What is a Reasonable Range Outline Range of

What is a “Reasonable” Range? Outline • Range of what? • For what purpose/use – • Whose range? • Transparent disclosure • Uncertainty, conservatism and bias 2006 CLRS What is a “Reasonable” Range – R. Blanchard 1

Range of what? Major issues in this area • Outcomes vs. estimates? • If estimates, what’s the measurement basis? • Speculative outcomes • Range endpoints 2006 CLRS What is a “Reasonable” Range – R. Blanchard 2

Outcomes vs. Estimates (part 1 - outcomes) Range of “reasonable outcomes” • “Reasonable outcome” = “Reasonably likely”? • If so, this implies a confidence interval, • implies a distribution, • but not necessarily a complete distribution Focuses on the middle of the distribution, not the tail, and middle should be more quantifiable 2006 CLRS What is a “Reasonable” Range – R. Blanchard 3

Outcomes vs. Estimates (part 2 – estimates) Range of estimates • Implies different opinions of what is likely • Not necessarily compatible to “confidence interval” presentation • Doesn’t require a distribution Not “what could happen”, but “what is reasonable to expect”. 2006 CLRS What is a “Reasonable” Range – R. Blanchard 4

Measurement basis What is a best estimate? “Best estimate” of what? • Mean? • • • Mode? Median? Central tendency? 75 th percentile? PML? Mean plus one standard deviation Could be any of the above, but does your audience know which one you’re estimating? Do you? (Communication opportunity? ) 2006 CLRS What is a “Reasonable” Range – R. Blanchard 5

Speculative Outcomes What is speculation? • If I can reliably quantify it, it isn’t speculation • Just because I can come up with a value, doesn’t mean it isn’t speculation. N = N* × fpm × ne × ng × fi × fc × fl × fm × fj × fme “Drake’s Equation” Key is reliable estimation. • Speculative outcomes may not be appropriate for inclusion in estimate Common rule – book estimates, disclose but don’t book speculation 2006 CLRS What is a “Reasonable” Range – R. Blanchard 6

Range Endpoints “Reasonable” is not a black and white standard This implies that the endpoints of a reasonable range are not exact or firm. 2006 CLRS What is a “Reasonable” Range – R. Blanchard 7

For What Purpose/use? The desired range may vary (or even be dictated) by the purpose/use. (I. e. , your options may be limited. ) 2006 CLRS What is a “Reasonable” Range – R. Blanchard 8

For What Purpose/use? – Part 1 Internal communication • No “a priori” limitations • Find out what is expected SEC filing • Range of reasonable estimates How much are current earnings a function of judgment? • Range of outcomes How much are future earnings at risk? 2006 CLRS What is a “Reasonable” Range – R. Blanchard 9

For What Purpose/use? – Part 2 Merger/acquistion • Buyer/seller may want to know many different things • Expected payments? • Worst case payouts? • Possible one year development? • other? • Find out what your audience wants Commutation work • See Merger/acquisition discussion above 2006 CLRS What is a “Reasonable” Range – R. Blanchard 10

Whose range? Some choices Reserving actuary’s Management’s Auditor’s The role of the range depends on whose range it is. 2006 CLRS What is a “Reasonable” Range – R. Blanchard 11

Whose range? – Reserving actuary’s? Range of estimates • Aids management in determining their best estimate • Should include discussion of drivers for various pts in range • Possible input for appointed actuary’s analysis Range of outcomes Part of risk analysis 2006 CLRS What is a “Reasonable” Range – R. Blanchard 12

Whose range? – Management’s? Auditor’s Management’s range • Used for determining what to book Auditor’s range • Used for determining reasonableness of management’s estimate (i. e. , what was booked) Concern is not whether management is continually booking near the low or the high of their range Concern is instead whether they are moving around in their range from year to year. (Note: Moving around in auditor’s range is not the same thing. ) 2006 CLRS What is a “Reasonable” Range – R. Blanchard 13

Transparent Disclosure – 4 major rules 1. Know your audience ! 2. Know your audience !! 3. Know your audience !!! 4. Your audience may not be fellow actuaries. 2006 CLRS What is a “Reasonable” Range – R. Blanchard 14

Transparent disclosure – a few other suggestions Be clear as to what kind of range it is Outcome? Estimates of what? Be clear as to how reliable it is Be ready for outcomes outside your previous range May want to disclose that possibility in advance. Be ready for misinterpretations (innocent and otherwise) 2006 CLRS What is a “Reasonable” Range – R. Blanchard 15

Uncertainty, Conservatism & Bias With all this uncertainty, shouldn’t I be a bit conservative in my numbers, and include a margin for adverse deviation in my range? Are you adjusting for bias? Or adding bias? Proposed GAAP – “this framework does not include prudence or conservatism as desirable qualities of financial reporting information. ” Current GAAP thinking clearly favors neutrality (lack of bias) over conservatism 2006 CLRS What is a “Reasonable” Range – R. Blanchard 16

Uncertainty, Conservatism & Bias – part 2 Existing U. S. Stat. (NAIC’s Emerging Accounting Issues Working Group – INT 01 -28: Margin for Adverse Deviation in Claim Reserve) “The working group reached a consensus that the concept of conservatism is inherent to the estimation of reserves and as such should not be specifically prohibited in the consideration of management’s best estimate. On the other hand, the working group does not believe there should be a specific requirement to include a provision for adverse deviation in claims as the application of estimates varies greatly from company to company and requires the careful judgement of management. ” Conclusion: GAAP intends to prohibit any “bias”, Stat. allows but doesn’t require upward bias. 2006 CLRS What is a “Reasonable” Range – R. Blanchard 17