Mortality improvements and Longevity Risk. Javier H Osorio. Senior Marketing Actuary L&H LatAm Swiss Re

Transcription

1 Mortality improvements and Longevity Risk Javier H Osorio. Senior Marketing Actuary L&H LatAm Swiss Re

2 Agenda Mortality improvements Casual drivers of Mortality Improvements Understanding longevity risk A possible longevity risk solution Final Comments 2

3 Mortality Improvements / (Setting the scene) 3

4 Longevity risk is a mixture of uncertainty about the present Estimation of Current Rate of Mortality Predictive model Experience analysis Cohort life expectancy at 65 for male Social-Security covered workers by birth years and earnings group(1) Base table estimation(2) Earnings in top-half of distribution Best Estimate c 17% probability Earnings in bottom-half of distribution % -70% -60% -50% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50% 60% 70% 80% Birth year Difference of 5.5 years between top-half and bottomhalf of earnings distribution % difference from best estimate base table With a fund of 10,000 pensioners there is a 1 in 6 chance that the base table overstates current mortality rates by 10% or more Even with detailed analysis, there remains statistical uncertainty around the current rate of mortality (1) US Social Security Administration Office of Policy; ORES Working Paper No. 108, October 2007 (2) Swiss Re Calculations 4

5 and uncertainty about the future Predicting the future, remaining Period Life Expectancy at age 65 for US males: Observed or projected by the SSA Study 120 published 2005 Study 116 published 2002 Study 107 published 1992 Study 89 published 1983 Study 72 published 1974 Study 62 published 1966 high-low average Study 46 published 1957 high-low average Study 33 published 1952 high-low average Remaining years of life 18 Even if confidence were to be gained regarding the base table, the rate of future improvement is even more uncertain Source: US Social Security Administration Office of Policy, Literature Review of Long-Term Mortality Projections,

6 Longest-lived life expectancy increases at 3 months per year for the last 150 years Record female life expectancy from 1840 to present Horizontal black lines show asserted ceilings on life expectancy, with a short vertical line indicating the year of publication Dashed red lines denote projections of female life expectancy in Japan published by the United Nations in 1986, 1999, and 2001 Even if confidence were to be gained regarding the base table, the rate of future improvement is even more uncertain Source: Oeppen & Vaupel, broken bounds to life expectancy, Science

7 Observed Mortality improvements Chile, Males Source: HMD data, Swiss Re Calculation 7

8 Observed Mortality improvements Chile, Females Source: HMD data, Swiss Re Calculation 8

9 Causal Drivers of Mortality Improvements 9

10 Recent high improvement has been heavily driven by the improvement of circulatory diseases Source: WHO database, HMD data, Swiss Re calculation 10

11 Improvements driven by reduction in deaths from circulatory and lung cancer Mortality rates - USA Males (aged 50-84) Circulatory Lung Cancer 15 Other Causes Source: Centers for Disease Control and Prevention, National Center for Health Statistics. Compressed Mortality File. Human Mortality database. Swiss Re calculation 11

12 Causes of death vary by age Source: WHO database, ICD10 file. Argentina, Brazil, Chile, Colombia, Mexico, Peru and Venezuela included. 12

13 Smoking Trends Smoking prevalence rate is decreasing Source: OECD Health Data

14 Obesity Trends Obesity rate in increasing Source: OECD Health Data

15 Difference in Gender 15

16 Difference in Improvements by Gender Canada Patterns tend to be quite different males more prone to periods of high improvements Source : HMD data, Swiss Re Calculation 16

17 Difference in Improvements by Gender Switzerland Source : HMD data, Swiss Re Calculation 17

18 Income, Health and Mortality 18

19 Health Status by SES Class age Better Socio-class correspond with lower disease burden Source : GPRD data, Swiss Re Calculation 19

20 Mortality as a function of pension size Source : CMI Working Paper 65 20

21 Mortality at advanced ages 21

22 Health Status by SES Class age At higher ages multiple disease burden dampens ability to intervene medically Source : GPRD data, Swiss Re Calculation 22

23 UK historical mortality improvement by gender and age band Historically lower improvements at advanced ages Source : HMD data, Swiss Re Calculation 23

24 Improvements are significant at post-retirement ages Male Mortality improvements by age 5% 4% 3% 2% 1% 0% % -2% US England & Wales Japan Source: lifemetrics 24

25 Convergence of Insured and Population tables? Generally arbitrary grading at high ages general expectation of convergence to population at highest ages Source : ONS and CMI data, Swiss Re Calculation 25

26 Understanding Longevity Risk 26

27 The holders of longevity risk Government State pension; public sector pensions; long-term and medical care Longevity risk transfer solutions Reduced benefits Individual Risk of outliving their assets Insurance industry Annuity portfolios Long-term care Longevity risk transfer solutions Closure of defined benefit schemes Employer Defined benefit pensions; retiree medical benefits 27

28 The global longevity risk issue is huge Global longevity exposure estimated as approx USD 20 trillion of pension assets 90% of exposure related to pension funds;10% pension insurance contracts Insurers exposure dominated by markets with compulsory annuitisation Corporate sponsors exposed to defined benefit pension schemes Governments committed to fund old-age pension/health benefits Source: Organisation for Economic Co-operation and Development (OECD) Global Pension Statistics 28

29 Risk Management Strategies Understand Limit Diversify Hedge Exit Identify risk drivers Quantify exposure Analyse risk scenarios Restrict sales/new entrants Redesign products/benefits Volume, markets Other product lines Internal Negative correlated risks External Re/insurance, Capital Markets Sale, Buy-Out Commutation 29

30 Historically, longevity risk has been underestimated Actual and projected life expectancy at birth, UK males 30

31 Mortality Trend Modelling Statistical/Stochastic models e.g. Lee-Carter models. ARIMA process on observed trends Actuarial Models e.g. CMI model, grade to assumed long term assumptions Medical Models e.g. Assumed changes in disease diagnosis and prognosis 31

32 Netherlands improvements % -4.0 % 2.0 % -3.0 % 1.0 % -2.0 % 0.0 % -1.0 % -1.0 % -0.0 % -2.0 % % -3.0 % % Source: Centraal Bureau voor de Statistiek data, Swiss Re calculation 32

33 NL improvements Projected to %-4.0% 2.0%-3.0% 1.0%-2.0% 0.0%-1.0% -1.0%-0.0% -2.0%--1.0% -3.0%--2.0% Source: Centraal Bureau voor de Statistiek data, Swiss Re calculation Projection from 1992, Lee Carter Model with cohort term (M3 in Lifemetrics) 33

34 NL improvements Actual %-7.0% 5.0%-6.0% 4.0%-5.0% 3.0%-4.0% 2.0%-3.0% 1.0%-2.0% 0.0%-1.0% -1.0%-0.0% -2.0%--1.0% -3.0%--2.0% Source: Centraal Bureau voor de Statistiek data, Swiss Re calculation 34

35 Statistical methods Projection LeeCarter Fit Males ( ) Source : ONS data, Swiss Re Calculation 35

36 Statistical methods CMI Projection UK Males Source : CMI 36

37 Forward looking scenario modelling Integrated Risk factors and Impairment Scenarios Complementary approach to stochastic mortality models blending between current mortality improvements and long-term assumptions over defined horizons Bringing together: Swiss Re experience (data and expert knowledge) Large patient medical databases in different countries External networks capturing expert opinion Causal-based mortality predictions, evaluating factors such as: Promotion and adoption of healthy lifestyle choices Advances in screening and diagnostic technology Pharmaceutical pipeline and its likely impact 37

38 IRIS forward-looking scenarios Different layers in disease-centred mortality model General drivers to diagnosis and survival Individual risk factors Age, gender, diet, smoking smoking considerations: Taxes and restrictions Current treatments (buproprion) Future treatments (vaccines) Healthcare funding Public vs private funding Disease-based patient advocacy groups' influence Allocation of resources towards cure vs prevention Circulatory Stroke, angina, heart attack Disease types and disease progression Healthy Risk factors Family history Obesity Having children later in life Not breast feeding Factors involved in assessing specific example disease Cancer Lung, colorectal prostate, breast Early detection Digital mammography MRI for high-risk Gail algorithm (own factors) Klaus algorithm (family history) Patient interaction Health awareness Trust and confidence in advice given Use of clinical guidelines to improve quality of care Respiratory Chronic obstructive pulmonary disease Research & development Public vs commercial sponsors Regulators' attitude to developments Disease-focused approach vs global impact of ageing Multiple diseases Death Neurological Dementia, Alzheimer's, Parkinson's Current approaches Targeting DCIS Surgery with node follow-up Adjuvant radiotherapy Herceptin, Tamoxifen Breast cancer Medical innovations Growth factor inhibition Future of personalised medicine (eg tumour profiling) Clinical trials pipeline Phase II (230 trials*) Phase III (56) eg pertuzumab (limits cancer growth) 38

39 Medical Models : IRIS schematic model operation 39

40 Statistical methods IRIS Projection UK Males Source : ONS and GPRD data, Swiss Re Calculation 40

41 Statistical methods CMI Projection UK Males Source : CMI 41

42 Statistical methods Projection LeeCarter Fit Males ( ) Source : ONS data, Swiss Re Calculation 42

43 Longevity modelling behind other disciplines Actuarial subjective projections Actuariallyadjusted historic projections Stochastic modelling Interdisciplinary cause-based projections Maturity of development Longevity modelling Natural perils modelling 43

44 Mortality trend is systematic in nature (no diversification from volume or geography) 20 Period life expectancy at age UK West Germany 2000 Canada USA Japan France Sweden Source: 44

45 A possible longevity risk solution 45

46 Catalysts for risk transfer Motivated buyers and sellers Risk transfer Credible, timely and consistent data Well-understood and widelyaccepted risk models 46

47 Longevity Risk Market Longevity Risk is heavily concentrated in a few agents Volume is huge This represents an interesting challenge 1 year deviation in longevity assumptions implies a 5% impact on the liabilities 47

48 Longevity Risk Market There is a large demand for longevity risk Transfer Current supply is available to meet this demand Reinsurers have limited capacity for longevity risk It is a developing market 48

49 How does a longevity swap work? With a longevity swap, a life insurer protects a part, or all, of the portfolio against an unexpected increase in life expectancy. The client provides his expectations about the expected future annuity or pension payments to Swiss Re, together with relevant experience studies. Swiss Re makes an assessment of its own view on the expected future annual annuity or pension payments for the underlying book and adds margins for capital costs, expenses and profit. 49

50 How does a longevity swap work? Annual premium leg Annual pension payments Scenario 1: Net difference paid by the insurer to the pension plan Scenario 1: Illustrative increased annual pension payments if longevity gets longer Scenario 2: Illustrative lower annual pension payments if longevity gets shorter Scenario 2: Net difference paid by the pension plan to the insurer 2010 Certainty Pension plan liability is the annual premium payment regardless of changes in longevity Swap Format Insurer covers the cost if longevity increases, and gains if longevity decreases

51 How does longevity insurance work? Alternative presentation Swiss Re Proposal Current Difficult to budget and invest for the future with this degree of uncertainty Uncertain annual cashflows over an unknown future timeframe become certain annual insurance premiums over a known premium term April 2013 Swiss Re Life & Health Business Development 51

52 Final Comments 52

53 Final Comments Understanding longevity risk continues to be a challenge for both buyers and sellers of risk Historic models have failed need for forward looking cause based models Longevity risk is systematic scale and portfolio diversification of limited use as risk mitigation tools Reinsurers providing some new capacity Future life expectancies are highly uncertain, potentially resulting in significant financial impacts Swiss Re concluded the world's first publicly disclosed longevity swap in 2007 and its first longevity risk transfer with a pension plan in 2009 Swiss Re is working with insurers and multinationals to construct solutions for their annuity and pension exposures 53

54 Thank you

55 Legal notice 2012 Swiss Re. All rights reserved. You are not permitted to create any modifications or derivatives of this presentation or to use it for commercial or other public purposes without the prior written permission of Swiss Re. Although all the information used was taken from reliable sources, Swiss Re does not accept any responsibility for the accuracy or comprehensiveness of the details given. All liability for the accuracy and completeness thereof or for any damage resulting from the use of the information contained in this presentation is expressly excluded. Under no circumstances shall Swiss Re or its Group companies be liable for any financial and/or consequential loss relating to this presentation. 55