Reserve Variability Working Party Kickoff

1
Reserve Variability Working Party Kickoff

• Phil Heckman
• Aon Risk Consultants
• Roger Hayne
• Milliman USA
• CLRS 2003

2
Agenda

• CAS Centennial Goal (CG)
• CAS Research Overhaul
• Working Party brainstorming

3
CAS CG

• The CAS will be globally recognized as a
  pre-eminent resource in educating casualty
  actuaries and conducting research in casualty
  actuarial science. CAS members will be recognized
  as leading experts in the evaluation of hazard
  risk and the integration of hazard risk with
  strategic, financial and operational risk.

4
CAS CG Implications

• pre-eminent resource in conducting research
  in casualty actuarial science.
• leading experts in the evaluation of hazard
  risk and the integration of hazard risk with
  strategic, financial and operational risk.

5
Problem Statements

• No keepers of the state of the science
• Need for survey papers, syllabus material
• Research overload via Call Papers
• Role/function of the PCAS unclear

6
Proposed Solutions

• De-emphasize Call Paper programs
• Establish Working Paper and Model repository on
  the CAS Website
• Evaluate CAS Publications
• Develop a CAS Research Taxonomy
• Establish Research Corners and Working Party
  sessions at the major seminars
• Institute Working Parties

7
Whither Call Papers?

• Bottom-up, fast-track research source.
• Stimulate communication, discussion and sharing.
• Good in concept, in practice has been another
  story.

8
Call Papers

• Not a professional journal
• Not peer reviewed.
• Inclusive editorial policy.
• Inconsistent review and prize standards.
• Inconsistent appearance and structure of papers.
• Contributes to members filtration and overload
  problems.

9
Call Papers

• Not generating discussion
• Solitary practitioners produce, present
• No context, follow-up, formal discussion
• Not leading to systematic progress of the science
• No referencing standards, context
• No clear advancement of the science

10
Call Papers ? Working Papers

• CPs are the equivalent of working papers within
  academia
• Posted on websites and discussion forums
• Works-in-progress, on their way to peer-reviewed
  journals
• We can still have bottom-up idea generation, idea
  sharing, and discussion by establishing a Working
  Paper (and Model) repository on the CAS Website

11
Working Paper Repository

• Categorized by research area
• Members can post and comment on posts
  (mini-reviews)
• Items receiving a lot of activity can be the
  material for the Research Corners at the major
  seminars

12
Publications Task Force

• Impact and notoriety of PCAS outside the CAS is
  NIL
• Forum is used / abused
• Large bodies of work published without formal
  peer review
• Considering some radical surgery
• Maybe we join the NAAJ

13
Research Taxonomy

• Categorization scheme for our research
• Should dramatically improve searches
• Precursor to having effective referencing
  standards
• Under development by the Research Chairs, with
  the assistance of Erin Clougherty and Mike Boa of
  the CAS Office

14
Working Parties

• Essentially a collective call paper task force
• Collective group effort, single group work
  product
• Ideas come from the members attending major
  seminar
• Seminar has presentation of prior years work,
  selection of next years topics

15
Working Parties

• Group effort forces discussion during the
  production of the product
• Oversight by research committee
• Can enforce editorial standards, referencing,
  ensure that current state of the science is
  documented, as well as context and scope of new
  research

16
Working Parties

• An easy to implement answer that helps on many
  fronts
• Solitary ? Group
• Bottom-up Top-down
• Consistency in format, referencing, etc.
• Member involvement
• Natural seminar cycle supports it

17
Working Parties

• Be prepared for some differences
• Progress will be made
• Ties will be broken
• Oversight by RD committees will be managed
  consensus
• Stronger editorial hand

18
Working Parties

• Piloted at 2003 RCM Seminar
• Work product to be presented at 2004 RCM Seminar
• Involves two proactive research committees
  (Reserves, COTOR)
• Also to be introduced Ratemaking (March 2004)

19
Reserve Variability WP

• Perennial problem solve piecewise
• Most techniques kill variability
• Not enough info in a set of triangles
• Need to learn to combine data from many sources
• Need to think of loss development as a stochastic
  process, increments not independent

20
What does cohort agg. dist. look like?
21
Cohort Aggregate Losses

• Distribution will have a probability mass for no
  further payments that becomes large in the tail.
• Severity stays up, frequency decreases with age
  of cohort.
• Need to consider this to fit tail
• Here ends my 2 worth.