Canadian Individual Payout Mortality Table

1
Canadian Individual Payout Mortality Table

• IP-6

2
A new table

• Why?
• What?
• So what?
• What next?
• How?

3
Why a new table?

• No Canadian table
• Improvement varies by age
• CIA Study since 1980s

4
What is the table like?

• Male Female
• Ages 0-115
• Insurance mortality to age 55
• Flat at high end

5
CIP2005
6
Male
7
Male 60-100
8
Female
9
Female 60-100
10
Annuities Male _at_ 5

• Age 83B 2005 CIP2005
• 60 12.79 12.89
• 65 11.36 11.49
• 70 9.79 9.93
• 75 8.09 8.22
• 80 6.37 6.40

11
Annuities Female _at_ 5

• Age 83B 2005 CIP2005
• 60 13.65 13.80
• 65 12.31 12.48
• 70 10.79 11.03
• 75 9.08 9.36
• 80 7.25 7.48

12
So what do I do now?

• Pricing
• Valuation

13
Pricing Issues

• Small but significant difference
• Represents industry experience
• Lower mortality for higher amounts
• Review pricing assumption

14
Valuation Issues

• Not a valuation table
• MfAD in base or projection?

15
PfAD vs MfAD
16
PfAD vs MfAD
17
PfAD vs MfAD
18
PfAD vs MfAD
19
PfAD vs MfAD
20
PfAD(t)/PfAD(0)
21
PfAD(t) / PfAD(0)
22
PfAD vs MfAD
23
What next for the table?

• Projection scale
• Add more data
• Some changes in method
• Your input

24
How was the table constructed?

• Technical part
• Described in paper
• Consider fundamentals
• Consider turning points

25
Technical Agenda

• Data
• Graduation Method
• Graduation Parameters
• Young and Old
• Improvement

26
Data

• 13 years
• 80
• Reconciled by policy
• Single / Joint / Survivor

27
Exposure
28
Deaths
29
Graduation Method

• Smooth over fluctuations
• Whittaker-Henderson

30
Graduation Formula

• Minimize

31
Graduation Method

• Smooth over fluctuations
• Whittaker-Henderson
• Same deaths, age

32
Graduation Method

• Smooth over fluctuations
• Whittaker-Henderson
• Same deaths, age
• Limited age range

33
Graduation Parameters

• Order of difference
• Weights
• h

34
Playing with h in Excel
35
Younger Ages

• Too little data
• Use a proxy
• 75 of CA 86-92 Ultimate

36
Older Ages

• Too little data
• Traditional approach
• Flattened curve
• Cubic to flat

37
Improvement

• Data centred on 1995
• Improvement to 2005
• CPP/QPP data
• Not IAMS data
• Warning not for general use

38
References

• My paper on CIA web site (soon)
• See www.howardfamily.ca/bob
• Or email bob_at_howardfamily.ca

39
(No Transcript)