managing longevity risk

8/14/2019 Managing Longevity Risk

1/30

Younger Members Convention, Birmingham

Managing longevity risk

Stephen Richards4th December 2007

Copyright c Stephen Richards. All rights reserved. Electronic versions of this and other freelyavailable papers and presentations can be found at www.richardsconsulting.co.uk


2/30

Managing longevity risk plan of talk1. Investing in heart disease2. Capital-market solutions3. Socio-economic group4. Bespoke risk models5. Summary and questions

Slide 1 www.richardsconsulting.co.uk


3/30

1. Investing in heart disease



4/30

Reliance of pension sponsor on CHD

Cause 4564 6584 85+

Infectious diseases 6 30 160Injury and poisoning 37 62 330Respiratory diseases 45 587 3,882Cancers 240 1,382 3,394Circulatory diseases 225 1,784 7,813All causes 644 4,316 18,803

Source: Selected causes of death for males, 2003, ONS data for England and Wales.



5/30

20th century mortality

1920 1940 1960 1980 2000 2020

0

200

400

600

800

Year

M o r t a

l i t y p e r

1 0 0

, 0 0 0

qq q

q q q q q q qq q q

q q q qq q q q q q q q q q q q q q q q q q q

q q q q q q q q q q qq q q q q q q q

q q q q q q q qq q q q q q q q q q

q q q q q q q q qq q q q q q q q q q

q

q qq

q

qq

q

q

q

q

q

q

q

q

q

q

q

q

q

q

q

q

qq q

q

qq

q

q

q

q

qq qq

q

q

q

q

q

q

qq q q q

q

q

q qq

q q

q

q

q q q

qq q q q

q

q qq q q q q

qq q

q qq

q q q

qq

q q q qq

q

q q

q

q

Circulatory diseasesCancerRespiratory diseasesInfectious diseases

Source: ONS data



6/30

Italian mortality, 5074 (males and females)

1980 1985 1990 1995 2000

0

100

200

300

400

500

600

700

Year

D e a

t h s p e r

1 0 0

, 0 0 0

q q qq

q q q q qq q q q

qq

q qq q

qq q

q

q q q q qq q q q q q q q q q q q q q q q q q

q

q

CirculatoryCancerLung cancerOther

Source: Cocevar, 2007



7/30

Rise and fall of heart-disease mortality

Year

R a

t e p e r

1 0 0

, 0 0 0

1850 1900 1950 2000

0

100

200

300

400

500

600

700

CHD mortality rateInfluenza mortality rate

Source: Azambuja and Levins (2007). Data from Massachusetts Department of Public HealthRegistry of Vital Records and Statistics.



8/30

Adequacy of improvement basis

q

q

qq q

q q q q q q q qq

q

q

q

q

q

q

q

qq

Age

A n n u a

l i m p r o v e m e n

t

50 60 70 80 90 100 110

0%

1%

2%

3%

4%

5%

q

q

q

q

q

q

q

q

q

q

q

q

q

q

qq

qq q q q q q q q q q q q q q q q q q q q q

q

q

Smoothed actual improvementsMediumcohort projectionPspline projection2.5th percentile97.5th percentile

Source: Richards et al (2007)



9/30

2. Capital-market solutions



10/30

Basis risk

Scheme Members Concentration

E 40 11%H 800 12%C 5,300 6%

Largest scheme (C) pays 50% of all pensions to just 6% of members.

Source: Richards Consulting calculations using Prudential data.

Concentration is the percentage of members accounting for half of all pensions in payment.



11/30

Concentration of income by membership decile

1 2 3 45

6

7

8

910

7.0%

15.1%62.2%

Source: Richards Consulting calculations using Prudential data for Scheme C.



12/30

Concentration of risk Gini co-efficient measures income inequality 36.8% for UK, one of the more unequal countries

70.8% for Scheme C

Source: mortalityrating.com



13/30

Current scheme options Buy-out: total defeasance of legal and economic risks LDI: some hedging of economic risks, none of longevity or legal risk Securitization Options and reinsurance Derivatives and indices



14/30

Risks with buy-out Individual buy-out policies upon wind-up are safe Buy-out policies as ongoing scheme asset still risky Uberima des applies to scheme, not insurer



15/30

Options and reinsurance

Scheme Members Option cost

E 40 7.7%H 800 2.6%C 5,300 0.6%

Source: Richards Consulting calculations using Prudential data.Option cost is percentage of reserves calculated at 5% interest using 90% PNxA00 with medium-cohort improvements with a 1.5% p.a. oor. Pensions assumed escalating at 2.5% per annum.



16/30

Securitization Transfer as much economic and longevity risk as you want Trustee-friendly as AAA-rated

Enhances employer covenant Investor-friendly as limited longevity risk Not insurance, so lower capital requirements



17/30

Legal wrinkles with securitization LDI leaves scheme with all legal and economic risks Insurers may not honour scheme-asset buy-outs in extreme cases Securitization not yet recognized in PPF process for employer insol-vency



18/30

Derivatives and indices Index-based contracts based on population statistics Examples: LifeMetrics (JPMorgan) and Credit Suisse

Substantial basis risk Partial coverage (e.g. 30%) often most efficient



19/30

3. Socio-economic group


l f b


20/30

Retirement life expectancy by socio-economic group

M a

l e l i f e e x p e c

t a n c y a

t a g e

6 5 i n y e a r s

I

II

IIIN

IIIM

IV

V

197276 197781 198286 198791 199296 199799

11

12

13

14

15

16

17

18

11

12

13

14

15

16

17

18

Source: ONS Longitudinal Survey.


Fi i l i f lif l


21/30

Financial impact of lifestyleFinancial impact of mortality rating factors

Factor Step change Reserve Change

Base case - 13.39 -Gender Female-male 12.14 -9.3%

Lifestyle Top-bottom 10.94 -9.9%Duration Short-long 9.88 -9.7%

Pension size Large-small 9.36 -5.2%Region South-North 8.90 -4.9%Overall - - -33.6%

Source: Richards and Jones (2004), page 39.


Wh f d i i li bl


22/30

Why fund size is not reliable Stakeholder fund of 8,583 Poor? Higher-mortality group? But AVC fund elsewhere of 42,808. . . . . .giving total fund of 51,391. . .

. . .so not poor and likely light mortality!


H t t d t d ti


23/30

How not to do postcode rating Beware any solution based on regional or population data! Test: EH4 2AB v. EH11 2AS Rating the two the same is simply wrong


4 B k i k d l


24/30

4. Bespoke risk models


Limit tions of postcode onl sol tion


25/30

Limitations of postcode-only solution Scheme has credible data: 9,000 lives A/E suggests 150% PCMA00 Postcode model suggests 110% PCMA00 Which is correct?


Li it ti f A/E l i


26/30

Limitations of A/E analysis Vulnerable to shifts in membership prole Barber judgement De-industrialization


Survival models


27/30

Survival models Statistical models, so better than A/E calculations Full and efficient use of data, so better than GLMs

Can handle age, gender, postcode, pension size . . . also cohort, region, time trend and duration Models can be tted weighted by pension size or reserve Bootstrapping can check nancial applicability


5 Summary and questions


28/30

5. Summary and questions


5 Summary and questions


29/30

5. Summary and questions Strong pattern of mortality improvements likely to continue Medium cohort inadequate for best-estimate purposes

Capital-market alternatives to buy-out Postcode rating for socio-economic mix Survival models best for bespoke risk analysis


References


30/30

ReferencesLongevitas 2006 Modelling pensioner mortality , www.longevitas.co.ukCocevar, P. 2007 An analysis of recent mortality trends in the Italian population using penalised B-spline regression , submitted paperRichards, S. J. and Jones, G. L. 2004 Financial aspects of longevity risk , SIASRichards, S. J., Ellam, J. R., Hubbard, J., Lu, J. L. C., Makin,S. J. and Miller, K. A. 2007 Two-Dimensional Mortality Data: Pat-terns and Projections , British Actuarial Journal (to appear)


managing longevity risk

Documents