Joint AAA-SOA Preferred Mortality Study Project Update Larry - - PowerPoint PPT Presentation

joint aaa soa preferred mortality study project update
SMART_READER_LITE
LIVE PREVIEW

Joint AAA-SOA Preferred Mortality Study Project Update Larry - - PowerPoint PPT Presentation

Joint AAA-SOA Preferred Mortality Study Project Update Larry Gorski, FSA, MAAA Vice-Chair, AAA Life Practice Council Chair, LCAS Economic Scenario Subgroup Consulting Actuary, Claire Thinking Inc American Academy of Actuaries LHATF, June


slide-1
SLIDE 1

American Academy of Actuaries LHATF, June 2006 1

Larry Gorski, FSA, MAAA

Vice-Chair, AAA Life Practice Council Chair, LCAS Economic Scenario Subgroup Consulting Actuary, Claire Thinking Inc

Joint AAA-SOA Preferred Mortality Study Project Update

slide-2
SLIDE 2

American Academy of Actuaries LHATF, June 2006 2

Where it fits in the process?

  • Data Validation
  • Underwriting Criteria
  • Experience Analysis
  • Valuation Basic Table
  • Valuation Table Implementation
  • Valuation Table
slide-3
SLIDE 3

American Academy of Actuaries LHATF, June 2006 3

Underwriting Criteria Team

  • Chaired by Al Klein (Tillinghast)
slide-4
SLIDE 4

American Academy of Actuaries LHATF, June 2006 4

Background

  • Goal of UCT is to create an algorithm to

group together like preferred risk classes by underwriting criteria in order to help create a preferred mortality table with multiple risk classes based on actual experience

– Still considerable differences among companies

  • Team consists of actuaries, underwriters and

SOA staff

slide-5
SLIDE 5

American Academy of Actuaries LHATF, June 2006 5

Activity to Date

  • Initial meeting — October 27-28, 2005

– Designed algorithm

  • Preliminary scoring meeting — April 10-11,

2006

– Fine tuned algorithm and began scoring process

  • Final scoring meeting — May 17-18, 2006

– Completed scoring process

slide-6
SLIDE 6

American Academy of Actuaries LHATF, June 2006 6

Basic Framework for Algorithm

  • Identify criteria commonly used in the

preferred underwriting decision

  • Determine its relative importance

among all of the criteria

  • Determine how to score each individual

criterion

slide-7
SLIDE 7

American Academy of Actuaries LHATF, June 2006 7

Underwriting Criteria considered in the scoring algorithm

  • Cholesterol
  • Personal medical history
  • Alcohol/drug use
  • Build
  • Blood pressure
  • Family history
  • Driving
  • Tobacco use
  • Aviation, avocations, citizenship, foreign travel,

hazardous activities, residence

slide-8
SLIDE 8

American Academy of Actuaries LHATF, June 2006 8

Valuation Basic Table Team

  • Chaired by Doug Doll (Tillinghast)
  • Vice-chair – Mary Bahna-Nolan (North

American Company for Life & Health)

slide-9
SLIDE 9

American Academy of Actuaries LHATF, June 2006 9

Background

  • Charged with developing the new

valuation basic table to be the industry mortality experience table with multiple preferred risk classifications

slide-10
SLIDE 10

American Academy of Actuaries LHATF, June 2006 10

Issues to Address

  • Data quality

– Not all companies provided underwriting criteria – Mortality results may not match guidelines due to underwriting exceptions

  • Term mortality after level premium period

– Not certain can identify – Can we/should we ignore anti-selective mortality from experience

slide-11
SLIDE 11

American Academy of Actuaries LHATF, June 2006 11

Issues to Address (cont’d)

  • Term Conversions

– Higher mortality due to anti-selection – Included in permanent experience – Should this be included

  • are we overstating mortality for products that do not

allow conversion?

  • Term v. Permanent

– Much of preferred experience driven from term plans – Are there significant differences between permanent and term mortality?

slide-12
SLIDE 12

American Academy of Actuaries LHATF, June 2006 12

Issues to Address (cont’d)

  • Number of Classes

– AAA’s LRWG wants sufficient number of tables to accurately reflect industry experience for most super preferred to standard tobacco risks (e.g., 10 tables) – Do we have sufficient credibility to dissect data so finely

  • May have to happen over time

– Blended mortality may not add back to aggregate mortality table