You are on page 1of 265

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
5 April 2005 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 A2005

Faculty of Actuaries
Institute of Actuaries

Company A is a large international insurance and reinsurance group whose statutory


returns show annual gross earned premiums in excess of $5,000m. The company
writes all classes of business from a large number of offices widely spread across the
globe.
You are an actuary within Company A. The company measures the performance of
its underwriters on taxable profit, including credit for investment return. Investment
return is credited at a notional rate of interest decided before the start of the financial
year.
Company A is aiming to increase taxable profit in the next financial year. The budget
for each portfolio of business written in Company A must be submitted in June so that
the group can finalise the overall figures by the end of July.
The underwriting manager of a niche commercial insurance business within Company
A is responsible for $200m of annual gross premium which has been stable at this
level for many years. Company A has set a budgetary target for this portfolio of
increasing taxable profit from $20m to $24m in the next financial year. This business
is underwritten in 6 offices each of which operates in a different country. Over half of
this business is written in the London Market. Most of the policies in this portfolio
have two year exposures and the risk exposure increases from zero at a uniform rate
over the policy period. Policies are written evenly over the course of the year.
For this portfolio acquisition costs have consistently been 20% of gross premiums and
reinsurance spend has been 17% of premiums. The reinsurance has been well used
and in the manager s opinion has been key to smoothing profits. Company A has
asked the underwriting manager to halve the amount of reinsurance purchased.
Internal expenses have been cut over the past two years from 15% to the current year
budget figure of 13% of gross premiums. The mean duration of claim settlement is
four years from policy inception.
This underwriting manager has asked you for a report outlining the options open to
him to achieve the company profit target explaining the advantages and disadvantages
of the options.
(i)

Discuss the possible options open to him that you would consider for your
report.
[47]

The underwriting manager has expressed dissatisfaction at the rate of investment


return credit, telling you that he is getting more on his own fixed interest investments.
He also disagrees with Company A s decision to cut his reinsurance spend by a half.
(ii)

SA3 A2005

Explain why the underwriting manager s view may be different from


Company A s view.
[18]
[Total 65]

The FSA is in the process of changing the capital requirements from solvency margin
to Individual Capital Assessment (ICA).
(i)

Describe the core elements of the new proposals under the headings of ECR,
ICA and ICG.
[4]

(ii)

Describe the three components of the ECR calculation.

(iii)

Discuss the advantages and disadvantages to a general insurance company of


ICA compared with ECR.
[6]

(iv)

Outline the main areas that are expected to be covered in the ICA format
proposed by the FSA.

(v)

[4]

[9]

For each of the risk types itemised under these proposals:


(a)

State each risk and suggest a basis for modelling each of them.

(b)

Explain, with examples, how each risk may be correlated with the
others.
[7]

(vi)

You are the actuary for a London market company writing commercial
property business both insurance and reinsurance. Describe how you would
parameterise the model of the insurance risk.
[5]
[Total 35]

END OF PAPER

SA3 A2005

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
April 2005

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT

Introduction
The attached subject report has been written by the Principal Examiner with
the aim of helping candidates. The questions and comments are based around
Core Reading as the interpretation of the syllabus to which the examiners are
working. They have however given credit for any alternative approach or
interpretation which they consider to be reasonable.

M Flaherty
Chairman of the Board of Examiners
28 June 2005

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

The examiners were generally disappointed with the majority of solutions given to this
question.
In particular there was a lot of information given in the question which many
candidates did not seem to use to form their solution. There were several references
to figures in the question which candidates did not refer to in their answers, in
particular regarding the pattern of earning of premium, and the value for money of
reinsurance.
Owing to the lack of reference to information given candidates did not gain marks for
mentioning the likely relevance of some of their options in achieving the target profit.
When candidates did mention the possibilities of meeting the target many considered
the long term rather than the short time window of the next financial year. Hence
some of the solutions suggested options which would not deliver the profit target in
the following year.
Their seemed to be very little reference to the fact that the expense action already
taken may not have been fully reflected in the profit figure and hence this was a
source of further improvement for the following year.
Some candidates seemed to be unsure what areas of the business the underwriter
would be able to influence.
Solutions to part (ii) were generally very good.
(i)

Options which could increase earned profit next year are:


1.
2.
3.
4.
5.
6.
7.
8.
9.

Increasing written and therefore earned premium.


Reducing acquisition expenses.
Reducing internal expenses.
Reduce the cost of reinsurance.
Writing to a lower loss ratio (for new / renewing business).
Reduce the cost of claims for claims settling next year.
Speeding up the earnings pattern of the portfolio.
Getting a higher investment credit.
Change to reserve assumptions and methodology.

For each option assume everything else unchanged:


1.
For two year policies: earned in yr 1 = 50% * 1yr * 50% pol exposure =
25%. Therefore 75% earned in year 2. (At 18 mths = 50% * 1.5yr * 75%
pol exposure = 56% )
Policies to be written in the forthcoming year will commence uniformly
over the year and will on average be written half way through the year.
They will thus be exposed for a quarter of their policy term. In the

Page 2

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

following year the earned premium on these policies will thus be 6.25% of
that written.
The portfolio premium would need to grow by $640m.
This is a very substantial increase which may be part of the plan for this
portfolio.
But may not be previously planned and could be a tough target.
Can the individual markets/offices deliver this?
Increase in premium likely to be restricted as niche market.
Over half is written in the London Market
subscription market
therefore could increase its written lines on business.
But the company could be already writing large lines with cedants and
brokers unwilling to give Co A larger lines.
The company could also write larger risks
Could diversify the types of risks written and by channel
Would the loss ratio suffer by taking on larger volumes of business which
may not be as good quality as the existing business.
Would the acquisition costs suffer by accepting business from higher cost
sources.
Can the front office / back office cope? Would direct expenses increase to
cope with the new volume?
Will the reinsurance programme be as efficient? How does the RI
premium adjust? Will RI aggregates be breached? How does the new
volume and business mix compare to the RI submission information?
Will the terms of the RI treaty have to be renegotiated?
May breach statutory premium limits in some countries.
Requiring the injection of capital into some countries.
2.
Acquisition expenses will be different from the different sources.
Some sources will be higher than others. Some may be relatively very
expensive.

Page 3

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Could cut the amount of business from the highest cost sources.
This may improve the percentage profit but not necessarily the absolute
profit
Could take a tough negotiating stand with all brokers to cut brokerage.
Which may be possible if the company dominates this niche market.
But may result in lower volumes if the brokers can place the business
elsewhere.
Restructure commission arrangements to give incentive to provide higher
volumes
However there are sensitivities regarding volume related commissions
(Spitzer)
Could push for (more) direct business with very low acquisition costs.
But again this may annoy brokers who may take business elsewhere.
Could make commissions profit related so higher commissions only paid
on profitable business.
But this will reduce the benefit of better business.
3.
Reduced over the last two years
made?

have all efficiency savings already be

It has taken 2 years to reduce expenses by 2%, there may be some savings
in the next year from previous/current actions.
You are unlikely to achieve the full 20% improvement with expenses
alone, however a further 1% reduction in expense ratio would contribute
$2m to the required increase in profit.
Niche product, therefore specialists required in underwriting / claims,
massive savings probably not possible.
Internal expenses consist of direct and indirect, much of the internal
expense will be out of the control of the portfolio manager.
Closing an office or other drastic measure would probably not deliver the
earned profit in the required timescale
redundancy payments etc.

Page 4

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Introduce new admin procedures to cut costs, but again this is unlikely to
be achievable in the required timescale.
4.
RI well used and key to smoothing profits so we expect a fairly high
recovery rate, assume 50%, but could be higher or lower.
Must reduce the net cost of the reinsurance, so take into account
reinsurance recoveries as well as the outward premium. Not buying any
RI would not reduce the costs by 17%. If recoveries average 50% then
cutting all RI would be expected to increase profit by 8.5%.
Reinsurance outward premium is $200m * 17% = 34m assuming 50%
recovery means that the cost will be $17m. Halving the reinsurance spend
will reduce the cost from 17m to 8.5m
increasing profits by 8.5m.
RI looks to be the solution to the manager s problem so we need to know
what the current arrangement is and what loss ratio it is running at.
Depending upon the insurance cycle may be able to cut reinsurance
premiums payable.
It is possible, but unlikely, that reinsurance provides a net benefit to the
portfolio.
Need to know when the RI incepts and any existing agreements
if
incepts mid year then will only be able to reduce the costs for a part year.
Long term agreements may mean that no changes are possible.
Reducing Xol reinsurance spend will increase the % volatility of the
account.
If Xol Reinsurance cover is reduced by:
cncreasing retention
reducing the limit
or retaining a self insured share across the programme
Purchase reinsurance in-house and avoid paying brokerage.
Reducing QS reinsurance spend will not change the % volatility of the
account.
Ads / disads.
This may not be what the manager is comfortable with. Manager may
prefer a less volatile / more certain result.

Page 5

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Reducing the amount of reinsurance purchased may reduce the other


perceived benefits obtained from reinsurers e.g. market knowledge.
Purchasing less reinsurance might also lead to higher capital requirements.
5.
Could write business to a lower Loss Ratio by tightening policy
conditions, increasing rates, and better risk selection.
However, the team should already be writing the best business they can so
this could be a difficult task.
Increasing premiums may lead to loss in business overall, depends upon
elasticity of demand.
Alternatively the LOB may be writing to significant losses.
So improvement should be much easier.
The market cycle may be hardening making the task easier,
or v.v.
It may be possible to be more rigorous with renewals, exiting poor
business.
Statistical analysis of data may identify both good and bad business which
could be targeted for appropriate action.
But writing higher business volumes works against writing to lower loss
ratios.
May identify one office with a higher loss ratio,
leading to difficult decisions which the manager may not want to take.
May mean axing some policies which have been written for many years,
and breaking longstanding relationships
difficult.

Page 6

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

6.
Profit is measured on an earned basis. So reducing incurred claims next
year would increase profits.
Reduce margins in reserves would increase taxable profit in that year,
although this would increase the risk of reserve inadequacy.
Could be more active in claims adjusting, settling outstanding claims for
less.
This may be possible in a specialist / niche LOB.
But this should be being done already.
This may increase loss adjustment expenses.
But should be more than compensated by lower claims.
But the strategy may not be successful.
7.
Accelerating the earnings pattern of new / renewing business will
accelerate earned premium next year.
Thus accelerating earned profit.
Assuming business written to a profit.
But this can only be done once per policy.
But may not be possible in this niche Lob.
8.
Not a lot can be achieved here. Investment return is credited according to
a company allocation.
This may be achievable but is not an exercise which adds value to
company A. It benefits the manager at the expense of someone else.
Increasing this portfolio s credit would mean reducing the credit for
another portfolio.
If the total investment return is to equal that earned.
9.

Page 7

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Change Reserving Methodology to reduce claims incurred in the financial


year (or their present value). The extent to which this may be possible
depends on the manager s sphere of responsibility or influence and the
legal constraints in the territories concerned.
Discount reserves at a higher rate of interest if this is allowed.
Extend mean duration of discounted business
slower payout pattern.

discount according to a

Reduce case estimates.


The above may be perfectly legal if a conservative approach is currently
taken, but care does need to be taken.
These options do not add value to Company A, they impact the timing of
the emergence of profit.

(ii)

Conflicts of interest / opposing views


Reinsurance spend
Investment return
1. RI Spend
Manager views RI as crucial to smoothing results. Elimination of large
losses is achieved by Xol not QS RI.
17% of $200m is $34m which is a large cash spend.
What the manager views as a large loss may be insignificant to the
company.
Manager may be buying down to very low retentions which Company A
would rather retain.
Manager may be protecting events which may be more efficiently
protected at group level.
This may be even more acute for the smallest premium volume offices
where a moderate sized claim could lead to an underwriting loss.
Reinsurers aim to make a profit so the manager is ceding profit over the
long term.
RI arrangements could involve funded/finite deals which smooth profits
but do not benefit Company A.
Ceded RI may be with companies less financially secure than Company A.

Page 8

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Some RI purchases may have more to do with relationships than any added
value.
2. Investment return
Not a function of underwriting so can argue should not be part of the
performance assessment.
Manager may not be taking into account currency matching of reserves
or matching by duration.
Some currencies e.g. Japan, USA have much lower interest rates than the
UK.
London Market business does not have to be denominated.
Matching investments by duration, the manager s view may be much
shorter term.
Investment managers should have their own targets, LOBs should not
benefit / suffer from any deviation from target by the investment
managers.
Managers view may not take into account the investment costs.
Or the investible percentage of funds.
The quality of investments may be different.
The company is not necessarily invested in fixed interest investments and
for the time being return on their investment return is set at a lower
notional rate but in the long term will be higher.
May be regulatory restriction for the company on what they can invest in
and this affecting the notional return.
Effect of tax on investment returns may be strict for the company than the
individual thus affecting net of tax return.

The better candidates showed that they had both read the notes and could
demonstrate an understanding of the bookwork, and that they had prepared for the
exam by reading around the subject in respect of important factors affecting the GI
industry at the moment.
The examiners were disappointed to see though that a number of candidates did not
appear to have studied the course notes on this subject.

Page 9

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

(i)
(1) A new risk-based enhanced capital requirement (ECR) based
on capital charges to be applied to asset values and insurance
premiums and reserves.

ECR

(2) Insurers will be required to undertake their own assessment of


their capital needs according to the size and nature of their
business taking into account major sources of risk (including
systems and controls and operational risk) and calculate their
individual capital assessment (ICA) using either stress and
scenario tests or stochastic modelling. FSA guidelines state the
minimum risk appetite should be 99.5% confidence of capital
adequacy over a 1 year time horizon

ICA

(3) Individual capital guidance (ICG) will be set by the FSA. ICG
is based on the FSA s own view of how much capital
individual insurers should hold, taking into account insurers
assessments of their own capital needs, risk and capital
management processes.

ICG

(ii)

The ECR is an FSA-prescribed, risk-sensitive calculation, made up of the sum


of various capital charges based on asset categories and underwriting risk:
Asset-related values
Insurance related values
Net written premium

related asset factor (%)


relevant technical provisions factors (%)
relevant premium factors (%)

Asset-related values are calculated after the application of valuation and


admissibility rules.
Insurance-related values comprise (for each class of business), the total
technical provisions, including outstanding claims, IBNR and IBNER
claims, unearned premiums reserves and any additional unexpired risk
reserves. Claims reserves are net of anticipated reinsurance recoveries and
premium reserves are net of deferred acquisition costs.
Premium values are gross written premiums net of reinsurance but before
deduction of commission.

Page 10

Subject SA3 (General Insurance Specialist Applications)

(iii)

April 2005

Examiners Report

Advantages
Risk based so assessment and framework needed so controls and systems.
Company specific data can be usetd so should be more accurate.
Company calculates so forces to assess and understand the risks in more
detail
may lead to better management.
Rigorous documentation.
Company can continuously monitor the amount they need and calculate
changes from changes in strategy i.e. buying more reinsurance changing
portfolio of business previous backward looking.
Higher level of capital may reduce insolvencies.
ICA does not penalise conservatively reserved and sufficiently premium
rated companies.
Disadvantages
Complex, difficult to model and to validate model, i.e. higher costs etc.
May not be transparent.
Different companies may take different approach to modelling
assumptions (for example tails or correlations), which may lead to
different levels of capital being set.
Need lots of data.
Even with large database some risks are very subjective for example cats
and operational risk.
Expected higher capital may put UK companies at a disadvantage in
producing returns for shareholders.

(iv)
A summary of the financial position of the firm at the time the report is
constructed and the risks to which the firm is subject.
The firm s proposal for ICA expressed as a proportion of its ECR
calculation.
Relevant historical development of the firm and any conclusions that can
be drawn from that development which may have implications for the
future of the firm.

Page 11

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

The business profile of the firm, the environment in which it expects to


operate, and its projected business plans, projected financial position and
future sources of capital.
A detailed review of the capital adequacy of the firm. This analysis could
include a commentary and opinion on the applicability of the ECR to the
firm s own capital position and its appropriateness compared to its own
capital assessment. It could involve an analysis of the current capital
levels and movements in solvency levels during the past years, future
capital requirements, and general outlook.
An identification of the major risks faced in each of the following
categories: credit risk, market risk, insurance risk, operational risk liquidity
risk and group risk, and the extent to which a firm holds capital in
response to each risk.
The quantitative results of stress tests and scenario analyses carried out by
the firm and the confidence levels and key assumptions behind those
analyses.
Identification of any risks (for example, systems and controls weaknesses)
which in the firm s opinion are not adequately captured by the ECR and
the firm s assessment of how the firm is responding to those risks and if
through holding capital, the amount.
If the firm uses more sophisticated modelling approaches, the FSA would
also expect a statement on the confidence level and other parameters that
have been used in the model.

(v)

Market risk This risk could be modelled using an economic model that
would enable a stochastic asset valuation at the end of the period by
simulating returns for each asset type. Alternatively could use asset
movement stress and scenario testing. Such risk could be correlated with
insurance risk as dependent upon inflation. Also could be correlated to
liquidity risk for example in the case of reinsurance failure.
Insurance Risk
This risk is likely to separate catastrophe claims experience
from attritional experience. Likely to model risk in respect of claims, reserves
etc using ECM (economic capital model). Correlations with most other risks,
e.g credit risk following effect of cat claim, liquidity risk for reinsurance
failure
Credit Risk Simulate counterparty risk using an ECM stress and scenario
test. Credit risk arising from a reinsurer failure could have an impact on other
insureers leading to a market risk, and also as mentioned above could lead to a
liquidity risk.
Liquidity Risk
In this case would likely to produce a cashflow model. As
stated above this is closely linked to insurance risk if large claim occurs.

Page 12

Subject SA3 (General Insurance Specialist Applications)

April 2005

Examiners Report

Group Risk
In this case each part of the group would be modelled
separately, i.e. parent or subsidiaries. There is a link here with market risk if
an event takes place that affects all areas of the group.
Operational Risk
This risk is likely to occur as a result of poor management
leading to monetary loss. As such stres and scenarion testing against the risk
register is likely to be used there is a correlation with Insurance risk as bad
management could lead to high loss ratios.

(vi)

For the insurance risk need a frequency severity model. Model gross claims
and also insurance and inwards reinsurance separately. Most likely to model
attritional, large and cat losses separately. Use historic data converted to
current times allowing for claims inflation, line size change, profile of the
account and rate changes (to capture relative exposure), and insurance cycle.
This will allow an estimate of freq and severity for the large loss, and
distribution for the attritional claims. Parameters of the distributions should be
varied and parameters should be estimated by fitting distributions to the data
using goodness of fit tests. RMS or other cat modelling software can be used
for the cat part.
The assumptions should be discussed with management and the underwriters
and checked against the business plan. This gives the gross losses. Set up a
stochastic model and model explicitly any recoveries from the reinsurance
program including reinstatement premiums.

Page 13

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
6 September 2005 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 S2005

Faculty of Actuaries
Institute of Actuaries

You are a consulting actuary for a medium sized UK general insurance company
writing personal and commercial motor, household buildings and contents,
commercial property, employers liability, public liability and professional indemnity
business. The company does not write reinsurance business and the business consists
mainly of UK risks.
(i)

Discuss the likelihood of occurrence of large individual claims and


catastrophes in each of the company s classes of business, giving examples of
how they may arise in each case.
[16]

(ii)

Explain why it may be necessary for this company to adjust claims data for
pricing and reserving in respect of large individual claims and catastrophes.
[8]

(iii)

(a)

Describe different ways of defining and treating large individual


claims when considering the extraction of them from the claims
triangulations for the purposes of establishing reserve requirements for
the personal motor business.

(b)

Set out the advantages and disadvantages of each approach.


[15]

In June 2002 the company recruited a new senior claims handler with responsibility
for improving the company s motor bodily injury claims reserving. Within six
months of joining the company the new recruit had implemented new claims handling
practices and had reviewed all claims larger than 50,000. The company believes this
review was responsible for significant increases in case estimate amounts.
You have been asked to review the company s personal motor reserve projections.
The methodology has been to remove claims with incurred cost greater than 100,000
from the claims triangles, project non-large claims using standard chain ladder
methodologies and include reserves for the large losses as the sum of the large loss
case estimates. You have been provided with the large claim listing shown opposite in
respect of the personal motor classes of business.
Total paid claims in respect of these large losses amount to 3,339,000 as at
31 December 2004.
(iv)

Explain the disadvantages of the company s approach to reserving for this


personal motor account.
[5]

(v)

Estimate a gross of reinsurance claims reserve for the personal motor large
claims using the claims information above, including an allowance for the
future emergence of large claims in relation to expired risk as at 31 December
2004, explaining any assumptions made.
[16]
[Total 60]

SA3 S2005

Notified claims at year-end (000)


Date of Loss

1998

1999

2000

2001

2002

2003

2004

12/02/98
01/05/98
02/06/98
27/11/98
1998 total

12
125
25

89
150
120
100
459

315
155
120
175
765

315
136
120
175
746

315
136
150
132
733

315
136
150
132
733

315
136
150
132
733

350

750
102
852

750
102
852

1250
120
1370

1250
102
1352

967
102
1069

5
55

40
70
225

60

250
585

60
125
225
200
325
935

110
125
196
210
325
966

110
122
196
210
325
963

425
150
45
620

2500
120
45
2665

2500
90
150
2740

62

80
30
116

450
312
824

117
350
425
312
1204

100
25
125

130
200
330

162

13/01/99
22/04/99
1999 total

350

03/03/00
21/05/00
05/08/00
17/09/00
13/12/00
2000 total
02/07/01
24/10/01
09/12/01
2001 total

300
10
310

30/03/02
03/05/02
18/11/02
26/12/02
2002 total
02/02/03
17/02/03
2003 total
24/06/04
2004 total
All losses

SA3 S2005

400
400
162

809

1677

2493

3774

6665

7439

PLEASE TURN OVER

You are the actuary for a proprietary company, company Z, based in the UK, which
writes only treaty reinsurance business, but does not write retrocession business.
The company will consider writing reinsurance of any standard class of general
insurance business.
It currently underwrites a small (around 100) number of large treaties, both on a
proportional and non-proportional basis.
The company s senior management has requested that the monthly management
information reports provided are changed to provide clearer output upon which to
base their decisions.
The first request that has been made is to provide an index, which will illustrate the
average year-on-year change in expected profitability to the company of each contract
written, sub-divided by line of reinsurance business written.
(i)

(ii)

Discuss how you may go about constructing such an index, detailing the
following:
the issues you would consider
assumptions that you would make
the data that you would require
the limitations that you would wish to highlight to the senior management
[25]
Suggest the sorts of other management information with regards to the
underwriting and claims processes that could be usefully monitored on a
monthly basis for this type of reinsurance company.
[6]

One of the company s cedants buys an excess of loss policy with four reinstatements
from company Z, each at 100% of original premium. They have asked whether your
company can provide a form of insurance known as reinstatement premium
protection, which would cover their obligations to pay reinstatement premiums in the
event of them making a claim on the original policy.
(iii)

Compare this reinstatement premium protection approach with the alternative


of adjusting the original contract to provide free reinstatements.
[5]

The original contract provides 1 million of cover with four reinstatements. The
reinstatements are for the same amount as the original contract
250,000 each.
Loadings to the risk premium make up 20% of the price actually charged.
(iv)

Calculate the theoretical risk premium for the reinstatement premium


protection cover, if it covers one loss with three paid reinstatements at 100%
of original premium.
[4]
[Total 40]

END OF PAPER

SA3 S2005

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
September 2005

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however
given credit for any alternative approach or interpretation which they consider to be
reasonable.

M Flaherty
Chairman of the Board of Examiners
29 November 2005

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

The examiners were generally disappointed with the answers to this question.
Although most candidates managed to answer part (i) reasonably well there were
generally very poor solutions given regarding how to deal with large claims in parts
(ii) and (iii). Part (v) was also badly answered with many candidates blindly using the
Chain Ladder method without consideration of the data available. A few candidates
could not even use the Chain Ladder correctly which is very basic methodology for GI
reserving. Ignoring the data given in part (v) and inventing own data did not gain any
marks.
The use of abbreviations without definitions does not help the examiners in assessing
if the candidate understands the issues being considered. Using non-standard
abbreviations even if definitions are given is not welcomed. The increase usage of text
speak is also most unwelcome as such speak would not be tolerated in the business
environment.
(i)

Personal and commercial motor

Might expect a reasonable proportion of total claims cost to arise from


large individual claims greater than 100,000
Large individual claims are likely to arise due to bodily injury rather than
due to property damage (unless very high value or large commercial
vehicles)
Likelihood of large claim usually higher for commercial owing to mileage
driven, although will depend upon experience of driver
Likelihood of large claim higher for young drivers which may generally
have non-comp rather than comp insurance
Likelihood of large claims is increasing owing to Court Awards and
general litigiousness
Catastrophes may arise from, say, a motorway pile up
or weather events such as floods
Potential accumulation of risk is greater for Commercial Motor
although these are likely to have less impact on overall claims costs than
large individual claims

Household buildings and contents

Page 2

Most household contents claims are small


as mostly property damage claims
More household buildings claims are large
e.g. total destruction due to fire, or total rebuild for subsidence
But generally the proportion of large individual claims is smaller than for
motor (or some other valid comment about relativity to other classes)
Catastrophes are a significant feature for household insurance, being a key
driver of profitability for a particular accident year
These generally arise due to weather conditions
such as flood, storm, freeze
Subsidence claims are not generally particularly large (on average about
10,000).

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

although they tend to aggregate regionally (due to type of soil).


and their occurrence is strongly linked to weather conditions
and therefore a bad year for subsidence may be considered a catastrophe
year
Possibility, although very unlikely to have a large PL claim

Commercial property

Large individual commercial property claims are common


As a proportion of total claims cost large individual claims are more
significant for this class than for motor or household (or some other valid
comment about relativity to other classes)
Potential for very large claims depends on nature of portfolio (e.g. retail,
industrial, small/medium/large assureds)
Large claims can arise when there is significant property damage
E.g. fire resulting in destruction of whole building
But also from business interruption claims if this cover is included within
the contract
Catastrophes generally arise due to weather conditions
Potential for accumulation of losses owing to proximity of risks

Employers liability

Employers liability gives rise to bodily injury claims of various sizes,


including some very large ones
Large individual claims can arise where bodily injury is such that cost of
medical care is very high e.g. back injuries
or employees salary is high
or employee is young
and therefore loss of future earnings when unable to work is high
The most serious asbestos-exposure claims (e.g. mesothelioma) can give
rise to individual claims in excess of 100k
The likelihood of some large claims (e.g. asbestos) will depend upon size
of past exposure and trades covered
Occasionally catastrophes can affect this class, although this is less of a
feature than for household business (or some other valid comment about
relativity to other classes)
E.g. Piper Alpha oil rig disaster in 1988 did find its way into employers
liability accounts / other suitable example
Catastrophes will depend upon trades covered

Public liability

Public liability gives rise to property damage and bodily injury claims of
various sizes, including some very large ones
Likelihood will depend upon business covered, e.g. major sporting event
Claim size distribution is generally more skew for public liability than for
employers liability

Page 3

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

Sometimes public liability includes product liability cover; this can lead to
aggregation of claims (e.g. product recall)
or large individual claims (e.g. pharmaceutical products)

Professional indemnity

(ii)

Outstanding Claims Reserves calculations

Page 4

Claim sizes generally depends on professions covered within account


Likelihood of a large claim depends upon policy terms and conditions and
generally frequency is more variable than for other classes
A professional negligence claim against a large firm of accountants may
result in a very large claim if a company became insolvent as a result of
negligent advice
Market-wide issues such as pensions misselling claims on professional
indemnity for IFAs, may be considered as catastrophe claims

If left unadjusted in aggregate data, individual large claims might distort


the projection of the OCR
This is the case if individual large claims have a different claims
development pattern than non-large claims.
and the mix of non-large and large claims varies from year to year (due
to random large loss experience i.e. if frequency of large claims is low)
...then leaving large claims in the aggregate data could result in unstable
chain ladder development factors
and average development factors for each development year might be
distorted by unusually high or low large loss experience in recent years
and even when the averages are not distorted, applying an average chain
ladder development factor might be inappropriate for those years of
account with unusually high or low large loss experience.
Catastrophes can cause a similar problem to individual large claims
although the various individual claims arising from a catastrophe may
develop at a similar speed to non-catastrophe claims
they may bias the average date of occurrence
e.g. storm occurring at the end of an accident year for the household
account might result in year being less mature than normal
claims resulting from storm and flood catastrophes tend to be reported
very quickly and therefore distort reporting pattern, whereas claims from
subsidence catastrophe tend to be reported quite slowly splitting of such
claims leads to greater accuracy within modelling
Catastrophes may lead to greater claims leakage owing to pressure of
making payments this distorting the true payments pattern
The inflationary effect on a large claim is likely to be different to that on
smaller claims

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

Rating factor relativities

If left unadjusted in aggregate data, individual large claims would unduly


dominate the experience of the risk group.
and might lead to inequitable pricing
which in turn might lead to antiselection
This is particularly relevant for rating cells/risk groups with lower volumes
of data
e.g. 80 year old drivers for private motor insurance
where the presence of a large claim is more due to random occurrence
rather than systematically bad experience
Could create non-competitive premiums

Reinsurance calculations

It might be necessary to assess current and future recoveries on excess of


loss and catastrophe reinsurances
And this may be easier to do by removing the elements of large claims that
are recoverable and projecting them separately

(iii)

1.

There are various different ways that large losses can be extracted from the
claims triangulations and
there are different definitions for a large individual claim
Different extraction approaches include:

Do not extract large claims from data


+ Simple and quick
+ Fairly robust if large claims experience has been fairly stable
from year to year
+ Ensures reasonable allowance for unreported large claims
May result in over/underestimation of IBNR if large loss
experience has not been stable
Does not recognise trends in large claim experience

2.

Extract whole of each large claim and associated history if its incurred
claim amount exceeds a certain threshold e.g. 100,000
+ Non-large claims triangulation is not distorted by part-history
of large claims
Will need to restate history of non-large triangulation each year
as non-large claims become large
So difficult to reconcile with last years data
Difficult to allow for claims currently classified as non-large to
become large

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

3.

Once large always large => even if incurred claims for a loss falls
back below threshold, still treat as large
+ Reduces need to amend history of non-large triangulation each
year
+ Recognises the potential for large claims to become non-large
and therefore avoids over-estimation of reserves for large losses
May distort any large claim average cost analysis

4.

Only extract claim from the point that it become large i.e. history of
claim before large remains in the aggregate data
+ No need to amend history of non-large triangulation each year
May be sharp reductions in claims in non-large triangle from
one development year to next
Development factors that rely too heavily on such an instance
would result in optimistic non-large IBNR estimate

5.

Apply indexing to the large claim definition e.g. 100,000 for losses
occurring in 2000, 105,000 for losses in 2001 etc.
+ Ensures that large loss definition maintains real value over time
+ If there were no indexation, there would be very few claims
extracted from early years of account compared to later years
and this would reliability of development analysis
+ Can make definition coincide with excess point for excess of
loss reinsurance
Indexation introduces complexity
Inflation hard to measure

6.

Only extract the part of each large individual claim that is in excess of
the threshold
+ The non-large aggregate claims history does not then change
over time
+ If threshold is in line with excess point for excess of loss
reinsurance, then reinsurance IBNR can be identified more
easily
Might be harder for systems to extract the excess over the
threshold

(iv)

Page 6

Depending on exactly how the large claims have been extracted, there may
be no explicit allowance for non-large claims to become large
There is no allowance for unreported large claims
and the large claims listing clearly shows that some large claims do not
become so until 2 or 3 years after the accident date
There is no allowance for development of existing large claims

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

(v)

Although the case reserving may be stronger since the arrival of the new
claims handler, there is still potential for case reserves to increase
eg due to unforeseen deterioration in medical condition
but there is also evidence of case reserve reduction in later years as some
claims are settled favourably
Chain ladder methodologies on the non-large claims may not be
appropriate for the more recent years of account
As development factors applied may be quite large
And no use is made of exposure information such as premiums or vehicle
years
There is no indexation of large losses
General comments about the disadvantages of the Standard Chain Ladder
method

Chain ladder methods less likely to be as reliable as claims handling


practices, and hence shape of development curve, have changed
Reasonable to take ultimates on 19982001 as current incurred
as little evidence to suggest that there is development after year 4
.=> 5505 total ultimate for 19982001
For the 2000 and 2001 accident years, calculate the % developed at
development year 3 over current incurred: (935+2665)/(963+2740) = 97%
Apply this to the 2002 total developed at year 3 to estimate unltimate
incurred as 1204/0.97 = 1241

Pure IBNR

Need to allow for pure IBNR i.e. large losses not on the list but which
have occurred prior to 31 December 2004
Use an average frequency average cost approach
No details about premium volumes/size of account over time so assume
stable
Calculate current average claim numbers and average cost each accident
year correct numbers below give
From table, large losses appear to be notified within 3 years of start of
accident year
Therefore establish average number of claims in 19982002 years as fully
developed in terms of number of claims
= 3.6
But allow for fact that large loss definition has not been inflation adjusted
=> round up to 4
2003 and 2004 notifications to date do not appear out of line with this total
annual number of claims
From table, large losses appear to be mostly developed within 4 years of
start of accident year
Therefore establish average cost of claims in each of 19982001 years as
fairly fully developed in terms of cost of claims correct numbers below
give

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

Large claim definition has not been inflation-adjusted so need to allow for
effect of inflation
E.g. at 7% for motor (or something similar)
correct numbers below give
Gives average of 599 in 2004 terms
Multiply by 4 claims to get est for 2003 and 2004
correct numbers below give

Reserve calculation

Total ultimates of 11538. Total paid of 3339. Reserve of 8,199,000


7%
Inflation
adj

Acc Yr

No.

Av Cost

1998
1999
2000
2001
2002
2003
2004

4
2
5
3
4
2
1

183.3
534.5
192.6
913.3
301.0
165.0
400.0

Avg
Sel

3.6
4

455.9

Future
no.

1.6
2.6

Avg Cost

Ult

275.0
749.7
252.5
1118.9
344.6
176.6
400.0

733
1069
963
2740
1241
2396
2396
11538

4 599
4 599

599.0
599

This question was also generally not well answered. The level od detail given by most
candidates in the main part of their solutions fell well short of what was expected to
gain sufficient marks to pass.

(i)

Quite possible that the data does not yet exist to do this accurately
Should examine the data available and identify gaps for further review
Impact split between change in value of treaty, and change in value of
underlying
There is a requirement to allow for year on year changes in expense allocation
There is a requirement to allow for year on year changes in cost of capital
Change in value made up of price and T&C
For treaty itself T&C changes could include:

Page 8

Current incurred
Current incurred
Current incurred
Current incurred
Incurred / 0.97

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

For non-proportional
Per claim limit
Feed from pricing model or apply increased limit factors from external data
Per claim excess
Feed from pricing model or apply increased limit factors from external data
Aggregate limit
Needs explicit pricing stochastic model ideal stress test if not priced
stochastically
Aggregate deductible
Needs explicit pricing stochastic model ideal stress test if not priced
stochastically
Per event limit
Approximate adjustment based on historical catastrophe experience
Types of claim covered
Market data apply external benchmarks for new/discarded heads of claim
Territory covered
Market data apply external benchmarks for new/old territories
Term of policy
Longer inflation adjustment
Inflation clauses (e.g. severe inflation clause)
Can adjust value compared to inflation assumptions actually made
Profit commission
Calculate explicitly based upon expected results + stress test
Brokerage
Calculate explicitly as discount/load to price
Risks attaching/Losses occurring nature
Calculate approximate change in earnings profile to assess change
Reinstatement terms
Calculate explicitly using pricing model assumptions
For Proportional
Profit commission
Calculate explicitly as discount/load to price
Over-ride commission
Calculate explicitly as discount/load to price
Brokerage
Calculate explicitly as discount/load to price
Classes covered
Market data external benchmarks review expected profitability of
added/lost underlying business.
For the underlying business, changes could include
Underlying exposure volumes
Assuming price per exposure, apply multiplicative factor
Pricing of underlying direct business

Page 9

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

Depends on how it feeds through if treaty price is % of underlying then


multiplicative, if fixed amount then no change. If combination of both then
submission needs to be examined
Types of business underlying
Market data apply external benchmarks to adjust for new/discarded
business
Policy wordings / exclusions underlying
Approximate adjustment qualitative view based on change in terms if
removing head of claim (e.g. asbestos exclusion) this may be possible to do
more accurately
Territories
Market data apply external benchmarks to adjust for new/old territories
Deductibles
Feed from pricing model using increased limit factors to reflect change in
risk
Exposure measure (e.g. payroll/turnover switch)
Approximate adjustment qualitative view of change in underlying price
Other issues will include:
Look at price as rate per exposure rather than just premium amount
Consider expected changes in frequency and severity (claims inflation)
Allowance for year on year tax changes
Adjustment for investment income depending upon how the product is priced
Method
Group contracts into broadly homogeneous categories
Need to consider materiality want usefully large groups, not every contract
Consistency with reserving groups feedback from actuarial control cycle
For each contract need to quantify overall change in value to company
Use solutions suggested above, or other similar appropriate to calculate impact
of T&C changes
Request that underwriters record their own estimate of changes at point of
underwriting
Can calculate alternative index of these changes as a comparison
Importance of high level common sense checks
Consider changes in total 1st loss rate on line for non-proportional book, for
instance
Check versus business plan for degree of consistency or otherwise
A full statistical model may not be practical given the monthly reporting
requirement
Data
Details of all the above for own treaty from underwriting system
Ensure that underwriting system is capturing underlying changes from broker
submissions in future
Market data may assist in calculating simple impacts
e.g. RAA data
e.g. Increased Limit Factor tables

Page 10

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

e.g. Relative performance of direct business in different territories


e.g. Market pricing indices for direct business brokers, ABI, Lloyds etc
Assumptions
In the absence of information about the underlying business on individual
treaties
then the market T&C changes are a good guide to average movements.
Changes in claim severity (inflation +)
Changes in claim frequency (possibly inflation driven on excess of loss)
Limitations
Indicative answers only because:
Data collection may be inadequate
Insurance business is volatile price increase is no guarantee of profit
increase
Assumptions may not hold, especially inflation
Heavily reliant on information provided by insureds
Indicies by line of business will not reflect client profitability (if there are
cross subsidies)
Danger of spurious accuracy
(ii)

Other types of underwriting/claims MI could include:


Premium written per month
Number of policies written per month
Deviations on income v expected by contract
Large claims advised
Large claim movements in the month
Any claim notified against large exposures
Top exposures written during month
Actual v expected claims development by class
Impact on reserving of monthly movements in claims
Concentration of risk exposure by territory/class etc
Total claims notified by class, territory etc.
External market information updates
Quotation conversion rate
Business by source

(iii)

Advantages of RPP approach


Keeps a client happy strengthens relationship
Additional income for company (as policy not previously written)
Artificially inflates premium further still book reinstatement premiums as
well as additional premium
this is because a reinstatement premium protection recovery is a claim on a
policy and not a premium refund so there are equal increases in claims and
premium produced

Page 11

Subject SA3 (General Insurance Specialist Applications) September 2005 Examiners Report

Profit (net of any additional attritional costs) remains the same with either
approach
Disadvantages of RPP approach
Increases exposure to the same risks for the company which may lead to
reinsurance exhaustion
May not be authorised strictly not reinsurance no direct insurance
contract to protect
Second contract cost of issuing lots of contracts where one would do
May not be covered under companys outwards reinsurance programme (as
not reinsurance)
Reinsurance cost may actually increase
Company may not have the choice of share of RPP as the policies may be
broked separately
(iv)

RPP cover responds to the same losses as the original policy.


Cover also responds in same proportion a half limit loss triggers a half
reinstatement.
Structure of RPP contract mirrors original too therefore risk cost should
follow same proportions.
Limit of policy is 250,000 ( of original).
Theoretical risk premium would be 250000/1000000*250000*0.8=50000

END OF EXAMINERS REPORT

Page 12

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
3 April 2006 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 A2006

Faculty of Actuaries
Institute of Actuaries

Packit is a mutual insurance company incorporated in the UK and owned by importers


who are retailers and wholesalers based in the UK. The importers have their goods
shipped from around the world by reputable carriers. The carriers choose the most
suitable transit methods for the goods by land, sea and air. The importers buy an all
risks cargo policy from the point at which they purchase the goods to the point at
which they enter their stores and warehouses in the UK. Packit was formed when
importers felt that insurance rates were too high, and that they could save significant
costs through a mutual. Packit underwrites all proposals prior to writing the business.
Although there is no restriction on what goods can be insured, in practice they are
exclusively high street non-perishable goods including clothing, electronics and white
electrical.
Recently there has been a major catastrophic loss at a large distribution hub resulting
in losses to warehoused and containerised goods largely as a result of water damage.
The size of the loss has exhausted Packit s reinsurance programme resulting in a large
net loss.
Following this event Packit has asked for a comprehensive review of its operations.
You are the consulting actuary advising Packit.
(i)

State how the various UK tax rules apply to Packit and the policies it sells. [3]

(ii)

Describe how the current loss should be recovered from the member
companies and steps which may be taken to reduce the size of loss.

(iii)

[16]

(a)

State the desirable characteristics of rating factors.

(b)

List the main risk factors which are likely to be used for rating this type
of business.

(c)

Describe why your chosen factors are likely to be the main factors.

(d)

Explain why some of your risk factors will not be used as rating factors.
[14]

(iv)

Describe the different ways in which differing policy deductibles and limits
requested by the different members could be priced equitably.
[4]

(v)

Describe the considerations that Packit needs to evaluate when designing next
year s reinsurance programme.
[12]

Based on recent studies you have found that the expected cost of claims to Packit is a
function of the maximum sum at risk and the selected attachment point. The
percentage of claim cost below the attachment point
= 39.81 * (attachment point * 100 / maximum sum at risk)0.2.

SA3 A2006 2

You are pricing a quote. The importer wants to reduce its insurance cost and thinks
that a total loss is unlikely. It has therefore asked for a policy which protects it for
40% in excess of 10% of its maximum exposed values.
(vi)

(a)

Calculate the percentage of claim cost retained by the importer.

(b)

Comment on the suitability of the importer s policy request.


[9]
[Total 58]

Payfast is a general insurance company writing only payment protection insurance.


Payment protection insurance protects a borrower s ability to maintain loan
repayments should he/she be unable to keep up his/her repayments due to accident,
sickness or unemployment. Payfast writes its business through retailers selling
furniture and electrical goods via various credit schemes.
For each payment protection plan taken out, the customer pays a one-off premium,
typically about 10% of the amount borrowed. In return the insurer will cover the
customer s monthly loan repayments:
for the period of the remaining loan term while the customer is hospitalised; and
for a maximum of 12 months in the event of unemployment or disability
The majority of the payment protection plans cover level monthly repayments.
Payfast pays a commission to the retailer of between 30% and 60% of the gross
premium paid by the customer. The level of commission varies by retailer. The terms
of the credit schemes generally range from 3 to 5 years.
Payfast s earned premiums are based on the assumption that risk reduces uniformly
during the term of the payment protection policy. Premiums earned in 2005 totalled
11.3m gross of commission.
The company has been writing business since 1 January 2003. Paid claims have
totalled approximately 3%, 5% and 10% of gross earned premium in 2003, 2004 and
2005 respectively. Payfast currently books an outstanding claims, IBNR and claims
handling expense reserve of 2.0m as at 31 December 2005. It has shareholders funds
of 20.0m at that date.
From your market research you have established that:
a current expected ultimate loss ratio for an account of this nature is 20%, gross of
commission
IBNR claims at the end of an accident year are typically equal to 25% of the
ultimate claims for that year
claims handling expenses are typically 20% of claims payments; and
it is common for payment protection insurers not to establish case estimates

SA3 A2006 3

PLEASE TURN OVER

(i)

Describe the risks relating to premiums and claims that are faced by Payfast.
[5]

(ii)

Suggest steps that Payfast could take when underwriting the business in order
to mitigate some of these risks.
[6]

(iii)

(a)

Suggest the factors that would influence the earning pattern for each
individual policy.

(b)

Comment on Payfast s assumption that risk reduces uniformly during


the term of each individual policy.
[7]

(iv)

Describe how you would segregate Payfast s claims and policy data for the
purposes of monitoring profitability.
[6]

(v)

Comment on the reasonableness of Payfast s outstanding claims, IBNR and


claims handling expense reserve of 2.0m as at 31 December 2005.
[10]

(vi)

Suggest, with approximate calculations, how you might expect Payfast s


shareholders funds to change between 31 December 2005 and 31 December
2006 assuming that the company stopped writing business on 31 December
2005.
[8]
[Total 42]

END OF PAPER

SA3 A2006 4

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
April 2006

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT
Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

M Flaherty
Chairman of the Board of Examiners
June 2006

Comments
Individual comments are shown after each part question.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications)

(i)

April 2006

Examiners Report

Underwriting losses and profits are regarded as arising from mutual trading
and hence are exempt from tax.
The investment return is taxed independently.
The return from loan relationships will be taxed as income on a mark-tomarket basis, unless accounts used amortised cost.
However, with respect to equities, realised investment gains are subject to
capital gains rules and hence indexation relief applies.
No relief is given for expenses, which are assumed to be part of the mutual
trade.
VAT is not payable on insurance premiums
Insurance for commercial ships and aircraft, commercial goods in international
transit and risks located outside the UK are exempt from IPT.

Comments on question 1(i): Generally well answered although many candidates did
not realise it is the mutual trading which is the reason for no taxes on profit and also
many got IPT wrong.
(ii)

The rules may define what happens in these circumstances and therefore leave
little scope for flexibility. The rules of the mutual must be followed.
There may be a request for a special premium (effectively a capital injection)
or an explicit capital injection.
This may require that any special premium be calculated in proportion to the
capital provided, or to the insurance premiums paid in the last year.
If the loss is recovered by higher premiums in future then this may be by
adding a larger profit margin (percentage) to the premiums charged.
Or by explicitly adding an amount in respect of the loss.
The insurance premiums charged should take account of the post loss
commercial premiums. The mutual may have the opportunity to recover its
losses through higher premiums whilst still offering good value to the
members.
If the mutual were to charge much more than commercially available then
member companies would have a good incentive to purchase elsewhere
although financial obligations to the losses would remain.
In the long run this may result in the mutual being wound up.

Page 2

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Some of the goods may be salvaged


This is the only realistic way of mitigating the loss
Clothing can be washed, ironed and re-packed.
None of the electrical or electronic goods will have been powered when the
water damage occurred and so some may be recoverable.
Water damaged goods could be moved quickly and dried quickly to avoid
additional damage being caused.
Retail packaging and instructions would be seriously damaged and would need
replacing.
These steps may be quite costly and require using specialist companies but
could significantly reduce the size of the gross loss.
As the loss has exhausted the RI programme any reduction in gross loss will
be beneficial to the mutual.
Once the gross loss is within the RI programme then the benefit to the mutual
will be a percentage of the gross saving.
The mutual will save on its payment of reinstatement premiums
or, if it retained some of the programme then it will save on this portion.
The remaining benefit will be taken by the reinsurers.
Even so the mutual should benefit in the cost of future reinsurance, as the
current loss will be reduced, and therefore have a smaller impact on future
pricing calculations.
This salvage is likely to be part of the reinsurance terms and conditions and
therefore something that the reinsurer insists upon.
Secondary ways of mitigating the loss size include being tough on claims.
However, as a mutual the aims of the insurer are slightly different and this
may be more difficult to do than with a proprietary insurer.
E.g. if a member company had not paid its premiums, it may be difficult to
void the coverage. (Non-payment of premium is not a valid reason for
voiding a claim under English law.)
Policies may be voided if specified policy warranties have been breached or
policy exclusions may reduce the insured loss.
Investigate precise cause of incident
seek to offset costs initiate
proceedings against culpable parties (candidates should be clear that this is
not Packit s property policies but warehouse owners whose policies may pay)

Page 3

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Investigate T&C of existing reinsurance coverage and ensure that all possible
recoveries under all policy sections are made.
A claim may be reduced due to under-declaration of values (similar to
averaging; although this is unlikely)
The claim may be recovered out of capital or other assets set aside for this
purpose.
Investigate potential Govt/State disaster compensation.
Comments on question 1(ii):
Many candidates did not mention that there would be rules governing the operation of
the mutual. These rules must be followed. Many candidates however suggested
sensible approaches.
Salvage was mentioned in passing but given the line of business and the nature of the
loss far more should have been said. A couple of different examples would have
shown that the candidate understood the problem and was tailoring his solution. This
was especially important given the number of marks for this part of the question.
Little mention was made of the fact that the claim had exhausted the reinsurance, and
the financial impact this would have on the mutual if the gross loss could be reduced.
(iii)

(a)

Rating factors should define the risk, i.e. be a proxy for the risk factor
not correlate too closely with other rating factors
and that they are practical (simple)
Objective
Easily measurable.
Acceptable to the policyholder
Verifiable (desirable but not essential)
Each additional rating factor should, therefore, be chosen to remove as
much of the residual heterogeneity as possible.
This approach should also help to avoid having too many rating factors
and so cause practical problems due to lack of data for analysis of each
cell.

(b) & (c)


With respect to the importers the loss frequency and severity will be
affected by the type of goods: clothing, electronic or white goods.
These may even be subdivided into smaller sub-categories.
All these goods are subject to the same major perils: fire, water
damage, theft and physical damage due to impact or crushing.

Page 4

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Different goods behave differently to these different perils and so the


severity of loss by type and peril will be different.
Commodity / Peril example 1
mobile phones will be more likely to
be stolen than washing machines or Clothing will be much less
susceptible to impact and crushing damage.
Commodity / Peril example 2
TVs may be more combustible than
cotton jeans or Dried foodstuffs will be very susceptible to water
damage.
Country of origin.
Whether the transport is primarily by air or sea.
Rating factor example 1
Air crashes are much less common than
ships sinking so the frequency of loss will be lower for air freight. Or
Airports tend to be subject to higher levels of security than sea ports
and so theft risk will be lower.
Rating factor example 2
But, given a loss an aircraft will be more
likely to suffer total loss so severity higher.
The Shipping Company may be used as a risk/ rating factor if this is
known.
The total value of goods at risk in the year.
This is a measure of the exposure to the policy and is a proxy for the
number of transits which take place.
The maximum single sum at risk is another exposure measure and
defines the largest loss which could occur.
A larger sum at risk has the possibility of a larger loss than a smaller
sum at risk.
The amount of policy deductible.
For a given set of circumstances a higher policy deductible will give
rise to a smaller claim size than a smaller deductible.
The points of largest risk will probably be during loading and
unloading. Once an item is on a ship or aircraft the theft and damage
risks are small, so distance travelled is likely to have a smallish effect.
(Distance could be a rating factor as it has an impact on risk and may
be a proxy for the number of trans-shipments).

Page 5

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

The number of times a cargo is moved between storage and a method


of transport will be an important risk factor but is not a practical rating
factor. It is almost impossible to determine in advance.
The Shipping Company may not be used as a risk / rating factor if this
is unknown.
The security of the transit locations could be important but very
difficult to measure as the goods owners will have little say in the
transit route, that being controlled by the shippers. So not a practical
rating factor.
Comments on question 1(iii): The bookwork part of this section was well answered
but the application and higher skills parts were poor.
(iv)

Policy limits and deductibles will be best dealt with by either an exposure
measure or a frequency / severity model.
Exposure based methods would be either an exposure curve or increased limit
factors.
Exposure curves describe the claim severity distribution:
this could be a loss size distribution
or may express the deductible as a percentage of the sum insured (i.e. first
loss curves)
This claim size or percentage can then be read from the curve to give the
proportion of claims cost which is retained within the deductible.
Increased limit factors work in a similar way but the limit and deductible are
looked up as factors. The factors at these amounts are then used to calculate
the proportion of claim cost retained by the deductible. (Mention of Limited
Expected Values score here.)
There may be more than one claim distribution needed to fully describe the
observed claims and therefore to calculate the equitable portion of claims
within the deductible.
Experience methods on observed historic claims for the cedant can be used but
are unlikely to give equitable answers as the observed claims are unlikely to
be a good representation of all the possible claims outcomes. (Burning cost
methods do not score as being equitable.)

Comments on question 1(iv): This question was exploring the higher skills and wider
reading of candidates. Candidates generally scored low marks on this section with
very few referring to exposure based methods (either exposure curves or increased
limit factors) or frequency/severity models.

Page 6

Subject SA3 (General Insurance Specialist Applications)

(v)

April 2006

Examiners Report

Considerations:
Cost of cover
following a large cat loss the cost of RI is likely to go up
following the exhaustion of the RI programme the cost will go up even for
high layers which were not purchased previously
can the mutual afford to buy the RI it would like to?
can the mutual afford not to buy and run the risk net?
Expected recoveries
linked to the cost, the mutual will want value for money
these will be evaluated by amount using both frequency and severity
May need to reparameterise frequency severity distributions following the
experienced losses
for accurate analysis.
Alternatives to traditional reinsurance
development covers etc.

ART, financial assistance, loss

Availability of cover
what cover is available following the cat
in terms of capacity (amount)
capacity may be severely restricted following a large cat
and in terms of coverage
Likelihood of a similar event
if the last event was seen as a remote event e.g. 1/10,000 years then the
mutual may not want to buy the cover
this will be location and peril specific
Exhaustion of the current programme
Vertical exhaustion
by how much?
wanting to ensure that a future loss does not exhaust the programme
Sideways cover
The number of reinstatements needed for multiple events in the same
policy period.
Level of exposure for the following year

sums insured

Maximum accumulations of value next year


per location and
geographically
although this may be hard to determine when the mutual has little
knowledge of the shipping details
Risk appetites of the member companies
more risk averse will want to buy more protection or vice versa
Rating agencies security status/rating of available reinsurers

Page 7

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Diversification of reinsurers (reciprocity does not apply for a mutual)


Relationships with reinsurers
Advice through purchasing reinsurance (most likely the advice will come from
the broker)
Reinsurance used by other mutuals (e.g. cover for a group of mutuals)
Regulatory requirements
Amount of capital
free reserves - could raise loan capital & reduce the
need for RI
Comments on question 1(v): Many candidates failed to pick up marks by not
exploring a wide enough range of options.
(vi)

Percentage of claims cost below deductible = 39.81*(10% *100)0.2


Percentage = 63.09
Percentage of claims cost below the limit = 39.81*(50%*100)0.2
Percentage = 87.05
So the percentage retained is the deductible = 63.09
Plus the proportion above the limit = 100 87.05 = 12.95
Giving a total retained percentage of loss = 76.04%
It is unlikely that the importer would really want to retain this much of the
risk.
As the shareholders of the component companies will want a steady trading
profit from its core business rather than an uncertain profit driven by claims
which could be insured.
Even a 1% of maximum value deductible would result in nearly 40% of the
claim cost being retained. (Or similar calculation.)
The claims distribution is therefore very heavily weighted towards smaller
claims.
Only insuring up to 50% of the maximum value without deductible means that
87% of the claim cost is covered. (Or similar calculation.)
This is a high proportion of the total claim cost and could be justified. It
depends how frequent large losses are. If they are very infrequent and the
financial implications have been evaluated then the importer may feel that it is
worth the risk.
Even though the premium is higher than the expected claims due to expenses
the cover provided may be very beneficial to the importer due to the reduction
in volatility.

Page 8

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

If the mutual is inefficient, or another insurance company takes a different


view of risk then the insurance cost may be high and cheaper cover may be
available elsewhere.
Comments on question 1(vi): Most candidates scored well on the calculation,
although disappointingly a number failed to perform this simple task. A common
error was to misinterpret the 40% in excess of 10% layer as 40% of an unlimited
excess of 10% layer. Conclusions were not well drawn. Candidates often did not
notice that this loss distribution was very heavily weighted to smaller claims or that
the very high proportion of retained claim cost is likely to be unsuitable for an
importing company.

(i)

Risks relating to premiums


Policy is longer than annual in that premium charged at start of policy has
to cover risk of accident, sickness or unemployment over 3 to 5 year
period so harder to get premium rating right.
And Payit is a new company so the lack of data is especially problematic.
Downturn in economic environment .sales of goods reduce hence
volumes of payment protection insurance business reduce.
Change in retailers sales procedures may lead to loosening of
underwriting conditions
e.g. more selection/moral hazard: policyholders expecting
unemployment or deliberately becoming unemployed.
Retailers may take their business elsewhere or demand higher commission
or not sell enough policies.
Credit risk with failure of retailers to return premiums.
Greater scrutiny by regulators of levels of commission being charged =>
potential damage to reputation, reduction in market size as customers
choose not to insure themselves.

Page 9

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Risks relating to claims


Unexpected downturn in economic environment.
.leading to higher than expected unemployment rates.
Increases in morbidity experience
Pandemic
Propensity to claim
Reputational risk (media) may mean paying claims that you would
otherwise have excluded
Policy wordings not holding up in court e.g. unfair policy exclusions.
Comments on question 2(i): This largely bookwork question was answered fairly
well. A common shortfall was to simply say that moral hazard, or the economy were
risks. It is an unexpectedly high level of moral hazard, or an unexpected downturn in
the economy which are risks.
(ii)
Establish agreed sales procedures with retailers.
A void accumulations by retailer and or region.
Introduce profit commission terms to encourage retailers not to underwrite
poor risks.
Using the information gathered for credit scoring etc.
Demographic information of retailer impacting sickness
--- driving premium rates by store and region
Implement exclusions in policy wordings
Examples:
- Deliberate or wilful acts of self-injury.
- Pre-existing conditions.
- Acts which result in self injury
for example drinking alcohol or
drug abuse.
- Any mental or nervous conditions unless under the supervision of a
psychiatrist.
- Backache or related conditions without medical certification.
- Unemployment within the initial exclusion period.
- Voluntary unemployment.
- Unemployment known before the start of the policy.
- Casual, seasonal or temporary employment.
- No payment for any period where policyholder is paid salary in lieu of
notice.
- Loss of job through any fault of policyholder s.
- Or any other sensible exclusion e.g. waiting period.

Page 10

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Or refer special cases to insurer for underwriting.


Comments on question 2(ii): Most candidates got the basic ideas here but did not
adequately elaborate on the policy exclusions which are so important to this policy
(iii)
Nature of loan payments: If level, then the outstanding repayment amount
reduces uniformly with each payment hence risk reduces uniformly during
the term of the loan..
Nature of loan payments: If not level, (e.g. no repayments for a year or low
start repayments rising later ) Then the outstanding repayment amount
does not reduce uniformly and hence the risk does not reduce uniformly
during the term of the loan
Insurance terms: period over which payments will be made (e.g. nothing to
pay for the first year) ; are there limits? Waiting period?
Term of policies.
Morbidity rates rise as people age. Or Gender Or Age or Occupation.
Changes in state of the economy.
If underwriting is applied at the time of sale the risk will be lower initially
Selection: does experience indicate that selection occurs earlier in the
policy?
Propensity to claim: does this reduce as policy approaches expiry?
Unemployment and disability: makes sense to assume closer to level risk
profile as payment is limited to 12 months and hence reduction of
exposure would only occur in last 12 months of the policy.
Comments on question 2(iii): Many candidates did not seem to understand that it
was the repayments which were being guaranteed. Loans on this type of purchase
are normally by level repayment which does not change with the interest rate. In this
case the outstanding repayments and therefore risk reduce uniformly through the
policy term. Unemployment and disability pose a level risk for the majority of the
policy term. Identification of these key features allowed some candidates to score
well but many candidates did not seem to understand risk exposure.
(iv)
Monitor on monthly basis because company has only been writing for a
few years.
Segregate the policies by retailer as different commission levels.
Segregate by country or region
Split by any extra rating factor gathered at point of sale.

Page 11

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

Segregate policies by size of loan or type of credit scheme as some may


encourage more selection/moral hazard than others.
Analyse by policy term to estimate claims development and earnings
pattern.
Analyse sickness, accident and unemployment separately.
Compile accident year triangles of paid loss ratios to monitor profitability
on earned basis.
Compile underwriting year triangles of paid loss ratios to monitor
profitability on ultimate basis.
Comments on question 2(iv): Generally answered well.
(v)
Company s outstanding claims reserve is likely to be about 80% of the
2m claims and expense reserve => 1.6m.
This represents 14.2% of earned premium gross of commission.
Plus 10% paid loss ratio gives 24.2% ultimate: higher than market loss
ratio of 20%
It is not entirely clear whether the IBNR contains allowance for the claim
handling expense. If not then the claims handling reserve will also need
to be added & will make the ULR higher still.
Paid loss ratio in 2005 was 10%; allowing for 25% IBNR => chain
ladder ULR of 10% / 0.75 = 13.3%.
This is less than market loss ratio of 20%.
Company s outstanding claims reserve is higher than predicted by chain
ladder and by market average.
However market loss ratio may be based on portfolios with very different
commission levels.
Portfolio likely to be less mature than market average with higher IBNR as
percentage of ultimate.
Plus actual experience may be poor guide to IBNR.
Note than paid loss ratios have been increasing, which may suggest that
earning pattern is inappropriate i.e. perhaps earning premium too fast.
Lot of uncertainty because insufficient history on which to base
projections.

Page 12

Subject SA3 (General Insurance Specialist Applications)

April 2006

Examiners Report

On the available information it is difficult to comment on the


reasonableness of the outstanding claims reserve. However as Payfast is
estimating a ULR approximately 20% worse than the market it suggests
that they are not being unduly optimistic.
Comments on question 2(v): Most candidates used the market loss ratio as their
starting point. Very few candidates performed an independent calculation and then
used this as a basis for comparison to the market. Any reasonable comparison scored
well.
(vi)
No profits from new business.
Assuming commission of 50%, say, and claims (+ claims handling
expense) loss ratio of 20% => 30% profit (before other expenses) on
earned premium.
Premium earned in 2006 likely to be lower than that earned in 2005
as some of the 3 year policies written in 2003 would earn little in 2006
plus earning pattern reduces with term
so assume EP in 2006 is less than half that in 2005 => 5m.
Assuming 30% profit, this 5m will earn at least 1.5m of profit
which is taxable at say 30% (or any reasonable tax assumption)
Thus increasing shareholders funds by about 1m.
Any reasonable expenses assumption
Assumes no dividends.
Any reasonable assumption on investment returns on free assets
Also, any reasonable investment assumption on technical provisions
Comments on question 2(vi): Many candidates got bogged down in calculating the
earned premium in detail and a number of candidates did not make any attempt at
this section. A sensible estimate using the understanding of part (iii) was all that was
required. The remainder of the calculation was then straightforward. In estimating
the profit margin, many candidates forgot to include the commission terms. Given the
significant size of these this was a serious error.

END OF EXAMINERS REPORT

Page 13

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
11 September 2006 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 S2006

Faculty of Actuaries
Institute of Actuaries

You are the newly appointed actuary for a small UK general insurance company that
writes liability insurance through regional UK brokers and Lloyd s. The company is
newly established having started writing risks on 1 July 2005. The professional
indemnity class of business currently accounts for most of the business written by the
company.
One of your first tasks as the company s actuary is to estimate outstanding claims and
IBNR requirements as at 31 December 2005 for the purposes of the company s yearend accounts. The company accounts for its business on a one-year accounting basis.
In addition to the company s own detailed policy and claims information, you have
decided to obtain benchmark claims development and loss ratio information from
publicly available sources. You have collected paid, incurred and ultimate loss ratios
by year of account from the UK statutory returns submitted by 10 different insurance
companies. You have selected each of the 10 companies because they present a
professional indemnity classification within their returns.
Having collected these data, you realise that it may not be appropriate to include data
relating to all of the 10 companies within your benchmarks. Your company s
professional indemnity underwriter has agreed to assist you in identifying those
companies whose data should be used for benchmarking purposes.
(i)

Suggest two main reasons why UK statutory returns are more useful than
Companies Act accounts for deriving reserving benchmarks for your
company.

(ii)

[2]

Explain the distinguishing aspects of each professional indemnity account that


you would wish to investigate when selecting the most appropriate loss ratio
benchmarks for reserving purposes.
[18]

Following your discussions with the professional indemnity underwriter, you have
made your informed selection from the 10 benchmark accounts and you have
combined the information, weighted by volumes of business, to produce the following
summary table.
Accident
Year
2001
2002
2003
2004

2002 Returns
PLR
ILR
ULR
21% 88% 125%
5% 36%
91%

2003 Returns
PLR
ILR
ULR
43% 105% 123%
21%
64%
85%
7%
29%
70%

2004 Returns
PLR
ILR
ULR
85% 118% 122%
38%
76%
86%
23%
46%
65%
2%
23%
60%

Notes to table:
1.

PLR = paid loss ratio = paid claims at valuation date divided by earned
premium at valuation date

2.

ILR = incurred loss ratio = incurred claims at valuation date divided by earned
premium at valuation date

3.

ULR = ultimate loss ratio at valuation date

SA3 S2006

The underwriter has additionally provided you with his estimates of market premium
rate changes over the past 3 years and these are shown in the table below.
Underwriting Years
2001 2002
2002 2003
2003 2004
2004 2005

Premium Rate Change


+ 50%
+ 25%
+ 10%
5%

(iii)

Describe how you might use the information provided in the tables above to
estimate a benchmark ultimate loss ratio for the 2005 accident year, stating
additional information that you might require from the underwriter.
[14]

(iv)

(a)

Estimate average paid and incurred claims cumulative development


percentages for an accident year at the end of the first, second, third
and fourth development years.

(b)

Comment briefly on the reliability of your results.


[10]

(v)

Suggest, with reasons, the extent to which you would expect to use your
company s own claims data versus benchmark information for reserve reviews
as at 31 December 2005 and 31 December 2006.
[14]
[Total 58]

You are an actuary working for a large UK general insurance company.


A motor manufacturer has approached your company to underwrite a new scheme that
provides gap insurance cover to purchasers of new vehicles sold by its network of
dealers. Under the terms of this policy, the insurance company will pay out the
difference between the original purchase price and the amount paid out by the
insured s own private car insurance policy in all events of the vehicle being written
off. The insurance cover is for the first three years of vehicle ownership, provided the
vehicle remains under the ownership of the original vehicle purchaser. The
manufacturer has expressed a preference that all purchasers of vehicles in the UK pay
the same flat monthly premium for the cover. The premium will be payable for a
period of 12 months from vehicle purchase.
(i)

Discuss the advantages and disadvantages to the new vehicle purchaser of this
policy.
[6]

(ii)

Describe risks that are likely to exist in this scheme for the insurer, suggesting
ways in which these risks can be mitigated.
[16]

(iii)

Discuss the data you will require to calculate the risk premium, including
possible sources for this information.
[14]

(iv)

State the other data or information you would consider in deciding whether or
not to underwrite the scheme.
[6]
[Total 42]

END OF PAPER
SA3 S2006

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
September 2006

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT
Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

M Stocker
Chairman of the Board of Examiners
November 2006

Comments
Individual comments are shown after each part question.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

(i)

Companies Act accounts do not show results by class of business whereas


UK statutory returns show results by class and risk group.
Companies Act accounts show figures by revenue year whereas statutory
returns are divided by underwriting year or accident year.

These points are contained within the core reading, although surprisingly few
candidates gained full marks on this bookwork question. A number of
candidates made incorrect comments on the different levels of prudence within
the outstanding claims and IBNR reserves between the accounts and the
returns.
(ii)

Mix of business by profession


The ultimate loss ratios (ULRs) may differ between, for example,
solicitors, architects, brokers, actuaries, accountants, IFAs.
This may be due to different levels of competition in these different
sectors
due to different levels of risk appetite e.g. IFA professional indemnity
experience has been poor in recent years due to pensions mis-selling.
Some insurers may concentrate on only one profession
.e.g. writing solicitors business through a lineslip.
And others may exclude certain professions
.e.g. no big-4 accountancy firms.
Some professions may incept at certain times of the year
e.g. solicitors business incepting in September/October
which is particularly relevant to consider for a company that has only been
writing since 1 July.
Size (turnover or number of partners) of assureds
Large practices may have very different claims experience to smaller
practices,
Which is particularly true for sole practitioners.
Large practices usually have higher fees and larger potential losses.
Different size assureds are likely to have different risk management
standards.
Different assureds have different clients and therefore different services
are provided (some are more risky).
Territory in which business written
Lloyds business may be different in nature to UK regional e.g. due to size
of risks.
Lloyds business may include non-UK risks.
US professional indemnity business in particular may perform differently
to UK (e.g. due to litigation).
Currency issues need to be considered.
Source of business
Different coverages or exclusions

Page 2

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

Size and type of lines written


Some insurers may write mainly primary layers whereas others write
excess layers.
Limits and deductible may differ.
Claims made vs claims occurring policy
Business written on a claims made basis may be different in nature to that
written on a claim occurring basis
e.g. volatility of ULRs due to dependency on legislative changes.
Different underwriting philosophy
Some insurers trying to build market share may be happier to write at a
higher loss ratio.
Underwriter may have views on quality of underwriting done by
competitors.
Different rate change history
Different reserving philosophy
Drivers for different reserving strengths may be different case reserving
practices/philosophies.
Some companies ultimates may show a systematic downwards trend in
ULRs over time for a given year of account.
Underwriting year vs Accident Year
If written on a funded basis, business will appear in statutory returns
classified by underwriting year.
If written on a one-year basis, business will appear in statutory returns
classified by accident year.
We would prefer to use accident year benchmarks as this is consistent with
companys one year accounting.
Although 2005 accident year ULRs could perhaps be estimated by
averaging 2004 and 2005 underwriting year ULRs.
Impact of exceptional/large claims that the underwriter would know about.
Size of account
Size of account impacts credibility for benchmarking purposes.
Smaller accounts may be more volatile in terms of claims experience.
Maturity of account
A growing account may have different characteristics to a stable one.
New accounts are likely to have less historical data on which to base ULR
projections that more established accounts.
This question was reasonably well answered, with candidates able to generate
a wide range of points. The stronger candidates were able to tailor their
answer to give Professional Indemnity specific points. Some candidates
appeared to be unrealistic about the information that they would expect the

Page 3

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

company actuary and/or underwriter to know about the benchmark


companies. The question asked about benchmarking for loss ratios and not
for development patterns, although some answers focused on the impact on the
latter. Most candidates did not explain why different claims characteristics
for different professions would automatically mean different ultimate loss
ratios.
(iii)

Starting point is to make a choice about which base years to use.


Professional indemnity is reasonably long-tailed so that would lead us to
place less reliance on ULRs on immature years.
At the end of the first development year, the incurred claims seem only to
be about 40% of ultimate claims.
So suggest not relying too heavily on 2004 booked ultimate loss ratio.
Historically, it can be seen that the ULR for a particular accident year can
drift over time.
It may be that companies book a pessimistic ULR initially to avoid poor
run-off.
For example, the 2002 accident year ULR has reduced over time.
2001 may be too old to be representative[the examiners accepted other
sensible comments about the reducing relevance of older years]
terms and conditions are likely to have changed alongside the premium
rate changes
and there may have been a shift in the mix of business with new
capacity entering the market.
Establish from underwriter whether there are any large losses/events that
might have distorted any of the figures in the market information.
May use 2002 and/or 2003 accident years as starting points. [or other
sensible conclusion]
Roll forward for premium rate changes
Definitely need to adjust for rate changes as they havent been flat and will
therefore have a big impact.
Need to understand from underwriter more about what the premium rate
change information represents:
.before or after allowing for claims inflation?
.and what would a typical claims inflation assumption be?
.before or after allowing for exposure (e.g. assured fees) inflation?
.has allowance for changes in terms and conditions e.g. reduced
coverage been made in the rate change information?
One approach is to estimate the earned premium rate change between year
x and x + 1 as the average of the underwriting year rate changes between x
-1 and x + 1.
Although we would need to discuss with underwriter how business incepts
and earns throughout the year.
We may need to make an adjustment for changes in the environment,
target market, longer term trends e.g. legislation, impact of the insurance
cycle etc

Page 4

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

The table below is an example of how the ULRs shown in the 2004
returns may be rolled forward to 2005 terms, assuming premium rate
changes include exposure inflation.
Accident
Years
2002
2003
2004
2005

Starting
ULR
86%
65%
60%

Earned Rate
Change From
Previous Year
NA
38%
18%
3%

Inflation
Assumption
8%
8%
8%
8%

ULR in
2005
Terms
65%
63%
63%

This uses the most recent ULRs (from the 2004 returns) although these
may be prudent for the more recent accident years
A sensible claims inflation assumption might be e.g. 5% to 10%.
Select the average of the 2002 and 2003 ULRs => 64% as 2004 too
immature to rely on 2004 ULR at this stage (or other sensible selection and
justification).
We would need to understand the treatment of commission: is premium
gross or net of commission in the statutory returns?

There was a wide variation in the quality of answers for this question.
Although the question did not specifically ask for ULR calculations, the best
descriptions were those that used the data provided in the question; this
enabled students to demonstrate the various steps required more clearly.
Some candidates appeared not to have learnt from recent SA3 questions on
calculating ULRs. This is worrying given the practical use of this technique
and the importance of being able to sense-check loss ratios against the
backdrop of information on premium rate changes and claims inflation. Good
candidates were able to use the claims development information provided in
the question in order to make a judgement about which base years to use.
Some candidates went further than this part question intended and tried to
estimate the effect on the loss ratio of the company having only written
business since 1 July 2005. In these cases, the examiners gave some credit
accordingly under Q1(v).
(iv)

Divide PLR by ULR to get sample % developed from data on 2001 to


2004 accident years. Use the ULRs from the same returns as the PLRs in
order to avoid distortions from changes in earned premium over time.
Average development %s across accident years, shown in the tables
below

[Note: an alternative approach to the above calculations is to divide the PLR or ILR
by the latest estimate of the ULR in each case, rather than the ULR estimated at
previous year-ends. This approach assumes that the earned premium in the
denominator is consistent over time (i.e. no late bookings or misstatements of
premium). This alternative approach was given full credit by the examiners only if
that assumption was stated.]

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

Paid cumulative development percentages


Accident
Year

Development Year
1

2001

16.8%

35.0%

69.7%

44.2%

2002

5.5%

24.7%

2003

10.0%

35.4%

2004

3.3%

Average:

6.3%

25.6%

39.6%

69.7%

Incurred cumulative development percentages


Accident
Year

Development Year
1

2001

70.4%

85.4%

96.7%

88.4%

2002

39.6%

75.3%

2003

41.4%

70.8%

2004

38.3%

Average:

39.8%

72.2%

86.9%

96.7%

Comments on reliability of results

Development year 4 position is based on only one sample point (relating to


2001) so this is likely to be less reliable.
For incurred claims, there seems to be a reasonable amount of consistency
between sample points at the same development period => reliable.
For paid claims, there appears to be some evidence of claims speeding up:
for second development year, % developed goes up from 16.8% on
2001 accident year to 24.7% on 2002 and 35.4% on 2003.
for third developments year, % developed goes up from 35.0% on 2001
accident year to 44.2% on 2002.
Using averages wont reflect any such trends.
May prefer to use incurred patterns rather than paid patterns (especially
as incurred is more mature).
However incurred claim development percentages depend on consistency
of case estimate strength.
The existence of the tail factor in the projections increases uncertainty.

The numerical parts of this question were not very difficult and a lot of marks
were available for good quality comments on the results. There were many
candidates, however, who failed to gain many marks on the numerical parts or
who were not able to make observations on key features of the development

Page 6

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

patterns. Some candidates tried to adjust the data for premium rate changes,
which was unnecessary. Some candidates gave development of loss ratios
which was not requested.
(v)

Generally
One key factor is speed with which claims are expected to develop.
Would expect to move gradually from benchmarks to companys own data
as time goes on.
Perhaps using a Bornhuetter-Ferguson type of approach.
Likely to rely on incurred claims data before paid as paid will be too
immature initially.
Depends on confidence in own case estimation.
Unlikely to rely on paid chain ladder for the first three development years
as less than 40% developed.
Benchmark claim development patterns suggest that incurred claims are
reasonably mature by the end of the second development year and hence it
likely that at least some reliance would be placed on chain ladder from this
point.
Likely to recognise differences between company and market experience
earlier if company experience appears worse than markets.
Would also treat very large losses separately.
As at 31 December 2005
Only need to calculate IBNR in respect of 2005 accident year.
May want to calculate 2006 ULR to check UPR is sufficient to cover
unexpired risk.
2005 is not a typical accident year as business only written from 1 July
2005.
even if business written evenly between 1 July and 31 December, earnings
will not be even over the second half of 2005.
therefore average accident date likely to be biased towards November
rather than mid-year.
Our 2005 accident year is even less mature own claims experience
multiplied by market development factors will underestimate ultimate.
Even if use Bornhuetter-Ferguson, far bigger weighting would go to
independent (market) ultimate loss ratio than to chain ladder.
So rely heavily on market ultimate loss ratio as claims to date of little use
(except if the account suffers large losses or very bad experience).
As at 31 December 2006
2006 accident year will still not have an even earning pattern as missing
accidents relating to business written in the first half of 2005
.plus likely that business volumes still increasing over 2006 as company
establishes itself
Therefore average accident date is likely to be biased towards second half
of year.
And it will not be easy to compare actual claims developments on 2005
and 2006 accident years with those from the benchmark data.

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

So likely to give significant weight to market loss ratio on both accident


years.

The strongest candidates understood that both the accident years would be
more immature than a typical accident year because no policies were written
in the first half of the first underwriting year. A number of candidates talked
at length about the validity of benchmarks but did not appreciate that there
was little alternative but to rely on them.

(i)

Advantages:
When an accident involving a total loss of the vehicle occurs, road risk
insurance companies will usually pay out the actual value of the vehicle.
This will usually be considerably less than the original purchase price of
the vehicle due to market value depreciation within the UK motor sales
market.
Depreciation of the vehicle market value will be particularly steep in the
first three years of a vehicles life before slowing down thereafter.
Vehicle purchasers will have financial peace of mind in the knowledge
that they will be covered in the event of accidents where they were not to
blame or driving at the time (e.g. theft of the vehicle) or involved complete
or irreparable damage to the vehicle (e.g. fire, total write-off of vehicle in a
road accident).
Road risk car insurers will not take into consideration any outstanding loan
on the vehicle purchase.
Holding gap insurance cover will assist the policyholder in paying off any
outstanding amount of the loan rather than be in a position of negative
equity.
The policy will pay out in full if the vehicle is written off by an untraced
driver or failure to pay by the owners insurance cover for another reason.
If the cover is free then this is good for the purchaser, but the cost may
already be included in the cost of the vehicle.
Premiums are known in advance in terms of level and period of payment
so will assist in financial planning (or could be regarded as an additional
monthly element to budget for in addition to road risk premium and loan
repayment).
No need to argue with the motor insurer on the value of the car.
Disadvantages:
The rates charged for gap insurance are often high.
The cover may be of limited use to a car purchaser who changes his car
frequently, e.g. once a year.
Value depends on risk appetite of car purchaser: he may not wish to pay
for something considered unlikely.
No benefit is provided if there is a bad accident but the car is not written
off.
The customer may not want to pay three years cover in the first year.
There is no value in the first year if the motor insurer gives new for old
cover in the first year

Page 8

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

The amount recovered in total from the gap policy and private car
insurance would not be sufficient to replace the car with an equivalent new
model due to price inflation.

This question was well answered by many candidates.


(ii)

Risks and mitigants


The main risk is that the premium rates charged under-estimate the risk.
This is a particular problem here as the premium assumptions are based on
information which is difficult to get
..and the premium needs to cover 3 years of exposure.
The lack of rating factors suggests possible anti-selection could occur on
this scheme, particularly if the cover is optional.
Mitigate by stricter underwriting.
For example, different age groups are more/less prone to total vehicle
write-off accidental damage claims so should be charged more than those
age groups which are have a lower total loss frequency.
Similarly different makes of vehicle have different depreciation levels
resulting in very different claim severity.
Similarly theft claim frequency is very different by postcode so possibly
more likely to pick up business only in areas where there is a higher theft
frequency. Is the location of the network of dealers biased towards high
theft areas?
Claim size and variance of size increases over the three year period of the
policy.
Depreciation of the vehicle value is also based on mileage. The more
mileage driven, the lower the market value and hence lower settlement
from road risk insurer.
Mitigate by having more sophisticated rating structure with rating factors
similar to those used in road risk insurance
and have the manufacturer pay the theoretical premiums.
Monitor frequencies and severities and adjust future premiums as
necessary.
Potential accumulation of risk with existing motor portfolio where the
company insures both the road risk and gap insurance element when a
significant event occurs (e.g. localised flood involving write off of many
flooded vehicles).
Mitigate by aiming for more diverse portfolio in terms of geographical
location of road risk and gap insureds.
Potential for moral hazard if insured is less likely to look after vehicle
knowing that he/she is covered for the full amount of the purchase price
for the first three years of the vehicle ownership.
Potential for fraud if insured decides to write off vehicle.
Risk of not receiving the whole annual premium during the first year of the
policy if total loss occurs and policyholder cancels the policy after
settlement. or car is resold.
Mitigate by collecting premium up front or over fewer months rather than
spread over 12 months.
Level of new car sales and economic conditions influencing these?
Page 9

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

The volume of business could be much higher or lower than expected.


Too little business might not cover the expenses incurred in writing a new
line of business.
So part of the agreement might be early cessation of the contract if
volumes do not exceed a certain amount to cut losses early.
Higher could entail more risk than desired.
Mitigate by imposing a pre-determined limit in the contract with the motor
manufacturer.
Economic conditions impacting on market values of vehicles and therefore
severity of claims for gap insurance are leveraged. For example, a 5%
increase in depreciation value has a greater than 5% increase in claim
severity.
Impacts on business plan of a downturn in volumes of business and
implications on expenses versus premium income to meet these?
Potential failure of vehicle manufacturer and impacts on market values
resulting from this.
Mitigate by modelling potential scenarios within Business Plan allowing
for variation in volumes of business written.
In the event of a total loss, the insured has no incentive to dispute the
amount paid by his private car insurer as will receive same amount in total
whatever amount paid: could mean higher than expected payout from gap
insurance.
Could mitigate by clause in contract enabling gap insurer to negotiate
amounts payable by communication and agreement with private car
insurer.
Could be moral hazard from underinsurance on car policy (even not
comprehensive).
Mitigate by clause in gap policy stating that car insurance policy should be
comprehensive and car fully insured.

The best answers to this question were those that focused on the particular
risks of gap insurance. Many candidates did not appreciate the geared impact
on claim severity of higher than expected depreciation and therefore did not
give enough attention to risk factors impacting market value of the vehicles.
(iii)

Page 10

Claims and exposure data are required for both the frequency and severity
elements.
No historic data exist for this scheme as it is new.
Internal motor data could be used to determine possible claims experience
on this scheme.
Total loss claims frequency for each of the first three years of ownership
of a vehicle could be derived from the companys motor experience if
there is a reasonable history.
This should be split down between theft, total fire/damage and flood if
data allows and is credible.
Use existing policy data to determine frequencies by rating cell (e.g. by
age/gender of customer, location, security features).

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

Consider trends in write off experience e.g. as a result of anti-speeding


initiatives or increased security.
Claims severity data: need to separate out property damage part of claims.
Claims severity data: allow for excesses on motor policy.
Severity will be modelled using purchase price and the market value of
each type of vehicle.
Purchase prices could be obtained from the manufacturer for each vehicle
type.
This data will need to be projected forward according to purchase price
inflation expected in future.
Market value by vehicle type likely to be available from Claims
department who will have information on expected settlement costs for
total losses.
Alternatively, industry data may be available (e.g. GLASS) providing
information on trade-in prices of second hand vehicles in the motor market
as these will be similar to the payout on total loss claims by the road risk
insurer.
Residual value decay information can also be obtained from published
information on the subject.
Inflation assumptions required for claims settlement costs of total losses.
Look at existing data to determine any trends with vehicle price inflation.
Check to see if there have been any historic changes in motor claims
handling procedures.
Obtain expected sales by vehicle from the manufacturer
...and likely customer profiles.
Claims handling costs may also be included in the risk premium.
Allowance should be made for reduced exposure due to resale of cars
within three years, if cover ceases on resale.
Allowance may also be made for the probability that not all premiums are
received (e.g. if a total loss occurs or car is resold or policyholder dies and
the premium payments stop). Or other monthly payment adjustments.
Information on resale volumes should be available from companys own or
industry data e.g. from DVLA.

Some candidates discussed data required for calculating the office premium
(e.g. fixed expenses, loading for profit) which gained no credit as this question
was specifically on the risk premium. A number of candidates listed a lot of
detailed data items, e.g. those used for a standard motor insurance policy, that
would not realistically be used in this situation. The stronger candidates were
more commercial in their answers, and more realistic about data that would
be publicly available. The weaker candidates talked about using the data
from other insurers, which showed a lack of commerciality.
(iv)

Other data when considering whether to underwrite:


Likely changes in economic factors that may influence car sales (e.g. price
inflation, wageroll inflation, unemployment levels).
How long has the motor manufacturer been trading?
Existing relationship with manufacturer.
Opportunity for cross-sell.
Page 11

Subject SA3 (General Insurance Specialist Applications) September 2006 Examiners Report

Market potential for this kind of business.


Insurers strategy.
Market share of manufacturer and credibility in the motor market.
Projected future sales by dealer.
Geographical spread of sales across the UK.
Policy wording and cover conditions to be used.
Expected conversion rate of business (policy to car sale ratio)/demand for
insurance product.
Commission rate payable to motor manufacturer.
Expenses of running the scheme split between acquisition, administrative
and claims handling costs.
Investment return expected on premiums in order to discount.
Profit share arrangements.
Competitors rates on similar products if available.
Any regulations or legislation that may impact on the selling methods for
this type of cover.
Can the IT systems cope with this product build?
Desired return on capital/profit requirements from the scheme.

This was a fairly easy question and candidates generally scored well. The best
candidates were able to state the other ingredients in the office premium and
talk about some of the other commercial considerations.

END OF EXAMINERS REPORT

Page 12

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
18 April 2007 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 A2007

Faculty of Actuaries
Institute of Actuaries

You are the reserving actuary at a small UK general insurance company writing just
private household buildings and contents insurance. Having joined last year, you
have just completed your first review of the latest year-end figures. Your company
uses an annual basis of accounting and the latest financial year ended on 31 March
2007. You have been asked by the Board to provide your best estimate of the
anticipated loss ratio for the year-end accounts and explain the level of uncertainty
inherent in your estimate.
You are also aware of the following information/facts:

Gross Written
Premium (m)
Booked Accident
Year Loss Ratio

2000

2001

Financial year ending 31 March


2002
2003
2004
2005

2006

2007
estimate

66.51

70.50

74.73

79.21

83.96

89.00

94.34

100.00

n/a

98%

102%

105%

120%

94%

95%

Note: For many years Gross Written Premium has been increasing at 6% per annum.

The policy cover was changed significantly on 1 October 2003 as follows, with
your predecessors estimate of the associated expected impact on incurred claims
in a typical year on an additive basis shown in brackets:
Excess for subsidence claims increased from 500 to 2,500 (15% reduction).
New exclusion of subsidence due to defective design / workmanship in the
first ten years (15% reduction).
Minor additions to cover (5% increase).

The various analyses you have undertaken lead you to believe that the underlying
trend in incurred claims (after allowing for prior year adjustments) has been rising
at an average rate of about 10% per annum, assuming no changes in cover.

A heat wave was recorded in August 2003, along with rainfall well below
average.

July 2006 was the hottest month on record, along with rainfall well below
average.

Ignoring the effects of any reinsurance:


(i)

List the different types of reserves that you might include in your estimate of
provisions.
[3]

(ii)

Discuss the advantages and disadvantages of reserving on an underwriting


year basis.
[4]

(iii)

Define the term loss ratio, giving examples.

(iv)

Calculate the expected accident year loss ratio for the 2007 financial year,
detailing any assumptions you make.
[18]

SA3 A20072

[3]

(v)

Discuss:
(a)

the extent to which any assumptions you have made may not be valid

(b)

any potential areas of uncertainty associated with the assumptions in


(a), and

(c)

how you might seek to improve the reliability of the estimates you
make by adjusting the assumptions in (a).
[10]
[Total 38]

Megasure Group was established in Japan over 100 years ago. Its headquarters are
still in Japan and it has become one of the worlds largest general insurance and
reinsurance organisations. Although it does not write any business in Europe at
present, it has operations in many other countries worldwide. Megasure Group writes
most types of insurance and reinsurance business, although not every product is
written in every country.
Following a strategic review, Megasure Group is considering writing general
insurance business in Europe. You are a consulting actuary and have been engaged to
advise Megasure Group. You have been told that:

It has initially decided to write only commercial property business. However


insurance will only be provided on large commercial risks, including large office
blocks or factories.

Both insurance and reinsurance business will be written.

The business will be written in London.

Megasure Group is considering two methods of entering the European market:


A

Becoming a corporate name on a Lloyds syndicate writing commercial


property business.

Establishing a wholly owned subsidiary company in the London market to


write this business.

Initially they would like advice to help them choose between the two market entry
options above.
(i)

Define the following terms:


(a)
(b)

Lloyds
The London Market
[2]

(ii)

Describe the main features of Lloyds.

SA3 A20073

[8]

PLEASE TURN OVER

(iii)

Describe the main differences between becoming a corporate name on a


Lloyds syndicate and establishing a wholly owned subsidiary. For each
difference, explain why it may be important for Megasure Group in deciding
its preferred method of entering the market.
[12]

Megasure Group subsequently decides to establish a wholly owned subsidiary,


Megasure Insurance Company Europe (MICE). You have been approached to assist
MICE in estimating its capital requirements for regulatory purposes. MICE will start
underwriting in 2008. You have produced the following estimates for the first three
years of operations, based on MICEs business plan:
TABLE 1
Estimated Capital Requirements for MICE
Amounts in millions
Year
2008
2009
10
20
MCR
10
20
RMM
25
35
ECR
35
45
ICA
(iv)

2010
35
35
45
50

State the meaning of each of the abbreviations MCR, RMM, ECR and ICA
and give a brief description. Details of how these amounts are calculated are
not required.
[4]

Megasure has considered the amounts shown in Table 1. Megasure needs to transfer
capital to MICE in 2007 in order that MICE can obtain approval from the FSA to start
writing business in 2008.
(v)

Describe the factors that Megasure should consider when deciding on the
initial capitalisation of MICE. Your answer should consider each of MCR,
ECR and ICA, in addition to any other relevant factors.
[13]

In its business plan, MICE assumed it would purchase excess of loss reinsurance and
that the net cost of this reinsurance would be 1 million per year. You allowed for
this reinsurance when estimating the ICA amounts shown in Table 1. MICE has now
asked you to produce an ICA estimate assuming that no reinsurance is purchased.
Your revised estimates are shown below, together with the original estimates from
Table 1.
TABLE 2
Alternative ICA Estimates for MICE
Amounts in millions
Year
ICA if reinsurance is purchased
ICA if no reinsurance is purchased
(vi)

2008
35
40

2009
45
50

2010
50
55

MICE is considering whether it should buy reinsurance or not. Discuss the


matters that MICE and its parent, Megasure, should consider.
[8]

SA3 A20074

MICE is preparing to submit its authorisation application to the FSA. An FSA


official is quoted in the press as saying that London Market insurers must monitor
their pricing more closely. MICE has called you to discuss this matter.
(vii)

List ways in which MICE can make sure it charges adequate premiums,
commenting on the practicality of each suggestion.

[9]

Following authorisation from the FSA, MICE is considering writing another line of
business in addition to commercial property.
(viii) Discuss the issues that MICE should consider before it decides whether or not
to write an additional line of insurance.
[6]
[Total 62]

END OF PAPER

SA3 A20075

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
April 2007

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

M A Stocker
Chairman of the Board of Examiners
June 2007

Comments
Individual comments are shown after each part-question.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

(i)

Outstanding reported claims reserve


IBNR
IBNER
Claims handling reserve
Reopened claims reserve
Catastrophe reserve
Equalisation reserve
Unearned Premium Reserve
Additional Unexpired Risk Reserve

Comments on question 1(i): Bookwork generally well answered by most candidates.


(ii)

Advantages:
Avoids heterogeneity arising due to:

mix of policy coverage (including terms and conditions)


mix of underwriting basis / use of rating factors
changes in the mix of business

Easier to allocate to underwriting year if date of loss is unclear.


IBNR is automatically included in any reserve estimate.
URR is automatically included in any reserve estimate.
Disadvantages:
Takes longer for each year to become fully runoff, giving rise to greater
potential uncertainty within the estimates.
Does not match the accounting procedures used by the company.
Can mask any potential changes in IBNR patterns.
Combines periods of potentially different claims environments.
May spread the incidence of a catastrophe or major claim event, potentially
reducing the apparent magnitude of the event.
Comments on question 1(ii): Bookwork with most candidates making a reasonable attempt
(iii)

Loss ratio: otherwise known as Claim ratio.


It is the ratio of the cost of claims to the corresponding premiums.
Claim ratios may relate to periods other than a year.
e.g. Incurred claims / earned premium for a given accounting year

Page 2

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Or estimated ultimate claims / earned premium for a given accounting


year.
Or incurred claims / written premium for a given underwriting year.
Estimated ultimate claims / estimated ultimate premium for a given
underwriting year.
Incurred claims and earned premiums for a given accounting year should
include estimated changes in estimates from prior years.
Comments on question 1(iii): Better candidates got full marks. Weaker ones just described
one version of the loss ratio.
(iv)

Assume business is written evenly through the year.


Assume that all the policies underwritten are annual.
Assume predecessors estimates are correct.
Assumption relating to how 6% change in premium applies.
Implies earned premium in year to 31 March 2007 =
50m + 50m/1.06 = 97.17m.
Assume loss ratio is calculated on an incurred claims / earned premium basis.
Implies incurred claims
2001:
2002:
2003:
2004:
2005:
2006:

0.98 97.17 / 1.066 = 67.13m,


1.02 97.17 / 1.065 = 74.06m,
1.05 97.17 / 1.064 = 80.82m,
1.20 97.17 / 1.063 = 97.90m,
0.94 97.17 / 1.062 = 81.29m,
0.95 97.17 / 1.06 = 87.09m.

Need to assess trend in claims incurred.


Up to and including 2003 can be taken at face value as there are no known
mitigating factors, other than natural variation.
Claims incurred from 2004 onwards need to be adjusted to put them on a
constant cover basis.
Aside from the change to cover, assume the same levels of exposure are
covered each year (that is, assume no other changes).
Assume mix of business unchanged following changes in cover.
Cover came in half way through 2003/04. Therefore the impact may be
assumed to affect 12.5% of claims.

Page 3

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

However, changes in cover on foundations brought in before the likely influx


in claims due to adverse weather conditions, so potential impact of hot
summer should be mitigated on a greater percentage of incurred claims. Say
15% (though anything up to, say, 25% may also be considered reasonable).
The defined claim event date will be significant.
This assumes that the incidence of subsidence claims increased significantly
because of the heat wave in August 2003 but after the policy change came
into force.
Implies claims incurred 2004 would have been (1 / 0.75) 0.15 97.90 + 0.85
97.90 = 102.80m in the absence of any changes to the cover provided.
Likewise, some claims in 2005 will have originated from the original policy
wording. So the impact may be assumed to affect 87.5% of claims.
However, the heat wave in August 2003 will affect this assumption, with
proportionately more claims likely to be affected by the change in cover. Say
95% (though anything from, say, 90% may also be considered reasonable).
Implies claims incurred 2005 would have been (1 / 0.75) 0.95 81.29 + 0.05
81.29 = 107.03m in the absence of any changes to the cover provided.
All of 2006 will be on the new cover = (1 / 0.75) 87.09 = 116.11m.
Estimates for new claims incurred = average of all years based on constant
(old) cover, reduced to new cover basis.
However, also need to adjust for higher than average incurred in 2003/04 due
to hot, dry summer.
By between 15% and 25%, judging by the numbers given say 20%
= 0.75 x [ (67.13 1.106 + 74.06 1.105 + 80.82 1.104 + 102.80 1.103 /
1.2 + 107.03 1.102 + 116.11 1.10) ] / 6
= 90.97m. (or alternative sensible calculation based on past results)
However, July 2006 was also exceptionally hot and dry
so the loss ratio is likely to be markedly higher than the long-run average.
However, the effect could be compounded on top of the 2003 conditions,
making the impact worse
or the impact may be less as many problems will already have been
highlighted in 2003/04
and losses will be limited by the changes to cover in 2003.
So, probably by between 15% and 25% judging by the impact of the
conditions in 2003 say 20%.
Page 4

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

=> adjusted estimate for new incurred claims = 90.974m 1.2 = 109.17m.
Expected loss ratio = incurred claims / earned premium
= 109.17m / 97.17m = 1.123.
As a cross-check 1.123 seems reasonable given the trend in loss ratios given in
the question after suitable adjustments have been made.
Comments on question 1(iv): Generally poorly answered. The question required candidates
to perform a number of separate calculations and adjustments to historical estimates to
obtain a reasonable estimate of the 2007 loss ratio. Candidates who took the time to set out
their solution logically tended to identify more of the steps required and so achieve the
highest scores. Candidates were asked to detail any assumptions made. Some candidates
oversimplified the question by inappropriate choice of assumptions, e.g., assuming that
subsidence claims should be ignored, missing the effects of the policy wording or
contradicting an assumption that was stated in the question. Many candidates attempted to
adjust the 2003 and prior years incurred claims to put them on a constant basis, rather than
adjusting the later years. Although it is possible to produce a reasonable estimate using this
approach, few candidates were able to apply it successfully. In both the 2004 and 2005
accident years, some claims occur from policies written under the new conditions and some
from policies under the old conditions. Candidates who attempted to adjust only the 2003 and
prior years generally did not identify these features of the 2004 and 2005 accident year
claims.
(v)

Business may not be written evenly in the year as more homes are bought and
sold during the period April September, so policy cover is more likely to
start in this period.
It will be important to analyse historical patterns of business written to
determine a reasonable assumption.
This area of uncertainty will be particularly relevant in light of the timing of
the hot dry periods and the timing of the change in cover as this is likely to
impact the various proportions of policies on risk affected.
Mix of business may have changed.
A more granular split of data could be used, for example, separating
subsidence claims.
The level of exposure accepted may have changed over time.
Example of investigation into level of exposure.
Assumptions regarding the impact of the hot dry summers will be critical to
the estimates,
but are likely to contain significant elements of uncertainty
in particular, the effect of the 2006 summer, for which little data will exist.
It will be important to analyse the pattern of weather / subsidence related
claims before and after the dry summers to ensure estimates are as reliable as
possible,
and discuss any technical engineering issues with relevant experts.

Page 5

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Your estimates of the impact of the cover changes may contain significant
degrees of uncertainty.
Any case estimates may be a particular source of additional uncertainty.
Your assumption regarding the underlying increase in incurred claims is likely
to contain significant uncertainty.
However, it should be possible to assess a range of reasonable estimates and
assess the impact of difference assumptions on your reserve estimates.
Your predecessors estimates of the impact of the changes in cover are likely
to contain a degree of uncertainty.
However, good quality data should be available to assess the two subsidence
related items as claim amount and age of home should be available for all
homes.
May reduce uncertainty by reviewing your predecessors analyses of these
changes.
Obtaining second opinions from a peer reviewing actuary should help to
improve reliability and the level of confidence in any estimates, thereby
reducing some of the uncertainty.
Estimates can be compared to benchmarks, industry sources, etc.
Comments on question 1(v): Again generally poorly answered with many candidates
answers largely limited to repeating the assumptions listed in part (iv) and noting that each
may not be valid in practice. Some candidates produced lists of "typical" areas of
uncertainty, for example, noting that it was possible that reinsurance arrangements or
taxation may have changed. Such answers typically did not sufficiently consider the extent to
which assumptions might not be valid, areas of uncertainty and possible ways to improve the
reliability of estimates, as requested in the question. Stronger candidates generated a wide
range of points.

(i)

Define Lloyds
A society that provides a market place and regulatory framework
within which individual and corporate members may participate
in the underwriting of insurance risks on their own account.
Define London Market
That part of the insurance market in which insurance and reinsurance business
is carried out on a face-to-face basis in the City of London.

Comments on question 2(i): Bookwork mostly answered well.

Page 6

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

(ii)

Describe the main features of Lloyds


Lloyds does not act as an insurer in its own right and, therefore, carries no
insurance risk.
The Council of Lloyds is responsible for management and supervision of the
market.
The Council delegates day-to-day running to the Committee of Lloyds, who
are responsible for administrative matters.
There is statutory actuarial involvement, and a Lloyds actuary
The FSA regulates Lloyds, (as well as Lloyds managing agents, members
agents and Lloyds brokers).
Members are grouped into syndicates.
Members are known as names.
Names can be individual or corporate.
Corporate names have limited liability.
Most individual names have limited liability
and there can be no new unlimited liability names.
Names are represented by members agents.
Each syndicate is run by a managing agent.
a company appointed to manage the affairs of the syndicate, appoint the
underwriter, and provide technical and administrative services.
Some managing agents are quoted companies listed on the stock exchange,
others are private companies.
In some instances, managing agents act as capital providers to the syndicates
they manage so have a dual role as corporate members of the market and
managing agents.
Business is written through the slip system.
Most policy and claims administration is performed by LPSO/Xchanging
The members of a syndicate share the risks written by the syndicates
underwriters.
However, if a member defaults on their liabilities, the other members of the
syndicate are not responsible for them there is no joint and several liability.
Because of this, each member is required to provide capital (Funds at
Lloyds) as security to support their total Lloyds underwriting business.
These funds can be drawn on in the event that the member defaults.
The level of funds required depends on the perceived level of risk in the
business which they underwrite, and the amount of business written.
The Central Fund is available at the discretion of the Council of Lloyds to
meet any valid claim that cannot be met by the resources of any member.
Compensation may now also be available from the FSCS.
Each syndicate year of account is allowed to remain open, usually for a
period of three years, before a profit or loss can be determined for that year.

Page 7

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

During that time, premiums received on business written in the year are
accumulated in a fund, out of which claims and expenses are paid.
At the end of the three year period, the fund would usually be closed by
estimating the value of the outstanding liabilities and reinsuring them into the
subsequent open year of the syndicate.
The reinsurance premium for this is known as reinsurance to close (or RITC).
Once this transaction has occurred, the final result of the closing year can be
determined, as can the profit or loss attributable to each member.
If liabilities are particularly uncertain, the year of account may be left open
longer than 3 years.
(Almost all) business is written through brokers.
Lloyds writes all classes of business,
in particular special and unusual risks
and some business in most countries of the world.
Comments on question 2(ii): This was a straightforward bookwork questions. Full marks
were available for candidates who wrote brief notes covering most of the main features. The
candidates who failed to achieve a good mark typically wrote answers that contained
inaccuracies, or focussed on a small number of features of Lloyd's
(iii)

Ease of entering market


Megasure Group could participate in an existing Lloyds syndicate by just
providing capital.
Establishing a wholly owned subsidiary would be more complicated.
would need to set-up admin processes / do admin
may not have necessary expertise in house/need to hire staff
and so this may take more time.
This may also lead to a difference in start-up costs of entering market (that is,
start-up costs likely to be lower for Lloyds).
Risk of entering a new market likely to be greater if start-up costs are greater.
Barriers to Entry
It may not be possible for Megasure Group to join the syndicates it wishes to
(they may have enough capital already).
There may not be syndicates that Megasure Group would like to participate in.
As a major international insurance group, there should not be significant
problems in obtaining permission to establish a subsidiary.
Should also consider ease of leaving market may be easier at Lloyds.
Control
Lloyds syndicates are run by managing agents, who make key decisions such
as appointing underwriters.

Page 8

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

The ability of individual names to control the syndicate may be limited


(although Megasure Groups influence will be greater if it provides a
significant amount of the syndicates capital).
Megasure could control every aspect of a wholly owned subsidiary, subject
only to regulatory constraints.
Megasure may have confidential information that a subsidiary could exploit. If
Megasure participated as a name on a syndicate, it may be reluctant to share
this with the managing agent.
Access to preferred risks
Underwriters on a syndicate may already be writing the type of business
Megasure wishes to write.
There may be advantages to writing renewal business on a syndicate than
considering risks for the first time at a new subsidiary.
The Lloyds credit rating may assist Megasure in accessing business.
Lloyds may give the start-up more credibility than would be attached to a
small start-up (value of Lloyds brand).
Ability to benefit from Lloyds licences.
Existing Lloyds syndicates would have links to brokers to access risks, and
underwriters may have special relationships.
Diversification
Megasure Group may be able to participate in a number of syndicates,
obtaining diversification.
Future strategy
Consider the long term strategy for the European operations. For example, if
more lines of business are planned in the future, may prefer to operate a
subsidiary.
Regulation
There may be differences in the regulatory requirements that make one option
preferable.
There may be differences in capital requirements.
There may be differences in the permitted assets.
Tax
There may be differences in the tax that make one option preferable.
Example of differences.
Expected profitability
There may be differences in the expected profitability that make one option
preferable.

Page 9

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Comments on question 2(iii): This question required candidates to apply their knowledge of
Lloyd's and the London market set out in parts (i) and (ii) to advise Megasure. Candidates
with a good knowledge of the main feature of these markets were able to identify the key
differences that would likely be of most interest to Megasure and so scored well. A surprising
number of candidates indicated that Megasure could only invest with limited liability through
becoming a Lloyd's name. Others suggested that a London market company would not be
able to write risks through a slip system. Candidates who did not understand the markets
therefore achieved lower scores.
(iv)

Definitions of RMM/MCR, ECR, ICA


RMM Required minimum margin.
MCR Minimum capital requirement.
RMM and MCR are different abbreviations for the capital requirement.
RMM/MCR is the greater of the GICR (general insurance capital
requirement)/RMS (required margin of solvency) and the minimum guarantee
fund (MGF) set by the EU.
ECR Enhanced capital requirement.
A more risk sensitive measure than the current EU directive minimum.
ICA Individual capital assessment.
This is a type of capital assessment introduced by the FSA.
Insurers are required to carry out regular assessments of the amount and
quality of capital that is adequate for the size and nature of their business.

Comments on question 2(iv): Bookwork, however many candidates did not know that RMM
is the same as MCR. Most candidates were able to define ICA, and many mentioned that this
was a type of capital assessment that firms made by considering the risks faced by their
business.
(v)

Initial capitalisation of MICE


The MCR/RMM has the force of EU directives, that is, it has the force of law.
This represents the minimum level of capital that must be held.
The ECR is currently only a private reporting requirement rather than a hard
test.
However, this is under review / may change in the future.
In any case, the ECR will be used as a basis of discussions between firms and
the FSA.
The ICA represents a firms own assessment of its capital requirements, so it
is unlikely it would want to hold less than this.
In practice, it is highly unlikely that the FSA would permit MICE to only hold
the MCR/RMM.
The ICA is the largest of the three estimates for MICE, so this is probably the
minimum level of capital the FSA would permit.
The FSA will review the ICA and issue individual capital guidance.

Page 10

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

ICG will be expressed by the FSA as a percentage of the ECR.


The firm would almost certainly want to hold at least the ICG.
We dont know what the ICG will be until the FSA has reviewed the ICA.
Therefore it is not possible to know with certainty the capital that MICE will
need.
We could try to obtain information on ICG from other sources to estimate the
capital requirements.
We may want the company to hold more capital than it is required to hold by
the regulator.
Additional capital will provide extra flexibility to management.
For example, greater flexibility with investment policy.
The ICA is an estimate. The company may wish to hold more for prudence.
If the company if very thinly capitalised, it may receive unwanted regulatory
attention which could distract management.
Consider how much capital the parent (Megesure) has.
Consider the cost of capital of Megasure and the return of MICE under various
levels of capitalisation. Consider costs of over-capitalisation.
Opportunity cost consider any alternative uses of capital that Megasure has,
and the returns on those activities.
Consider alternatives to parent providing capital.
The parent could provide a letter of credit rather than capital.
The parent could provide addition reinsurance to MICE to reduce its capital
requirements.
However, these alternatives may not be acceptable to the FSA.
Rating agencies what level of capitalisation does MICE require in order to
obtain the desired rating.
Do potential policyholders or investors require a particular level of capital /
rating in order to place business with the company?
Consider the capitalisation/ratings of competitors.
Where possible, consider future changes in regulation (for example, Solvency
II)
The estimated capital requirements are higher in 2010 than in 2008.
The company should consider whether it needs to fund future capital
requirements initially.
There may be options that dont require this to be funded initially, for
example, through retained profits anticipated in the business plan.
Megasure will also wish to finance 2007 start-up costs of MICE.

Page 11

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Comments on question 2(v): Poorer candidates lack of knowledge really showed here as
they were not able to write very much. Often there was a standard list of things you would
consider when deciding how much capital to hold. Better candidates tailored this to the
specific situation and demonstrated their understanding of the capital requirements
necessary now and in the future. Many candidates stated that firms should consider factors
such as the potential for accumulation of risk or large losses. Such factors would have been
considered in producing the ICA, and candidates were expected to refer to this in their
answers. Candidates who produced standard lists of factors to consider when determining
capital requirements needs without acknowledging where these were included in the ICA
generally scored poorly.
(vi)

Should reinsurance be purchased?


Purchasing reinsurance with a net cost of 1 million reduces the ICA by 5
million.
Many of the matters to be considered in deciding on the reinsurance purchase
will have been quantified in the ICA.
The company should consider the risk reward trade-off when purchasing
reinsurance, that is, the cost of the reinsurance and the benefit to MICE.
MICEs expected profitability will be higher if reinsurance is not purchased
(reinsurance premium is saved).
This reinsurance does not seem to reduce MICEs capital requirement by very
much.
As a result, MICE may be better off not purchasing this reinsurance.
Note that there may be a difference between MICEs capital requirements and
the ICA estimates (e.g. ICG). 5 million may not be the figure for Megasure to
consider.
Consider whether the parent has the extra capital available that would be
required if reinsurance is not purchased.
and what is the opportunity cost of the extra capital.
MICEs profits will be more volatile without reinsurance.
Consider whether Megasure (and its shareholders) prepared to accept more
volatile profits in return for higher expected profitability.
Consider any rules Megasure has regarding the reinsurance purchased by its
subsidiaries.
Megasure is also a reinsurer, so accepting an internal reinsurance may be
consistent with its risk tolerance.
The ICA amounts are estimates, and may have underestimated the benefit of
the reinsurance.
Although the reinsurance may not reduce the ICA by much, there may be
other benefits.
Example of other benefits: e.g. may not reduce the ICA by much, but may
reduce ruin probability significantly.
The reinsurance premium estimated in the business plan may be wrong.

Page 12

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

There may be an alternative reinsurance strategy that would better suit MICE.
Example of alternative reinsurance strategy: purchase higher limits, purchase
more reinstatements.
Another example of alternative reinsurance strategy: purchase reinsurance
from counterparties with better credit quality.
MICE could consider securitisation/ART.
Consider the availability of reinsurance.
Consider what the regulator might think about MICE not purchasing
reinsurance.
Consider rating agency views.
Consider the views of others, e.g. the underwriter, MICE board.
Consider competitors reinsurance strategies.
Consider possible tax differences.
Comments on question 2(vi): Weaker attempts just gave standard lists, hence missing the
specifics that applied in the situation, and scored poorly. Again candidates who
produced standard lists of factors to consider when determining reinsurance needs without
acknowledging where these were included in the ICA generally scored poorly. Some marks
were available for comment on the ICA calculations, e.g. consideration of capital
requirements if alternative reinsurance was purchased.
(vii)

Ways to ensure premium adequacy.


Detailed repricing of each individual risk offered to MICE.
May be too time consuming and expensive to do this for every risk.
MICE may not have sufficient internal expertise, especially initially.
Data for individual repricing may not be available.
May not be possible if need to respond to brokers quickly.
Only write in follow market with trusted lead underwriters.
Availability market may be limited.
Buy or build pricing software.
Consider the output from catastrophe models and location models, for
example, RMS.
Build in set underwriting protocols/guidelines and rating factors.
Agree policy wording protocols and exclusions.
Software may not work as intended.
Cost of buying/building software may be prohibitive.
Build tool to monitor the profitability of business being written.
Own data will be limited initially.
Some third party data will be available.
As business is short-tail, own data can be gathered relatively quickly.
Hire experienced staff (underwriters, pricing actuaries etc.) with detailed
market knowledge.
Or engage external consultancy to review processes, controls and outputs.

Page 13

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Need to find a way to assess experience.


Experienced staff will be expensive.
Have internal peer review process.
Should focus most attention on largest risks.
May not be possible if need to respond to brokers quickly, but can apply
results of peer review to future underwriting decisions.
Add margins to rates to reduce possibility of rate inadequacy.
Rates may become uncompetitive.
Ensure adequate premium loadings (commissions, expenses, reinsurance etc.)
Comment on practicality of monitoring loadings.
Monitor changes in volumes and other movements.
May be difficult to identify reasons for any changes.
Monitor insurance cycle.
May be difficult to accurately determine the position of the market in the
cycle.
Monitor prices charges by competitors as a check on your own model.
It may not be possible to access this information.
Competitors may be charging the wrong rates.
Comments on question 2(vii): The key features of MICE are that it is a new company, and it
is writing London market property business. Candidates were expected to identify methods of
monitoring pricing that are appropriate to such a company, and comment on the practicality.
Ensuring the reasonableness of the loadings for investment income and fixed expenses would
tend to be less important for MICE than some of the other factors identified in the solution.
Some candidates devoted most of their answers to expenses and investment income and so
omitted to mention other important considerations. Weaker candidates produced answers
that didn't reflect the nature of commercial property insurance or were a standard how to
price a product answer.
(viii) Extra line of business
Regulatory considerations.
FSA would have to reconsider the authorisation (unless it was already
authorised for the new line as part of the original process).
A sudden change in the business plan may create a bad impression with
regulator.
Characteristics of new market
For example, size, growth, ease of entry, relationships.
Capital requirements of new line of business.
Is this capital available?
Opportunity cost of capital.

Page 14

Subject SA3 (General Insurance Specialist Applications) April 2007 Examiners Report

Expected profitability/return on capital of new line of business.


Consider whether new line is consistent with MICEs risk appetite.
There may be a diversification benefit of writing lines of business with low
correlations.
Synergies with existing book.
Does company have sufficient expertise to write this line of business?
Would writing a new line of business distract the company from its
commercial property targets?
It might be better to defer introducing a new line until the company has been
established for a couple of years.
Consider views of parent.
For example, in raising the profile of the new business.
Consider views of rating agency.
Reinsurance requirements of new line of business.
Consider potential for cross-selling.
Consider whether system changes would be required.
Consider any tax issues.
Comments on question 2(viii): This question was generally well answered.

END OF EXAMINERS REPORT

Page 15

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
1 October 2007 (am)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt all 3 questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the
Formulae and Tables and your own electronic calculator.

SA3 S2007

Faculty of Actuaries
Institute of Actuaries

A student has produced the following tax calculation for a UK resident proprietary
general insurance company. The company writes only UK business.
Investment Returns
Purchase price of fixed interest securities
Market value of fixed interest securities as at 1/1/2006
Market value of fixed interest securities as at 31/12/2006
Accrued investment income during 2006

Accounting
Data
500.0
625.0
620.0
25.0

Return on fixed interest assets


Purchase price of UK equities
Market value of equities as at 1/1/2006
Market value of equities as at 31/12/2006
Net dividend income during 2006

145.0
30.0
55.0
60.0
2.0

Return on UK equities
Underwriting Result
Unearned premiums b/fwd
Written premiums
Unearned premiums c/fwd

32.0

45.0
100.0
50.0

DAC b/fwd
Acquisition costs paid
DAC c/fwd

9.0
18.0
10.0

Reinsurance purchased on 1 January for 2006

20.0

Earned Premiums net of DAC and Reinsurance


Outstanding and IBNR claims b/fwd
Net Claims paid in 2006
Outstanding and IBNR claims c/fwd

56.0
485.0
60.0
475.0

Allowable Equalisation reserve transfer


Additional provision for future catastrophes

10.0
4.0

Claim handling expenses incurred in 2006


Provision for future claim handling expenses
for claims incurred in 2006

3.0
5.0

Increase in claims and claims provisions


Staff and buildings costs

72.0
10.0

Underwriting Result
Tax @ 30%
Underwriting result after tax

Student's
Calculation

151.0
45.3
105.7

Correct the students tax calculation and state the underlying principles that explain
how each accounting item should be treated. You may ignore any interim
arrangements and can assume that the accounting items given are correct.
[11]

SA3 S20072

Company A is a large insurance group with worldwide operations. One of its small
subsidiary companies, Company B, is a UK company that started writing international
property and liability reinsurance business in 1950. Company B took the decision to
pull out of the US reinsurance market in 1980 and since then has focused on writing a
small and very profitable European property reinsurance book. Company Bs
discontinued portfolio currently consists largely of US Asbestos, Pollution and Health
Hazard (APH) liabilities. In recent years, Company A has needed to inject more
capital into Company B following reserve deteriorations on the US APH liabilities.
The technical reserves for Company B are established on a discounted basis.
The Board of Company A is currently considering selling Company B in view of
Company As worldwide strategy to focus on writing direct business.
(i)

Describe the major areas of risk facing Company A in respect of the APH
claims liabilities of Company B.
[7]

(ii)

Suggest the benefits to Company A of selling Company B.

(iii)

Discuss alternative options open to Company A to remove the risks of the


APH liabilities within Company B, explaining their advantages and
disadvantages. Company A has already dismissed the use of an adverse
development cover.
[10]

[5]

Company C has shown an interest in acquiring Company Bs APH liabilities from


Company A. Company C has proposed that it acquires the APH liabilities in return
for the transfer of assets equivalent to the discounted claims reserves (including
claims handling expenses) on a High estimate basis, where High estimates are deemed
adequate in 90% of future possible outcomes.
The company actuary for Company B has provided the following estimates in respect
of the APH liabilities as at 31 December 2006.
Figures in US$m

Case
reserves

Asbestos
Pollution
Health Hazards
Subtotal
Claims handling expenses
Total

12.8
12.1
1.0
25.9

Annual
Average Paid
(last 3 years)
2.7
4.5
0.8
8.0
1.0
9.0

Best estimate reserves


Undiscounted Discounted
4%
51.3
24.1
2.0
77.4
3.1
80.5

34.5
19.8
1.8
56.2
2.5
58.7

High estimate reserves


Undiscounted
Discounted
4%
77.0
36.2
3.0
116.1
4.6
120.7

48.8
29.7
2.5
81.0
3.8
84.8

Company Bs assets have produced investment return averaging 4.2% per annum over
the last 3 years.
(iv)

Suggest the benefits to Company C of acquiring the APH liabilities.

(v)

(a)

Estimate the discounted mean terms of each claim type on each of the
Best estimate and High estimate bases.

(b)

Comment on the reasonableness of the relative length of these


discounted mean terms.

(vi)

[5]

[9]

Set out the challenges that you would make in regard to the numbers in the
table above when assisting Company C in its initial price negotiations for this
portfolio.
[9]
[Total 45]

SA3 S20073

PLEASE TURN OVER

You have recently joined a large general insurance company and you are determining
the reserves for a portfolio of business in the first quarter of 2007, using annual data
as at 31 December 2006. The company business plan for the 2004 year shows a
planned ultimate loss ratio of 85%. The company stopped writing this portfolio at the
end of 2004.
You have the following summary information readily available:
Underwriting Years
2000 2001 2002 2003 2004 2005 2006
Premiums million
Reported loss ratio
Basic chain ladder
ultimate loss ratio %
Booked ultimate loss ratio %
A priori * loss ratio %
Mean term of paid claims
In years
Development year
Age to age factors n to
n + 1(reported claims)

17.2
87.0
90.5

17.4
82.0
93.0

17.6
69.0
90.0

18 20.1
50.0 34.6

91.0
91.0
4.0

93.0
93.0

89.6
89.6

92.0 85.0
92.0 85.0

10.00

3.00

1.80

1.40 1.15

1.09

1.03 1.01 1.00

* a priori in this context refers to the initial estimate ultimate loss ratio used in the
Bornhuetter-Ferguson calculation
As part of your review you investigate other sources of information. You discover
that:

The 2003 and prior years are running off as planned.

A company pricing database shows that the average rate change across policies in
2004 is minus 10%.

The ex-underwriter of the account tells you that he feels that rates were showing a ten
to fifteen percent reduction and that some policies had wordings which gave wider
coverage than in previous years.
(i)

(a)

Calculate the Bornhuetter-Ferguson and chain ladder estimated


ultimate loss ratios for the 2003 and 2004 years.

(b)

Comment on the difference between the Bornhuetter-Ferguson and


chain ladder expected ultimate claims.
[6]

(ii)

(a)

Derive three alternative a priori loss ratios for use in the BornhuetterFerguson method for the 2004 year.

(b)

Calculate three alternative Bornhuetter-Ferguson ultimate loss ratios


using your estimates from part (a) and tabulate your results.
[5]

SA3 S20074

(iii)

(a)

State the concerns that you would have with the 2004 initial a priori
loss ratio and the new a priori loss ratios you have derived in part (ii).

(b)

Comment on the calculated ultimate loss ratios.

[9]

The company finance director tells you that it has always been company
practice to use the budgeted loss ratios in the Bornhuetter-Ferguson
calculations. He has seen your table of results and states that the BornhuetterFerguson method is not credible as it can give any result that you may want.
He tells you that he intends to use the ultimate claims using the budgeted loss
ratio in the Bornhuetter-Ferguson calculation.
(iv)

(v)

Discuss the finance directors comment on the credibility of the


Bornhuetter-Ferguson method.

[5]

(a)

Select one a priori ultimate loss ratio for the 2004 year and calculate
the increase in reserves needed if the company changes its a priori loss
ratio to your figure and

(b)

Comment on the materiality of this difference if the company follows


its usual practice.
[5]

A broker has approached the company offering to arrange an adverse


development cover to reinsure this portfolio at a price that you estimate to be
best estimate loss cost plus 30% load plus brokerage.
(vi)

Explain the term adverse development cover.

[2]

(vii)

Explain the risks to the ceding company of reinsuring this portfolio and the
risks to the reinsurance company of writing this portfolio and explain what
each company can do to mitigate its risks.
[9]

(viii) State with reasons whether you think that such a policy would be appropriate
in this situation.
[3]
[Total 44]

END OF PAPER

SA3 S20075

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
September 2007

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

M Stocker
Chairman of the Board of Examiners
December 2007

Comments
Individual comments are shown after each part-question

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

1
Investment Returns
Purchase price of fixed interest securities
Market value of fixed interest securities as at 1/1/2006
Market value of fixed interest securities as at 31/12/2006
Accrued investment income during 2006

Corrected
Calculation
500.0
625.0
620.0
25.0

Return on fixed interest assets =


{Accrued investment income} + {MV @ 31/12/2006} - {MV @ 1/1/2006} =
25 + (620 - 625) = 20
20.0

Purchase price of UK equities


Market value UK equities as at 1/1/2006
Market value of UK equities as at 31/12/2006
Net Dividend income during 2006
Return on UK equities =
{Net dividend income} + {MV @ 31/12/2006} - {MV @ 1/1/2006} =
2 + (60 - 55) = 7

30.0
55.0
60.0
2.0

7.0

Underwriting Result
Unearned premiums b/fwd
Written premiums
Unearned premiums c/fwd

45.0
100.0
(50.0)

DAC b/fwd
Acquisition costs paid
DAC c/fwd

(9.0)
(18.0)
10.0

Reinsurance purchased on 1.1 for 2006

(20.0)

Earned Premiums net of DAC and Reinsurance =


{Unearned Premiums b/f} + {Written Premiums} - {Unearned Premiums b/f}
- {DAC b/f} - {Acquisition costs paid} + {DAC c/f}
- {Net Cost of Reinsurance} =
45 + 100 - 50 - 9 - 18 + 10 - 20 = 58
58.0

Outstanding and IBNR claims b/fwd


Net Claims paid in 2006
Outstanding and IBNR claims c/fwd
Allowable equalisation reserve transfer
Additional provision for future catastrophes

Claim handling expenses paid in 2006


Provision for future claim handling expenses for claims incurred in 2006

485.0
(60.0)
(475.0)
(10.0)
0.0

(3.0)
0.0

Increase in Claims and claims provisions =


{OS & IBNR b/f} - {Net claims paid} - {OS & IBNR c/f}
- {Allowable equalisation reserve transfer} - {CHE} =
485 - 60 - 475 - 10 - 3 = -63

(63.0)

Staff and buildings costs

(10.0)

Page 2

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report
Technical Result =
{Fixed interest return} + {Equity Return} + {Net Earned Premiums}
{Increase in paid claims and claims provisions paid} =
20 + 7 + 58 63 10 = 12
of which no further tax due on dividend income

2.0

Taxable result = 12 2 = 10
Tax @ 30% = 10 0.3 = 3.0
Underwriting result after tax = 12 3 = 9.0
Alternative solution:
Gross Earned Premium
Less Earned RI
Net Earned Premium
Incurred Claims
Allowable Expenses
Increase in DAC
Underwriting result
Taxable fixed interest return
Taxable return on equities
Equalisation reserve provision
Taxable result
Tax @ 30%
Franked income from equities
Technical result after tax

12.0

10.0
(3.0)
9.0

100-(50-45) = 95
(20)
= 75
485-60-475 = (50)
31
1
= (5)
20
5
(10)
= 10
(3)
2
=9

Investment Return
From Bonds or Loan relationships is taxed on the total returns whether
realised or unrealised. The bonds are valued on a mark to market basis.
Dividends from UK equities are not taxed further. Gains are taxed on a mark
to market basis.
Underwriting result taxable as follows:
Earned Premiums (net of RI), so unearned premiums net of DAC tax
deductible
Expenses are tax deductible (acquisition costs, running costs of the business,
)
Less claims handling expense provision is allowable to the extent that the
expenses relate directly to claims for which claims provisions have been
accepted by HMRC.
Less paid claims
Less change in o/s claims carried forward supported by case estimates or
statistical projections
Change in IBNR subject to justification

Sensible mention of discounting.


The insurance technical provisions claimed for tax purposes are net of amounts
recoverable from reinsurers.
For tax purposes it is necessary to assume that all amounts due will be recovered.
A deduction is allowed for specific provisions for amounts estimated to be
irrecoverable from reinsurers but a general provision is not allowed.
In the UK, insurers are required to establish an Equalisation Reserve/provision over
and above their claims provisions in respect of certain classes of business (regarded as

Page 3

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

being potentially volatile). Statutory rules govern the calculation of transfers to the
Reserve (which are tax deductible) and transfers from the Reserve (on which tax is
payable).
Provisions for future catastrophe losses are not allowable except where required by
law
Comments on Q1. The template given should have guided the layout of the answer
required. A minority of students followed this template and scored well. Those who
chose their own format often made mistakes and the answers were harder to follow.
Most candidates showed a reasonable grasp of the basics concerning equities and
bonds returns and made a decent effort at correcting the calculation. However, very
few candidates made many sensible comments on the principles of taxation ( e.g.
IBNR is deductible only subject to justification, change in outstanding claims is
deductible provided that supported by case estimates or projections) so very few
candidates scored highly on this question.

(i)

APH risks
Claims reserves very uncertain because APH liabilities are long tailed due
to long latency period (could be inadequate)
Liabilities stem from early years when policy records may be incomplete:
difficult to assess full extent of exposures
Liabilities increase leaving the reserves inadequate due to: (need reason)
Risk of legal judgements increasing liabilities
Change to regulatory environment
Risk of new latent diseases emerging thus increasing liabilities
Propensity to claim
Claim inflation higher than expected
Gearing for RI policies
Risk of accumulations
Risk around the value of the discount within the discounted reserves:
+ Timing of claims payments is uncertain can be difficult to assess
cashflows
+ May not achieve return on assets implied by discount rate.
Currency risk if liabilities not matched
Mismatching of assets and liabilities
Risk of disputes and bad debt on outwards reinsurance given the age of the
liabilities
Claims handing cost could be higher than expected (for various reasons,
legal, cost of specialist handlers etc)

Comments on Q2(i). This was answered well by most candidates.


(ii)

Page 4

Benefits of sale
Removes risk of further reserve deterioration on APH
More stability (less volatility)
No longer distracts Company A management from ongoing business
frees up resources (future strategy)

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

Brokers, potential customers and rating agencies may have undervalued


Company A because of APH liabilities of B
Sale of B may result in enhanced sale value of A
and increase ability to issue debt
Administrative savings
Extract capital from Company B (less capital needed)
May get a good deal

Comments on Q2(ii). This was answered reasonably well.


(iii)

Alternative options

Part VII transfer to external company (or, alternatively, for Loss Portfolio
Transfer and Novation)
+ legal liability transferred
+ so employee and shareholder rights not affected
+ may improve the sale terms of Company B
+ can transfer to specialist APH run-off company
+ dont have to transfer non APH policies
need to get regulatory/court approval which can be time consuming
possible reputational risk
will need to commission an independent expert to opine on
policyholder protection
this could be expensive
Proactive commutation of policies
+ does not normally require regulatory approval
+ opportunity to make profits on individual policies
can be time consuming and needs senior input
will be impractical to commute all the policies and unlikely to be able
to remove all exposures this way
Scheme of arrangement
+ Can achieve finality
+ Do not have to get agreement from every policyholder
May be reputational issues for Company A if scheme fails or seen to
be unfair
Can take some time to set up
May not have expertise in house to plan or execute

Comments on Q2(iii). This was poorly answered. Many candidates gave answers that
either would not remove the liability from the company (e.g. ring fence within company)
or were very unlikely to be realistic or practical. Those candidates that did suggest a Part
VII transfer or a scheme of arrangement often did not give sufficient further detail. The
wording of many candidates answers gave the impression that the portfolio could be sold
and all associated capital would be released. If such a transfer were to take place then a
premium over the reserves will almost certainly be paid and this could be greater than
the capital held. Loss portfolio transfers answer the question correctly only if they
transfer legal liability. In the UK this is called a Part VII transfer, other legislation
existing in other counties.

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

(iv)

Benefits of acquiring

Make a profit and/or diversify but VERY unlikely


best estimate reserves may be prudent following reserve injections
actual investment return is greater than assumed in the value of the
discount
deal done on high estimate basis, for which reserves are an additional
50% on best estimate
Expertise to be able to assess risks and price portfolio
May have experience of doing schemes of arrangement
May have own APH department so gain from economies of scale
May have specialist claims handling and commutations expertise
Bargaining power for class actions, set-off rights with brokers and
reinsurers
Better bargain with claimants if not involved in writing current business

Comments on Q2(iv). Most candidates did not give enough weight in their answers to the
benefits of scale, expertise and bargaining power of a specialist company taking on this
portfolio.
(v) (a)

(b)

Estimate discounted mean term


Discounted reserves ~ undiscounted reserves
(1 + interest rate)-DMT
So DMT can be estimated by
ln(undiscounted/discounted)/ln(1 + discount rate)
Best
High
Liability Type
DMT
DMT
Asbestos
10.1
11.6
Pollution
5.0
5.0
Health Hazard
2.7
4.6
Claim Handling
Expenses
5.5
4.9

Reasonableness:
Asbestos DMT higher than others reflecting longer latency period of asbestos
Some of pollution liabilities relate to clean up costs, which are not bodily
injury claims so slightly shorter than asbestos
Pollution mean term could be a bit short
Pollution best = high could be an error (or other sensible comment)
Health hazards DMT would depend on claim type but expected to be shorter
tail than asbestos
Health Hazard estimated DMT seems too low: may be error
Health Hazard is more dependent on latent claims (IBNR) as low ratio of paid
claims to case reserves therefore expect a bigger difference between best DMT
and high DMT
Asbestos High DMT > Best DMT

Page 6

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

May be reasonable if assume longer payment pattern accompanies


deteriorating experience

Because it would take time to deteriorate


Reasons for deterioration: more new claims reported than expected

inflation of average claims costs higher than expected

more mesothelioma claims (longer latency period and higher average


cost)
Comments on Q2(v). A surprising number of candidates were unable to calculate the
DMTs and those that were able to do the calculation typically made only brief
comments.
(vi)

Challenges

Company C wants the amount of assets transferred to be as high as


possible therefore challenges should focus on increasing discounted High
estimates
Discount rate may be too high and therefore discounted High reserves too
low
need to allow for risk of past returns not being achieved in future
especially as need to consider long future payment patterns
Nature of liabilities => outcome is very uncertain
Calculation of IBNR to OS ratios
Calculation of survival ratios
Ratio between High and Best reserves is the same at 1.5 for all claim types
How likely is it that the High estimates represent the 90th percentile?
Data provided are inadequate would want to see lots more (need valid
example of extra data)
Asbestos probably more uncertain than pollution due to longer latency
and mean term to settlement
=> may need to increase High estimate for asbestos
Based on average claims payments, Health Hazards reserve could be
exhausted in 2.5 years.
How to allow for possibility of later emergence of claims e.g. lead paint,
new claim types?
Survival ratio for Claims handling expenses is just over three years, far
less than the DMTs would suggest. Claims handling expense reserve
appears to be understated.
The ratio of High to Best undiscounted reserves is the same for claims
handling expenses as for claims reserves. You might expect claims
handling expenses to be proportionately higher than this in view of
reducing economies of scale as time goes on
What downside scenarios have been considered in assessing the range?
E.g. US legal developments
Rapid increase in claims inflation

Comments on Q2(vi). Candidates missed many obvious challenges in this question and
most did not consider calculating survival or IBNR to outstanding ratios

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

(i)

development year
Age to age factors n to
n+1(reported claims)
Age to ult factor

10.00

3.00

1.80

1.40

1.15

1.09

1.03

1.01

1.00

3.286 1.826 1.304 1.134 1.040

1.01

1.00

1/cumf
1 - 1/cumf

0.304 0.548
0.696 0.452

BCL for 2003


BF for 2003
BCL for 2004
BF for 2004

= 50 1.826
= 50 + 0.452 92
= 34.6 3.286
= 34.6 + 0.696 85

= 91.3
= 91.6
= 113.7
= 93.7

Chain ladder does not take account of prior expectations and therefore projects the
ultimate without adjustment.
2003 91.6 because a priori loss ratio in line with experience
2004 113.7 because a priori not in line with experience
It is the same as using the bf method with an a priori ultimate of 113.7%
The BF method is credibility weighted to take account of prior knowledge.
BCL could be distorted by one or two large claims hence big difference
If the a priori estimate is closer to the real ultimate then theBF method will give a
more accurate result and vice versa
Which means that great care needs to be taken in selecting the a priori estimate.
Comments on Q3(i). All candidates should have scored full marks for the technical part
of the reserving calculations. However, some candidates did not appear to know how to
perform a BF calculation and others calculated it in a very inefficient way thus wasting
valuable time. Also basic errors were made by some candidates.
(ii)

(2003 and 2004 booked starting points given in the table below. 2003 is a
better starting point given the implausible 2004 number. There are other valid
starting points including prior year basic chain ladder estimated ultimate loss
ratios.)
(Increase in reserves shown here for part v calc)
There are alternative starting points e.g. 2003, BCL.
Correct application of rate reductions:
Calculation of ultimates:
Clear tabulation of results

Page 8

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

a priori LR bf ult
using 2003 as a base
with 10% rate reduction
with 15% rate reduction
with 20% rate reduction
with 30% rate reduction

increase in
reserves from bf
ult of 93.7
(using an 85 a
priori LR)

increase in
reserves
from 85%
booked
loss ratio

92
92
92
92

--->
--->
--->
--->

102.2
108.2
115.0
131.4

105.7
109.9
114.6
126.0

2.41
3.25
4.20
6.49

4.16
5.00
5.95
8.25

85
85
85
85

--->
--->
--->
--->

94.4
100.0
106.3
121.4

100.3
104.2
108.5
119.1

1.32
2.10
2.97
5.09

3.08
3.85
4.73
6.85

using 2004 as a base


with 10% rate reduction
with 15% rate reduction
with 20% rate reduction
with 30% rate reduction

Comments on Q3(ii). Almost all candidates were able to select three ratios for use in the
BF. Some candidates were happy to select unhelpful a prioris (ignoring information
given) or made selections which were near-identical (e.g. comparing a volume weighted
average with a simple average) missing the much greater uncertainty in other
assumptions. The better candidates recognised that the key issue is one of uncertainty and
calculated a range of a priori estimates accordingly. Mistakes were often made in the
calculation of rate change impact. The better candidates recognised that there may well
be claims trend in addition to the rate weakening. Many candidates dropped easy marks
for not tabulating the results as instructed.
(iii)

Original loss ratio


Original (budgeted) a priori loss ratio is now over three years old.
And we have newer information which suggests that this estimate is
optimistic.
And could easily be 30 points out, significantly distorting the ultimate
Need to ask questions about currency of claims as could cause distortions
New estimate a priori loss ratios
The pricing database is an unknown to us and we do not know how the figures
are calculated. Example required of why this may be a problem, e.g.
granularity.
Assuming the calculations are correct then this estimate is very different to
that assumed in the budget
Budget assumes approximately 8% rate increase compared to 10 point
decrease, an 18 point gap.
It is unlikely that the rate changes due to wording changes will be accurately
recorded, if at all.
The rate reduction may well be calculated as a change in premium per unit of
exposure which will not take into account inflation of loss cost over the year,
in which case the rate change will be worse by the amount of loss inflation
(e.g. 4%)

Page 9

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

The underwriters estimate will probably be influenced (biased) by the


calculated rate change and the poor experience of the portfolio to date, and it
is likely that his figure is based on some of the larger and more memorable
policies in the portfolio. However, his judgement seems to be credible.
The difficulty is in knowing how much to add for the widening of policy
coverages but bearing in mind that rate reductions are normally underestimated in a softening market
The true rate softening could easily be double the estimated 10%. This is more
worrying given the sudden increase in the portfolio premiums from an
otherwise stable position.
Ultimate loss ratio
The reported to date loss ratio for 2004 would suggest an ultimate loss ratio
significantly higher than 85% using BCL
Using the 10% reduction gives a loss ratio 15 points higher (or alternative
suggestion)
With different plausible assumptions there is a large range in the calculated
answers
BCL could be distorted by one or two large claims hence making a big
difference so it could be difficult to settle on one estimate
The range gives us significant concern over the budgeted ultimate.as 85% is
below the bottom of our range.
Comments on Q3(iii). Very few candidates raised questions on the pricing database and
the way it calculates a rate change. Also many candidates did not discuss the difficulty of
taking account of the underwriters comments appropriately. Several candidates chose to
dismiss the underwriters views entirely and thus discarded a valuable source of
information. Most candidates got very low marks on this part.

(iv)

It is true that the BF method relies heavily, even critically, on the a priori
estimate which is a strength if the a priori is chosen well and a weakness if
chosen badly.
Indications are that the selected a priori estimate needs revision as evidenced
by
the high reported loss ratio for a year at the end of 36 months
the rate reduction estimates in the database
and from the underwriter
The company should be trying to book best estimate results which means
using the best information available and should take precedence over an
established practice which in this instance looks like it is not giving a good
solution.

Comments on Q3(iv. Most candidates made some reasonable points in discussing the BF
reliance on its a priori. However, few candidates stated clearly that reserves should be
calculated as best estimates using the most up to date information. Most candidates gave
a general description of the BF method and failed to tailor their answer to the specific
circumstances given in the question and thus missed many easy marks. There was
significant misunderstanding of what independent means in relation to the BF method.

Page 10

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

A commonly held incorrect view was that no information derived from the year in
question could be used. Rate changes, claims inflation, wider wordings and increasing
premium volume are all independent and should be used in determining a best
estimate a priori.

(v)

For calculation of reserve deterioration see above tabulation


Selecting one number: e.g. 20% rate reduction on 2003 gives a priori loss ratio
of 115%. This gives an ultimate loss ratio of 114.6%
The increase in reserves is the change in IBNR
So the reserve deterioration is the premium multiplied by the change in ULR
= 20.1 x (114.6% -85%) = 5.95m
Calculating reserve deterioration:
Ultimate claims are approximately 90m 92% = 83m for the portfolio
Assume paid claims are at 25% which means reserves are about 62m
Then the increase in reserves will be about 5.95/62 = 9.6%
This has borderline materiality for this portfolio but will be immaterial for the
whole company. One should check that this is an isolated instance and not
systematic, in which case the overall impact may well be material.

Comments on Q3(v). Most candidates were able to calculate the reserve deterioration
but very few considered paid claims and therefore the likely impact on unpaid claims.
Further a surprising number of candidates did not pick up that for a large company this
deterioration in itself is unlikely to be significant.
(vi)

Adverse development cover pays for the deterioration in claims over a


specified period in excess of a pre-agreed amount.
The deterioration can be measured using paid claims or reported claims
The specified period would normally be more than one year.
It is usual for the ceding company to retain a share of the claims i.e. less than
100% is ceded.

Comments on Q3(vi). Most candidates appeared to know broadly what an adverse


development cover was, but missed the detail.
(vii)

Risks to ceding company:


Claims experience is better than expected and the company pays the claims
and the reinsurance premium
Claims could deteriorate outside of the period of cover and therefore not be
reinsured.
Claims could deteriorate badly and the companys retained share could still be
significant.
The reinsurer could fail

Page 11

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

Mitigated by:
Negotiating the lowest premium possible
Retaining the minimum possible share
Spreading the ceded risk amongst several reinsurers
Using only top credit rated reinsurers
Risks to the reinsurer
Underwriting risk takes on a risk which is worse than anticipated at too low
a premium
Timing of claims happens inside the covered period
Legislation/judicial decisions/interpretation mean that the original policies
have wider coverage than expected, and hence claims more than expected.
which may accumulate with other risks that it has.
There may be significant latent claims in the portfolio.
Currency fluctuations could make the claims larger than expected
Invested assets may perform worse than assumed in any calculation of
premiums.
Moral hazard from the cedant, manipulating claims payment/reporting to fall
within the policy period.
Mitigated by:
Thorough analysis and investigation of the policy and claims files
Purchasing reinsurance
Careful monitoring and limiting of aggregates
Hedging of currency
Diversification of assets
Participation clause
Increase profit margin
Claims audits to verify no changes in claims procedures.
Comments on Q3(vii). Most candidates gave reasonable answers, but only identified a
few of the risks to the reinsurer.
(viii)

Page 12

I would not think that buying this policy would be appropriate in this instance.
(or clear statement of opposite opinion)
All years except for 2004 seem to be running off to the recognised pattern
With the exception of 2004.which will attract more scrutiny and a higher
premium anyway.
The total reserve size is small and any adverse development is likely to be
much smaller

Subject SA3 (General Insurance Specialist Applications) September 2007 Examiners Report

So unless the companys solvency is very tight this does not look like a good
option as the company will be ceding profit and paying brokerage when the
company should be well able to pay the claims without reinsurance.
Comments on Q3(viii). Many candidates answered this well. However some lost marks
by not stating a clear preference on whether the cover is appropriate in this case.

END OF EXAMINERS REPORT

Page 13

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
14 April 2008 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 A2008

Faculty of Actuaries
Institute of Actuaries

You are the actuary responsible for estimating the reserves for a large UK general
insurance company with several lines of business. Below is the gross claims
development for the public liability class of business and your estimates of the gross
ultimate claims at year end 2006.
Figures in 000s
Policy
Year
2002
2003
2004
2005
2006

4
5,691
5,649
8,652
10,428
11,513

Notified Claims
Development Quarter
8
12
16
10,256
11,956
11,421
3,146
3,265
3,684
11,423
10,875
23,453

20
11,565

Figures in 000s
Year End Year End Estimated
Policy
Notified
Paid
Ultimate
Year
Claims
Claims
Claims
2002
11,565
7,716
12,409
2003
3,684
2,431
4,265
2004
10,875
3,898
13,285
2005
23,453
3,596
38,202
2006
11,513
1,006
35,125
Total
61,089
18,648
103,286

At year end 2006 you estimated the following range of gross ultimate claims for this
class of business:
65%
105,352

Range of estimated ultimates ( 000s) - Percentiles


70%
75%
80%
85%
90%
115,680 121,878 126,009 131,173 136,338

95%
146,666

The underwriter for this line of business calls you to explain that policies are written
on a claims made basis and have been since 2002. He is worried that you are not
taking account of this in your projections as you seem to have significant IBNR in
prior years.
(i)

Define the term claims made policy.

[1]

(ii)

List the advantages and disadvantages to an insurer of this type of wording. [3]

(iii)

Outline the points that you would make to the underwriter in explaining your
estimate of ultimate claims in prior years.
[4]

(iv)

Explain whether the actuary responsible for reserves should provide a range of
estimates in a reserve report.
[5]

You estimated the range at year end 2006 using scenario testing and your judgement.
(v)

Explain the advantages and disadvantages of this approach in calculating a


range.
[4]

SA3 A20082

You have been asked to consider two further ways of deriving a range: Bootstrapping
and Macks method.
(vi)

Describe each of these methods, giving its advantages and disadvantages. [14]

You now have six additional months of data. The latest notified claims and your new
estimated ultimate claims as at 30 June 2007 are as follows:
Figures in 000s
Policy
Notified
Year
Claims
2002
11,565
2003
3,774
2004
10,848
2005
50,910
2006
14,542
Total
91,640

Estimated
Ultimate
12,409
4,350
13,030
65,687
41,256
136,732

You mention this to a colleague who comments that:

the ultimate level is inconsistent with the range that you provided at year end
2006.
this might have a significant impact on the companys ICA.

(vii)

State with reasons whether you agree with these comments.

[10]

The premium for each of the years is shown below:


Figures in 000s
Gross
Policy
Written
Year
Premium
2002
15,649
2003
12,156
2004
21,635
2005
25,265
2006
35,684
Total
110,389

You have been asked by the finance director whether this is a sufficiently profitable
line of business following the change to the claims made basis as from 2002.
(viii)

Outline the points that you would make to your finance director and detail any
further information that you might require.
[18]
[Total 59]

SA3 A20083

PLEASE TURN OVER

Rapidco is a large industrial company based in the United Kingdom. The company
has seen rapid expansion over the past ten years, doubling its output through
automation of key processes while keeping its workforce at similar levels and
increasing the number of manufacturing sites it owns by over fifty percent.
You are an independent actuary engaged by the companys directors. The directors
believe that the company has a very low reported claims record over the past ten
years. They believe that the insurance premiums are too high as they are paying their
insurers standard book rates with the exception of the fleet portfolio which is
experience rated. Its main other insurances are predominantly commercial fire,
business interruption, employers liability, public liability and goods in transit. The
company does not buy product liability cover.
(i)

Discuss the features of past loss experience and exposure that you would need
to consider in advising the companys directors on its insurance premium. [18]

The company is part of a world-wide manufacturing group with subsidiaries in other


countries. The group is considering establishing a captive in one of the non EU
countries in which it operates.
(ii)

Define the term captive.

[2]

(iii)

State the advantages and disadvantages to the group in setting up this captive.
[6]
You have been asked to advise on:

(iv)

the capital requirements of this general insurance subsidiary.


the establishment of its reinsurance programme.
Comment on the factors that you would take into account in formulating your
recommendations.
[15]
[Total 41]

END OF PAPER

SA3 A20084

Faculty of Actuaries

Institute of Actuaries

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT
April 2008
Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

M A Stocker
Chairman of the Board of Examiners
June 2008

Comments
Individual comments are shown after the solutions to each part question that follows.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

(i)

A claims made policy is one that covers all claims reported to an insurer
within the policy period irrespective of when they occurred.

Comments on Question 1(i): Bookwork question with most candidates scoring well.
(ii)

+ Limits exposure to latent claims via recent retroactive date


+ Quicker reporting of claims to ensure that coverage is triggered
+ Greater clarity on which insurance period is triggered by a claim
+ Reduced chance of (expensive) legal action between insurers to assess who
is on-risk for a claim
+ Insurer can determine profit/losses more quickly less uncertainty
+ Potentially easier to reserve
- Notification of greater number of claims/circumstances
- Risk of moral hazard where an insured takes out a policy being aware that a
claim has incurred in order to claim under the claims made policy if the
insured has not been on claims made cover before
- LOD is the norm so out of line with the competition
- claims emerge from different periods of exposure

Comments on Question 1(ii): Generally well answered by most candidates. Better


candidates recognised that exposure to latent claims could be reduced via a
retroactive date. Poorer candidates incorrectly stated that moving from a losses
occurring basis to a claims made basis created gaps in cover.
(iii)

Although all claims have to be reported within the exposure period of the
policy, this does not mean that they will be paid within this period so
uncertainty remains
Most annual policies will still have exposure after end of year
There may be policies for longer than one year
or binders/lineslips meaning further claims possible after more than one year
Claims could re-open
May be options for extension of notification period
Might expect more notifications to come late in the policy term
IT delays may cause further delay
There may be little information available on a claim / notification of
circumstance when it is first reported, so it may be difficult to set up an
appropriate reserve
This could be especially true where companies laundry list claims
attempting to ensure that all potential incidents which could lead to a claim
(however unlikely) are reported to the insurer so that they would be covered
under the claims made trigger
The claims made trigger has been in place for some years, and there are
increases in incurred claims in years 4 and 5 of development
Public liability covers damage to 3rd party property and bodily injury claims.
The latter can be subject to a high degree of uncertainty and may take a long
time to settle

Comments on Question 1(iii): Reasonable attempts were made by most candidates


on this question. Better answers appreciated the potential for new claims on the most

Page 2

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

recent year even though the basis is "claims made", and used the data provided to
demonstrate the need for IBNER.
(iv)

Range requirement
Reserve report likely to be formal in GN12 terms
GN12 states that report should normally indicate the nature, degree and
sources of uncertainty surrounding the results
and sensitivities to key assumptions
Uncertainty should normally be quantified where practicable
but otherwise should normally be reported using an appropriate descriptive
summary
This would suggest that it might be necessary to provide a quantitative range
of reserves in order to be compliant with Institute guidance except where
impractical when a description might suffice
Consider GN50 which states that uncertainty surrounding advice or opinions
formed must be considered and communicated appropriately
The less likely the audience is to appreciate the importance or extent of
uncertainty the greater is the need for this to be communicated
In this case the audience may be unlikely to appreciate extent of uncertainty
without some quantitative calculation e.g. a range

Comments on Question 1(iv): Most candidates mentioned GN12 and some of the
advice therein. Few candidates demonstrated full knowledge of the relevant parts of
GN12 and GN50. Better candidates identified the need to consider communication in
the context of the audience of the report.
(v)

+ No need for any proprietary software or programming skills


+ Very flexible to different situations
+ Easy to involve other stakeholders e.g. underwriters
+ particularly important for this line of business as limited data may be
available due to change to claims made in 2002
- Highly dependent on the experience of the actuary applying the judgement /
not objective
+ This could be an advantage where actuary has appropriate experience/skill
+ Should be easier to explain to any audience than more technical methods
+ Potentially allow for model error as well as parameter and process error
- Difficult to avoid being anchored to estimates provided in different
context
- little data on extreme events so difficult to model catastrophes
- difficult to model correlations
- difficult to check / peer review

Comments on Question 1(v): Many candidates scored well here. Better candidates
went beyond general comments such as "easy to do" (which is far from clear!) and
gave opinions as to which aspects might require specialised judgement and the
associated difficulties.

Page 3

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

(vi)

Bootstrapping characteristics:
A method of estimating the parameter uncertainty surrounding an estimate of
the reserves
To estimate process uncertainty need to use in conjunction with e.g. Mack or
over-dispersed Poisson model
Estimation achieved by repeated re-sampling with replacement from the
historic data to produce a large number of alternative pseudo-data sets
consistent with the original data.
Each of these alternative data sets is projected using the chosen projection
method to give an alternative reserve estimate for each re-sampled set of data
By repeating this process thousands of times we can generate standard
deviations, confidence intervals
Can be applied to paid or incurred data, and accident year or underwriting year
cohorts
Open to manipulation
Bootstrapping- pros/cons
+ Easy to apply for most datasets
+ customisable
- No allowance for tail factor
- Basic method very restrictive in terms of how development factors are
selected
+ however method can be applied to subjectively derived development factors
Macks Method characteristics:
An analytical method based on the chain ladder for estimating the uncertainty
inherent in the reserve estimate for a given accident or underwriting year
A standard chain ladder method is applied to the cumulative triangle to
determine the incremental development factors.
Variability between the actual and expected development at each point in the
triangle is calculated.
Then the variability across the rows is aggregated to produce a standard error
for each accident/ underwriting year
Can extend to derive a standard error of the overall reserve estimate
However, if percentiles are required, in order to produce a range, a distribution
needs to be assumed via a deterministic calculation or bootstrap approach
Based on chain ladder so assume underlying chain ladder assumptions
appropriate
Can be applied to paid or incurred data, and accident year or underwriting year
cohorts
Macks method pros/cons
+ No assumption of prior distribution
+ A tail factor can be incorporated as a deterministic multiple
-Limited judgement possible
Both methods pros/cons
+ Require few assumptions

Page 4

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

+ Easy to use can be run in Excel or proprietary software may contain a


version of the method
+ Increasing usage of methods in insurance industry
- Dependent on the quality of data used
- Output may reflect variability of data which is a feature of data
errors/inconsistency rather than the underlying claim features
- Any variability not included in the data will not be reflected in the derived
range
- This is a particular problem where limited data are available which is likely
to be the case for this line of business
- Difficult to explain to non-technical audience
+ Objective
+ Can audit and peer review
Comments on Question 1(vi): Generally poorly answered. Alternative acceptable
answers were given equivalent marks for this question, e.g. giving a brief example of
bootstrapping in place of an explanation. Many candidates gave a reasonable
explanation of bootstrapping; fewer demonstrated knowledge of Mack. Most
candidates scored poorly on identifying the advantages and disadvantages of each
method. Better candidates identified that Macks method could be used as part of a
bootstrapping exercise. Not many candidates picked up on the marks available for
non-technical observations such as the need for sufficient data, failure to reflect
variability not in the data etc.
(vii)

Inconsistent with range


Latest ultimate is not inconsistent with range
Latest ultimate is greater than 90th percentile at year end
One in ten years, might expect ultimate to be greater than 90th percentile
Ultimate is below 95th percentile so still within this part of range
Change in ultimate largely driven by huge (27.5m) increase in incurred
claims in 2005 policy year
This might be caused by a single unexpected large claim, accumulation of
claims, class actions
Public liability is always exposed to such claims
2006 policy year has also seen worse than expected development
Earlier years have not seen large increases
so may not be indication of need for heavier tail factor
Impact on ICA
Unlikely (in itself) to have significant impact on ICA
Only one line out of several other lines may have seen better than expected
development
ICA calibrated to 1 in 200 year event
Increase may not be that significant in comparison to overall reserve size
Diversification benefits from multiple lines
Need to consider impact of reinsurance
In particular if increase caused by single large claim with XOL reinsurance
Many other elements than risk of reserve deterioration included in ICA:
Operational risk
Credit risk
Page 5

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Market risk
Liquidity risk
Group risk
Might impact ICA if change not because of volatility but e.g. legal ruling
Comments on Question 1(vii): Many candidates failed to answer the question as
they did not specify whether they agreed with the comments. Some also did not
explain the significance of other influences on the ICA.
(viii) Observations
Loss ratio is volatile over time
Varies from 36% to 260%
Loss ratio for 2006 should be treated with caution however as at an early stage
of development
Last 2 years it is significantly above 100% suggesting unprofitable
Consideration of non-claim elements (expenses, investment income highly
unlikely to compensate for loss ratio > 100%)
But potentially a large claim in 2005 distorts numbers
Trend in loss ratio appears to be upwards
Since this is associated with increasing premium volume this is a particular
concern
Might be growing book by offering lower rates
Or trying to expand book in soft market
Or antiselection / competitor rating
Claims made features
Business written in early years might be from clients moving from loss
occurrence to claims made
In this case early years might have low exposure to claims (due to slow
emergence of claims so few claims reported which do not trigger prior
insurance)
As account matures there is a full pool of earlier years generating claims
so greater number of claims trigger claims made policy
If these features are not appropriately allowed for in the pricing, a worsening
profitability trend might be observed
Further information
Need further information before any strong conclusions
Knowing the reinsurance structure for the class and the reinsurance spend/
recoveries in past is crucial
Net loss ratio could be much lower due to e.g. excess of loss protection
But if this is the case reinsurance premium likely to increase in future
Need benchmark profitability requirement e.g. X% ROE over market cycle
Must estimate the capital required for business
Taking account of diversification
Investment income estimated
e.g. based on mean term from payment pattern
and interest rate
Expenses investigation required
At least estimate of expense ratio

Page 6

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Ideally split of expenses into constituent parts


Fixed/variable etc.
Estimate claims handling costs
Commission level unknown (assuming premium is gross of commission)
Rate change information / rate adequacy on new business
Further information on large claims
Need discussion with underwriter of account
and claims handler of account
in particular to understand drivers of poor performance in 2005
Credit terms for premium
Tax rate
Monitor attachment/limit profile
Mix of business change e.g. territory, industry
Consider position in market cycle
More detailed exposure information (e.g. terms and conditions)
Split of business into new/renewal
Undertake full profit testing exercise
Benchmark against competitor loss ratios (if possible)
Any reasons for running a loss leader, e.g. need to offer product to secure
profitable business on other lines
Investigate any changes in legislation
Number of claims triangulation
this would assist in observing separate frequency / severity trends
Comments on Question 1(viii): Candidates generally scored well on this question,
with the better candidates considering the potential impact in performance of
changing from claims occurring basis to claims made basis in 2002. Most candidates
showed that although results appeared poor for the last two years, more information
was needed to put this in context. Some candidates suggested that a reason for the
good results in the first two years was because of having a choice of who to claim
from, i.e from the current claims made insurance or previous losses occurring. This is
highly unlikely as a retroactive date would be used to avoid overinsurance.

(i)

Some of the classes of insurance will have seen claims experience change in
character over the last ten years because of the companys expansion. Others
will have only been subject to moderate change.
If there have been no claims, it will be difficult to allow for changes in
exposure
Commercial Fire
The number of sites has increased by over 50% over the past ten years,
implying that the latest claims experience arising from exposure now may be
very different to that of ten years ago.
There may have been a number of site sales and acquisitions over the period,
adding further exposure changes and therefore impacts on claims experience.
There may have been changes in other risk factors.
Example of other change in risk factors, e.g. age of buildings different

Page 7

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

As the business has expanded, productive use per unit floor space has most
likely increased.
Need to consider the number and amount of gross losses per location over
period
And sum insured
Review of sum insureds over 10 year period
Consider relevant rating factors
Key aspect in the past claims experience will be whether or not any large
single fire losses have occurred during the past ten years.
Unless the company has been unlucky with large losses during period, the
claims experience will most likely be very low and not take into consideration
the additional premium to charge for expected large losses.
Do any market statistics exist on much greater size portfolios of similar
exposure mix with more credible large loss experience?
Is the cost of small claims abnormally low?
If it is then justifies a lower premium rate
Are there particularly good safety procedures in place?
Does the type or age of construction of the buildings warrant a lower or higher
premium than a traditional book rate?
Change in socio-economic factors
Business Interruption
A loss of this kind is only likely to occur after a major or total loss as the
company may well now be potentially large enough to have business
continuity plans in place.
Therefore premium may look high in relation to past claims
Furthermore, if a claim does occur now, it is likely to be much greater now
than it would have been ten years ago because of far greater output.
Existing claims experience is unlikely to be adequate to be able to use in
isolation for assessing a suitable premium rate.
Specialized machinery requiring a long lead time for replacement could
increase potential risk.
Exposure measure turnover / profit
which could be volatile
Consider relevant rating factors
Employers Liability
If the exposure available was over a long period, with stability in numbers of
employees and working conditions, the companys own experience may well
be a good guide for small claims
E.g. payroll may be used as an exposure measure.
Analysis of clerical versus manual payroll
Consider relevant rating factors
However, regard must be made to the chances of unsuspected industrial
disease claims such as deafness or vibration white finger.
10 years is insufficient to establish the potential for such latent claims
It may also be necessary to allow for the presence or absence of any abnormal
claims or accidents to several employees at the same time.
Changing environment and technological progress may also have an impact on
claims experience.

Page 8

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Awareness of health and safety issues might have improved claims experience
Greater productivity per employee may imply a higher real salary than ten
years ago, as more responsibility is placed on individuals. This may result in
much larger claims now than ten years ago.
Public Liability
Exposure measure turnover
Consider relevant rating factors
This is likely to cover mainly premises risks.
Also external factors such as environmental exposure
The number of premises has doubled over the past ten years and with floor
space increasing and potentially a greater number of third parties visiting each
premise, the experience now will be very different to that of ten years ago.
Changes in internal environment may impact on claims experience.
Claims experience may have been very light during the ten years particularly
for large losses.
It will be important to analyse trends in claims experience compared with the
number of premises or other exposure measures such as turnover.
Goods in Transit
Acceptable exposure measure e.g. sum assured
Consider rating factors e.g. distance travelled, frequency of travel, hub in
travel, methods of transport
Lots of new premises might mean more internal shipping rather than external
shipping
Past experience should be a reasonable guide unless rating factors have
changed significantly
Fleet
Vehicle year as exposure measure
Consider relevant rating factors
The experience over the past ten years should be a good guide as the
workforce has been stable and therefore the number of vehicles should be
broadly similar.
Key issue will be how the mileage per driver has changed over the past ten
years.
Are vehicles travelling much greater distances now as a result of road
improvements or has the company been using alternative forms of transport
such as rail, shipping or air?
Has changed manufacturing processes impacted on the type of transport used?
Ideally consider losses broken down into vehicle type
Are larger vehicles being used now compared to ten years ago?
Different drivers to previous years
Need to consider losses split into physical damage and bodily injury due to
different level of inflation
Large claims will need to be truncated at a certain level to remove abnormally
large claims with suitable adjustment for a long term allowance for large loss
derived from a number of years data within the premium rate calculation.

Page 9

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

General consider for all products


Claim frequency
Average severity
Analysis by cause
Analysis by year
Trends on all the above
Comments on Question 2(i): This question was poorly answered. Few candidates
considered the context of the question. Many did not answer the question being asked,
instead deciding that this was asking for an explanation of how to conduct a pricing
review. Better candidates demonstrated an understanding that each class of business
would exhibit differing experience in losses and exposure during a period of rapid
expansion with a stable workforce.
(ii)

A captive is an insurer wholly owned by an industrial or commercial


enterprise and set up with the primary purpose of insuring the parent or
associated group companies
and retaining premiums and risk within the enterprise / form of self
insurance
Some captives are set up with the primary purpose of selling insurance to the
customers of the parent.
or alternatively they may insure other non Group companies if they have
the expertise

Comments on Question 2(ii): Bookwork question. Better candidates identified that


the captive could sell insurance to customers of Rapidco or other non-related
companies.
(iii)

Page 10

Advantages
Rather than passing insurance profits onto external insurers, the company
retains these.
This could therefore further improve the profitability of the groups accounts
on a consolidated basis
Could also benefit from further profits by selling products to customers of the
company e.g. warranty
Promotes greater awareness to senior management of managing risk within the
company rather than passing to an insurer.
A captive may pass on good experience and risk management improvement
savings through lower premiums quicker than an external insurer
Direct access to reinsurance markets and expertise.
Or obtaining cover that cannot be obtained by a direct insurer.
Assists in negotiating desired cover, terms and conditions with reinsurers
There may be tax advantages if located in certain domiciles, e.g. Bermuda,
Channel Islands
Reserves/premiums are built up as pre-tax profits
Reduces insurer credit risk exposure
Could select against insurance market by increasing retention in captive when
in hard market

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Disadvantages
Ties up the groups capital so potential conflict of opportunity cost of captive
versus alternative better returns in group activities
Increased volatility in group results dependent on reinsurance retention levels
i.e. lack of risk-transfer
Costly and complex to set up and subsequent running costs
May pull management time and resource away from main Group activities.
Location may be a factor here.
Might not have the economies of scale of an external insurer so may overall be
more expensive even though not giving away profit margin
Accumulation of risk
Comments on Question 2(iii): Bookwork question, generally well answered by most
candidates.
(iv)

Capital requirements
This will depend upon whether the subsidiary is to be restricted to the
manufacturers group business or not.
If it is restricted, the need for capital may be fairly low since this selfinsurance arrangement is basically equivalent to internal group accounting.
Although the actual capital requirements will depend on the domicile of the
captive
The statutory minimum requirement is required as long as the parent realises
that it must subscribe more capital in order to maintain the minimum at least at
every year end for the purposes of declaring financial accounts to the
regulators and shareholders.
If the company is to be seen as an independent trading entity which is
accepting business from elsewhere, then it needs enough capital to show the
level of solvency margin expected of other insurers.
This will almost certainly be greater than the statutory minimum requirement.
Capital and solvency levels will be a critical factor in the captives ability to
attract business as a measure of its security to meet claims as they fall due.
The regulators of the country in which the captive will be set up may also
demand a far greater level of security for independent policyholders
The parent will need to balance the capital employed against return achieved
against the alternative use of capital within the manufacturing company.
Consider changes in future solvency regulation.
Generic capital points e.g.
volumes & growth
initial costs
adequacy of premium rates
use of reinsurance
expected volatility
adequacy of reserves
investment strategy

Page 11

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Reinsurance programme
The reinsurer will require full details of all the group insurances it will write in
order that it can rate them.
Setting the level of retention will be important
a lower retention will attract increased claims supervision by reinsurers.
. as they will want to ensure that adequate risk management procedures are
in place to limit claims in size and volume.
The company may wish to purchase reinsurance initially to benefit from
technical assistance
and to provide financial support.
The costs in providing reinsurance for this captive may be prohibitive if it is
small in relation to other insurance companies who have greater purchasing
power.
It might be very difficult to place proportional reinsurance since the company
could write business at artificially low premiums in agreement with the
holding company
.. with all subsidiaries being then required to pay supplementary premiums if
necessary.
If the captive could persuade reinsurers that it was rating risks on standard
market rates and underwriting correctly, it might obtain reasonable terms.
Its requirements would then be fairly standard, e.g.
Surplus treaty for commercial fire and business interruption
Excess of loss for motor, liability and goods in transit
Quota share outward/inward with other insurers/captives may also be an
option to reduce claims volatility and reduce accumulation of risk
Catastrophe excess of loss to reduce exposure to concentration of risk in
one area (all property and casualty covers)
Generic reinsurance points e.g.
Consider the groups overall risk appetite
Consider what reinsurance is available and the cost
including possible profit-sharing arrangements
Consider alternatives to reinsurance available, e.g. more capital from parent.
Consider net impact on capital requirements
Consider security status of available reinsurers.
Fronting could be used by captive if convenient or cost-effective
Comments on Question 2(iv): Although reasonably well answered, most candidates
missed marks for identifying the differences in capital requirements between open and
closed captives. A number of candidates discussed in detail the current UK
regulations on capital and Solvency II when the question clearly states that the
captive will be set up in a non EU country rendering these comments largely
redundant. Few candidates discussed the pros and cons of setting the desired
reinsurance retention limits. There was a tendency to concentrate on general capital
and reinsurance points rather than considering the specific issues for the captive.

Page 12

Subject SA3 (General Insurance Specialist Applications) April 2008 Examiners Report

Comments on Question 2: This question was based on a question from an early


90s exam paper as was coincidentally an ActEd assignment. Very few candidates
who may have done the assignment appear to have benefited from this.

END OF EXAMINERS REPORT

Page 13

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
22 September 2008 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 S2008

Faculty of Actuaries
Institute of Actuaries

You are an actuary working for a large, diversified insurance group based in the UK.
You have been recently transferred to the Megarisks Division (MAD). MAD writes
high layer product liability insurance. It accepts business via the slip system in the
London Market. It tends to take large lines on the policies it writes, and does not
participate on policies with excess points less than 100 million. The policies cover a
range of major multinational companies.
The director of MAD has asked for your assistance. Like you, the director has only
recently transferred to MAD. He previously managed the UK household property
division, and has no experience of product liability business. The director would like
to understand more about product liability insurance.
(i)

Define product liability insurance.

[1]

(ii)

Give an example of a claim that could be covered by an insurance policy


written by MAD.

[1]

(iii)

List the main factors MAD would be likely to consider when assessing the risk
under each policy, stating why each factor is relevant to the risk.
[9]

MAD started writing business in 2002. The written premium, paid and incurred
losses for each underwriting year as at 31 December 2007 are summarised below.
Amounts are in millions. MAD does not purchase reinsurance.
Underwriting Year

Written Premium

Paid Claims

2002
2003
2004
2005
2006
2007

5.0
10.0
15.0
20.0
20.0
15.0

0.0
0.0
0.0
0.0
0.0
0.0

Incurred Claims
0.0
30.0
0.0
0.0
0.0
0.0

There is only one claim for which a reserve for indemnity is held. This claim is on
the 2003 underwriting year and was reported during 2005.
(iv)

Describe the claim characteristics of the product liability business written by


MAD.
[4]

(v)

Describe how you would produce a best estimate of unpaid claims for this
business, in respect of all underwriting years as at 31 December 2007. Both
discounted and undiscounted estimates are required. Note that you are not
required to produce an estimate, but rather to describe how you would go
about producing one.
[19]

SA3 S20082

The director of MAD has told you that he does not want the reserves to deteriorate.
He would like to understand how confident you are that the actual unpaid claims will
be no more than your undiscounted best estimate.
(vi)

Outline points you would make in a meeting with the director regarding:
(a)

The requirements in GN12 for an actuary to comment on uncertainty in


a formal report.

(b)

The likelihood of future claims payments for this portfolio differing


significantly from the estimated undiscounted reserve.
[7]

The director is keen to monitor changes in the adequacy of premium rates in the
portfolio.
(vii)

Describe ways in which you could monitor changes in the adequacy of


premium rates from 2007 to 2008, commenting on the practicality and likely
effectiveness of each.
[16]

The director has recently read an article about insurance-linked securities (ILS). He is
considering creating a bond which could be sold to the capital markets, which would
provide funds to MAD in the event of a deterioration in the reserves.
(viii)

Comment on whether the directors plan is likely to be viable for MAD,


including a recommendation to the director on whether to proceed.
[5]
[Total 62]

SA3 S20083

PLEASE TURN OVER

Quote-u-online is a new UK general insurance company that has recently been set up
with the aim of selling personal lines insurance direct to the UK public exclusively
through the internet. The company will advertise in newspapers, on television and the
internet and will also be included on all the major price comparison websites. All
quotes given and claims notified will be through online screens only. Telephone
support is only offered to sales and claims enquiries.
The companys underwriters devised the premium rating structure and rating levels.
At the time of the launch, the Chief Executive (CEO) has asked you, as a consulting
actuary, to explain the areas of uncertainty that may exist within the premium rates.
He has not shown you the proposed premium rates but asked for your report to focus
on the issues rather than the proposed rates.
(i)

Discuss the matters that you will need to cover in your report, with reference
to the main products that are likely to be sold.
[18]

It is now two years since the launch and the company has been successful in acquiring
business, especially in the private motor account.
The CEO has requested that you provide an independent assessment and report of
how profitable the current private motor account is. You will be given access to
whatever company data you require. However, the CEO confides in you that he
believes that the information from the companys case estimators is of little use.
(ii)

(iii)

Explain what GN12 requires you to include in your report in respect of


information and data you use.

[5]

Describe how you would produce such an assessment assuming that the case
estimate information is of poor quality.
[15]
[Total 38]

END OF PAPER

SA3 S20084

Faculty of Actuaries

Institute of Actuaries

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT
September 2008

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

R D Muckart
Chairman of the Board of Examiners
December 2008

Comments
Individual comments are shown after the solutions to each part question that follow.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

(i)

Define Product Liability Insurance


Insurance that indemnifies the insured against legal liability for death or
bodily injury to a third party, or for damage to property of a third party, that
results from a product fault.

Comments on Q1(i): This definition question was generally well answered, although
a disappointing number of candidates simply regurgitated the question with some
form of This offers cover for liabilities arising from products without mention of
the nature of the liabilities and many did not mention third party.
(ii)

Example Loss
Possible examples include faulty motor components causing a large number of
motor accidents, pharmaceutical product liability relating to a widely
distributed medicine, or failure of a single very expensive product (such as an
industrial turbine). Many other examples are possible, but the example should
be extreme enough to potentially produce losses excess of 100 million, and
be within the scope of product liability insurance.

Comments on Q1(ii): The majority of candidates correctly identified that the


company in question wrote only at very high attachment points and gave an
appropriate example (most frequently a pharmaceutical claim) although some
candidates did miss the point of the question giving examples of events that would be
unlikely to lead to any claims of a remotely high enough order of magnitude.
(iii)

Rating Factors
Nature of product/industry type
Certain products tend to experience a higher frequency and severity of losses,
e.g. stationary manufacturers would tend to have a lower loss potential than
pharmaceutical companies.
Turnover or payroll
Requires high turnover to pose a realistic risk to the high layer
Geographic location of sales and geographic location of manufacture and
related quality control laws
E.g. litigiousness in the US / separate US and non-US turnover figures may be
requested
Latent claims / amount of the product already sold and used
Packaging instructions and reason
Subjective factors are also likely to be considered by the underwriter, e.g. his
understanding of the insureds risk management systems.
Claims history may be considered
Possibly using a lower claim threshold
However, claims history will frequently be limited, or of little relevance to the
current risk environment.
Consider claims history of similar companies

Page 2

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Limit/line/excess
Attachment points impact likelihoods of claims / large claims
Whether cover is on a claims made or losses occurring basis
Whether a sunset clause is included and/or RDI.
Writing business on a claims made basis/with a sunset clause allows the final
underwriting result to be determined more quickly
Under occurrence business, new claim notifications may be received many
years after the policy has expired.
Whether losses can be aggregated by event
Whether there is an aggregate deductible
Level of aggregation with that risk and the rest of MAD's portfolio
Treatment of legal expenses (excluded, included in addition to limits, included
within limits)
Comments on Q1(iii): Candidates almost invariably identified such key factors as
the locations sold and type of industry/products produced, although many went into
extensive detail on the industry/product type while missing other key aspects of the
risk such as the policy terms or attachment points.
(iv)

Claim Characteristics
Liability claims tend to be long-tailed, i.e. claims may take many years to be
notified.
The slow notification is likely to be a pertinent feature of MADs experience as
products that have continued to be used for a significant time before discovery
and notification are more likely to hit the excess point
If MAD writes policies on a claims made basis this will affect the
development profile of the risk, depending on the time limits for reporting
Case estimates are often highly uncertain.
Uncertainty in respect of reported losses relates to the existence of liability as
well as its quantum
Settlement can be a lengthy process involving legal action, particularly for
claims of this magnitude
Claims are heavily affected by legislative changes. There may be issues that
lead to claims purely on this basis.
Claims are heavily affected by inflation, including general, wage and court
award types, and inflation is heavily geared for MAD as it is an Excess writer
The outcome of the settlement process might be that the insurer is not liable
for the claim, e.g. because it is not covered by the policy, or because the final
claim is below the excess point of the layer.
However, the insurer would likely incur legal and other costs in handling some
of the claims received, even if no indemnity is ultimately payable.
The majority of claims hitting MAD's excess point are likely to be the result of
catastrophes and accumulations
Re-opened claims.
Latent claims can be an issue, with claims often not noticed for a while.
Payment characteristics - periodic / lump sum
Litigiousness
MADs claim frequency is likely to be low because few claims would be
expected to exceed the high excess points at which it writes.

Page 3

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

MADs claim severity is likely to be high as if a claim reaches the high layers
insured it is likely to be very large.
The data given for MAD appears to be consistent with low frequency, high
severity claim experience, and losses taking several years to settle.
Comments on Q1(iv): Candidates generally scored well on this question, although
the number of potential points available was well in excess of the maximum mark
leaving a high score comparatively easy to obtain.
(v)

How to produce a best estimate of unpaid claims


Review of Outstanding Claims
The 30 million outstanding claim requires separate analysis.
Investigations could include discussion of the claim with claim staff,
discussion with other experts such as underwriters, and review of legal or
other expert opinions that have been obtained by the company.
Consideration may be given to the overall market loss with MAD's exposure
then worked out as a proportion depending on the line size and attachment.
points.
The loss is unlikely to settle for the current case reserve and may settle higher
or lower.
General Analysis
The actuary should meet with underwriters and claim staff to understand the
business written in more detail.
Review previous reserving methods and assumptions.
Data splitting / portfolio segmentation
Split of Exposure by country/region/currency
Split of Exposure by industry/product type
Split of Exposure by size of insured
Split of Exposure by limit/excess
The characteristics of the portfolio may have changed over time (e.g. change
in mix of business), and such features should be understood.
Premium changes and reasons
It is necessary to understand how the underwriters price the business.
The premium written has varied from year to year. It would be useful to
understand the reasons for this.
How much of the change is due to increases in the size of the portfolio.
How much is due to a change in mix of business.
How much of the change is due to the insurance cycle.
Additional data, particularly lower layer information
The underwriters or claims staff may be aware of potential claims which are
not included in the outstanding claims data.
E.g. they may have been notified by the broker of losses on lower layers
written by other insurers, which are considered to have the potential to
deteriorate to higher layers written by MAD.
Techniques such as extreme value theory could be used to estimate extreme
values from a limited data set.

Page 4

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

This could be used in conjunction with a stochastic method to estimate the


potential excess losses based on the lower layer experience.
Loss curves used in underwriter rating guides could also be used to adapt low
layer experience to an estimate of MAD's exposure
Specific IBNR estimates can be produced in respect of such losses.
There may be awareness of industry/product types that require separate
analysis, because there is considered to be a particular risk of claims.
Policy wording (e.g. sunset clause) may mean that no further claims are
possible on some of the older underwriting years.
For such years, reserves would only be required for reported claims, and the
required reserve may be nil.
Projection methods
Limited claims data means it is not possible to apply standard projection
techniques such as chain-ladder to the account.
It would be useful to apply a number of different methods to estimate unpaid
claims.
Techniques based on initial expected loss ratios could be used, e.g. the
Bornhuetter-Ferguson method.
Loss ratios used would need to reflect the insurance cycle.
Frequency-severity modelling
A frequency-severity model (or average cost per claim model) could be
constructed.
Although even at the lower layers claims data are unlikely to be sufficient to
accurately parameterise this.
Exposure based methods
An exposure based method based on the underwriter's rating model could also
be used.
This would involve rating each risk individually against a rating model
assumed to produce a particular loss ratio and aggregating the data to portfolio
level.
The capital model may also provide some estimates of claims levels.
E.g. the models may contain information on pricing and loss assumptions.
External data sources and benchmarking
relevant internal data from other areas of the business
cedant/policyholder data
industry data/benchmarks
data from regulatory returns
reinsurers data: must specify that it is parent group's reinsurance as MAD
doesn't purchase
expert judgement
engage external expertise (e.g. consultants)
rate change indices
External benchmarks might not be representative of this portfolio, and so
require adjustment before they can be used.

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

It may not be possible to objectively adjust the data so some subjective


assumptions will be required..
Discounting
Benchmark payment patterns could be used to discount claims.
Various legislation will affect this depending on the purpose
Discounting rate should be set with reference to the investment returns
achieved
Comments on Q1(v): This question appeared to trouble many candidates who clearly felt
uncertain how to proceed when standard statistical methods were not practical due to a lack
of claims experience, although the majority of candidates at least correctly identified that
statistical methods were not practical and many were able to suggest sources of information
that could help the actuary to produce an estimate, for example, benchmarks or expert
advice. Better candidates considered methods such as review of lower layers experience or
stochastic modelling. A surprising number of candidates however failed to make any
comment about what the actuary might do with the data already available to produce an
estimate, in particular review of the single existing claim which would be a natural first step
in any such claims review.
(vi)

(a)

GN12 Uncertainty Requirements


The report should normally indicate the nature, degree and sources of
uncertainty surrounding the results and sensitivities to key
assumptions.
Uncertainty should normally be quantified where practicable, but
otherwise should be reported using an appropriate descriptive
summary.
If there are specific features of the business that present potential
concerns or significantly increase the uncertainty of the results, beyond
that which an informed reader of the report would reasonably expect,
then this fact must be clearly highlighted in the corresponding
reservations or limitations of scope, included in the report.
If there is a substantial probability of material adverse deviation from
modelled results, attention should normally be drawn to this in the
report.

(b)

Uncertainty in this portfolio


The claim characteristics of the business written are such that the
difference between the actual unpaid claims and the best estimate
unpaid claims may be large.
It would be impossible to hold large enough reserves to guarantee no
adverse deviation as claims can effectively be unlimited.
In any case, accounting regulations may prevent such a reserving
policy.
The expected claim frequency for high-layer product liability business
is low, which makes the overall claim numbers significantly more
volatile than more attritional business..

Page 6

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

If the portfolio written is concentrated in certain areas the uncertainty


in the reserves is increased.
E.g. concentration risk may occur if a large number of assureds make
similar products, or very large limits are written for some assureds. (Or
other relevant example.)
Any deterministic reserve estimate only gives limited information for
an account of this type and a stochastic approach would give a better
indication of the scope of the volatility.
Individual claim sizes are uncertain, and can be very large. This means
that a single claim could have a very large effect on the liabilities of
MAD.
The high attachment point means that a small variation in the ground
up claim amount will have a disproportionate effect on MADs
liability.
It may be possible that no payments are ultimately required on the
outstanding claim. This could mean that the estimated unpaid claims
are overstated.
Because MAD does not purchase reinsurance any improvement or
deterioration in the claims will be entirely for the account of MAD,
rather than possibly being shared with reinsurers.
Comments on Q1(vi): In spite of GN12 featuring regularly in SA3 exams and
knowledge of professional guidance being a critical requirement for qualification as
an actuary, many candidates displayed only the most basic grasp of the requirements
set out for commenting on uncertainty. The most common error was to go into
extensive detail about tailoring communication to the audience involved. Many
candidates also missed the point of the second section, commenting in great detail
about aspects of general product liability while failing to comment at all on the
uncertainty and volatility caused by such high attachment points.
(vii)

General comments
It should be noted that premium rating on this account is necessarily
subjective.
As a result of this subjectivity, any method of premium rating will be
approximate.
Year to year premium comparisons
For stable portfolios, the overall premium can be compared from one year to
the next
Alternatively, for risks written in both 2007 and 2008, compare the premium
charged in each year to get a risk level movement and aggregate it to portfolio
level.
These approaches are simple and practical but require reasonable levels of
stability in the portfolio.
The methods will not pick up the effects of new and lost business.

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

If a material proportion of the portfolio is stable one can use movements on


the renewing element of the business as a basis for estimates of movements on
new and lost business.
These methods become more useful the more stable the business is.
Adjustments for exposure changes for year to year comparisons
When looking at movements on individual risks, policy conditions such as
limits and deductibles will almost certainly change from year to year, and the
effect of such changes on rate adequacy will need to be considered.
A table of increased limit factors could allow the effect of such changes to be
quantified.
Such tables could be applied simply, but may not be accurate.
The nature of the risk insured will change from year to year.
E.g. the insureds turnover may increase, or it may start manufacturing new
products (or other example).
Allowances for the effect of such changes on rate adequacy may rely heavily
on judgement.
Rate-on-line comparisons
An index could be produced showing rate-on-line or some equivalent measure
charged in 2007 and 2008.
There may be issues in producing this measure where liability is unlimited;
and for business at these attachment points such a measure is often of limited
value
This index could consider both new business and renewals.
Separate indices would need to be produced for different rating cells, since a
different rate-on- line would be charged for different types of assured, so a
change in mix of business could distort the calculation.
Such indices could be produced relatively easily from a detailed policy
database.
It may not be possible to fully remove heterogeneity from the data without
reducing the credibility in each sample.
Pricing tool / individual risk pricing
Underwriters may use a pricing tool to assist them in pricing risks.
A pricing tool/software could be produced to give an indication of the
technical price for each risk.
An estimate of the technical rate might be produced by a pricing actuary for
each risk.
To monitor rate adequacy, the actuary could compare the rates actually
charged to an indicated technical rate for both 2007 and 2008 underwriting.
It may only be practical to apply this to a sample of policies.
This method could consider both new business and renewals.
The indicated technical rate should allow for rating factors such as excess,
limit, industry type and size of insured, so the rate change produced should be
sensitive to changes in the mix of business in the portfolio.
However, the indications of technical price may not be accurate.
This process may also be expensive.

Page 8

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Subjective underwriter comments


The underwriter could be asked to comment on the change in rate adequacy in
the portfolio.
This approach makes use of the underwriters expertise.
The rate change indication is quick and easy to obtain.
The underwriter may be able to quantify the effect of variables that are
difficult to capture in a model, such as changes in the overall risk
environment:
e.g. changes in propensity to claim
Legislative changes
Changes in coverage e.g. costs in addition
Or movements in the insurance cycle / changing levels of competition etc.
It is difficult to judgementally allow for all the factors that could have an
effect on rate adequacy.
The results of this approach may be difficult to verify for auditors, reinsurers,
management and others who are interested in rate adequacy.
This could, however, be a useful check on the results of other methods.
The underwriters assessment may potentially be subject to bias (which could
be introduced intentionally or unintentionally).
Other miscellaneous factors
Underwriting file review/ peer review.
Portfolio movement analysis may highlight areas where rates are out of line
with the market, and identify specific areas to focus on
Claims inflation will affect rate adequacy.
Because of the high attachment point the effects of claim inflation will be
highly geared
Appropriate indices of prices should be monitored to allow an adjustment to
be made.
Examples of other factors that could affect rate adequacy include:
Changes in tax rates
Changes in expenses
Changes in commissions
Compared to the adequacy of technical rates, the allowance for such factors
will typically be relatively straightforward.
Comments on Q1(vii): This question was extremely poorly answered by the majority
of candidates. Most candidates amazingly made no reference whatsoever to such
absolute basics of rate monitoring such as looking at changes in premium or exposure
from one year to the next. A number of candidates made very confused comments
about monitoring the insurance cycle to find out what was happening with rates
rather than considering how they would review their own companys data to find out
the changes within their own portfolio.

Page 9

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

(viii) Insurance-Linked Securities (ILS)


A more usual type of ILS is a catastrophe bond, which the market is likely to
have more appetite for
Under these, if a defined index is triggered, the bondholders typically forfeit
the interest and principal on the bond to the insurer.
Unlike catastrophe bonds, the trigger for the ILS in this case is unclear.
It would be difficult to construct an objective method of establishing the
required change in reserves for this account.
Capital market investors may therefore be reluctant to buy the securities
because of the moral hazard.
Also, because this is an unusual type of bond, the expense of structuring and
marketing the security may be especially high.
May be difficult to persuade capital markets to invest in risks where standard
quantification tools have not been developed.
Capital markets may also demand a higher return because this is an unusual
type of risk with which they are unfamiliar.
However, this risk may provide diversification from other assets in investors
portfolios.
The portfolio written is fairly small, so the costs of an ILS may be prohibitive.
.Investors would probably want MAD to maintain an interest in the reserves,
so that claims are managed properly.
There may be alternatives that better meet MADs need, e.g. an internal
reinsurance within the group, or conventional external reinsurance.
Use of SPV would mitigate counterparty risk.
Another alternative could be to purchase adverse development covers.
Adverse development covers written in the market typically protect against
losses above a specified percentile.
Although issue of an ILS for MAD would be possible in theory, it is unlikely
to be viable in practice.
The director should be advised against proceeding with a capital markets
solution at this time.
Comments on Q1(viii): This question was generally reasonably well answered. Many
candidates correctly identified the key issue of the trigger for the ILS not being a clearly
defined and objective index and were able to provide an appropriate recommendation
to the director and the reasons for their advice.

(i)

General comments on internal & external data


No historical claims experience on which to base premium rates.
Where benchmark data have been used there will be uncertainty as to the
quality of the adjustments made to them.
Has it managed to recruit good quality underwriters
and are they basing premium rates solely on those of its competitors?
Internet/distribution method factors
The level of moral hazard associated with Personal Lines business may
introduce a level of uncertainty to the premium rating.
Moral hazard risk is increased for an internet channel as it is easy to just adjust
quote inputs to get different quotes out e.g. what if I said the car was in a
garage overnight?

Page 10

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Marketing/brand risk from site crashing or if it takes too long to get a quote.
Risk of expenses being greater than expected due to handling lots of telephone
calls from internet referrals.
Potential impact on fraudulent claims from economic downturn.
Risk of antiselection if entering the market with less sophisticated rating
structure than competitors.
This is a potentially serious issue due to the level of referrals from aggregator
sites which naturally highlight underpriced areas of the rating structure.
Motor
Volume is a key factor. If the company writes too little business, the fixed
expenses allowed for in the premium rates may not be recouped.
Persistency is also critical in the longer term as renewal expenses will be
smaller compared to initial expenses as well as broker market costs built into
their premium rates.
Claims experience is usually not very volatile with the exception of very large
individual losses.
Antiselection because of the poor rating structure is a more critical issue for
motor business than for others
The impact of investment returns is not a critical issue as motor mainly
consists of short tail damage claims with usually a smaller amount of longertailed liability claims.
Claims inflation is a material issue for the bodily injury claims.
Household
As with motor, volume is a key factor.
Household business tends to have better persistency than motor so may have
more difficulty breaking into this market with established players and a
sizeable market attached to the building society/mortgage channel.
Customers who regularly shop around direct through aggregator channels may
result in a lower persistency level once business is gained.
There is greater uncertainty from year to year on claim amounts than motor
due to more of the claims being linked to uncertain weather conditions such
as:
freeze leading to burst pipes
storm damage to properties
flood damage to properties
long, dry summer leading to subsidence
There are fewer critical rating factors compared to motor (location and sum
insured being the most important) and therefore less uncertainty from the
rating structure.
As this is generally a short tail class, investments are not a major issue.
Rebuild costs may be linked to inflation, so can be a significant risk
For both classes, the use of excess of loss and catastrophe reinsurance will
help limit some of the uncertainty in the claims experience. However this will
be at a cost.

Page 11

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Risk that notional sum insured used in rating engine for household quotes is
incorrect
Other classes
Other main personal lines classes of business that the company may be writing
include:
Travel
Pets
Creditor
Personal Accident
Warranty
The premium rating structures of these products are usually less sophisticated
with all insurers using broadly the same rating factors.
Claims are for relatively small amounts and are usually very short tailed.
Medical inflation (for Travel) and vets fees inflation (for Pets) are potential
areas of uncertainty as recent trends have shown an increase in these.
Expenses
Amount of expenditure on acquisition can be controlled by the company.
However volumes of business emerging from the advertising are highly
uncertain. So expenses per policy are often very difficult to predict.
The total level of expenses may differ from the amount assumed in the
business plan.
For example, the cost of hiring staff or obtaining premises may exceed
expectations (or other example).
The insurer may not be able to raise premiums to cover expense inflation due
to competition in the market.
Assessing the level of uncertainty with respect to volumes of business
Need to consider trends in methods of buying and selling insurance.
The stage in the insurance cycle will affect the volumes of business that these
particular rates will generate.
Current and likely future trends between the high street broker market,
telesales brokers and internet operations need to be forecast.
Assess the current number of internet based operations and their financial
position if known.
Likely future number and size of internet based insurers.
The history of existing internet based insurers (how quickly they gained
critical mass, how many have failed)
Impact of premium rates on volumes
Conversion rate is critical to volumes achieved.
Conversion rate will be highly price sensitive due to method of sale.
This will depend on the extent and speed with which the company revises its
premium rates in response to experience.
E.g. a company may find it can reduce its rates in a certain rating cell to
increase conversion rate without impacting significantly on profit.

Page 12

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Other more generic Risk factors


Reinsurance cost
Mix of business sold, if profit not uniform between rating classes and
products
Changes in legislation
Levels of crime
3rd party bad debt rates
Propensity to claim
Level of contingency margin appropriate for the company's attitude to risk
Comments on Q2(i): Answers to this question were frequently disappointing with
many candidates giving very generic answers more appropriate to an ST3 exam while
completely failing to exercise any higher order skills. The first paragraph of the
question set out a number of features of the company and their method of sale that
were intended to steer candidates into giving consideration to particular uncertainties
and issues that might arise from internet operations, although most candidates made
little or no mention of internet issues in their answers. Candidates failed to give due
consideration to the specific phrasing of the question which asked for the
uncertainties within premium rates, commenting on such components of office
premiums as taxes and MIB contributions which are readily available information
that no credible insurer would be uncertain of. Also, relatively few candidates
identified which were the key areas of uncertainty, and how the key areas of
uncertainty varied between products.
(ii)

The report must indicate the sources of the data used


and the extent to which the user of the data takes responsibility for data
accuracy or completeness.
The analyst may need to rely on or use the work of other people.
If there is a risk of confusion as to the division of responsibilities between
themselves and other persons or organisations, the respective responsibilities
must be made clear in the report.
The analyst must draw attention to any material limitations in the available
data.
In particular the companys case estimators.
Including the effect on the appropriateness of the data of changes in the way
the business analysed has been conducted.
The analyst must make reference to limitations in the data that have materially
added to the uncertainty surrounding the results of the work carried out.
The report must describe the criteria used for subdividing data into groups
Where the member makes adjustments to the data the nature, amount and
rationale for the adjustments must be clearly stated.
The concerns about the accuracy of the companys case estimators have
materially added to the uncertainty

Comments on Q2(ii): This question was generally well answered, although a number
of candidates were only vaguely aware of some of the detail of GN12, which should
be an essential part of any candidates preparation for SA3.

Page 13

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

(iii)

The main emphasis at this stage will be on how well the company has
achieved its planned loss ratio, i.e. estimated ultimate claims against earned
premium.
Other main factors
Expenses
Expenses incurred to date are likely to be very high due to development and
initial running costs of the new company and this should be adjusted for to
arrive at a long term assessment of profitability.
Example of expense breakdown (loss adjustment, fixed, variable etc)
Investment income
Reinsurance cost
Capital costs / profit margin / requirement etc
Data issues
The absence of good quality case estimates is an issue, primarily due to the
difficulty in assessing the tail development of longer tail claims such as third
party bodily injury.
We only have 24 months of development experience and earned exposure in
the initial 6 months will be very low, meaning the credibility of the paid
development at the tail will also be questionable.
The claims handling procedures would also likely take some time to stabilise.
Earning patterns and run-off for the first few months would be unstable and
this should be borne in mind in any analysis (any relevant attempt to correct
for this)
As the company has been successful, there may be more reasonable patterns of
development from later months. It may be possible to use these to fix the run
off pattern for the first accident year.
If PD and BI are considered together then the resulting paid claims
development will include virtually no third party bodily injury claims.
Third party property damage claims may also be under-represented unless it
adequately allowed for with a tail factor
If they are considered separately then there will be little to no data at all for the
bodily injury.
Therefore the use of paid triangulations with chain ladder techniques alone
may well under-project by significant amounts.
Other methods will need to be used to determine appropriate levels for IBNR
and IBNER on longer tailed claims.
General data splitting
Estimated ultimate claims will need to be calculated by subdividing the claims
into homogeneous groupings of similar development profile.
Splits are likely to be (in ascending order of tail) own damage, third party
damage and third party bodily injury.
If the data allow, it may be possible to separately analyse gross payments and
recoveries received from third parties.
Further sub-analysis may be taken between comprehensive and noncomprehensive business.

Page 14

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

Accident month data should be viewed to determine if there are any trends
emerging from the data.
PD / BI data splitting
One could develop own damage and third party damage separately, using any
additional company or industry-wide data available to derive a suitable
IBNR/IBNER.
For bodily injury, an addition to the above could be applied by either:
projecting bodily injury claim numbers to ultimate and then applying a
market average cost of bodily injury claims
recruiting or requesting a claims assessor to concentrate on coming up
with a sensible estimate for each of these claims (as there may only be a
few hundred)
For the latter, it will be particularly useful in identifying particularly large
potential outstanding liability claims as this could have a significant impact on
profitability.
Benchmarking
Compare the loss ratio against other insurers in the market.
Ensure that loss ratios are defined in a consistent manner e.g. treatment of
claims handling
Assess long term profitability by assuming similar loss ratios in the long term
along with expected long term expense ratios, e.g. by considering the expense
ratios of other insurers who are long established in the internet market,
together with investment rates achievable.
One possibility is to make use of market data to select a suitable tail factor.
E.g. FSA Returns for a similar company or a similar set of companies may
allow development factors of paid to ultimate to be derived.
Miscellaneous key factors
The mix of business (comp versus non-comp)
A review of key exposures in the account may highlight areas of weakness /
concern.
The potentially different terms and conditions between insurers.
Any changes to the rates and structure since the company started writing.
Mix by source:
the different claims handling procedures
the actual mix of claims
Exposure-based / BF type methods.
An exposure-based approach could be used.
This could involve taking projections from more developed accident months
divided by associated earned exposure and applying this risk premium to
more recent, less developed accident months.
Bornhuetter-Ferguson methods could be used on the claims splits by taking
initial assumptions about the loss ratios split between own damage, third party
damage and bodily injury and split between comp and non-comp business

Page 15

Subject SA3 (General Insurance Specialist Applications) September 2008 Examiners Report

from the underwriters with credibility factors derived from a combination of


internal development factors and external market or industry wide data.
Case estimates
It may be possible to look at a sample of case estimates to try to understand
better what the problems may be with them following the CEOs comment.
This may enable the case estimates to be adjusted to derive more reliable
incurred claims figures.
General comments
In practice, a combination of methods is likely to be used to determine the
expected profitability of the company.
Other stakeholders e.g. underwriter, internal actuaries etc may have relevant
information about the account.
Making the company aware of the additional uncertainty in the results as a
result of poor quality estimate capture could enforce better estimation in
future.
Comments on Q2(iii): As with question 1(v), candidates clearly struggled with a
situation where standard statistical projection techniques were not ideal. As standard
statistical techniques can easily be replicated with software packages, perhaps the
most critical area of the actuarial skill set is the capacity to identify weakness in
statistical methods and to exercise judgement in working around such difficulties, and
this skill appears to be lacking in the majority of the candidates. Answers that simply
listed a standard step by step account of a statistical projection method were common
for this question, with few candidates giving due consideration to specific features of
the question such as the fact that the company is growing (distorting the development
pattern) or to methods by which they could compensate for the poor case estimates. It
is not enough to just say use benchmarks/consultants or that someone else must know
the answer.

END OF EXAMINERS REPORT

Page 16

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
27 April 2009 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes before the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only, but notes
may be made. You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 A2009

Faculty of Actuaries
Institute of Actuaries

The Republic of Merino is a developed country with a range of small, medium and
large general insurance companies.
Regulations currently require general insurance companies in Merino to establish
claims reserves on a best estimate basis. Thus the claims reserves held by companies
represent the mean of the range of possible future outcomes, without margins.
Two years ago, one of the largest general insurance companies in Merino became
insolvent. In response to public concern about the state of the insurance industry, the
government has proposed new regulations. These would require general insurance
companies to establish claims reserves that include risk margins. It is proposed that
general insurance companies hold claims reserves equal to the 75th percentile of the
range of possible future outcomes.
Company A is a general insurance company writing private motor insurance. It has
been writing this business for 20 years and is the largest insurer in Merino.
Company B only started writing insurance one year ago. It writes only professional
indemnity insurance.
(i)

Give an example of a method that is likely to be suitable for each of the


companies A and B to estimate the 75th percentile risk margin, giving a reason
for your choice in each case.
[2]

The government is keen to understand how general insurance companies are likely to
view this proposal.
(ii)

Discuss the advantages to the general insurance industry of requiring insurers


to hold 75th percentile risk margins.
[9]

(iii)

Describe the concerns that the insurance industry may have with this proposal.
[18]

The government announces that it will go ahead with its proposals. However, general
insurance companies will be allowed to choose the percentile risk margins that they
wish to adopt. The adopted risk margin must provide a probability of sufficiency
greater than or equal to 60%. General insurance companies must publicly disclose the
percentile that they have adopted as their risk margin.
(iv)

Describe the issues that a general insurance company should consider when
deciding the probability of sufficiency to adopt.
[14]
[Total 43]

SA3 A20092

A small UK general insurance company writing UK and other European business


wishes to expand its employers liability book of business. It is considering building
a new rating model for this portfolio.
The underwriter responsible for this portfolio explains that:

he is the only underwriter for the employers liability book

he has recently hired a junior assistant

he sometimes uses a US workers compensation rating model that he was given


five years ago as a guide to setting rates and relativities between different trades

he has never possessed a rating model that he has trusted and does not believe that
a rating model can be better than his own judgement

his incurred loss ratios over the last ten years have been at a level that he believes
has made money in every year except one

(i)

Explain the risks to the company of not building a new rating model.

(ii)

Discuss the advantages to the underwriter of building a new rating model. [8]

(iii)

Outline the information that would be required to build the rating model, in
particular considering the following areas:

[12]

Underwriting
Reinsurance
Claims
Finance
[17]

(iv)

Outline the problems that might be encountered in building the rating model.
[4]

The company has had discussions with a UK based managing general agent (MGA)
that specialises in casualty insurance. The MGA offers underwriting and claims
handling services. It is looking for a general insurance company to provide capital by
taking a coinsurance follow line on its portfolio of employers liability business
through a binding arrangement.
(v)

Discuss the advantages and disadvantages to the company of taking a follow


line on this coinsurance arrangement.
[16]
[Total 57]

END OF PAPER

SA3 A20093

Faculty of Actuaries

Institute of Actuaries

Subject SA3 General Insurance


Specialist Applications
EXAMINERS REPORT
April 2009

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

R D Muckart
Chairman of the Board of Examiners
July 2009

Comments
Individual comments are shown after the solutions to each part question that follows.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

(i)

Example Methods
Company A
Any relevant example likely to be one of the standard stochastic methods.
Relevant reason, e.g. company A likely to have good data or suggestion is a
widely recognised method.
Company B
Any relevant example judgement, scenario analysis, or a statistical method
applied to benchmark data.
Relevant reason, e.g. company B has limited internal data or necessary to use
judgement.

Comments on Q1(i): This was generally well answered by most candidates,


although some candidates did not suggest a method as requested in the question.
(ii)

Advantages of Margins
Reduces the likelihood of holding inadequate claims reserves.
Relevant examples where reserves might deteriorate.
Should have a positive impact on credit ratings/share price support/attracts
investors/international trade. This may make the general insurance industry
appear more secure.
This is an advantage to insurers. People will only buy insurance if they believe
the insurer will be able to pay any claims. This is particularly relevant for
Merino, as one of the largest general insurance companies has recently
become insolvent
The regulations would be likely to promote a greater use of stochastic
modelling within companies. Although stochastic modelling would not
necessarily be appropriate for every portfolio the proposals would require
companies to give greater consideration to reserve uncertainty, thus increasing
management awareness
A greater understanding of reserve uncertainty will provide valuable
information to managers (or other advantage of stochastic modelling)
Consideration of reserve uncertainty may align well with other analyses
undertaken by insurers, e.g. DFA, estimating capital requirements.
Reserving above the best estimate is increasingly regarded as best practice.
The insurance industry may consider it advantageous for local regulation to be
in line with international standards. This will particularly be the case if local
market participants also operate internationally.
All companies are required to carry out these calculations - neutral to
competition.
Defer profits and thus defer payment of tax.
Encourages more appropriate use of reinsurance.
Makes benchmarking easier/consistency in the market.

Page 2

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

Expected lower contributions to solvency protection fund.


Better match to risk.
Relevant example of a country where general insurance companies are
required to hold risk margins, or such a change is proposed, e.g. Solvency II,
Australia
Comments on Q1(ii): A number of candidates failed to understand
what a percentile is.
(iii)

Some companies may not have sufficient data to apply a stochastic method.
Some companies may not have sufficient internal expertise to calculate the
required estimates. External expertise may be expensive to engage or
unavailable.
Additional costs to companies in doing the calculations
These additional costs would need to feed into premium rates with potential
impact on level of sales.
Risk margins are only one part of the regulations that ensure the solvency of
insurers (e.g. operational risk, credit risk)
Arguably it is the total amount of capital held by insurers that is of greatest
importance, not whether the capital is held as free reserves or reserve margins
Significant reserve increases can be difficult to explain to shareholders and/or
other stakeholders
However, an annual release of margins from prior periods would become
expected.
Some exposures are particularly difficult to estimate margins for, an example
being latent claims.
There are many approaches to calculating a 75th percentile.
Different methods can produce very different results.
Different methods are appropriate for different lines of business.
There is not yet general agreement among actuaries as to which is the best
approach for a given class of business.
Naive application of methods can provide misleading results.
From a regulatory perspective, it may not be possible to determine whether
companies estimates really do represent the 75th percentile.
Increased regulatory costs will feed back to general insurance companies.

.
Although it is also impossible to say whether a best estimate really is the mean
of the distribution, uncertainty is greater for more extreme percentiles.
It can be difficult to compare results between insurers, e.g. if company A has a
lower margin than company B, it will be unclear whether this reflects the
riskiness of the liabilities, method selection or a difference in judgement? (or
other relevant example)
Such uncertainty may undermine confidence in the insurance industry.

Page 3

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

The assessment of diversification between classes of business is one of the


more difficult areas.
It is difficult to estimate correlations between lines of business based on
observed experience.
It is often necessary to use general reasoning to estimate correlations.
Regulators may not have the necessary level of expertise and resources to
review the risk margins estimated by general insurance companies.
The calculations used to produce the risk margins may be difficult to explain
to management.
Other stakeholders may also be confused, e.g. policyholders, investment
analysts.
There may be a false confidence held by stakeholders.
A 75th percentile means there is still a 1 in 4 chance that reserves will be
deficient.
Capital requirements should be revised to be consistent with the new reserving
requirements, e.g. if companies are required to hold capital to the 99.5th
percentile, then the regulations should reflect the fact the reserves are already
at the 75th percentile.
If current capital regulations are set at a very low level, some companies may
not have enough capital to reserve to the 75th percentile.
Companies may appear less solvent, since they are required to hold higher
reserves.
If reserves are overstated results may look worse than expected, and tie up
assets that could be better used (increased costs of locking in assets, conflicts
with growth/acquisition opportunities)
Companies may be unhappy if the profits cannot be distributed until later
(because profits are used to fund reserve margins)
General insurance companies would be concerned that the new regulations
result in a less favourable tax treatment.
Tax and regulatory issues for insurance companies who already hold implicit
reserve margins.
Companies may reassess their mix of business in light of these regulations
75th percentile may not be relevant for all classes of business
Comments on Q1(iii): This was poorly answered by many candidates who failed to
generate a broad range of answers for the number of marks available. The marks
awarded for this section were generally low relative to the rest of the paper.
(iv)

Page 4

General insurance companies would want to know the level competitors and
other countries are adopting.
Consider any views that have been expressed in the market, e.g. by
regulator/government, rating agency, competitors, industry or professional
bodies.

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

If most companies decide to hold 80%, a company holding 70% might appear
unattractive to customers or other stakeholders (or other relevant example).
However, stakeholders may be much less interested in reserving probability of
sufficiency than other factors (e.g. policyholders may buy based on price).
Companies would certainly want to hold at least the minimum specified by
law.
Consider uncertainty in best estimate and distribution of uncertainty to decide
on level of percentile, e.g. if there is a greater than usual level of uncertainty in
the reserves, companies may wish to hold a higher margin in order to increase
the likelihood that reserves will be adequate.
However, the financial cost of holding reserve margins at a higher level of
sufficiency will be greater for the most uncertain liabilities.
The cost of holding risk margins to a particular probability of sufficiency will
also depend on the diversification within the portfolio.
Insurers should not adopt extreme risk margins, e.g. 99.9%.
Such percentiles cannot reliably be calculated for many classes of business.
It would be damaging the general insurance companys reputation if reserves
deteriorated beyond an extreme margin.
As the selected percentile moves above the mean, the cost of providing for the
risk margin increases significantly, e.g. it generally takes more capital to move
from the 90th to the 95th, compared to moving from the 70th to the 75th.
For some highly skewed distributions, the mean might be above the 60th
percentile, e.g. reserves for asbestos claims (or other example of a highly
skewed distribution).
It would seem imprudent to hold less than the best estimate reserves.
The company should consider the link between reserve margins and total
capital requirements.
Consider the total amount of assets available and ability to raise capital.
The company may be relatively indifferent about the precise split of assets
between capital and reserve margins.
The company may choose to minimise risk margins to increase the apparent
level of capital.
This may be attractive if investment restrictions apply to assets backing the
free reserves and the reserve margins (or other relevant example).
Alternatively there may be some tax or other advantage in adopting a higher
reserve margin.
The company should consider its tolerance for reserve deterioration.
It is possible to imagine companies with a high degree of tolerance (e.g.
certain privately held companies) (or other example)
The company may have derived statements of its risk appetite as part of its
broader risk management framework.
Such statements may include a policy on tolerance for reserve movements,
which would drive the decision on risk margins.

Page 5

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

Consider whether reserves are to be held on a discounted or undiscounted


basis. Undiscounted reserves will contain margins, so will correspond to a
higher percentile of adequacy.
Position in insurance cycle, particularly if percentage can vary year by year
Opportunity cost of capital/other business opportunities.
Benchmarks used by smaller companies may not be available for different
percentiles.
Consider mix of business (classes of business, territories).
Cost and availability of reinsurance.
Time horizon over which adequacy is assessed: the shorter the timescale, the
less capital required to achieve a desired probability of sufficiency
Comments on Q1(iv): Most candidates were able to identify that the market
influences (e.g. competitors, rating agencies, position in insurance cycle) would have
a significant impact on the probability of sufficiency to choose. Better candidates
identified risk appetite, the increased cost of holding more extreme scenarios, reserve
discounting, existing available capital and insurer reputation as important
considerations. A large number of candidates suggested that a higher percentile
should be used for classes with greater variability demonstrating the lack of
understanding of percentiles.

(i)

Risks of not building rating model


The premium charged by the underwriter may be insufficient due to features
that would be identified by the rating model but not by the underwriter.
Lack of rating factors in existing rating may lead to risk of anti-selection, in
particular by other insurers who are collecting and analysing up-to-date data
and have better rating structures to the extent that the underwriters own
knowledge is insufficient for ideal rating.
Significant risk of concentration of knowledge in one underwriter.
If the underwriter was suddenly to leave the company, risk of not finding
replacement underwriter of similar expertise.
The junior assistant is unlikely to have the necessary underwriting experience
in the short term.
Very difficult for the underwriter to accurately adjust prices for changing
assumptions without a rating model, such as:

Page 6

expenses
commission rates
investment conditions
capital required
e.g. as nature/size of company changes
changes in regulatory environment
target profitability

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

change in claims environment


latent claims
inflation (claims, expenses)
changes in exposure
cost and structure of reinsurance

Subjective allowance for cross subsidies.


Difficulty in measuring profitability performance of this class of business.
Accurate rate change information unlikely to be available.
Lack of explicit rate monitoring processes and data capture may not meet
regulatory requirements.
Potentially higher solvency requirements from regulators due to Solvency II
Small company => claims volatility and rating accuracy are significant risks
Difficulty in setting appropriate reserves for account as difficult to gauge rate
adequacy against accurate technical rate.
Risk of writing loss making business if underwriter continues to use US rating
model as US claims experience may not be relevant to UK and other European
countries
Different coverage/terms and conditions in the US compared to UK and other
European countries and different legal/regulatory environment in the US
compared to UK and other European countries.
Pricing assumptions may not have been updated for 5 years.
In particular claims inflation may have been considerably higher during the
period.
Greater danger of being influenced by market cycle as no relevant technical
rate available to guide the underwriter.
The underwriter only occasionally uses the model, possibility selecting
occasions when US relativities are too low and possibly inconsistent with the
junior underwriters judgement.
Lack of clear audit trail for internal peer review process.
Rating models may not have been properly checked in the first place.
Risk of losing business to peers with accessible online systems.
Reinsurers may not trust rate adequacy resulting in higher reinsurance costs.
Comments on Q2(i): Too many candidates focused on what would make up a rating
model rather than considering the wider commercial and business implications of
having an effective rating process.
(ii)

Advantages to underwriter
More time to concentrate on underwriting more complex cases.
Simplified rating approach on standard risks.
Ability to quote quicker to assist expansion.
Potential to write higher volumes through efficiency savings or make more
careful selections for risks taken on.

Page 7

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

Potential to identify underpriced renewals to correct rating or decline at


renewal, or overpriced renewals to compete more aggressively at renewal.
Easier/quicker to demonstrate profitable performance of account.
Capture of granular exposure information would make it easier to adjust profit
assumptions e.g. in the light of a legal change.
Ultimate loss ratios described by underwriter are not necessarily an accurate
indicator, only backward looking and more easily attributable to good
fortune
Better understanding of account from rating exercise and perhaps formal
collection of additional rating information within rating tool and on IT systems
Less time explaining account to management.
Junior assistant requires lower training burden as knowledge within tool
Fewer risks with resourcing.
Rating model could provide a more robust benchmark price compared
toposition in market cycle including year on year price changes relative to
benchmark.
Consistency in approach to pricing risks (improves reputation with brokers).
May be able to negotiate better reinsurance terms using the better data.
Internal peer review easier.
Meets regulatory underwriting standards better.
Highlights areas where the underwriters judgement is out of line with market
prices which may improve underwriting quality.
Can establish web / auto quoting for small homogeneous risks.
Comments on Q2(ii): In parts (i) and (ii), better candidates appreciated that the
advantages to the underwriter of building a model were not simply a reverse of the
risks to the company of not building a model. Better candidates recognised the level of
subjectivity and lack of a clear audit trail of the underwriter's decision making
process as well as the potential risks involved with the use of an inappropriate and
out of date US rating model.
(iii)

Information requested
Underwriting:
Book rates and existing method of pricing adopted.
Trades/occupations covered/declined.
Existing rating factors considered/available.
Level of subjectivity needed.
Scale of size discounts.
Approach to pricing different limits of indemnity if applicable, e.g. increased
limit factor (ILF) curves used.
Distribution channel of business written.
Different policy wordings/terms and conditions used.
Exclusions and excesses for other European business.
Information provided by client when risk quoted, in particular the claims
history provided to assess what (if any) experience rating possible.
Frequency of claims.

Page 8

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

Exposure information for account.


Any limits to individual policyholder exposure (by manual/clerical wageroll).
Typical limits of indemnity / size of risks in portfolio.
Countries risks written in / currencies risks written in.
Views on market cycle (current prices soft/hard?)
Any current competitor rates available.
Time available per risk for pricing/time constraints.
Conversion rate and viability of quoting before/afterwards.
Expected future claims inflation to apply (will want to build this into the
model and will be necessary if integrating rate monitoring into the process)
Existing risk data that could be used to (at least partially) populate a rating
database for renewal risks to save time.
Potential latent issues not yet showing in claims experience.
Reinsurance:
Reinsurance structure.
Current costs of reinsurance, net of recoveries.
Expected changes in the above.
Any feedback from reinsurers on current portfolio against market average.
Claims
Fgu claims history for line of business.
Ideally 10 years of data.
Transaction level detail if possible.
Split between payment type (legal costs, compensation, loss of earnings, etc.)
and/or paid & incurred triangles.
Split by territory, industry, peril, rating cell.
Gross/net of reinsurance.
To understand development of claims for experience rating and payment
pattern to estimate investment income.
Details of historical changes in processes, controls, systems, external factors
(e.g. legislation).
Nil claims.
Catastrophes/unusually light or heavy experience.
Potential latent issues not yet showing in claims experience.
Any benchmark/market data that may assist in building model.
Cost of handling claims.
Information on large claims in market, not just own account.
Views on claims inflation split between legal fees, compensation, cost of care,
loss of income.
Finance:
Expenses including relevant allowance for inflation including any breakdown
of expenses available (direct/indirect).
Commission levels (e.g. any fixed fees, differential new business/renewal
commission rates).
Interest rate assumed / investment income information.
Capital requirements and return on capital required for this business.
Taxation (corporation tax, IPT).
Future agreed budgets/plans/forecasts for employers liability account.

Page 9

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

Comments on Q2(iii): This was answered well by most candidates. Better


candidates identified that most information would be required from the Underwriting
and Claims areas and therefore gave greater focus to these areas in their answers.
However, there was a seemingly universal lack of awareness of the statutory nature of
EL cover in the UK meaning that deductibles are not applicable and limits are
generally the same.
(iv)

Significant problems
Biggest problem is lack of data especially as small company.
Particularly in respect of disease claims.
Meaning little information in any rating cell and employers liability claims
frequency low so huge reliance on underwriter opinion.
May be few rating factors following rating model review that are hard to
define (underwriters view) .
Potential errors in the data.
Difficulty in building in different rules for each European country.
Future court awards unpredictable.
Costs of building rating models may be prohibitive (and lack of expertise to
build).
Added uncertainty around projected burning costs (appropriate allowances for
IBNR/IBNER).
Potential problem with underwriter buy in.
Particular problems if pricing higher limits of indemnity as incidence of large
claims even lower.
Latent claims are difficult to allow for objectively.
Potential IT problems if aiming to embed into existing systems.

Comments on Q2(iv): This was reasonably well answered, although few candidates
identified the lack of disease claims history and issues around different rules for
European business
(v)

Advantages:
Additional source of revenue with very little resource/cost required to set up
arrangement internally .
Greater expertise of MGA could provide potential to underwrite other types of
business not currently written by insurance company..
Relevant examples.
Based in the UK so no currency risks.
Marketing undertaken by MGA rather than coinsurers.
Diversify portfolio exposure into other types of business/countries thus
reducing accumulation risk.
Could provide expertise in handling claims for non-standard risks or claims
Provides policy and claims administration through its own IT systems so little
ongoing cost for insurer.
Lead insurer likely to be responsible for MGA audits so lower ongoing
administration costs.

Page 10

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

If MGA fails, the insurance company is only liable for its coinsurance share of
claims (i.e. lower risk than taking on whole risk and reinsuring).
If arrangement successful, could lead to additional capacity offered for other
classes of business (e.g. public and product liability) or cross selling.
Disadvantages:
Less control over underwriting decisions and risk that policy wording may be
undesirable.
Could lead to undesirable risks being underwritten.
Potential concentration of risk and/or aggregation of risk with own portfolio.
Certain risks may fall outside existing reinsurance limits or may expose
insurance company to risks not authorised to write by regulator.
Capacity limits would have to be imposed due to limited capital of insurance
company which may in itself restrict the type of business written
MGA may be competing for same target market as insurance company which
may result in company effectively competing against itself for same business
or could cause broker relationship issues if through same distribution channel.
MGA likely to request higher commission rates or fees compared to company
selling through brokers.
May not provide same return on capital compared to existing portfolio.
Potential loss of or reduced level of management information if MGAs IT
systems are substandard .
Less experienced MGA claims handling staff may result in higher ultimate
claims costs.
Potential lack of experience in handling particularly complex claim cases
involving serious injuries.
Claims handling fees may be higher than equivalent cost of insurance
companys own claims handlers.
Different claims handling philosophy to insurance company may imply
difficulty in establishing accurate claims reserves.
Especially if MGA is relatively new with little historical claims development.
Likely to be contractually tied into arrangement for an agreed period so
difficult to walk away immediately from arrangement if results are poor.
Increased delay between policyholder paying premium and receipt of
premiums from MGA due to credit agreements between broker, MGA and
insurance company meaning potential loss of investment income.
Revenue stream and service elements dependent on MGA staying in business.
Potential credit risk with binding arrangements.
Time delay on claims bordereaux will mean that internal development patterns
will not work well for the binding risks; may have to project separately.
Harder to exit the market rapidly if needed.
Reinsurer may not like the third party involvement.
May cause resentment with internal underwriter over loss of control or may
impose additional supervisory burden on in house underwriter to keep track of
the business.
Moral hazard dependent upon MGA share of risk (e.g. underwriting risk)

Page 11

Subject SA3 (General Insurance Specialist Applications) April 2009 Examiners Report

May be difficult to assign ownership; if the in house underwriter is responsible


for the business then he may object to the risk to his bonus.
Comments on Q2(iv): This was poorly answered by many candidates. A significant
number of candidates did not generate a broad range of advantages and
disadvantages. Many candidates were unfamiliar with the concepts of binding
arrangements and managing general agents which are very topical subjects.

END OF EXAMINERS REPORT

Page 12

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
6 October 2009 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes before the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only, but notes
may be made. You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 S2009

Faculty of Actuaries
Institute of Actuaries

Lloyds managing agency A is considering taking over a smaller Lloyds managing


agency B whose single syndicate has recently had some extremely poor experience.
This has left it without the capital strength necessary to compete effectively and
attract business, although it is currently still solvent.
(i)

Set out the key issues and risks that need to be considered in this takeover,
indicating any specific aspects that could materially impact these risks. [22]

One of the units in Bs syndicate that is causing particular concern is the Special
Situations Unit. Agency B uses this unit to group a variety of individual niche
products that dont fall into the target markets of their main underwriting units,
including a number of contingency risks.
(ii)

Outline the general advantages and disadvantages of such a business unit. [5]

(iii)

(a)

Define contingency business.

(b)

Give two examples of the types of risk that could fall into this category
(excluding prize indemnity insurance).
[3]

One major source of premium written to this unit is a portfolio of prize indemnity
insurance risks. The portfolio is written under a binding arrangement with a third
party coverholder that manages the underwriting and claims handling with the
syndicate providing the capital backing.
.

The policies written cover fixed prize lotteries each with their own weekly draw and
various levels of ticket sales, with tickets often sold through affiliate groupings or
business connections. All the draws are on the same basis, with five unique numbers
per 1 ticket chosen from 50 possible numbers (the order of the numbers is not
important) and three fixed prize levels as follows:

All five numbers: 100,000


Any four out of five: 1,000
Any three out of five: 100

The insurance covers the top two prizes at a premium of 25p per ticket, 20% of which
is paid to the binding authority underwriter to cover commission and expenses.
(iv)

Deduce the planned loss ratio that Agency A would estimate for this business.
[5]

(v)

Discuss the likely advantages and disadvantages of continuing to write this


type of risk.
[8]

Analysing the business in more detail has shown that one of the major sources of the
recent poor experience has been its professional indemnity business. This originally
focused on high value international clients, and has recently expanded rapidly to
develop a strong regional presence underwriting small to medium sized businesses,
particularly in the solicitors indemnity market.

SA3 S20092

As the board of agency A has little experience of this business it would benefit from
an introduction to the type of experience that might be expected and the way that this
business is written.
(vi)

List the key features of professional indemnity insurance.

[6]

The board would also like to understand the risks relating to this book of business,
including any key historic and ongoing features of this particular account and any
current market conditions that might affect these risks.
(vii)

Outline the key exposures to Agency A from this book of business and the
main investigations into this account that they should consider in order to
understand the risks better.
[14]

The board of Agency A has decided to go ahead with the takeover, but is unsure of
the ways in which it could be done and the key regulatory requirements that must be
observed in setting up a combined Managing Agency.
(viii) Outline the options available for a takeover, including the regulatory issues
that should be considered.
[6]
[Total 69]

SA3 S20093

PLEASE TURN OVER

An investment banker is considering two UK general insurance companies as


potential acquisition targets for a client. The client is a multinational group of
insurance companies. The following information on the target companies has been
provided. The project is at an early stage so that at present only limited information is
available on the targets.
Amounts in millions
MCR
ECR
ICA
ICG

Company A

Company B

10
15
20
20

10
17
16
18

Note: Both target companies have the same premium income.


(i)

State the meaning of each of the abbreviations MCR, ECR, ICA and ICG and
give a brief description. Details of how these amounts are calculated are not
required.
[5]

The bank is meeting with its client tomorrow, and capital requirements will be
discussed at the meeting. The banker would like to understand the likely reasons for
differences in the numbers, and in particular the following features of the data:

Company B has a higher ECR, although both companies have the same MCR.

Company A has a higher ICA than company B.

Company B has an ICG greater than its ICA, but the ICA and ICG are the same
for company A.

(ii)

Write notes describing the possible key reasons for the differences noted in the
list above.
[18]

The banker believes that the client may want to consider alternative levels of
capitalisation, other than the capital requirements given in the table above, when
considering which company to acquire.
(iii)

Discuss examples of capital requirements, other than those listed above, which
should interest the client.
[8]
[Total 31]

END OF PAPER

SA3 S20094

Faculty of Actuaries

Institute of Actuaries

Subject SA3 General Insurance


Specialist Applications
September 2009 examinations
EXAMINERS REPORT

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.
R D Muckart
Chairman of the Board of Examiners
December 2009

Comments for individual question are given with the solutions below.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

1
(i) Key issues to investigate
Reserve adequacy company is currently solvent under their reserving
methodology, but capital backing is obviously weak.
If over-reserved there is a potential benefit to the company but given the
company is in a weak capital position that is leaving it vulnerable to takeover
it is unlikely to be materially over-reserved as they would most likely have
moved to a best estimate basis to remain solvent enough to continue trading /
attract a more favourable takeover settlement, leaving it far more likely that
any risk is to the downside due to the company being under-reserved
One aspect that materially affects this level of risk is the length of tail and
level of uncertainty of the business lines being written with shorter tail classes
such as property posing significantly less reserving risk than longer tailed
classes.
along with the length of time the syndicate has been in operation if there are
any longer tailed lines as that increases the number of years on which reserves
could potentially deteriorate
including any inherited lines and discontinued business units
Reasons for the poor loss experience
e.g. was it due to a series of unfortunate major claim events e.g.
catastrophes, higher than anticipated latent claim amounts
in which case one would need to consider whether or not the business
is adequately diversified / protected against such volatility
and whether any lack of diversification is simply due to the small size
of the syndicate and once integrated into your larger syndicate such
volatility wouldnt be so material
and is the likelihood of any further such volatility within your agencys
risk tolerance
and if due to major claim events were these market claim events that
materially affected their peers or were they isolated events of bad
underwriting judgement / luck for this syndicate
or was it due to poor underwriting controls allowing the unit to write
excessively large lines on individual risks
or a failure to properly manage aggregate limits in particular areas
or was it simply due to inadequate rating levels
or was it poor claims controls
or an increase in fraud
and if so can these be addressed with improved internal controls /
rating models
and are the market conditions going forward likely to offer a more
favourable rating environment

Page 2

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Is any poor experience confined to only some of the underwriting units


and if so is there any scope for discontinuing these units and going
forward with the profitable underwriting teams only
although the impact on morale of this move must be considered
Synergies with the existing business
e.g. with the IT systems (or other appropriate example)
most importantly is the business written likely to complement the
business your agency already writes
and add to diversity
and add genuine growth opportunities
and/or to only retain those underwriting teams whose market segments
are not already covered
Do the underwriting teams that your business might seek to retain have
valuable expertise / intellectual capital
and how much of their client / premium base is likely to be loyal and
easily transferable to your new business
and the impact on broker relationships
Are there potential expense savings / economies of scale from
integrating the two businesses
with particular consideration as to whether your company already has
the necessary expertise to support any new business lines / claim types
etc.
Difficulties in letting go / retaining their current management team
Capital savings from integrating the two businesses
in particular the level of diversification credit from adding a
complementary portfolio
and the reduced volatility from simply having a larger overall business
Price at which the business could be obtained
Opportunity cost of alternatives
Any other potential buyers and the impact this will have on the price
e.g. risk of paying too much
Costs associated with the take-over e.g. administration, advisory
and the receptiveness of their staff base to the takeover opportunity
along with moral issues for your managing agency in such a move
Free capital for such a takeover opportunity
or the scope to raise additional debt to finance it / price of such debt
impact on As credit rating
Page 3

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Attitude of other stakeholders to the takeover including: your


shareholders / the general market, names, your debt providers, your
reinsurers, Lloyds / FSA, rating agencies
Security / value of assets held
e.g. are they readily tradable liquid assets whose value is clear.
and are they adequately matched to the liabilities
both by currency and term
Security of any outwards reinsurance contracts on the inherited
business
adequacy of bad debt reserves
which could be a material issue if the poor loss experience was due to a
market wide issue that would impact reinsurers
or if the poor experience relates to older years for which the reinsurers
may be less likely to still be trading
Debt held by the syndicate and whether there is any opportunity to
refinance and improve profitability there
especially if your managing agency has a superior credit rating
Any tax benefits to be gained
Any other relevant suggestions
There was a significant variation in the quality of responses to this question. Candidates
generally fared better on the more generic issues to consider in a takeover (synergies,
diversification, reactions of stakeholders etc.) although not in general covering sufficient
issues. However, a disappointing number gave little or no consideration to the most
pressing issues in the situation described in the question: what caused the poor
experience, how much potential is there for the existing experience to deteriorate further
and how much potential would there be to turn the experience around and have a
profitable business going forward. Candidates should clearly recognise the far greater
risks in taking over long-tailed business with significant latent potential suffering from
significant under-reserving issues than in taking over short-tailed business suffering as a
result of a recent catastrophe event with inadequate reinsurance or diversification. Few
candidates gave more than a cursory consideration to these most critical aspects,
suggesting either poor exam technique or a lack of awareness of the relative significance
of different types of risk. A surprising number of candidates also failed to address the
single most important point, i.e. the price at which Agency B could be purchased.
(ii) Advantages / disadvantages (+/)
+ Niche business can be significantly more profitable than higher
volume business lines
+ due to the lack of competition in the market

Page 4

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

+ Having a separate specialist team to pass unusual but profitable risks


to allows the main underwriting teams to focus on their areas of
expertise
+ without turning down good business
+ Capital requirements are often low when written as part of a larger
business as niche business is frequently uncorrelated with other lines
+ / Requires very specialist underwriting
although this could be an advantage if the specialist skill is available as
it reduces the scope for new entrants into the market
+ Reinsurance may be easier to obtain given specialist knowledge of
the team to find the right reinsurer and package
- alternatively reinsurance may be harder to obtain for niche products
despite internal specialist skill
Often difficult to analyse as all cases are so individual that few data
are available
so true underlying profitability can be difficult to ascertain
along with difficulties in business monitoring including rating
and reserving projections
Contingency business in particular can be notoriously difficult to
analyse and the loss record in the market is often poor as a result of
this
Difficult to set standard underwriting controls as each case is so
different
meaning that the unit may well require a high level of management
supervision / expense compared to the premium volumes written
This question was generally interpreted as to why you may want a Special Situations Unit
separate from the main business (as opposed to why have one at all). Both
interpretations are valid and many of the relevant points are valid under either
interpretation. Most candidates recognised the advantage of being able to give unusual
risks the focus they need although many failed to address the key reason for writing a
niche product being the potential for higher profits because of a lack of competition and
many failed to point out the difficulties in pricing and reserving because of heterogeneity
and paucity of data.
(iii) Contingency insurance provides financial compensation for specific insured
events that cause the insured delay, expense or an inability to continue current
professional activities
Examples include:
Event cancellation due to weather / venue damage etc.

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Film risks including death of cast members, damage to props etc.


Non appearance e.g. musician refusing to turn up / falling ill etc.
Legal delays costs incurred if a case overruns due to death or damage of a
participant
Product recall sometimes
Specialist life / disablement cover J-Los bottom, that wine tasters nose
etc.
Although this risk type is not considered in the core reading and has not featured in
past exam questions, at SA3 level candidates are expected to display general market
knowledge including awareness of less mainstream product lines such as contingency
insurance which is on offer from a number of different product providers and features
regularly in the media when coverage is offered for highly specific and unusual risks,
often with a celebrity focus. Because of the lack of a specific, core reading definition
some level of credit was given for any reasonable attempt at a definition, although
many candidates guesses were either contradictory, incomprehensible or
inappropriate as a form of insurance.
(iv) Expected claims cost from the top prize = chances of getting all 5 correct out
of 50:

5 4 3 2 1
50 49 48 47 46
= 1 in 2,118,760
expected cost of 4.72p per ticket to cover this prize
Expected claims cost from the second prize:
5 possible ways of picking 4 of the 5 correct numbers
Each of those can be combined with any of the 45 remaining possible numbers
Therefore 5 45 = 225 possible ways of picking 4 out of 5 from any
combination
2,118,760 possible overall combinations
225 / 2118760 = 10.62p per ticket
Net premium after binding authority underwriters commission & expenses =
20p per ticket
You work for a Lloyds syndicate so all accounting would be done on a net of
commission and external expenses basis (award marks for alternative
appropriate assumption stated)
Expected loss ratio = (4.72 + 10.62)/20 = 76.7%

Page 6

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Candidates at this level should possess the appropriate understanding of probability to


answer this question, and although many candidates could produce reasonable answers
for the probability of all five numbers remarkably few were capable of calculating the
probability of four out of five numbers. Some papers also included comments that the
calculation would in some way depend on the number of tickets sold. Many candidates
failed to carry out even a basic sense check on their numbers. A number of candidates
failed to pass comment even after producing loss ratios of less than 1%, and while the
marking schedules do not explicitly penalise such an omission it does form part of the
subjective assessments made of candidates close to the borderline.
(v) Advantages / disadvantages
+ Some profit to be made
especially since internal expenses are likely to be low as underwriting
is done externally
also potential for increased profits if some winning tickets are lost / not
claimed
+ Purely mathematical calculation, no subjectivity at all so profitability
can be assessed with confidence
which also makes it easy to set clear parameters for the binding
authority to write to a target profitability for any other business
+ Easy to obtain reinsurance given that the product is easy to price
+ Extremely short tailed business (claim event known at the end of the
draw) leading to reduced uncertainty
+ Should be entirely random
therefore utterly uncorrelated with everything else
meaning capital requirements for this business would be very low if it
is a small part of the overall premium base, because of diversification
credit
and also meaning all separate risks and individual draws are utterly
independent so there is no risk of aggregation
+ Premium volumes should be reasonably stable as it is an aggregation
of thousands of individual ticket purchases
+ As the question specifies this comprises a major source of
premium then at 20p premium per ticket it is likely that volumes are
high suggesting any volatility would be pretty low, at least over a 1
year timescale
+ Rating levels may be fairly stable as it is a fixed priced product
Risk of moral hazard depending on the independence of the draw
would want to check how the draw is handled and/or insist on an
independent witness of your own choosing to supervise each draw
although this may add a prohibitive level of expense

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Would also request a database of all tickets to be sent through before


each draw
Risk of fraud e.g. printing of tickets
Loss ratio is comparatively high, may not meet internal requirements
although given the low capital needed on a ROC basis this is unlikely
Potentially high annual volatility particularly if ticket volumes lower
than expected
or if popular numbers are selected e.g. birthday dates (award point for
mentioning accumulation of risk due to numbers selected)
Potentially high weekly / monthly volatility even with higher sales
volumes
e.g. even 5m of annual premium suggests 25m tickets sold a year or
500k a week, so the top prize particularly can be volatile in the short
term with only about 1 claim expected every 4 weeks
and possible cashflow issues resulting from that
although the managing agency is large enough that it is unlikely to be
an issue
unless there is material risk in different currencies to the rest of the
business
Very few candidates understood the issue being tested here of the implications of a
completely random (barring fraud) and mathematically calculable product and simply
offered a generic answer on binding arrangements. Many candidates commented that
there was a risk of ceding control of underwriting under binder arrangements, without
considering that underwriting simply is not a factor for a policy where the risk can be
calculated exactly. The majority of candidates failed to recognise that a purely random
product like this would by its very nature offer complete diversification with every other
line of business, making at best generic comments that it might offer diversification
depending on other lines of business written. A number of candidates also wrongly
commented that this would not constitute an appropriate product as the insured had no
financial interest in the outcome, in spite of the insurance contract clearly being with the
lottery provider who is responsible for paying the claims and who clearly has a
significant financial interest. Also points about moral hazard and what action could be
taken were generally not made.
(vi) List some key features of professional indemnity insurance (key points)
Covers professionals against the consequences of flawed professional advice
Specialty business generally written through brokers
Sometimes compulsory cover for certain professions
There are often cover restrictions and exclusions
Key exposure measure is usually fees

Page 8

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Generally low frequency


Claims can be high when they do occur
and subject to court award inflation and earnings
Accumulation risk following on from precedent cases
Development is often slow, medium to long term business
Affected by changes in legislation/regulation
Any other reasonable comments e.g. changes in legislation affecting advice or
actions retrospectively, latent claims in medical malpractice, moral hazard
Additional points
Solicitors business tends to renew on October
Type of cover E&O
Type of cover breach of duty
Type of cover Civil Liability
Example of a profession
2nd example of a profession
Sometimes rated on turnover
Sometimes rated on number of partners
Generally rated in more detail according to the split of fees by different types
of activity
Example of type of activity (e.g. conveyancing)
Sometimes rated on experience basis for larger clients
Development longer for some professions where the consequences of poor
advice take longer to be realised (e.g. architects)
Legal costs are usually covered
Written to defined policy limits no unlimited liability
Can be written as direct or reinsurance business
e.g. reinsurance of a professional body
Generally claims made (although can be losses occurring)
Claims made will generally have retroactive date
Comments on Q1(vi): Generally adequate answers on a largely bookwork question,
although a majority of candidates suggested turnover or even wage-roll as the most
common exposure measure with few mentioning fees.
(vii) Potential concerns about professional indemnity business:
Long-tail class so claims may take a while to emerge
Page 9

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Even those claims that do emerge can cover fairly unique situations
that are often tested in court requiring highly subjective valuations based on
good understanding of the risks and situations in question
which makes the quality of the claims assessors used by the company critical
in understanding the level of risk in the existing business
Given the high value international clients there could be large claims and
precedent cases
and deep pockets syndrome
and more general reserving methodology
and development profiles for the business
Particular attention to any changes in claim assessment methods / staff that
might change the development profiles of the business
Investigate reinsurance programme
Given the slow development, rate changes are of critical importance for this
business for early years of development as emerging experience will be
insufficient / lack of data on solicitors
New business strain as it's a growing book
rapid expansion may be a result of under-pricing leading to solvency issues
Coverage changes can also have significant impact on rate movements for this
type of business
e.g. costs in addition coverage
Investigate potentialrecent significant market events and the impact (if any) on
this book
Investigate rate monitoring quality and methodology and portfolio analysis
including allowances for any coverage changes
Coverage offered can have a significant impact on the development profile of
the business
claims made versus losses occurring in particular
Investigate the type of coverage offered and any trends or changes in coverage
historically given change in mix of business
Potential accumulation risk given strong regional presence
Investigate level of expenses in running the business
Currently in a very soft market for UK PI business
particularly for solicitors business
with a number of insurers reducing or withdrawing their product range
with low profits or even losses expected to emerge from recent years across
the market as a result of the low rates
Business is also highly influenced by macroeconomic events

Page 10

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

which are currently extremely unfavourable and likely to lead to a significant


influx of claims
such as valuation claims for surveyors
or claims on conveyancers
following the property market collapse
or mis-selling claims for IFAs following losses on the stock market / savings
etc.
or from mortgage brokers for inappropriate sale of mortgage products /
checking of documentation etc.
This risk is increased given that there are more SME that will be more
impacted by a recession
Critical to understand the make up of the account
and any overexposures to certain areas, country or currency
e.g. is the solicitors book heavily weighted towards high risk areas such as
conveyancing
It was mainly generic bookwork answers that were produced for this question, including
that it is highly influenced by macroeconomic events with very few candidates passing
any comment whatsoever on what these recent macroeconomic events have been. With
the first major recession since most candidates started their professional lives having
dominated media coverage for more than two years we would have hoped that candidates
had spent many months considering recessionary issues in their daily working lives and
indeed that candidates would be expecting such a topical question. Many candidates
failed to address the specifics in the question and simply gave a generalised list of
potential risks for a book of business. The word outline in the question is being used to
suggest a brief summary but many candidates gave very lengthy answers as to how they
would perform a reserving analysis, often writing several pages, which only earned a
couple of marks on the schedule.
(viii) Methods of transferring the business:
Insured is another Lloyds syndicate so assets and liabilities can be transferred
as needed through a reinsurance to close premium to one of the syndicates
your agency manages
Although the timing of this may not be ideal as that would require waiting
until the years of account are to be closed
A Novation e.g. part VII transfer or LPT would allow the assets & liabilities to
be transferred in entirety
If the managing agency is a listed company their shares can be purchased on
the open market or via a takeover arrangement
which would give instant access to the employees / intellectual capital /
premium base etc. of the company

Page 11

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

and allow the practical process of merging the companies to begin / the
underwriters to continue writing to your agencys current syndicates
Key stakeholders would need to be involved, namely Lloyds / FSA, both of
which would have minimum capital requirements for the merged entity which
would need to be met
requiring a new capital assessment to be put forward for the combined
business to demonstrate adequate capital backing for the takeover
The reserves of the newly combined entity will need to be reviewed and
signed off
Mentions of GN 12 / 20 / 33 & 50 as needed: professional guidance with
relevant examples.
This was not generally well answered, with many candidates demonstrating little
knowledge of potential routes to purchase and giving little consideration to capital
considerations that would form the cornerstone of regulatory interest in a takeover,
often considering competition rules or treating customers fairly while not even
mentioning capital. Very few mentioned professional guidance. A number of candidates
discussed whether the syndicate should be run separately or incorporated into
Syndicate A.

2
(i) Definitions of MCR, ECR, ICA, ICG
MCR Minimum capital requirement.
MCR is the greater of GICR (general insurance capital requirement) and the
minimum guarantee fund (MGF/BCRR) set by the EU.
Formula based calculation
Essentially comprises capital charges as a percentage of claims or premiums.
The capital charges only reflect the relative riskiness of different categories of
claims and premiums to a very limited degree.
Calculated is retrospective
ECR Enhanced capital requirement.
A more risk sensitive measure than the current EU directive minimum.
Comprises capital charges as a percentage of claims, premiums and asset
values.
The capital charges reflect the relative riskiness of different categories of
assets, claims and premiums.
Currently a soft test of solvency not a hard test
ICA Individual capital assessment.
This capital assessment was introduced by the FSA.

Page 12

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Insurers are required to make their own regular assessments of the amount and
quality of capital that is adequate for the size and nature of their businesses.
Expressed as a percentage of ECR
Aimed to be held at the 99.5th percentile level
ICG Individual capital guidance
The FSAs view of the level of capital that should be maintained.
Based on the FSAs review of the firms assessment of its capital needs and its
risks.
Most candidates scored highly on this very bookwork question.

(ii) Differences Between MCR, ECR, ICA, ICG


Reasons why ECR differs but MCR does not
The ECR considers the assets held, as well as premiums written and reserves
held
Company B may hold riskier assets than company A/hold assets subject to a
higher capital charge.
Company B may hold more assets than company A
The ECR charge factors for premiums written and reserves vary between
classes.
The MCR factors for premiums written and reserves do not vary by class
with very few exceptions
although there is some variation in MCR factors according to size of
premiums and reserves.
Company B may write more premium than company A in classes with higher
charge factors.
Company B may hold more reserves than company A in classes with higher
charge factors.
As the premium income is the same the MCRs will be equal if a premium
basis was used, but the ECRs will still differ due to some of factors used in the
calculation
for example B may use a stronger reserving basis than A (or other relevant
example)
A might have suffered particularly bad claims experience recently, which has
increased the MCR (as it's a retrospective calculation) to the same level as B's
and has had no impact on A's ECR level.
If A has suddenly contracted GWP then its ECR would reduce but MCR
would be subject to the minimum brought forward GICR from last year (less
reduction in o/s claims), which would keep it high. (or equivalent example)

Page 13

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

Reasons why Company A has a higher ICA than Company B


The insurance risk for company A may be higher than for company B.
e.g. compared to Company B, Company A might:
Write more volatile classes of business, e.g. more liability business, or
more reinsurance
Conversely, B may write more of classes where the ECR/MCR factors
are high but the internally modelled risk is lower
Write the same lines of business but select more volatile risks,
e.g.
- there may be more geographic concentration risk in Company As
portfolio
- Company A may be more exposed to natural catastrophe losses
Have a different reinsurance strategy
Have a different reserving policy
There may be differences in the credit risk accepted
Company A might purchase more reinsurance
Reinsurers of Company B may have better credit ratings
Company B has lower outstanding balances with other debtors such as
brokers, because of different distribution strategy (or other example)
There may be differences in the market risk accepted
Company A may have more foreign currency exposure (or other relevant
example)
Company B may have better quality assets (or other relevant example)
There may be differences in liquidity risk accepted
Company B may hold higher cash balances (or other relevant example)
There may be differences in the operational risk or group risks accepted
Company A may be reliant on legacy computer systems (or other
relevant example)
Company A may be more exposed to key policyholders or brokers (or
other relevant example)
The companies may be making different allowances for diversification in
their ICAs
Company B may write a range of uncorrelated classes of business,
whereas company A might only write a single class (or other relevant
example)
The companies may be making different allowance for expected profitability
in their ICAs

Page 14

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

This may be because the classes of business written are at different points in
the insurance cycle.
The difference in ICAs may in part be due to differences in the judgements
made by each company, rather than the inherent riskiness of the businesses.
Considerable judgement is required as part of an ICA assessment.
Some of the most critical areas require the most judgement, e.g. correlation
assumptions.
Alternatively, the difference may arise due to capability/quality or type of
modelling techniques used, rather than judgement.
The companies may be using a different time horizon in their ICA
calculations, although in theory the different time horizons permitted should
be equivalent.
Reasons why Company B was given an ICG greater than its ICA
For company B, the FSA took a different view of the capital required than was
produced by the company.
The FSA may have thought that some of the assumptions made by company B
were too optimistic.
The FSA may have thought that some risks in the business had not been
identified or adequately assessed
Some candidates scored extremely well by considering the various moving parts of an
ICA in detail. Knowledge of the more prescribed nature of the ECR and MCR was more
patchy, although even this section was still well answered. Perhaps the most common
reason for scoring low marks on the question was the allocation of time between the three
aspects, with many candidates giving almost equal weight to all three. Good candidates
recognised that there is really only one reason why the ICG would be higher than the
ICA, i.e. that the regulator thinks the ICA is too optimistic! Other common errors were
simply regurgitating bookwork points about the various capital measures without
addressing them to the question in hand.
.
(iii) Alternative Capital Measures
May wish to consider future changes in statutory capital requirements.
e.g. Solvency II
Company will be concerned about return on future capital requirements, not
just present requirement.
However, there will be uncertainty regarding what the future requirements
may be.
Capital requirements indicated by an internal model (other than the one used
for ICA).

Page 15

Subject SA3 (General Insurance Specialist Applications) September 2009 Examiners Report

The company may use different risk tolerances from FSA regulatory capital
(which is at the 99.5th percentile).
The company is likely to want to hold more capital than the ICG, as this is a
fairly low level of capitalisation.
May wish to hold more capital to avoid regulatory intervention/interference.
The internal model will allow for diversification of the target with existing
business.
One target company may have a much lower capital requirement than the
other if it is uncorrelated with the purchasers existing activities.
Capital requirements in other jurisdictions
Purchaser may be regulated outside UK
Capital required by rating agencies.
Insurer may need to retain certain rating to attract business.
Different rating agencies may impose different capital requirements.
Consider amount of capital needed to purchase the target
Need to get return on purchase price, not just the capital held by the insurer.
May be interested in capital requirements of target on some other basis
e.g. allowing for purchasers future business plans, such as increasing
premium or changing reinsurance requirements.
Level of capital held by its competitors so that it is not out of line with them
This was poorly answered. Many candidates focused only on formal capital requirements
with acronyms, talking at length about GICR or MCR etc., giving no thought to wider
business objectives and the role that capital considerations might play.

END OF EXAMINERS REPORT

Page 16

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
26 April 2010 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes before the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only, but notes
may be made. You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 A2010

Faculty of Actuaries
Institute of Actuaries

(i)

Outline the basic structure of Solvency II.

[2]

(ii)

Describe the role of the Actuarial Function.

[3]

(iii)

Identify the key regulatory regime changes to Solvency II from Solvency I. [5]

An actuary joining a London market general insurance company has been appointed
to manage the Actuarial Function under Solvency II and has been asked by the Chief
Executive Officer to assist in the setting of the business plans for the next three years.
This planning process will include a major review of the business as the company is
currently experiencing poor results, partly as a result of the economic climate, with a
steady deterioration in profits over the last two years.
The company writes three of its five classes of business through brokers (commercial
property, EL and D&O), and the other two directly via internet sales (motor and
personal creditor). The gross written premium is similar for each class. The D&O
and personal creditor lines have experienced poor and deteriorating underwriting
results over the last three years, the commercial property and motor lines have
experienced volatile results, while the EL line has seen small but steady profits.
The overall results for the last two years have led to returns on shareholders capital
of 15% and 20% respectively. The corresponding budget figures for these two
years were +5% and 0%.
From the initial work on looking at the problems, it has been noted that:

there has been poor risk selection with a large influx of unprofitable business
the D&O line has experienced three very large losses
a blanket 20% rate increase has been applied across all lines of business
the average cost of claims has increased across all lines of business
prior year losses are emerging owing to under-reserving for reported claims
expenses are escalating and exceeded the budget by 30% last year

The objectives of the company are to turn the situation around over the next three
years, with target returns of 7%, 0% and +10% respectively on the shareholders
capital. The financial planning process used historically has been as follows:

The starting point is the total earned premium and aggregate earned to incurred
loss ratio for the previous 12 months.

Premiums are adjusted in proportion to rate changes, in this case a 20% increase.

Claims are increased by 3% for claims inflation.

Expenses are assumed to be the same monetary value as the actual expense cost
for the previous 12 months.

Commission and investment income are assumed to be the proportion of


premiums and assets respectively as in the previous plan.

A draft report is required by the Board of Directors explaining the problems


associated with the current method used for financial planning.

SA3 A20102

(iv)

Outline the key issues which should be discussed in this report, including any
potential compliance issues under Solvency II.
[17]

The Board recognises that the current planning process does not provide adequate
insight into the business and is unlikely to be acceptable under Solvency II.
Suggest key considerations for improving the planning process, including potential
actions and appropriate measures to track the effectiveness of the process, with
respect to the following aspects of the plan:
(v)

Run-off of existing gross reserves.

(vi)

Future business volumes and gross performance.

(vii)

Reinsurance issues.

SA3 A20103

[11]
[9]
[10]
[Total 57]

PLEASE TURN OVER

A UK general insurance company has recently formed a new business division,


known as the Special Investments Division (SID). SID accepts transfers of portfolios
of business from other general insurance companies.
(i)

Give reasons why other general insurance companies may wish to transfer
portfolios of business to SID.
[5]

(ii)

Give reasons why SID may want to accept transfers of portfolios of business
from other general insurance companies.
[5]

Mojo Insurance is a large international insurance business. Mojo is financially strong,


with a Standard & Poors credit rating of AA. Mojo operates in almost every country
in the world through a number of subsidiary companies. Each of its subsidiaries
underwrites a number of lines of business.
Following a strategic review, Mojo has decided to stop underwriting two lines of
business:

Travel insurance Mojo currently sells annual travel insurance policies in several
countries. All policies are sold direct to customers via a website which is owned
and operated by Mojo.

Employers liability reinsurance Mojo underwrites excess of loss reinsurance in


the London market for general insurance companies writing EL for UK
employers.

Mojo would like to reduce its exposure to these businesses as soon as is reasonably
possible while still maintaining a reasonable return on capital.
(iii)

Describe possible exit strategies for the above lines discussing for each
strategy and each class of business whether the exit strategy is likely to be
attractive to Mojo.
[25]

SID is currently considering accepting two portfolios of business. The table below
shows details of each of the portfolios under consideration.
Portfolio
Liabilities (best estimate)
Transfer Premium

$2.0 million
$1.0 million

$0.1 million
$1.0 million

The liabilities have been estimated by the general insurance companies that wish to
transfer the business to SID, and the transfer premium is the amount in excess of the
best estimate quoted by SID to assume the liabilities.
A director of SID has expressed surprise that the transfer premium proposed by SID is
the same for both portfolios A and B, given that the estimated liabilities for portfolio
A are much larger than those of portfolio B.
(iv)

Suggest possible reasons why the two transfer premiums are different relative
to the best estimates of the liabilities.
[8]
[Total 43]

END OF PAPER
SA3 A20104

Faculty of Actuaries

Institute of Actuaries

EXAMINERS REPORT
April 2010 examinations

Subject SA3 General Insurance


Specialist Applications

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

R D Muckart
Chairman of the Board of Examiners
July 2010

Comments
Individual comments are shown after the solutions to each part question that follows.

Faculty of Actuaries
Institute of Actuaries

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

(i)

Basic structure of Solvency II:

Based on 3 pillars:
Pillar 1: Minimum capital requirements
Solvency Capital Requirement (SCR) and Minimum Capital Requirement
(MCR)
Pillar 2: Supervisory Review Process
Supervisors may require additional capital against risks not covered in
Pillar 1.
Pillar 3: Disclosure
To harness market discipline by requiring firms to publish certain details
of their risks, capital and risk management.

Comments on Q1(i): Generally well answered, although many candidates could not
describe Pillars 2 and 3 and a number of candidates were not able to name the three
Pillars.
(ii)

To undertake to:

Coordinate/oversee the calculation of technical provisions;


ensure the appropriateness of the methodologies and underlying models
used as well as the assumptions made in the calculation of technical
provisions;
assess the sufficiency and quality of the data used in the calculation of
technical provisions;
compare best estimates against experience;
inform the administrative or management body of the reliability and
adequacy of the calculation of technical provisions;
express an opinion on the overall underwriting policy;
express an opinion on the adequacy of reinsurance arrangements;
contribute to the effective implementation of the risk management system,
in particular with respect to the risk modelling underlying the calculation
of capital requirements.

Comments on Q1(ii): Very few candidates could give a full range of requirements for
Actuarial Function. Candidates were expected to be sufficiently up to date on such issues as
Solvency II to answer this question. Many candidates descriptions of the Actuarial Function
were rather lightweight. Most recognised the actuarial role in overseeing the technical
provisions but failed to recognise the wider aspects of the role, e.g. the input into the overall
management functions and confused the formal role of the actuarial function with the day to
day work of actuaries.
(iii)

Page 2

The Solvency II regime will be maximum harmonizing, aiming to apply


consistent regimes across all member states
These new solvency requirements will be more risk-based and more
sophisticated than in the past, enabling a better coverage of the real risks
run by any particular insurer

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

The new requirements move away from a crude one size fits all
approach to estimating capital requirements to more entity specific
requirements
More comprehensive covering liabilities & assets, e.g. harsher asset
recognition rules
New total balance sheet approach, with interactions considered
Now need to hold capital for insurance, market, credit, liquidity and
operational risk
New rules will compel insurers specifically to focus on and devote
significant resources to the identification, measurement and proactive
management of all risks
Prospective basis
Introduction of Own Risk and Solvency Assessment, or ORSA
Different disclosure rules,
including Solvency and Financial Condition Report, or SFCR
More recognition of the differences between insurers and reinsurers
Penalises underpricing/ underreserving

Capital requirements:
MCR Minimum Capital Requirement if net assets fall below which a
firm will be de-authorised
SCR Solvency Capital Requirement further up regulatory ladder of
intervention
These can either be based on standard formula or an internal model
Comments on Q1(iii): The main thrust of this part was to recognise the more risk focussed
regime under Solvency II, and while most candidates commented to this effect, only a
minority were able to provide sufficient detail. Candidates generally did better on many of
the high level differences but many failed to give the detailed, deeper points required.
(iv)

The actions and measures should be sufficient to ensure that the operation is
on target to follow its strategy and meet its goals.
If the measures being used to monitor the progress of the plan dont produce
the results expected, then there may be a need to revise the key actions in the
light of the new knowledge to ensure that the plan is back on track.
Problems with the current method
Aggregation across business lines / mix change / rate change issues

Even for internal purposes, there are clear weaknesses with aggregating
business lines in this manner/better to take each class separately
It is likely that different lines of business will be at different levels of
profitability, and as such any mix changes could significantly affect
average loss ratios.
There would be shifts in mix between years of account with different
strategic focuses in response to market opportunities and challenges
Mix changes are particularly to be expected following the blanket
premium increases imposed,

Page 3

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

as competition is ignored,
and the insurance cycle is rarely consistent across classes and so the
change is likely to be too high for some classes and too low for others
There may well be mix change effects even within lines of business
E.g. creditor business has been displaying significantly worse profitability
in the construction sector following the recent property market issues
The premium increases may also have unexpected effects on profit even
within classes, as good business is cherry-picked by competitors with such
high rate increases mainly only achievable on distressed or undesirable
business

Issues with starting loss ratio

A single years experience is highly unlikely to be representative as even


the largest portfolios will display some degree of volatility
This is in particular the case for the lines of business covered here with
exposure to:
- Large claim events from D&O
- Catastrophe claims from commercial property
- Aggregation issues on creditor business due to macroeconomic
conditions
Some credit for other suggestions, e.g. macroeconomic factors on EL or
natural catastrophes having some impact on motor ( each, max )
Using a single year makes no allowance for potential trends
The previous 12 months data may also be difficult to interpret without
some consideration of trends for longer tailed classes in particular, as
business earned over that year may come from a number of different origin
periods, e.g. with different coverage conditions.
and on an earned premium basis the rate increases will not be fully in
effect over a year
The historical under-reserving of the business suggests that there may be
further weaknesses in the reserves for the previous 12 months for which
some adjustment should be made
These under-reserving factors may also lead to distortions in the data
because of the effects of prior years, and the analysis may be clearer if the
results are separated.

Claims inflation issues

Page 4

Use of single static claims inflation figure whichever year the planning
exercise is considering is clearly inappropriate
The claims inflation assumption should be updated regularly to take
account of any mix changes as the expected inflation is likely to be
significantly different
depending on the classes of business (so rather than using one average
claims inflation figure better to plan each class separately each with its
own inflation figure)
and within classes by claim type
and within a class/claim type, from one year to the next

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Claims inflation fluctuates in proportion to expected claims costs, so the


weighting between different business lines is likely to be different to the
weighting by premium for average rate increases
Most critically, creditor business is enormously sensitive to
macroeconomic conditions,
e.g. with the recent economic downturn producing dramatic increases to
loss ratios across the market
with the possibility that this continues to deteriorate in the future
The higher average claims costs across all lines of business suggest that
other classes are also experiencing some degree of claims inflation beyond
that expected

Other issues

Assuming the same absolute expense costs as in the previous year is likely
to be wholly unsuitable
Also as expenses exceeded the budget in the past they might also do so in
future
There is likely to be some degree of expense inflation over the year
Expenses can be split into variable and fixed
The variable part would not be expected to scale with any increases in
premium purely attributable to rate change
and is dependant on whether new or renewable business
There may also be one-off items present or absent in the previous year that
should be adjusted for appropriately
There are also issues with focusing on absolute monetary amounts for
expenses as combined ratios are then heavily dependent on correct
premium volume assumptions
In this instance, the method used is highly simplistic and likely to
significantly overstate premium volumes following such a high requested
rate increase
Average commission rate may change due to mix change effects
May also be higher on average if brokers request better commission rates
for successfully implementing the relatively high rate increases requested
Investment returns could vary significantly between years according to
market conditions / prevailing base rates etc.
and are likely to be depressed because of the current economic climate
and would be expected to vary if there were any changes in the mix of
assets backing the liabilities
The plan also makes no consideration for any changes to reinsurance
arrangements or the related costs.
Monitoring should be implemented more frequently than annually and
should allow for seasonality

Solvency II considerations

A key consideration in Solvency II is that all financial planning must


closely match the underlying business.
Page 5

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

This would ideally involve splitting the business into relatively


homogenous groups
At a minimum there is a standard segmentation into broad business
categories that must be observed for SCR purposes
Although this does not need to be matched for internal reporting
As such the aggregation of all lines of business for planning would clearly
not be compliant.
If an internal model is approved for calculating SCR this would be
required to be used as part of the planning process and so a clear link
would need to be evidenced
Other appropriate comments linking the internal issues discussed below to
a Solvency II framework

Comments on Q1(iv): A number of candidates set out a proposed methodology rather than
focussing on the shortcomings of the current process. Whilst there was some overlap with
what the question required, this approach cost many marks. Also, many of the candidates
were trying to solve the problem of why the company's results were poor, and what to do to
turn the results around, rather than considering the problems associated with the planning
process. Most focused on the need to analyse at a more granular level, but many failed to go
into sufficient detail of how the analysis would need to change. A number of candidates
pointed out the problems with the analysis but did not suggest any suitable improvements to
be Solvency II compliant.
(v)

Run-off of existing gross reserves


General points

Past under-reserving issues suggest that latest reserves may still be


significantly flawed and a thorough reserve review should be performed

Priority should be given to the areas of the business / aspects of the


reserving process responsible for the most material past shortfalls
Especially where such issues may affect business going forward.
Reserves would need to be set on a discounted best estimate basis and a
risk margin assessed for Solvency II
Consideration should also be given to reserve uncertainty to ensure there is
adequate capital backing for any further deterioration
Stress and scenario tests/stochastic modelling should be used
Allow for past & future inflation appropriately.
Allow for any recent events which may be likely to produce losses which
have not yet shown up in reserves, e.g. recessionary impacts on D&O
Expense analysis, including finding the cause of the escalation and
reviewing expense allocation
Consideration of split of claim amounts:
- Opening
- Closing
- Reopening
- Speed of settlement

Page 6

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Frequency vs severity
Nil claims
Court award fees
Any other reasonable suggestion

Individual case estimates

Consideration should be given to the overall quality of case estimation,


particularly for large claims
Review all large case reserves above a particular threshold
General spot checks on all reserves

Initial loss ratio expectations

Longer tailed classes such as EL are likely to use loss ratio expectations
for immature years, review these
Such a review should take account of
- Historical experience
- Adjusted as appropriate for large claims etc.
- Any viable benchmarks
- Rate changes
- Including coverage changes
- And claims inflation
- Mix changes
Any assumptions used should be consistent with internal modelling for
Solvency II

Individual case estimates/Initial loss ratio expectations:


Watch for fraud
Ongoing individual case review
Creation of a large claim monitoring committee
Peer review process for claims departments
External consultants
Development projections

Obtain triangles of data in order to assess the adequacy of the reserves for
each class of business.
Possible projections include:
- Paid
- Incurred
- Premiums
- Numbers
- Gross & Net
- Accident or underwriting year cohorts
- Other valid suggestions
These triangle should be segmented at an appropriate level balancing the
need for homogeneity with credible data volumes

Page 7

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Although more granular segmentations could be used with aggregated or


benchmark development profiles
These more detailed segmentations could be used to identify or examine in
more detail any problem areas
Using such segmentations in any actual vs expected analysis might also
highlight more quickly areas of the business where existing development
profiles are inappropriate.
Key segmentations would include
- Business line
- Sub-category (e.g. EL / PL, comp / 3rd party)
- Types of claim (property damage / injury)
- Broker
- Large / attritional / cat
- Nil claims (claim number projections)
- Other valid suggestions

Distortions from the presence or absence of large claims / catastrophes /


heavy / light experience should be adjusted for
Along with any changes in claim reserving methodology / recording of
claims etc.
Coverage changes should also be considered, especially when they affect
development profile (e.g. claims made vs losses occurring)
Inflationary distortions to development profiles should be considered
(likely to only be an issue where there are changes to settlement rate or
levels of claims inflation)
Also:
Regular reserve review
AvE analysis
Establishment of a reserve committee with primary responsibility for
reviewing the actuarial numbers
Greater use of external auditors / peer reviewers / consultants etc.
Comments on Q1(v): This was generally poorly answered, with many candidates only
detailing high level issues. Very few candidates suggested the projections which should be
considered, or the segmentations which may be useful and to give detail of how a reserve
review would be carried out.
(vi)

Future business volumes and gross performance


General points

Page 8

Speak to management/ marketing/ underwriters / consultants / analysts etc.


Review underwriter incentives for accurate planning, e.g. charge for
capital on the basis of original plan, make bonuses conditional on
performing close to plan (whether over or under)
Integrate seasonal targets, road maps etc. so that performance can be
tracked on e.g. a monthly/quarterly basis to give warnings about the full
year
Implement DFA

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Projecting future volumes:

Starting point for projections should be existing business volumes


This information should be segmented as far as is practically feasible into
groups where the expected changes going forward are reasonably similar.

And where starting profitability is similar for the future profit expectations
(or regrouped to fit as appropriate)
Segmentations that bear some resemblance to rating factors used in pricing
can be of value, particularly in future monitoring.
Expiring business should be adjusted for expected retention levels
This should take account of the following issues:
- Normal level of churn based on average historical experience
- Likely impact of rating levels / rate movements relative to the market
- Any strategic changes such as non-renewal of specific areas
Adjustments should be made for likely changes in premium volumes on
retained business:
- Average exposure growth, e.g. from wage-roll inflation / sum insured
increases
- Consideration should be given to unusual economic conditions that
might make these averages inappropriate
- Any other exposure changes such as limits, line sizes, deductibles etc.

- Any rate increases after allowances for such exposure changes should
follow through as increases to premium volumes
- Currency movements will produce changes in volume in the base
accounting currency.
New business should be allowed for
- Changes to market share
- Influenced by price differentials
- quality of coverage
- broker relationships
- direct marketing / website quality / advertising spend
- Changes to overall market volumes (economic cycles, expanding
levels of coverage etc.)
- Strategic considerations, particularly on a segmental level
- Competition
- Elasticity of demand
- Position in market cycle
Movement analysis: lapses, cancellations, conversion rates.

Future business profitability

Reserving analysis in part (v) should suggest appropriate starting point for
projections of future profitability
Note that such projections should consider average expectations such as
IELRs for BF NOT actual experience for the most recent year
Although they should be appropriately adjusted to the relevant rating
environment

Page 9

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

And claims environment


From that starting point, projections with the appropriate segmentations
above should produce reasonable results
Although differential rate movements by segment should be considered if
market conditions are likely to produce this
Along with different levels of claim inflation
Antiselection
Consider uncertainty related to creditors because of industry concerns, e.g.
Competition Commission.

Comments on Q1(vi): This was not particularly well answered, with many candidates failing
to suggest issues which can affect new business volumes, and relatively few mentioning how
you should look at future business profitability. Very few candidates gave any consideration
to the key elements of any planning process the expiring volumes and the likely retention,
growth and new business given the current market conditions.
(vii)

Reinsurance
General points

All internal or external models used for reinsurance analysis for planning
should be consistent with those used for capital modelling in order to be
solvency II compliant.
Consider capital implications of reinsurance purchasing.
hire consultants to benchmark any of the aspects below
define regular review of any parameters, perhaps quarterly, with the
process documented for solvency II compliance
Alternatives to reinsurance: derivatives, cat bonds etc.
Technical assistance from reinsurers
Other appropriate actions or measures not covered elsewhere (must be
specific and workable approaches)

Estimation of reinsurance spend

Page 10

Starting point would be:


assessing value for money
reviewing existing arrangements, covers, limits etc.
reinsurance spend from current year adjusted for intended changes to
programme
Adjustments should allow for both changes in modelled loss experience
and changes to likely capital charges from reinsurers
Capital charges should particularly be considered if the level of attachment
for the programme is changing significantly.
Also consider level of recent claims passed to reinsurers and any potential
resulting rate increases / decreases
and whether reinsurance programme has been effective with respect to
recent large claims etc.
Along with major market events that might impact capacity even if these
did not impact the company

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Consider the position in the reinsurance cycle.


which may be at different stages for different lines of business
If the programme is being changed significantly, consider any discounts
for volume or cross-subsidies that may exist between layers / business
lines etc in the current reinsurance purchases, e.g. costs for other layers
may increase if one layer is dropped.
Consider cost allocation between units for shared programmes
Speak to reinsurance brokers / reinsurers for indicative quotes
review planning assumptions whenever there are company or market
events which would affect this or when intended purchase changes

Estimation of reinsurance recoveries

Model recovery rates for the historical and planned future purchase
using a stochastic model, e.g. the internal capital model for Solvency II
compliance
The cost of purchase should give some indication of likely recovery rates
based on some assumptions about target loss ratios / capital costs for
reinsurers this can be used as a benchmark to the internal model.
Past recoveries (adjusted for programme changes as necessary) can also be
used as a sense check on the internal model although experience may not
be particularly credible or stable.
External software such as RMS can also be used for the catastrophe
accounts.

Bad debt adjustments

Consider potential bad debt issues that might reduce the recovery rate /
increase capital costs
Some allowance should possibly be made for current economic conditions
with defaults arguably more likely at what is currently a low point in the
economic cycle, with additional capital difficult to raise by the company or
its reinsurers.
Historical and current Credit ratings for all reinsurers will be available,
along with market information such as S&P default studies to suggest
likely default rates.
Consideration should be given to any interaction between extreme events
for the company and events that might trigger defaults for reinsurers for
example high CAT losses might be combined with reinsurer default
whereas individual risk losses on EL are unlikely to be significant for any
reinsurers.

Comments on Q1(vii): A significant number of candidates answered incorrectly parts (v) to


(vii). Specifically, rather than commenting on the planning process, they focussed on the
underlying issues themselves. In part (vii) a discussion of how to monitor the effectiveness of
reinsurance was required, not a detailed description of the types that may be appropriate.

Page 11

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Many candidates appeared to be answering a standard "how would you decide on the
appropriate reinsurance programme" question.

(i)

Reasons for Transferring General Insurance Business


Where the businesses is no longer a core part of the sellers portfolio.
For example:
The business represents a small proportion of the total business, but requires a
disproportionate amount of management time or capital.
The business was purchased as part of a larger acquisition, but was not the
reason for undertaking the transaction
The business is, or is anticipated to become, not sufficiently profitable for the
current owner. This may be because it has been making losses for a number of
years, or regulatory/other changes mean the outlook is unfavourable
Easing balance sheet strain (particularly an issue where reserves are
discounted)
Simplification of company structures, e.g. portfolio diversity
Where the company or owner is in run-off, e.g. has become insolvent and is
unable to write business.
If the company lacks the support expertise to manage the business effectively
(perhaps due to the loss of a key underwriter)
The efficiency savings made by SID allow them to offer a better price than the
expected value of the business
Presence of the business line is viewed unfavourably by the stock market /
regulators / rating agencies
Freeing up capital would allow the company to take part in other upcoming
market opportunities
The business is in a different currency to the core business causing significant
exchange rate issues
Reinsurance is no longer available or is too expensive.
To remove uncertainty, e.g. for latent claims.
Other appropriate reasons

Comments on Q2(i): A bookwork question which was generally well answered.


(ii)

Reasons for Accepting Transfer


SID will be attempting to make a profit from the transfer, that is, obtain a
satisfactory return on the capital employed by SID on the transferred business.
Through portfolio transfer, SID may obtain a viable underwriting business.
Purchasing a whole company gives SID the benefit of company assets, for
example, existing licences, intellectual property, renewal rights, skilled staff
etc.
This allows SID rapid entry into a market.
May be lower cost than starting business from scratch.

Page 12

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Businesses which are not performing satisfactorily for the seller may become
viable with an alternative management/more efficient management.
SID will benefit from diversification by buying a number of unrelated
portfolios.
In the case of portfolios in run-off, it may cost less to administer several
portfolios together through economies of scale
Diversification could result in reductions in capital requirements for a larger
group of businesses than for each business individually.
By combining businesses with common insureds, SID would have greater
negotiating power in commutations.
Acceptance of portfolios in this manner is core to SIDs business and hence
utilises internal resource (legal departments etc.)
Acquisition of a substantial share of a particular pool of risks may give SID
significant bargaining power to run the business off profitably
It may also accumulate sufficient in house expertise to price and reserve the
portfolios accurately
Cross-selling opportunity.
May be tax advantages. (with reasonable explanation, e.g. arbitrage).
Comments on Q2(ii): A bookwork question which was generally well answered.
(iii)

Possible exit strategies


(Marks are given to the opposite opinion on whether a strategy is attractive or
vice-versa for each of Travel/EL Re if the reason given is valid)
Run-off to exhaustion
Cease to write any new business or renewals in the lines being exited, but
continue to retain responsibility for administration and claim payments for the
existing business.
Once the business has run down to a certain point, expense and management
costs involved in running off the remaining exposure are likely to be
sufficiently high that it will become an ineffective way of maximising return
on capital, although can outsource
Re-entry to the market easier than with other strategies
Travel strategy likely to be attractive
Short tail business all claims should be paid in not much more than a year
Mojo is a large business, so should have the capacity/resources to complete
the run-off (even when the volume of claims has declined to very low levels).
This does however waste the value inherent in the website which could be sold
separately
EL Re strategy unlikely to be attractive
Mojo wishes to eliminate all its exposure as soon as reasonably possible. Runoff for this business will take a long time.
Reinsurance
Fully reinsuring all future claims under business written historically.

Page 13

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Risk transfer may be partial, e.g. there may be a cap on the reinsurers
liability, or the reinsurer may only pay part of the claims.
The insurer remains ultimately liable for the claims cost, e.g. if the reinsurer
becomes insolvent.
In some cases, the reinsurer may also administer the claims run-off.
Cover for the policyholder is maintained.
Travel strategy unlikely to be attractive
Mojo is financially strong, so unlikely to need reinsurance (which would pass
profit outside the group).
EL Re strategy unlikely to be attractive
Given these are long tail liabilities, counterparty default risk cannot be
ignored.
Or: EL Re strategy likely to be attractive
Simple solution subject to the reinsurers credit rating being high enough to
suggest that claims will be paid in the long term
Commutations
Insurance (or reinsurance) policies are cancelled with the agreement of both
parties, subject to a return premium, so that no further claims can be made
under the policies.
Active commutation strategies can accelerate the run-off of the businesses.
Mojo would likely attempt to commute both inwards and outwards business.
Could be antiselection as agreement to commutation is at option of insured
Travel strategy unlikely to be attractive
Not practical, as agreement of all policyholders required
EL Re strategy unlikely to be attractive
As insurer is unlikely to agree as may then wish to commute the underlying
business which requires the individual agreement of every policyholder.
Novation
The complete transfer of insurance business from one insurer to another, with
the agreement of all three parties (insured, old insurer and new insurer).
The old insurer is replaced with the new insurer, with no contractual liability
remaining with the old insurer.
The old insurer pays the new insurer to make this arrangement.
Cover for the policyholder is maintained.
Travel strategy unlikely to be attractive
Not practical, as agreement of all policyholders required
EL Re strategy likely to be attractive
Mojos contractual liability completely removed but depends on the terms.
Insurance business transfer or Part VII Transfer
Complete transfer of business from one insurer to another, so that no
contractual liability remains with the original insurer.

Page 14

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Cover for the policyholder is maintained.


Can be used to smooth the process to a scheme or arrangement, e.g. by
removing business that cannot be part of the scheme.
There is no voting mechanism for policyholders either (unlike schemes of
arrangement). However, policyholders are entitled to be heard by the court
sanctioning the transfer.
Achieves the same effect as a novation, but can be effected for a large number
of policies at the same time and does not require the agreement of
policyholders.
Can transfer reinsurance asset.
Consideration should be given to the premium charged by a third party to take
on exposures such as these (which can be highly volatile and have significant
latent potential)
Onerous legislation and disclosure requirements/ action required by FSA,
court, independent actuary, company lawyers
Transfers subject to FSMA 2000
which requires court sanction/approval
Travel strategy unlikely to be attractive
While it would achieve Mojos aims of totally eliminating exposure, there
would likely be considerable costs involved. Therefore it would be more
efficient to allow the business to run off.
EL Re strategy unlikely to be attractive
Because of onerous legislation etc.
Or: EL Re strategy likely to be attractive
Completely removes Mojos contractual liability and has less problems than a
novation, dependant on price/terms
It may be best to retain the exposures for a few years until they have largely
stabilised before taking this approach
Schemes of arrangement
Effectively a mass commutation of policies of an insurance company.
Must be sanctioned by the court.
Specified majorities of policyholders (both by number and value) must vote in
favour of the scheme in order for it to proceed.
Mojo would not need the individual agreement of all policyholder affected by
the scheme in order for it to proceed.
All creditors are bound by the scheme once it has been approved by the court
(even if they voted against the scheme, or were unaware of the scheme).
Can be put in place for both solvent and insolvent insurers.
It is possible to include all policyholders or just some policyholders. The
scheme document will specify which policyholders are included.
Reinsurers are not contractually bound by the scheme.
Travel strategy unlikely to be attractive
While it would achieve Mojos aims of totally eliminating exposure, there
would likely be considerable costs involved. Therefore it would be more
efficient to allow the business to run off.
Personal lines customers may find the process confusing.

Page 15

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

EL Re strategy unlikely to be attractive


Given Mojo is financially strong, many policyholders would prefer to
maintain cover than receive a payout. It may be difficult to obtain the required
votes.
Reputational risk to Mojo by removing cover from insureds. May have a
negative impact as Mojo continues to write other lines of business.
In any case, preparing for the scheme and gathering the required votes could
take a long time. Mojo wishes to eliminate exposures as soon as possible
Sale of business
Sell the whole company, which achieves finality for the seller.
Not viable when a company only wishes to cease underwriting part of its
business.
Sales of certain business assets are also possible.
Examples of assets that can be sold are the renewal rights or reinsurance
recoveries.
Travel consider sale of renewal rights or website
We do not know why Mojo has exited the business. The business may be
profitable (or could be made profitable), and another company could be
interested in purchasing.
EL Re consider sales of renewal business
However, because this is London market business the broker, rather than the
insurer, owns the account. This can limit the value of the renewal rights.
Combinations of the above strategies are also possible.
e.g. pursue run off initially while agreeing commutations where possible, then
sell business.
Comments on Q2(iii): A methodical approach helped many candidates score well in terms
of discussing the different strategies although too many failed to actually comment on the
attractiveness of each for the two classes.
(iv)

Loadings will need to be applied to the best estimate cost to allow for SIDs
profit, contingencies, and expenses, which would differ for A and B
The type of portfolio transfer mechanism may be different for portfolios A and
B.
Transferring company may retain a share of portfolio A
E.g. a large/particularly uncertain claim may be retained (or other example)
One of the transferring companies may continue to do administration,
resulting in different expense loadings.
SID may charge a minimum premium for any transactions to cover its
overheads, or the work required to prepare a quote.

Page 16

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

SID may disagree with the best estimate prepared by the company transferring
the liabilities.
E.g. one company may tend to over reserve, one under reserve. Estimates may
or may not be discounted (or other example).
There may be differences in the nature of reinsurance on each portfolio.
Differences include whether reinsurance is being transferred to SID, the
quality of the reinsurers, and whether there are any disputes with the
reinsurers.
Also whether claims have exceeded or are close to any policy or aggregate
limits
SID may be applying different contingency loadings for each portfolio.
E.g. one portfolio may be considered to be more uncertain, because of:
long tail vs short tail,
e.g. asbestos liabilities, / latency
the quality of data being poor
currency considerations
in run-off or ongoing
any other reasonable suggestion
SID may not want to acquire Portfolio B, but be keen to acquire Portfolio A,
for strategic reasons.
Competition premium will reflect market conditions, that is, willingness of
buyers and sellers to reach a deal.
SIDs costs may be very different for different lines of business due to
synergies with existing liabilities
Similarly, SIDs capital costs may be very different due to different
diversification credits
Consideration should be given to the number of years of business included in
the transfer and the total original liability although a portfolio may be largely
run off, any late stage or latent development could well bear more
resemblance to the original liabilities than the current best estimates
Liabilities may have been assessed on a discounted basis and the choice of
discount rate may vary between the two portfolios
There may be an error in the figures
Comments on Q2(iv): Marks for this part were low in general. Very few candidates
covered all the obvious points, e.g. different methodologies and bases used by the companies,
different level of uncertainty in the outstanding claims (for many reasons) and the impact of
competition on the price or considered the wider issues around the potential strategic or
operational issues for SID. Some candidates concentrated on the possible volatility of the
portfolios and didn't cover anything else although they generally remembered that SID may
not agree with the beat estimates of the liabilities..

Page 17

Subject SA3 (General Insurance Specialist Applications) April 2010 Examiners Report

Overall: Q2 was generally better answered than Q1. Marks for the Q1(v) to Q1(vii) part of
the paper were particularly low, to such an extent that the average mark awarded for Q1,
with 57 available, was lower than for Q2, with 43 available.

END OF EXAMINERS REPORT

Page 18

Faculty of Actuaries

Institute of Actuaries

EXAMINATION
6 October 2010 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes before the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only, but notes
may be made. You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 S2010

Faculty of Actuaries
Institute of Actuaries

Safetysure is a medium sized UK commercial general insurance company writing


several lines of business, with a focus on longer tailed liability and speciality lines.
The company has been in operation for more than 30 years with a management team
comprising entirely former underwriting staff. Collective responsibility is taken by
this team for all management decisions, on which the executive directors vote before
being signed off by the board of non-executives, who are themselves retired
executives.
Historically, results have been good and the company has been well capitalised.
Management oversight has focused purely on underwriting issues, with no interest in
the risk management processes. This area has been a low priority for intensive
regulatory oversight, and the companys shareholders have had little incentive to
question the management style.
The management team has used external actuarial consultants to meet all reserving,
capital and compliance requirements, taking little interest in their feedback. All
pricing has been done internally by underwriting teams with no actuarial involvement.
Safetysure has just appointed its first internal actuary who has been asked to review
all aspects of the company, starting with the companys management arrangements.
(i)

Outline the key standards for management arrangements, and any potential
concerns with the current approach, in respect of both regulatory compliance
and good corporate governance requirements.
[10]

(ii)

Describe how your answer to part (i) would change with the adoption of
Solvency II.

[3]

The company has no rate monitoring system in place and has not historically captured
any rating data, although most of the underwriters do have pricing models and rating
guides that they often use in their work. Safetysures underwriters have historically
used benchmark market rate movements which they believe adequately reflect their
book.
The actuarys recommendation that a formal pricing and rate monitoring process be
put in place has met with some resistance. The board feels strongly that underwriting
should be left to underwriters and actuarial staff should focus on reserving and capital
modelling.
(iii)

Outline the key benefits of a formal and well documented pricing and rate
monitoring process.
[8]

(iv)

Suggest ways in which actuaries can add value to pricing and rate monitoring
processes.
[7]

The board accepts that there may be some benefits to actuarial review of the pricing
and rate monitoring processes and has requested a review of their UK Employers and
Public Liability book which the company first underwrote in 2001.

SA3 S20102

The reported results for this book have historically been profitable, but the consultants
reviewing the reserves have expressed concerns about the emerging experience. As
the account is still relatively young they have been using market benchmarks for the
development profiles as well as the rate movements. Relative to these benchmarks,
older years, which they would have expected to have run off, are continuing to show
movements and more recent years have shown an unusual level of claims for the
current stage of development.
The claims department has also raised concerns about the volume of small claims
coming through for this book, which they are under-resourced to handle, and has
suggested imposing higher minimum deductibles.
(v)

Describe, giving your reasoning, the investigations that should be undertaken


in order to understand and improve the account.
[23]

The liability portfolios include a number of scheme arrangements with third parties,
who underwrite risks on Safetysures behalf. A number of these schemes contain
profit commission arrangements, under which the third parties receive additional
commission if the scheme meets certain performance criteria. Currently the
underwriters make no allowance for these profit commissions in pricing, reserving or
planning as the performance criteria are set at a level where they would only become
payable if a scheme outperformed its targets.
(vi)

Explain the problems with this approach.

(vii)

Suggest alternative approaches that might allow for profit commissions more
appropriately, with reference to the key factors to consider and any data
requirements.
[19]
[Total 73]

SA3 S20103

[3]

PLEASE TURN OVER

An article has appeared in an international insurance journal describing Takaful. The


article included the following information.

Takaful refers to insurance that is compliant with Islamic law, also known as
Shariah law. Takaful products, which are already well established in parts of
Asia and the Middle East, are increasingly becoming available in Western
countries. A number of major international insurance groups have entered the
Islamic insurance market in recent years.
In addition to the normal management structures present in financial services
companies, Islamic financial institutions appoint a Shariah board. The Shariah
board consists of recognised Islamic scholars. The Shariah board determines
whether the companys products and practices are consistent with Islamic
teachings.
Islamic law forbids the payment or receipt of interest. Investments related to
anything considered unlawful under Shariah law are also prohibited. Contracts
involving uncertainty or speculation are forbidden, as is the use of capital
simply to obtain a return. However, some Islamic scholars are of the opinion
that insurance can be written in ways that adhere to Shariah law. The sharing
of risk within a community is considered to be beneficial to society, and so is
acceptable to Islam.
Shariah law permits the use of capital for social or ethical purposes, and
sharing profits from social or ethical investments is permitted when the risks
are also shared. In addition, it is permissible for a company to be rewarded for
organising the pooling of risk, or for sharing risks with policyholders.
According to most Shariah scholars, Muslims have a religious obligation to
purchase Takaful in preference to conventional insurance.
While many Islamic financial businesses operate on an essentially mutual
basis, new ventures are increasing on a for-profit basis. However,
policyholders are entitled to share in any surplus, and a share of any profits
would typically be given to charity.

Vibe Insurance is a large general insurance company operating only in the country of
Texel. Vibe is a mutual insurer, and specialises in personal motor insurance. Around
5% of the population of Texel are Muslims. There are no Takaful products currently
available in Texel.
The managing director of Vibe has read the article on Takaful. She has asked Vibes
business development team to consider launching a subsidiary to write Takaful in the
country of Texel.
(i)

Define the term mutual insurer.

[1]

(ii)

Describe why Vibe may wish to provide Takaful.

[5]

SA3 S20104

(iii)

Describe the difficulties that Vibe may face in adapting an existing product
offering to be Takaful compliant
[8]

Vibes business development team has prepared a business plan for a Takaful
business. The business would initially only underwrite personal motor insurance.
The table below shows the premium income, claim outgo and expenses that have been
assumed for the Takaful business. For comparison, Vibes current business plan for
personal motor is also shown. The current plan excludes the Takaful business.
An actuary has been asked to review the Takaful business plan and prepare a report
on the business plan to assist the board of Vibe.
Takaful Motor
2011
2012
2013
Premium

20

100

200

1,200

1,260

1,320

Claims

18

80

140

1,080

1,134

1,188

10

120

126

132

Expenses
(iv)

Other Personal Motor


2011
2012
2013

Describe the matters that the actuary should draw to the attention of the board
in his report.
[13]
[Total 27]

END OF PAPER

SA3 S20105

INSTITUTE AND FACULTY OF ACTUARIES

EXAMINERS REPORT
September 2010 examinations

Subject SA3 General Insurance


Specialist Applications

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

T J Birse
Chairman of the Board of Examiners
January 2010

Institute and Faculty of Actuaries

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

(i)

The key guidance is the Senior Management Arrangements, Systems &


Controls (SYSC) standards.
High level standards under Principle 3.
These cover areas including:

Business structure & contingency planning


Training, competence & expertise
Compliance, internal audit & financial crime
Risk control
Outsourcing
Record keeping
Conflicts of interest
Other appropriate points

They encourage firms directors and senior management to take appropriate


practical responsibility for their firms arrangements on matters likely to be of
interest to the FSA.
This includes taking reasonable care to establish and maintain appropriate
systems and controls.
The nature and extent of the systems and controls will depend on a variety of
factors, including:

Nature of business
Scale of business
Complexity of business
Diversity of business types
Geographical diversity
Volume of transactions
Size of transactions
Nature of transactions
Degree of risk of each area of operation
Other appropriate points

Currently management takes no interest in risk management processes, which


should be the practical responsibility of directors and senior management.
Perhaps reflecting the fact that the company is well capitalised
This lack of involvement leads to a significant operational risk for the
company
Left unchecked this could ultimately lead to regulatory, legal and reputational
risks
Even if no operational risk losses occur, the company should still be holding
additional operational risk capital to reflect the low quality of their
management processes, reducing the companys capital efficiency and return
to shareholders.

Page 2

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

The SYSC standards also recommend that companies should vest


responsibility for effective and responsible organisation in specific directors
and senior managers
As the directors take collective responsibility for these issues, they are not
following best practice, and the lack of individual responsibility for any
particular issue may lead to ineffective management.
The SYSC standards also recommend that a firm should segregate duties of
individuals and departments in a way that reduces opportunities for financial
crime or the contravention of regulatory requirements.
This management approach may also affect the costs of D&O cover for the
management team
Management expertise is currently too heavily weighted towards underwriting.
The team should incorporate a wider range of skills and backgrounds.
Actuarial involvement in pricing for example would be of value.
Examples of other potential board members.
The over-representation of former underwriters on the management board may
also lead to biased judgements and conflicts of interest.
This is especially the case for the longer tailed business, where directors may
not objectively consider the emerging experience on business they themselves
wrote
Where bonuses are dependent on underwriting experience, this may produce
an even stronger vested interest.
The close links between executive and non-executive directors lead to little
diversity of opinion, and lessen the likelihood of appropriate challenges to the
executive directors from the board of non-executives.
As this is the key reason for the involvement of a board of non-executives, this
does not constitute best practice.
The fact that the executive directors are retired may also be an issue as they
may be out of touch with the market.
Need to review work of external consultants.
Choice to outsource means limited involvement.
Comments on Q1(i): Very few candidates were aware of the relevant legislation. The
majority of candidates did however manage to use the information in the question to form
some sensible conclusions about the key weaknesses in the management structure proposed
with better candidates addressing each of the issues in turn:
no interest in risk management
collective responsibility
they are all underwriters
non-executives are ex-underwriters
A number of candidates focussed on other issues such as ensuring adequate capital is held or
acting in the interests of shareholders, missing the focus of the question.

Page 3

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

(ii)

Impacts from solvency II


The supervisory review process covered by Pillar II will involve increased
regulatory oversight of companies internal processes
This includes having internal audit and actuarial functions with sufficiently
skilled personnel in-house to oversee and challenge, as necessary, the work of
the company
Where deficiencies are identified in these processes, regulators can require
additional capital to be held under Pillar II
Under Pillar III requirements on Solvency II, firms will be required to publish
details of their risk management processes
In order to publish explicit details of risk management processes, management
will need to develop and be able to articulate a formal risk management
approach
These should address the different risks faced, considering for each risk
category the risk exposure, concentrations, mitigation potential and
sensitivities.
Solvency II also sets higher standards for awareness of capital considerations
among senior managers,
which currently appear to be low as all capital work is outsourced with
little interest in any feedback.
Appropriate mention of ORSA
Appropriate mention of Use Test or IMAP process
This is a particular issue if the company wishes to use an internal model as
they are unlikely to pass the use test.

Comments on Q1(ii): Candidates displayed a disappointing knowledge of Solvency II, a


highly topical issue that we would have expected candidates to be largely prepared for,
particularly as it has been part of the exam for the past two sessions. Few candidates made
reference to ORSA or the Use Test. A number of candidates offered some general comments
on Solvency II with little effort to tie this back to the question.
(iii)

Formalised data capture at individual risk level produces more accurate


information than subjective judgements on the account as a whole
This would not only generate the rate movement information that the
consultants are benchmarking
but can also be used to estimate the impact of:
lost business
mix change on renewing business
new business
This information is of great value for other areas of the business,
e.g. for reserving, capital modelling or business planning
Management should also have the best information possible on market
directions so that they the company can be refocused appropriately.

Page 4

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Capture of data at individual risk level can improve understanding of the


account
Mix change effects in particular would not be picked up by the benchmark
index approach.
This would reduce risks of anti-selection / competitive issues.
Having a more formal and documented rating process will encourage the
underwriters to use their rating models for all their policies
This will reduce the risk of errors in judgement or mental calculation
Formal process and consistent models could be more easily updated for
changing assumptions.
A formal rating process also ensures greater consistency between:

Years
Underwriters
Risks
Classes

Robust rating processes are viewed favourably by key stakeholders including


reinsurers, brokers, regulators, rating agencies and capital providers /
shareholders
This could lower the cost of reinsurance or capital,
attract more and better quality business
and lessen regulatory issues.
Under solvency II, poor pricing processes could lead to a higher capital
requirement.
A formal rating process produces a clear audit trail that allows better internal
controls
This audit trail will also assist with business continuity / handover in the event
of a change in underwriting staff
A project to formalise rating processes would most likely identify other areas
of work which may yield other benefits, e.g. more streamlined writing of
business, refinements to the rating model etc.
Comments on Q1(iii): This question was generally comparatively well answered. Some
candidates did give too much focus on what to take into account in setting a technical price,
whereas the question very specifically focused on the pricing process.
(iv)

Technical Pricing Assistance


The actuary can assist with calculation of the technical rate
Helping with loadings (expenses, capital etc)
Modelling Latent claims
Modelling Large claims & CATS
Triangle projections
Sensitivity and scenario testing

Page 5

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Relevant generic examples


and development of underwriting models
Relevant generic examples
particularly if the company has a sufficient volume of credible data to analyse
The actuary can act as a second pair of eyes
or in an audit capacity
This may be of particular use in checking that the models already in use are fit
for purpose
Actuaries can offer assistance with pricing of individual larger risks or
schemes
This can be of particular value when the risk has some unusual features which
are not always allowed for accurately by underwriters
e.g. profit commissions, aggregates, excess layers or other relevant examples
Actuaries can help ensure that appropriate allowances are made for claims
inflation
This could incorporate their knowledge of regulatory and legislative changes
Examples PPO / Ogden tables
Actuaries could also contribute awareness of fiscal changes e.g. tax laws
Allowance could also be made for changes to the general claims environment
e.g. because of recessionary effects
Process Changes
They can design rate monitoring processes
and data systems to capture the data
They can help monitor aggregations and accumulations
Mention of actuarial control cycles
General support / communication
The actuary can act as a second pair of eyes
or in an audit capacity
This may be of particular use in checking that the models already in use are fit
for purpose
They can offer sign off of pricing and other terms and conditions as part of the
license process
Actuarial involvement is viewed favourably by key stakeholders including
reinsurers, brokers, regulators, rating agencies and capital providers /
shareholders.
They can communicate with third parties around rating issues including
regulators, reinsurers, agents / brokers, senior management, clients, finance
staff, claims departments, and other actuarial functions such as reserving and
capital modelling.
Comments on Q1(iv): A number of candidates focussed on the individual components of a
technical rate and neglected to even mention practical issues such as development of pricing
models.

Page 6

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

(v)

Specific considerations
Mix of business
One key issue is the mix of trades underwritten
This should be considered throughout all years of account as far as the data
allows as mix changes over time could lead to a significantly different
development profile.
The lack of adequate rate monitoring processes could potentially exacerbate
the impact of mix changes, with the potential for the historical experience to
have been benign due to cross-subsidies from areas of business that no longer
make up such a high proportion of the account.
The high early stage development of more recent years compared to the high
late stage development for earlier years suggests that there may well be some
mix change factors which should be accounted for.
Benchmark issues
The benchmark may be inappropriate for a number of reasons

Mix / Class
Length of tail
Out of date
Territories
Coverage
Different claims handling process
Just plain wrong

Late stage development / potential latency issues


Also, the development profile appears to be longer than a benchmark market
development curve, suggesting that the mix of trades might be biased towards
longer tailed exposures.
One example might be trades with high levels of chemical exposures causing
latent illnesses not identified for many years after the policies are written
or pollution liabilities where the environment damage is not immediately
apparent
or industrial diseases such as vibration white finger
Any other relevant examples
As the account has only been in existence since 2001, there could still be
significant exposure to risks such as these so it is critical to understand the
historical mix of business.
Once the potential exposures are understood it should be possible to make use
of market benchmarks for APH development to supplement the internal data.

Page 7

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Discussions with the consultants should indicate the level of potential latency
they have been allowing for in the previous reported results.
Depending on the quality of data systems, analysis of the late stage claims by
type of claim should give some indication as to the source of the late
development.
Failing that, discussions with the claims department about some of the larger
individual cases may give an indication.
Consideration should be given to any changes in policy wording that may
affect development profiles
Although it should be noted that no EL business can currently be written on a
claims made basis
Potential weaknesses in claim estimates
The unusual late stage development could also reflect poor quality reserving
and claims estimation processes
Analysis of the late stage movements should give some indication as to
whether this is the case. If the movements are the result of increases in
reserves for claims already recorded, this would suggest a deficiency in the
claim estimation process,
Whereas late notification of claims would be more likely to indicate latent
exposures
The combination of this feature with the high level of early reserves for more
recent years suggests that claim estimation processes across all years of
account may have moved to a different basis in recent years.
Claims inflation may also have been higher than anticipated.
Legislative impacts may also have increased reserves.
If reserves were discounted in the past, need to know the levels of discount
applied and their changes over time. Need to check that the market data used
for benchmarking is on the same basis.
Other examples of changes to the reserving process, e.g. Levels of prudence
All such changes should be fully investigated as they are critical to any
analysis of the claim development.
Regardless of any changes to processes, some assessment of the
appropriateness of the current claim reserves would be of value
Spot checks on individual claim reserves may assist with this, perhaps using
an independent external claim assessor.
Influx of small claims
EL & PL are low frequency classes so claims volumes would not generally be
expected to be significant.
An analysis of the nature of these claims should be undertaken in order to
understand them better

Page 8

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Potential explanations may include:

A weakness in the policy cover


Selection effects with a particular type of client
Perhaps due to poor pricing structure (particularly as regards premium
loadings to act as a disincentive to make small claims)
or the quality of vetting of clients health and safety processes
or unusually low levels of deductibles or warranties
The company may have a reputation for poor claims control leading to an
increase in fraudulent claims
This will be a particular issue with the recent recession
although a recession on its own may lead to a higher propensity to make
valid smaller claims
There may be a few key insureds driving this experience
Claims farming activity
Claims leakage / poor claims handling
Greater awareness of potential to claim / litigiousness
Other valid explanations

If the issues are driven by specific issues different process could be explored

Changing deductibles on PL
Sunset clauses
Different pricing structures
Other appropriate examples

It is important to note that deductibles can NOT be imposed on EL policies,


only warranties
As EL is statutory cover to protect employees from the risk of negligence on
the part of the employer, all financial liability for the claim must pass to a
regulated insurer
Policies can however be written with warranties under which the employer
commits to reimburse the insurer for the first tranche of any claim.
In most cases this acts in the same manner as a deductible, but in the event of
failure of the employer the employee would still be covered for the full
quantum of claim without any reduction from the deductible.
At an account level therefore, warranties would produce broadly similar
benefits to deductibles in terms of disincentives for the coverholder to make
small claims.
If the claims appear to originate from a few key insureds, risk types or trade
types, consideration should be given to non-renewing those clients, tightening
the policy wording or declining / increasing rates on those trade types.
General investigations
Undertake internal reserving analysis
Review current pricing structure & models
Individual analysis of any particularly large cases

Page 9

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Cost / benefit analysis for any outwards reinsurance


Analysis of backing investments
Assessment of capital costs (with reference to mix type / weighting to large
claims etc.)
Review of potential accumulations.
Generic expense analyses
Benchmarking exercises against competitors
Profitability investigations
Review past assessments of rate change
Comments on other factors that should be taken into account in analysis of
rate movements (claims inflation, propensity to claim, macroeconomic issues)
Comments on Q1(v): Candidates generally provided a relatively comprehensive but
disappointingly generic answer. A number of candidates did provide challenge to the
suitability of the benchmarks and consideration to potential latency issues with EL claims.
Few candidates mentioned the most obvious initial investigation to narrow down the cause of
the deterioration in older years: identifying whether the deterioration is down to new claims,
suggesting latency issues, or to deteriorations on existing claims, suggesting under-reserving
issues. Many focussed almost exclusively on a claims investigation and went into great detail
on this. Clearly this is relevant but is just part of an answer that needs to consider much
wider issues such as reserving adequacy, reinsurance, expense analysis etc.
The majority of candidates suggested raising deductibles to address the small claims issue, in
spite of a previous examiners report that clearly drew candidates attention to the fact that as
EL is a statutory cover there are no allowable deductibles.
(vi)

Issues with the underwriters approach


Taking a single scheme in isolation over a single year, the underwriters
approach is not too damaging
However, any expected loss ratios used in planning / reserving / capital
modelling / pricing for an account will most likely consider the aggregation of
a number of schemes,
with the summary longer term averages also considering the aggregation
across different accident years as well.
Such loss ratios are not expected to apply to every risk within the account, but
are an average of some schemes making higher than average profits while
others make losses.
This effect is particularly marked for low frequency / high severity business,
where average loss ratios are likely to comprise the majority of accounts
outperforming and earning profit commissions while only a few accounts
significantly underperform with large claims / catastrophes.
Failure to allow for these issues not only distorts the summary level estimates
used elsewhere in the business, i.e. pricing, reserving planning.
It also leaves the underwriter unable to fairly compare the relative long term
profitability of different risks.

Comments on Q1(vi): Candidates almost universally missed the point of this question, with
most identifying only that it might lead to a degree of understatement for reserving / pricing /

Page 10

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

planning. Very few candidates recognised that the issue with the proposed approach is that
most expectations, e.g. reserves, would tend to include some risks that outperform and other
risks that underperform. As such the proposed approach is only reasonable for a single
scheme in a single year, and when aggregated totals are considered it would clearly be
flawed. Some candidates failed to answer the question asked and described the problems of
using profit commission rather than the problems with the underwriters not making
allowance for it.
(vii)

Allowing for profit commissions


Methods
Stochastic
Stress testing
Full DFA
Subjective margins
Replicating portfolio / reinsurance arrangement
Volatility issues
Impact of profit commissions is highly dependent on the volatility of the
business
The greater variance there is in loss experience, the greater the proportion of
the account that will outperform sufficiently to trigger profit commissions.
Analysis of past data could provide some indication of the overall volatility of
the account.
Barring accounts with relatively stable and credible data however, volatility
can be significantly harder to estimate accurately than a mean loss ratio might
be.
At the simplest, the standard deviation shown by past loss ratios will give a
very crude indication of volatility.
Unless the account has been running for a number of years however, there will
not be enough data points from the overall account to produce a statistically
reliable estimate
This is an even greater issue for longer-tailed classes where loss ratios for
more recent years will be smoothed estimates due to lack of data
Perhaps reserved using a BF method for example.
Such an approach would not be a fair reflection of the potential volatility once
that year of account has matured.
Care must also be taken when analysing loss ratios from different years of
account where effects of the underwriting cycle can easily distort the
underlying volatility.
Past loss ratios can be adjusted for rate movements which should substantially
reduce this effect
Although loss ratio cycles have historically shown greater amplitude than
rating indices, and the extent of this deviation is extremely difficult to allow
for.
Reserving cycles can also affect the validity of past loss ratios and there are
unlikely to be relevant indices that can be used to adjust for this.

Page 11

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Differences between different risks:


In practice, differing volatility assumptions are likely to be needed for
different schemes as the volatility for each scheme can be influenced by a
number of factors:

The size of the scheme is critical all else being equal, the larger the
scheme is the more stable the experience will be.
For lower frequency business, small schemes could well run largely claim
free over a number of years purely due to good fortune
The level of diversity within the scheme will also have a material effect undiversified schemes could produce a significant risk of accumulations
hitting a particular time period, with other times producing low claim
volumes
The nature of the business covered will also have a material effect, with
some risk types being more weighted towards a volume of stable,
attritional experience than others.
If a risk includes any profit commission sub-pools this needs to be allowed
for as it is the size of each distinct segment which is critical
Any other relevant factors

Grouping of risks
One potential approach might be to group the schemes to produce pools of
data where broadly similar volatility might be expected.
Aggregating the different loss ratios within these pools would produce a more
stable set of data points from which volatility could be estimated.
As profit levels can be significantly different between different schemes
however it may be best to rebase the loss ratios for each scheme to produce the
same averages so that this feature does add to the derived volatility
The grouping of schemes is likely to be a highly subjective process
and may be complicated by some schemes having years of account which
would be expected to show significantly different volatility
perhaps due to start up years where volumes were low as the scheme grew /
other relevant example
Although this method is not perfect, it is relatively easy to implement in
practice and to communicate to underwriters,
and would encourage explicit consideration of the volatility of the account
when underwriting
data requirements to implement this would also be minimal
Volatility model for aggregation
Another method would be to derive a loss model which could be applied to the
specific exposures for each account.
Such a model could estimate the frequency relative to the exposure measure
used along with a severity curve for example

Page 12

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Again some differentiation between risks may be appropriate, with some risk
types having different loss profiles
Risks of each type could be aggregated from a number of schemes for the loss
model to be parameterised.
While this method is potentially more accurate than the alternatives, it does
pose some significant data requirements
Many insurers have not historically captured exposure data at a sufficiently
granular level for analysis of this type
The method would also be harder for the underwriters to manage and maintain
themselves on an individual account basis.
General data requirements on an individual risk basis
Profit commission rate
Threshold at which profit commission becomes payable
Details of any escalation in profit commission with more significant
outperformance
Expected loss ratio for the scheme for comparison with the profit commission
thresholds
Expenses payable to the insurer when making the profit calculation
Policy conditions
Sensible comments about ways of tying in insureds (tie-ins)
Sensible comments about aggregating a number of years in the calculation
(carry overs etc.)
Number of years before any earned commission is released.
Any clawback provisions if experience deteriorates after commission is
released.
Creditworthiness is a potential issue in any clawback
The scope of business which products, distribution channels, territories or
currencies are included/excluded.
Will there be a single pool or several sub-pools?
The duration of the agreement.
When will the profit commission be calculated?
Will the profit commission be paid in a single installment or in several
payments?
Termination terms and profit commission payments after the termination of
the relationship.
Arbitration arrangements.
Profit commission formula.
Sensible comments around appropriate reserves in profit commission formula.
Comments on Q1(vii): This was the most poorly answered question. Candidates almost
universally missed the point of this question, with the majority failing to even recognise that
the key consideration for profit commissions is the volatility of the individual risks. As such,
candidates missed the majority of marks for this question, which considered the extent of the
possible data issues for volatility and how these issues might be resolved in practice.
Disappointingly, the majority of answers to this question were clearly too short for the marks

Page 13

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

available, suggesting that candidates should have known they were missing something.
Few candidates even considered the most basic of practical issues for profit commissions :
the terms on which the commission is set and payable. In spite of the data requirements being
specifically prompted in the question, candidates appeared unable to propose anything other
than the generic data items they would need for e.g. a reserve review.
Overall, this question was highly disappointingly answered for the final paper before
qualification. While this is not a subject that is directly in core reading or a subject that has
been examined before, it is something that candidates should have been capable of forming
sensible answers for based on their understanding of the underlying risks and principles.
Candidates often recognised that one might wish to create a stochastic model or derive
appropriate subjective risk margins. At this level candidates should be able to display higher
order thinking, and anticipate how they would go about e.g. parameterising a stochastic
model what problems they might encounter in practice and how they might resolve them.
A number of candidates went into great detail in terms of the necessary data, however it
appeared they were answering a question about a claims investigation rather than the more
broader points appropriate for a question concerning profit commission.
Some candidates suggested not using profit commission and gave alternatives. This was not
what the question required.
Some students went into great detail about capital allocation methods. These were not
relevant to this question.

(i)

Mutual
An insurer owned by policyholders,
to whom all profits ultimately belong.

Comments on Q2(i): Many candidates only managed half the definition of a Mutual Insurer:
owned by the policyholders but no mention of the profits ultimately belonging to them.
(ii)

Reasons to provide Takaful


Muslims are required to purchase Takaful in preference to conventional
insurance.
5% of the Texel population are Muslims, but there are no Takaful products
currently available in Texel.
There is therefore likely to be a level of demand for Takaful in Texel.
Potentially compulsory so demand may be significant
There may be a first-mover advantage for Vibe if it decides to launch its
product now.
Being seen as innovative/cutting edge may improve Vibes standing in the
insurance market.
Cross selling
Expand to other countries
Possible diversification benefits risk profile may be different
Less drink driving so may be lower risk
Many Islamic financial businesses operate on an essentially mutual basis.
Vibe is a mutual, so may be better suited to this venture than commercial
insurers.
In other countries, new Takaful ventures are increasing on a for-profit basis.

Page 14

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

While Vibe is a mutual, it may still price on a commercial basis. Takaful may
provide an attractive additional revenue stream.
Elements of Takaful may appeal to the non-Muslim population, who may
decide to purchase the products,
e.g. policyholders are entitled to share in any surplus.
a share of any profits would typically be given to charity.
certain unethical investments are forbidden.
Comments on Q2(ii): This question was relatively well answered, although there were a
significant number of points on the schedule available for restating points raised in the
question itself and the additional points were identified more rarely.
(iii)

Difficulties
As the first Takaful business, Vibe should expect the authorisation process to
be considerably more difficult than for a conventional insurance start-up.
It is unclear how regulations that have been written to apply to conventional
insurers should apply to Takaful.
The application of tax and other legal requirements may also be less than
straightforward.
Investigations may indicate that these regulations put Takaful at a
disadvantage to conventional insurers, which would affect the competitiveness
and viability of the business model.
The costs of investigating these issues would be in addition to the usual
expenses of starting a business.
A viable business model requires a large pool of potential customers.
It cannot be assumed that all Muslims will follow religious guidance.
Some Muslims do not agree that the Takaful models used in western countries
are Shariah-compliant.
Some Muslims may be prepared to purchase conventional insurance, if the
available products are cheaper than Vibes offering.
Some Muslims may simply choose to go without non-compulsory insurances,
particularly if they have not previously had cover.
In order to price the business, it would be necessary to consider the risk profile
of the potential market and the extent to which this is reflected in premiums.
This information may be difficult to collect.
A potential niche is considered attractive if the market is currently charging a
high premium relative to the risk.
It would be difficult to attract business if conventional insurance premiums for
the groups being targeted are already competitive.
Vibe may be concerned that a Takaful business is not run solely by the
directors, since it must adhere to the rulings of its Shariah board.
Shariah-compliant reinsurance (retakaful) may not be easily available in
Texel.

Page 15

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Vibe may have to reinsure on a conventional basis or put up additional capital


to ensure that it remains solvent.
Shariah-compliant investments may not be easily available in Texel.
Restricted investment freedoms may reduce profits or have other impacts
There may be difficulties in recruiting the board of Islamic scholars
There may be reputational issues for Vibe
The target market for this product may require significantly different
distribution channels and may have low awareness of the existing Vibe brand.
Separation of funds
Sensible comments around issues resulting from separation of funds
Comments on Q2(iii): Most answers were disappointing here missing many obvious points
such as the potentially greater cost of double compliance and difficulties with sourcing
Takaful compliant investments.
Some students seemed to forget that Vibe was a mutual and started talking about
shareholders or about the need to set up a mutual company.
(iv)

Business Plan
Premium Comments
Significant premium growth is anticipated for the Takaful product in the first
three years.
Minimal growth is anticipated for the existing personal motor business.
The projected 2013 Takaful premium income appears large relative to the nonTakaful premium, given that Vibe is a large insurer specialising in personal
lines.
However, its not possible to be sure from these numbers.
Loss Ratio Comments
Planned loss ratio for existing motor business is 90% in each year.
Planned loss ratio for Takaful is 90% in 2011, reducing to 70% in 2013.
It is not clear how Vibe will achieve a lower claims ratio for the Takaful
business than the existing business.
Vibe would also need to explain the downward trend in loss ratios from 2011
to 2013.
Alternatively, the high profits may suggest that Vibe believes this product will
appeal to a profitable niche of customers.
Specific features of the likely customer base may make takaful more profitable
e.g. reduced likelihood of drinking
Differences in investment income may account for the differences in pricing
The directors would want to be sure that Vibe has sufficient data to be
confident in this assumption.
The relevant data may be difficult to obtain.

Page 16

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Expense Ratio Comments


Planned expense ratio for existing motor business is 10% in each year.
Planned expense ratio for Takaful is 5% in each year.
It is not clear how Vibe will achieve a lower expense ratio for the Takaful
business than the existing business.
The absolute amount of expenses is low in the first year (at $0.3 million).
It appears unlikely that the costings allow for significant spending on launch
and set-up costs.
The level of growth anticipated would likely require a significant amount of
marketing spend, and possibly investment in new infrastructure such as call
centres.
As a consequence of the claims and expense ratios, the gross and net profit
margins for Takaful are much higher than for the existing business.
The net margins for the existing business are nil, and the net margins for
Takaful increase to 25% in 2013 (or other relevant numerical example).
This may suggest that Takaful is being sold at a higher price than conventional
insurance.
Customers may not be prepared to pay a premium for this product, even if
they are obliged to purchase by their religion.
Obtaining higher profits from Takaful may not be perceived as being
consistent with the nature of the product (reputation risk for Vibe)
Obtaining higher profits from some groups of customers rather than others
may not be consistent with Vibes constitution as a mutual.
If the market has the levels of profit shown, Vibe should expect other insurers
to enter the market.
Competition could mean that the level of growth and profits assumed may not
be sustained in to 2013.
The loss ratios as currently displayed for Takaful may be before any
distribution of profits and/or charitable donations and the net results at a
company level may be significantly different after this
As such distribution of profit is contingent on the profits being made,
however, the capital requirements may be comparatively lower on this
business
A full business plan would need to include additional information
e,g. tax, investment income, capital requirements, reinsurance.
These may differ between the Takaful and existing motor business.
Sensible comments about potential alternative ways of interpreting the
numbers.
Based on the information available, the Takaful business plan would not
appear to be plausible unless there are valid explanations not reflected in the
information given.

Page 17

Subject SA3 (General Insurance Specialist Applications) September 2010 Examiners Report

Comments on Q2(iv): Too many candidates missed some very easy marks through either
not calculating the expense and claim ratios for each year or making careless errors when
doing so. Weaker candidates seemingly took the figures at face value and commented just on
their implications rather than questioning their validity too.
A significant number of candidates commented that it would be unrealistic for Vibes Takaful
premium income to increase to 13% in 2013 on the basis that Muslims only make up 5% of
the population and therefore the difference would have to be made up from non-Muslims
seemingly missing the point that the 13% applied to Vibes business and not to the total
population.

END OF EXAMINERS REPORT

Page 18

INSTITUTE AND FACULTY OF ACTUARIES

EXAMINATION
27 April 2011 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes at the start of the examination in which to read the questions.
You are strongly encouraged to use this time for reading only, but notes may be made.
You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 A2011

Institute and Faculty of Actuaries

Champion is a Lloyds managing agent, managing the affairs of the CHUM


Syndicate. A new Chief Financial Officer (CFO) has joined the managing agent and is
reviewing the business written by CHUM and its underwriting agencies under binding
authorities.
CHUM writes business through the slip system as well as through two underwriting
agencies, Agency A and Agency B. The following table has been prepared by the
CFO for the year ending 31 December 2010:
Source of Business

Agency A

Agency B

Premium Income ($ millions)


Loss Ratio
Expense Ratio

40
70%
20%

10
10%
10%

Non-Agency Business
(slip system)
50
50%
15%

Total
100
54%
17%

Having seen these figures, Champions CFO has said CHUM should stop doing
business through Agency A so as to improve the syndicates financial performance.
However, CHUMs senior underwriter believes that Agency A improves the financial
performance of the syndicate.
(i)

Explain why CHUMs senior underwriter may be correct.

[8]

Another underwriting agency, Agency C, is negotiating with CHUM to obtain a


binding authority. Agency C is a long established business selling insurance to large
hotels located throughout Africa.
(ii)

Describe the main features of a binding authority.

(iii)

Outline and explain the investigations CHUM should undertake before


deciding to issue a binding authority to Agency C.

[4]

[25]

Agency C has provided CHUM with some analysis of potential claim costs, obtained
using a recognised natural catastrophe loss model. Agency C writes $20 million of
premium per year.
Scenario
1 in 5 year loss
1 in 40 year loss
1 in 200 year loss
1 in 1,000 year loss

Modelled Claim Costs ($ millions)


2
10
15
25

(iv)

Discuss why the natural catastrophe model may mis-estimate the potential for
large losses from the portfolio.
[8]

(v)

Describe approaches that could be used to obtain alternative estimates of the


loss potential.
[7]
[Total 52]

SA3 A20112

A consulting actuary is advising a regional transport authority in a major city. The


transport authority is introducing a new bicycle hiring scheme. A large pool of
bicycles is to be made available across the city, stationed at a number of speciallyerected docking stations. There are to be two categories of scheme participants: feepaying account holders and casual users. Account holders register their details online
and pay daily, weekly or annual fees. In return, they are issued with a key to access
the bicycles. Casual users may access the bicycles on a pay-as-you-go basis,
requiring a bank card with sufficient funds to cover a deposit as well as the access fee.
The bicycles, docking stations, web service and other equipment and services have
been provided by a number of specialist suppliers.
(i)

Outline the different insurance needs of:


(a)
(b)
(c)

the scheme provider


the specialist suppliers of the scheme
cyclists, as road users taking part in the scheme

[6]

The scheme has considered ways in which these insurance costs can be passed on to
the schemes users. Two alternative ways of charging for the insurance costs within
the hire fees have been proposed differential pricing or a flat-rate fee.
(ii)

Outline the factors and further considerations relevant to:


(a)
(b)

(iii)

a differential pricing framework


a flat-rate fee

[8]

List the advantages and disadvantages of each approach.

[3]

The scheme provider has ruled out differential insurance charges for users. Fees will
however incorporate incentives to encourage users to register online and take up the
weekly or annual option.
(iv)

Outline alternative options for the charging framework and scheme design to
pass on insurance risk to the participants.
[3]

The transport authority has established a captive insurer to take on the insurance risks.
(v)

Explain what a captive insurer is and give reasons why this may have been the
preferred route for insuring the risks of the scheme provider.
[4]

(vi)

Outline the authorisation process which would be required to set up such a


captive.
[4]

(vii)

List the reinsurance options for the captive, setting out the benefits and capital
implications of each option.
[10]

SA3 A20113

PLEASE TURN OVER

The bicycle scheme was introduced on 1 April. It has now been running for six
months. The take-up has been successful and it has been necessary to increase the
number of bicycles by 50% available in the pool. A comprehensive reinsurance
programme was effected when the scheme started with both proportional and nonproportional cover. The results and key information for this business at 30 September
are as follows:
All amounts in 000s
Insurance Liabilities
Net Claims Reserve
700
Net Premium Reserve
400
Total Reserve
1,100
Additional information:

Net written premium


Expenses
Loss Ratio

Actual to
30 Sept
1,000
300
85%

Original
Planned
1st Year
1,450
360
70%

Original
Planned
2nd Year
1,950
480
65%

The Finance Director of the captive has asked for an assessment of the projected yearend profitability of the scheme for the insurance company.
(viii) Outline the factors that would be considered in assessing the profitability and
the sources of uncertainty attaching to this assessment.
[10]
[Total 48]

END OF PAPER

SA3 A20114

INSTITUTE AND FACULTY OF ACTUARIES

EXAMINERS REPORT
April 2011 examinations

Subject SA3 General Insurance


Specialist Applications

Introduction
The attached subject report has been written by the Principal Examiner with the aim of
helping candidates. The questions and comments are based around Core Reading as the
interpretation of the syllabus to which the examiners are working. They have however given
credit for any alternative approach or interpretation which they consider to be reasonable.

T J Birse
Chairman of the Board of Examiners
July 2011

Institute and Faculty of Actuaries

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

(i)

Financial Performance of Agency A


The loss ratio statistics may have been prepared on a different basis.
It is not clear whether the figures are gross or net of reinsurance.
The figures might be discounted or undiscounted.
The other loss ratios (Agency B and Non-Agency Business) might not include
IBNR.
Agency A may tend to include margins in its reported loss ratios (or other
example).
If the loss ratios exclude IBNR, Agency As higher loss ratio could simply
reflect the fact that the business written is significantly shorter tail.
If this is the case, it would most likely be accompanied by lower capital
requirements which would further affect the profit relativities.
The figures might be calculated incorrectly.
The 2010 loss ratios may not be reflective of the underlying, long term
profitability of the business. Taking an average over several years, Agency A
might be more profitable (or equivalently, other business less profitable).
For example, Agency A might have been impacted by a single exceptional
loss during the year.
The rating cycle might also be at a lower point for the business written by
Agency A than it is for business written by other agencies
Agency A might support the financial performance by providing business that
is uncorrelated with the Syndicates other activities.
Agency As business could be more stable than Agency B's, which might be a
catastrophe account.
The expense ratios may not be comparable.
A large proportion of the profits on Agency B might have to be paid out in
profit commission.
Agency A might carry out more of the administration than Agency B, or
require less oversight from Syndicate staff (or other valid example).
Agency A produces 40% of total syndicate premium income. Agency A may
make a significant contribution to covering the syndicates fixed costs.
This $40m premium income may help support the Syndicate's market
standing.

Page 2

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
The greater volumes of business may also help with data volumes/expertise.
Based on the table the business written by Agency A appears to be profitable
(albeit less profitable than the other business).
Total syndicate profits would be lower if Agency A had not been
underwritten.
The business written by Agency A may have lower capital requirements than
the other Syndicate business.
It may in any case serve to reduce capital requirements for other business
(especially if relatively uncorrelated).
It is possible that Agency A might have generated the highest return on
capital.
Even if Agency A generates a lower return on capital than other activities, it
may still be considered a good use of capital if no superior opportunities are
available.
The relationship with Agency A might be important to the syndicates overall
financial result.
For example, CHUM may be allowed to underwrite Agency B if it agrees to
underwrite agency A (or other example).
Agency A may write longer tailed business on which greater investment
income can be earned.
Agency A might have different investment strategy/experience.
Agency A might have growth plans that improve the financial forecasts of the
Syndicate.
Agency A may boost the liquidity of the Syndicate.
Miscellaneous points that could explain differences: reinsurance programme,
accommodation business, tax
(ii)

Main Features of Binding Authorities


Binding authorities are contractual agreements setting out the scope of
delegated authority, allowing cover holders to enter into contracts of insurance
and to issue insurance documents on behalf of Lloyds managing agents.
Underwriting agencies represent a very large source of London Market
business.

Page 3

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
The contract will specify the period for which insurance can be placed
the classes of business covered
and the policy wordings that are to be used.
Many agencies are paid a percentage of premiums as commission.
This causes a potential conflict of interest for the agency because it has an
incentive to increase premiums without sufficient regard for the profitability of
business.
Many of these agencies were formed by brokers.
In some cases, a company may establish a specialist agency to underwrite
risks on behalf of an insurer.
Some agencies have been formed with specialist risk management functions
in-house to write specialist business on behalf of an insurer.
Miscellaneous points: how they are authorised, whether the agency handles
claims and to what extent, reporting of summarised premium and claims
information.
(iii)

Agency C investigations
Investigate Agency Cs historical performance.
Investigate historical premium volumes.
Consider both the total premium written each year, and how business has
varied from period to period.
It may not be worthwhile for CHUM to consider an underwriting agreement if
the amount of business written is too small.
Alternatively, the amount of business written may exceed CHUMs available
capacity.
High growth rates may show that the agency is a successful business, and so a
good agent for CHUM.
However, operational risk issues may be of more concern for a high growth
company than one with a stable portfolio.
Investigate historical claims costs.
Generic points about projecting to ultimate
These should be adjusted for any rating movements over the period
Consider the degree of variation in historical claims results

Page 4

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Consider the potential for worse financial performance than has been seen in
the data historically
Large losses are likely to be a particular feature of this type of business as fire
claims for large hotels could make up a significant proportion of the overall
exposure
PML reports, frequency severity modelling, and discussions with agency staff
could be useful here.
Nature of business / business mix
What classes of business do they/are they going to underwrite?
Consider whether the nature of the risk changed over time.
This could mean that the historical data is not an appropriate indicator of
future claims experience.
Consider the location of insured risks.
This may indicate the potential for large losses due to accumulation of risk.
Consider the number of individual insureds covered by the agency
This will indicate whether premium income and claims experience is
dependent on a few large accounts, or there is a good level of diversification
of risk.
This would impact the volatility of claims and premium income.
Conduct some general market research to understand features of the business
written e.g. by interviewing Agency C:
e.g. consideration should be given to local building standards for the territories
covered as this may significantly impact potential risks
e.g. employment practice/court awards for injuries/sensible alternative point.
Agency C expenses, including any profit contingent commissions.
Ideally the expense structure should better align the interests of agency and
underwriter, on both the downside and the upside.
Review Agencys C future plans
Could new business help defray syndicate overheads?
Agency C growth and strategy
Review Agencys C future plans.

Page 5

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Consider the point in the rating cycle for this type of business.
Areas of focus include forecast growth, profitability and portfolio
composition.
It would be necessary to understand the assumptions underlying any forecast
and how they have been derived.
Agency Cs market share should be considered when analysing any growth
forecast
As should the potential growth in overall market volumes (with Africa being
an emerging market with significant growth potential).
Consider whether the agency plans to significantly change its business model.
Agency reputation
Consider the reputation of the agency in the insurance market.
For example, consider whether the agency is known to CHUM staff.
Consider whether the agency is involved in any legal disputes.
Company background how long established, where are offices located, etc.
Reason for agency leaving current underwriter or being interested in the deal.
Consider the qualifications and experience of Agency C staff.
Consider whether the staff currently involved been responsible for historical
results.
Qualitative factors will play a part in CHUMs overall assessment of
Agency C.
Consider the reputation of the agency amongst hotels.
Underwriting results may be better if the agency has long established
relationships and is a dominant player in the market.
A relatively new agency may need to offer competitive rates to establish client
relationships.
Operational risks
Investigate the agencys risk management and assessment capabilities.
Where these are well developed, this would add comfort that the business is
well run and that risks are properly considered

Page 6

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

However if these are less well developed there is greater scope for enhancing
underwriting quality, further improving any historical performance
Depending on scale and resources, there may also be value to be added by
working with coverholders to improve safety features
Processes and procedures consider whether the operating procedures of the
agency are consistent with CHUMs own policies.
Examples of policies to consider include quality control, safe custody of the
insurers money, issuing contracts on time, settlement of ex gratia claims.
Investigate how Agency C sets premiums.
Consider how CHUM will assess the adequacy of rates offered.
Consider claim settlement procedures.
Consider the claim authorities currently offered to Agency C, for example,
what is Cs authority to settle claims be, and what claims need to be referred to
the underwriter.
Monitoring and reporting issues
Investigate whether Agency C is set up to provide sufficient information to
meet Lloyds reporting requirements.
Extra half mark for mentioning the PMD.
Consider whether the agency will be able to provide the information needed
by the syndicate for monitoring and reporting and in good time.
The required information would include:
Claims data for various purposes, for example, to allow monitoring of loss
settlement, provide reports to reinsurers, monitoring of profitability, year-end
reserving.
Exposure data, for example, to assess aggregations and use in catastrophe
modelling.
Does the syndicate have enough expertise to effectively monitor the
underwriting agent? Extra staff may be required.
Overseas location could cause potential issues with effective monitoring, or at
least add to the resource costs involved
How much management time will be required to oversee the business?

Page 7

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Syndicate strategy
Consider whether the business has a unique distribution model which allows it
to access business not otherwise available to the syndicate.
Consider whether there are other forms of insurance not currently written by
Agency C that could be sold to their existing client base (particularly
insurance written by other agencies of CHUM).
Consider whether underwriting through Agency C is consistent with the
Syndicates business strategy.
Does CHUM want exposure to African hotels?
Does the business provide diversification from or aggregation issues with the
Syndicates other activities?
What currency is the business written in and how does this fit with CHUMs
business model?
Would writing the business be consistent with the syndicates risk appetite?
Does the syndicate have sufficient reinsurance?
Capital issues
Estimates of historical capital requirements, and therefore historical return on
capital.
Historical variation in profitability should be considered in the context of the
return on capital.
Does the syndicate have sufficient capital (regulatory and economic capital
requirements would need to be considered)
If existing capital is insufficient, can capital be raised at an appropriate rate to
finance this venture
Does the business meet syndicate targets for return on capital?
Opportunity cost of capital is this the best use of syndicate resources?
Other issues
Regulatory/legal issues, for example, consider whether Lloyds is authorised
to write in all the countries that Agency C writes in.
Local regulatory risks for example, could the claims environment change
significantly due to developments in local legal systems.
Political risks

Page 8

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
What are the premium payment terms and consequent credit risk?
Other valid points ( mark per point, max of 1)
(iv)

Underestimation by Model
All models are uncertain. The results rely on judgements which may not turn
out to be correct.
In particular, little or no data will be available to guide the most extreme loss
scenarios (1 in 1000).
Underestimation of the tail is very common
both in terms of severity and frequency.
Sensible comment about the numbers, such as there is surprisingly little
difference between the modelled claims costs for a 1 in 40 year loss and a 1 in
200 year loss.
Historical data for Africa is also likely to be of significantly lower quality than
the equivalent data for e.g. the US or Europe
The model may not simulate losses for all the perils that the portfolio is
exposed to.
For example, the model may only cover certain natural catastrophes. Other
natural catastrophes such as bush fire may not be included.
Even where perils are included in the model, assumptions might not be
appropriate for the risks insured by Agency C.
Most modelling expertise is focussed on areas with greatest insured risks
(rather than African hotels).
Mapping software is also of significantly greater granularity for developed
world risks and the CAT model may not accurately reflect the specific risk
profile.
This is a natural catastrophe model and does not cover manmade losses, for
example a large fire could be expensive to cover.
CHUM should check that the results were produced using the latest versions
of the catastrophe models.
Models are updated following cats to allow for new data, and this can change
the model estimates.
Climate change effects may also be significantly altering the underlying CAT
risks
Data issues There may have been parts of the portfolio that could not be
modelled due to missing data.

Page 9

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
The portfolio may have changed significantly since it was modelled.
For example, new or lapsed policies, changes in limits, deductibles or terms
and conditions.
Hotel risks can also demonstrate significant variations in underlying exposure
in short timescales (e.g. following major refurbishment works)
Currency fluctuations may significantly alter the weighting of values between
different countries and CAT zones.
Some problems identified with cat models include:
Unmodelled elements of a modelled loss, e.g. storm surge.
Unmodelled parts of modelled policy, for example, any business interruption
offered to the hotel, marine policies for hotel pleasure craft.
Modelling errors.
Model may not adequately allowed for any frequency trends.
Model may not have allowed for severity issues such as demand surge
following a catastrophe.
(v)

Alternative Estimate of Loss Potential


Interview underwriters and other experts at agency C. Produce a loss scenario
that reflects the judgement of the staff.
Contact the provider of the current proprietary CAT model to understand how
their model works and what data it is based on.
Most importantly, identify whether there are specific loss types that are
excluded from the model so that models for these risk types can be added to
the existing model projections.
Identify whether there are alternative proprietary models that cover these
additional risk types.
Discuss development of bespoke models for these additional risk types with
experts e.g. underwriters, geologists, meteorologists etc.
Review historical loss experience, for example, the worst loss that Agency C
has ever had.
E.g. using Extreme Value Theory.
Consider whether this loss is exceptional, for example, the Agency C may no
longer write risks like the one that produced the loss.

Page 10

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Alternatively, the historical experience may only provide a lower bound on
possible future losses.
Review indicators of possible losses for significant insured risks.
For example, review any surveyors reports and consider PMLs and other
indications of possible losses.
Use this data to perform scenario analysis.
Alternatively a simple simulation model could be produced.
It would be important to consider the location of hotels in the analysis. For
example, if several hotels are part of the same complex they could all be
impacted by a claim.
Repeat the analysis with other proprietary catastrophe models, and consider
the range of results produced.
This comparison should adjust for any differences in loss types covered
Apply a top down approach.
Apply a bottom up approach.
Other third party sources of possible expertise include reinsurer, Lloyds,
brokers, other experts with relevant knowledge
This question was generally reasonably well answered by candidates. Most were able to think
rationally about the numbers presented in part (i) and consider the wider context that might
surround high level numbers of this kind. In part (ii) candidates broadly understood the key
features of binding arrangements, although many appeared to think that the binding
authority was the coverholder rather than the agreement itself.
Part (iii) was well answered by the standards of open ended questions of this manner, with
many candidates displaying the ability to think on their feet and make sensible comments
about potential issues with African business. Overall, this was still one of the lower scoring
sections however, with many candidates simply not generating a sufficiently wide range of
points particularly on strategic points and around wider operational issues with this kind of
deal. For questions of this type, candidates may find it helpful to take a step back to consider
both what companies are looking to accomplish with any business deal and how any deal
might actually be practically implemented. This ability to think in a wider business context is
something we aim to test in SA3, with a view to producing better rounded qualified actuaries
who are capable of adding commercial as well as technical value to their employers.
Part (iv) was again relatively well answered given the open ended nature of the question,
with most candidates identifying the general uncertainty issues with catastrophe modelling as
well as some more specific challenges with African CAT modelling given lower data volumes.
References to the numbers provided were rare however.

Page 11

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Part (v) was less well answered, with many candidates outlining investigations to clarify
attritional losses rather than focusing their attention on large and catastrophe losses.

(i)

(a)

Scheme provider:

(b)

Suppliers:

(c)

Page 12

Theft stolen bikes will need to be replaced


Damage repair and/or replacement of damaged bikes
CAT cover e.g. damage following flood or storm
Docking stations
o Property damage
Public Liability resulting in injury to third parties or damage to
third party property caused by the docking stations
Loss of or damage to keys to cover the cost of replacing/repairing
them
Employers Liability - to cover the employees of the scheme
provider
Insurance against a delayed launch to cover the expenses
associated with the over-run (e.g. launch event, additional
marketing), any claims from potential participants
Compensation for loss of confidential information e.g. payment
details

Product Liability in the event of third party damage or injury


resulting from faulty bikes
Functionality failure of bikes/docking stations/keys
Fleet insurance cover for lorries transporting the bikes between
locations
Commercial property insurance
Goods in transit cover
Delay or failure of registration website
or overcharging errors and related administrative and compensation
costs.
Employers Liability - to cover the employees of the suppliers
Fraud insurance? If key system is hacked and bikes can be used for
free could be a number of potential costs
Delayed delivery of bikes/docking stations/keys

Scheme participants:
Theft to cover the cost of any payment required towards the cost
of replacement of the bike or key
Damage to cover the cost of any payment required towards the
cost of repair/replacement of the bike or key
Public liability/third party liability to cover against third-party
claims for injury to other persons or damage to other peoples
property following an accident caused by the participant

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

(ii)

Personal accident covering the cyclist for a fixed amount in the


event of death or disablement to themselves.
Breakdown cover may also incorporate cover for any delay in
returning the bike back to the docking stations and compensation
for any fees incurred
Compensation for charges e.g. if unable to find a free docking
station, if liable for faulty docking station etc.

In both cases the insurance cost will be determined by reference to the


estimated number of participants and anticipated frequency/duration of usage.
To identify users and track their experience, good information systems are
required and users need to be identified through online details (for registered
users) or payment details (for casual users).
(a)

Differential pricing key considerations


Substantial data requirements covering exposure, claims and other
costs, including:

age/sex
cycling experience/proficiency
intended use: commute/pleasure/peak-time/route
casual vs registered users
seasonal use
use of helmet
holds UK drivers license
number of bikes on risk
duration of each journey
distance travelled

Claims history of the cyclist would be important.


NCD could be an incentive to register.
As this is a new scheme and new insurer, there will be no data
...beyond any data obtained from advance registrants or market
research
Market data from other cities or countries will be used
This may not be directly comparable
Consultants may provide some assistance
It may be worth starting to collect data online while applying flat fee
approach ready for differential pricing later.

Page 13

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Inadequate data may lead to deficient rates, resulting in:

Underwriting losses
Poor take-up, leading to unsustainable running costs
Attracting undesirable risks, resulting in poor experience and costs
Insufficient funds to meet running costs
Scheme may need to be subsidised by other means

To offset data limitations, conservative assumptions or additional


loadings could be applied.
But care needs to be taken to ensure that the charges are not so
excessive as to impact take-up.
To analyse and collate the information, the scheme will need to have in
place appropriate systems.
This will add to the administration and cost of the scheme.
Consideration will need to be given to casual users who effectively
need to be underwritten at the docking station. Practicality of seeking
information at point of access?
Refusal of significant number of potential casual users at point of
access may lead to bad press.
(b)

Flat-rate fee additional considerations


Above data requirements are relevant plus the additional assessment of
the appropriate mix of each factor in order to determine an appropriate
average or flat rate.
Assumption of number of participants will be critical
Flat fee could sit alongside a schedule of post-claim penalties for
damage to bike, for example
The cost can be split between the upfront access fee the hourly hire
cost.
In both cases the insurance costs of the scheme provider will need to
be factored in.

(iii)

Differential

Page 14

More equitable
Avoids subsidising of poorer risks
Reduces the risk of anti-selection
More responsive to underlying risks
More complicated to administer

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

Not viable for the casual users


Data not immediately available

Flat rate

(iv)

Simple and cheaper to administer


More straightforward and easier to market
Risk of greater proportion of poorer risks
Can be enhanced with explicit schedule of post-claim charges

A key aim would be to ensure that participants are responsible for a portion of
the cost in the event of loss/damage to minimise moral hazard.
A fee could be charged before the key is issued, and for additional keys issued.
In the event of loss or damage to the bike/key, this fee when combined with
the net insurance proceeds would cover the cost of repairing or replacing the
bike/key.
Casual users (i.e. those that are not registered on-line) could be required to
provide card details before hiring bikes. The amount of the charges for
loss/theft/damage must be available on the payment card, and deductible if the
bike is not returned in good condition
Alternatively casual users could be asked to put a full deposit up, with scope
to have this being more substantial than the flat fees charged for insurance.
In the event of a public liability claim, the insurance taken out by the regional
authority would pay out, but participants could be required to pay the first part
of any such claim, so minimising moral hazard risk.
Participants could be asked to sign terms and conditions requiring them to
observe basic road and bike safety principles e.g. wearing a helmet, observing
traffic regulations
. . . . and in the event that they didnt any personal accident element would be
invalid.
The scheme provider could reserve the right to decline an individual request to
hire due to previous claims/poor track record in maintaining the property
Opt in/opt out choice / sign a waiver.

(v)

An insurer wholly owned by an industrial or commercial enterprise and set up


with the primary purpose of insuring the parent or associated group companies
and retaining premiums and risk within the enterprise.
Reasons for establishing:

To fill gaps in insurance cover that may not be available from the
traditional insurance market.
Page 15

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

(vi)

To allow the transport authority to better manage its total insurance spend.
To enable the enterprise to buy cover directly from the reinsurance market.
To focus effort on risk management.
To gain tax and other legislative or regulatory advantages.
Although tends not to be the primary reason anymore.
In addition to accepting the risks of their parent companies, captives may
also accept external risks on a commercial basis e.g. insuring the third
party suppliers.
Although regulatory benefits would not be possible in this case.

The captive would need to follow the local regulators (e.g. the FSAs)
authorisation process.
This could involve a process similar to the FSA's:

(vii)

Submitting a regulatory business plan outlining the business which the


captive intends to underwrite.
Scope of permission required, i.e. the regulated activities which the captive
will carry on, the client types etc.
The financial resources of the captive, including financial projections over
at least a 3 year period, the type of assets to be held, and the amount of
capital the captive believes is appropriate given the regulatory business
plan proposed.
List of personnel in terms of controlled functions.
Compliance arrangements
Estimate of FSA fees and other levies
Other reasonable points e.g. demonstration that captive at arms length (half
each to max of 1)

Quota Share
May be desirable to spread risk for this new class of business
In return for expertise from reinsurer
Allows writing of more risks for the same capital
can retain a smaller share of a larger number of risks
useful if expecting the portfolio to grow significantly
and hence increase the diversification within its portfolio, reducing the
volatility of the overall result
this would reduce the capital requirement of the captive/improve solvency
ratio
most useful for more predictable risks e.g. damage/theft
administratively simple
commission may help with cash flow
Surplus is only relevant for non-liability non-cycle hire risks

Page 16

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
Excess of loss
Caps exposure to larger risks
Risk xl may be useful for specific larger risks e.g. related to launch date
delivery or website failure
Aggregate xl may be appropriate to cover a collection of bodily injury
claims e.g. if thefts/accidents spiral to unprecedented levels
Cat xl provides cover in the event of flood or windstorm
Stop Loss
Will protect against loss ratio higher than expected
But unlikely to be available at reasonable price
Financial reinsurance may improve liquidity/cashflow and apparent solvency
position.
There would be capital implications regarding credit risk with any reinsurance.
No reinsurance is an option all profit is kept but will require more capital.
(viii)

Figures shown indicate the business is loss making


However reserve figures seem inconsistent with LRs as UEP is 600 and claims
reserve is 700.
We need more information on what these figures represent e.g. definition of
net.
COR of 115% (85% LR plus 30% Expense Ratio)
This compares with planned COR of 95% after 12 months and 90% after 2
years
The year-end position will reflect the development of premiums, expenses and
claims
[assume limited impact of investment income]
Premium and uncertainties
The major increase in take-up has resulted in significant increase in
insurance premium income for the captive
Nearly 70% of the full year premium income has been achieved in the first
6 months
YE profit will be impacted by NEP for the year
Earning of the premium will depend on the underlying mix of exposures
(annual fee payers vs casual users for example)
Assumption needed for any further growth in premium in the last quarter
o discuss plans and forecasts with the regional authority
o seasonality considerations; take-up may fall away as weather declines
o any increases anticipated to scheme charges may directly or indirectly
captive premium income
o sensitive to any negative press, infrastructure developments

Page 17

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011

Net premium subject to any changes in reinsurance e.g. increase in nonproportional coverage, if there are an increased number of larger risks
presenting

Expenses and uncertainties


Expense ratio is running at 30% compared to the year 1 plan of 25%
This is despite the increase in premium volume
May indicate
o higher than expected set-up costs
o or set up costs biasing the expense ratio for the first 6 months
o re-negotiated commission rates
o inexperienced claims staff
o more complicated claims than expected
May take some time for expenses to stabilise/reduce.
Claims and uncertainties
Experience of first 6 months has been worse than planned: 85% COR cf
70%
Claims experience is likely to be seasonal.
Profile of participants may be different to expected:
o Age
o Gender
o Cycling proficiency
o Casual vs registered
More BI claims than anticipated
Road infrastructure unable to cope with influx of cyclists leading to greater
hazard than anticipated
May reflect larger claims in relation to website failure or launch failure,
equipment failure (although would expect to receive recoveries from
suppliers)
Theft/accidents greater than anticipated
Major cat event/weather event
For YE projection, will need to anticipate the extent to which the poor
claims experience is likely to continue or can be considered to be one-off
Also need to assess the possibility of the claims already incurred
deteriorating e.g. poor case estimation, adverse court settlements, adverse
claims inflation
Generic point about investments.
Generic point about tax.
Performance on this question was mixed, although overall candidates tended to provide
reasonable answers. Most candidates produced a reasonable range of points, although as
with Q1(iii) candidates would benefit from taking a step back to really envision how the
business would work; for example almost no candidates considered fleet insurance for the
transport of bikes between stations (something that London candidates at least should have
thought of).
Parts (ii) and (iii) were reasonably well answered, although many candidates offered
relatively generic points with limited effort to tailor it to the question specifics. In part (iv)

Page 18

Subject SA3 (General Insurance Specialist Applications) Examiners Report, April 2011
many candidates spent time discussing forms of differential charging such as linking
premiums to past experience, in spite of this being specifically excluded.
Questions (v) and (vi) were standard bookwork and were answered well. In part (vii), a
number of candidates went into detail on the nature of the reinsurance contracts written. As
the question only required candidates to list options before detailing the specific benefits and
implications in the situation posed, this additional detail gained no marks. Part (viii) was one
of the lowest scoring parts, although this appeared to be more due to an abnormal number of
candidates simply running out of time. Few candidates made any specific comments on the
numbers provided either.

END OF EXAMINERS REPORT

Page 19

INSTITUTE AND FACULTY OF ACTUARIES

EXAMINATION
5 October 2011 (pm)

Subject SA3 General Insurance


Specialist Applications
Time allowed: Three hours
INSTRUCTIONS TO THE CANDIDATE
1.

Enter all the candidate and examination details as requested on the front of your answer
booklet.

2.

You have 15 minutes before the start of the examination in which to read the
questions. You are strongly encouraged to use this time for reading only, but notes
may be made. You then have three hours to complete the paper.

3.

You must not start writing your answers in the booklet until instructed to do so by the
supervisor.

4.

Mark allocations are shown in brackets.

5.

Attempt both questions, beginning your answer to each question on a separate sheet.

6.

Candidates should show calculations where this is appropriate.

AT THE END OF THE EXAMINATION


Hand in BOTH your answer booklet, with any additional sheets firmly attached, and this
question paper.
In addition to this paper you should have available the 2002 edition of the Formulae
and Tables and your own electronic calculator from the approved list.

SA3 S2011

Institute and Faculty of Actuaries

Sunland is a small, prosperous country with a population of around 2 million people.


It is mandatory for employers in Sunland to purchase employers liability insurance
from SunInsure, a state-owned not-for-profit monopoly insurer.
The premiums charged by SunInsure vary depending on the insurers assessment of
the riskiness of the industry in which the employer operates. However, all employers
in the same industry pay the same premium as a percentage of workers wages.
SunInsure sets premiums to cover only its expected payments in the following year
(referred to as a pay as you go or PAYG basis). The PAYG premiums also include
an adjustment to allow for any under or over collection of premium in the previous
year.
Injured employees may be entitled to compensation payments many years after an
accident has occurred. SunInsure does not hold assets to cover the future cost of such
claims, but will instead collect premiums each year from its current customer base to
cover any emerging required payments.
A number of private insurers operate in Sunland, and provide all insurances other than
employers liability. The regulations that apply to private insurers are different to
those that apply to SunInsure.
Insurance regulations require private insurers to hold assets to cover all the future
costs for claims that have occurred, even if payments are not expected to be made for
many years. Premiums estimated to cover all future claim costs are referred to as
being on a premium funded basis. There are also capital requirements for private
insurers.
The Minister for Work has issued the following press release:

The good people of Sunland have had enough of SunInsure. I believe that SunInsure
has become inefficient because it has no competition, and so employers are paying
higher premiums than are necessary.
I am proposing that private insurance companies be allowed to enter the employers
liability insurance market. Employers will be able to choose whether to purchase
insurance from a private insurer or from SunInsure, which will continue to operate as
it does now.
Private insurers will be free to decide what premiums to charge. Private insurers can
vary the premium charged to each employer based on an individual risk assessment.
Since private insurers set premiums on a funded basis, the new arrangements will
provide a high level of security to employers and to injured employees.
I believe that this scheme will reduce employers liability costs for the businesses of
SunLand.

SA3 S20112

(i)

State the costs that will need to be covered by employers liability premiums
charged by:
(a)
(b)

SunInsure (PAYG basis)


private insurers (premium funded basis)
[4]

The Minister said I believe that this scheme will reduce employers liability costs for
the businesses of SunLand.
(ii)

Discuss whether the Ministers statement is likely to be correct.

[10]

The SunLand Employers Association (SEA) has provided a response to the


Ministers press release. SEA is interested in the proposal, but believes that many
practical issues would have to be considered in assessing whether the plan to
introduce private competition will work in practice.
(iii)

Describe the practical issues that will need to be considered to assess whether
the plan will work in practice. Your answer should include reasons why the
issues are important.
[20]
[Total 34]

SA3 S20113

PLEASE TURN OVER

Manufast is a large industrial manufacturing business based in the UK. It has


expanded rapidly over the past 15 years, tripling its output through improvements and
innovation in the key processes, whilst also broadening its product range. During this
time staffing levels have increased only marginally so the majority of the expansion is
due to processes and products.
An actuary has been engaged by the company to review its insurance requirements..
Although processes and products have been continually changing the company has
maintained a very low claims record and is questioning whether it may be paying too
much for its insurance.
The company currently purchases the following covers:

fleet insurance for its vehicles


employers liability
commercial fire
business interruption
goods in transit, and
public liability (excluding product liability)

All of these are priced using standard book rates with the exception of the fleet
insurance which is experience rated.
(i)

Discuss how the actuary might assess the expected level of profit ceded on
each of the lines of insurance purchased and what data might be used to
support these assessments.
[19]

The companys parent is considering setting up a captive insurance subsidiary in the


UK to cover the insurance requirements for its entities.
(ii)

Define the term captive.

[2]

(iii)

Explain the advantages and disadvantages to the group of setting up this


captive, taking into account the current changes in the regulatory climate. [10]

(iv)

Discuss the factors that should be taken into account when advising on the
capital requirements for this new insurance subsidiary
[10]

(v)

Discuss the administrative issues that should be taken into account in setting
up a new insurance subsidiary.
[3]

In light of the various changes in its business, Manufast is also considering adding
product liability to its insurance coverage. It has asked the actuary to advise on this as
part of the considerations of setting up the captive.
(vi)

Describe the general features of product liability cover

(vii)

Discuss product liability pricing issues

(viii) List issues that may be encountered with product liability claims.

END OF PAPER
SA3 S20114

[7]
[11]
[4]
[Total 66]

INSTITUTE AND FACULTY OF ACTUARIES

EXAMINERS REPORT
September 2011 examinations

Subject SA3 General Insurance


Specialist Applications
Purpose of Examiners Reports
The Examiners Report is written by the Principal Examiner with the aim of helping
candidates, both those who are sitting the examination for the first time and who are using
past papers as a revision aid, and also those who have previously failed the subject. The
Examiners are charged by Council with examining the published syllabus. Although
Examiners have access to the Core Reading, which is designed to interpret the syllabus, the
Examiners are not required to examine the content of Core Reading. Notwithstanding that,
the questions set, and the following comments, will generally be based on Core Reading.
For numerical questions the Examiners preferred approach to the solution is reproduced in
this report. Other valid approaches are always given appropriate credit; where there is a
commonly used alternative approach, this is also noted in the report. For essay-style
questions, and particularly the open-ended questions in the later subjects, this report contains
all the points for which the Examiners awarded marks. This is much more than a model
solution it would be impossible to write down all the points in the report in the time allowed
for the question.

T J Birse
Chairman of the Board of Examiners
December 2011

Institute and Faculty of Actuaries

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

General comments on Subject SA3


Consistent with previous examiners reports, the examiners would offer candidates two key
pieces of advice read the question properly and take the time to think about what is going
on.
On the first issue, candidates should always work on the assumption that the question
wording has been carefully chosen. It is therefore essential to read the question accurately.
If something is not asked for then candidates will waste valuable time writing answers that
will gain no marks. These broader answers may be a logical next step to the question and so
may be appropriate for candidates to discuss in a professional context. However in an exam
with a finite number of marks available the scope must necessarily be limited and specifically
defined.
If a question specifically mentions something, candidates should also assume that there are
definitely marks available for this aspect of the question. During the exam setting process,
any content that is superfluous should have been removed.
Wording of question parts should also be considered in the context of the position within the
overall question. Where new question information is provided between parts, candidates
should recognise that this information is specifically relevant to the following part or parts.
Various examples from this paper of recurrent failure to read the question are below.
On the second issue, candidates should note that SA3 is the key paper at which candidates
broader thinking is tested. This is generally the final paper before qualifying as a
professional, and a capacity for broader thinking is one of the best indicators of a candidates
suitability to act in a professional capacity once qualified.
As such exam papers are designed so that it is difficult to pass without displaying some
capacity for independent and broad thinking, as well as to reward instances where these skills
are displayed. When reviewing past papers, candidates should assume that the marks
available for generic points are substantially less than those awarded for the more challenging
points that would be the mark of high quality professional insight in a practising actuary.
Even among passing candidates, this capacity for broader thinking is not always in evidence.
The examiners strongly recommend that candidates step back and take the time thoroughly to
think about what is actually going on in the question situations posed rather than simply
considering numbers to be analysed with standard techniques. For example, candidates might
stop to think in detail about claims for a particular class of business, considering factors such
as what causes the claim, who brings the claim, how it is dealt with once brought, what
makes one claim small while another is substantial, etc.
This more grounded, real world perspective will help candidates to consider such things as
practical issues, stakeholders involved and their potentially diverging objectives, wider
impacts, regulatory or ethical issues, inappropriateness of certain actuarial techniques for the
specific situation, current economic or cyclical effects, etc. This is likely to lead to generation
of broader points (and indeed reflects the thought processes of the examiners in drafting the
questions and solutions) and a more rounded understanding of the underlying risks and

Page 2

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

dynamics which should also be of value to candidates when dealing with different
stakeholders in their professional life.
Again, some examples of this failure to think more widely on the current paper are below.
As in every subject, there is a need to employ basic exam techniques such as well structured
answers and effective time management.
Comments on the September 2011 paper
Overall performance on this paper was reasonable, as reflected in the above average pass rate
for this session. Performance on each of the questions was similar, with the more unusual
question 1 appearing to have sufficient depth of content in the question to allow competent
candidates to respond to an unusual situation.
On question 1 however, some candidates did not seem to understand fully the nature of
funded / PAYG cover in spite of the detail given in the question. Even where candidates did
understand, only the better candidates displayed an ability to think through the full range of
implications of different funding bases to the various stakeholders, although most candidates
were good on more practical and generic issues.
Candidates are advised to take the time to consider scenarios presented in SA3 papers and
think about what is happening throughout the full process and for all stakeholders. This
consideration should lead to generation of the wider points necessary to demonstrate higher
order skills and obtain a high mark in this paper.
In question 1 part (ii), some candidates discussed the entire press release rather than the
immediate statement before this section of the question. In question 1 part (iii), a number of
candidates discussed solutions where the question only asked for details of the issues and
their importance.
In question 2 part (i), many candidates did not correctly identify that the question considered
expected rather than past profits and so missed most points available for issues relating to
changes in risk. In question 2 part (iii) a number of candidates failed to mention Solvency II
in spite of this being very clearly indicated within the question.
More generally on question 2, the better candidates structured their answers well and
considered issues relating to each business line in turn rather than providing generic answers
across all lines. This generated more points and tied their answers more closely to question
specifics, which is highly advisable exam technique for all papers, particularly SA3.

Page 3

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

(i)

Premium Components
(a)

SunInsure Costs
Premium needs to cover:
Expected claim payments during the coming year (regardless of when
the accident occurred)
SunInsures expenses during the year
Costs can be net of investment income earned during the year

(b)

Private Insurer Costs


Premium needs to cover:
Expected claim payments for accidents occurring during the year,
whether or not those amounts will actually be paid during the year
Claims handling expenses, including a provision for expenses payable
in future years
Commissions and brokerage
Net cost of reinsurance
Profit margin (possibly including an allowance for contingencies)
Costs can be net of investment income earned on reserves

Candidates would have benefited from a clearer articulation of claims costs and timings
under each of the two systems. Many candidates failed to discriminate between costs PAID
for SunInsure and costs INCURRED for private insurers.
(ii)

Comparison of private and SunInsure premiums


PAYG and funded represent very different bases of setting premiums.
The two bases cannot be directly compared.
Differences in premium in the current year reflect timing differences in
funding claims costs, rather than indicating that one sector is more efficient
than the other.
Example of how PAYG and funded premiums might compare. For example,
in a fast growing population, PAYG premiums would be less than funded
premiums.

Page 4

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

It is unclear whether PAYG premiums would be lower than funded premiums


at the present time.
There may be differences in efficiency of private sector insurers and
SunInsure.
Private insurers may have less leverage in purchasing services from providers
than SunInsure, due to loss of scale. This could result in higher claim costs for
private insurers.
It is plausible that the motivation of higher profits could lead to greater
efficiency in terms of claim costs and claim handling expenses under a private
sector than a public sector model.
The number of accidents that occur at work is a key driver of claim costs. It is
not clear privatising the workers compensation insurance market will have
any impact on this.
Private insurers may however be able to provide more innovative products
which incentivise lower claims, for example, offering no claim discounts.
SunInsure charges the same premium to every employer in an industry, so
there would appear to be limited reward for good claims experience under the
current system.
There would also be no recognition of other distinguishing risk features
For example some companies may have greater automation and their wageroll
may be more skewed towards high value added development work rather than
higher risk direct manual work.
If Industry groupings are relatively broad, there may also be significant
differences in the underlying work carried out between different businesses in
the same industry.
Conversely, the move to a private sector approach would mean each insurer
having significantly lower data volumes and may reduce the accuracy of the
overall market premiums.
Overall, the scope for savings under a private sector model would depend on
the current level of efficiency of SunInsure, which is unclear.
The private and SunInsure models have different expense bases.
Under the private system there would be:
Duplication of overheads with multiple operators for example,
multiple IT systems.

Page 5

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Commissions and brokerage, which are not necessary where there is a


monopoly provider.
Marketing expenses for private insurers to attract customers.
Private insurers will need to charge a profit margin to achieve a return on
capital, whereas SunInsure is a not-for-profit business.
If private insurers purchase reinsurance, there will be profit margins in the
reinsurers premiums.
In terms of expenses and profit margins, a public sector monopoly represents a
lower cost model than private competition.
Overall, it is impossible to say whether the Ministers statement is correct
given the information presented.
Some limited credit available for additional valid points not on the marking
schedule.
Many candidates failed to offer a balanced assessment of how costs might increase or
decrease in both the short and long term, with PAYG systems often having hugely distorted
short term cashflows.
(iii)

Scheme viability
Residual Funding Issues:
Need to consider how claims for prior accident years will be paid.
SunInsure does not have any assets to pay these claims, as it collected
premiums on a PAYG basis.
Under the PAYG model, only employers that remain with SunInsure will need
to pay for these claims.
The greater the number of employers leaving SunInsure, the higher the
premiums needed for remaining policyholders.
These high premiums will encourage even more employers to switch to
private insurers, leading to further price increases and lapses.
If SunInsure loses all its policyholders it will be unable to pay claims, meaning
workers will not receive compensation.
Government may be called upon to pay for these liabilities.
Possibly a separate levy could be made on all employers to cover historical
liabilities.

Page 6

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Funding Future Accident Years:


The proposal might also be more viable if SunInsure moved to a funded basis
of premium calculation.
SunInsure premiums may be much lower than private insurer premiums, or
vice versa.
This could be because of the difference between PAYG and funded premiums.
Alternatively, this could be because of other structural differences,
for example, the required profit loadings
Or other relevant example of efficiency
Consequences of differences in premiums:
Few employers would choose a private insurer if SunInsure was cheaper.
In this case, SunInsure would remain a monopoly as private insurers would be
unable to compete.
Having insurers with funding policies (PAYG and FF) competing is unlikely
to work in practice.
Under the Ministers proposal, SunInsure will continue charging average
premium to each industry, whereas private insurers can risk rate individual
employees.
This will allow private insurers to cherry pick the employers with better than
average claims experience.
SunInsure will be left with poorer than average risks, and need to increase
premiums commensurately.
Increases in premiums means the private insurers appear more competitive,
leading to the loss of more business for SunInsure.
Insurer Failure:
Need to consider what will happen if a private insurer fails.
The Minister should consider the compensation arrangements that are
appropriate in this case.
Suninsure has implicit government guarantee as a public insurer, so may be
regarded as a safe option for employers.

Page 7

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Data / Competition Issues:


Consider what information will be made available to private insurers to
determine premium rates.
In particular, consider whether SunInsure will be required to share its
historical data.
If not required to share data, SunInsure may have an unfair advantage.
Data may not be available for risk rating individual employers, as SunInsure
has not previously priced at this level.
Private insurers may include large contingencies in the premiums if good
pricing data is not available.
Sunland is a small country will enough insurers be interested in writing
employers liability for the market to be viable?
Consider whether private insurers have the necessary pricing skills.
Specialist claim handling and other staff may also not be available.
Consider whether private insurers have sufficient capital available.
Will reinsurance be available for the insurers?
If there are too few competitors, profits may be excessive, or there may be no
incentive for efficiency.
Need to ensure that all employers will be able to get insurance, so that
compensation will be available to workers.
For example, will private insurers be obliged to provide cover to any employer
requesting it.
Consider whether there should be maximum premiums set by government.
Consider the level of regulation needed.
Will all private insurers be allowed to write employers liability insurance, or
will there be an insurer accreditation scheme.
Consider whether existing private insurer capital requirements are appropriate
for writing employers liability business.
What is the minimum level of cover that private insurers need to provide.

Page 8

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

May be a poor outcome for workers if employers may purchase inferior cover.
Need to make sure private insurers are incentivised to return workers to work
(optimal for workers and society) rather than just minimise cost.
SunInsure will need to be regulated too.
Private insurers will need to understand the level of political
interference/oversight in the market.
The best candidates recognised that the key practical issue was the potential gap in coverage
caused by switching from SunInsure to a private insurer. If SunInsure was to cover incurred
claims this would have a knock-on effect as the costs would have to be met by the remaining
policyholders only which in turn would exacerbate their funding problems.
Most candidates covered the more practical issues such as lack of data, Regulation and
systems and personnel requirements although not usually in sufficient depth to score well.

(i)

The experience of the 6 classes of insurance will have seen some changes in
claims experience over the last 15 years due to the companies expansion,
especially those related to processing, which has driven a large proportion of
the expansion, i.e.
EL, Business Interruption and PL
Some of the other classes will not have been impacted as much.
The changes in claims experience will have to be analysed, and if there have
been nil claims for any classes this will prove difficult.
Also, where classes would be expected to be exposed to spike or catastrophe
losses volumes of these claims may be low.
Claims should be analysed in terms of claims frequency, severity, causes, and
trends.
In each class they would need to consider relevant rating factors and exposure
measure.
Fleet Insurance
Exposure measure in this case will be vehicle years.
The experience over the past 15 years should be a good guide allowing for the
marginal increase in staffing levels, and therefore marginal increase in
vehicles being used. Although it should be fairly easy to get this accurate data
from the company itself.

Page 9

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Even if number of vehicles is similar the average mileage per vehicle would
need to be checked, especially if changes in processes have meant changes in
types of vehicles used and distances of distribution.
The more granular the claims analysis the better.
This can be achieved by splitting by:

Vehicle type
Claim type (bodily damage/property damage/ TPL)
Size of claim (attritional/large/catastrophe)

(the benefits of such analysis will be being able to identify the differing levels
of inflation required especially between PD and TPL, and between attritional
and large claims)
Employers Liability Insurance
The exposure measure to use here would be payroll, which would be easily
available.
Although they would need to allow carefully for the amount of salary inflation
over the last 15 years as this will significantly increase the likely claims sizes.
Again the experience over the past 15 years may be a useful starting point for
looking at attritional type claims,
as long as the marginal change in staffing has been even across all employee
types,
and that there has been no significant changes in working conditions, and
health and safety.
Rating factors would need to be considered in terms of any changes to
processes impacting on types of industrial diseases which may arise, e.g. have
innovations increased the likelihood of industrial deafness, or chronic
obstructive pulmonary disease for example?
It may be difficult to say much about the likely impact of such latent claims,
given you only have 15 years data.
But could benchmark against industry data especially for companies using
similar material & processes.
They will need to allow for any large /abnormal claims in the data, or lack
thereof.
They changes in processes could have also resulted in a change in the mix of
staff and hence payroll, so an analysis of payroll by skill types would help
identify change in exposures.

Page 10

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Commercial Fire
Exposure measure sum insured
Manufast has tripled its output in 15 years, so they will need to check whether
this has been achieved with increases in manufacturing sites even if the
staffing levels have only marginally increased.
This will dictate how different the exposure change would be.
The rating factors would need to be considered in terms of:

Construction types of buildings


Age of buildings
Health and safety procedures e.g. fire doors in place

Will need to consider whether the innovation in production processes have


changed the risk profiles of each building.
This could include:

Changes in value of machinery


Changes in types of material used
Any other appropriate examples

As in other lines, the analysis would need to consider any large single fire
claims in the past data. If these do not exist in the data, they could look at
market statistics for similar types of operations to see if the premium loadings
are appropriate for expected large losses.
Business Interruption
Exposure measure would most likely be turnover or profit.
This could have been volatile in the past 15 years.
The fact that the company has expanded 3 fold in the last 15 years will need to
be factored into the analysis as this will result in larger claims from such
cover.
This type of loss will only occur after a major loss, and as such the firm may
not have any in their claims experience to date,
Although this is likely have been allowed for in their premiums.
Given its expansion it is likely that the company have changed their BCP
plans over the last 15 years and as such the premiums for such cover may
appear inflated.

Page 11

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

The fact that the firm have improved processes may mean that the new
processes are more difficult to replace/restart etc., and this will need to be
allowed for.
Have they specialised e.g. do they do different bits of a process at different
sites with dependencies between sites so that fire in one site prevents others
from operating effectively.
How many different sites do they have & safety processes (overlap with
commercial fire)
Are stocks held separately
How specialised is the machinery can it be replaced relatively quickly or
does it need to be built from the ground up
What stock levels do they operate with / how long could they continue to trade
for using stock products
How specialised are the products i.e. can customers just switch to someone
else leading to long term consequential loss of business
What coverage terms are in place indemnity periods, retention of part of the
claims costs
Trade type often used as a rating factor
Contract terms with suppliers / break clauses / penalties etc.
Goods in transit
Exposure measure would be sum assured for shipping, number of journeys or
overall turnover.
Would need to consider relevant rating factors:

Method of transport
Distance travelled
Frequency of transit
Internal/external shipping
Retention of an interest in the goods shipped
Choice of packing method
Use of in-house / contract delivery teams
Recovery potential from third party shippers / transfer of liability
Salvage potential for goods damaged in transit

The previous claims experience should be a good guide as long as rating


factors have not changed significantly.

Page 12

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

But this would depend on the nature of the expansion and process changes
which have taken place.
Public Liability
Exposure measure would be turnover.
Would need to consider relevant rating factors:

Number of premises
Number of external visitors
Environmental exposure
Excesses
Deductibles
Policy limits
Territories covered (based in UK but potentially has sites outside?)

The experience now is likely to be very different from that 15 years ago, due
to potentially increased floor space, processed used, number of external
contractors etc. As such these will impact on claims experience likely in the
future. So analysing trends in claims experience relative to square footage/
number of contractors as well as turnover will be important.
A number of candidates suggested looking at insurer FSA returns to identify expected profit.
It is unclear how this would be of use unless this was the insurers only customer.
(ii)

A captive is an insurer wholly owned by an industrial or commercial


enterprise and set up with the primary purpose of insuring the parent or
associated group companies.
Thus retaining premiums and associated risks within the group, as such this is
a form of self insurance.
Some captives are set up with the primary purpose of selling insurance to the
customers of the parent;
Or they may insure other non-group companies if they have the expertise.

(iii)

Advantages:

Retains insurance profit within the group, which could potentially improve
profitability for the consolidated group.

Could increase awareness of risks within the company and as such


improve risk management practices

Captive would be subject to Solvency II and as such should be deemed


equivalent to many reinsurance companies regulatory requirements
around the world.

Page 13

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

Reserves and premiums are built up as pre-tax profits

Could also sell products to customers of the company, hence improving


profits further.

Gives them direct access to the reinsurance markets with all the associated
expertise.

Could select against the insurance market during harder markets, by


increasing retention in the captive.

May allow access to wider range of insurance products.

Although Manufasts parent is considering setting captive up in UK, if


they are flexible in terms of where the captive is domiciled they may be
able to take advantage of certain tax benefits, e.g. in Bermuda

Reduces insurer credit risk

Disadvantages:

Could increase volatility in groups results depending on the amount of


reinsurance purchased by the captive.

There may be a large accumulation of risk, depending on which other


group entities the captive is insuring.

Complex to set up

Costly to set up and subsequently run


o
o
o
o

Lack of internal expertise,


Lack of policy/claims administration
Lack of claims handling etc.
May be mitigated to a degree by getting external management to run
it?

Need to think of requirements of Solvency II if authorising in the UK,


which brings its own issues:
o How will the proportionality principal of Solvency II be applied to
captives?
o The issue of standard formula versus internal model for calculating
capital requirements would need to be considered , and if internal
model route is taken this could be very costly in terms of validation.

Page 14

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

o If they chose the standard formula and opted for the simplified basis
the captive would only be allowed to underwrite risks for entities
owned by the parent.
o The governance requirements would increase and the captive would
need to develop policies and systems around such things as
transparency, risk management and compliance functions.

(iv)

Ties up group capital which may reduce returns in terms of opportunity


costs

May be a stretch on management time which could be best utilised else


where.

The level of capital held by the captive is an important consideration as it will


impact on the captives ability to attract business (if trading independently
from the group).
The parent will need to think carefully about the balance between capital
employed in the captive against the return achieved elsewhere within the
business.
To assess the capital required to set up the captive they will have to think
about the current capital regime, but also how Solvency II would impact the
capital required (and also system requirements)
Current requirements:

The capital requirements will firstly depend on whether the captive will
only be used as a pure captive, i.e. writing only business for the parent, or
whether it will write business outside of the group.

If it will operate as a pure captive then the capital requirement may be


relatively low, depending on where the group is domiciled. This is due to
the fact that as a pure captive this self insurance will be equivalent to
internal group accounting.

It will be required to hold the MCR (statutory minimum capital


requirement) and test this against an ICA. If ICA is larger then its ICG will
be set at this amount. The level of ICG will be determined by the risks
which the captive writes, as for any other insurer. The captive will have to
ensure that it can cover its capital requirement at all times.

If it is going to write business outside of the group then it will have to


assess its capital requirements under ICAS, and this is likely to be more
thorough than the assessment for a pure captives, due to the policyholders
being wider than just the group.

As such it is likely that the capital requirement for a pure captive will be
less than that for an independent trading entity.
Page 15

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

To assess the likely quantum of capital required they will need to look at:
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o

Initial costs of setting up the captive


Lines of business to be written
Initial volumes of premium
Growth targets
Profit margins within premium rates and related adequacy of premium
rates
Tax on those profits (varying by region)
Volatility expected within the business lines
Reinsurance program likely to be used
Technical help from reinsurer
Retained risks
Value of RI
Reduction in volatility via use of reinsurance
Associated credit risk
Reserving methodology including consideration of adequacy of
reserves
What investments they are likely to hold, and how the investment
strategy will be set
Potential for liquidity risk should be assessed
What operational risks are likely to be present and how these are best
mitigated

They will also have to think about the requirements placed on captives
from Solvency II and how this may change the capital requirements.

The captive will need to consider whether it will use the standard formula
or an internal model to calculate its SCR. This will depend on the size and
complexity of the writes being underwritten in most cases.

It is likely that using the standard formula will result in an increase in


capital requirements due to the typically undiversified nature of captive
business.

The best candidates recognised that they needed to consider both the general issues relating
to capital required (line of business, reinsurance etc.) as well as issues specific to this
subsidiary being a captive.
(v)

When setting up a new subsidiary there are many administrative issues that
need to be considered as part of the application process.
Policy administration is this going to be handled internally or outsourced? If
internal what administration system
Policy wordings
Claims management is this going to be handled internally or outsourced?

Page 16

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

If internal, what system will be used to administrate, and what claims


adjusters/claims handlers will be used?
Expense administration
Premium collection
Administering legal fees
HR issues (e.g. different reward / performance management structures
required for an insurance business)
Regulatory administration (e.g. application for licenses)
Few candidates scored well with most missing obvious practical and administrative issues.
(vi)

Product liability insurance indemnifies the insured against legal liability for
the death of, or bodily injury to a third party, or for the damage to property
belonging to a third party, that results from a product fault.
The basic benefit provided by such liability insurance is an amount to
indemnify the policyholder fully against a financial loss.
This benefit may be restricted by:

A maximum claim amount per event


A maximum aggregate over 1 year
An excess
Insured retaining a proportion of any claim
A maximum amount per claim

The perils covered will depend on the nature of the products being
manufactured, but may include

faulty design,
faulty manufacture,
faulty packaging,
incorrect or misleading instructions

The policy will also normally cover legal costs.


Policies are normally written on a claims-made basis.
Some policies may also include product recall for products which have not
actually caused any damage.
Overlap of content between parts (vi)(viii) was taken into account in marking, with
candidates scoring the relevant marks whichever question the points were recorded against.
Generally candidates performed well on this section understanding the difficulties of low
data volumes and potential latency.

Page 17

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

(vii)

Pricing issues:

Firstly need to consider what rating factors will be used.

These are likely to include:


o Nature of product
o Distribution channel used
o Whether there is US exposure, due to the risk of greater litigation, torts
etc.
o Use of product
o Materials used within product
o Testing completed on product
o And associated risks of failures

In order to derive a risk premium, there is unlikely to be any internal data


available to assess the cost of claims.

Therefore need to look to an external market source to identify appropriate


data to use, this could be via reinsurers or external consultants.

Should consider premiums being quoted by other insurers for similar


products, taking into account whether there exists differences in the scope
of the cover offered.

Once a technical risk premium has been estimated would need to allow
for:
o
o
o
o
o
o
o
o

Page 18

Commissions
Expenses
Profit
Contingencies
Tax
Reinsurance costs
Capital requirements
Investment return

To make these allowances for expenses would need to look at projected


volumes of exposure in this case turnover, and then consider per policy,
and per premium loadings. The cost of handling claims and administering
policies can be compares with that for similar products.

Would need to look at whether this resulting premium was be competitive/


any anti-selection issues.

Subject SA3 (General Insurance Specialist Applications) Examiners Report, September 2011

(viii) Claims issues:

Several issues need to be considered when looking at any liability claim,


i.e.
whether or not there has been a loss,
whether the insured is actually liable,
and the quantum of the loss

Need to consider what claims information should be kept for


administration purposes but also for future rating of the product.

Need to consider the likely quantum of notification and settlement delays


involved.

Aggregation of claims, if one particular product fails.

Likely legal costs involved in litigating claims.

END OF EXAMINERS REPORT

Page 19

You might also like