Public Spending and the Effect of Changes in Life Expectancy. Tuesday June 16 th 2009

Transcription

1 Public Spending and the Effect of Changes in Life Expectancy Tuesday June 16 th 2009

2 Professor Mike Murphy LSE

3 ILC-UK and Actuarial Profession Joint Seminar Public Spending and the Effect of Changes in Life Expectancy 16 June 2009 What alternative methods of estimating life expectancy are available? Mike Murphy LSE

4 MAP2030 Programme Objectives To produce high quality analysis to inform public debate and development of future long-term care and pensions policy up to 2030 by: projecting the numbers, disability status, family circumstances, income, savings and care needs of older people assessing the affordability and distribution of costs and benefits of combined policy options for pensions and long-term care accounting for links between care needs and economic resources in later life

5 MAP2030 Structure Mortality trends & implications linkages

6 MAP2030 Research Teams Mike Murphy, Mariachiara di Chesare (LSE) Carol Jagger, James Lindesay, Ruth Matthews (Leicester) Emily Grundy & Sanna Read (LSHTM) Ruth Hancock & Marcello Morciano (UEA) Raphael Wittenberg, Adelina Comas-Herrera, Linda Pickard, Derek King, Juliette Malley & Megan Challis (PSSRU) Chris Curry, John Adams, Sean James (PPI) Supporting partner Department for Work and Pensions

7 Why looking at life expectancy matters Dependency time-bomb: Britain must urgently take steps to prepare for the seismic impact of a rapidly ageing population George Magnus (senior economic adviser for UBS investment bank) The Guardian, Wednesday 4 February 2009

8 Projections of population aged 85 & over in 2011, 2031 and 2041, alternative projection base years, England & Wales Source: author's calculations

9 Observed mean absolute error for life expectancy at birth for men in 14 countries Nico Keilman. UK national population projections in perspective: How successful compared to those in other European countries? Population Trends 129, Table 3

10 England & Wales standardised mortality rate annual smoothed improvement (%) (2006-based principal projection) Author's calculations based on WHO European Standard

11 Sex Ratio (Males per 100 Females), England & Wales 2001, 2011 & 2021 (2001 Census & 2006-based projections) 100 & over Age grou

12 The choices There is the possibility of lower incidences of cancer, heart disease and strokes through changes in lifestyle and medical advances On the other hand some demographers believe that a law of diminishing returns will apply to death rate reductions at advanced ages (Office for National Statistics, 2008, National Population Projections 2006-based Series PP2 No 26, ed. Helen Bray, p. 33)

13 Main drivers of past mortality - and future also? Medical Advances Smoking trends Obesity Infectious diseases Uncertainty at young ages (Karen Dunnell, 2008, Ageing and Mortality in the UK - National Statistician's Annual Article on the Population. Population Trends 134: 6-23, p. 19)

14 The Golden generations Those born during the period (and centred around 1931) have exhibited greater rates of improvement over the last 25 years than those born on either side. There is currently no evidence that these differentials are declining. Similar cohort effects seen in other countries suggest that these differentials may persist well into the oldest ages. (1931 Cohort assumed to improve 1.5% p.a. more than underlying 1% value: PP2, 2008, p. 26)

15 Actual and assumed overall average annual rates of mortality improvement, England & Wales (Per cent) Series PP2 No 26 Table 7.3 Males Females Past (actual) Future (assumed) Past (actual) Future (assumed) Last/next 24 years Last/next 44 years Last/next 74 years

16 The role of obesity Obesity Increasing trend Potentially important in substantially reducing e 0 But Extrapolation of trends is over-simplified Magnitudes of estimated effects are inconsistent Major route of obesity -> mortality is CVD

17 The first generation to live less than its parents? The Government Actuary s Department currently predicts that life expectancy will rise in the next 50 years by around eight years for men and seven years for women. the role of the increase in obesity will have surprisingly little impact (less than a year) on the life expectancy of the population (Government Office for Science. Tackling Obesities: Future Choices. Foresight Project Report 2nd Edition p. 38)

18 Reasons to expect mortality improvements to decline Negative cohort effect dies out smoking cessation bonus disappears early life factors less relevant Positive obesity lack of blockbuster drugs evolution

19 Reasons to expect mortality improvements to increase Negative smoking effects among older people over-rated early life effects Positive improvements in prevention and treatment empirical lack of limits

20 Life Expectancy at Birth, UK & Japan 2006 Principal Projections Years 80 UK Male Japan Male UK Female Japan Female Year

21 Life expectancy at birth, alternative variants, UK Users of the projections can gain some insight by considering the high and low life expectancy variants intended to represent plausible alternative assumptions and are far from reflecting the extremes of thinking on future mortality. (Office for National Statistics, 2008, National Population Projections based Series PP2 No 26, ed. Helen Bray, p. 33) Variants: underlying rate of improvement Low 0% Principal 1% High 2% Very high 3%* All include cohort term (maximum 1.5% p.a. improvement over underlying value in 1931) (*Own projection, others are taken from PP2 No. 26)

22 Life expectancy projections: US (2008) and UK (2006-based) variant values (years) Males Value Change from US (Principal) UK (Principal) UK (High) UK (V. high) Females US (Principal) UK (Principal) UK (High) UK (V. high)

23 National variability: scope for improvement? Life expectancy at birth with 95% confidence intervals for Scottish Index Multiple Deprivation 2006 Decile area, (Males & Females)

24 International & national context British mortality projections are among the most optimistic in the World for in the next 40 years or so The differences in life expectancy between 2010 and 2050 even for the high variant are similar to that between values in the top and bottom deprivation deciles in Scotland

25 Are sustained high rates of mortality improvement possible? US female single-year-of-age death probabilities, observed and under assumption of 2% p.a. improvement, with period from which particular causes would be eliminated S. Jay Olshansky and Bruce A. Carnes, Demographic Perspectives on Human Senescence Population and Development Review, 20(1):57-80 (1994), Figure 8

26 Are sustained high rates of mortality improvement possible? US female single-year-of-age death probabilities, observed and under assumption of 2% p.a. improvement, with period from which particular causes would be eliminated 2004 q q S. Jay Olshansky and Bruce A. Carnes, Demographic Perspectives on Human Senescence Population and Development Review, 20(1):57-80 (1994), Figure 8

27 Some issues in predicting future life expectancy that should be considered In the absence of clear consensus on empirical balance of competing future effects, why not retain current patterns? Accelerating rates of improvement Cohort effects not proven More attention to whether cohorts are pseudo-cohorts Sensible to investigate implications of populations with higher rates of mortality improvement

28 What alternative methods of estimating life expectancy are available? Thank you for your attention

29 Raphael Wittenberg Personal Social Services Research Unit LSE

30 PUBLIC EXPENDITURE ON LONG- TERM CARE FOR OLDER PEOPLE: THE EFFECT OF CHANGES IN LIFE EXPECTANCY 16 June 2009

31 MAP2030 WORKPACKAGE 5: LONG-TERM CARE AND PENSIONS ADELINA COMAS-HERRERA, PSSRU, LSE CHRIS CURRY, PENSIONS POLICY INSTITUTE RUTH HANCOCK, UNIVERSITY OF EAST ANGLIA SEAN JAMES, PENSIONS POLICY INSTITUTE DEREK KING, PSSRU, LSE JULIETTE MALLEY, PSSRU, LSE MARCELLO MORCIANO, UEA LINDA PICKARD, PSSRU, LSE RAPHAEL WITTENBERG, PSSRU, LSE JOHN ADAMS, PENSIONS POLICY INSTITUTE MEGAN CHALLIS, PSSRU,LSE 31

32 OVERVIEW OF PRESENTATION POLICY BACKGROUND METHODOLOGY BASE CASE VARIANT POPULATION PROJECTIONS FUNDING REFORM OPTIONS WANLESS PACKAGES OF CARE CONCLUSIONS 32

33 LONG-TERM CARE POLICY CONTEXT Concern over future affordability of longterm care and pensions increasing numbers living to late old age uncertainty over numbers who will need care potentially rising expectations Debate about the appropriate balance between public and private funding 33

34 LONG-TERM CARE FINANCING SYSTEM Health care free of charge at point of use throughout UK Nursing care in nursing homes now also free throughout the UK Personal care free in Scotland but subject to user charges in rest of the UK Hotel costs in care homes and domestic help subject to charges throughout UK Disability (cash) benefits are not subject to means test, throughout the UK Govt Green Paper on Care and Support expected to be published shortly 34

35 METHODOLOGY Use of three linked models: PPI suite of models on pensions Caresim, UEA, model of resources of older people and charging for care PSSRU at LSE model of long-term care Innovative combination of macro and microsimulation models 35

36 PSSRU MODEL The PSSRU model aims to make projections of: Numbers of disabled older people Long-term care services and disability benefits Long-term care expenditure: public and private Social care workforce 36

37 CARESIM MODEL Uses micro data on older people s income and capital assets (Family Resources Survey) Simulates the means-tests for residential care and for home care Uncompleted lengths of stay assigned randomly and contribution to fees calculated for that point Calculates what each older person in the sample would pay for care should they need it Does NOT predict who or how many will need care 37

38 LINKING THE CARESIM AND PSSRU MODELS PSSRU model provides: age/gender/marital status/housing tenure weights within type of care for each projection year up to 2032, to act as weights CARESIM model provides to PSSRU model projected trend in % of care home residents and home care clients eligible for state support projected average % of care home and home care fees met by state supported service users Projected average % of user charges and private payments met using disability benefits 38

39 CENTRAL BASE CASE Official principal population projection, by age, gender and marital status Unchanged age-specific disability rates No change in patterns of formal and informal care Unit costs rise by 2% per year in real terms (but constant for non-staff, non-capital costs) No change in financing system from current system in England 39

40 BASE CASE EXPENDITURE Public expenditure on long-term care (including disability benefits used to fund care) is projected to rise from around 12bn in 2007 to over 25bn in 2027 and 32bn in 2032, at constant 2007 prices; Private expenditure on long-term care is projected to rise from around 7.3bn in 2007 to around 18bn in 2027 and 23bn in 2032, at constant 2007 prices. 40

41 BASE CASE PROJECTIONS: PUBLIC EXPENDITURE AS % GDP 2.0% 1.8% 1.6% 1.4% 1.2% % GDP 1.0% 0.8% 0.6% 0.4% 0.2% 0.0%

42 VARIANT POPULATION PROJECTIONS ONS official high life expectancy population projections (HLE) show numbers of older people rising by 71% between 2006 and 2032 as against 66% under the principal projection. Very high life expectancy population projections (VHLE) prepared by Mike Murphy show numbers of older people rising by 77% between 2006 and

43 IMPACT OF VARIANT POPULATION PROJECTIONS POP 65+ % INCREASE POP 85+ % INCREASE DISABLED POP 65+ % INCREASE TOTAL EXPENDITURE 2032 bn TOTAL EXPENDITURE 2032 %GDP CLE HLE VHLE 66% 71% 77% 145% 168% 192% 78% 86% 95% 53bn 56bn 59bn 2.90% 3.05% 3.25% 43

44 PROJECTED PUBLIC EXPENDITURE ON LONG-TERM CARE UNDER DIFFERENT LIFE EXPECTANCY VARIANTS, % GDP 2.10% Expenditure as a percentage of GDP 1.90% 1.70% 1.50% 1.30% 1.10% 0.90% 0.70%

45 REFORM OPTIONS Free personal care a non means-tested state contribution to care home fees of ( fixed care costs) costs of assessed needs for personal care at home met by state Wanless partnership proposal user pays 1/6 of personal care costs as copayment; state pays remainder user contributes to hotel costs as under current rules, first deducting 1/6 co-payment from income. 45

46 REFORM OPTIONS (cont) Disregard of housing assets for care component of care home fees (with income used first towards hotel costs) Doubling of Personal Expenses Allowance From to in April 2007 prices 46

47 PUBLIC EXPENDITURE COSTS OF DIFFERENT FUNDING REFORMS The cost to public expenditure of the different funding arrangements, under the principal population projection, would be: Free personal care, around 1.7bn in 2007 rising to some 4.2bn in Disregard housing assets in means test for care in care homes around 700m in 2007 rising to 1.7bn in 2032 Wanless partnership funding arrangement, around 1.2bn in 2007 rising to some 2.9bn in 2032 Doubling the personal expenses allowance around 275m in 2007 rising to some 725m in

48 COSTS TO PUBLIC EXPENDITURE OF DIFFERENT FUNDING REFORMS 2007 AND Public expenditure, m Free personal care Housing disregard Partnership PEA doubled

49 CARE HOME FEES UNDER REFORM OPTIONS Self-funder fees are higher than LA-funded fees Shifts in balance of numbers self- funded to LAfunded residents over time leads to Change in average provider income per person Change in total provider income Alternative scenarios explored assuming that average provider income per person remains constant, but these are not discussed in this presentation. 49

50 COSTS OF DIFFERENT FUNDING REFORMS UNDER DIFFERENT POPULATION PROJECTIONS While free personal care would cost 4.2bn in 2032 under CLE, it would cost 4.5bn under HLE or 4.8bn under VHLE, relative to projection for current funding system. The total cost above current system CLE for free personal care would be 6.4bn under HLE and 8.5bn under VHLE. 50

51 COSTS TO PUBLIC EXPENDITURE IN 2032 OF DIFFERENT FUNDING REFORMS UNDER DIFFERENT POPULATION PROJECTIONS 10,000 9,000 8,000 Public expenditure, m 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 Current system Free personal care Housing disregard Partnership model PEA doubled CLE HLE VHLE 51

52 WANLESS PACKAGES OF CARE The Wanless Social Care Review recommended a substantial increase in community-based services (over 900,000 rather than 650,000 users) and a reduction in residential care services (250,000 rather than 340,000 users) The public expenditure costs of implementing the Wanless patterns of care would be around 3.2bn in 2007 rising to around 8bn in 2027 and 10bn in 2032, under the current funding system and principal population projections. 52

53 WANLESS PACKAGES OF CARE The cost of Wanless packages of care and partnership funding system would be 13.4bn in 2032 under CLE rising to 18.4bn under VHLE. The cost of Wanless packages of care and free personal care would be 15.7bn in 2032 under CLE rising to 21.0bn under VHLE (above current system CLE). 53

54 COSTS TO PUBLIC EXPENDITURE OF DIFFERENT FUNDING REFORMS UNDER DIFFERENT POPULATION PROJECTIONS AND PACKAGES OF CARE, ,000 Public expenditure, m 20,000 15,000 10,000 5,000 0 Current Wanless Current Wanless Current Wanless Current system Free personal care Partnership CLE HLE VHLE 54

55 CONCLUSIONS (1) Public expenditure on long-term care projected to rise from 0.95% of GDP in 2007 to in 2032: 1.75% of GDP under the ONS principal population projection; 1.85% under the ONS high life expectancy projection; 1.95% under a very high life expectancy projection. 55

56 CONCLUSIONS (2) The costs of reforms to the funding system would in 2032 be significantly higher under the high or very high life expectancy projections, eg for free personal care 6.4bn and 8.5bn as against 4.2bn. The costs of Wanless packages of care would in 2032 also be significantly higher under the high or very high life expectancy projections, 12.3bn and 14.6bn as against 10.1bn. 56

57 Chris Curry Pensions Policy Institute

58 PENSIONS POLICY INSTITUTE PPI The impact of longevity on pension projections Chris Curry Pensions Policy Institute International Longevity Centre 16 June

59 Impact of longevity on pension projections PENSIONS POLICY INSTITUTE PPI Higher than expected life expectancy: Would increase spending on state pensions Could increase or reduce income from private pensions Would increase spending on meanstested benefits 59

60 About the PPI models Aggregate model, based on the labour market, for BSP, SERPS/S2P and private pensions Distributional model, based on the Pensioners Incomes series, for Pension Credit and income distributions PENSIONS POLICY INSTITUTE PPI Static analysis no behavioural changes assumed in saving or working 60

61 The Baseline projections Government spending on state pensions increases as the numbers over pension age increase and pensions become more generous Income from private pensions increases initially, but starts to fall back as DB is replaced by DC Spending on Pension Credit falls as it is made less generous and more people receive better state pensions PENSIONS POLICY INSTITUTE PPI 61

62 Spending on state pensions is projected to increase Government spending on pensions (Basic state pension, SERPS/S2P, Pension Credit and other benefits), % GDP, baseline projection PENSIONS POLICY INSTITUTE PPI

63 Income from private pensions rises then falls Income from private pensions, % GDP, baseline projection PENSIONS POLICY INSTITUTE PPI

64 Means-tested benefit expenditure increases slightly in the long term Government spending on Means-tested Benefits (PC, HB, CTB)for pensioners, % GDP, baseline projection PENSIONS POLICY INSTITUTE PPI

65 Higher life expectancy increases state spending PENSIONS POLICY INSTITUTE PPI Individuals receive the same levels of state pension as in the baseline But they receive pensions for longer 65

66 Higher life expectancy increases state spending Government spending on pensions (Basic state pension, SERPS/S2P, Pension Credit and other benefits), % GDP 6.5% 6.0% 5.8% PENSIONS POLICY INSTITUTE PPI 6.1% 5.9% 5.5% 5.0% 4.5% 4.5% 4.5% 4.7% 4.9% 5.0% 5.2% 4.0% 3.5% 3.0% Baseline High Very High

67 Higher life expectancy can increase or decrease private pension income PENSIONS POLICY INSTITUTE PPI People live longer and receive pensions for longer DB pension pay outs are unaffected for individuals, but increase overall as they are paid for longer But DC pension pay outs would be lower for individuals if annuity providers know about higher life expectancy overall impact is ambiguous 67

68 Higher life expectancy can increase or decrease private pension income Income from private pensions in 2032, % GDP PENSIONS POLICY INSTITUTE PPI

69 Higher life expectancy increase spending on means-tested benefits PENSIONS POLICY INSTITUTE PPI Although more state pension is paid, individuals do not see higher state pensions Some state pension income in payment is still uprated in line with prices Private pension income may be lower for individuals Older pensioners are more likely to have low income and be eligible for means-tested benefits 69

70 Higher life expectancy would increase spending on meanstested benefits Government spending on means-tested benefits (PC, HB, CTB) for pensioners, % GDP PENSIONS POLICY INSTITUTE PPI

71 Impact of longevity on pension projections PENSIONS POLICY INSTITUTE PPI Higher than expected life expectancy: Would increase spending on state pensions Could increase or reduce income from private pensions Would increase spending on meanstested benefits 71

72 Higher life expectancy increases state spending Government spending on pensions, means tested benefits and long-term care, % GDP PENSIONS POLICY INSTITUTE PPI