Final Assessment — The Return Period

Before You Begin

20 multiple-choice questions covering all 5 modules
Passing score: 70% (14 out of 20 correct)
No time limit — take your time and think carefully
Questions cover: fundamentals, model anatomy, metrics, perils, and professional practice
Review your lesson notes before starting if needed

// Module 01 — Introduction

Fundamentals of Catastrophe Modelling

Question 1 of 20

Which of the following best explains why natural catastrophe losses do NOT follow the law of large numbers?

ACat events are too expensive for insurers to model accurately

BCatastrophe losses are spatially correlated — a single event generates many simultaneous claims across a portfolio

CHistorical data is available but ignored by the industry

DCatastrophes only affect uninsured populations

Question 2 of 20

Hurricane Andrew (1992) is considered the catalyst for widespread cat model adoption. Which outcome most directly drove this?

AIt caused more deaths than any previous hurricane

BNine insurers became insolvent, demonstrating that historical data was inadequate for managing catastrophe exposure

CThe U.S. government mandated the use of cat models by law

DIt was the first hurricane to be accurately predicted by computer models

Question 3 of 20

What is the 'protection gap' in the context of natural catastrophe risk?

AThe gap between a model's estimate and the actual observed loss

BThe difference between total economic losses from a catastrophe and the insured portion of those losses

CThe regulatory capital buffer required above a modelled PML

DThe period between a catastrophe event and the settlement of claims

Question 4 of 20

Two conditions must be met for a risk to be insurable. Which pair correctly identifies these conditions?

ALow frequency and high severity

CHistorical data availability and government backing

BAbility to quantify the risk and ability to set a viable premium

DReinsurance availability and regulatory approval

// Module 02 — Anatomy of a Cat Model

The Four Model Components

Question 5 of 20

What are the four primary modules of a catastrophe model, in the correct order of the loss estimation process?

AFinancial → Vulnerability → Inventory → Hazard

BHazard → Inventory/Exposure → Vulnerability → Financial

CInventory → Hazard → Financial → Vulnerability

DVulnerability → Hazard → Inventory → Financial

Question 6 of 20

A hurricane causes $80,000 of physical damage to a property worth $400,000. The policyholder has a 10% hurricane deductible. How much does the insurer pay?

A$80,000

B$40,000

C$72,000

D$0 — the loss is below the deductible

Question 7 of 20

What is geocoding, and why is its accuracy critical in catastrophe modelling?

AGeocoding is the classification of buildings into vulnerability classes; accuracy matters because different classes have different damage ratios

BGeocoding is the assignment of geographic coordinates to properties; accuracy matters because hazard intensity can vary enormously within a single ZIP code

CGeocoding is the encryption of sensitive exposure data; accuracy matters for regulatory compliance

DGeocoding is the mapping of policy terms to building types; accuracy matters for correct financial module application

Question 8 of 20

A vulnerability function (damage function) relates which two variables?

AInsurance premium and expected annual loss

BHazard intensity at a site and the expected damage ratio for a given building type

CEvent magnitude and the probability of occurrence

DReturn period and the exceedance probability

Question 9 of 20

The financial module applies policy conditions to ground-up losses. Which sequence correctly represents the output hierarchy?

ANet loss → Insured loss → Ground-up loss

BGround-up loss → Insured loss → Net loss (after reinsurance)

CInsured loss → Ground-up loss → Net loss

DNet loss → Ground-up loss → Insured loss

// Module 03 — Key Metrics & Outputs

Reading and Interpreting Cat Model Results

Question 10 of 20

A cat model EP curve shows a 1-in-100 year loss of $20 million for a portfolio. An insurer's appetite is $15 million at this return period. Which statement is correct?

AThe portfolio is within appetite and no action is needed

BThe portfolio exceeds appetite; the insurer could reduce the portfolio, buy reinsurance, or purchase a cat bond to cover the excess $5 million

CThe insurer should immediately exit the entire portfolio

DThe EP curve must be wrong because the loss is too high

Question 11 of 20

What is the key difference between OEP (Occurrence Exceedance Probability) and AEP (Aggregate Exceedance Probability)?

AOEP is used by primary insurers; AEP is used by reinsurers

BOEP measures the largest single annual event; AEP measures the total of all events in a year

COEP applies to natural perils; AEP applies to man-made perils

DOEP is a deterministic metric; AEP is probabilistic

Question 12 of 20

Under EU Solvency II, what return period loss must insurers demonstrate sufficient capital to withstand?

A1-in-50 year

B1-in-100 year

C1-in-200 year

D1-in-500 year

Question 13 of 20

A portfolio has an AAL of $5 million and a TIV of $500 million. What is the loss cost?

A0.01%

B1.0%

C10%

D0.1%

// Module 04 — Peril Deep Dives

Earthquake, Hurricane, and Flood Science

Question 14 of 20

Ground Motion Prediction Equations (GMPEs) are described as "region-specific." What is the primary reason for this?

ADifferent regions have different building codes requiring different equations

BCrustal geology, fault mechanisms, and local soil conditions vary by region and affect how seismic waves attenuate

CRegulators in each region require bespoke equations for compliance

DEarthquake magnitudes are measured differently in different regions

Question 15 of 20

In hurricane modelling, what is meant by the 'filling' of a storm?

AThe process by which rainfall fills flood plains in the storm's path

BThe rate at which a hurricane's central pressure rises and winds weaken as the storm moves over land

CThe infilling of data gaps in historical hurricane track records

DThe process by which the eye of a hurricane fills with cloud

Question 16 of 20

Paleoseismology is an important data source for earthquake modelling. What does it study?

AAncient weather patterns that may indicate historic earthquake-triggered tsunamis

BPrehistoric earthquakes, including their location, timing, and size, using geological evidence such as fault offsets and liquefaction artifacts

CThe economic impact of historical earthquakes on ancient civilisations

DThe study of seismograph technology developed before electronic instrumentation

Question 17 of 20

Which secondary peril associated with earthquakes involves saturated soils losing bearing strength and behaving like a liquid?

ATsunami

BFire following earthquake

CLiquefaction

DSubsidence

// Module 05 — Professional Practice

Model Limitations, Uncertainty, and Workflow

Question 18 of 20

A cat model is updated by its vendor, resulting in a 25% increase in the estimated 1-in-100 year loss for a major insurer's portfolio. Which limitation of cat models does this illustrate?

AThe difficulty of collecting building characteristic data

BModel update volatility — software changes can cause large swings in estimated losses from year to year

CAdverse selection in the insurer's portfolio

DThe protection gap between economic and insured losses

Question 19 of 20

In the standard cat model workflow, what is the purpose of the 'results review' step before reporting?

ATo ensure the model run time was within acceptable limits

CTo redact proprietary model assumptions before sharing with clients

BTo check for unexplained changes from prior results, identify outliers, and validate that output is reasonable before communicating to stakeholders

DTo select the model that produces the lowest loss estimate for competitive pricing

Question 20 of 20

Which statement best describes the appropriate use of catastrophe models for risk management decisions?

ACat model outputs should be accepted as definitive and used without question for all pricing decisions

BCat models are too uncertain to be useful; experienced underwriters should rely solely on their judgment

CCat models are powerful decision-support tools; their outputs should be interpreted with an understanding of limitations and uncertainty, and combined with expert judgment

DCat models should only be used by regulators; commercial insurers should use historical loss data

Answer all 20 questions before submitting.