Past papers 2015

Question 1

Profit commission

Answer 1

Profit commission is commission paid by the reinsurer to a cedant.

The commission is DEPENDANT ON THE PROFITABILITY or claims experience of the total business ceded during each accounting period.

Question 2

Risks-attaching basis

Answer 2

The basis under which reinsurance is provided for claims arising from policies COMMENCING DURING THE PERIOD to which the reinsurance relates,

irrespective of when the claims are incurred or reported.

A.k.a. Policies-incepting basis.
Corresponds to an UNDERWRITING PERIOD COHORT.

Question 3

Reinsurance contracts to which profit commission is most applicable

Answer 3

Profit commission is most likely to be used for:

• Proportional business (i.e. quota share and surplus); or
• Low excess layers of XL reinsurance (i.e. working layers)

Question 4

Reinsurance contracts to which a risks-attaching basis is most applicable

Answer 4

A risks-attaching basis is a natural arrangement for PROPORTIONAL reinsurance.

Question 5

Pure risk premium

Answer 5

The premium required to cover the EXPECTED CLAIM AMOUNT ONLY.
No allowance is made for expenses or profit.
We may express it as a nominal amount, but it is usually expressed as a rate per unit of exposure.

Question 6

Steps involved in calculating a pure risk premium

Answer 6

• collect relevant data, including past exposure data and claims arising from that exposure
• adjust the data to make them more relevant, e.g. if policy conditions have changed
• group the data into risk groups (if there are significant differences between groups)
• select the most appropriate rating model or estimation process for the specific case.
• analyse the data
• set the assumptions required by the model or process
• test the assumptions for goodness of fit or likelihood probability
• run the model or process to arrive at an estimate of future claims costs.
• perform sensitivity and scenario testing, or apply other methods, to check the validity of the estimate

Question 7

burning cost method

Answer 7

EXPERIENCE-BASED method that takes the actual cost of claims during a past period of years, expressed as an annual rate per unit of exposure.

This could apply to a single risk or to a portfolio of similar risks.

The technique may be
purely based on past claims without adjustment, although an improvement would be to
adjust past claims for trends and develop the claims to ultimate, but often this is not done in practice. If trending is applied to claims, exposure should also be adjusted.

The burning cost approach is commonly applied to aggregate claims, but may also be applied to frequency and severity separately.

The burning cost method is most suitable when there are lots of credible data. When the data are not credible, the burning cost premium should be combined with book rates using credibility techniques to obtain a more accurate premium.

Question 8

Original loss curves

Answer 8

Original loss curves are an EXPOSURE-BASED rating method.

The main principle of exposure rating is to not use historic claims experience at all, but instead to base premium rates on the amount of risk (ie exposure) that policies bring to the portfolio.

In exposure rating, we USE A BENCHMARK to represent a market severity distribution for the line of business and territory being covered. The benchmark may even be directly derived from the market severity distribution.

Original loss curves (exposure curves or ILF curves) are used to estimate the cost to the layer based on the exposure and premium information provided by the cedant rather than the actual cost and past exposure.

In particular, we commonly use original loss curves in excess of loss insurance pricing to infer prices for layers at which the data are too sparse to derive a credible experience rate.

So for example we might use them in place of a burning cost approach (which requires lots of credible data) when calculating the risk premium net of a layer of reinsurance with a high excess point, or perhaps even calculating the risk premium for the layer of reinsurance (from the reinsurer’s perspective).

Question 9

Benefits provided under personal accident cover

Answer 9

The insurance provides a fixed amount in the event that the insured party suffers the loss of one or more limbs or other specified injury, or accidental death.

Such cover is usually included in:

• comprehensive motor vehicle insurance
• household (contents) insurance
• employers liability
• marine and aviation liability

Question 10

Factors impacting whether an insurer may decide not to match its liabilities

Answer 10

• FREE RESERVES: Greater free reserves allow the company to mismatch assets and liabilities.
• BUSINESS PLANS: if the company experiences rapid growth and declining free reserves in future years, the scope for mismatching will reduce over time. Excess free assets are more likely to be used to fund business growth rather than be used as a buffer for mismatch risks.
• The extent to which the insurer can rely on premium income to meet short-term expenses and claims may allow it to mismatch; there is likely to be a high degree of uncertainty for a new insurer, and hence a need for close matching.
• NEED FOR DIVERSIFICATION to reduce specific risks from overexposure to a particular
  asset or asset class.
• The extent of REINSURANCE may increase investment freedom.
• The company’s ATTITUDE TO RISK and access to parent company resources may influence investment freedom.
• The outlook for returns for various asset classes may lead to short term tactical decisions leading to mismatching.
• Any REGULATORY REQUIREMENTS, including admissibility rules, may force the insurer to mismatch

Question 11

Characteristics of the liabilities under Personal Accident cover

Answer 11

• Claims are usually reported quickly, as the incidence of an event is usually very clear. However, with accidental death claims the insured’s dependants may not always know that the policy exists, and may discover their entitlement after an extended period, resulting in a reporting delay.
• The claims may be settled quickly, although if a claim is for permanent total disability, it may be necessary to wait several months or years for a claimant’s condition to stabilise.
• The claim frequency tends to be very stable.
• Claims can be large: cover of millions of rands per person is not uncommon.
• Benefits are fixed in a monetary amount and not exposed to inflation.
• Currency of the liabilities will be local.

Question 12

Matching assets for Personal Accident cover

Answer 12

CASH / MONEY MARKET INSTRUMENTS

• highly liquid and can be used to pay claims and other expenses
• capital values are stable, making them suitable for short-tailed claims, as there is no risk that assets will be sold at depressed market values.

SHORT-DATED (< 3-YEAR) BONDS

• These also tend to be very liquid, so they can be easily sold if a claim needs to be settled.
• As they provide a return that is fixed in nature, they provide a good match to fixed benefits.
• They can be used to match claims of slightly longer tail.
• They should provide a slightly higher expected return compared to cash.

Question 13

5 Different types of technical reserve

Answer 13

IBNR
IBNER
UPR
URR
Catastrophe Reserve

Question 14

IBNR

Answer 14

A reserve to provide for claims in respect of claim events that have occurred before the valuation data but have not yet been reported to the insurer at the valuation date.

Question 15

IBNER

Answer 15

A reserve reflecting expected changes in estimates for reported claims only.

Question 16

UPR

Answer 16

The amount set aside from premiums written before the valuation date to cover risks incurred after the date.

Question 17

URR

Answer 17

The reserve required to cover claims and expenses that are expected to emerge from unexpired periods of cover.

Question 18

Catastrophe reserve

Answer 18

A reserve built up over period between catastrophes to smooth the reported results over a number of years. The purpose of the catastrophe reserve is smoothing, not solvency.

Question 19

Explain one important consideration you should apply when using industry benchmarks

Answer 19

It is important to consider whether the source from which the benchmarks are derived has characteristics that are appropriate to the business for which the reserves are being derived.

E.g.

• similar claim reserving philosophy
• similar underlying nature and mix of business
• whether the benchmark is up to date
• benchmark may be based on more data than the target’s and hence exhibit less volatility

Question 20

6 ratios/quantities you could compare to industry benchmarks when assessing the level of IBNR reserves.

Answer 20

• Ultimate to paid
• Ultimate to incurred factors
• Ultimate loss ratios
• IBNR as a percentage of paid
• IBNR as a percentage of outstanding claims
• IBNR as a percentage of premium
• IBNR as a percentage of total technical reserves
• Average claim cost and frequency
• Survival ratios
• Incurred loss ratios
• Age-to-age development factors

Question 21

6 Aspects for which an insurer might be able to receive technical assistance from a pure reinsurance company

Answer 21

PRODUCT DESIGN, including advising on appropriate terms and conditions

PRICING,
the reinsurer will be able to provide past claims data to assist in setting pricing assumption.

ASSIST WITH DEVELOPMENT OF MODELS for pricing, catastrophe modelling, etc

RESERVE ASSUMPTIONS

INITIAL UNDERWRITING POLICIES

CLAIMS CONTROLS

Refining POLICY WORDING to avoid ambiguities.

Development of suitable ADMINISTRATION SYSTEMS

MARKETING STRATEGY - distribution channels and selecting target markets.

Question 22

Discuss the initial analyses which should be conducted on the data and explain why they are done before conducting a GLM analysis for pricing with multiple rating factors.

Answer 22

Before conducting an analysis using a GLM it is appropriate to firstly check and have an understanding of the raw data that will be used in the model.

The raw data need to be :
- checked for COMPLETENESS,
- checked that it has NOT BEEN CORRUPTED
- compared with that used in a previous review to ensure that the most recent and appropriate data are
being used.

• checked for the CREDIBILITY of data in cells
• and HOMOGENEITY WITHIN CELLS should also be conducted.
• rating factors chosen should have as much explanatory power as possible and remove the heterogeneity within each risk group.
• In selecting rating factors, we usually do an analysis of variance (ANOVA) exercise.
  This ANOVA can be a one-way analysis, two-way analysis or multivariate analysis.

We need to achieve a balance between the number of rating factors and the homogeneity of the risks.
- choose each additional rating factor to remove as much of the residual heterogeneity as possible.

If the factors do not sufficiently distinguish between
different levels of risks, insurers are likely to attract the underpriced risks and lose the over-priced ones.

However, if too many factors are chosen, insurers may experience difficulty due to high administrative costs and resistance of the market and brokers.

Having a good understanding of the underlying data will help when deciding on the appropriate balance of the number of rating factors to include into the analysis. A correlation analysis will explain why the multivariate results for a particular factor differ from the univariate results. It also indicates which factors may be affected by the removal or inclusion of any other factor in the generalized linear model. Cramer’s V statistic can
be used to understand the correlation between two variables.

We should also assess the validity of other risk groupings by stochastic analyses to test for differential results. We should adjust the theory for practicalities, including the availability of information and the applicability of systems. If we can’t get the data in a
reliable easy-to-use format, we may need to compromise our calculations.

Finally, a DISTRIBUTION ANALYSIS for claim amounts could be analyzed for various segments of the data.

• identifying presence of large losses
• better understand the statistical distribution of the claim amounts.
• reveal unusual patterns that may need to be investigated.

Question 23

One-way ANOVA

Answer 23

In the case of one-way ANOVA, we investigate the amount of variability explained by each factor without taking into consideration the correlation between factors.

We may find that when we split policyholders into different age groups, for example, the variability of claims experience within each age group is small relative to the variability in the overall portfolio of risks. Hence the factor “age” helps to ‘explain” the variability
because, after grouping the policyholders by age, there is little residual variability left within the groups.

Question 24

Two-way ANOVA

Answer 24

In a two-way analysis of variance, we investigate each factor and the correlations between any two of the factors.

This can explain the variability better than a one-way analysis.

For example, the one-way analysis may show that the size of a household claim is highly related to both the number of bedrooms and the value of the contents.
A two-way analysis may reveal that these two factors are in fact highly correlated, so that only one should be
included in the pricing factors.
It also helps in identifying the interaction effect between factors and can reveal the exposure and claim numbers for various combinations of levels from a pair of factors.

Question 25

Claims characteristics

Answer 25

The ways and speed with which claims

• originate
• are reported
• are settled
• on occasion, are reopened

Claim frequency and amount as well as potential accumulations are also relevant.

Question 26

Describe the claim frequency and severity of Motor insurance

Answer 26

 High expected frequency, with many accidents for small claim amounts.
 There may be variability in claim amounts due to aggregations, for example a hail storm. Full accident write-offs or theft of high value vehicles may result in large claims, although not as large as the largest buildings insurance claims.

Question 27

Describe the claim frequency and severity of Household contents insurance

Answer 27

 Lower frequency of claims than motor, but may be higher than buildings due to frequency of theft.
 There may be little variability in claim amounts, having limited scope for very large claims being personal lines business, except maybe for specie/art.
 There is some risk of aggregation of claims due to catastrophe such as hail/earthquake.

Question 28

Describe the claim frequency and severity of Household buildings insurance

Answer 28

 Low frequency of claims, although geyser – related claims (which fall under buildings cover) are common. Some buildings damage as a result of break-ins, so the frequency of such claims will be related to theft frequency.
 High variability in the size of losses, as there will be many small claims with the possibility of very large claim amounts these could be caused by one off large losses such as subsidence, conflagration (fire) or earthquake.
 Catastrophes such as earthquake may result in large losses from single events (geographical accumulations).

Question 29

Discuss the loss components that will need to be modelled separately

Answer 29

Attritional:
 Attritional loss can be modelled as an aggregate distribution using a Lognormal/similar.
 Due to the high number of losses classed as attritional, the stability of the distribution means that an aggregate distribution is appropriate.
 The loss distribution can also be linked to premium income (where all else equal, high growth results in a higher absolute loss)

Large:
 Large losses are best modelled using a frequency-severity approach. This allows more details of the available information to be used, especially important where there is scarce data.
 The additional detail is worth the effort due to the larger values and will also enable reinsurance limits to be tested more adequately.
 Frequency can be modelled using a Poisson/Negative Binomial.
 Severity can be modelled using a Pareto/LogNormal or other heavy-tailed distribution.
 Need to decide on a threshold per class of business above which large losses will be modelled, to not allow large losses to unduly affect attritional claims distributions.
 Large losses will be important for household buildings and motor (1st and 3rd party).

Catastrophe:
 Catastrophe losses will be modelled using different approaches depending on the peril.
 Earthquake – may use a catastrophe model which applies losses to the exposure base.
 Factors which will affect the size of the loss include event magnitude and location.
 Other perils such as hail/flood etc. may be modelled using a frequency severity approach.
 This could be based on historic losses experienced by the insurer/industry, or purely a theoretical forecast model.
 Motor may be more affected by hail/flood damage than earthquake, whereas buildings are most affected by earthquake.

Question 30

Outline how the correlations between the different lines can be allowed for in the capital model.

Answer 30

 Can correlate lines of business in attritional losses and large loss number/size of losses. This depends very much on historic experience.
 Catastrophe model will automatically apply same scenario, so there is some correlation built in.
 Underwriting cycles may introduce correlation between the lines of business, as well as dependency as a result of economic conditions.
 Stronger correlation between buildings and contents than the other lines of business, as contents are normally in the insured building.
 Diversification between lines will serve to reduce the overall capital requirements.

Question 31

5 Accounting concepts

Answer 31

• going concern
• accruals basis
• consistency
• prudence and realisation
• separate valuation of assets and liabilities

Question 32

Going concern

Answer 32

The enterprise will continue in operational existence for the foreseeable future

Question 33

Accruals basis

Answer 33

Revenue and costs are recognised as they are earned or incurred, not as money is received or paid.

Question 34

Consistency

Answer 34

There is consistency of accounting treatment of like items within each accounting period and from one period to the next.

Question 35

Prudence and realisation

Answer 35

Revenue and profits are not anticipated (i.e. they must be realised) and provision is made for all known liabilities, whether the amount of these is known with certainty or is a best estimate in the light of the information available.

Question 36

Separate valuation of assets and liabilities

Answer 36

When determining the aggregate amount of any item, the enterprise must determine separately the amount of each individual asset or liability that makes up the item.