Upload
jerome-cameron
View
219
Download
0
Embed Size (px)
DESCRIPTION
Headlines n People are living longer nMore people survive to retirement nPeople then tending to live longer n Most improvement for males n PA90 tables have the wrong shape
Citation preview
The Younger Members Convention
2-3 December 2002, Warrington
Who wants to live forever? Impact of mortality improvements
James DaviesHewitt Bacon & Woodrow
Agenda
Headlines More detail on recent mortality experiences Implications / effects of changes Experience of self administered pension schemes Future changes Questions / discussion
Headlines
People are living longer More people survive to retirement People then tending to live longer
Most improvement for males PA90 tables have the wrong shape
Changes in mortality of male assured lives
0
50
100
150
200
250
300
350
400
17 22 27 32 37 42 47 52 57 62 67 72 77 82 87 92 97
Age
Perc
enta
ge o
f 196
7-70
1924-29
1949-52
1967-70
1991-941979-82
More detail on recent mortality experiences
CMIB – some background
Been collating data since 1955 Data collected by Amounts and by Lives Pensioners of life office pension schemes (DB & DC) Includes most of the major insurance companies Lots of data – will it reduce in the future?
Recent history of tables
a(55) – based on 1949-52 data (projected) PA(90) – 1967-70 data projected to 1990 P?A80 – 1979-82 data, two-way table P?A92 – 1991-94 data, two-way table
100A/E against PA(90) – all ages
40
60
80
100
120
1964 1968 1972 1976 1980 1984 1988 1992 1996
100
A/E
MAFAMLFL
100A/E against PA(90) – all ages
40
60
80
100
120
1964 1968 1972 1976 1980 1984 1988 1992 1996
100
A/E
MAFA
100A/E against PA(90) – MA – various age groups
40
60
80
100
120
1964 1968 1972 1976 1980 1984 1988 1992 1996
100
A/E
66-7076-8086-90All ages
100A/E against PA(90) – FA – various age groups
40
60
80
100
120
1964 1968 1972 1976 1980 1984 1988 1992 1996
100
A/E
61-6571-7581-85All ages
100A/E against 80 series – recent experience only
70
80
90
100
110
1983 1985 1987 1989 1991 1993 1995 1997
100
A/E
MAMLFAFL
100A/E against 92 series – recent experience only
80
90
100
110
1992 1993 1994 1995 1996 1997 1998
100
A/E
MAMLFAFL
Ratio of Pxx92C1992 (the base values) to PA(90)-2
0.2
0.4
0.6
0.8
1
1.2
40 50 60 70 80 90 100 110
Rat
io
Female AmountsMale Amounts
Criticisms of 92 series
Don’t know enough about “why” No allowance for “cohort” effect
Group born 1925-45 Biggest improvements in morality rates Not adjusted for => may overstate future improvements
Wider uncertainty about future morality improvements
Implications / effects of changes
The cost of living
People are living longer Scheme are paying out pensions to more members Pensions are paid for longer So pensions cost more (or you can buy less pension if DC)
Valuation strains Deficits in DB schemes Greater uncertainty for DC members?
Life expectation for male aged 60
17
19
21
23
25
27
29
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 2022 2026 2030
Exp
ecta
tion
of li
fe (y
ears
)
1968 data1980 data1992 data
Life expectation for female aged 60
17
19
21
23
25
27
29
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 2022 2026 2030
Exp
ecta
tion
of li
fe (
year
s)
1968 data1980 data1992 data
Life expectation for female (and male) aged 60
17
19
21
23
25
27
29
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 2022 2026 2030
Exp
ecta
tion
of li
fe (
year
s)
1968 data1980 data1992 data
Life expectancy - males
0
5
10
15
20
25
30
35
40
50 60 70 80 90 100
Exp
ecta
tion
of li
fe (y
ears
)
PA90PA90 -2PA90 -4PA92C2000PA92C2020
Life expectancy - females
0
5
10
15
20
25
30
35
40
50 60 70 80 90 100
Exp
ecta
tion
of li
fe (y
ears
)
PA90PA90 -2PA90 -4PA92C2000PA92C2020
Life expectancy – ratio to PA90 - males
1
1.1
1.2
1.3
1.4
1.5
50 60 70 80 90 100
Rat
io
PA90 PA90 -2PA90 -4 PA92C2000PA92C2020
Life expectancy – ratio to PA90 - females
1
1.1
1.2
1.3
1.4
1.5
50 60 70 80 90 100
Rat
io
PA90PA90 -2PA90 -4PA92C2000PA92C2020
Examples of changes in life expectancy
Male aged 65 PMA80 says survive for just over 15 years PMA92 says survive for just under 20 years
Male aged 35 1980 data says survive until 80y 1m 1999 data says survive until 85y 1m
Effect on pension schemes
Value of liabilities increases Past service deficits Higher cost of future service benefits Impact depends on where you start from and what
table you want to use for future
What tables to use …
Should actuaries use two-way tables? Scheme or industry experience Is life office data totally relevant ? Calendar year vs year of birth Imprudent to ignore future improvements? Allowance for future change is an imprecise science
Eg effect of future medical advances
Ratio of joint life annuities to PA90M-4 (3% interest)
0.8
0.9
1
1.1
1.2
50 55 60 65 70 75 80 85 90 95 100
Rat
io
PA90M-2 PMA92C2000PMA92C2020 PMA92B1945
Ratio of joint life annuities to PA90F-4 (3% interest)
0.8
0.9
1
1.1
1.2
50 55 60 65 70 75 80 85 90 95 100
Rat
io
PA90F-2 PFA92C2000PFA92C2020 PFA92B1945
Actives – past service and future service
Example of impact on actives for a scheme Effect not necessarily obvious => care needed when
choosing assumptions Sensitive to a number of factors
PA65PS FS (AA)
92C2000 -0.2% -0.8%
92C2020 +6.0% +5.9%
92B1945 +8.9% +8.8%
Actives – past service and future service
Example of impact on actives for a scheme Effect not necessarily obvious => care needed when
choosing assumptions Sensitive to a number of factors
PA65 PA60PS FS (AA) PS FS (AA)
92C2000 -0.2% -0.8% +0.8% +0.5%
92C2020 +6.0% +5.9% +6.5% +6.5%
92B1945 +8.9% +8.8% +10.1% +8.5%
Mortality is not the whole story …
Falling yields have had a big impact … £100 pa pension (male, LPI, 50% spouse’s pension)
Approx cost in 1985 was £1,000 (’80’ tables) Cost risen to £1,500 by 2001 due to yields falling Rises further to £1,800 when use ’90’ tables
Pension scheme liabilities have risen even more than just the amount due to mortality
Experience of self administered pension schemes
CMIR number 20
Request from Pensions Board Pilot investigation of mortality of pensioner of self
administered schemes Comparison to standard tables
Data
From 7 offices, 13 schemes (one unusable) Various formats – required some work by CMIB No consistently usable information on industry / type Included ill health retirements? “… the data could have been better”
PROFESSION
Findings
Males – experience between 80 and 92 series Shape different to PA90
A/E < 100 at younger ages A/E > 100 at higher ages
Females – experience heavier than 80 and 92 series Amounts vs Lives effect similar to main investigation Results at younger ages suggest data included IH
… but removing under age 56 has little effect on A/Es Broadly similar results for all 12 schemes used
100A/E for males, amounts
Age group PA(90) –2 PMA80C1997 PMA92C1997
46-50 287 350 896
51-55 73 98 224
56-60 69 109 201
61-65 60 88 137
66-70 73 91 124
71-75 86 94 117
76-80 94 93 109
81-85 102 95 106
86-90 105 95 103
91-95 105 97 103
96-100 106 100 107
101-105 109 104 116
56-105 88 94 113All ages 88 94 114
100A/E for males, amounts – by age group
40
60
80
100
120
140
46-50 56-60 66-70 76-80 86-90 96-100
100A
/E
PA(90) –2PMA80C1997PMA92C1997
100A/E for males, amounts – by age group and total
40
60
80
100
120
140
46-50 56-60 66-70 76-80 86-90 96-100
100A
/E
PA(90) –2PMA80C1997PMA92C1997
Broad conclusions
Data could be better Experience not same as life offices Shape similar to that seen in main investigation Males improving more than females
Future changes
Where will it end?
With the knowledge that is accumulating now… we could add 30 years to human life in the next
decade Dr William Regelson, Professor of Medicine, Virginia
Life expectation – cost of 10% improvement (males)
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
8.00%
9.00%
50 55 60 65 70 75 80 85 90
Incr
ease
in a
nnui
ty (
with
spo
use'
s pe
nsio
n) PA90M-2PMA80C2000PMA92C2000
Life expectation – cost of 10% improvement (females)
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
8.00%
9.00%
50 55 60 65 70 75 80 85 90
Incr
ease
in a
nnui
ty (
with
spo
use'
s pe
nsio
n) PA90F-2PFA80C2000PFA92C2000
Consequences
DB schemes unaffordable in current form Are there solutions?
Lower benefits Fewer early retirements on generous terms Increase NPA Increase contributions
Companies more nervous due to increased risk
Questions / Discussion