrepresentation, and strategic disclosure.
• Privacy restrictions that limit data availability.
• Conflicting evidence from apparently credible sources.
• Overconfidence in numerical scores.

Good underwriting does not attempt to eliminate uncertainty. It identifies uncertainty, estimates its significance, applies appropriate safeguards, and prices or manages it transparently.

Best-Practice Checklist

Before finalising a material underwriting decision, the underwriter should ask:

1. Is the exposure clearly defined?
2. Are the main perils, hazards, vulnerabilities, and controls identified?
3. Is each important data source current, accurate, relevant, and traceable?
4. Have contradictions been investigated?
5. Are assumptions clearly separated from verified facts?
6. Has potential bias or proxy discrimination been considered?
7. Is personal and sensitive information handled lawfully and securely?
8. Are terms proportionate to the evidence?
9. Can the decision be explained to an auditor, regulator, intermediary, and customer where appropriate?
10. Is post-bind monitoring required?

Chapter Summary

Underwriting is an evidence-based process for identifying exposures, evaluating hazards and controls, estimating expected loss, and deciding appropriate insurance terms. Proposal forms provide structured declarations, inspections provide physical evidence, financial statements reveal capacity and incentives, medical evidence supports life and health assessment, loss histories reveal frequency and severity, geospatial data identifies environmental and accumulation risks, and credit or external data may provide additional context.

The quality of an underwriting decision depends on more than the quantity of information. Data must be accurate, complete, timely, relevant, traceable, and lawfully obtained. Underwriters must distinguish correlation from causation, facts from assumptions, and legitimate risk differentiation from unfair discrimination. Privacy, confidentiality, explainability, human oversight, and auditability are essential, especially when automated systems and artificial intelligence are used.

A professional underwriter integrates multiple sources, investigates contradictions, recognises uncertainty, documents reasoning, and applies proportionate terms. Sound underwriting is therefore both a technical discipline and a governance responsibility.

Key Terms

Underwriting: The process of evaluating risk and determining insurance terms.

Exposure: The person, property, activity, liability, or portfolio subject to possible loss.

Peril: The event that directly causes loss, such as fire, theft, or flood.

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

Moral hazard: Increased risk arising from dishonesty or intentional misconduct.

Morale hazard: Carelessness or changed behaviour because insurance exists.

Proposal form: A structured document used to collect information from an applicant.

Materiality: The extent to which information could influence an underwriting decision.

Inspection: A physical or specialist examination of an insured risk.

Loss history: A record of previous claims and related financial information.

Frequency: The number of claims relative to an appropriate exposure base.

Severity: The cost of an individual claim or group of claims.

Expected loss: The statistically estimated average loss for a defined exposure and period.

Geospatial data: Information connected to geographic location or spatial relationships.

Credit indicator: Information about payment performance, financial reliability, or creditworthiness.

Data quality: The accuracy, completeness, consistency, timeliness, validity, and traceability of information.

Proxy variable: A variable that indirectly represents another characteristic, potentially creating unfair outcomes.

Data minimisation: Collecting and retaining only information reasonably necessary for a legitimate purpose.

Model drift: Declining model performance as underlying conditions or relationships change.

Audit trail: A record showing information, actions, decisions, and changes over time.

Review Questions

Short Answer Questions

1. What is the difference between a peril, a hazard, and an exposure?
2. Why should an underwriter distinguish observations from inferences in an inspection report?
3. Identify four dimensions of data quality.
4. What is the difference between claim frequency and claim severity?
5. Why is geospatial information not automatically proof of an individual property’s loss outcome?

Long Answer Questions

1. Explain how a proposal form should be designed and reviewed to support fair and reliable underwriting.
2. Discuss the uses and limitations of financial statements in commercial insurance underwriting.
3. Explain how an underwriter should analyse loss history, including frequency, severity, exposure changes, and claims development.
4. Discuss the ethical and privacy issues associated with medical evidence, credit indicators, and external data.
5. Evaluate the benefits and risks of artificial intelligence in underwriting and propose an appropriate governance framework.

Analytical / Application Questions

1. A warehouse has a low five-year claims frequency but stores increasingly large quantities of combustible goods. Explain why historical claims alone may be misleading and identify additional information required before renewal.
2. An insurer’s location-based model produces higher premiums in a postcode that is correlated with a protected characteristic. Design a review process to determine whether the model is justified, biased, or using an inappropriate proxy.
3. An applicant declares stable operations, but inspection, financial, and geospatial evidence indicate increased production, cash-flow pressure, and growing surface-water exposure. Construct a defensible underwriting response, including verification steps, possible terms, referrals, and monitoring requirements.

Multiple Choice Questions with Answers

Question 1

Which statement best defines an exposure in insurance underwriting?

A. The event that directly causes a loss

B. A condition that increases the chance of loss

C. The person, property, activity, or liability subject to possible loss

D. The amount added to premium for operating expenses

Correct Answer: C. The person, property, activity, or liability subject to possible loss

Question 2

A portfolio has 1,500 exposure units and an annual claim frequency of 0.02. What is the expected number of claims?

A. 30

B. 75

C. 750

D. 3,000

Correct Answer: A. 30

Question 3

Which inspection finding is an observation rather than an inference?

A. The premises has an unacceptable fire risk

B. The alarm system is inadequate

C. Combustible packaging was observed near a heat-producing process

D. The business will probably suffer a major fire

Correct Answer: C. Combustible packaging was observed near a heat-producing process

Question 4

Which financial indicator primarily assesses short-term liquidity?

A. Operating margin

B. Current ratio

C. Debt-to-equity ratio

D. Interest coverage

Correct Answer: B. Current ratio

Question 5

Why must loss frequency be calculated using an appropriate exposure denominator?

A. To eliminate all random variation

B. To ensure claims are always predicted exactly

C. To make comparisons meaningful across different exposure sizes

D. To replace the need for claims investigation

Correct Answer: C. To make comparisons meaningful across different exposure sizes

Question 6

Which is a major geospatial data-quality concern?

A. A policy has a deductible

B. A hazard map may be outdated or have unsuitable spatial resolution

C. A claim has a reserve

D. An applicant signed a proposal form

Correct Answer: B. A hazard map may be outdated or have unsuitable spatial resolution

Question 7

What is a proxy variable in underwriting?

A. A variable that directly measures claim severity in every case

B. A variable that indirectly represents another characteristic and may create unfair outcomes

C. A variable used only for accounting purposes

D. A variable that has been deleted from a database

Correct Answer: B. A variable that indirectly represents another characteristic and may create unfair outcomes

Question 8

Which principle supports collecting only information reasonably necessary for a legitimate underwriting purpose?

A. Data minimisation

B. Adverse selection

C. Indemnity

D. Subrogation

Correct Answer: A. Data minimisation

Question 9

If expected claim frequency is 0.05 and average severity is CU 20,000 for one exposure unit, what is the simplified expected loss?

A. CU 400

B. CU 1,000

C. CU 20,050

D. CU 400,000

Correct Answer: B. CU 1,000

Question 10

Which control is most appropriate when an automated underwriting model produces an unusual high-impact decision?

A. Automatically accept the model output without review

B. Delete the applicant’s record

C. Refer the decision for human review and document the reasons

D. Ignore all external data permanently

Correct Answer: C. Refer the decision for human review and document the reasons