Intermediate Track I

1. Intermediate Track I Considerations in Evaluating Reserves 2000 CLRS Minneapolis, Minnesota

2. Introduction Purpose: To discuss and illustrate twelve concepts that, if not understood, could cause a reserver to draw faulty conclusions Intermediate Tracks II through IV will provide additional insights and techniques useful in addressing several of these issues

3. Considerations 1. Average Closed Value is not the same as Average Open Value 2. Closed Savings are not the same as Open Redundancy 3. Early Reported Claims are not the same as Late Reported Claims 4. The ratio of Paid Allocated Loss Expense (ALAE) to Paid Loss increases over time 5. Segregate ALAE into Components 6. Rate adequacy can impact reserving 7. Higher limits mean more future development 8. Higher attachment points mean more future development 9. Positive Development does not mean a Claim Department problem 10. Operational changes affect reserving 11. Incomplete accident years can be deceiving 12. Tail development is important

4. Consideration #1 The average value of claims currently closed is often a poor estimator of the ultimate average settlement value of claims still open.

5. Consideration #1 (cont.) Accident Year 1985 Why might this frequently be true?

6. Consideration #1 (cont.) Claims that close early are often smaller. • Case A Should settle quickly • Case B May not settle quickly, and may settle for substantially more than the insurer’s estimate.

7. Consideration #1 (cont.) Workers Compensation: The cases that close quickly are for minor injuries, and may involve just medical-only costs. The cases open for a long period represent severe injuries and may include: • Major Medical Expenses • Lifetime Pension Benefits The longer a case is open, the more periodic indemnity payments will be made, and the medical bills likely will be more costly.

8. Consideration#2 Savings on closed claims may be a poor estimator of case reserve adequacy for open cases.

9. Consideration #2 (cont.) Examples include: (1) Often, cases that are more adequately reserved settle more quickly (a closer “meeting of the minds”). • Case A Will likely settle first, and show savings. • Case B May fare differently, and settle much later for more than the case reserve.

10. Consideration #2 (cont.) (2) Back injuries for Workers Compensation An adjuster may review 20 back cases. Those who return to work earlier than expected close for “Savings”. Others may later be classified as permanently disabled, at a much higher cost than initially expected and not closed for decades.

11. Consideration #2 (cont.) (3) Changes in Case Reserves Before Settlement It is not unusual for an estimate of an individual case reserve to be strengthened just prior to closing, and then show “savings”. This may/may not reveal anything about case reserves further from settlement. Therefore, a comparison of settlement costs to the corresponding prior case reserves just before closing can be a very misleading comparison.

12. Consideration #3 The average costs for late reported claims may differ materially from those reported earlier.

13. Consideration #3 (cont.) Reason: Often, late reported claims have a very different nature than those reported early. (1) General Liability: • Product Liability cases are often reported later • Product cases are often more complex, requiring expert testimony and lengthy litigation • Product cases reported very late may involve latent injury or cumulative exposure, cases which are difficult to define in terms of: • Date of loss - Party at Fault • Number of occurrences - Type or extent of injuries

14. Consideration #3 (cont.) (2) Workers Compensation: Most Workers Compensation cases are reported within the first 18 months. However, when there are late reported claims they often involve occupational diseases, rather than trauma that is quickly identified and assignable to a single accident date and/or policy.

15. Consideration #4 For an accident year, the future ratio of ALAE to loss may be materially higher than has been true for payments to date.

16. Consideration #4 (cont.) Reasons: 1) Cases open for lengthy periods often involve costly litigation. 2) Legal payments are occasionally disbursed later than loss payments.

17. Consideration #4 (cont.)

18. Consideration #4 (cont.) This pattern by company can be influenced by many factors, such as the mode of payment of legal bills, which may vary by company between: • Interim Case Billing • End of Case Billing Other influences can include: • Geographical Differences • Use of Staff Counsel vs. Outside Counsel • Classes of Business • Primary vs. Excess Contracts

19. Consideration #5 Where ALAE costs are rising, it may be useful to split ALAE into components such as: • Outside Attorney Fees • In-House Attorneys • Other Legal • Expert Witnesses • Medical Audits/Reviews

20. Consideration #5 (cont.) Reasons: (1) Legal expense are typically the fastest growing component of ALAE, with a growth rate exceeding trends in loss costs. (2) Many companies have attempted cost savings steps such as: • Use of staff counsel, rather than independent attorneys, in some situations • Use of companies which audit legal bills • More vigorous defense (which may slow payment patterns on loss side) • Initiating contact with the claimant sooner

21. Consideration #5 (cont.) Note that the recent change in the definition of ALAE and ULAE may result in distortions in ALAE development. This recent definition re-defined ALAE to include expenses associated with defense, litigation or medical cost containment expenses, whether internal or external.

22. Consideration #6 If loss ratios based on prior years’ experience are to be used in reserving, they must be adjusted for any material changes in rate adequacy.

23. Consideration #6 (cont.) Rate adequacy often changes over time for reasons including: • Industry pricing cycles • Internal company initiatives • Growth strategies etc.

24. Consideration #6 (cont.)

25. Consideration #7 When reinsurance retentions and/or policy limits increase, the portion of ultimate losses that are reported at each given maturity tends to decrease.

26. Consideration #7 (cont.)

27. Consideration #8 When attachment points increase for reinsurance, excess, umbrella or self-insured coverages, then the percentage of ultimate dollars that is reported at each given maturity tends to decrease.

28. Consideration #8 (cont.)

29. Consideration #9 Upward case development does not necessarily demonstrate something “needs fixing” in the Claims Department.

30. Consideration #9 (cont.) • To Illustrate, Assume: • The company’s reserving philosophy is to set the best estimate of ultimate cost given solely the facts at the time. • For an accident year, there are 100 back injury cases (Workers Compensation) • At 24 months, all injuries appear similar given medical evaluations (etc.) at that time. • However, the Claim Department accurately estimates 97 of the 100 at 24 months, but as facts emerge over the ensuing years, 3 workers are determined to be permanently disabled.

31. Consideration #9 (cont.)

32. Consideration #10 Internal company changes can dramatically affect patterns in reserving data, and distort the result of basic reserving methodologies.

33. Consideration #10 (cont.) Examples of Such Changes: • Mix: Until 2 years ago, the company’s workers compensation book primarily covered flower shops. It now contains significant exposures for general contractors. • Likely implications: The more recent accident years will involve more serious injuries and longer future development • Growth: The company is expanding, with an increasing percent of the book being new business. This new business may come with higher loss ratios and may have an uneven distribution of writings throughout the year. • Claims Management: • A new computer system is implemented, speeding up the recording of claims and payments • A new case reserving form is developed to help claim technicians more readily consider all costs. • A change in TPA (third party claims adjuster)

34. Consideration #10 (cont.) For example, suppose the company changed TPA’s 12 months ago, and now has the following triangles:

35. Consideration #10 (cont.) Paid to Reported Ratios are an example of a diagnostic tool which can be used to check for: • Case reserve strengthening (this example) • Case reserve weakening • Change in rate of payment Later sessions will discuss methods, such as the Berquist & Sherman approach to correct for these kinds of changes.

36. Consideration #11 Estimating ultimate losses for an incomplete accident year may require special adjustments.

37. Consideration #11 (cont.) Factors to consider in estimating ultimate losses for an incomplete accident year: • Loss development may not be even throughout the accident year: • Seasonality in accident dates, reporting patterns, or settlement patterns • Exposure changes during the year • When the traditional loss development method is applied to accident year data, ultimate losses are projected for the entire accident year, even if losses are evaluated at an interim date

38. Consideration #11 (cont.)

39. Consideration #11 (cont.)

40. Consideration #11 (cont.) NO. Estimated ultimate losses for accident year 2000 as of 9/30/00 are for the first 9 months only. The projection of $409,599 (130,000 x 3.15) is for the whole accident year (12 months). One solution, which does not consider seasonality or growth:

41. Consideration #12 “Tail Development” can have a dramatic effect on reserve needs.

42. Consideration #12 (cont.) The Need For “Tail Factors”

43. Products Workers Compensation Medical Malpractice Complex issues (Who’s liable? How to prove injury was caused by product? Date of loss? Occupational Disease Life pension cases, with escalation clauses in some states’ benefit structures Medical costs on life pension cases. Child injured at delivery reaches legal age Delayed manifestation, with subsequent complex issues. Consideration #12 (cont.) Some examples of when development occurs beyond 10 years

44. Consideration #12 (cont.) Techniques To Derive Tail Factors 1. Examine broader data sources e.g. ISO, NCCI, RAA, AM Best (Caution: Learn the limitations of such data) 2. Curve Fitting 3. Generalized Bondy Method

45. Consideration #12 (cont.)

46. Consideration #12 (cont.)

47. Consideration #12 (cont.) CURVE FITTING

48. Consideration #12 (cont.) There are numerous types of curves which are typically used to project tail factors. One of the common ones is the POWER MODEL: Y = A ^ (B ^ X) Y variable = Loss Development Factor X variable = Time (in months) A, B are constants.

49. Consideration #12 (cont.) To determine the best curve for a set of data, we need to determine the correct A and B parameters. For the Power Model, this is done by converting the equation to a linear form and then determining the slope and intercept of the resulting linear equation. We want an equation of the form Y' = intercept + slope * X where Y' is called a "transformed value" of Y

50. Consideration #12 (cont.) Y = A ^ (B ^ X) 1. Take the Natural Log of both sides of the equation. LN (Y) = LN (A^(B^X)) According to the rules of logarithms, this is equivalent to: LN (Y) = (B ^ X) * LN (A)