Case for Alternative Risk Finance

1
Case for Alternative Risk Finance

• Appendix A
• Prepared by Aon

2
Case for Alternative Risk Finance

• This document has been presented to Capgemini and
  was prepared by Aon
• Information in the final report regarding
  Alternative Risk Finance was supported by
  material extracted from this document
• Details on the implementation and Way Forward
  for Alternative Risk Finance can be found in the
  final report

3
ARF Position Statement

• Establish a Mutual or Government Pool to finance
  LEA insurance
• A new insurance facility can be created within 3
  years whereby Local Authorities will self-insure
  a greater proportion of their Education risk by
  sharing their property and liability risks and
  resources. This facility will provide similar
  levels of cover whilst optimising cost
  efficiencies.
• The objective is to capture the underwriting
  profit that has increased in the recent hard
  market while transferring the unwanted
  catastrophe risk. Within this three year period,
  an annual cost reduction of 35 million has been
  targeted, leading to over 100 million for the
  three year period to 2009.
• The extent of the solution will be decided by the
  ability of the stakeholders to drive through
  change as this level of risk-sharing is a
  fundamental deviation from existing procurement
  arrangements.
• Rather than arranging conventional insurance
  cover, the Local Authorities will buy insurance
  from the new facility, receiving policies and
  similar terms and conditions as before at a
  similar cost.
• Over time, the underwriting profits and
  investment income will be earned by the facility
  instead of the insurers.
• There are several facilities that could be formed
  such as a mutual, a Government pool or an
  insurance company (only writing school risks and
  often called a captive).
• This facility can be formed within one year, once
  the business case has been completed. It can be
  located in the UK or offshore, typically
  somewhere like Guernsey or the Isle of Man.
• Implementation would be phased over a two / three
  years period. Within this three year period, an
  annual cost reduction of 35 million has been
  targeted.

Excess surplus will be returned to schools to
fund front-line service provision and targeted
Risk Management investment.
4
What does the current structure look like?
Current Structure in 2004

• Irrespective of risk exposure, by relying on the
  commercial insurance market, schools are faced
  with volatile pricing which is driven by the
  markets capacity to write insurance cover
  (supply and demand).
• Currently the market decides what cover to write
  and who to write it for and therefore favours the
  insurer rather than the insured, especially for
  the education sector where competition is so
  limited.
• There is considerable variation in the financing
  of risk and insurance at regional, LEA and school
  level. Inherent in this confused picture lie many
  inefficiencies that when aggregated, appear to be
  resulting in poor value for money.
• Across both Property and Liability, from a sample
  of three regions, LEAs appear to be spending more
  than twice as much on premium (139m) than is
  being recovered in claims payments (58m).
• The majority of claims are now retained within
  the various increased excesses to the policies
  and with recent large percentage increases in
  premium, insurers have reduced their loss ratios
  considerably, clawing back losses from prior
  underwriting years.

Traditional Market EL/PL and PD/BI Premium c.
139 m Claims c. 58 m
LEA Retention
School retention
Standing still is not an option change now for
the future
5
There is a fundamental need for change

• Despite best efforts of brokers and insureds, no
  new markets have entered this sector in recent
  years and there continues to be little interest.
• Our recommendation is to force change upon the
  market. Create a new market that is positioned to
  provide the policy cover that is required, while
  profiting from the current high premiums caused
  by the insurance cycle.
• The key area to target is the transferred cost to
  the insurance market. One of the primary
  principles of risk financing is to retain the
  attritional (expected) losses and to buy
  insurance for only the catastrophic risk as the
  purchase of insurance is essentially inefficient.
  In this case due to the limited capacity only
  around 7 of every 10 is recovered in claims.
  Many LEAs and schools have already moved in this
  direction either through choosing a larger excess
  or having one imposed. Most Alternative Risk
  Finance (ARF) structures operate with an expense
  ratio of less than 10, making them more
  efficient to run than conventional insurance. 
• The insurance market however is limited in the
  education sector, removing control from the
  procurer and resulting in terms being imposed
  that may be more attractive to the insurer than
  the insured. This is not an ideal market. 
• Forming a new structure to retain a significant
  amount of risk will re-design the proposition for
  either insurers (including new players) or
  reinsurers. Aligned to risk management
  improvements and investments, the additional
  savings will fall back upon the school, LEA or
  DfES.  
• The biggest potential gains can be realised by
  targeting the whole insurance cost across the
  education sector for major risk classes. The
  concept is to replace conventional risk transfer
  by alternative risk transfer in the form of
  significant risk retention.
• Due to the proposed changes in procurement of
  Supply Teacher insurance, the proposed solution
  does not include this risk class. The proposed
  focus is on Liability and Property classes where
  the greatest expense is incurred.
• The NHS Litigation Authority was formed in the
  late 1990s under conditions where insurers were
  perceived to be profiting from insufficient
  competition and those profits were better
  utilised to provide public services.

Regain control through creation of new insurance
capacity
6
Why make fundamental changes now?

• There is a strong case to support a very
  aggressive amendment to the way that insurance is
  currently procured. The cost of doing business in
  an inconsistent, devolved manner with a limited
  insurance market that has vastly increased its
  costs in recent years, makes the existing
  proposition untenable in the long-term.
•  
• Conditions will deteriorate before they improve
  unless we drive change. The reasons for this are
  as follows
• There are only a few insurers willing to cover
  education risks and the limited competition will
  keep prices high. The worst-case scenario
  involves a further withdrawal from the sector to
  the extent that a monopoly occurs, or indeed no
  market this scenario is not unheard of what
  happens then?
• The cost of all claims is increasing for a
  multitude of reasons.
• The DfES is unable to evaluate the existing
  costs.
• The current procurement model leaves schools and
  LEAs vulnerable to the volatility in insurance
  pricing from the commercial insurance market,
  largely driven by fluctuations in market
  capacity.
• There are currently significant inefficiencies in
  the procurement of insurance which are absorbing
  funds better used to finance improvements to the
  education service. By extrapolation of the data
  in the sample given, we forecast that the cost of
  these inefficiencies in the current procurement
  model will be over 35 million per annum by 2007.
• In addition to the potential savings of 35
  million per annum to be achieved through a more
  efficient risk financing model, there is
  potentially a much higher level of saving to be
  achieved through improved and consistent risk
  management practices which could be introduced to
  all 27,000 schools.
• The best way to capture this is through
  self-insurance as in the short-term any reduction
  in the cost of claims will be for the benefit of
  insurers, not schools.

The recent hard market has left a cost base that
contains significant profits to be targeted
within a workable timescale
7
The level of Local Authority compulsion is a
critical consideration in the design of the
facility
Level of Local Authority Compulsion
Mandatory - Compulsory for all LEAs and Schools

Voluntary - Optional at LEA level

• Selection of LEAs or regions forming a facility
  to insure certain risk classes
• Retain a common and consistent amount of risk,
  whilst protecting against adverse performance
  through the purchase of aggregate stop loss and
  excess reinsurance
• The selection of members would be agreed based on
  pre-determined criteria such as
• loss ratios
• loss history
• common size and exposures
• This type of structure could be created in many
  ways, for example
• Geographically diverse - split England and Wales
  into four districts (NSEW) and select a set
  number from each district
• Geographically close - select the members from
  similar areas
• By risk class - based on selection criteria, e.g.
  a facility for property, one for liability, etc.
• By school type - form one facility for all
  schools of a similar nature

• Making it mandatory for all schools to insure
  chosen risk classes via their LEA
• Scheme will be a Government owned / sponsored
  facility, such as a pool or mutual
• The procurement process will be simple, rated
  fairly, consistent and accountable
• This provides the greatest potential for
  cost-reduction, through best economies of scale,
  greater level of risk-sharing and improved
  incentives.
• Uncertain of reaction of existing insurers to
  this withdrawal from market

The voluntary option contains many challenges
e.g. who joins, procurement, insurer reaction.
The NHS Litigation Authority, an example of a
mandatory facility, was formed in the late 1990s
under conditions where insurers were perceived to
be profiting from insufficient competition and
those profits were better utilised to provide
public services.
8
What will the new structure look like?

• For the schools, the cover will look remarkably
  similar to existing cover. Funding will be to a
  level that covers most loss scenarios with
  reinsurance protection above this to a
  catastrophe loss level.
• Claims will be paid by the facility in return for
  an annual premium, which will not be subject to
  the market swings that have been suffered
  recently.
• The facility will be limited to writing this risk
  only eliminating reliance on other groups or
  risks.
• There will be real consistency in the financing
  of risk and insurance at regional, LEA and school
  level, as shown opposite. The procurement picture
  at the highest level will provide efficiencies
  that will drive best value for money.
• For both Property and Liability, LEAs will
  continue to spend similar amounts on premium
  (140m) but any profits will be retained in the
  facility.
• The majority of claims will be retained within
  the LEA retention and the pool, with only
  catastrophic losses being covered by the
  reinsurance market.

Example of Structure proposed for 2007
1 bn
Traditional Reinsurance Market
500,000/ 1,000,000
Pooling arrangement
100,000
LEA Retention
500
School retention
9
What benefits can be achievedwithin three years?

• Making such radical changes will provide
  challenges but the change scenario is very
  appealing. A suitable structure will
• Return excess profit to schools to fund
  front-line service provision
• Provide similar levels of cover, more efficiently
• Decrease the overall cost of risk by eliminating
  insurer expense and therefore run at lower cost
  than private sector insurer
• Recognise the implementation of risk mitigation
  processes financially
• Allow access to greater market capacity at
  catastrophe levels
• Improve transparency of pricing through visible
  reporting
• Better understand the key claims issues with
  larger population
• Achieve more consistency in insurance procurement
• Give schools more choice and scope over the
  insurance they purchase
• Reduce the majority of price volatility from risk
  financing which will make budgeting easier
• Common ownership will enable better sharing of
  best practice in terms of risk management to
  reduce overall total cost of risk all in the
  same game

Eliminate the fat in the system to deliver
savings to invest in front-line transaction
10
The sample results from 3 regionsprovide
evidence of savings

• During the eight week exercise, a representative
  sample has been utilised to form an opinion on
  the results of the whole system for school
  insurance across England and Wales.
• From the sample of three representative regions,
  the extrapolated results indicate potential
  savings of around 35 million per annum within
  three years. This is a high level view and
  contemplates a radical switch in procurement
  policy that may take some time to develop across
  major classes of insurance, namely Property
  Damage / Business Interruption (PD/BI),
  Employers and Public Liability (EL and PL).
  Supply Teacher (ST) insurance has been removed
  due to its imminent restructure.
• What are the major reasons for drawing these
  conclusions?
• Our sample loss ratios for three regions are as
  follows PD/BI- 50 EL/PL - 25 - much of the
  difference flows to the insurer(s) by way of
  profits
• We appreciate that historically loss ratios have
  been much higher for PD/BI, however the recent
  premium increases indicate these current loss
  ratios when comparing 2003 premiums with average
  claims.
• The aggregation of buying power of all LEAs will
  deliver economies of scale to minimise costs -
  insurance procurement is devolved to varying
  levels - bundling will provide efficiencies
• There are potential gaps and duplication of cover
  through the existing procurement process wrong
  level of cover being purchased
• Schools and LEAs are not reaping premium
  discounts for risk management improvement there
  is little incentive to improve
• The limited market, which varies by insurance
  class, removes competition and increases pricing
  increase supply to reduce pricing
• By retaining a higher aggregate level of risk for
  a much larger single group of schools reliance on
  the external commercial insurance market will be
  substantially reduced
• The model we advocate would also exert huge
  leverage on the market for risk control services
  and physical risk improvement products and
  services for the benefit of individual schools.

Detailed results are a compelling reason for
change see next section
11
Assumptions for projection of future premiums and
claims cost

• Data Utilised
• Premium extrapolated from 11 premium data sets
• Claims extrapolated from a loss ratio calculated
  from weighted average VFM figures from 2000 to
  2003 25 for Liability and 50 for PD/BI. As
  previously advised, Supply Teacher insurance has
  not been included in these projections. This is
  the Expected case. Pessimistic and Optimistic
  scenarios have been considered with loss ratios
  of 40 and 20 for Liability and 60 and 40 for
  PDBI respectively.
• Assumptions
• Data used is representative and hence
  extrapolation can provide realistic picture
• Extrapolation undertaken on the basis of number
  of schools, using those that were included in the
  sample, as provided from government statistics as
  at January 2004
• Average retained claims have been calculated over
  a period of 1999 to 2003
• With some LEAs increasing their deductible and
  some decreasing to a common level, premium to
  insurers and retained costs will remain at the
  calculated costs in 2004
• Future inflation of premiums and claims assumed
  is summarised below
• It is believed that the premiums are at the top
  of the cycle and will decrease before increasing
  to the level of 2004 by 2009, as indicated by
  Zurich
• Claims have been assumed to continue to increase,
  as indicated by Zurich
• Reinsurance costs are a guestimate without any
  markets being approached and are illustrative
  rather than robust
• Administration costs estimated at 10, which is
  very conservative.

12
The case to include Liability Insurance

• GENERAL
• Includes Employers Liability (EL) and Public
  Liability (PL)
• In general the claims are of high frequency and
  low severity
• The relatively predictable levels of future
  claims provide accurate funding levels for
  alternative solutions
• The sample does not include any noticeable latent
  claims. If any alternative solution is
  established on an occurrence basis then it will
  only be liable for current exposures
• Long-tail nature provides cash-flow benefits i.e.
  investment income on funds until claims are paid
• SPECIFIC based on our sample of 3 regions
• Approximately 40 of liability premiums are
  allocated to EL 60 to PL
• Approx. 60 of claims cost from 1999 to 2004 is
  due to PL
• Approx. 80 of the current claims cost from 2000
  onwards is outstanding
• An average of 80 of claims cost is retained
• Deductibles vary between nil and 250,000
• Based on our sample current loss ratios are low,
  from 13 to 38 i.e. claims reported v premiums
• There are 4 Insurers in our sample, with ZM
  having 40 of this market in terms of number of
  schools
• Collecting claims (and even premium) information
  on these classes has proved very tough
• The largest claim since 2000 in our sample of 15
  LEAs schools, was a PL claim for 503,000
• Premiums have increased by 110 from 2000 to 2004
• Average loss ratio from 2000 to 2003 for the
  sample (including development) is approx. 25
  (40 in first two years)

Liability classes are ideal for funding through
an alternative structure
13
What would a new Liability facility look like?
Example of Structure proposed for 2007
Current Structure.. in 2007
1 bn
Traditional Reinsurance Market Premium example
7m
Traditional Market Premium 46m Claims 11.5m
500,000
Pooling arrangement
100,000
LEA Retention
LEA Retention
500
School retention
School retention
Risk transfer premiums reduced substantially from
46 million within three years
14
Expected case example financials for Liability
Insurance

Reinsurance example purely for illustrative
purposes and subject to change depending on
reinsurance structure and available markets.
Target surplus of 19 million p.a. within three
years
15
The case to include Property Insurance

• GENERAL
• Includes Property Damage and Business
  Interruption (PD/BI)
• This risk although having a high frequency of low
  severity losses such as Theft also has the
  potential for the low frequency and high severity
  claims such as a rebuild after arson
• Despite investments in risk management little
  recognition has been received from the insurance
  market
• For some LEAs the major costs come from ICT,
  Water damage and Theft despite having a smaller
  average cost
• Some LEAs will rebuild schools while others will
  consider this not worthwhile and incorporate the
  pupils into other schools, particularly with a
  general trend in declining number of pupils.
• Volatile levels of current and future claims
  provide difficulties with funding levels for
  alternative solutions over the short term, with
  potentially more stability over the long term
• Short tail (quick settlement) nature provides
  little cash-flow benefit
• SPECIFIC based on our sample of 3 regions
• PD/BI premiums are approximately twice as high as
  liability premiums
• Approx. 20 of the current claims cost from 2000
  onwards is outstanding
• Deductibles vary between nil and 500,000
• Based on our sample, current loss ratios vary
  considerably from 13 to 165. The average
  historic loss ratio from 2000 to 2003 is 85.
  However, premiums have increase significantly and
  relative to the current premiums is a lot lower.
• Collecting claims (and even premium) information
  on these classes has proved very tough
• The largest claim since 2000 in our sample of 15
  LEAs, was a claim for 7.4m.
• Premiums have increased by 120 from 2000 to 2004
  and therefore loss ratios based on current
  premiums would exhibit lower figures. Taking an
  average of inflated claims (3p.a) from 1999 to
  2003 from the VFM calculations and comparing to
  2003 premiums the loss ratio is approx. 50

More volatile PD/BI risks and historically high
loss ratios are unattractive - higher premium
rates in 2003/4 reveal greater reason for
Alternative Risk Financing
16
What would a new Property Insurance facility look
like?
Current Structure in 2007
Example of Structure proposed for 2007
TBA
Traditional Market Premium example 10m
Traditional Market Premium 93m Claims 46.5m
5m
Pooling arrangement
150,000
LEA Retention
LEA Retention
500
School retention
School retention
2004 premium of 93 million (estimated at 88m in
2007) is double the 2001 premium level, improving
loss ratios (premium/claims) significantly
17
Expected case example financials for Property
Insurance

Reinsurance example purely for illustrative
purposes and subject to change depending on
reinsurance structure and available markets.
Current premiums are at top of cycle good time
to retain these amounts and work towards reducing
losses to increase profitability
18
The case to exclude Supply Teacher Insurance

• GENERAL
• Supply Teacher Insurance includes the insurance
  for the supply teacher and other staff as
  required such as cooks and bursar
• Other than long term sickness, claims are of
  middle frequency and low severity of claims
• Across a portfolio, the level of claims should be
  reasonably predictable providing for accurate
  funding levels for alternative solutions. There
  is no catastrophic risk
• No information available on existing loss ratios
  other than from Doncaster who build in little
  margin of 5
• If the expected change in procurement of this
  type of cover does not occur, it contains good
  characteristics for ARF and could be added to the
  risks to be insured by the facility.
• SPECIFIC based on our sample of 3 regions
• Supply Teacher premiums are approximately 50 of
  total premium spend, as provided by Government
  data
• From our sample a number of LEAs self fund this
  risk already
• Doubt was expressed as to the meaning of these
  premium figures from the Section 52 data
• Excesses are expressed as days e.g. a 10 day
  excess means no cover until day 11
• Collecting claims (and even premium) information
  on these classes has proved especially tough,
  only 6 of 18 responded
• Great doubt over validity of data

It is generally accepted that the whole structure
of this cover will be amended in the next year or
two therefore, including this in future analysis
is unjustified.
19
Business Plan 2007 - 2009

• GENERAL
• Blend of insurance classes provides
  diversification of the portfolio
• Each class has its own characteristics which will
  decide whether it is ultimately suitable for
  inclusion or in particular the timing of its
  inclusion
• General market perception that all these classes
  are a bad risk e.g. stress and arson are
  major causes
• The limited interest of other insurers in
  entering the market enables the existing ones to
  price accordingly
• Further evidence of existing loss ratios and
  other trends will assist our understanding of the
  potential long-term benefits
• Certain Risk Management and IT improvements are
  required to enhance future returns
• A capital base of approximately 30m would be
  required to write this level of risk for both
  these classes
• SPECIFIC based on our sample of 3 regions
• Initial review of results provides cautious
  confidence in the potential to reduce costs over
  time, depending upon the incentives that can be
  built in
• Collecting claims (and even premium) information
  on these classes has proved very difficult
  systems appear to have flaws
• For a robust business plan, reliable claims and
  premium data is essential from the existing
  sample
• There is little consistency between LEAs in the
  risk information systems used and indeed how they
  use systems and record losses. This makes
  collation of consistent data time consuming.
• It would be worthwhile for all LEAs to work
  towards implementation of a risk information
  protocol to agree a common approach to the
  treatment and recording of risk information.

20
Expected case Business Plan 2007 - 2009

Reinsurance example purely for illustrative
purposes and subject to change depending on
reinsurance structure and available markets.
Improved efficiencies over time may well increase
savings from 100 million for three years from
2007
21
Expected case comparison of Do Nothing with
Pooling Arrangement
The following shows the financial comparison of
Do Nothing to a Pooling Arrangement
Total 105.6m
22
From 2007/8, an Alternative Risk Financing
vehicle could deliver significant benefits
Create an Alternative Risk Financing Vehicle for
EL/PL

• Key Assumptions
• Loss ratios of 40, 25 and 20 for Pessimistic,
  Expected and Optimistic case respectively, being
  the highest, average and lowest loss ratio from
  our sample
• Reinsurance premium of 7m purely to show that
  there will still be a cost of risk transfer but
  the scale cannot yet be determined with any level
  of authority
• Operating Costs - the type and size of facility
  will drive the administration costs, but 10 of
  gross premium is a conservative estimate
• Long term investment income rate of 4 per annum
  assumed
• 100 adoption rate by Local Authorities

Benefits yearly from 2005/6-2009/10
Create an Alternative Risk Financing Vehicle for
Property
Benefits yearly from 2005/6-2009/10

• Key Assumptions
• Loss ratios of 60, 50 and 40 for Pessimistic,
  Expected and Optimistic case respectively, being
  the average inflated historic claims from 1999 to
  2002, 1999 to 2003 and 1999 to 2001 compared to
  2003 premium
• Reinsurance premium of 10m for illustrative
  purposes, as noted above
• Operating costs of 10, as noted above
• Investment income of 4 per annum
• 100 adoption rate by Local Authorities

These benefits dont consider the impact of the
other propositions. In reality, the benefits from
reducing the claims costs will only enhance the
potential benefits of ARF.
23
Alternative Risk FinanceSupporting analytical
slides

• Appendix B
• Prepared by Aon

24
Data as at 10th December 2004
25
Data Issues

• Received 15 of the 18 questionnaires
• Premium and claims provided for different periods
  limiting inclusion to the shortest period 2000
  onwards (assumptions made for some to complete
  this period)
• Four LEAs excluded from value for money (VFM)
  calculation due to
• One LEA with no premium figures (North Yorkshire)
• One LEA with no claims figures (Cornwall)
• One LEA with only number of claims for EL risk
  and premium for one year only (East Sussex)
• One LEA with only a combined premium (South
  Gloucestershire)
• Limited data for Supply teachers
• Premium and claims information from six
• A number of LEAs appear to self fund this risk
  class
• Limited exposure data from which to draw trends
• Insurance recoveries not input correctly for some
  of the LEAs. This is crucial for the calculation
  of the VFM. IRMG have made changes where
  individual loss data was provided.
• We have relied on loss and exposure information
  provided by or made available without benefit of
  detailed verification or audit other than checks
  for reasonableness. We do not assume
  responsibility for the accuracy or completeness
  of the data or material provided to us.
  Additional information or change in assumptions
  may produce results completely different than
  displayed.

26
Number of Schools

• Includes all 15 LEAs
• Significant proportion of LEA schools, 82

27
Number of Schools

• Includes all 12 LEAs
• Largest LEAs
• Kent County Council - 543
• Hampshire - 521
• Staffordshire County Council - 411
• Smallest LEAs
• Bath and North East Somerset Council - 85
• City of York Council - 70
• Torbay Council - 43

Is this a representative sample in order to
extrapolate? Total number of schools is 3,588
28
2004 Split per insurer

• Premium figures exclude East Sussex and North
  Yorkshire.
• Based on the number of schools, Zurich Municipal
  has almost half of the exposure.
• Based on the 13 LEAs (premium for AIG not
  available) included in the premium calculation,
  Zurich, Zurich and CBMDC and St Paul
  International have approximately a third each.

ZM is the dominant insurer
29
Historic EL / PL Premiums for Portfolio

• 12 LEAs (excludes South Gloucestershire, East
  Sussex, North Yorkshire)
• Assumption made for Staffordshire premium for
  2004 based on average increase across the LEAs
  from 2003 to 2004
• Split EL / PL is stable the last 2 years.
  Approximately, a third is EL premium.
• Increase in premium from 2003 to 2004 15
• Increase in premium from 2001 to 2004 110

30
Historic PDBI Premiums for Portfolio

• 12 LEAs (excludes East Sussex, South
  Gloucestershire and North Yorkshire)
• Assumption made for Staffordshire premium for
  2004 based on average increase across the LEAs
  from 2003 to 2004
• Approximately PDBI total premium is double of EL
  / PL premium.
• Increase in premium from 2003 to 2004 is
  reasonably low.
• However, the increase in premiums over the last 4
  years is 120.

31
Historic Supply Teacher Premiums for Portfolio

• Based on 6 LEAs (Devon, Doncaster, South
  Gloucestershire, Lewisham, Sheffield and
  Hampshire)
• Doncaster fully self-insured
• At this stage, significant increases have not
  been observed
• Due to the lack of data no further analysis has
  been carried out at this stage

32
2004 Deductible Levels
Self Insured

• Different levels across the portfolio, from nil
  to 500,000
• For 3 LEAs, the deductibles are the same for all
  risks
• The LEAs with the largest deductibles are
  Doncaster, Devon and Lewisham, The two LEAs with
  nil deductible are Cornwall and Bradford

A common deductible would be required for the
operation of an Alternative Risk Financing
Vehicle (ARF)
33
2004 Deductible Levels and Aggregate Stop Loss
34
2004 Limits

• EL and PL
• Across both classes, limits are common
• Minimum 20 million, maximum 50 million
• PDBI
• From the data available, limits are approximately
  200 million and above

What limit would be required for an ARF vehicle ?
35
Historic EL / PL Claims for Portfolio

• Transferred Costs Insurance Recoveries
• 14 LEAs (excludes Cornwall)
• Claims for Lewisham for 2000 estimated from
  average of 2001 and 2002, split between paid and
  outstanding estimated from other LEAs
• Large proportion of outstanding claims -
  approximately 80 over the period
• The number of claims would be affected by late
  reporting

36
Historical Claims ExperienceIncurred But Not
Reported (IBNR)
Current position does not represent ultimate cost
IBNR
Movement in current case reserves
Late reporting of claims
Examine average pattern of development data from
Aon public sector database Factors selected for
each LEA according to their renewal date and data
valuation date.
Assumed the same factors for retained claims
and for transferred claims
Apply to each LEA aggregate loss per policy year
37
Historic EL / PL Claims for Portfolio

• Developed costs for 2004 should be treated with
  caution due to immaturity
• Retained costs are relatively stable
  (approximately 78 of the claim costs retained)
• For these 14 LEAs, the 2000 2004 average
  developed claim cost is approximately 3.1
  million (0.7 million transferred)

38
Historic PDBI Claims for Portfolio

• 13 LEAs (excludes Cornwall and Hampshire)
• Hampshire only losses above 100,000 were
  available
• This risk is more volatile, minimum ground-up
  claims of 4 million in 2004, maximum 13.8
  million in 2002
• The majority of ground-up cost is paid (only 14
  outstanding 2000 to 2004)
• The 1999-2004 average number of claims is 3,706,
  largely due to Doncaster (on average 1,800 claims
  per year)
• The 1999-2004 average ground-up claims is 7.5
  million with a particularly bad year in 2002
  (largely due to Arson claim, 7.4 million).
• Are all claims being reported?

39
Historic PDBI Claims for Portfolio

• Retained costs are more volatile (approximately
  57 of the claim costs retained over the period).

40
Large EL / PL and PDBI losses

• The large liability losses above 250k and PDBI
  losses above 1.5m are shown from 1998 onwards
• The size of EL / PL claims is smaller
• Other samples may well contain larger individual
  claims
• The two largest EL / PL claims are from lack of
  supervision cause
• Large PDBI claims are all fire with 4 out of the
  8 being arson
• There are various reporting periods by LEA, e.g.
  1992 large losses provided by Bradford

41
Value for Money Comparator

• Loss ratio
• Transferred cost/Premium to insurance market.
• Where loss ratios are greater than 100, shown in
  red overleaf
• Ratios shown by risk class and as a portfolio.
• Two approaches
• Current position of losses as they are, without
  any assumption about their future development.
• Losses for EL / PL on a developed basis, allowing
  for movement in case reserves and IBNR.
• For each LEA, only the specific years where both
  the premiums and claims were available have been
  used for the Value for Money calculation

42
Value for Money Comparator EL / PL

• Excludes Cornwall, South Gloucestershire, East
  Sussex and North Yorkshire for all years
• Nota Bene LEAs have only been included in the
  years where both premium and claim information is
  available therefore, each year may not exactly
  mirror the specific LEAs included in each loss
  ratio calculation
• Loss ratios appear very low with an average
  across the period 2000-2003 of 21 rounded up to
  25 for extrapolation purposes
• 2004 figures have decreased because Staffordshire
  is excluded for this year as no premium figure
  was available

Based on this sample, there appears to be a
strong case for further analysis to explore the
financial viability of an ARF vehicle.
43
Value for Money Comparator EL / PL

• The figures are shown on a developed basis
• Surplus premium to insurers recoveries from
  insurers
• Surplus in 2002 is 2.5 million for a premium
  level of 3.1 million (81)
• In every year, there is a significant gap between
  premiums paid and recoveries received
• This does not represent good value for money over
  the last 6 years
• Again, is this a representative sample?

44
Value for Money Comparator PDBI

• Excludes Cornwall, South Gloucester, East Sussex
  and North Yorkshire for all years
• NB LEAs have only been included in the years
  where both premium and claim information is
  available therefore, each year may not exactly
  mirror the specific LEAs included in each loss
  ratio calculation
• Loss ratio is more volatile with an average
  across the 2000-2003 period of approximately 85.
  If claims are adjusted for inflation (3p.a) and
  compared to 2003 and average loss ratio of
  approx. 50 arises. This lower figures has been
  assumed for future projections
• 1999, 2001 and 2002 severe loss ratios are due to
  four large arson from one LEA
• We do not have full premium information for 2004,
  however, this year has incurred an arson loss of
  2 million

Loss ratios vary significantly, but current
premium levels appear to have been adjusted to
account for poor underwriting years.
45
Value for Money Comparator PDBI

• With three years having a loss ratio greater than
  100, historical premiums have been a little low
  up to 2001. A significant increase has been
  observed since 2001.
• Compared to average claims (4.2 million) the
  2003 premium is more than sufficient. However, as
  seen in 2002, there is potential for very large
  losses, which could erode any apparent surplus.

With such a volatile risk, there is uncertainty
about the probability of surplus
46
Current Long Term Agreements

• Three of the fifteen have LTAs running beyond 2007

The existence of LTAs may affect the potential
formation date of any ARF vehicle