Download September 16, 2010, 10:30 am

THE RISE OF THE INTERGENERATIONAL STATE:
AGEING AND DEVELOPMENT
Tim Miller and Jorge Bravo
Word count: 9,296
Text: 6,546
Tables: 5, word equivalent = 1,250
Figures: 6, word equivalent = 1,500

The authors would like to thank the excellent assistance of Elisenda Rentería in compiling and evaluating
an extensive amount of the information from various sources, presented succinctly in the introduction.
1
1.0 Introduction: Public transfers in an international perspective
The State currently transfers a sizeable share of resources, within and across generational
groups. The degree of involvement of the public sector in resource reallocation varies
widely across countries and regions, and has evolved substantially over time. Among
NTA countries, public transfers represent 21% to 80% of all transfers and x% to z% of
GDP. Although there have been short and medium-term movements in different
directions, two broad long-term trends are apparent: the State has generally expanded its
weight in the economy, and has become an increasingly important player in reallocating
resources across generations.
1.1 The weight of the State in the economy over time
The expansion of the public sector over the 20th century is tightly linked to its
changing role. Its primordial function of provider of essential public works, the
maintenance of law and order and national defense, shifted and expanded to a larger
presence and activity in the development of infrastructure, intervention in preventing or
moderating large cyclical fluctuations in the macro-economy, and in the regulation of the
financial and productive sectors (Tanzi 1997; Tanzi and Schuknecht 2000). Also, and
very significantly, the public sector has greatly expanded its role in using taxes and other
revenues to finance expanded spending in social programs such as education, health and
pensions, which are clearly age-specific.
There are different ways to measure the weight and evolution of the government in
the economy, and in the transfer of resources across generations. Looking at aggregate
figures, we first examine general government expenditures (GGE), a broad measure of
2
public spending that includes all levels of government, which is available for long periods
of time for the more developed countries.
Table 1 shows that in industrialized countries, GGE represented a little over 10 per
cent of their GDP at the end of the 19th century, it more than doubled to about 22.8% of
GDP before World War II, and almost doubled again to 43.1% of GDP by 1980. The
growth of government spending in OECD countries has decelerated since, and in some
countries has even declined during the last few decades (OECD 2009).
[Table 1 here].
Within GGE, government real expenditure (i.e., spending in wages, salaries and
goods) increased steadily and vigorously during the 20th century, but the main driver of
the expansion of the public sector was the growth of cash subsidies and transfers, from
one or two percentage points of GDP in the early part of that century, to almost 1/4 of
GDP towards the end of it. As can be inferred from Table 1, subsidies and transfers in the
industrialized countries surpassed real expenditure sometime around the 1980s, and by
the end of the 20th century, they represented more than 1/2 of general government
expenditure.
These subsidies and transfers reflect to some extent increases in producer subsidies
and, more importantly, the expansion of welfare spending and the public provision of
education and health, social security, including pensions, all major programs that
effectively reallocate resources across generational groups. The growth of public
spending has been evidenced most markedly in the more developed countries, but is also
3
apparent in many developing countries in different regions of the world over the last
several decades.
Government spending can change significantly even over the course of a few years
as a result of short-term national budgetary priorities and policies, or international
macroeconomic swings. For example, in recent years, government spending as a percent
of GDP has increased in many countries because of the global financial and economic
crisis that unfolded in 2007 and that slowed the world economy during 2008 to 2010. The
crisis led to a contraction of GDP or slower growth, while government spending has
either increased after stimulus plans have been implemented in some countries, or
remained stable or has fallen less than GDP in others. Furthermore, data in Table 2 shows
that, before this crisis, OECD countries were taking different paths during the 1980s and
1990s, that reflect mostly purposeful short and medium-term economic policies:
countries like Belgium, Ireland, New Zealand, The Netherlands, Sweden and the United
Kingdom sustained significant reductions in government expenditure as a percent of
GDP, while others like France, Greece, Iceland, and Portugal progressively increased it,
in some cases to high and unsustainable levels as evidenced by recent debt crises.
[Table 2 here].
As for the future of public spending, there are widely differing views on the
expected levels and trends, from a limited role of the State in total spending and in
reallocating resources across generations, to an increased or at least stable public
spending in proportion to the economy. It is certain, however, that population aging will
bring significantly increased pressure for expansion of the public sector spending over the
next few decades, as discussed in Section 3.
4
1.2 What do governments spend on?
Table 1 above showed that, in industrialized countries, spending in all major social
programs, such as education, health and pensions have expanded substantially over past
decades. Some specialists (Sanz and Velazquez, 2001; Castles, 2007), have noted that
government spending is showing some signs of international convergence, in particular as
regards the composition of expenditure. Of direct interest to this paper, total social
expenditure in OECD countries has risen from 18.5% of GDP in 1980 to 22.4% of GDP
in 2003, and have come to represent nearly one half (48%) of general government
expenditures (Castles 2007, Tables 2 and 4).
Public social spending began with poor relief in the late 19th century, continued
with public schooling which expanded to represent more than 1% of GDP by the early
20th century. Health and pensions programs started to absorb an increasing fraction of
public spending, and by mid-century out-paced education. In the second half of the 20th
century, health and pension expenditures continued to rise as these programs greatly
expanded their coverage and population ageing became more significant (Lindert 2004).
1.3 Government spending in richer and poorer nations
Total central government expenditures (CGE) do not cover all the levels of
government, and it thus provides more limited information than GGE, especially for
federal states. But data on CGE are available for many more, richer and poorer nations in
different regions of the world. CGEs include public consumption and investment, as well
as cash transfers and subsidies to individuals provided by the central government.
5
The data shows some, but rather weak correlation of CGEs with the level of income
(United Nations 2001); for example, in 1997, average central government spending
varied from nearly 38% of GDP in developed countries, to 32% in Transition Economies,
27% in Africa, and to 20%-23% of GDP in Asia and Latin America. Within OECD
countries (see Table 2) CGEs are highest in Austria, Belgium, Denmark, France and
Sweden, all countries with per capita GDP at or above the OECD average, but most the
countries with the lowest levels of government spending (Australia, Ireland, Japan,
Korea, Switzerland and Unites States) have as high or even higher per capita income,
with the exception of Korea, that has low CGE (less than 30% of GDP) and lower than
OECD average per capita GDP. Within Latin American countries, there is not either a
clear correlation between government spending and per capita income.
1.4 Regional differences in public consumption
Another measure of the weight of the government sector in the economy that
affects resource reallocations is public consumption (also called government
consumption expenditure), a key component of the NTA framework taken from the
System of National Accounts (SNA). Data on public consumption are available for more
countries than either GGEs or CGEs, and since these figures are produced annually using
the common conceptual and accounting framework of National Income and Product
Accounts (NIPA), they are also better suited for comparison across a wider spectrum of
countries and regions.
Public consumption includes the value of goods and services provided by the
government collectively or to groups of individuals. It encompasses all in-kind transfers
6
from the government to individuals, like public consumption in health, education or
defense, but not cash transfers (United Nations 2008; International Monetary Fund 2009).
The figures of public consumption for 2005 (IMF 2009), at current prices, show
that the majority of countries that had public consumption of less than 10 per cent of
GDP are from Asia, Latin America and Africa, although Luxembourg and Ukraine are
also in this category. The countries where public consumption represents between 10 and
15 per cent of GDP are a mixed group. There are many that are Asian countries, such as
China, Korea or Thailand, while some others are from Latin America, like Argentina,
Chile or Mexico, and Africa, including Benin, Senegal or Mali. One European country,
Ireland, is also in this group. The majority of European countries and developed countries
have a level of public consumption that is in between 15 and 20 per cent of their GDP,
but so do some countries from Asia and Africa. Among the countries that spend more
than 20 per cent of their GDP in public consumption, there seem to be two different
kinds: those that have a big Welfare State, like Sweden, France, Denmark and Belgium,
and those who devote a large share of public consumption to defense, like Israel and Iraq.
Average public consumption by region has not experienced sharp changes during
1970 to 2005 (Table 3). The region that has changed the most is Europe, from 15 per cent
of GDP in 1970 to 20 percent in 1995-2000. In all the other regions, average government
consumption for the most part increased until the 1980s and fell afterwards, in varying
degrees, most pronouncedly in Northern America.
[Table 3 here].
7
The overall levels and trends of public consumption by country and region, are
roughly in line with the data on GGE and CGE previously shown. Asia has the lowest
percentages of government consumption, and Latin America comes next, followed
closely by Africa. Australia and New Zealand started in 1970 with a very low share of
public consumption, but they increased it considerably in the 1980s, to levels similar to
those of Europe and Northern America.
1.5 “Social” public consumption and transfers
This section briefly reviews government social expenditures in health, education and
social protection, using data from the OECD (2009), ECLAC (2006) and The System of
National Accounts (SNA), United Nations (2009). Despite the somewhat different
expenditure concepts, some broad differences between countries can be noted.
The data from OECD and ECLAC show that: a) the majority of countries with a young
age structure, many of them less developed, devote a large part of their government
expenditure to education; b) In these younger, less developed countries, spending in
education is larger than in health, with the exceptions of Argentina and Uruguay, two
countries of intermediate level of development and with relatively aged populations; c)
Among the more developed countries of the OECD, almost the exact opposite situation is
observed for the majority of the countries, with the clear exception of Korea (which
spends more in education than in health), and the noteworthy cases of Finland and
Sweden, that have high levels of spending in public health but of similar magnitude than
those in education.
8
We next focus on public consumption only for the 19 NTA countries for which we found
reliable SNA data . The SNA data corroborates the observation already made on the basis
of data from OECD and ECLAC, that those countries with a younger population and
lower per capita income spend more on public education than on public health. Other
countries, like Austria, Argentina, Costa Rica, Finland and Sweden, are spending about
the same amount on public spending on health and education, although their overall level
of public consumption is quite different. For the all the other developed countries, public
expenditure in health is notably higher than in education.
A particularly interesting case is that of the Republic of Korea (Figure 1), which is fairly
representative of the broad patterns of change in government expenditure in connection
with changing population age structures. Until 1980, the Korean government devoted a
very small proportion of its budget to social protection or health, while education got the
lion’s share. As time went by, and the Korean population continued to age, health and
social protection consumption increased progressively, to the point that, by 2007, health
consumption was close to catching up with education consumption.
[Figure 1 here].
The trend of increasing health expenditures, in absolute and relative terms, has been
pervasive over the last few decades; even in developed countries like France, Finland or
Germany, which already have high levels of public spending in health and aged
populations, health spending has continued to increase in recent years. This continued
expansion is partly due to population ageing. And health care projections for the US, for
example, show that about one-half of the forecast increase of the share of health
9
expenditures in GDP stem from population aging (Lee & Miller, 2002). In most OECD
countries ageing plays a significant role, although the recent increases are mostly due to
increasing costs of health care (Hagist and Kotlikoff, 2005, OECD, 2006). But some
international differences are more clearly aligned along regional, demographic and
development groupings: in the three aforementioned countries, health represents about 30
per cent of public consumption (while education is near 20 per cent); conversely, in
countries like Mexico, Thailand, Kenya or India, with lower per capita income and
younger populations, health expenditures are less than 10 per cent of total public
consumption.
In sum, in most of the documented national cases, public transfers have become larger
over time, in proportion to GDP and of general government spending, although in some
developed countries total government spending declined towards the end of the XXth
century, and has fluctuated during the most recent global economic crisis. Public transfers
in industrialized countries, which represented just around 1% of GDP at the beginning of
the 20th century, have expanded to become nearly one fourth of GDP and more than half
of the total government budget by the end of the 1990s. Public transfers in the social
sectors tend to be heavily concentrated in programs that finance education, health
services, and public pensions, all of which have defined age profiles. Thus population
ageing has accompanied and affected the evolution of the level and composition of public
spending, and policy decisions regarding these sectors have direct and important
consequences on the intergenerational distribution of resources.
We found that there is a rather weak association between level of development (indexed
by per capita GDP) and total or central government spending. Despite substantial
10
international variation, regional differences are slightly better defined, with countries
with low spending in public consumption found mostly in Asia, Latin America and
Africa, and with most European countries having relatively high levels of public
consumption. The highest public spenders though, include two different kinds: some with
extended Welfare States (e.g., Sweden, France, Denmark) and some who devote a large
share of their government budgets to defense (e.g., Israel, Iraq). Among NTA countries,
there appears to be a fairly systematic relationship between the degree of ageing, level of
income and the level of social public consumption and transfers. Wit a few exceptions,
younger, less developed countries devote a large share of their public budgets to
education, which are almost always substantially larger than spending on health. In the
great majority of older, richer nations, expenditures in social security, in particular in
public health, are notably higher than those in education. We turn now to examine the age
dimension of public transfers as revealed by the NTA database.
2. The measurement of public transfers in the NTA framework
National Transfer Accounts define three mechanisms by which economic resources
are reallocated over the life cycle: public transfers, private transfers, and asset-based
reallocations. Governments use two of these reallocations mechanisms: a) public
transfers by which they tax and transfer resources across age groups, and b) borrowing,
lending, and other asset-based operations by which they transfer resources over time and
between generations. The focus of this chapter is public transfers; asset-based operations
of the government are discussed in Chapter 9.
There are four important conceptual features of public transfers. First, public
transfers include in-kind as well as cash transfers. In most of the economic literature,
11
public transfers are defined more narrowly to include only cash transfers – typically
social welfare payments and social insurance payments. Public transfers in the NTAs are
much more comprehensive, as they include items such as public education, public health,
and other public spending on goods and services such as defense, transportation, general
government administration and operating costs.
Second, in NTA, the receipt of public transfers is always assigned to specific
individuals. For example, those students attending public schools receive public
education benefits. Some public benefits are difficult to assign to specific individuals –
either because we may lack good data sources on the users of particularly public services
(for example, public transportation subsidies) or in the case where such spending is
arguably a public good (such as operating costs of government or defense spending). In
these cases, spending is assigned equally to all individuals on a per-capita basis.
Third, as in private transfers, the total amount of public transfers received must
equal the total amount of transfers given. For every recipient, there is a giver. However,
in public sector transfers, these transactions are much more impersonal, as the transfers
flow from the ensemble of taxpayers to all beneficiaries. Public transfer outflows
include all taxes and contributions from current tax payers as well as a public transfer
deficit as discussed in the NTA methodology presented in Chapter 3. This public
transfer deficit is assigned to current taxpayers as part of public transfer outflows and the
asset-based transactions which fund this deficit are accounted for as public asset-based
reallocations as discussed in Chapter 9. Ultimately, these public transfer deficits are
financed either by past generations of taxpayers (via accumulations reflected in public
asset income) or are passed on to future taxpayers (in the form of public debt).
12
Fourth, contributions for social insurance are considered to be functionally
equivalent to taxes and are included as part of public sector outflows. In particular, in
pay-as-you-go pension systems, the contributions from current workers are used to fund
retirement payments to current retirees, and thus NTA treats them exactly the same way
as taxes used to pay for public education. Indeed, a great strength of the NTA system is
that it views social insurance contributions and payouts as part of the overall public
transfer system, taking a neutral view of the competing demands of health care,
education, and pensions. In sorting out competing demands, governments may well
decide, for example, that claims of elderly for retirement benefits should weigh more
heavily than claims for education or health care, but it is important that our accounting
system be policy-neutral.
3. International comparisons.
3.1 Typical age pattern of public transfers
The typical age profiles of per capita public transfer inflows and outflows are
presented in Figure 2. These are based on an unweighted average across 20 NTA
countries. The solid line represents per capita public transfer outflow by age. The bulk of
public transfer outflows are from working-age adults – with the peak in the early 50s.
Public transfer outflows decline sharply thereafter – mainly a reflection of withdrawal
from the labor market and the cessation of paying social insurance contributions for
public retirement programs. Public transfer outflows increase sharply around the 20s as
individuals enter the labor market and begin paying social insurance contributions from
their labor earnings. The fact that children have public transfer outflows is a reflection of
13
sales taxes. Children, like persons of any age in the population, are assumed to be paying
taxes based on their consumption of taxed goods.
In summary, public transfer outflows are the least for children, low for the elderly,
and highest for working-age adults. This is a reflection of the degree to which specific
economic activities are taxed. Reliance on taxes on labor earnings leads to higher public
transfer outflows from working-age adults, while taxes on property tend to increase
public transfer outflows from elderly, and taxes on consumption tend to be spread more
equally across age groups. The mix of revenue sources differs across countries and so the
public transfer outflows by different age groups will differ from country to country. But
the general pattern is that public transfer outflows are drawn primarily from the
population in the main working-ages.
[Figure 2 here].
The dashed line represents per capita public transfer inflows at each age. This
presents an almost mirror image of public transfer outflows by age. Working-age adults
are the age group that pays the most taxes and are also the age group that receives the
least amount of government benefits. Children receive public transfers mainly in the form
of public education, which accounts for about 45% of public transfers to children in NTA
countries. Accordingly, a high plateau in public transfer inflows is seen in Figure 2
extending from age 6 to age 18 – corresponding to the ages of primary and secondary
schooling. Public benefits begin to decline at age 18, a reflection of the decline in per
capita public education benefit that is due to the fact that not everyone attends college
and, in some countries, a significant share of students attends private schools. Thereafter,
a long plateau in per capita benefits in the early working-years is evident. This is a
14
reflection of the per-capita share of general government spending (on the administration
of public services, defense, transportation, etc) which as noted earlier are distributed
equally to everyone on a per-capita basis. In the later working-years, public transfer
inflows begin to rise – a reflection of growing receipt of public benefits from health and
disability payments. Very sharp increases are seen in the late 50s and early 60s reflecting
the increasing receipt of retirement benefits – as well as the increasing use of public
health services in later life. These age patterns differ across countries depending on the
role of the public sector in provision of education, health care, and pensions.
Nevertheless, the typical pattern across these countries is for low levels of benefits to the
working-age population and high levels of benefits being directed toward children and
the elderly.
Net public transfers are presented in Figure 3 and are calculated as the difference
between public transfer inflows and public transfers outflows across the 20 NTA
countries. There are two periods of net receipt of transfers. In early life up until age 22,
individuals receive more in benefits than they pay in taxes. Thereafter follows a period of
38 years in which individuals pay more in taxes than they collect in benefits. In later life,
starting at age 60, individuals begin receiving more in benefits than they pay in taxes.
Based on this average pattern across the countries, the most costly age group for
governments are those 90 and older – mainly because they receive a large amount of
pension and health benefits but also, as was noted earlier, because they pay little taxes.
[Figure 3 here].
The likely fiscal impact of population aging is readily apparent in this graph. In the
initial stages of population aging during the period of the “demographic dividend” the
15
working age population is growing faster than the youth population. These demographic
changes lead to an easing of fiscal burden on governments. In federal systems such as
those of Brazil and the US, this would be more notable in state and local governments
that have traditionally provided the main source of funding for public education. This is
not to say that the easing of fiscal burdens resulted in an actual reduction in taxes. Indeed,
in many countries, funding in public education was maintained or increased, which has
allowed expanded enrollment and increased investment per student. As populations
continue to age, the period of the demographic dividend draws to a close as the elderly
population begins to grow more quickly than the working-age population. Governments
then face decades of continually increasing fiscal pressures from population aging.
As a first approximation to assess such pressures, one can use the age profiles of
public transfers to calculate a “fiscal support ratio,” defined as the ratio of aggregate
taxes to aggregate benefits. In the absence of government asset-based reallocations such
as borrowing and debt repayment, this ratio would be 1 with aggregate taxes equal to
aggregate benefit payments. As the population ages, the fiscal support ratio would decline
as the number of taxpayers declines relative to the number of public transfer
beneficiaries. The changes in the fiscal support ratio indicate the relative size of the tax
increase or benefit cut needed to return to the initial fiscal position.
Calculations of the change in the fiscal support ratio over a 100-year period (19502050) for a select group of NTA countries are shown in Table 4. The estimates are based
on the age profiles of public transfer inflows and outflows observed in each country for a
recent year with estimates and projections of the population by age from 1950 through
2050. The fiscal support ratio is set to 1.0 in the base year of 2010, corresponding to a
16
situation in which aggregate taxes equal aggregate benefits. Countries are ranked in the
table according to the size of the fiscal adjustment needed by 2050 as a result of changes
in the age structure of the population brought about by population aging. Of all NTA
countries, the fiscal impact of population aging is projected to be most severe in Brazil.
Population aging combined with the current tax and benefit policies would lead to a 31%
decline in the fiscal support ratio by 2050. Either transfer benefits would need to be cut
by 31% before 2050 or taxes would need to increase by 45% or some combination of the
two.
Brazil is by no means alone in facing these mounting fiscal pressures. In Europe,
declines in fiscal support range from 28% in Slovenia to 14% in Sweden. Declines in
fiscal support ratio among other Latin American countries range from 31% in Brazil and
28% in Chile to 10% in Uruguay. This underscores the point that population aging is a
worldwide phenomena not restricted to Europe. Therefore it should not be surprising that
some of the most severe fiscal impacts of aging are projected to occur outside of Europe.
The fiscal support ratio in the US is projected to decline by 11% by 2050 – slightly less
severe than Sweden and slightly more severe than Uruguay. In Asia, the top three
countries with the most severe projected fiscal impacts are Japan (a decline of 26%) and
China and South Korea (with declines of 20%). Equally significant for several of these
countries is the fact that, not only is the fiscal support ratio deteriorating rapidly in the
future, but this represents a distinct break from the past decades in which fiscal support
ratios were stable or improving. For example, Brazil, Chile, Slovenia, Spain, China,
South Korea, and the United States all reached the point of minimal fiscal pressure from
demographic change during the last 10 years. A final group of countries shows increases
in the fiscal support ratio – a reflection of the fact that under current tax and spending
17
programs in these countries, the elderly are net taxpayers or only have very moderate net
fiscal costs. These are Thailand, Indonesia, and the Philippines. Underlying these
calculations is the assumption that the age shapes of benefits and taxes remain fixed over
time, with their absolute levels increasing at the same rate with economic growth. In this
way, the needs for fiscal adjustment in benefits or taxes can be derived directly from the
changes in the fiscal support ratio.
More realistically, we expect the age shape of public benefits to shift over time –
for example, with increased investment in schooling for youth and increased investment
in health care for the elderly. In addition, reforms to public pensions programs towards
funded programs – such as those undertaken in Chile, Germany, and other Latin
American and European countries – will lead to a significant shift in the burden of
population aging away from the public transfers towards private transfers and asset-based
reallocations. NTA data can be readily combined with these more realistic assumptions
about changing public benefits to produce medium and long run fiscal forecasts for
governments (see for example, Miller, Mason, and Holz 2010 for fiscal forecasts for ten
Latin American countries; Miller and Castanheira 2010 for fiscal forecast for Brazil).
In addition to projections, NTA public transfer data can be used to develop
historical accounts. When these are combined with projected data, it is possible to
estimate net public transfers received and taxes paid by generations. These permit the
analysis of how different generations fared over the course of economic development as
public transfer systems were expanded (for an analysis of US generations born between
1850 through 2090 see Bommier et al. 2010).
[Table 4 here.]
18
3.2 Net public transfers to older persons and to youth
Having seen the typical pattern of net transfers based on the simple average across
20 countries, we now examine the impact of each governments tax and spending policies
on per capita benefits received by youth compared to those receive by elderly. As is
evident in Figure 4, there is a great diversity in these policies. Per capita net transfers
received by youth (ages 0 to 19) range from 6% of per capita labor earnings for adults
aged 30-49 in China to 29% in Finland. Similarly, per capita net transfers received by
elderly (ages 65 and older) range from -2% of per capita labor earnings for adults aged
30-49 in Thailand – where elderly pay more in taxes than they receive in benefits— to
87% in Brazil, due its generous government pension program.
[Figure 4 here].
The median values of net public transfers to youth (16% of per capita labor
earnings of adults ages 30-49) and to elderly (38%) divide Figure 4 into 4 quadrants. The
European countries with the exception of Spain all lie in the upper right quadrant of the
graph – with high net transfers received by children (public education) and by elderly
(public pensions and public health). Among Asian countries, Japan shares the
characteristics of European countries. But most Asian countries are found in the lower
left quadrant of the figure with low levels of transfers to both children and elderly. One
Asian country (Taiwan) lies in the upper left quadrant – with high levels of transfers to
children and low levels to elderly. The United States is similarly characterized by high
levels of per capita transfers to children with public transfers to elderly slightly below the
median value among NTA countries. All Latin American countries have levels of public
investment in children below the median among NTA countries. Three Latin American
19
countries (Chile, Costa Rica, and Brazil) stand out among the NTA countries for having
low levels of public transfers to children coupled with high levels of public transfers to
elderly. Brazil is a clear outlier among NTA countries with per capita public transfers to
elderly amounting to 86% of per capita labor earnings of adults ages 30 to 49 – more than
9 times the per capita public transfer to youth.
The general pattern that emerges in Figure 4 is the strong positive correlation
between public transfers to youth and those to elderly. The simple correlation between
the two transfers is +0.46, rising to +0.67 when excluding the outlier Brazil. This crosssectional evidence is consistent with a view of government action as the outcome of
cooperation between generations. It is generally consistent with ideas of Becker and
Murphy (1988) in which generations cooperate via the public sector to overcome
inefficiently low-levels of income security in old-age and educational investment in
youth. That is, generations cooperate to vote for higher taxation to provide for both public
education and public pensions. An alternative view was put forward by Preston (1984) in
describing public transfers in the United States. Under this view, generations compete for
scarce public sector resources and as population aging increases the voting power of the
elderly, public transfers shift toward elderly at the expense of children. Arguing against
this hypothesis, one notes that those societies in the lower left quadrant of Figure 4 – with
high per capita transfers to elderly and low to youth – are Latin American countries with
relative low percentage of elderly. In Costa Rica and Chile, net public transfers to the
average elderly person are about 4 times greater than those to youth and in Brazil, they
are 9 times as greater. The tax and spending pattern in these countries which favor elderly
20
may have more to do with high levels of income inequality in these societies and the
political power of the wealthy, as suggested by Turra and Queiroz (2005).
We can also rank governments by the aggregate amount of net public transfers to
elderly versus youth as shown in Table 5. Germany ranks first with aggregate net public
transfers to elderly 2.56 times as great as those to youth. This is a product of both the
relative generosity of per capita net transfers to elderly in Germany (2.9 times the per
capita public transfer to youth) as well as the large elderly population (88% the size of the
youth population). Japan, with the largest elderly population among NTA countries
(102% of its youth population), also ranks highly with net public transfers to elderly 1.7
times those to youth. Governments in all European NTA countries spend more on elderly
than on youth. Brazil is an interesting case in that it ranks with European countries with
aggregate net public spending on elderly exceeding that of youth – despite the small size
of its elderly population (13% of its youth population). This is due to Brazil’s generous
public pension programs and its relatively low levels of public educational investment
that results in per capita net public transfers to elderly more than 9 times greater than
received by youth. Chile and Costa Rica are also notable in this regard: the per capita net
public transfer to elderly is nearly 4 times that to youth – leading to large aggregate
public transfers devoted to elderly despite the relative small size of the elderly population
in these societies (20% and 15% of the youth populations respectively).
[Table 5 here].
3.3 Net public transfers and consumption
Per capita consumption by elderly and by children relative to working-age adults
vary greatly among countries. Here we examine the relationship between per capita
21
consumption (private and public consumption combined) and public transfers. Figure 5
shows a strong positive correlation (r=+0.69) between per capita consumption by elderly
(relative to working-age adults) and per capita net public transfers to elderly. Countries in
which elderly tend to consume much more than working-age adults also tend to be
countries in which the elderly receive relatively high levels of net public transfers. Two
notable countries in this regard are Sweden and Brazil. In Sweden, the average elderly
consumes 39% more than the average working-age adult and receives annually net public
transfers equivalent to 77% of GDP per working-age adult. Much of this transfer is in the
form of public health. In Brazil, the average elderly consumes 30% more than the average
working-age adult and receives annually net public transfers equivalent to 87% of per
capita labor earnings of adults ages 30 to 49.
[Figure 5 here].
Among youth, we also see a strong positive correlation (r=+0.47) in Figure 6
between per capita consumption relative to working-age adults and per capita net public
transfer received. Those countries with high per capita net public transfers to youth also
tend to have high levels of per capita consumption by youth. The European countries are
clustered in the upper right quadrant – with the average European youth consuming just
15% less than the average working-adult while receiving on average net public transfers
equivalent to 22% of per capita labor earnings of adults ages 30-49. Latin American
countries lie at the opposite extreme – with the average Latin American youth consuming
30% less than the per capita working-adult while receiving net public transfers equivalent
to 10% of per capita labor earnings of adults ages 30-49. Clearly, many factors contribute
to these low levels of consumption among youth in Latin America relative to other NTA
22
countries – but low levels of public investment in education are certainly an important
cause.
[Figure 6 here].
4.0 Concluding remarks
We began this chapter reviewing the historical evidence for the expansion of the
government’s role in the economy and in particular its growing importance in
reallocating resources across generations. Examining NTA data on public transfers we
reported on the typical role of the state in transferring resources from the working-ages
toward youth and the elderly. Using population projections, we forecast the likely fiscal
impact of population aging given each country’s unique tax and benefit structure. Half of
those countries facing severe fiscal impacts are in Latin America – underscoring the
widespread nature of population aging throughout the world. We observed significant
differences among governments in terms of their treatment of youth and elderly – as well
as significant differences in well being as measured by relative consumption levels. At
higher levels of economic development, relative consumption by youth and by elderly are
higher and so too, are public transfers to these groups with the state playing an
increasingly important role in the lives of youth and the elderly.
The NTAs are an important tool for informing public policy choices. Governments
play a large and increasingly important role in providing economic support to youth and
the elderly. The NTAs provide a comprehensive and coherent measure of this expanding
role. They enable governments to monitor the full scope of its policy actions – accounting
for the impact of all government spending programs and all taxation. Additionally, they
allow governments to perceive the roles played by other economic actors (financial
23
markets, families, civil society) in providing support for youth and the elderly.
Furthermore, because the NTA project has assembled a diverse group of countries,
governments can draw on international comparisons of their policies and their impact on
the well being of youth and the elderly.
NTAs also provide the basis for long-run fiscal forecasts for governments. These
long-run forecasts readily reveal the major economic transformations brought about by
slow but inexorable social forces such as population aging, the epidemiological
transition, or shifts in the educational distribution. The adoption of long-run forecasts by
governments is important in two key respects. First, some of the policy choices with the
largest payoffs (such as investment in public education) have very long delays between
investment and payoff. A short-run policy focus will bias decisions by failing to measure
the full extent of the return on long-run investments such as investing in youth. Second, a
long run policy focus promotes marginal changes in policy (i.e., “course corrections”)
which are politically more feasible to implement and provide smooth transitions in
spending and tax policies that are less likely to unfairly burden any particular generation
– avoiding the need for draconian policy responses to avert fiscal crises.
24
5.0 References
Becker, Gary and Kevin Murphy. 1988. “The Family and the State”. Journal of Law and
Economics 21.
Bommier, Antoine, Ronald Lee, Tim Miller, and Stéphane Zuber. 2010. “Who wins and
who loses? Public Transfer Accounts for US generations born 1850 to 2090.”
Population and Development Review 36(1): 1-26.
Castles, Francis G. 2007. “Patterns of State in Europe and America”. Paper presented at
the European Conference on The attractiveness of the European and American
Social model for new members and candidate countries of the European Union.
Berlin, 7-8 May, Social Science Centre.
Christian Hagist and Laurence J. Kotlikoff. 2005. Who’s going broke? Comparing
healthcare costs in ten OECD countries. National Bureau of Economic Research
Working Paper 11833, http://www.nber.org/papers/w11833.
Economic Commission for Latin America and the Caribbean (ECLAC) 2006. Panorama
Social, LC/G.2326-P/I, Santiago, Chile, December 2006.
International Monetary Fund. 2009. Government Finance Statistics. In:
http://www.imfstatistics.org/imf/ (accessed on 09/15/09).
Lee, Ronald D. and Tim Miller. 2002. “An Approach to Forecasting Health
Expenditures, with Application to the U.S. Medicare System”. Health Services
Research 37(5): 1365–1386.
Lindert, Peter H. 2004. Growing Public: Social spending and economic growth since the
eighteenth century. Volume I: The story. Cambridge University Press.
25
Miller, Tim, Carl Mason, and Mauricio Holz. 2009. The fiscal impact of demographic
change in ten Latin American Countries: Projecting public expenditures in
health, education, and pensions. Paper presented at Demographic Change and
Social Policy – A LAC Regional Study, Authors’ Workshop, World Bank,
Washington, July 14-15, 2009.
Miller, Tim and Helena Cruz Castanheira. 2010. The fiscal impact of population aging in
Brazil. Paper presented at Brazil Country Study on Aging Authors’ Workshop,
Brasilia and Washington, April 6-7, 2010.
National Transfers Accounts Project (several authors). http:// www.ntaccounts.org
(accessed on 08/1/10).
OECD. 2006. Projecting OECD health and long-term care expenditures: What are the
main drivers? Economics Department Working Papers no. 477. Available at
http://www.oecd.org/dataoecd/57/7/36085940.pdf.
OECD (2009) Stat Extracts, accessed on 1 July 2009.
http://stats.oecd.org/Index.aspx?DataSetCode=SNA_TABLE11
Preston, Samuel H. 1984. “Children and the Elderly: divergent Paths for America’s
Dependents”. Demography 21(4): 435-457.
Sanz, Ismael and Francisco J. Velásquez. 2001. The evolution of the convergence of the
government expenditure composition in the OECD countries: an analysis of the
functional distribution. Policy Modelling for the European and Global Issues,
Brussels, 5-7 July.
Tanzi 2007
26
Tanzi, Vito and Ludger Schuknecht 2000 Public Spending in the 20th Century: A global
perspective. Cambridge University Press.
Turra, Cassio M. and Bernardo L. Queiroz. 2005. Intergenerational Transfers and
Socioeconomic Inequality in Brazil: a First Look. Paper presented at the Taller
sobre Transformaciones Demograficas, Transferencias Intergeneracionales y
Proteccion Social en America Latina, CELADE, July 6-7, Santiago, Chile.
United Nations. 2001. World Public Sector Report – Globalization and the State.
Department of Economic and Social Affairs. United Nations Publication.
United Nations. 2009. System of National Accounts. Statistical Division of the
Department of Economic and Social Affairs.
http://data.un.org/Browse.aspx?d=SNA (accessed on 09/15/09).
27
Table 1. Evolution of Government expenditure in Industrialized Countries, 1870-1995 (as
percent of GDP)
c. 1870
General government
expenditure
Government real
expenditure
Subsidies and
Transfers
1960
1980
1990
c. 1995
22.8
27.9
43.1
44.8
45.6
4.6
11.4
12.6
17.9
17.4
17.3
1.1
4.5
9.7
21.4
23.2
3.7
3.4
2.5
2.1
3.5
5.8
6.1
0.4
2.4
5.8
6.4
1.9
4.5
8.4
0.3
0.9
10.7
1913
12.7
1920
18.7
1937
Public spending on
selected
sectors/programs*
Defense
4.0
Education
0.6
Health
0.3
Pensions
0.4
Unemployment
1.3
2.4
1.3
1.2
2.0
8.9
9.6
1.6
* Note: spending in these programs can be in-kind or provided in cash; e.g., defense, health and education
are mainly in-kind, pensions and unemployment consist mostly of cash benefits.
Source: Tanzi and Schuknecht, 2007, last row of tables I.1 to II.X.
28
Table 2. Evolution of General government expenditure as % of GDP, OECD
countries,1980-2005
Year
1980
1985
1990
1995
2000
2005
63.5
53.6
45.7
49.7
55.9
40.8
40.1
64.2
55.8
51.8
53.2
58.4
49.8
46.3
60.7
55.9
49.5
51.5
52.3
52.9
47.9
65.1
59.2
54.4
56.1
52.1
52.5
61.6
54.8
55.2
29.2
57.3
41.1
54.9
44.9
34.3
39.4
40.5
56.4
45.7
54.5
43.4
55.5
54.2
51.6
50.8
49.2
48.0
47.7
47.6
47.3
45.4
45.3
44.5
44.4
55.2
52.7
53.6
49.7
52.1
48.2
50.5
46.8
50.0
45.2
43.1
44.9
47.6
46.1
43.6
53.3
35.7
41.8
38.0
48.1
41.5
49.7
48.0
50.9
47.7
42.7
48.0
44.3
44.1
43.8
42.7
41.8
38.4
42.1
43.3
42.4
46.5
34.7
33.9
46.3
33.9
42.7
33.9
33.6
39.5
36.6
54.2
21.2
53.8
20.0
42.2
32.1
42.8
37.7
53.5
36.4
36.9
30.3
42.8
20.0
44.1
36.5
44.4
39.7
41.4
37.4
37.0
35.0
41.1
20.8
41.1
38.6
38.6
38.1
37.7
35.6
35.0
34.8
33.3
25.0
38.5
41.9
39.9
34.7
36.5
35.3
33.8
28.9
42.5
46.0
44.8
47.1
43.1
43.8
Sweden
Denmark
France
Austria
Belgium
Italy
Finland
Germany
Hungary
Netherlands
Greece
Czech Republic
Portugal
Slovak
Republic
Norway
Poland
Iceland
Canada
United
Kingdom
Japan
Spain
Luxembourg
New Zealand
Australia
United States
Switzerland
Ireland
Korea
Average
(unweighted)
19801990
-2.8
2.3
3.8
1.8
-3.6
12.1
7.8
Change during:
199019802000
2005
-5.2
-8.3
-1.7
-1.0
2.1
7.9
-0.7
0.0
-3.1
-3.8
-4.9
7.4
-0.2
10.4
-0.3
15.6
-9.6
0.5
-10.0
13.9
6.3
3.9
13.3
7.2
-9.2
-4.0
5.8
7.9
1.2
-7.9
6.7
-4.3
-2.6
9.0
-11.4
-1.2
-1.1
6.5
-4.2
0.4
-15.8
-0.8
-1.8
4.5
-9.5
5.0
-20.4
7.7
2.3
-1.7
1.3
2.5
3.2
4.6
0.8
2.9
Source: OECD (2009)
29
Table 3. Public consumption by region, 1970-2005
A. Public consumption as percentage of GDP.
Asia
Europe
UE
Africa
Latin America
Northern America
Australia/New Zealand
Oceania wo A/NZ
1970
12.7
14.8
15.6
17.2
15.9
19.5
14.9
28.1
1975
13.8
16.9
18.1
17.7
16.3
19.9
18.2
27.3
1980
13.2
17.6
18.5
18.1
16.7
19.0
19.3
35.2
1985
16.1
18.2
19.0
18.8
18.4
19.7
19.1
37.9
1990
15.9
19.2
18.5
17.2
16.5
19.6
18.8
37.4
1995
14.4
20.8
20.2
17.6
15.6
18.3
17.9
34.2
2000
14.9
19.7
19.4
17.0
16.3
16.5
17.8
33.3
2005
15.0
19.3
19.9
16.6
16.3
17.5
18.0
35.1
Source: United Nations (2009)
B: Public Consumption as a percentage of total consumption.
Asia
Europe
UE
Africa
Latin America
Northern America
Australia/New Zealand
Oceania wo A/NZ
1970
18.9
20.1
20.7
20.4
18.8
24.7
19.9
28.8
1975
20.8
22.3
23.2
20.8
19.6
25.2
23.4
30.0
1980
20.9
23.0
23.8
21.0
20.2
24.7
24.9
32.2
1985
22.4
23.6
24.5
21.3
22.4
24.8
24.8
33.3
1990
21.4
24.4
24.6
19.3
20.9
24.3
24.0
33.2
1995
19.3
25.3
25.9
19.9
20.3
22.8
23.4
31.2
2000
21.2
24.5
25.3
20.2
20.6
21.2
23.1
30.1
2005
21.4
24.7
26.0
20.1
20.0
22.1
23.7
30.0
Source: United Nations (2009)
30
Table 4. Fiscal support ratios: 1950-2050 (Countries ordered by severity of projected
fiscal impact in 2050).
Country
Brazil
Chile
Slovenia
Spain
Austria
Japan
Germany
Costa
Rica
Hungary
China
South
Korea
Finland
Mexico
Sweden
US
Uruguay
Thailand
Indonesia
Philippines
Fiscal support ratio
Year of most
favorable age
strucutre
Year
Support
Ratio
1950
2010
2020
2030
2050
1.00
0.94
1.01
0.94
1.08
0.91
1.11
0.89
1
1
1
1
1
1
1
1
0.94
0.93
0.91
0.96
0.93
0.92
0.94
0.97
0.86
0.83
0.81
0.87
0.83
0.87
0.84
0.91
0.69
0.72
0.72
0.73
0.74
0.74
0.75
0.76
2000
2004
2002
2010
1950
1976
1950
2012
1.02
1.01
1.04
1.00
1.08
1.15
1.11
1.00
1.06
0.93
0.76
1
1
1
0.97
0.94
0.97
0.93
0.87
0.89
0.77
0.80
0.80
1950
2007
2008
1.06
1.00
1.00
1.08
0.85
1.15
0.99
1.08
0.66
0.79
0.87
1
1
1
1
1
1
1
1
0.92
1.02
0.96
0.96
1.00
1.04
1.06
1.06
0.87
0.99
0.90
0.92
0.98
1.04
1.10
1.11
0.83
0.86
0.86
0.89
0.90
1.04
1.08
1.16
1991
2019
1950
2006
1959
2039
2033
2050
1.11
1.02
1.15
1.00
1.09
1.04
1.10
1.16
Source: Authors calculations using population estimates and projections from the United Nations and age
profiles of public transfers from NTA.
31
Table 5. Age orientation of public sector transfers
Country
Germany
Sweden
Austria
Japan
Hungary
Spain
Slovenia
Uruguay
Finland
Brazil
Chile
China
US
Costa Rica
South Korea
Mexico
Taiwan
Indonesia
Philippines
Thailand
Aggregate net public
transfers to elderly
relative to youth
Elderly population
relative to youth
population
Per capita net public
transfers to elderly
relative to youth
2.56
2.38
2.32
1.67
1.65
1.62
1.58
1.39
1.34
1.21
0.79
0.73
0.63
0.55
0.26
0.23
0.22
0.01
-0.01
-0.03
0.88
0.72
0.67
1.02
0.73
0.79
0.73
0.42
0.67
0.13
0.20
0.23
0.44
0.15
0.25
0.12
0.27
0.15
0.07
0.21
2.91
3.33
3.48
1.65
2.26
2.05
2.16
3.29
1.99
9.53
3.93
3.11
1.42
3.67
1.07
1.90
0.84
0.08
-0.09
-0.16
Source: Authors calculations using population estimates from the United Nations and age profiles of public
transfers from NTA.
32
Figure 1: Health, education and social protection expenditure as a percentage of
public consumption in Korea, 1970-2005
25
20
15
10
5
0
1970
1975
1980
Health
1985
1990
Education
1995
2000
2005
Social protection
Source: SNA UN
33
Figure 2. Per capita public transfer inflows and outflows by age
Figure 2. Per capita public transfer inflow and outflows by age
0.60
Relative to per capita labor earnings at ages 30-49
0.50
0.40
Per capita public
transfer inflows
Per capita public
transfer outflows
0.30
0.20
0.10
0.00
0
10
20
30
40
50
60
70
80
90
Age of individual
Source: Authors calculations based on NTA data.
34
Figure 3. Per capita net public transfers by age
Figure 3. Per capita net public transfers by age
0.50
Relative to per capita labor earnings at ages 30-49
0.40
0.30
0.20
0.10
0.00
0
10
20
30
40
50
60
70
80
90
-0.10
-0.20
-0.30
Age of individual
Source: Authors calculations based on NTA data.
35
Figure 4. Per capita net public transfers to elderly and to youth
Figure 4. Per capita net public transfer per elderly and per youth
0.30
Per youth (Relative to per capita labor earnings at ages 30-49)
FI
HU
JP
SE
SI
US
0.20
TW
DE
ES
AT
KR
TH
CL
MX
CR
0.10
BR
ID
PH
UY
CN
0.00
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
1.20
Per elderly (Relative to per capita labor earnings at ages 30-49)
Source: Authors calculations based on NTA data.
36
Figure 5. Net public transfers to elderly and per capita consumption.
Figure 5. Per capita net public transfer and per capita consumption by elderly
Per caputa consumption by elderly
(Relative to per capita consumption by adults ages 20 to 64)
1.50
1.40
SE
US
JP
1.30
BR
1.20
SI
1.10
PH
1.00
CR
UY
ES
DE
HU
FI
AT
CL
CN
TH
TW
KR
0.90
MX
ID
0.80
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
1.20
Per capita net public transfer per elderly
(Relative to per capita labor earnings at ages 30-49)
Source: Authors calculations based on NTA data.
37
Figure 6. Per capita net transfers to youth and consumption.
Figure 6. Per capita net public transfer and per capita consumption by youth
Per capita consumption by youth
(Relative to per capita consumption by adults ages 20 to 64)
1.10
SI
1.05
1.00
SE
0.95
TW
0.90
JP
KR
0.85
ES
CN
AT
0.80
FI
UY
HU
TH
0.75
US
ID
0.70
DE
CL
PH BR
MX
0.65
CR
0.60
0.05
0.10
0.15
0.20
0.25
0.30
Per capita net public transfer by youth
(Relative to per capita labor earnings at ages 30-49)
Source: Authors calculations based on NTA data.
38