Slides from The Fiscal Implications of Demographic Change in the Health Sector presented by Paul Goldrick-Kelly

1. The Fiscal Implications of
Demographic Change in the
Health Sector
Paul Goldrick-Kelly
NERI (Nevin Economic Research Institute)
Dublin
PaulGK@NERInstitute.net
www.NERInstitute.net

2. Outline
• Introduction
• Data
• Model Components
– Demographic Cost Drivers
– Income and Residual Cost Drivers
• Projection Model
• Assumptions
• Results
• Conclusions

3. Introduction
• Context
– Substantial increases over two decades preceding crisis of 2008
from low base (Wren,2004).
– Fiscal retrenchment results in reduction in spending after crash.
Expenditure plans of previous government imply restrained
spending growth.
– Demographic change proceeds apace, population (still)
expanding and ageing. This has associated costs.

4. Figure 1: Real Current Public Health Expenditure Per Capita 2000-
2013 SHA(2011) (pg. 4)
0
500
1000
1500
2000
2500
3000
3500
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
RealCurrentPublicExpenditurePer
capitaatConstant2013€Prices
Source: CSO (2015) Ireland’s System of Health
Accounts, Annual Results 2013 (Preliminary)

5. Current Health Expenditure as a Percentage of GDP 2013
SHA(2011)
(EU15 Members denoted by *)
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
CurrentHealthExpenditureasa%ofGDP
Current Public Expenditure Other Current Expenditure
Source: Eurostat, CSO (2015) Ireland’s System of
Health Accounts, Annual Results 2013
(Preliminary),Author’s Calculations

6. • This working paper aims to:
– Quantify the costs associated with demographic change on
overall current public health expenditure using a simple model.
– Produce projections of public health spending to 2035.
– Briefly highlight likely compositional changes to health spending
induced by an ageing and expanding population.

7. Data
• Data Sources include:
– CSO (2013) Population Projections M1F1 and M2F2 with
attendant life expectancy and mortality rate projections by year
cohort.
– EU-15 Average Age and Gender Specific costs (EU AWG, 2012).
 Irish data unavailable
– NERI (2015) and OECD (2013) estimates of GDP growth.
– SHA (2011) estimates of Current Public Health Expenditure
(CSO, 2015).

8. Figure 2: Age and Gender Specific Average per Capita Current
Public Expenditure EU-15
2013 Adjusted (pg.6)Under1year
3years
6years
9years
12years
15years
18years
21years
24years
27years
30years
33years
36years
39years
42years
45years
48years
51years
54years
57years
60years
63years
66years
69years
72years
75years
78years
81years
84years
87years
90years
93years
96years
99yearsandover
€0.00
€2,000.00
€4,000.00
€6,000.00
€8,000.00
€10,000.00
€12,000.00
Male Female
Source: EU Commission 2012 Ageing Report

9. Model Components
• Demographic Component
– Age Composition - Literature indicates not very important
(Zweifel et al.,1999; Anderson and Hussey, 2002).
– Death Related Costs – More significant predictor of health costs,
proximity to death. High concentration of cost immediately
preceding death. (Stooker et al.,2001; McGrail et al.,2000)
– Morbidity – Will increases in life expectancy result in a larger
portion of life spent ill or healthy? (Expansion vs Compression)

10. Demographic Cost Drivers
• Age Composition
– Most intuitive demographic driver of demand increases.
– Given high relative cost for certain cohorts (infants, women of
child-bearing age, elderly) change in fertility and life expectancy
will likely impact healthcare utilisation and cost.
– Common assumption that increases in the elderly population will
result in higher average expenditures.

11. • Death related costs
– Hypothesis states that relevant factor in cost incidence and care
demands is proximity to death rather than age per se (Gray,
2005).
– Majority of lifetime healthcare costs accumulate at the end of an
individuals life.
– Death related costs higher for younger patients than old ones.
Generally offset by lower mortality rates.

12. • Morbidity Compression or Expansion
– Debate as to whether healthcare improvements characterised by
expansions or contractions with respect morbidity.
– Morbidity expansion implies that increases in life expectancy
translate into more years spent in ill-health. Chronic illness and
disability drives increases.
– Morbidity compression proponents suggest that number of years
spent in good health will increase as life expectancy increases.
The relative portion of life spent in ill-health shrinks.
– Some evidence for morbidity compression outpacing life
expectancy gains within “treatment groups”-Hubert and Frees
(1994), Chakravarty et al. (2008).

13. Income and Residual Cost Drivers
• Income
– Response of health demanded to income changes. Measured via and
income elasticity of health demand. For every 1% change in income,
what is the percentage change in health demanded?
• Residual
– Expenditure growth left unexplained by demographic or income related
factors.
– Thought to include factors such as:
I. Technology
II. Relative Prices
III. Institutional Arrangements and Policies
– Difficult to forecast these factors. Relative Prices, “Baumol Effect” may
be relevant for particular sectors, but can’t be assumed for health
service as a whole.

14. Current Public Health
Expenditure
Demographic
Factors
Age
Composition of
Population
Death Related
Costs
Morbidity
Income
Residual
Technology
Relative Prices
Institutional
Arrangements
Determinants of Current Public Health
Expenditure Growth

15. Projection Model
∆ ln
H
N t
= ∆ ln Dt + ε∆ ln
Y
N t
+ ∆ ln γt
Per capita current public spending cost growth is a function of
demographic cost pressure, the response of health demand to changes
in income (Real GDP per Capita) and residual growth factors
(OECD, 2015).

16. • Demographic cost growth:
– Death Costs → Multiply age/gender specific per capita death
costs by decedent population by cohort. Mortality rates from
CSO.
– Survivor Costs → Subtract cohort specific death costs from total
health costs for that cohort and divide by the surviving
population.
– Under morbidity expansion, survivor and death per capita costs
remain constant. Under compression, per capita survivor costs
adjusted according to LE gains. Constant ratio between per
capita survivor and death related costs used to calculate
adjusted death costs in accordance with survivor adjustments.

17. Assumptions
• Demographic Component
– Central Scenario uses M1F1 projections. Sensitivity analyses
uses M2F2.
– Death related costs= 4 times cost of oldest cohort (100+) for all
below 60. Thereafter coefficient allowed decline linearly to unity
for oldest cohort. (OECD 2015)
– Morbidity Compression= constant cost profile across forecast
period.
– Morbidity expansion= gains in life expectancy translate into
healthy years 1 for 1 over 65. Coefficient declines to zero at 44,
i.e.LE gains don’t result in adjusted costs for individuals under 45
(Caley and Sidhu, 2010).

18. • Income Component
– Central scenario sets ε at 0.8. Sensitivity performed for
elasticities of 0.6 and 1.
– M1F1 growth equals QEO projections to 2017, 3% annually
thereafter (OECD).
– M2F2 growth same to 2017. Rest of forecast
growth=Employment Growth + Productivity Growth (1.5%).

19. • Residual Component
– Residual set at 1.5% annually. Country specific residual from
growth accounting not included because period analysed
potentially atypical and would result in explosive cost growth.
– 2 scenarios are presented. First has residual constant for entirety
of forecast period. Second sees residual decline linearly to zero
by 2035. Second scenario models an implicit policy response to
curb cost growth.

20. Scenario Gains in Life
Expectancy=
Gains in
healthy life
years
Cost Profile
Remains
Constant
Residual stays
a constant
1.5%
Residual
declines from
1.5% in 2013 to
0 in 2035
Scenario 1
Morbidity
Expansion
 
Scenario 1
Morbidity
Compression
 
Scenario 2
Morbidity
Expansion
 
Scenario 2
Morbidity
Compression
 
Table of Nomenclature for Results

21. Results
Years
(Inclusive)
Million € Spending Pressure
(at 2013 prices)
% Overall Annual Expenditure
Growth due to demographics
Morbidity
Expansion
Morbidity
Compression
Morbidity
Expansion
Morbidity
Compression
Population
Projection
M1F1 M2F2 M1F1 M2F2 M1F1 M2F2 M1F1 M2F2
2013-2015 €152.37 €142.42 €121.35 €111.40 1.16% 1.08% 0.92% 0.85%
2016-2018 €195.91 €167.46 €162.53 €134.16 1.44% 1.24% 1.21% 1.00%
2019-2021 €237.49 €189.23 €202.03 €154.01 1.67% 1.35% 1.44% 1.11%
2022-2024 €265.95 €202.07 €228.21 €164.80 1.78% 1.38% 1.56% 1.15%
2025-2027 €281.94 €210.23 €243.00 €171.96 1.79% 1.38% 1.58% 1.16%
2028-2030 €295.92 €220.25 €256.14 €181.40 1.78% 1.39% 1.59% 1.18%
2031-2033 €303.43 €223.18 €264.15 €185.17 1.73% 1.35% 1.57% 1.16%
2034-2035 €306.85 €222.16 €267.87 €184.82 1.68% 1.30% 1.53% 1.13%
Table 1: Average Annual Direct
Demographic Cost Pressure (pg. 14)

22. Figure 5: Cumulative Demographic Cost
Pressure at 2013 Prices (pg.15)
0
1
2
3
4
5
6
CumulativeDemographicCostPressure2013
Prices(€Billions)
M1F1 Morbidity Expansion M1F1 Morbidity Compression M2F2 Morbidity Expansion M2F2 Morbidity Compression

23. Figure 6: Decomposition of Real Average per Capita
Current Health Expenditure growth 2013-2035 (pg.18)
Expanded Morbidity
Scenario 1
Compression of Morbidity
Scenario 1
Expanded Morbidity
Scenario 2
Compression of Morbidity
Scenario 2
Average Real GDP
Growth Per Capita
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
3.50%
4.00%
AnnualAveragePerCapitaCostGrowth
Demographic Component Income Component Residual Component

24. Figure 7: M1F1 Projections 2035
Sensitivity Analysis according in Income
Elasticity (pg.19)
7.29% 7.29% 7.29% 7.29%
2.91% 2.48%
1.34% 0.97%
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
Mobidity Expansion Scenario 1 Mobidity Compression Scenario
1
Mobidity Expansion Scenario 2 Mobidity Compression Scenario
2
PublicHealthSpendingasa%ofGDP
Current Public Spending as % GDP 2013 Relative Current Public Spending Increase % GDP 2035

25. Figure 8: M2F2 Projections 2035
Sensitivity Analysis according in
Income Elasticity (pg.21)
7.29% 7.29% 7.29% 7.29%
2.91% 2.48%
1.34% 0.97%
0.21%
0.18%
0.18%
0.16%
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
Mobidity Expansion Scenario 1 Mobidity Compression Scenario
1
Mobidity Expansion Scenario 2 Mobidity Compression Scenario
2
CurrentPublicHealthSpendingasa%ofGDP
Current Public Spending as a % GDP 2013 Relative Increase in Spending as % GDP M1F1
Relative Deviation from M1F1 Estimates as % GDP

26. Figure 9: Age Decomposition of Current
Health Expenditure Cost Growth (pg.22)
70.00%
58.68% 59.63% 56.16% 57.06%
30.00%
41.32% 40.37% 43.84% 42.94%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2013 2035 M1F1 Morbidity
Expansion
2035 M1F1 Morbidity
Compression
2035 M2F2 Morbidity
Expansion
2035 M2F2 Morbidity
Compression
Under 65 Over 65

27. Conclusions
• Significant demographic cost pressures exist in the
forecast model.
– Annual cost pressure exceeds €100 million in all cases.
– Doesn’t attain €200 for MC M2F2. In all other cases, exceeds
€200 million from 2019. Reaches €300 million in 2034 in ME
M1F1.
– Cost growth ≈ 1% annually
– Cumulative demographic costs exceed €1 billion between 2019
and 2021. Range from €3.57 billion to €5.66 billion in 2035.

28. • Current Public Health Spending will increase to 2035.
– Range of central estimates between 8.3 and 10.2% of forecast
GDP in 2035, Though this is higher under the most pessimistic
assumption set (maximum ≈ 11.6%).
• Changes in composition of public health spending.
– Over 65s go from 30% to over 40% of spend 2013 to 2035.

29. • Complicating factors to consider include:
– Possible inefficiency within the current system upwardly biasing
forecast estimates.
– Absence of Irish Specific cost profiles.
– Residual cost growth accuracy given its lack of explanatory
power.
– European and domestic rules restraining expenditure growth.
– Likely endogeneity between components.
– Political Choices.

30. Policy Questions
• What are appropriate investment levels?
• Are there efficiencies that can mitigate residual cost
growth?
• What is the vision for the health service?
• What is a reasonable time frame for such a vision’s
realisation?

31. www.NERInstitute.net