Tsa 34 Cumulative Antiselection Theory (Cast) In Ihi

Front 1

The Cumulative Antiselection effect is more common in HI and DI than in LI, because:

Back 1

 ph’s know their claims probability
 elective nature
 higher lapse rates
 b/c of short-term contracts

Front 2

Problems with the Classical Loss Ratio

Back 2

 LR deterioriates
 LR projections judgment-based
 LR’s distorted by: Seasonality; Reserving method
 Upward Rate Spiral:
 failure to notice an LR trend
 Inflation weights later claim costs more
 premiums must be raised
 Antiselective lapses  claim costs rise
 premiums must be raised again

 But: Advantages of LR:
 single, simple number
 widely accepted
 used by regulators

Front 3

Problems with Classical Select-and-Ultimate Reserving Methods

Back 3

 account for UW wear-off, but not for antiselective lapses

Front 4

THE GOALS OF CAST THEORY

Back 4

 To mathematically predict antiselective lapses
 e.g. on a specific premium increase
 To calculate a “Durational Antiselection Reserve”, that stabilizes LR
 To determine a better initial premium
 To keep LR as a single, simple number

Front 5

THE CAST MODEL
 The CAST model is a stochastic model that follows a cohort of lives beginning at Time 0.

Assumptions and Parameters:

Back 5

 two types of lives, “Active Healthy” and “Impaired”.
 Perfect UW means all lives begin Healthy.
 Transition probability of healthyimpaired.
 net of recoveries
 Impaired claim costs > Healthy claim costs
 Impaired lapse rate < Healthy lapse rate
 Deaths are lapses
 Interest and Inflation are ignored.

See numerical example and mathematical notation in chapter notes

Front 6

The Effects of a Premium Increase

Back 6

The CAST model successfully predicts the following:
1. a one-time increase in lapsation.
2. higher cost per remaining policy
3. Lapses release some reserves,
4. but remaining reserves must be raised
5. LR increases.

Front 7

CAST Theory Differs from Classic Select-and-Ultimate Theory in that:

Back 7

 There is no “ultimate”
 two sets of lives: healthy and impaired.
 includes antiselective lapses as well as UW wearoff.
 CAST LR’s increase exponentially; classical’s more level
 recommends larger net terminal reserves.
 recommends a higher initial premium

Front 8

Problems with the CAST Model

Back 8

Nonrealities
 constant select-and-ultimate multiplier
 Interest and inflation are ignored.

Implementation Problems
 Regulators’ conversion would be difficult
 Pay-as-you-go systems are more favored
 higher initial premiums make a company uncompetitive, whether it is “correct” or not.
 How would the “Durational Antiselection Reserve” be taxed? Accounting?

Front 9

Recommendations for Implementing the CAST Model:

Back 9

 a YRT pricing approach
 make HI into a health form of Universal Life
 nonstandard renewal actions may be req’d
 develop computer systems to track loss by duration.

(See numerical examples in chapter notes.)


Done.