012 Acta všfs, 1/2010, vol. 4
Intergenerational Solidarity of the Public Health
Care Systems in Europe
Mezigenerační solidarita v systémech veřejného
zdravotnictví v Evropě1
Abstract
Ageing of the population has become one of the important topics in developed countries
in recent times. Health care and pension systems based on solidarity are especially vulnerable
to demographic changes associated with falling fertility rates and increasing life
expectancy. The main task of the paper is to answer and quantify the question who bears
the costs and who benefits from the public health care systems and reveal possible future
imbalances. We introduce a long-term projection technique enhanced by some elements
of generational accounting approach with intent to express the ageing problem in the
health care sector. We explore both the revenue side of the public health care systems as
well as the health care expenditures from the perspective of separate generations. Following
countries have been selected as the representatives of the European health care
systems: the Czech Republic, Denmark, Estonia, France, Italy, the Netherlands, Switzerland
and the United Kingdom. The model points to un-sustainability of the current public
health care systems if the effective tax rates did not increase. If the demand for health care
had to be satisfied, the health care systems financed through social contributions and
income-based taxes would be faced with increasing burden on working population.
Keywords
health care financing, sustainability, generational accounts, population ageing
JEL classification
E62, H51, I18
Abstrakt
Stárnutí obyvatelstva se stalo v poslední době jedním z důležitých témat v mnoha rozvinutých
zemích. Důchodové systémy a systémy financování zdravotnictví založené na
solidaritě jsou výrazně citlivé na demografické změny související s klesající porodností
a rostoucí střední délkou života. Cílem tohoto článku je ukázat a vyčíslit možnou budoucí
nerovnováhu v systémech financování veřejného zdravotnictví a prezentovat jej
i z pohledu jednotlivých generací. Abychom lépe vyjádřili problém stárnutí obyvatelstva
v systémech financování veřejného zdravotnictví, obohacujeme tak standardní přístup
k dlouhodobým projekcím o prvky generačního účetnictví. Srovnání systémů napříč Evropou
umožňuje výběr reprezentativních systémů financování zdravotnictví, a to České
1	 Award-winning essay of Prof. F. Vencovský 2009 Prize.
KATEŘINA PAVLOKOVÁ
013Acta všfs, 1/2010, vol. 4
republiky, Dánska, Estonska, Francie, Itálie, Nizozemí, Švýcarska a Velké Británie. Výsledek
modelu ukazuje na neudržitelnost současných systémů financování. Má-li být poptávka
po zdravotních službách uspokojena, budou systémy založené na veřejném zdravotním
pojištění nebo na daních z příjmu klást rostoucí nároky na pracující populaci, a zvyšovat
tak pnutí a požadavky na další redistribuci podmíněnou rostoucí daňovou zátěží, aby
nedošlo k omezování zdravotních služeb. Míru ohrožení starší populace omezením poskytované
zdravotní péče v jednotlivých letech a zemích vyjadřujeme indexem soběstačnosti,
který ukazuje na některá rizika vybraných systémů.
Klíčová slova
financování zdravotnictví, udržitelnost, generační účty, stárnutí obyvatelstva
Introduction
Ageing of the population has become an important topic in developed countries in recent
times. European health care and pension systems based on solidarity are especially vulnerable
to demographic changes associated with falling fertility rates and increasing life
expectancy. Even though new approaches to the modelling of health care sector expenditures
predicting lower increases in ageing-related health care expenditure than previously
assumed have emerged, health care systems will be affected by population ageing.
Modifying a long-term projection approach with some fragments of the generational accounting
methodology, we explore the revenue side of public health care systems as well
as the health care expenditure side from the perspective of separate generations. Due to an
increasing dependency ratio, health care systems financed through social contributions and
income-based taxes will be faced with an increasing burden on the working population if
the demand for health care is to be satisfied. In the long term, however, health care expenditure
will always tend to equal the revenues of the health care system. In other words, if revenues
remain at current levels, some of the demand for health care will not be satisfied.
In order to represent the differences between the various systems of health care financing
we have chosen eight representative countries from all over Europe. These countries are:
Denmark, Estonia, France, Italy, the Netherlands, Switzerland, the United Kingdom and
the Czech Republic.
The plan of this study is as follows. First of all, we briefly discuss economic implications
of the phenomena known as ageing, which reflects most of major changes the study
count with. Thereafter we continue with a detailed description of the long-term projection
technique developed for the study and give the descriptive statistics of the main variables.
The third part then summarises results and conclusion discusses the implications
for policy-makers.
1	 Ageing society and its basic economic implications
Ageing of the population may be characterised as a long-term process where the structure
of the population changes due to falling fertility rates or increasing life expectancy or
both, on condition that net migration does not contribute too extensively to reverse the
014 Acta všfs, 1/2010, vol. 4
process. In such a population the group of elderly, i.e. people usually older than 64 years
or even better those of retired from the economic point of view2
, considerably enlarges
and becomes a significant social, economic and political unit.
The "graying" of the population in the developed part of the world is a result of three
factors. First, post-war "baby boom" generation, i.e. generation born between 1946 and
1964, comprising now 25% (Eurostat, 2009) of the European population is approaching
retirement age (Chamberlain and Prante, 2007), second, rising living standard and
female labour force participation supported by higher education cause that families
have decided to have fewer children which results in decline of the population growth
rate (Day and Dowrick, 2004).3
And last but not least, higher quality of health care, generous
social systems guaranteeing substantial minimum living standard and environment
without any serious war conflicts or environmental disasters, jointly contribute to
longevity.
Turner et al. (1998) provides a number of main channels the ageing affects the macroeconomics
through. First, changes in the labour force are supposed to be most directly
involved by ageing as labour force is expected to shrink and change its structure. Second,
increasing share of older workers could possibly have adverse effects on productivity since
they might be "less dynamic and innovative". On the other hand, the labour shortage
might work as an incentive to technological investments4
and findings of new substitutes
to scarce labour. Third, the "life-cycle" hypothesis reflects the influence on private saving
behaviour. However, the effect is ambiguous as it may stem from lower fertility or
greater life expectancy in the ageing process. "Under the life-cycle hypothesis, a rise in
life expectancy would lead individuals to save more during their working years in order to
maintain consumption over a longer retirement period, and so generate higher aggregate
private savings. By contrast, a decline in population growth due to lower fertility would
leave individual savings profiles unchanged, but lead to lower aggregate savings as the
proportion of the low-saving elderly in the population increased…" (Turner et al., 1998,
p. 12). Momentous changes in national savings might further influence flows of goods,
services and capital amongst economies and thus contribute to exchange rate and balance
of payments adjustments. The final channel is a frequent topic in nowadays academic
and policy research studies – the pressure of ageing on public finance, the aspect we
deal with in this study from the public financed and provided health care point of view.
Higher pension and both acute health and long-term care5
expenditures, possibly mitigated
by lower educational costs, represent the main expenditure drivers, which, along
with uncertain revenues, are challenging task for many researchers and policy-makers at
the present time.
2	 In demographic perspective, the end of reproduction period of women is taken into account (usually age of
50).
3	 Simultaneous influence of the effects of falling fertility and increasing longevity relates to a "double ageing".
4	 Cutler’s et al. (1990) evidence suggests that countries experience higher productivity growth if the labour
force develops slower.
5	 The term health care refers to the acute health and long-term care further in the text.
015Acta všfs, 1/2010, vol. 4
The gradually taking place ageing is likely to change demand for goods and services provided
by both private and public sector (Denton and Spencer, 1998). The product and
service structure is probable to be modified, some of them may disappear and others
may have to be developed and introduced entirely new. And even if the effect of some of
them is to appear in one, two or maybe three decades, it would be quite desirable to start
preparing new capacities as soon as possible in order to be ready for different demand (in
product market) and supply (in labour market) in every area of our life.
Ageing has far-reaching consequences associated with each of us. Changes in health and
long-term care might represent only a small fraction, however a significant one.
2	 Methodology
First of all we would like to point out that all calculations in the model are based on
demographic and economic projections which more or less predetermine the results. Estimates
of the future path of variables such as participation and unemployment rates,
labour productivity growth, life expectancy and total fertility rates play an important role
in the correctly calibrated model and its output. One should bear in mind that the results
of such studies are not an accurate forecast of the future. Rather, they present the possible
path stemming from the trend evolution of the projected variables and their mutual relationships.
Furthermore, they are based on a so-called "no-policy-change scenario", which
means that no future corrections and changes e.g. to tax bases, tax rates, transfer policies
or any other policy measures are taken into account.
The basic approach we have used for long-term projections of the economic variables
as well as the expenditures side of the public health care systems is the methodology of
the Ageing Working Group of the European Commission (2008b). However the model is
modified and augmented by the analysis of the revenues side of the healthcare systems
as well as by some features of the generational accounting method.
The projection model is further augmented by some elements of generational accounting,
a method originally developed by Auerbach et al. (1991) designed to evaluate and compare
the burdens of current and future generations stemming from recent government
policy. The method provides a very useful tool for public finance sustainability assessment
of the present arrangements. The calculations are thus made for every age and gender cohort
instead of approach per each year. Nevertheless, our technique allows us to calculate
the outcomes retrospectively in the other way as well.
The main idea behind the concept of the generational accounting is the fact that every individual
is both a taxpayer and, on the other hand, a transfer beneficiary during his life. In
the life cycle hypothesis6
, the various burdens and transfers arising out of the government
sector are age and gender specific. For example, personal income taxes are, in a simplified
way, paid only when an individual is working. Moreover, the amount of tax collected differs
according to the amount of the wage. Young and elderly people are assumed not to
pay these taxes. Alternatively, the higher life expectancy of women implies that old-age
6	 Developed by Ando and Modigliani (1963) – see Dybczak (2006), p. 3.
016 Acta všfs, 1/2010, vol. 4
pensions will be paid to women for a longer time than to men. Nonetheless, the average
old-age pension paid to men is supposed to be higher due to the higher levels of their
previously earned wages. In the health care sector the situation is in many aspects similar.
The retirees paying only consumption taxes are at major risk to a range of illnesses and
health difficulties, whereas people at working age bear the most costs connected with
the health care.
Broadly speaking, "…generational accounts indicate, in present value, what the typical
member of each generation can expect to pay, now and in the future, in net taxes (taxes
paid net of transfers received)" (Auerbach et al., 1994, p. 75). Considering the age dependency,
generational accounting takes into account the age profile of an average person,
presenting "…whether the tax and transfer-policy of a selected base-year can be maintained
into the indefinite future or whether sooner or later adjustments will be necessary
in order to meet the government’s intertemporal budget constraint" (European Commission,
1999, p. 1).
The following text describes how we have incorporated the generational aspect into the
projection technique, whereas six sections (modules) relating to the diverse projection
subjects are distinguished. The first one is devoted to the core element of all economic
projections – the demography development. The second module explains how to use
demographic projection to determine the future labour market development. Labour productivity
as well as gross domestic product are subjects of the third and fourth module
together with closer ruminating of constant factors of production shares. And last but
not least two final sections deal with health care expenditure and revenue estimates and
projections.
In this place, it is worth noting that the projection should not be affected by current
economic crisis as it is a long term projection. Values of selected variables will probably
differ from the observed ones also in a short time horizon but the aim of a long term projection
is not to predict future but to point out to possible problems arising from current
trends.
2.1	 Demography
The overall "exercise" starts with a demographic projection based on Eurostat’s EuroPop
2008 migration convergence scenario. The projection assumes that the socio-economic
and cultural differences between the EU Member States, Norway and Switzerland will
fade out in the very long run. This assumption implies convergence of demographic
values (see Table 1 for the Eurostat’s assumptions). The migration flows will converge
to zero net migration. The methodology consists essentially of setting the values of
the demographic indicators for the convergence year 2150, i.e. the year in which the
theoretical convergence would be achieved, and of appropriately interpolating from
the starting value for each country and each demographic component (fertility, mortality).
The national values for the year of interest (target year, 2060) are thus derived (in
appendix I “Population pyramids – ‘trees and coffins’ of life” can be found a graphical
description).
017Acta všfs, 1/2010, vol. 4
Table 1: Assumptions for the Eurostat’s EuroPop 2008 demographic projection
Country Total fertility rate
Life expectancy at birth Migration
assumptionsMales Females
Czech Republic 1.52 83.2 87.8 13.0
Denmark 1.85 84.3 88.4   6.5
Estonia 1.66 80.8 87.5  -0.1
France 1.93 85.1 90.1   6.0
Italy 1.55 85.5 90.0 19.9
Netherlands 1.77 84.9 88.9   3.0
United Kingdom 1.84 85.0 88.9 10.1
Switzerland 1.59 85.8 89.9 19.1
Source: Eurostat (2008c).
Additionally in Table 1, it is worth noting that the fertility rates are below the natural
population replacement rates, which are required to maintain population.
Table 2 gives a simple data description where economic, i.e. young and old-age, dependency
ratios are displayed. The economic (or total) dependency ratio reflects the decreasing
share and ageing of the working-age population. From the current level, ranging between
40.8% in the Czech Republic to 53.7% in France, the dependency ratio is supposed
to change dramatically. High increases in the Czech Republic (by 47.5 p.p. to 88.4%), Estonia
(by 35.8 p.p. to 82.8%) and Italy (by 33.0 p.p. to 84.9%) anticipate a rising ageing problem
that will have to be faced in the not-too-distant future. All the other countries considered
are confronted with a similar problem, even if some of them rather less so than the others.
Table 2: Economic dependency ratios (in %, p.p., 2008 – 2060)
Country 2008 2010 2020 2030 2040 2050 2060
Change
2060 – 2008
Czech Republic 40.8 42.1 53.8 57.0 64.8 79.9 88.4 47.5
Denmark 52.0 53.2 59.3 68.0 74.8 72.4 74.1 22.1
Estonia 47.0 47.3 56.3 60.1 63.9 75.0 82.8 35.8
France 53.7 54.5 63.2 69.9 76.1 77.6 77.6 23.8
Italy 51.9 52.4 56.7 64.0 77.8 84.2 84.9 33.0
Netherlands 48.6 49.2 56.3 67.9 76.5 75.1 76.7 28.2
Switzerland 47.0 47.5 52.7 61.9 67.4 69.6 73.3 26.4
United Kingdom 50.8 51.4 57.9 64.5 69.2 72.5 76.5 25.8
Note: The Economic dependency ratio is calculated as the ratio of the number of people aged
0 – 14 and 65 and older to those of working age (15 – 64). Since the Eurostat’s projection is
always dated to the 1st January, annual averages are presented in the table.
Source: Eurostat (2008c). Author’s calculations.
018 Acta všfs, 1/2010, vol. 4
2.2	 Labour market
Demographic projection is then used to calculate the total labour force ( ). Due to accessibility,
we use the European Commission’s age and sex-specific participation rates (defined
as the ratio of economically active men/women in a given year to the total number
of men/women of this age):
	 ,				 (1)
where stands for the total labour force and for the age ( x ) and sex ( s ) specific
participation rate in year t and xs
tPOP ,
is analogously the age ( x ) and sex ( s ) specific
part of the population. For the purposes of further proceedings we had to decompose
the labour force into employment and unemployment.
Several assumptions concerning employment and unemployment had to be made. First,
we hold constant the share of employees and the self-employed over the projection span.
Second, we do not distinguish between part-time and full-time work and all data are
taken for or recalculated to the full-time employment equivalent in the process. Next, we
presume that hours worked and the full-time and part-time ratio do not change. And finally,
the European Commission’s (2008a) NAWRU (Non-Accelerating Wage Rate of Unemployment)
values were considered as a proxy for structural unemployment rates (Carone,
2005). Current unemployment rates converge to these structural levels in the medium
term, with the speed of convergence derived from historical trends of unemployment
rates. The past and future estimated NAWRU values (the NAIRU – Non-Accelerating Inflation
Rate of Unemployment – was selected for Switzerland, because other data are not
available) can be seen in Table 3.
Table 3: Non-accelerating wage rate of unemployment (in %, 2000–2060)
Country 2000 2005 2008 2010 2015 2020 – 2060
Czech Republic   6.92 7.18 6.18 4.50 4.50 4.50
Denmark   5.17 4.30 3.77 3.25 3.25 3.25
Estonia 10.66 9.14 5.43 3.46 3.46 3.46
France   9.63 9.06 8.64 7.81 7.00 6.20
Italy   9.13 7.91 7.23 5.75 5.75 5.75
Netherlands   3.35 3.31 3.29 3.01 3.01 3.01
Switzerland*   3.27 3.73 3.69 3.69 3.70 3.70
United Kingdom   5.68 5.00 5.66 5.43 5.43 5.43
Note: *) Non-Accelerating Inflation Rate of Unemployment (NAIRU).
Source: For years 2000–2008 European Commission (2008a). For years 2010 – 2060 European
Commission (2008b). OECD (2008a) NAIRU values for Switzerland.
019Acta všfs, 1/2010, vol. 4
In order to preserve the NAWRU and NAIRU values over time, the age and sex-specific
unemployment rates were derived as follows (European Commission, 2008b, page 79):
	 			 (2)
where is the fixed unemployment rate for age x and sex s in year 2007 and
is the desirable sex and age-specific unemployment rate in year t . Therefore, the
structure of unemployment in 2007 is kept constant. The age and sex-specific employment
is then easily the product of the sex and age-specific participation, one minus unemployment
rate and the number of persons in the age-sex category:
	 .			 (3)
Total employment ( tL ) is calculated as the sum of all of the specific employments:
	 .					 (4)
2.3	 Labour productivity and gross domestic product
By multiplying total employment ( tL ) and labour productivity ( tPP ), we are able to compute
gross domestic product ( tGDP ) in real terms:
	 ttt PPLGDP ⋅= .						 (5)
For simplicity, we adopted the EC’s labour productivity growths summarised in Table 4.
A crucial assumption about labour productivity growth is that in all countries it converges
to 1.7% in 2050 (see European Commission, 2008b, p. 94). For short and medium-term development
the GDP growth was decomposed using the Cobb-Douglas production function
with constant returns to scale. The main drivers of labour productivity growth are thus
total factor productivity growth and capital deepening.
020 Acta všfs, 1/2010, vol. 4
Table 4: Labour productivity growth projection (in %, 2008–2060)
Country 2008 2010 2020 2030 2040 – 2060
Czech Republic 4.3 3.8 2.9 1.8 1.7
Denmark 1.9 1.9 1.8 1.7 1.7
Estonia 5.3 4.9 3.3 2.7 1.7
France 1.6 1.6 1.7 1.7 1.7
Italy 0.6 0.6 1.6 1.7 1.7
Netherlands 1.7 1.7 1.8 1.7 1.7
Switzerland 1.9 1.7 1.8 1.7 1.7
United Kingdom 2.6 2.1 1.8 1.7 1.7
Source: European Commission. Switzerland – Author’s calculation based on similar assumptions.
In other words, GDP growth ( tg ) is determined by the growth of total employment and
labour productivity growth, so the model does not assume any changes in the capitallabour
ratio:
	 ( ) 1
11
1 111 −
−−
− ⋅




 ∆
+⋅




 ∆
+=⋅+= t
t
t
t
t
ttt GDP
PP
PP
L
L
GDPgGDP .	 (6)
Moreover, the labour productivity development determines the progression of the average
real wage, which is essential for labour supply income afterwards (see below).
Nominal GDP ( N
tGDP ) is then a product of real GDP and the GDP deflator ( tdef ), which
converges to the European Central Bank’s inflation target, ensuring "…inflation rates close
to 2% over the medium term" (ECB, 2003):
	 ( )tt
N
t defGDPGDP +⋅= 1 .					 (7)
2.4	 Rationalising constant shares of factors of production
Arising from previous part of this chapter, the constant ratio of production factors per output
is considered in the long-term over the projection span. The stable shares assumption
might be verified directly or derived indirectly via theories of growth models. Herein, we
would like to warn of possible deceptive information given by all long-term time series of
indicators as their contents and definitions could have been changing in time.
As far as direct approach is concerned, Kaldor (1963) gives six basic "stylized facts" about
economic growth which also encounter constant shares of labour and physical capital in
national income. Barro and Sala-i-Martin (2003) then provide several other studies indicat-
021Acta všfs, 1/2010, vol. 4
ing the factors long-term stability. Gollin (2001) revised the Kaldor’s "fact" and assessed
that "…factor shares give estimates that are remarkably consistent with the claim that factor
shares are approximately constant across time and space…", moreover he suggests "…
to use models that give rise to constant factor shares" (p. 15). As well as Gundlach (2007)
asserts that "…the cross-country data on output per worker can be consistently summarized
by a specification that allows for international variation in technology conditional
on a constant capital output ratio" (p. 17). Analogously Bernanke and Gurkaynak (2001)
finds out "…the time series of labour shares by country tend to be quite stable, with no
systematic tendency to rise or fall over time" (p. 26).
Besides, some theories of growth models might support the stable shares hypothesis as
well – either those based on neoclassical Solow growth model, both basic and encompassing
Harrod-neutral technical change, as well as later Ramsey-Cass-Koopmans model
or Diamond’s overlapping generations model. The convergence to steady state observed
in reality is built in these models. The variables develop at constant rates in the steady
state and do not change the capital-labour ratio. Later theories coping with endogenous
growth of internalised technology could keep the ratio constant in the long-run, however,
there are some theories (e.g. pure "AK model") where the convergence to steady state is
broken and thus the ratio might vary in time.
Empirical verification of the growth models, which should help with selection of the right
model, is quite disputable. Some of them support the claim that the Solow model "…is
consistent with the international evidence if one acknowledges the importance of human
as well as physical capital…" (Mankiw, Romer, Weil, 1992, p. 433), nevertheless Bernanke
and Gurkaynak (2001) nail down that Mankiw’s, Romer’s and Weil’s "…basic estimation
framework is broadly consistent with any growth model that admits a balanced growth
path" (p. 1). Similarly, Gundlach (2007) concedes his empirical results are "…in line with the
Solow model…" (p. 17). Others like Okada (2006) contribute the Solow’s mechanism serves
as a good explanation of convergence amongst developed countries such as OECD but
not when taking into account less developed countries. To the contrary, study of Arnold et
al. (2007) suggests that the "…estimated speed of convergence appears to be too high to
be consistent with the human-capital-augmented version of the Solow model, but rather
support the endogenous growth model…" (p. 21).
Moreover, the verifications of the models with human capital seem to be problematic. Not
only they are tremendously data-consuming, they also often quantify qualitative categories
to specify human capital like level of knowledge, quality of health care, education, social security
environment, etc., substituted by questionable proxies in econometrical estimates.
Thus, analyzing developed countries, considering convergence of the economies and on
reflection of standard long-term projection methodology we have decided to develop
the model on the neoclassical economic framework with constant factor shares. The nonnegligible
impact of this choice will be presented in the chapter dedicated to results of
the projection model.
Having described the major macroeconomic prerequisites we can continue with government
health care expenditures and revenues.
022 Acta všfs, 1/2010, vol. 4
2.5	 Expenditure projection
In order to present different hypothesis about the evolution of the health care costs that
appear in the literature and to show possible ways of development arising from these
hypothesis we have developed four scenarios of health expenditure evolution based on
standard methods used in health care expenditure projection models such as the model
of the European Commission, the model of the OECD or the model of the WB.
The first scenario, called "pure ageing scenario", represents the basic projection method.
The health expenditure projection starts with age-related health care costs profiles for
acute and long-term care of the initial year that are applied to the demographic projection
for future years. The age-related health care cost profiles of the selected countries
expressed in per cent of GDP per capita denote average annual health care expenditure
for five-year categories. For the model these were smoothed to one-year categories. The
average health care costs for an individual of age x and sex s in the initial year 2007 are
multiplied by the number of individuals of age x and sex s in every year of the demographic
projection. The spending in a particular year can be described by the following
equation:
	 ∑∑= =
⋅=
2
1
110
0
,,
2007
s x
xs
t
xs
t POPcC 					 (8)
where
xs
c ,
2007 are the costs per capita for each sex and age that are given by the initial year
of the projection and
xs
tPOP ,
stands for the number of individuals in the defined age
and sex categories in year t given by the demographic projection.
Yet the above described method implies the so called "expansion of morbidity" hypothesis
that was first proposed already in 1977 by Grunenberg. According to the hypothesis, the
number of years of life spent in good health will remain constant and all the additional
years gained by increases of life expectancy will be spent in bad health due to chronic
illnesses and other causes. However more optimistic scenarios have been proposed in
the literature latter on. The so called "compression of morbidity" hypothesis, published
only three years later by Fries (1980), states on the contrary that the part of life spent in
bad health will remain constant and the health status of the population will thus improve.
In long term health expenditure projections, the hypothesis is represented by shifts of
the age-related health care costs profiles to the right in line with the growth of the life
expectancy. The medium way between the two hypothesis is the so called "dynamic equilibrium"
according to which the proportion of life lived in bad health will remain constant.
Last but not least, according to the death-related costs hypothesis according to which
the health expenditure is concentrated in last years of life the health expenditure will be
postponed to higher age in line with the growth of the life expectancy (for more details
see Pavlokova, 2009).
In order to represent the above mentioned hypothesis, the second scenario we develop
in the model is the so called "constant health scenario" taking into account the possible
improvements of the health status of the population or the postponement of health care
023Acta všfs, 1/2010, vol. 4
costs in line with the death related costs hypothesis. The constant health scenario represents
the dynamic equilibrium hypothesis (i.e. the medium way). The age-related health
care cost profiles are shifted by the change in life expectancy:
	
xs
texsxs
t cc
,
,
2007
, ∆−
= 					 (9)
where
xs
te ,
∆ is the change in life expectancy of a given gender s and age x between
year t and the base year 2007, i.e.
xsxs
t
xs
t eee ,
2007
,,
−=∆ . The final total costs tC can be
described by the following equation:
	
xs
t
s x
xs
tt POPcC ,
2
1
110
0
,
⋅= ∑∑= =
.				 (10)
The importance of the change in life expectancy is described by Table 5:
Table 5: Life expectancy at birth (in years, 2008–2060)
Country
2008 2030 2060
Diff. between
2060 and 2008
Males Females Males Females Males Females Males Females
Czech Republic 73.9 80.2 78.1 83.7 83.2 87.8   9.3 7.6
Denmark 76.4 81.0 80.0 84.5 84.3 88.4   7.9 7.4
Estonia 68.0 78.7 74.0 82.9 80.8 87.5 12.8 8.8
France 77.5 84.3 81.0 87.0 85.1 90.1   7.6 5.8
Italy 78.5 84.2 81.7 86.9 85.5 90.0   7.0 5.8
Netherlands 77.9 82.2 81.1 85.3 84.9 88.9   7.0 6.7
Switzerland 79.3 84.2 82.3 86.9 85.8 89.9   6.5 5.7
United Kingdom 77.4 81.5 80.9 85.0 85.0 88.9   7.6 7.4
Source: Eurostat (2008c).
Countries can be divided into three categories according to the projected change in life
expectancy at birth. The group of higher-than-average life expectancies consists of the
Czech Republic and Estonia, values around the average are represented by Denmark,
France, Italy, the Netherlands and the United Kingdom, and life expectancy at birth is
slightly below the average in Switzerland. The numbers fully correspond with the demographic
convergence scenario and the lower initial life expectancy values.
Furthermore, next important driver of health care expenditure, the income elasticity of
health care, cannot be omitted in a health expenditure projection model. The third and
fourth scenario called "pure ageing with elasticity" and "constant health with elasticity"
are an enlargement of the first and second scenario that take into account the income
024 Acta všfs, 1/2010, vol. 4
elasticity of health care expenditure converging in a linear manner from 1.1 to 1 in line
with the EC reference   scenario. The average health and long term care costs for gender
s and age x in year t are computed by multiplying the average costs for the age and
sex category from the previous year by sum of one and the multiple of the elasticity and
the growth of the GDP per capita between the two years:
	














−+⋅=
−−
− 1
/
/
1
11
,
1
,
t
N
t
t
N
t
t
xs
t
xs
t
POPGDP
POPGDP
cc e 		 (11)
where te represents the income elasticity of 1.1% diminishing linearly to 1.0% in 2060.
The final spending tC is described by equation 10.
Whether the income elasticity of health care is higher than one ore not has been subject
to large discussion in the literature with ambiguous results. The differences result broadly
from diverse methodological approaches used. We can however observe both in history
and in cross-country comparisons that the part of GDP spent on healthcare grows in line
with GDP per capita. This development can result also from another driver that can be
modelled by the above described methodology – the technological progress. The measurement
of the importance of the technological progress for health care expenditure is
however one of the very controversial ones. The importance of the technological progress
is usually measured as the residual in the regression model of total health care expenditure
on GDP per capita and an ageing-related variable. The use of the income elasticity
factor in our model with the methodology that is described above can be used as a proxy
for both factors.
2.6	 Revenue projection
The financing of the public healthcare systems differ by the mix of taxes in the selected
countries. Broadly, the general government sector tax-to-GDP ratio7 N
tt GDPT / might
serve as a tool in modelling government revenues. It can be disintegrated into three com-
ponents:
	 N
t
t
CN
t
t
KN
t
t
N
t
LN
t
t
GDP
C
t
GDP
K
t
GDP
LW
t
GDP
T
++
⋅
= 			 (12)
where tT represents the total amount of tax revenues in year t ,
N
tGDP is gross domestic
product in current ( t ) prices, Lt is the average effective tax rate of personal income
tax including social security contributions, N
tW is the average nominal wage, tL is total
employment, tK is the total value of capital, Kt is the average effective tax rate on capital,
tC is total consumption and Ct is the average effective tax rate on consumption.
7	 Tax-to-GDP ratio is the approximation of total government sector revenues.
025Acta všfs, 1/2010, vol. 4
The model requires associating the relevant types of taxes with age and gender. As for
government sector revenues affiliated with labour income assigned to every age and sex,
we determined: "1100 Taxes on income, profits and capital gains of individuals" and "2000
Social security contributions"8
. The sum of the variety of indirect taxes – "5111 Value added
taxes" and "5121 Excises" – is supposed to be born fully by individuals, which allows restoration
of their age profile. Table 6 then reveals the total amount of age-related taxes
included in the model.
Table 6: Age related taxes considered in the model (in % of the total tax revenues, 2006)
Country
Taxes on income
of individuals
Social security
contributions
Value
added tax
Excises
Total
age
related
taxes1100 2000 5111 5121
Czech Republic 11.5 43.9 18.0 10.1 83.5
Denmark 50.0   2.1 20.9   9.8 82.8
Estonia   7.1 40.3 32.9 12.1 92.4
France 20.1 37.0 16.4   5.5 79.0
Italy 25.7 29.9 14.9   5.4 75.9
Netherlands 18.9 36.4 18.7   8.6 82.7
Switzerland 35.0 23.6 13.2   5.5 76.9
United Kingdom 29.1 18.6 18.2   8.2 74.0
Note: Four-digit numbers and the revenue categories refer to the OECD international government
revenue classification.
Source: OECD (2008b). Ministry of Finance of Estonia (2008).
Final determination of the sources of financing needs to find out, what revenues are directly
linked to health care financing. That is why the following few short paragraphs
describe closer the systems of health care financing.
In the Czech Republic, social health insurance is compulsory and the contributions are
paid mostly by employees, employers and the self-employed. The State contributes in
the system on behalf of special groups of non-wage earners, such as children, pensioners,
parents on maternity leave, the unemployed, asylum seekers etc. Except for the State
contribution on behalf of non-wage earners, the State finances capital investments in
hospitals it manages, in training of medical personal etc. The social security contributions
cover thus only 69.6% of public health care spending.
The public health care system in Denmark is financed through general and municipal
taxes. Health care in the regions, which accounts for most of the health care spending, is
financed by four kinds of subsidies: a block grant from the State (75%), a State activityrelated
subsidy (5%), a local basic contribution (10%) and a local activity-related contribu-
8	 See OECD Revenue Statistics (2008) for the tax classification.
026 Acta všfs, 1/2010, vol. 4
tion (5%). The part coming from the State is financed mainly by health contributions that
account for 8% of wage.
In Estonia, the social health insurance contributions are paid by most of the employees
and the self-employed, the State contributes only for less than 3% of the non-wage earners
(individuals on parental leave, the unemployed, people receiving social benefits and
other minor groups). Major groups of non-wage earners (pensioners, children etc.) are
covered by the social health insurance without contributing and without the State contributing
for them.
In France, every working person and his/her dependants belong to a health insurance
scheme according to occupation.9
The revenues of the health insurance schemes are
raised from social health insurance contributions paid by employees, employers and selfemployed
(46% of revenues of the general scheme), from all income including retirement
pensions and invalidity and pre-retirement benefits and other replacement income,
investment income and property income (34%), from taxes on car insurance premiums
some alcoholic drinks, tobacco, advertising for pharmaceuticals and medical products etc.
(9%) and from other sources including government contribution (11%).
The main source of finance of the Italian National Health Service is general taxation but
the regions and autonomous provinces are entitled to set the level of regional taxes,
which account for 40% of public health care sector resources. The main regional tax dedicated
mostly to health care is the tax on productive activities (IRAP) which is a flat-rate
tax on the value added generated by all types of business and self-employed activities.
The second regional source of income of the Italian National Health Service is a personal
income surcharge.
The Dutch public health care system has two components, the first of which is the statutory
health insurance system financed by a mixture of social health insurance contributions
and nominal premiums paid by insurees and the second of which, covering the long
term care, is the AWBZ (defined under the Exceptional Medical Expenses Act) raising funds
purely from social health insurance contributions. The State contributes for the premiums
of children up to the age of 18 in the first component and a grant to the AWBZ.
In Switzerland, two thirds of public health expenditure is financed through mandatory
health insurance purchased on individual basis. The nominal premia differ among funds
and purchased policies and are not income-related. Children benefit from lower nominal
premia. The remaining one-third of public expenditure is financed by cantons from general
taxation and consists mainly of subsidies to institutional providers (hospitals, nursing
homes).
The National Health Service of the United Kingdom is funded mainly by general taxation
(76%), but also by national insurance contributions (19%) and user charges (5%). National
9	 In 2001, 84% of the population belonged to the general scheme (Régime Général) that covers employees
in commerce and in industry and CMU beneficiaries (CMU is a subsidiary system which provides medical
coverage for persons who do not benefit from any other existing medical coverage schemes).
027Acta všfs, 1/2010, vol. 4
insurance contributions are paid by employers and employees and are counted as general
government revenue in the National Health Accounts.
The systems based on social health insurance (Czech Republic, Estonia, France, Netherlands,
Switzerland) use mostly social security contributions to cover the public health
expenditure, however general taxation is often used in these systems to cover the costs
of selected groups of non-wage earners (children, pensioners) or capital investments in
publicly owned health care facilities. In Switzerland and partly in the Netherlands the
nominal premia replace social security contributions based on wages. The systems of National
Health Service (Denmark, Italy, United Kingdom) raise funds mostly from general
taxation but in some countries special taxes are assigned purely to health care.
Table 7 gives a quick preview of the structure of all considered tax revenues in 2007. The
structure of financing of 2007 as the initial year is preserved over time in the projection.
Table 7: Structure of the government sector sources financing acute and long-term care
(in %, 2007)
Country
Social security
contributions
Nominal
premia
Personal
income tax
Value added
tax and excises
Other tax
revenues
Czech Republic 69.6 -   6.2 15.2   8.9
Denmark   0.0 - 51.1 31.4 17.5
Estonia 82.3 -   3.8   6.1   7.8
France 93.6 -   2.1   2.2   2.1
Italy   0.1 -   9.6 72.7 17.6
Netherlands   0.0 31.4 62.0   4.1   2.6
United Kingdom   0.0 - 35.7 32.4 31.9
Switzerland   0.0 71.4 13.0   7.0   8.6
Source: OECD (2008b), OECD (2008c). Author’s calculations.
Now, step further to the modelling of every relevant type of taxes. Computing labour income
taxes, the effective tax rates of social contributions and personal income taxes are
applied to the tax base in every following year. The tax base for every age and sex category
is equal to the product of the number of employed persons in every age and sex category
and the modelled average wage in the age and sex category. The average effective tax
rates ( PITt and SSCt ) are at the same time held constant over the projected period.
	
xs
t
xsN
tSSCPITSSCPIT LWttTT ,,,
)( ⋅⋅+=+ 			 (13)
The age and sex-specific nominal wage ( xsN
tW ,, ) is supposed to grow in line with an average
nominal wage ( N
tW ) that rises by the percentage change of labour productivity
(  tPP  ) and the GDP deflator ( tdef ), which for the sake of simplicity is taken as the rate
of growth of the price level:
028 Acta všfs, 1/2010, vol. 4
	
( ) ( ) N
tttt
N
t
xsN
t WdefPPPPWW 11
,,
1/1 −− ⋅+⋅∆+== .		 (14)
Whereas indirect taxes are said to be typically proportional to their tax base (see André
and Girouard, 2005; or European Commission, 2005), personal income taxes are usually
constructed as progressive10
, i.e. the higher the tax base, the higher the rate of the tax
(the so-called marginal tax rate). In such a system of taxation, increasing the average
wage would raise the effective tax rate, which might boost the total tax burden and the
size of the government sector in the economy indefinitely. That is why we assume all effective
tax rates to be constant in the projection, rather than using constant statutory tax
rates. A similar problem arises for gradual tax progressivity (tax brackets) and taxflation,
the process of shifting of all incomes to the upper tax brackets and taxing them at higher
statutory marginal tax rates.
The age distribution of value added tax and excise revenues is calculated from the age
distribution of consumption as defined by COICOP11
per adult equivalent. The effective tax
rates for age categories are applied to the demographic projection. The ratio of consumption
to GDP is held constant over the projected period.
Other revenues, which are not connected with age or sex, are held constant to GDP over
the projected span and are proportionally allocated to the total population. Thus, we
consider them to be a lump sum proportionally divided per capita:
	 			 (15)
where denotes the age and sex-specific unit of other government sector revenues,
the total value of other revenues in 2006 and N
GDP2006
the nominal gross domestic
product in 2006.
In summary, tax incidence is another simplifying assumption, especially with regard to
taxes that cannot be simply assigned to a specific age and gender. Capital income taxes
represent one example. A variety of studies presume that capital income taxes are treated
in the same way as personal income tax (Cardarelli et al., 2000, van Ewijk et al., 2002, Gál
10	 A statutory flat tax rate does not necessarily mean that the tax is proportional to its base. Flat taxes are
usually computed as the product of a constant (the statutory tax rate) and the difference between income
(or another tax object) and a stable deductible item (alternatively there is a deduction from the calculated
tax). Thus, deductible items/tax deductions imply progressivity of the flat tax, since the higher the value of
the object of the tax, the higher the relative tax base (see Kim et al., 2006, for details).
11	 COICOP stands for Classification of Individual Consumption According to Purpose. The decomposition of
individual consumption and assignation of taxes was made on the basis of the following sources: Deloitte
& Touche (2004), Taxation.ch (2009), European Commission (2007), Eurostat (2008d), OECD (2008b) and
European Commission (2009).
029Acta všfs, 1/2010, vol. 4
et al., 2005). In other words, they assume these types of taxes might be divided according
to the nominal wage, thus fully borne by employees. Nonetheless, there are at least three
circumstances which complicate this decision. First, capital taxes may be partly included
in the price of current or old assets; second, the tax payments may differ from the income
on the assets. And third, the view of both the theoretical and empirical literature regarding
tax incidence is not unambiguous.
Now we are prepared to modify the equation No. 12 in order to express age and sexrelated
government revenues:
	 	 (16)
where PITt stands for the average effective personal income tax rate, SSCt for the average
effective tax rate of social security contributions (paid by employees, employers and
the self-employed), xsN
tW ,,
for the age ( x ) and sex ( s ) specific nominal wage,
xs
tL ,
for
employment broken down by age ( x ) and sex ( s ), ITt for the effective indirect tax rate,
xs
tC ,
for age and gender-specific household consumption and finally for all other
revenues.
With respect to the previously derived kinds of taxes, the average effective tax rates were
calculated as the ratio of the tax part used for health care funding and its tax base. In
the case of social security and personal income effective tax rates, the product of total
employment and the average nominal gross wage was applied as the tax base. Effective
consumption tax rates were calculated by dividing the part of value added tax and excise
revenues that serves for health care financing by the final consumption expenditure of
households.
The next section summarises and discusses the results and main findings from the projec-
tion.
3	 Results
Recapping expenditure scenarios, we have calculated four different expenditure developments:
conventional pure ageing scenario (no shifts of cost-profiles are considered),
pure ageing scenario enriched by heightened income elasticity of health care (the elasticity
gradually declines to 1.0 in 2060 from 1.1 in the base year), constant health scenario
taking into account life expectancy change reflecting improvements in health of
the population, and finally constant health scenario with increased health care income
elasticity. The results in the terms of deficits of public finance health care financing are
depicted in Table 8.
The average increase of health care expenditure per GDP equals 36% in the selected
countries. Nevertheless, different health expenditure scenarios change the picture quite
dramatically. Without adjusting a health cost profile, the pure ageing scenario belongs to
the middle pathway of the future health expenditure development in all cases. However
030 Acta všfs, 1/2010, vol. 4
together with the income elasticity, the scenario seems to be the most powerful expenditure
driver, where the elasticity contributes 1% GDP on average in the last projection
year, which is approximately 2.3 percentage points higher than the lowest scenario. Next,
similarly constant health scenario with elasticity assigns itself to the middle of considered
expenditure variants. And there is another interesting point we would like to emphasise.
Whilst the pure ageing scenario with elasticity and the constant health scenario
extend similarly in all selected countries, the pure ageing scenario without elasticity and
constant health scenario with elasticity develop differently. In almost all countries, with
the exemption of Estonia, the constant health scenario with elasticity exceeds the pure
ageing scenario in the beginning, which sooner or later transcends the former scenario,
as the non-changed health status starts dominating. On the other side of spectrum, the
constant health scenario represents the most modest expenditure development ranging
from 0.4 in the Switzerland, across 1.5 in the Czech Republic, Denmark or Italy, to 3.2 in
the Netherlands.
Constant shares of factors of production discussed in chapter 2.4 coupled with the assumption
of equal growth rate of consumption and the GDP implies, under the assumption
of constant effective tax rates, constant share of the revenues of the public health
care systems on GDP.
Actually, sources of the public health care systems raised by any kind of personal income
tax or social security contributions are proportional to the volume of wages and salaries
that evolves in line with GDP. Indeed, the GDP growth in the model is the result of the
labour productivity growth (and thus the growth of the real wage) and the total employment
that are the two drivers of the volume of wages and salaries. Constant share of
revenues coming from consumption taxes and excise duties results from the assumption
of constant effective tax rates coupled with constant share of consumption on GDP which
is assumed on the balanced growth path in standard growth models. The deficits of the
public health care systems the evolution of which is expressed in table 8 thus result purely
from higher growth of expenditures than the GDP.
Table 8 reveals the differences in the height of deficits amongst countries. Our projection
estimates suggest that the Netherlands is expected to face the largest pressure
on public finance: from 3.2 to 6.3% deficits in terms of GDP, i.e. more than double in
comparison with other countries. This striking imbalance has its roots in the overproportional
height of long-term care costs, which are more vulnerable to population
ageing and increase the sensitivity of the Dutch system to economic dependency ratio.
Actually, the AWBZ covers not only pure health care but also part of the social care
offered to the elderly that is financed differently in other countries. Slightly aboveaverage
public deficit might be envisaged in Denmark, conversely, below average in
Estonia, Italy, the United Kingdom and also in France, despite of high expenditureto-GDP
ratios. The quite plausible development in Estonia (a difference amounting
to 1.2% less than average in 2060) is mainly due to lower long-term care expenditure
per capita. The Czech Republic and Switzerland are the two countries with closed-toaverage
deficits in the area of public health care, with both revenues and expenditures
lower than usual.
031Acta všfs, 1/2010, vol. 4
Table 8: Deficits of the public finance health financing under different expenditure scenarios
(in % GDP, 2008 – 2060)
Country Expenditure scenario 2008 2010 2020 2030 2040 2050 2060
Czech Rep. Pure Ageing Scenario 0.05 0.16 0.66 1.30 1.86 2.30 2.69
Pure Ageing Scenario with Elasticity 0.10 0.30 1.13 2.00 2.76 3.32 3.79
Constant Health Scenario 0.03 0.08 0.33 0.68 0.99 1.23 1.47
Constant Health Scenario with Elasticity 0.08 0.22 0.77 1.33 1.79 2.12 2.42
Denmark Pure Ageing Scenario 0.03 0.10 0.79 1.78 2.44 2.96 3.16
Pure Ageing Scenario with Elasticity 0.07 0.22 1.18 2.43 3.31 3.99 4.25
Constant Health Scenario 0.00 0.01 0.33 0.95 1.29 1.52 1.50
Constant Health Scenario with Elasticity 0.04 0.12 0.70 1.54 2.06 2.41 2.43
Estonia Pure Ageing Scenario 0.02 0.06 0.24 0.44 0.76 1.03 1.26
Pure Ageing Scenario with Elasticity 0.09 0.24 0.71 1.10 1.57 1.94 2.22
Constant Health Scenario 0.00 0.02 0.06 0.11 0.26 0.41 0.55
Constant Health Scenario with Elasticity 0.07 0.20 0.50 0.73 1.00 1.23 1.39
France Pure Ageing Scenario 0.06 0.16 0.62 1.15 1.75 2.06 2.21
Pure Ageing Scenario with Elasticity 0.09 0.26 1.00 1.78 2.59 3.04 3.25
Constant Health Scenario 0.03 0.09 0.31 0.59 0.96 1.07 1.04
Constant Health Scenario with Elasticity 0.07 0.18 0.68 1.18 1.74 1.96 1.96
Italy Pure Ageing Scenario 0.05 0.15 0.59 1.08 1.63 2.10 2.25
Pure Ageing Scenario with Elasticity 0.08 0.23 0.91 1.61 2.32 2.92 3.12
Constant Health Scenario 0.03 0.09 0.32 0.59 0.94 1.26 1.30
Constant Health Scenario with Elasticity 0.06 0.17 0.63 1.09 1.58 2.00 2.07
Netherlands Pure Ageing Scenario 0.11 0.29 1.19 2.57 3.91 4.93 5.11
Pure Ageing Scenario with Elasticity 0.15 0.38 1.56 3.21 4.82 6.06 6.30
Constant Health Scenario 0.08 0.20 0.74 1.69 2.59 3.26 3.22
Constant Health Scenario with Elasticity 0.12 0.29 1.09 2.27 3.40 4.23 4.22
Switzerland Pure Ageing Scenario 0.05 0.13 0.60 1.24 2.05 2.57 2.84
Pure Ageing Scenario with Elasticity 0.09 0.24 0.95 1.79 2.80 3.46 3.78
Constant Health Scenario 0.02 0.06 0.29 0.66 1.21 1.49 1.55
Constant Health Scenario with Elasticity 0.06 0.16 0.62 1.17 1.89 2.28 2.37
United Kingdom Pure Ageing Scenario 0.04 0.10 0.41 0.90 1.45 1.90 2.19
Pure Ageing Scenario with Elasticity 0.07 0.20 0.76 1.46 2.21 2.79 3.14
Constant Health Scenario 0.01 0.03 0.08 0.28 0.61 0.84 0.93
  Constant Health Scenario with Elasticity 0.05 0.12 0.42 0.80 1.30 1.63 1.75
Source: Author’s calculations.
Generally, the gap between projected health care sector expenditures and revenues implies,
ceteris paribus, important fiscal imbalances. The question rising from the results is
whether society will choose to satisfy the growing demand for health care from public
sources and increase the effective tax rates or whether it will shift the burden onto individuals
or whether the demand for health care will be restricted in some way. The second
two alternatives seem to be inappropriate for the elderly, who cannot increase their effort
to raise their labour income.
032 Acta všfs, 1/2010, vol. 4
The division of expenditures and revenues between different generations in every year
of the projection (for details in all the selected countries see Annex II, with a graphical illustration
of the development of generational revenues and expenditures covering acute
and long-term health care) shows important differences between who bears the costs and
who benefits from health care sector expenditures. Unlike in pensions, the entitlement in
health care is never precise. The easiest way to cut expenditures in health care might thus
be simply not to provide care. The most sobering finding of the study is the vulnerability
of the large elderly populations to cuts in health care expenditures arising from the possible
unwillingness of young generations to cover the ever increasing health care sector
expenditures.
As for the illustration of the intergenerational burden carried by population, we have
calculated "self-sufficiency index", comparing revenues (in the view of public finance) and
public expenditures covering health care for those older than 65 year old.
In Table 9 one can distinguish four groups of countries. On the one hand Italy and Switzerland
show lowest intergenerational solidarity and thus highest self-sufficiency of the
elderly (approximately 33 and 30%). The fact stems from the use of the nominal premia
for financing healthcare in the Switzerland and from high proportion of consumption
taxes used to finance healthcare in Italy. On the other hand we find France, where the
self-sufficiency belongs to smallest in Europe (just above 5%) even if our calculations
estimate a tendency to increase up to 1.8 percentage points. High intergenerational solidarity
in France results from prevailing financing based on social security contributions
and personal income tax and it may induce future intergenerational imbalances and tensions.
Then, the "self-sufficiency index" in the Czech Republic, Denmark, Estonia and the
Netherlands moves predominantly inside of the 10 to 16% interval. And finally the United
Kingdom, a special case, whose index culminates around 21% and lies between the others
as the National Health Service is financed more or less uniformly by all the main tax
aggregates (see Table 7).
From the point of view of the different systems of financing healthcare we can state, based
on the index, that systems of National Health Service show generally higher self-sufficiency
of the elderly. Actually, these systems use more consumption taxes that are paid by all
the population. The public health insurance systems raise revenues mostly from social
security contributions (i.e. contributions based on wages) and they thus reveal higher
intergenerational solidarity. However, as we mentioned above, the division of the systems
is not so simple as the systems of public health insurance may use nominal premia
or partly general taxation as well and on the other hand the systems of National Health
Service are in some countries financed by personal income taxes assigned to healthcare
such as in Denmark.
The indexes seem to fluctuate nearly around a subtly decreasing trend in time, or possibly
around a growing trend in the Czech Republic and France (in Italy and United Kingdom the
constant health scenario indicates increasing direction as well) and there are no important
changes in the indexes such as in the economic dependency ratio. Thus, any considerable
improvement of the old-age sufficiency and decreasing vulnerability of the elderly to the
policy decisions about redistribution might not be expected under the current unchanged
033Acta všfs, 1/2010, vol. 4
conditions. The important question that arises now is whether the working population will
be able to cover the healthcare costs of the dependent population if, based on the index,
we do not expect the dependent population to cover more of their costs. The proportion
of the costs of the dependent population that can be covered by the working population
without changes in the effective tax rates is summarised in Table 10.
Table 9: Self-sufficiency index of the elderly people (65+) in financing health care (in %
expenditure, 2008-2060)
Country Expenditure scenario 2008 2010 2020 2030 2040 2050 2060
Czech Rep. Pure Ageing Scenario 10.4 10.5 12.8 11.7 11.9 12.7 12.2
Pure Ageing Scenario with Elasticity 10.3 10.3 12.0 10.7 10.7 11.4 10.9
Constant Health Scenario 10.4 10.6 13.3 12.3 12.8 14.1 13.6
Constant Health Scenario with Elasticity 10.3 10.4 12.5 11.3 11.6 12.6 12.1
Denmark Pure Ageing Scenario 16.2 16.4 16.4 15.8 15.8 14.5 14.6
Pure Ageing Scenario with Elasticity 16.1 16.1 15.7 14.8 14.5 13.2 13.3
Constant Health Scenario 16.3 16.6 17.4 17.4 17.8 16.8 17.3
Constant Health Scenario with Elasticity 16.2 16.3 16.6 16.3 16.4 15.3 15.7
Estonia Pure Ageing Scenario 12.8 12.8   9.8 10.1   9.6   9.8   9.5
Pure Ageing Scenario with Elasticity 12.6 12.3   9.1 9.0   8.4   8.5   8.2
Constant Health Scenario 12.8 12.8 10.1 10.6 10.3 10.7 10.6
Constant Health Scenario with Elasticity 12.6 12.4   9.3 9.5   9.1   9.4   9.2
France Pure Ageing Scenario   2.9 3.4   5.1 4.8   4.8   4.4   4.3
Pure Ageing Scenario with Elasticity   2.9 3.4   4.9 4.5   4.4   4.0   3.9
Constant Health Scenario   2.9 3.5   5.3 5.1   5.2   4.8   4.7
Constant Health Scenario with Elasticity   2.9 3.4   5.1 4.8   4.8   4.4   4.3
Italy Pure Ageing Scenario 34.4 34.1 33.5 33.4 34.0 33.1 32.4
Pure Ageing Scenario with Elasticity 34.3 33.8 32.1 31.3 31.5 30.3 29.6
Constant Health Scenario 34.5 34.4 34.6 35.4 36.8 36.3 35.8
Constant Health Scenario with Elasticity 34.4 34.0 33.2 33.2 34.1 33.2 32.7
Netherlands Pure Ageing Scenario 14.4 14.5 15.6 14.3 13.1 11.3 10.7
Pure Ageing Scenario with Elasticity 14.4 14.3 14.9 13.5 12.2 10.3 10.4
Constant Health Scenario 14.5 14.7 16.6 16.0 15.2 13.3 13.7
Constant Health Scenario with Elasticity 14.4 14.6 15.9 15.0 14.0 12.2 12.5
Switzerland Pure Ageing Scenario 31.1 31.1 31.4 31.1 29.2 27.4 27.1
Pure Ageing Scenario with Elasticity 30.9 30.6 30.0 29.1 26.9 25.1 24.7
Constant Health Scenario 31.2 31.4 32.8 33.5 32.1 30.9 31.1
Constant Health Scenario with Elasticity 31.0 30.9 31.3 31.3 29.6 28.2 28.3
United Kingdom Pure Ageing Scenario 21.4 21.3 21.8 21.3 20.7 19.6 20.1
Pure Ageing Scenario with Elasticity 21.3 21.1 20.8 19.9 19.1 17.9 18.2
Constant Health Scenario 21.5 21.6 23.1 23.5 23.6 22.9 24.0
  Constant Health Scenario with Elasticity 21.4 21.3 22.1 21.9 21.7 20.8 21.8
Source: Author’s calculations.
034 Acta všfs, 1/2010, vol. 4
Table 10: Proportion of the healthcare costs of the dependent population that the working
population can cover without changes in effective tax rates (in %, 2010-2060)
Country Scenario 2010 2020 2030 2040 2050 2060
Czech Rep. Pure Ageing Scenario 76.9 62.3 51.9 42.7 41.3 40.6
Pure Ageing Scenario with Elasticity 72.5 52.9 40.8 31.6 31.2 31.2
Constant Health Scenario 77.4 64.1 54.6 45.8 45.2 44.9
Constant Health Scenario with Elasticity 74.9 60.8 53.0 46.1 45.7 46.0
Denmark Pure Ageing Scenario 71.5 61.3 49.4 44.7 42.0 39.8
Pure Ageing Scenario with Elasticity 69.3 55.9 43.2 38.0 34.8 32.7
Constant Health Scenario 69.3 55.9 43.2 38.0 34.8 32.7
Constant Health Scenario with Elasticity 71.2 62.7 52.2 48.4 46.8 46.0
Estonia Pure Ageing Scenario 80.8 74.6 68.9 60.9 55.7 55.5
Pure Ageing Scenario with Elasticity 73.8 60.0 51.3 42.8 39.1 40.4
Constant Health Scenario 81.2 76.3 71.9 64.7 60.4 61.2
Constant Health Scenario with Elasticity 75.6 66.7 61.7 56.5 53.2 54.3
France Pure Ageing Scenario 88.6 78.1 71.1 65.2 62.4 61.1
Pure Ageing Scenario with Elasticity 86.6 71.6 62.5 55.8 52.3 50.8
Constant Health Scenario 89.2 80.4 74.9 69.8 67.7 67.2
Constant Health Scenario with Elasticity 88.2 77.1 70.8 65.5 63.8 64.2
Italy Pure Ageing Scenario 56.1 46.5 37.4 31.6 29.1 28.2
Pure Ageing Scenario with Elasticity 54.6 41.6 31.0 25.4 23.0 22.0
Constant Health Scenario 56.5 47.9 39.4 34.0 31.7 30.9
Constant Health Scenario with Elasticity 55.8 46.2 37.9 32.9 31.0 31.0
Netherlands Pure Ageing Scenario 75.4 59.3 46.3 39.4 35.2 34.9
Pure Ageing Scenario with Elasticity 75.3 55.8 41.9 34.8 30.4 29.3
Constant Health Scenario 78.1 64.4 52.4 46.1 42.1 41.6
Constant Health Scenario with Elasticity 77.0 61.6 49.5 43.0 39.2 39.1
Switzerland Pure Ageing Scenario 57.1 47.1 37.8 31.6 28.4 27.1
Pure Ageing Scenario with Elasticity 55.0 41.7 31.7 25.4 22.2 21.1
Constant Health Scenario 57.6 49.0 40.5 34.6 31.7 30.8
Constant Health Scenario with Elasticity 56.4 46.9 38.7 33.3 31.1 30.9
United Kingdom Pure Ageing Scenario 65.7 59.8 52.2 47.1 42.6 40.2
Pure Ageing Scenario with Elasticity 63.9 54.3 45.2 39.4 34.7 32.7
Constant Health Scenario 66.5 62.7 56.8 52.7 48.9 47.2
  Constant Health Scenario with Elasticity 65.4 59.7 53.3 48.4 44.8 43.5
Note: The values are calculated as the ratio of the difference of the revenues from the 15-64 years
old and their expenditures and the sum of the expenditures of the 0-14 years old and the 65+.
Source: Author’s calculations.
According to the Table 10, the proportion of healthcare expenditure of the dependent
population that can be covered by the working population without changes in the effective
tax rates will drop dramatically in all of the analysed countries. The phenomenon is in line
with the increasing economic dependency ratio as the demographic projections expect the
drop of the proportion of the economically active population on total population.
035Acta všfs, 1/2010, vol. 4
Based on the self-sufficiency index and the values from table 10, we can expect that if the
effective tax rates are not increased the large elderly population will be menaced by important
rationing of the healthcare. Countries with high intergenerational solidarity such
as France or Estonia are especially vulnerable to such a development.
Conclusion
Population ageing as a long-term process seems to be inevitable in the European economies
during next 50 years, at least according to all available demographic projections.
Ageing will have significant impact on the life of current and future generations affecting
the market of goods, services, inputs and financial products. The effects of ageing on longterm
and acute health care with a special focus on intergenerational solidarity have been
the main concern of this paper.
Having built a macroeconomic long-term projection model, we have outlined and computed
four expenditure scenarios, which might characterise future health care development
in the selected eight countries denoting different systems of the health care financing.
Population ageing coupled with increasing expectations of society concerning access
to and quality of health care represent an important driver of health care costs. In order
to satisfy the needs of the population, public health expenditure as a percentage of GDP
would have to grow by 36% on average in the selected countries. The results show that
the expenditures will be mostly boosted (up to 53%) by considering no shifts in costprofiles
and pondering initially higher elasticity to diminish to unit one, i.e. pure ageing
scenario with elasticity. On the contrary, the restrained expenditure development could
be expected if the constant health scenario projecting the life expectancy change into the
health care costs is the right future expenditure driving force. The expenditure driven by
this scenario may increase by 20%.
Bearing in mind constant tax-to-GDP development in the future, expenditure scenarios
imply noticeable fiscal imbalances. The average deficit necessary to cover health care expenditure
in 2060 varies from 1.4 to 3.7% GDP, depending on the selected scenario. However,
in some countries like the Netherlands, the deficit will most likely range from 3.2 to
6.3% GDP, which might be substantial burden for the public finance.
Yet the flip side of the health expenditure is who is going to cover the increasing costs.
Current revenue collection systems differ considerably across the European countries.
Some systems are dominated by intergenerational solidarity (the healthcare expenditure
is financed mostly by the working population) others raise revenues of the public healthcare
more equally either from nominal premia or from consumption taxes. In order to
express the intergenerational solidarity of the systems we have developed the so called
"self-sufficiency index". We have found that France is a leading country in solidarity of
financing health care. As well former communist countries like the Czech Republic and
Estonia show lower self-sufficiency index of the elderly. Denmark and the Netherlands
with indices between 14 to 16% also belong to the more "generous" countries. Then the
United Kingdom and in particular Switzerland and Italy have the least vulnerable elderly
people from the studied point of view. The current state of the self-sufficiency is not to
be changed markedly during next 50 years. The question we tried to answer to was thus
036 Acta všfs, 1/2010, vol. 4
whether the working population will be able to cover the remaining healthcare costs of
the dependent population. The answer to the question, under the assumption of constant
effective tax rates, is however one of the very unfavourable ones in all of the analysed
countries. The proportion of healthcare costs of the dependent population that the working
population could cover after covering their own healthcare will drop on average by
45% by 2060 (from current 74% to 41%).
The imbalance between who bears the costs and who benefits from the health care systems
reveals a possible future risk of health care being denied to the large elderly generations.
Situation may require special attention especially if the elderly people have objectively
some difficulties connected with earning additional income to ensure the level of
the health care they need. Decisions regarding this problem are not trivial and request
deep knowledge of the whole social system since the problematic is interconnected with
pension system and system of social network specific for every country.
The obligation of policy-makers is thus to design measures that will ensure accessible
health care for all. Possible measures include policies aiming at higher flexibility of the
labour markets allowing the elderly populations to remain in the labour force and thus
decreasing the negative impact of ageing, adequate settings of the pension systems motivating
to remain in the labour force if possible, support of the savings of households
aimed at financing the long term care when needed and last but not least standardizing
of the provision of healthcare in order to limit the rationing of healthcare for the elderly.
References
ANDRÉ, Ch.; GIROUARD, N. (2005). Measuring Cyclically-Adjusted Budget Balances for
OECD Countries. OECD Economics Department 2005, Working Paper No. 434, 40 p.
ARNOLD, J.; BASSANINI, B.; SCARPETTA, S. (2007). Solow or Lucas? Testing Growth Models
Using Panel Data from OECD Countries. OECD 2007, Economics Department Working Paper,
No. 592, 25 p.
AUERBACH, A.; GOKHALE, J.; KOTLIKOFF, L. J. (1991). Generational Accounting: A Meaningful
Alternative to Deficit Accounting. National Bureau of Economic Research 1991,
Working Paper No. 3589, 71 p.
AUERBACH, A.; GOKHALE, J.; KOTLIKOFF, L. J. (1994). Generational Accounting: A Meaningful
Way to Evaluate Fiscal Policy. The Journal of economic Perspectives, Vol. 8, No. 1,
Winter 1994, pp. 73-94.
BARRO, R. J.; SALA-I-MARTIN, X. (2003). Economic Growth. 2nd Edition, the MIT Press
2003, 672 p.
BERNANKE, B. S; GURKAYNAK, R. S. (2001). Is Growth Exogenous? Taking Mankiw, Romer and
Weil Seriously. National Bureau of Economic Research 2000, Working Paper No. 8365, 54 p.
CARDARELLI, R.; SEFTON, J.; KOTLIKOFF, L. J. (2000). Generational Accounting in the UK.
Economic Journal, Royal Economic Society, Vol. 110, No. 467, pp. 547-574.
CARONE, G. (2005). Long-Term Labour Force Projections for the 25 EU Member States. A Set
of Data for Assessing the Economic Impact of Ageing. European Commission 2005, Economic
Paper No. 235, 214 p.
CARONE, G.; COSTELLO, D.; GUARDIA, N. D.; MOURRE, G.; PRZYWARA, B.; SALOMÄKI,
A. (2005). The Economic Impact of Ageing Populations in the EU25 Member States. European
Commission 2005, Economic Paper No. 236, 57 p.
037Acta všfs, 1/2010, vol. 4
CUTLER, D. M.; POTERBA, J. M.; SHEINER, L. M.; SUMMERS, L. H. (1990). An Aging Society:
Opportunity or Challenge? Brookings Papers on Economic Activity, Vol. 21, No. 1, 1990,
pp. 1-74.
DAY, C.; DOWRICK, S. (2004). Ageing Economics: Human Capital, Productivity and Fertility.
Agenda, Vol. 11, No. 1, 2004, pp. 3-20.
Deloitte & Touche (2004). Federal Law with regard to Value Added Tax. Zurich, Deloitte &
Touche AG 2004, issued by the VAT competence centre of the Swiss Institute of Certified
Accountants and Tax Consultants, 44 p.
DENTON, F. T.; SPENCER, B. G. (1998). Economic Costs of Population Aging. Research Institute
for Quantitative Studies in Economics and Population, McMaster University, 1998,
Research Report No. 339, 59 p.
DYBCZAK, K. (2006). Generational Accounts in the Czech Republic. Czech Journal of Economics
and Finance, Vol. 56, 2006, No. 7-8, pp. 298-317.
ECB. (2003). The ECB’s Monetary Policy Strategy. Governing Council of the ECB, 8 May 2003.
ECONOMIC POLICY COMMITTEE AND EUROPEAN COMMISSION. (2005). The 2005 EPC
Projections of Age-Related Expenditure (2004-2050) for the EU25 Member States: Underlying
Assumptions and Projection Methodologies. Brussels, European Economy Reports and Studies
No. 4, 36 p.
EUROPEAN COMMISSION (2005). New and Updated Budgetary Sensitivities for the EU
Budgetary Surveillance. Brussels, European Commission, Directorate General, Economic
and Financial Affairs, 30. September 2005.
EUROPEAN COMMISSION (2006). The Impact of Ageing on Public Expenditure: Projections
for the EU25 Member States on Pensions, Health care, Long-term Care, Education and
Unemployment Transfers (2004-2050). Economic Policy Committee a European Commission
2006, Special Report No. 1, 207 p.
EUROPEAN COMMISSION. (2007). Study on Reduced Value Added Tax Applied to Goods and Services
in the Member States of the European Union. Copenhagen Economics, DG TAXUD, 2007.
EUROPEAN COMMISSION. (2008a). Annual Macro-Economic Database [on-line]. 23 October
2008, [cit. 15.9.2008], Internet access: <http://ec.europa.eu/economy_finance/ db_in-
dicators/db_indicators8646_en.htm>.
EUROPEAN COMMISSION. (2008b). The 2009 Ageing Report: Underlying Assumptions and
Projection Methodologies. European Commission 2008, Directorate-General for Economic
and Financial Affairs, European Economy No. 7, 221 p.
EUROPEAN COMMISSION. (2009). ValueAddedTaxRatesAppliedintheMemberStatesoftheEuropean
Community. European Commission 2009, Taxation and Customs Union, Doc. No. 2401.
EUROPEAN COMMISSION. (1999). Generational Accounting in Europe. Brussels, European
Commission 1999, Directorate-General for Economic and Financial Affairs, European
Economy No. 6, 215 p.
EUROSTAT. (2008a). Annual National Accounts [on-line]. 2008, [cit. 15.9.2008], Internet
access: <http://epp.eurostat.ec.europa.eu/extraction/evalight/EVAlight.
jsp?A=1&language=en&root=/theme2/nama/nama_gdp_k>.
EUROSTAT. (2008b). Employment and unemployment – Labour Force Survey [on-line].
2008, [cit. 15.9.2008], Internet access: <http://epp.eurostat.ec.europa.eu/extraction/evalight/
EVAlight.jsp?A=1&language=en&root=/theme3/lfsa/lfsa_agan>.
EUROSTAT. (2008c). Population Projection – Convergence Scenario [on-line]. 2008, [cit.
15.9.2008], Internet access: <http://epp.eurostat.ec.europa.eu/extraction/evalight/ EVA-
light.jsp?A=1&language=en&root=/theme3/proj/proj_08c2150p>.
038 Acta všfs, 1/2010, vol. 4
EUROSTAT. (2008d). Overall structure of consumption expenditure by detailed COICOP
level [on-line]. July 2008, [cit. 13.1.2008], Internet access: <http://epp.eurostat.ec.europa.eu/
extraction/evalight/EVAlight.jsp?A=1&language=en&root=/theme3/hbs/hbs_str_t211>.
Fries, J.F. (1980). Ageing, Natural Death, and the Compression of Morbidity. The New
England Journal of Medicine, No. 303, 1980, pp. 130-135.
GÁL, R. I.; MEDGYESI, M.; RÉVÉSZ, T.; TÖRZSÖK, A. (2005). Generational Accounts in
Hungary 1992–2001. Ministry of Finance, Hungary, Working Paper No. 14, 45 p.
GOLLIN, D. (2001). Getting Income Shares Right. Williams College 2001, Draft, 27 p.
GRUNENBERG, E. M. (1977). The Failure of Success. Milbank Memorial Fund, Q. Health Society.
No. 55, 1977, pp. 3-24.
GUNDLACH, E. (2007). The Solow Model in the Empirics of Growth and Trade. Kiel Institute
for the World Economy 2007, Working Paper, 24 p.
CHAMBERLAIN, A.; PRANTE, G. (2007). Generational Equity: Which Age Groups Pay More
Tax, and Which Receive More Government Spending? Tax Foundation, June 2007, Special
Report No. 156, 16 p.
KALDOR, N. (1963). Capital Accumulation and Economic Growth. In Lutz, F. A., Hague, D.C.
eds.: Proceedings of a Conference Held by the International Economic Association. London,
Macmillan, 1963.
KIM, K.; VARSANO, R.; KEEN, M. (2006). The "Flat Tax(es)": Principles and Evidence. International
Monetary Fund 2006, Working Paper, No. 218, 50 p.
MANKIW, N. G.; ROMER, D.; WEIL, D. N. (1992). A Contribution to the Empirics of Economic
Growth. The Quarterly Journal of Economics, Vol. 107, May 1992, No. 2, pp. 407-437.
MARTINEZ-MONGAY, C. (2000). The Long-Run Determinants of Government Receipts. In.
Fiscal Sustainability, Perugia, Bank of Italy 2000, pp. 381-421.
Ministry of Finance of Estonia. (2008): Macroeconomic Forecast [on-line]. Summer 2008
[cit. 22.1.2009], Internet Access: <http://www.fin.ee/doc.php?80642>.
OECD. (2006). Projecting OECD Health and Long-term Care Expenditures: What are the
Main Drivers? OECD 2006, Economics Department, Working Paper No. 477, 81 p.
OECD. (2008a). Economic Outlook – Annual Projections for OECD Countries. OECD 2008,
December, No. 84.
OECD. (2008b). Revenue Statistics 1965-2007. Paris, OECD 2008, 361 p.
OECD. (2008c). OECD Health Data 2008: Statistics and Indicators for 30 Countries [on-line].
10.12.2008 [cit. 2.2.2009], Internet Access: <http://www.oecd.org/document/30/ 0,3343,e
n_2649_34631_12968734_1_1_1_37407,00.html>.
OKADA, T. (2006). What Does the Solow Model Tell Us about Economic Growth? Contributions
to Macroeconomics, Vol. 6, 2006, No. 1, pp. 1228-1258.
PAVLOKOVÁ, K. (2009). Time to Death and Health Expenditure of the Czech Health Care
System. Charles University, IES FSV, Working Paper No. 5/2009, 19 p.
TAXATION. Ch. (2009). Value Added Tax [on-line]. 2008, [cit. 22.1.2009], Internet Access:
<http://www.taxation.ch/index.cfm/fuseaction/show/temp/default/path/1-541.htm>.
VAN EWIJK, C.; TER RELE, H.; WESTERHOUT, E. (2002). Ageing in the Netherlands. CPB,
Netherlands Bureau for Economic Policy Analysis 2002, Publication No. 25, 136 p.
Contact address / Kontaktní adresa
PhDr. Kateřina Pavloková, Ph.D.
Ministry of Health of the Czech Republic
katerina.pavlokova@gmail.com
039Acta všfs, 1/2010, vol. 4
ANNEX1: Population pyramids – "trees and coffins" of life
040 Acta všfs, 1/2010, vol. 4
041Acta všfs, 1/2010, vol. 4
Source: Eurostat (2008c).
042 Acta všfs, 1/2010, vol. 4
ANNEX2:Publichealthcaresectorrevenuesandexpenditures(constanthealthscenariowithelasticity)bygenerations
(in%GDP)
043Acta všfs, 1/2010, vol. 4
044 Acta všfs, 1/2010, vol. 4
045Acta všfs, 1/2010, vol. 4
046 Acta všfs, 1/2010, vol. 4
ANNEX 3: Public finance deficits scenarios (in % GDP)
Source: Author’s calculations.