Eur J Ageing (2013) 10:353–361

DOI 10.1007/s10433-013-0280-x

REVIEW

The effect of population aging on health expenditure growth:

a critical review
Claudine de Meijer • Bram Wouterse •

Johan Polder • Marc Koopmanschap

Published online: 15 May 2013

Ó Springer-Verlag Berlin Heidelberg 2013

Abstract Although the consequences of population aging aging moderately increases expenditures on acute care and
for growth in health expenditures have been widely strongly increases expenditures on long-term care. The
investigated, research on this topic is rather fragmented. evidence further shows that the most important driver of
Therefore, these consequences are not fully understood. health expenditure growth, medical technology, interacts
This paper reviews the consequences of population aging strongly with age and health, i.e., population aging rein-
for health expenditure growth in Western countries by forces the influence of medical technology on health
combining insights from epidemiological and health eco- expenditure growth and vice versa. We therefore conclude
nomics research. Based on a conceptual model of health that population aging will remain in the centre of policy
care use, we first review evidence on the relationship debate. Further research should focus on the changes in
between age and health expenditures to provide insight into health that explain the effect of longevity gains on health
the direct effect of aging on health expenditure growth. expenditures, and on the interactions between aging and
Second, we discuss the interaction between aging and the other societal factors driving expenditure growth.
main societal drivers of health expenditures. Aging most
likely influences growth in health expenditures indirectly, Keywords Population aging
Morbidity
Technology

through its influence on these societal factors. The litera- Health expenditures
Acute care
Long-term care
ture shows that the direct effect of aging depends strongly
on underlying health and disability. Commonly used
approximations of health, like age or mortality, insuffi- Introduction
ciently capture complex dynamics in health. Population
The societal consequences of population aging, the
Responsible Editor: H. Litwin.
increasing share of older people in the population, are the
subject of extensive public and scholarly debate in many
C. de Meijer (&)
M. Koopmanschap different fields. In health economics, the focus has been
Institute of Health Policy and Management, Erasmus University largely on the effect of population aging on health expen-
Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The
diture growth (see Payne et al. (2007) for a review). The
Netherlands
e-mail: demeijer@bmg.eur.nl aging of Western populations stems mainly from decreas-
ing fertility rates, the aging of the post-World War II baby
C. de Meijer
M. Koopmanschap boom generation and longevity gains in general. In Wes-
Institute for Medical Technology Assessment, Rotterdam, The
tern countries, average health expenditures increased from
Netherlands
approximately 5 % of GDP in 1970 to nearly 10 % in 2009
B. Wouterse
J. Polder (OECD 2011). The assumed relationship between old age
National Institute for Public Health and the Environment, and health expenditure has raised serious concerns about
Bilthoven, The Netherlands
the financial sustainability of health care systems in Wes-
B. Wouterse
J. Polder tern countries. Despite extensive research on the effect of
Tranzo, Tilburg University, Tilburg, The Netherlands population aging on health expenditure growth, consensus

123
354 Eur J Ageing (2013) 10:353–361

on its role has not yet been achieved. Instead, two contra between the risk of death, the risk of ill health and health
posing views coexist (Reinhardt 2003). Some analysts and expenditures. Here we discuss the, often implicit, relation-
policy makers view population aging as the major cause of ship between longevity and health in these models. Third,
the rapid growth in health expenditure. Others, mainly we briefly discuss other main determinants of health
health economists as well as scholars from other fields, expenditures, and their possible interactions with aging, to
have argued that population aging is largely irrelevant for put the effect of aging into context.
the growth in health expenditure. An important reason as to
why these views are able to coexist is a lack of research
combining epidemiological insights on the relationship Conceptual framework
between health and aging with health economic research.
Given that changes in underlying health importantly Figure 1 provides an overview of the different factors that
determine the effect of longevity on health expenditures, influence individual health expenditures. The figure is based
insights combining health economics and epidemiological largely on the behavioral model of health service use that
evidence are crucial. was developed by Andersen and Newman (1973). Since
It seems obvious that the relationship between age and these theorists focused on health care use rather than
health expenditures depends on health. As individuals age, expenditure, we add relevant links to the model, as for
their health generally decreases and this in turn leads to instance those between medical technology, wages and
increasing utilization of health care. Less clear, however, is prices, and health expenditure. Although not all the ele-
how expected increases in longevity relate to health, and to ments of the model are relevant for our discussion of the
health expenditures. Epidemiological research on the rela- role of aging, the model serves to put aging and the related
tionship between longevity improvements and health can be individual and societal determinants into context. Two main
broadly characterized by three hypotheses proposed in the features of the model are especially relevant for a discussion
1980s: expansion, compression, and postponement of mor- of the role of population aging. First, the classification of
bidity (Kramer 1980; Fries 1980; Olshansky et al. 1991; individual factors clarifies the relationship between age,
Payne et al. 2007). The expansion hypothesis assumes that health, and health expenditures. Second, the model distin-
longevity gains will increase the period of time lived with guishes between societal and individual factors that jointly
morbidity or disability. The compression hypothesis determine the level of individual health expenditures.
assumes that this period will shrink. In the postponement, or The individual determinants are classified into three
dynamic equilibrium, hypothesis longevity gains are groups: predisposing, enabling, and need factors. Predis-
expected merely to shift the period with morbidity or dis- posing determinants reflect the individual’s ‘‘propensity
ability to an older age, while its duration remains constant. toward use.’’ They influence the likelihood that one will
Projections of the consequences of aging for health expen- use health care, without being directly responsible for its
diture growth have traditionally been based solely on age, utilization. Examples of predisposing factors include age,
and thus on the implicit assumption that gains in longevity do sex, marital status, co-residence status, socio-economic
not alter age-specific risks of disease or poor health. During status, and living and working conditions. Essential for our
the last two decades, however, health expenditure models discussion is the fact that age is classified as a predisposing
have been developed that do allow for changes in the age determinant. That is, age itself is not a reason for seeking
profile of health expenditures, most notably by controlling health care. Instead, people in different age groups have
for proximity to death (Zweifel et al. 1999). In general, this different types and amounts of illness and consequently
line of research has found a much smaller effect of aging on different patterns of health care use. The role of age as a
health expenditures than the traditional age-based studies. predisposing determinant is important, because it suggests
Yet, most of these models have not made the link between that the effect of increases in longevity, or the number of
health and longevity gains explicit, and only a few include individuals at older ages, depends on the relationship with
direct measures of health. the underlying need determinants.
In this paper, we aim to improve understanding of the These need determinants constitute the reason why an
consequences of population aging for health expenditure individual, given the presence of predisposing and enabling
growth in Western countries. The paper is organized as determinants, seeks the use of health care. The main need
follows. First, we start with a conceptual model of health determinants are poor health and disability. Health is
expenditure which relates the effect of age and aging to comprised of various dimensions, e.g., the presence of
other determinants on both the individual and the societal (chronic) diseases, self-reported health, mental- and phys-
level. Second, we discuss how health economic research has ical illness. Disability reflects the way in which poor health
developed from age-based models of health expenditure limits the ability to perform (instrumental) activities of
toward models that allow for changes in the association daily living and mobility. Although health and disability

123
Eur J Ageing (2013) 10:353–361 355

Fig. 1 Conceptual model of

Societal determinants
individual health expenditures
National Income

Health Policy Medical Technology Organization of the health Wages and Prices
care (insurance) system

Individual determinants
Predisposing factors Enabling factors Need factors

Health behavior, living and Informal care availability Health

working conditions perceived health, (chronic) illnesses

Demography Individual or household income Disability

Age, sex, co-residence status (i)ADL, mobility

Socio-economic status Health insurance Mortality(TTD)

Health care use and expenditures

Acute care Long-term care Informal care

are related, both determinants have a different relationship At the top of Fig. 1, the societal determinants of health
to health expenditures. This relationship also differs expenditures are shown. In a collectively financed health
between types of care, especially between acute and long- care system, societal determinants are largely responsible
term care (De Meijer et al. 2011). Mortality is also included for the level of collective health expenditures, whereas
as a need determinant, even though mortality itself cannot individual determinants are mostly responsible for the
lead to health care use, because a large part of aging distribution of collective health expenditures between
research in health economics has focused on the role of individuals (Getzen 2000). Societal determinants influence
proximity to death, as a proxy for health, to explain health health expenditures directly and indirectly, through their
expenditures (Payne et al. 2007). We relate to mortality as interaction with the individual determinants of health care
a need determinant rather than a predisposing determinant, use. The societal determinants that play the largest role in
because, unlike age, death is a consequence of poor health health expenditure growth are national income, medical
and not the other way round. technology, and wages and prices (Reinhardt 2003).
Enabling determinants concern the resources available Although there is some evidence of cost containment
to satisfy a need regarding health care use. They include effects of health system characteristics and certain policies
the level of health insurance coverage, individual, or (e.g., Wagstaff 2009; OECD 2009), their long run effect
household income, and informal care supply. Informal care seems to be limited given an increasing willingness to pay
can serve as a condition or substitute for formal care, for health care (Woodward and Wang 2012). The relevance
particularly in less skilled long-term care (Bonsang 2009). of the societal determinants for this paper lies in their
Informal care supply depends largely on the population age interaction with age and health. Therefore, societal deter-
composition, which on its turn depends on fertility rates minants potentially reinforce the effect of population
and the number of children, and household composition, as aging, and vice versa.
informal caregivers are often children or partners. Popu- The determinants shaded in Fig. 1 are those that are
lation aging is therefore likely to impact future informal discussed more extensively in the remainder of this paper.
care supply. However, the direction and strength of this We believe that this selection captures the most relevant
effect is less clear. On the one hand, future older people are determinants needed for an analysis of the role of aging in
likely to receive less informal care from their children. On health expenditure growth. In the case of individual
the other hand, decreases in the male–female life expec- determinants, these are age, health, disability, and mortal-
tancy gap (OECD 2011), resulting in fewer widowers, ity. In the case of societal determinants, these are national
might increase informal care from partners. Apart from the income, medical technology, and wages and prices as there
issue of informal care supply, other enabling determinants is (indicative) evidence that these particular societal
have less direct relevance for the role of aging, and we do determinants reinforce the effect of population aging, or
not discuss them further in this paper. vice versa.

123
356 Eur J Ageing (2013) 10:353–361

Individual determinants of health expenditure: age, Werblow et al. 2007). For long-term care, in comparison,
mortality, and health expenditures still increase with age, although controlling
for time to death diminishes its effect (e.g., Häkkinen et al.
The impact of age and mortality 2008; Roos et al. 1987; Weaver et al. 2009; Werblow et al.
2007).
Until the end of the 1980s, studies on the consequences of
population aging for health expenditure growth were Reconsidering the impact of age and mortality:
mostly based on the observed relationship between age and controlling for health and disability
average health expenditures per year. This relationship
shows a strong increase of health expenditures with age. Cost of dying and time to death studies recognize the role
When cross-sectional age profiles of expenditures are of age as a predisposing determinant in the sense that the
combined with demographic projections, huge increases in age pattern of health expenditures largely reflects increas-
health expenditures due to population aging are predicted ing mortality rates with age. However, the latter relation-
(Longman 1987; OECD 1988). As pointed out in the ship is itself the result of deteriorating health. If mortality is
conceptual model, age is a predisposing determinant. indeed a proxy of morbidity in explaining health expen-
Therefore, age-based projections are only valid when it is ditures, the concentrated health expenditures observed at
(implicitly) assumed that increases in life expectancy do the end of life are merely due to a higher burden of disease
not change the relationship between age and the onset of at the end of life. When mortality approximates health,
health problems. However, studies that explicitly account predictions of the effect of population aging, and specifi-
for the high health expenditures in the final years of life cally longevity gains, are only accurate when longevity
contradict this assumption. gains do not change the relationship between mortality and
The idea behind the latter studies is that the age profile health. But, as mentioned earlier, epidemiological research
of health expenditures can be explained by the fact that suggests that this relationship might not be constant (Kra-
relatively more people in older age groups are in their final mer 1980; Fries 1980; Olshansky et al. 1991; Payne et al.
years of life than those in younger age groups. Cost of 2007). In such cases, models are needed that incorporate
dying studies have confirmed that health expenditures are direct measures of health.
considerably higher for decedents than for survivors (Sci- While time to death studies abound, only a few of them
tovsky 1984; Madsen et al. 2002; Yang et al. 2003; Polder explore the relationship between health, disability, and
et al. 2006). However, unlike acute care expenditures, individual health expenditures. The available studies can be
long-term care expenditures still increase significantly with divided into three categories. The first category comprises
age after controlling for the expensive final years of life studies that differentiate the effect of mortality by under-
(McGrail et al. 2000; Polder et al. 2006; Spillman and lying cause-of-death. For example, decedents from cancer
Lubitz 2000; Yang et al. 2003). The fact that expenditures and respiratory diseases have significantly higher end-of-
are largely concentrated at the end of life regardless of the life spending than decedents from heart disease, indicating
age at death suggests that this will reduce health expendi- that the precise effect of mortality depends on the specific
ture growth, as there will be fewer people in the last year of health problem (Bird et al. 2002; Seshamani and Gray 2004;
life due to falling mortality rates. Predictions of future Wong et al. 2011a), and also on the coexistence of other
health expenditures should account for this. health problems (e.g., Wong et al. 2011b; Häkkinen et al.
A more refined way of controlling for the high expen- 2008). Second, there are studies that use both mortality and
ditures near the end of life is offered by time to death general health indicators to explain health expenditures.
studies. Instead of comparing aggregated costs of dece- Dormont et al. (2006) in France, and Shang and Goldman’s
dents and survivors by age, time to death studies use (2008) study of Medicare in the United States, show that
individual data to model health expenditures as a function mortality has little impact on health expenditures after
of the time away from death, allowing for in-depth anal- controlling for morbidity, confirming that mortality largely
yses of the effects of approaching death on health expen- approximates the effect of morbidity in health expenditure
ditures over time. Time to death studies consistently models. Third, a few studies have analyzed the relationship
conclude that time to death and not age is the main between health status and cumulative health expenditures
demographic determinant of health expenditures (e.g., over the remaining lifetime. These studies demonstrated
Roos et al. 1987; Häkkinen et al. 2008; Seshamani and that improvements in health lead to longer life expectancy
Gray 2004; Werblow et al. 2007; Zweifel et al. 1999). For but not to lower health expenditures, in most cases (Lubitz
acute care, age has no effect or a negative effect on 2005; Wouterse et al. 2011).
decedents’ expenditures, and only a weak positive effect on Some studies examined the effect of need determinants
survivors’ expenditures (e.g., Häkkinen et al. 2008; specifically for long-term care (De Meijer et al. 2009,

123
Eur J Ageing (2013) 10:353–361 357

2011; Weaver et al. 2009). As in the case of acute care, the Few projections of future health expenditure control
effect of mortality turns out to depend on cause-of-death directly for trends in disability. Manton et al. (2006) and
(De Meijer et al. 2011). However, there is also a significant Manton and Lamb (2007) investigated how the decline in
difference: the relationship between age and long-term care disability among older Americans between 1982 and 1999
expenditures remains, even after controlling for disability affected future Medicare costs. Although their estimates
and general health. In addition to disability and age, are now outdated, they demonstrated that projections
informal care availability has been shown to decrease extrapolating the recent disability decline accurately
formal long-term care expenditures. approached the actual amount spent. De Meijer et al.
(2011) found similar effects of decreasing disability trends
Predictions of health expenditures growth based for future homecare expenditures in the Dutch population.
on individual determinants (age, mortality, and health)

Different types of expenditure models also lead to different Societal determinants: national income, technology,
predictions of the effect of population aging on future health and wages
expenditures. Predictions from age-based models (models
that do not account for mortality or health) implicitly The previous section showed that population aging, as
assume that gains in longevity do not influence the age- such, accounts for an annual real growth of health expen-
specific risks of diseases or poor health, and thus adhere to ditures of 0.5–1.0 %. As total annual real growth in Wes-
an expansion of morbidity scenario. In contrast, time to tern countries is on average 4 % (OECD 2011), population
death models assume that the high health expenditures aging is, thus, not the only important determinant of health
during the final years of life shifts equivalently with lon- expenditure growth. Societal factors found to be at least
gevity gains. Thus, mortality-based projections adhere to a equally important are national income growth, technolog-
postponement of morbidity hypothesis, as the period spent ical development, and rises in wages and prices (Burner
in poor health is assumed to be merely postponed to a later et al. 1992; Reinhardt 2003; Richardson and Robertson
age. Population aging has, therefore, a significantly lower 1999). However, as explained in the conceptual model,
impact on future health expenditures when accounting for there might be strong interaction effects between popula-
mortality instead of age only. tion aging and these factors. As a result of these interac-
The total annual real growth of health expenditure is, on tions, the age profile of health expenditures might change.
average, 4 % in Western countries (OECD 2011). Studies Given this, we first briefly discuss the main societal
on annual growth in health expenditure have found a rate determinants, and then turn to their interactions with age
falling roughly between 0.5 and 1.7 % due to population and health showing evidence for changes in the age profile
aging without accounting for decreasing mortality rates and of health expenditures.
a 0.1 and 0.5 % point lower growth rate when accounting
for reductions in mortality rates (e.g., Steinmann et al. 2007; Societal determinants
Miller 2001; Shang and Goldman 2008; Polder et al. 2006).
The extent of the variations in prediction between age-based In general, income reflects the ability and willingness to
and mortality models depends on the service under con- pay for health (care). On a national level, income growth is
sideration. The difference between age-based and mortality found to be the strongest explanatory factor for health
models is particularly large for hospital expenditures and expenditure growth (Newhouse 1977; OECD 2006). The
much less for primary care and pharmaceutical expendi- common explanation for the strong relationship between
tures (e.g., Seshamani and Gray 2004; Serup-Hansen et al. national income and health expenditure is that health is a
2002; Häkkinen et al. 2008; Kildemoes et al. 2006). luxury good: when income grows, individuals want to
Although mortality models adhere to the postponement spend a larger share of it on health care (Wildavsky 1977;
of morbidity hypothesis, recent evidence seems to support Hall and Jones 2007). Although the evidence for this rea-
the compression hypothesis, although not unambiguously soning is quite strong on a national level, it should be noted
(Martin et al. 2010; Christensen et al. 2009; Lafortune et al. that income is found to have a very small impact on an
2007; Parker and Thorslund 2007). Despite divergent individual level (Van Doorslaer et al. 2004; Getzen 2000).
results, the tendency seems to be that the time spent with This difference in results can be explained by the presence
(severe) disability remains constant or is shrinking. Popu- of collective health insurance that makes the price of health
lation aging has, therefore, a significantly lower impact on care use close to zero at the individual level (Getzen 2000).
future health expenditures in models that take disability Although it is sometimes stated that development of new
into account as compared to those based on only age or medical technology is a result of increasing willingness to
mortality. pay for health, medical progress itself is often mentioned as

123
358 Eur J Ageing (2013) 10:353–361

the most important driver of health expenditure growth, health expenditures or medical innovations are likely to be
and particularly of acute expenditures (e.g., Newhouse directed toward groups with relatively poor health.
1992; Weisbrod 1991). In fact, technological progress has Although there are examples of medical innovations aimed
two contrasting effects on health expenditure. It can miti- at age groups with relatively small health problems, older
gate expensive care and reduce costs, but it also tends to people are likely to disproportionally benefit from medical
increase use (Cutler and McClellan 2001; Lubitz 2005; innovations. Goldman et al. (2005) simulated the effects of
Cutler 2007). In general, the latter of these two effects ten key technologies on health expenditures in the U.S.
prevails, resulting in a rise in total health expenditure They found that most of these innovations resulted in a
(Bodenheimer 2005). Although intuitively strong, evidence shift of health expenditures to older age groups. Wong
on the impact of technological progress on health expen- et al. (2012) found similar results for the Netherlands. That
ditures is relatively scarce. Studies that do explicitly con- medical innovations are targeted more at older people also
sider technological growth, for example those that look at seems to be suggested by the fact that during the last few
research and development spending, generally find a strong decades, life expectancy at age 65 rose faster than at other
positive effect on health expenditure growth (Dormont ages (Christensen et al. 2009).
et al. 2006; Goldman et al. 2005; Jones 2002; Okunade and Third, changes in the population composition caused by
Murthy 2002; Suen 2005; Westerhout 2006). population aging will influence some of the societal factors.
The relative price of health care is also an important For instance, Murphy and Topel (2006, p. 884) have shown
driver of health expenditure growth. Health care tends to be that ‘‘aggregate willingness to pay for progress against a
relatively labor intensive and part of this labor cannot be particular disease will be highest when the age distribution
easily substituted by technology, especially in the long- of the population is near, but before, the typical onset of the
term care sector. As a result, labor productivity in health disease.’’ Therefore, population aging might increase the
care tends to develop more slowly compared to other demand for collectively financed health care aimed at older
industries. Since health care workers earn an income age groups. This effect, where population aging results in a
comparable to that in other sectors, the relative price of larger demand for health care aimed at the older popula-
health care increases. Although findings vary (Murillo et al. tion, resulting in further population aging, has been named
1993; Murthy and Ukpolo 1994; Okunade et al. 2004), the the Sisyphus syndrome. Empirical evidence on this phe-
effect of this so-called Baumol’s disease (Baumol 1967) on nomenon, however, is limited (Zweifel et al. 2005).
health expenditure growth tends to be as large as that of Finally, increases in the share of older people can also be
population aging (Hartwig 2008). expected to result in a further rise in wages for the health
care sector (Dixon 2003; Simoens et al. 2005).
Interactions Hence, the literature tends to suggest that the influence
of societal determinants is not age-neutral, and additional
The conceptual model suggests interactions between the health expenditure is likely targeted at older age groups. As
individual determinants age and health, and the societal a result, the age profile of health expenditure might steepen
determinants. Although empirical evidence is limited, we over time. Indeed, there is emprical evidence for this
discuss indicative findings on three possible interactions. steepening effect (e.g., Meara et al. 2004; Dormont et al.
First, one would expect that health expenditures will have a 2006; Breyer et al. 2010).
positive effect on life expectancy and/or health. In fact, one
of the most important criticisms of the time to death lit-
erature has been that it ignores this effect (e.g., Salas and Discussion and conclusion
Raftery 2001). However, for long periods of time, the role
of health care in decreasing mortality has been found to be Aging in perspective
small compared to other factors such as nutrition and
public health measures (McKeown 1976; Szreter 1988). This paper has reviewed evidence on the effect of popu-
For recent decades, a much stronger effect of health care on lation aging on health expenditure growth. We have
lowering mortality has been found, especially for the older focused on the relationship between improvements in
population (Bunker et al. 1994; Mackenbach 1996; Cutler longevity and trends in health, and its consequences for
et al. 2006; Mackenbach et al. 2011). Empirical evidence health expenditures. Furthermore, we have put the effect of
of the effect of increased health spending on health is population aging in perspective by looking at other major
restricted to disease-specific studies (e.g., Cutler et al. societal determinants of health expenditures and their
2008). interactions with aging. The paper started with two
Second, because the distribution of health expenditures opposing views: population aging is either the main factor
between individuals is largely determined by health, additional behind the rapid health expenditure growth, or largely

123
Eur J Ageing (2013) 10:353–361 359

irrelevant for this growth. What has our review of the lit- selective in what kinds of health care they want to fund
erature added to these views? by public resources.
First, evidence from the literature has shown the value The consensus from the literature on the relatively large
of explicitly considering health-related causes of the rela- role of age and disability for long-term care expenditures
tionship between age and health expenditures. As discussed suggest that prioritizing medical innovations aimed at
in the conceptual model, age is a predisposing determinant improvement of quality of life and functioning over those
that is not directly responsible for utilization of health care. aimed at postponement of death might be a promising
Instead, it is the relationship between age and need deter- approach to contain costs. Medical technology that allows
minants such as health and disability that explain the age people to remain living at home, even with chronic dis-
pattern of health expenditures. Predictions that account for eases, might lead to substantial savings in long-term care.
health, either directly or through the variable of mortality In addition, improved functioning could increase labor
as a proxy, tend to find that population aging only mod- participation at older ages, and increase the (health care)
erately contributes to health expenditure growth. The workforce.
results therefore seem to support the second view. How-
ever, the effect of population aging is much stronger for Conclusion
long-term care compared to acute care. In fact, the annual
health expenditure growth attributed to population aging is Health expenditures will continue to rise in the coming
found to be up to 1 %, which is far from trivial. decades. Although the direct effect of population aging is
Second, in discussing the effect of societal determinants modest, age and aging remain important factors in the
on health expenditure growth, the literature has shown that debate on health expenditure growth. Many drivers of this
the total annual health expenditure growth in Western growth interact with population aging, particularly health,
countries averages 4 %. National income growth, medical national income growth, technological progress, wages and
technology, and price and wage rises are the main drivers prices. Future health expenditure is likely to be targeted
of health expenditure growth on the aggregated level more toward the older population. If increases in health
(Burner et al. 1992; Reinhardt 2003; Richardson and expenditure reflect an increasing willingness to pay for
Robertson 1999). However, there is evidence for interac- health and solidarity, its growth may not necessarily be a
tions between population aging and these societal deter- problem. However, the larger extents to which health
minants. Whereas the level of collective health expenditure expenditure will be used by older people, in combination
is to a large extent driven by societal determinants, its with a financing system that distributes costs over the entire
distribution is largely determined by health and disability. population, can strain inter-generational solidarity. There-
Medical innovations and additional growth in health fore, population aging will remain in the centre of policy
expenditure are therefore more often targeted at older debate. Further research should focus on the changes in
people, reinforcing the effect of population aging. In a health that explain the effect of longevity gains on health
number or studies, a steepening of the age profile of health expenditures, and on the interactions between aging and
expenditures has indeed been found. other societal factors driving expenditure growth.
While the direct effect of population aging is less
dramatic than expected by some, policy makers should Acknowledgments This study was part of the projects ‘‘Living
longer in good health,’’ which was financially supported by the
keep in mind the interaction between population aging, Network for Studies on Pensions, Aging and Retirement (NETSPAR),
health, income growth, technological developments, and and ‘‘Healthy aging and health care expenditure,’’ financed by the
wages/prices. Although health expenditure growth might National Institute for Public Health and the Environment (RIVM). We
reflect a collective willingness to pay for health, we have thank two anonymous referees and the editor (Howard Litwin) for
their extensive comments on an earlier version of the paper.
discussed evidence that increased health expenditure is to
a relatively large extent beneficial to the older age groups,
which are growing in relative size. In a collectively
financed health care system, in which a substantial part of
References
health care premiums are paid by the working force
population, this might strain inter-generational solidarity. Andersen R, Newman JF (1973) Societal and individual determinants
Baicker and Skinner (2011) have found that OECD of medical care utilization in the United States. Milbank Mem
countries with higher tax burdens in 1979 experienced Fund Q Health Soc 51:95–124
slower health expenditure growth in subsequent decades. Baicker K, Skinner JS (2011) Health care spending growth and the
future of U.S. tax rates. NBER working paper no. 16772. Bureau
This observation seems to suggest that there are limits to of Economic Research, Cambridge
what individuals are willing to pay for collectively Baumol WJ (1967) Macroeconomics of unbalanced growth: the
financed health care. Policy makers should therefore be anatomy of urban crisis. Am Econ Rev 57:415–426

123
360 Eur J Ageing (2013) 10:353–361

Bird CE, Shugarman LR, Lynn J (2002) Age and gender differences Longman P (1987) Born to pay: the new politics of aging in America.
in health care utilization and spending for Medicare beneficiaries Houghton Mifflin, Boston
in their last years of life. J Palliat Med 5(5):705–712 Lubitz J (2005) Health, technology, and medical care spending.
Bodenheimer T (2005) High and rising health care costs. Part 2: Health Aff 24(Suppl 2):W5R81-5
technologic innovation. Ann Int Med 142(11):932 Mackenbach JP (1996) The contribution of medical care to mortality
Bonsang E (2009) Does informal care from children to their elderly decline: McKeown revisited. J Clin Epidemiol 49:1207–1213
parents substitute for formal care in Europe? J Health Econ Mackenbach JP, Slobbe L, Looman CWNL, Van der Heide A, Polder
28(1):143–154 JJ, Garssen J (2011) Sharp upturn of life expectancy in the
Breyer F, Costa-Font J, Felder S (2010) Aging, health, and health Netherlands: effect of more health care for the elderly? Eur J
care. Oxf Rev Econ Policy 26(4):674–690 Epidemiol 26:903–914
Bunker JP, Frazier HS, Mosteller F (1994) Improving health: Madsen J, Serup-Hansen N, Kragstrup J, Kristiansen IS (2002)
measuring effects of medical care. Milbank Q 72:225–258 Ageing may have limited impact on future costs of primary care
Burner ST, Waldo DR, McKusick DR (1992) National health providers. Scand J Prim Health Care 20(3):169–173
expenditures projections through 2030. Health Care Fin Rev Manton KG, Lamb VL (2007) Medicare cost effects of recent US
14(1):1–29 disability trends in the elderly: future implications. J Aging
Christensen K, Doblhammer G, Rau R, Vaupel JW (2009) Aging Health 19(3):359–381
populations: the challenges ahead. Lancet 374:1196–11208 Manton KG, Gu X, Lamb VL (2006) Change in chronic disability
Cutler DM (2007) The lifetime costs and benefits of medical from 1982 to 2004/2005 as measured by long-term changes in
technology. J Health Econ 26(6):1081–1100 function and health in the U.S. elderly population. Proc Natl
Cutler DM, McClellan M (2001) Is technological change in medicine Acad Sci USA 103(48):18374–18379
worth it? Health Aff 20(5):11 Martin LG, Schoeni RF, Andreski PM (2010) Trends in health of
Cutler DM, Rosen AB, Vijan S (2006) The value of medical spending older adults in the United States: past, present, future. Demog-
in the United States, 1960–2000. N Engl J Med 355:920–927 raphy 47(Supplement):S17–S40
Cutler DM, Landrum MB, Stewart KA (2008). Intensive medical care McGrail K, Green B, Barer ML, Evans RG, Hertzman C, Normand C
and cardiovascular disease disability reductions. In: Cutler DM, (2000) Age, costs of acute and long-term care and proximity to
Wise DA (eds) Health at older ages. The University of Chicago death: evidence for 1987–88 and 1994–95 in British Columbia.
Press, Chicago Age Ageing 29(3):249–253
De Meijer CA, Koopmanschap MA, Koolman XH, van Doorslaer EK McKeown TF (1976) The role of medicine: dream, mirage or
(2009) The role of disability in explaining long-term care nemesis. Nuffield Provincial Hospitals, London
utilization. Med Care 47(11):1156–1163 Meara E, White C, Cutler DM (2004) Trends in medical spending by
De Meijer C, Koopmanschap M, d’Uva TB, Van Doorslaer E (2011) age, 1963–2000. Health Aff 23(4):176–183
Determinants of long-term care spending: age, time to death or Miller T (2001) Increasing longevity and Medicare expenditures.
disability? J Health Econ 30(3):425–438 Demography 38(2):215–226
Dixon S (2003) Implications of population ageing for the labour Murillo C, Piatecki C, Saez M (1993) Health care expenditure and
market. Labour Mark Trends 111(2):67–76 income in Europe. Health Econ 2(2):127–138
Dormont B, Grignon M, Huber H (2006) Health expenditure growth: Murphy KM, Topel RH (2006) The value of health and longevity.
reassessing the threat of ageing. Health Econ 15(9):947–963 J Polit Econ 114:871–904
Fries JF (1980) Aging, natural death, and the compression of Murthy NRV, Ukpolo V (1994) Aggregate health care expenditure in
morbidity. N Engl J Med 303(3):130–135 the United States: evidence from cointegration tests. App Econ
Getzen TE (2000) Health care is an individual necessity and a national 26(8):797–802
luxury: applying multilevel decision models to the analysis of Newhouse JP (1977) Medical-care expenditure: a cross-national
health care expenditures. J Health Econ 19(2):259–270 survey. J Hum Resour 12(1):115–125
Goldman DP, Shang B, Bhattacharya J, Garber AM, Hurd M, Joyce GF Newhouse JP (1992) Medical care costs: how much welfare loss?
et al (2005) Consequences of health trends and medical innova- J Econ Perspect 6(3):3–21
tion for the future elderly. Health Aff 24(Suppl 2):W5R5-17 OECD (1988) Aging population: the social policy implications.
Häkkinen U, Martikainen P, Noro A, Nihtilä E, Peltola M (2008) OECD Publishing, Paris
Aging, health expenditure, proximity to death, and income in OECD (2006) Projecting OECD health and long-term care expendi-
Finland. Health Econ Policy Law 3(2):165–195 tures: what are the main drivers? Economics department working
Hall RE, Jones CI (2007) The value of life and the rise in health papers no. 477. OECD Publishing, Paris
spending. Q J Econ 122(1):39–72 OECD (2009) Achieving better value for money. OECD Health
Hartwig J (2008) What drives health care expenditure? Baumol’s model Policy Studies, Paris
of unbalanced growth revisited. J Health Econ 27(3):603–623 OECD (2011) Health at a glance 2011: OECD indicators. OECD
Jones CI (2002) Why have health expenditures as a share of GDP Publishing, Paris
risen so much? NBER working paper no. 9325. Bureau of Okunade AA, Murthy VNR (2002) Technology as a major driver of
Economic Research, Cambridge health care costs: a cointegration analysis of the Newhouse
Kildemoes HW, Christiansen T, Gyrd-Hansen D, Kristiansen IS, conjecture. J Health Econ 21(1):147–159
Andersen M (2006) The impact of population ageing on future Okunade AA, Karakus MC, Okeke C (2004) Determinants of health
Danish drug expenditure. Health Policy 75(3):298–311 expenditure growth of the OECD countries: Jackknife resam-
Kramer M (1980) The rising pandemic of mental disorders and pling plan estimates. Health Care Manag Sci 7(3):173–183
associated chronic diseases and disabilities. Acta Psychiatr Olshansky SJ, Rudberg MA, Carnes BA, Cassel CK, Brody JA (1991)
Scand Suppl 62:282–297 Trading off longer life for worsening health. J Aging Health
Lafortune G, Balestat G, Disability Study Expert Group Members 3(2):194–216
(2007) Trends in severe disability among elderly people: Parker MG, Thorslund M (2007) Health trends in the elderly population:
assessing the evidence in 12 OECD countries and the future getting better and getting worse. Gerontol 47(2):150–158
implications. OECD health working papers no. 26. OECD Payne G, Laporte A, Deber R, Coyte PC (2007) Counting backward
Publishing, Paris to health care’s future: using time-to-death modeling to identify

123
Eur J Ageing (2013) 10:353–361 361

changes in end-of-life morbidity and the impact of aging on Wagstaff A (2009) Social health insurance vs tax-financed health
health care expenditures. Milbank Q 85(2):213–257 systems: evidence from the OECD. Policy research working
Polder JJ, Barendregt JJ, Van Oers H (2006) Health care costs in the paper 4821. World Bank, Washington, DC
last year of life—the Dutch experience. Soc Sci Med Weaver F, Stearns SC, Norton EC, Spector W (2009) Proximity to
63(7):1720–1731 death and participation in the long-term care market. Health
Reinhardt UE (2003) Does the aging of the population really drive the Econ 18(8):867–883
demand for health care? Health Aff 22(6):27–39 Weisbrod BA (1991) The health care quadrilemma: an essay on
Richardson J, Robertson I (1999) Aging and the cost of health technological change, insurance, quality of care, and cost
services. Working paper no. 90. Centre for Health Program containment. J Econ Lit 29(2):523–552
Evaluation, Sydney Werblow A, Felder S, Zweifel P (2007) Population ageing and health
Roos NP, Montgomery P, Roos LL (1987) Health care utilization in care expenditure: a school of ‘red herrings’? Health Econ 16(10):
the years prior to death. Milbank Q 65(2):231–254 1109–1126
Salas C, Raftery JP (2001) Econometric issues in testing the age Westerhout WMT (2006) Does ageing call for a reform of the health
neutrality of health care expenditures. Health Econ 10(7):669– care sector? CESifo Econ Stud 52(1):1–31
671 Wildavsky A (1977) Doing better and feeling worse: the political
Scitovsky AA (1984) The high cost of dying: what do the data show? pathology of health policy. Daedalus 106(1):105–124
Milbank Mem Fund Q Health Soc 62(4):591–608 Wong A, Van Baal PHM, Boshuizen HC, Polder JJ (2011a) Exploring
Serup-Hansen N, Wickstrinm J, Kristiansen IS (2002) Future health the influence of proximity to death on disease specific hospital
care costs—do health care costs during the last year of life expenditures: a carpaccio of red herrings. Health Econ 20(4):
matter? Health Policy 62(2):161–172 379–400
Seshamani M, Gray A (2004) Ageing and health-care expenditure: the Wong A, Boshuizen HC, Schellevis FG, Kommer GJ, Polder JJ
red herring argument revisited. Health Econ 13(4):303–314 (2011b) Longitudinal administrative data can be used to examine
Shang B, Goldman D (2008) Does age or life expectancy better multimorbidity, provided false discoveries are controlled for.
predict health care expenditures? Health Econ 17(4):487–501 J Clin Epidemiol 64(10):1109–1117
Simoens S, Villeneuve M, Hurst J (2005) Tackling nurse shortages in Wong A, Wouterse B, Slobbe LCJ, Boshuizen HC (2012) Medical
OECD countries. OECD health working papers no. 19. OECD innovation and age-specific trends in health care utilization:
Publishing, Paris findings and implications. Soc Sci Med 74(2):263–272
Spillman BC, Lubitz J (2000) The effect of longevity on spending for Woodward RS, Wang L (2012) The oh-so straight and narrow path:
acute and long-term care. N Engl J Med 342(19):1409–1415 can the health expenditure curve be bent? Health Econ
Steinmann L, Telser H, Zweifel PS (2007) Aging and future 21(8):1023–1029
healthcare expenditure: a consistent approach. Forum Health Wouterse B, Meijboom BR, Polder JJ (2011) The relationship
Econ Policy 10(2):1–30 between baseline health and longitudinal costs of hospital use.
Suen RMH (2005) Technological advance and the growth in health Health Econ 20(8):9-685-1008
care spending. Economie d’Avant Garde research reports. Yang Z, Norton EC, Stearns SC (2003) Longevity and health care
University of California, Riverside expenditures: the real reasons older people spend more. J Ger-
Szreter S (1988) The importance of social interventions in Britain’s ontol Ser B 58(1):2–10
mortality decline 1850–1914: a re-interpretation of the role of Zweifel P, Felder S, Meiers M (1999) Ageing of population and
public health. Soc Hist Med 1:5–37 health care expenditure: a red herring? Health Econ 8(6):485–
Van Doorslaer E, Koolman X, Jones AM (2004) Explaining income 496
related inequalities in doctor utilisation in Europe. Health Econ Zweifel P, Steinmann L, Eugster P (2005) The Sisyphus syndrome in
13(7):629–647 health revisited. Int J Health Care Financ Econ 5:127–145

123