Asbestos Valuation

Document Sample
Asbestos Valuation Powered By Docstoc
					     Asbestos Valuation
        CLRS – Chicago; September 8, 2003

       Kevin M. Madigan, PhD, ACAS, MAAA
Vice President, Platinum Underwriters Bermuda, Ltd.

Claus S. Metzner, FSA, FCAS, MAAA, Aktuar – SAV
Principal and Consulting Actuary, Epic Actuaries, LLC
           Asbestos Valuation
What are we faced with?
    •   a large number of claimants
    •   multiple defendants
    •   long latency periods
    •   a number of coverage years
    •   different coverages (products; premises/ops, etc.)
    •   varying policy limits
    •   limited information
    •   complex financial arrangements
  What are we to do?
            Asbestos Valuation
• Develop techniques to deal with:
  – A gigantic allocation problem between
     •   The manufacturer/installer/distributor: the insured
     •   The insurers of the manufacturer, etc
     •   The re-insurer
     •   The retrocessionaire
• Need to obtain/quantify expert knowledge
  from claim and legal staff to master the
         Asbestos Valuation
• Step I: The Insured
  – Develop the “Universe” of claims for each
  – Consider the type of claim: products versus
    premises/operations, etc.
  – Consider the length of the reporting period
  – Consider legal costs
  – Consider inflation (all kinds)
  – Consider issues of bankruptcy
            Asbestos Valuation
• What do we know/what can we find out for Step
  –   Historical reporting patterns
  –   Historical frequency and severity
  –   Claim knowledge
  –   Legal knowledge
• Goal: develop a set of assumptions and a model
  for the projection of the ultimate claims and the
  allocation to coverage period
  – Quantifying the “expert knowledge”
          Asbestos Valuation
• After we have all the assumptions, how do we
  model the results ?(the assumptions are
  developed with the help of claim and legal
  – Assumptions are generally not just a best estimate
  – Assumptions have a distribution, a range, of possible
  – Model on a stochastic basis (note: less than perfect
    approximations are useful – the output of the model
    provides information as to the sensitivity of key
         Asbestos Valuation
• Step II: The Insurer
• Once we have obtained the model output
  (note – this is a range of outputs!) can
  apply coverage charts to obtain the range
  of ultimate losses
  – Policy limits important
  – Primary versus excess
  – Other special coverage conditions
    (manuscript forms)
             Asbestos Valuation
• Having developed the range of ultimate losses for each
  insured, can now subtract the payments to date to obtain
  a range of reserves for each insured
• Note 1: pay special attention to validation routines on an
  account by account basis – the “smell test” – using the
  collective knowledge of claim and legal experts
• Note 2: consider if a covariance adjustment is
  appropriate – i.e. will everything go bad at the same
  time? - if no co-variance adjustment is used, the range
  may be misstated on the high side; if complete
  independence assumed, the range may be misstated on
  the low side
   – Each company’s mix of business by state, by type of insured,
     etc. should be considered
          Asbestos Valuation
• “Smaller” insureds with reported asbestos claims
  – some alternatives for the valuation
  – Use the information gained from the modeling
    process for large accounts to group small accounts
  – Treat each group as if it were a single account and
  – Use a review of emerging claims to assess products
    and premises/operations coverage
  – Use a review of the types of accounts to evaluate
    potential impact of policy limits
           Asbestos Valuation
  – A very difficult subject but note the following
     • Major asbestos defendants have long since been identified
       and are presumably already modeled – see Step I
     • Minor defendants are emerging - model how many additional
       accounts may report claims in the future and apply
       frequency/severity assumptions per account – in the
       alternative, model number of emerging claims (note that
       policy limits may not apply – c.f. reinsurance impact)
     • Consider where you are on the “reporting” curve
     • May want to consider issues of “modeling risk” as part of
             Asbestos Valuation
• Step 3: Assess Reinsurance Recoverable
• Reinsurance Program for each coverage/underwriting
  year can be applied to each outcome of the ultimate
  losses to develop a range of reinsurance recoverable
  consistent with the gross ultimate losses
   – Consider: dimunition/exhaustion of coverage due to other types
     of claims and previous payments
   – Consider: facultative/treaty/etc.
   – Consider: what accounts are generating IBNR and will the
     emergence of the IBNR claims even trigger reinsurance
   – Consider: creditworthiness
        Asbestos Valuation
• The previous steps have taken us through
  the process of developing a gross and a
  net liability for the primary insurer
• Now we come to the re-insurer
            Asbestos Valuation
• What are some key differences between the
  primary insurance valuation problems and the
  reinsurance valuation problem?
  –   Re-insurer has less information than the cedant
  –   Potentially more heterogeneous book of business
  –   Generally higher attachment points
  –   Generally longer lags until claims are reported
  –   Potential that primary companies’ reserving problems
      lead to a (massive) understatement of reinsurance
      recoverable by primary companies and
      understatement of re-insurer’s liability
           Asbestos Valuation
• Possible remedies for the lack of data
  – For primary companies representing a large portion of
    the exposure, investigate who the major insured
    accounts were for the years in question – then follow
    primary process with superimposed reinsurance
  – Adjust reported data for
     • Perceived reserve inadequacies
     • For better/worse than average reporting of data
         – Quality and timeliness and relative case basis reserve
           adequacy are important
  – Consider the limits profile of the primary company
         Asbestos Valuation
• Retrocessionaire Issues
  – Even less data available
  – Longer report lags
  – “Spiral” may be critical – e.g. London Excess
• IBNR very important, but the least amount
  of information is available
• Consider if the market under review – e.g.
  LMX – can be viewed as a pool
          Asbestos Valuation
• Summary
 – Key issues
   •   Long latency period
   •   Multiple coverage years
   •   Complex legal issues
   •   Complex financial structures
   •   Lack of data
        Asbestos Valuation
• Summary
 – Key Approaches
   • Gain as much information as possible
   • Use expertise of claim and legal staff to develop
   • Model results stochastically
      – Assumptions are documented
      – Can measure impact of change in environment as
        reflected in revised assumptions
   • Use consistent underlying assumptions to model
     primary, excess, reinsurance layers
        Asbestos Valuation
• Summary
 – Expect to be “surprised”
 – Keep monitoring and updating
 – Stay in close contact with claim and legal

Shared By: