. 1
( 7)


This page intentionally left blank
Age in the Welfare State

The overwhelming costs of providing for aging populations have brought
many welfare states to the brink of insolvency. Now is the time to ask: how
did we get here? Age in the Welfare State explains how it came to pass that
some nations give the lion™s share of social bene¬ts to the elderly, while
others do more to protect children and working-age adults. A sweeping
work of historically and sociologically informed political science, Age in
the Welfare State offers a surprising challenge to the conventional wisdom
that welfare state policies are a result of either pressure-group politics or
the ideologies of parties in power. This vividly written and exhaustively
documented work draws on in-depth case studies of family, labor-market,
and pension policy making in Italy and the Netherlands, as well as broader
cross-sectional analysis of spending patterns in twenty OECD countries.
Scholars of social policy and comparative politics, practitioners, and policy
makers will be challenged by this book™s startlingly new insights about the
historical roots of current welfare state predicaments.

Julia Lynch is Assistant Professor of Political Science at the University of
Pennsylvania. Her recent dissertation, on which this book is based, garnered
the Gabriel Almond prize of the American Political Science Association for
the best dissertation in comparative politics. Professor Lynch was previously
a scholar in the Robert Wood Johnson Health Policy Scholars program at
Harvard University, and she has been a visiting researcher at the European
University Institute in Florence and the Luxembourg Income Study project
in Luxembourg.
Cambridge Studies in Comparative Politics

General Editor
Margaret Levi University of Washington, Seattle

Assistant General Editor
Stephen Hanson University of Washington, Seattle

Associate Editors
Robert H. Bates Harvard University
Helen Milner Princeton University
Frances Rosenbluth Yale University
Susan Stokes University of Chicago
Sidney Tarrow Cornell University
Kathleen Thelen Northwestern University
Erik Wibbels University of Washington, Seattle

Other Books in the Series
Lisa Baldez, Why Women Protest: Women™s Movements in Chile
Stefano Bartolini, The Political Mobilization of the European Left,
1860“1980: The Class Cleavage
Mark Beissinger, Nationalist Mobilization and the Collapse of the Soviet State
Nancy Bermeo, ed., Unemployment in the New Europe
Carles Boix, Democracy and Redistribution
Carles Boix, Political Parties, Growth, and Equality: Conservative and Social
Democratic Economic Strategies in the World Economy
Catherine Boone, Merchant Capital and the Roots of State Power in Senegal,
Catherine Boone, Political Topographies of the African State: Territorial
Authority and Institutional Change
Michael Bratton and Nicolas van de Walle, Democratic Experiments in
Africa: Regime Transitions in Comparative Perspective
Michael Bratton, Robert Mattes, and E. Gyimah-Boadi, Public Opinion,
Democracy, and Market Reform in Africa

Continued after Index
Age in the Welfare State


University of Pennsylvania
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press
The Edinburgh Building, Cambridge cb2 2ru, UK
Published in the United States of America by Cambridge University Press, New York
Information on this title: www.cambridge.org/9780521849982

© Julia Lynch 2006

This publication is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.

First published in print format 2006

isbn-13 978-0-511-21984-9 eBook (EBL)
isbn-10 0-511-21984-9 eBook (EBL)

isbn-13 978-0-521-84998-2 hardback
isbn-10 0-521-84998-5 hardback

isbn-13 978-0-521-61516-7 paperback
isbn-10 0-521-61516-x paperback

Cambridge University Press has no responsibility for the persistence or accuracy of urls
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
In memory of Rue Bunzelman Deutsch

List of Tables and Figures page x
List of Abbreviations xiii
Acknowledgments xvii


References 201
Index 219

Tables and Figures

2.1. Public Spending on the Elderly, per Person
Aged 65+ page 21
2.2. Public Spending on Unemployment and Active
Labor Market Policies, per Registered Unemployed 22
2.3. Public Spending on Occupational Injury and
Sickness Programs, per Member of the Civilian
Labor Force 22
2.4. Public Spending on Cash Bene¬ts and Services for
Families, per Person under 15 23
2.5. Public Education Spending, per Person Aged 0“20 24
2.6. Per Capita Health Spending Ratios 25
2.7. Elderly/Non-elderly Spending Ratio (ENSR) 30
2.8. ENSR with Education Spending 32
2.9. Social-Fiscal Measures as a Percentage of GDP 34
2.10. Tax Expenditures on Elderly and Non-elderly 35
2.11. Housing Policy Orientations 38
3.1. Percentage Change in ENSR, 1960“2000 45
5.1. Net Replacement Rates of Unemployment
Insurance Bene¬ts 113
5.2. Bene¬ts for the Unemployed in Italy and the
Netherlands 120
5.3. Age Distribution of Long-Term Unemployed
Population 125
5.4. Working-Age Disability Bene¬ciaries, by Age 125

Tables and Figures

1.1. Elderly/Non-elderly Spending Ratio (ENSR) 5
1.2. Age Orientation and Decommodi¬cation 6
1.3. Age Orientation and Demographic Structure 7
1.4. Age Orientation and “Level of Development” 7
1.5. Age Orientation and Total Welfare State “Effort” 8
3.1. Change in Economic Openness and Age Orientation,
1960“2000 51
3.2. Watersheds of Welfare State Formation 57
3.3. Age Orientation and Welfare State Program Structure 59
4.1. Public Spending on Families, per Person Aged 0“14,
Italy and the Netherlands 72
5.1. Percentage of Unemployed Receiving Unemployment
Insurance Bene¬ts, Italy and the Netherlands 123
6.1. Pension Spending as a Percentage of Total Nonhealth
Social Expenditures, Italy and the Netherlands 141
6.2. Pension Spending per Person Aged 65+ as a Percentage
of GDP per Capita, Italy and the Netherlands 144
7.1. Welfare State Spending and Poverty Reduction:
Children 182
7.2. Welfare State Spending and Poverty Reduction:
Non-elderly Adults 183
7.3. Welfare State Spending and Poverty Reduction:
Elderly Adults 184


ANF assegno per il nucleo familiare (family allowance)
CENSIS Centro Studi Investimenti Sociali (Center for Social Studies and
Policies, socio-economic research institute)
CGIL Confederazione Generale Italiana del Lavoro (left-leaning labor
union confederation)
CIG Cassa per l™Integrazione Guadagni (short-time earnings replace-
ment bene¬t program)
CIGS Cassa per l™Integrazione Guadagni Straordinaria (special short-
time bene¬t)
CISL Confederazione Italiana Sindacati Lavoratori (Catholic-
inspired labor union confederation)
CNEL Consiglio Nazionale dell™Economia e del Lavoro (National
Council on Labor and the Economy, tripartite consultative
CUAF Cassa Unica per gli Assegni Familiari (Uni¬ed Family Allowance
DC Democrazia Cristiana (Christian Democratic Party)
FNP Federazione Nazionale Pensionati (pensioners™ union af¬liated
with CISL)
FPLD Fondo Previdenza Lavoratori Dipendenti (largest state-
administered pension fund, covering dependent employees)
INCA Istituto Nazionale Confederale di Assistenza (social service
agency linked to CGIL)
INPS Istituto Nazionale della Previdenza Sociale (national social secu-
rity administration)


ISTAT Istituto Nazionale di Statistica (national statistical agency)
PCI Partido Comunista Italiano (Communist Party of Italy)
(P)DS (Partido) Democratici di Sinistra (Democratic Party of the Left,
formerly PCI)
PSI Partido Socialista Italiano (Socialist Party of Italy)
TS trattamenti speciali (special unemployment insurance bene¬ts)
UIL Unione Italiana del Lavoro (Centrist labor union confederation)

The Netherlands
ABW Algemene Bijstandswet (Unemployment Assistance Act)
AKW Algemene Kinderbijslagswet (General Family Allowance Act)
AOW Algemene Ourderdomswet (General Old-Age Pensions Act)
ARP Anti-Revolutionaire Partij (Protestant Reform Party)
CBS Centraal Bureau voor de Statistiek (national statistical agency)
CDA Christen Democratisch Appel (Christian Democratic Appeal,
formed in 1980 from merger of Catholic and Protestant parties)
CNV Christelijk Nationaal Vakverbond (Catholic trade union confed-
FNV Federatie Nederlandse Vakbeweging (largest Dutch trade union
JWG Jeugdwerkgarantiewet (Youth Work Guarantee Law)
KVP Katholieke Volkspartij (Catholic Peoples™ Party)
NVV Nederlands Verbond van Vakverenigingen (Socialist trade union
confederation, merged to form part of FNV in 1981)
PPR Politieke Partij Radicalen (Radical Party)
PvdA Partij van de Arbeid (Labor Party)
RWW Rijksgroepsregeling voor Werkloze Werknemers (Unemploy-
ment Assistance Act)
SER Sociaal-Economische Raad (Socio-Economic Council)
SVB Sociale Verzekeringsbank (Social Insurance Bank)
VUT Vervroegde Uittreding (private early retirement pension provi-
WAO Wet op de Arbeidsongeschiktheidsverzekering (Disablement
Insurance Act)
WIW Wet Inschakeling Werkzoekenden ( Job-Seekers Employment
WW Wet Werkloosheidsvoorziening (Unemployment Insurance


WWV Wijziging Wet Werkloosheidsvoorziening (Extended Unem-
ployment Insurance Act)

ILO International Labour Of¬ce
IMF International Monetary Fund
LIS Luxembourg Income Study project
OECD Organization for Economic Cooperation and Development

ENSR Elderly/Non-elderly Spending Ratio
GDP Gross Domestic Product

Country Abbreviations Used in Figures
AUS Australia
AUT Austria
BEL Belgium
CAN Canada
DEN Denmark
FRA France
FIN Finland
GER Germany
GRE Greece
IRE Ireland
ITA Italy
JPN Japan
LUX Luxembourg
NET Netherlands
NOR Norway
NZL New Zealand
POR Portugal
SPA Spain
SWE Sweden
UK United Kingdom
US United States


This book has been many years in the making, the vast majority of which
have been enjoyable. I attribute this in large part to the fact that it was
truly a joint effort. It would not have been possible for me to research,
write, or complete it without the invaluable contributions of various funding
institutions, mentors, colleagues, friends, and family.
I owe deep intellectual debts to many: to Gøsta Esping-Andersen, who
set me on the right track; to John Zysman, Jonah Levy, and Henry Brady,
who gave me the right tools and showed me how to use them; to Maurizio
Ferrera and Tim Smeeding, who always showed faith and backed it up
with good works; and especially to Karen Anderson and Sara Watson, who
have always been generous and who just keep getting smarter every year.
Numerous colleagues “ among them Melani Cammett, Andrea Campbell,
Anna Grzymala-Busse, Katie Carman, Evan Lieberman, Lauren Morris
MacLean, Paul Pierson, Mark Vail, Rob Weiner, Christa van Wijnbergen,
and Daniel Ziblatt “ gave intelligent feedback at crucial moments. I cannot
thank these wonderful people enough.
Field work for this project was ¬nanced with a National Science Foun-
dation Graduate Research Fellowship, a Social Science Research Council
International Dissertation Field Research Fellowship, and an Alan Sharlin
Memorial Award from the Institute for International Studies at the Univer-
sity of California, Berkeley. Financial support during writing and rewrit-
ing came from another Sharlin Award and a John L. Simpson Memorial
Research Fellowship, also administered by the Institute for International
Studies at Berkeley, and from a most generous fellowship from the Robert
Wood Johnson Health Policy Scholars program.
While in Italy, I relied heavily on the kindness of the European University
Institute in Florence. Since my ¬rst venture to Italy in 1993, the faculty,

staff, and researchers of the EUI have provided me with a home away from
home and a vibrant intellectual community. The Robert Schuman Centre™s
1998“9 European Forum on “Recasting European Welfare States” provided
an ideal environment for testing out new ideas during my stay in Italy and
beyond. I would particularly like to thank Stefano Bartolini and Martin
Rhodes for making the resources of the EUI available to me on repeated
My research in Italy would not have been possible without the support
of Maurizio Ferrera, whose kindness, generosity, and belief in the project
has sustained me through many rough patches. I would also like to thank
Marino Regini and Daniele Franco, as well as librarians Peter Kennealy
at the EUI, Oreste Bazzichi at Con¬ndustria, and Mila Scarlatti at the
Centro Studi CISL. Aedin Doris, Jackie Gordon, Ann-Louise Lauridsen,
Dan Oakey, Jacobien Rutgers, and Joanna Swajcowska supplied moral and
immoral support at crucial moments during my Italian sojourns.
In the Netherlands, Anton Hemerijck was a welcoming beacon, setting
me up with logistical support from the University of Leiden and provid-
ing me with the feedback and intellectual support necessary to research
a case study effectively in a short period of time. The staff of the library
at the Ministry of Social Welfare and the experts gathered at the Hugo
Sinzheimer Institute in Amsterdam amazed me with their patience and
expertise. Nelleke van Deusen-Scholl and Heleen Mastenbroek managed
to teach me workable Dutch in a period of about ¬ve months, a feat that
shall never cease to amaze me. Karen Anderson, Jacobien Rutgers, and
Bauke Visser helped to make my time in the Netherlands pleasant as well
as productive.
Logistical support in the ¬nal phases has come from my terri¬c TS-
CS guru, Ben Goodrich; from Todor Enev, whose deeply intelligent data
sleuthing makes the term “research assistant” utterly inadequate; and
from Melanie Daglian, whose positive energy made preparing the ¬nal
manuscript an enjoyable task. Tom and Emma were there at the beginning
and saw it through to the end. There are no two better friends with whom to
go through life. Last, this book is dedicated to my very special grandmother,
Rue Bunzelman Deutsch. As we both got older, she became my partner in
crime, showing me that sometimes the true meaning of intergenerational
solidarity comes down to poking fun at the middle generation.

Philadelphia, Pennsylvania
November 2005


Welfare states work better for some age groups than for others. Social pro-
grams in the United States and Italy, for example, do little to raise children
out of poverty, but elderly citizens are made better off by the substantial
bene¬ts available to them. In other countries, such as Norway and Portugal,
senior citizens™ incomes on average are lower than in the United States or
Italy, but low-income workers, families with children, and the long-term
unemployed receive signi¬cant support from the welfare state. Across the
industrialized countries, social programs such as public pensions, family
allowances, and bene¬ts for the unemployed vary signi¬cantly, with conse-
quences for the well-being of different age groups in the population.
This book asks how social policies in rich democracies buffer and channel
risks for the aged, the young, and working-age adults. What do different
welfare states do for their elderly and non-elderly citizens? Why does the
age orientation of social policies vary from country to country and over
time? And what are the political consequences of different strategies for
redistributing resources across different age groups in society? How and
why welfare states distribute resources to different age groups is linked to
broader questions of theory in comparative politics: What are the important
dimensions of similarity and difference among different modes of economic
regulation? Which actors impact political-economic outcomes? What is the
relative importance of social and economic structures, political practices,
and institutional legacies in determining the policies pursued in different
The welfare state™s role in caring for young people and the elderly plays an
important part in political debates about welfare reform. An alleged elderly
bias in American social spending has, during recent years, nourished intense
political debates about generational equity. In many European countries,
Age in the Welfare State

relatively high incomes from pensions and increasing rates of child poverty
provide a fertile environment for the emergence of a parallel discussion.
Unequal bene¬ts for the old and the young provide ammunition for those
who advocate providing more support for people at all stages of the life
course, but also for those who wish to cut existing bene¬ts in the name of
intergenerational equity. These inequalities also serve as a reminder that
welfare states can differ objectively and dramatically in their ability to insure
diverse age groups in society against risks such as poverty, ill health, or social
This book begins with an analysis of social spending patterns in twenty
industrialized democracies. Welfare states do in fact differ quanti¬ably
in the age orientation of their social policies. The ¬rst half of the book
establishes a strategy for conceptualizing and measuring these differences
(chapter 2), and then explores a series of competing hypotheses about why
countries might vary in the age orientation of their social policy regimes
(chapter 3). The second half of the book ampli¬es and tests these rival
hypotheses systematically using paired case studies. Case studies of the
development of three key social programs in Italy and the Netherlands “
family allowances (chapter 4), unemployment bene¬ts (chapter 5), and old-
age pensions (chapter 6) “ demonstrate the path by which two countries,
sharing a set of common ideological orientations and facing similar labor
market and demographic conditions in the immediate postwar period,
arrived at welfare states that allocate very different roles to the state in
distributing resources across generations.

Why Study the Age Orientation of Welfare States?
Welfare states vary in the extent to which they protect older and younger
citizens. But traditional theories of welfare state development neither notice
nor explain this variation. If welfare state scholars have until now preferred
to focus on the cross-class, cross-occupation, or cross-gender distribution
carried out by social policies, why should we now be concerned with the age
pro¬le of welfare states? Put simply, it is because changing socio-economic
conditions mean that how welfare states cover the risks associated with
different stages of the life course has become more important.
Advanced industrialized societies today are aging. At the same time,
labor markets are changing, and family structures evolving. The male-
breadwinner model of social organization, premised upon stable, lifelong


employment for men, has given way to more frequent or longer periods
of unemployment. Families, long called upon to provide for needs not met
in the marketplace or by the state, are stretched to new limits. But this is
occurring just as their capacity to respond is reduced by increasing female
employment outside the home, divorce, and changing fertility patterns. In
the context of current demographic, labor market, and family changes, how
welfare states address the risks faced by people at different stages in the life
course affects both citizens™ lives and the capacity of national economies to
adapt to new conditions.
Demographic, social, and economic transformations confronting even
the most “traditional” of Western societies affect the foundations of the
political economic orders established in the period after the Second World
War. How will welfare state institutions, which were created under radi-
cally different demographic, social, and economic circumstances, respond
to these changes? How well will traditional institutions of social policy
buffer citizens as they adapt their lives to the new social risks associated
with changing work patterns and family demands? Will political sponsors
of the welfare state be able to balance pressure from constituencies to both
maintain established entitlements and meet new needs?
To evaluate how welfare states will stand up to these new pressures, we
need to understand how they address the risks encountered by people at dif-
ferent stages in the life course. Quite apart from normative concerns about
intergenerational justice, it is worth understanding how welfare states treat
different age groups because this affects crucially the decisions individuals
make about labor market participation, family organization, and invest-
ment and savings strategies. When welfare states direct resources toward
families with children, for example, it can affect fertility rates, female labor
force participation, and the professional preparedness of young adults. The
division of labor among family, market, and state in caring for young chil-
dren or the frail elderly may affect both women™s emancipation and the
quality of care provided. The structure and extent of public pension sys-
tems of course has consequences for labor costs and ¬nancial markets,
but can also set limits on the speed and ¬‚exibility with which welfare
states retool to meet new needs that affect adults during their working
years. In sum, the capacity of welfare states to respond to new challenges
depends critically on a characteristic that has received almost no attention
in the literature on comparative social policy: the age orientation of social

Age in the Welfare State

Why Does Age Orientation Vary? Some Preliminary Evidence
and Hypotheses
The age-orientation of social policies, as chapter 2 demonstrates in some
detail, varies dramatically across advanced industrialized countries and in
ways that upset our traditional notions of family relationships among differ-
ent types of welfare states. Figure 1.1 shows the average for the years 1985
to 2000 of the ratio of direct social expenditures on the elderly (pensions
and services for the elderly) to spending on the non-elderly (unemployment
bene¬ts, active labor market policy, family allowances, and family services),
adjusted for the relative size of elderly and non-elderly populations in each
of twenty OECD (Organization for Economic Cooperation and Devel-
opment) countries. I call this measure the Elderly/Non-elderly Spending
Ratio, or ENSR. It allows us to estimate the relative weight of spending
on the elderly “ people aged sixty-¬ve and above or in formal retirement “
versus that on working-age adults and children. This spending measure is
of course only an approximation of the full range of services and bene¬ts
offered to different groups, many of which we consider in more depth in
chapter 2. But the ENSR serves to introduce us to the range of variation
across countries in the age orientation of social policies.
The most striking feature of the age orientation of welfare states is its
transgression of the boundaries set by Esping-Andersen™s (1990) seminal
division of advanced countries into three “worlds” of welfare capitalism.
The least elderly-oriented countries among the twenty OECD nations
considered here are a mix of his “Liberal,” “Conservative-Corporatist,”
and “Social Democratic” regimes. At the same time, two of Esping-
Andersen™s Liberal regimes, the United States and Japan, are clearly among
the most elderly-oriented. Likewise, Conservative-Corporatist regimes run
the gamut from relatively youth-oriented Belgium and the Netherlands to
elderly-oriented Italy and Austria. The lack of correspondence between the
ENSR and Esping-Andersen™s key concept, decommodi¬cation, is easy to
see in Figure 1.2. The relief from market forces that social policies pro-
vide is surely an important measure of the welfare state. But it is not
enough to ask how much welfare states decommodify; we must also ask
whom they decommodify.
Alternative typologies fare no better when confronted with the data on
age orientation. “Christian Democratic” welfare states (van Kersbergen
1995) are as likely to be youth-oriented (the Netherlands) or age-
neutral (Germany) as they are to throw their support to the elderly (Italy).











Figure 1.1 Elderly/non-elderly spending ratio (ENSR), average 1985“2000. Sources: Spending data from
OECD 2004; demographic data from OECD 2003b.

Age in the Welfare State






13 27 39

Decommodification score

Figure 1.2 Age orientation and decommodi¬cation. Sources: Spending data from
OECD 2004; demographic data from OECD 2003b; decommodi¬cation scores
from Esping-Anderson 1990.

Neither do Mediterranean countries cluster neatly, contrary to scholarship
suggesting a distinctive Southern European welfare state type (Leibfried
1992; Ferrera 1996c; Rhodes 1997). Italy and Greece look like classic “pen-
sioner states” (Esping-Andersen 1997), but Portugal resembles Canada, the
United Kingdom, and Germany more closely than it does its Southern
European neighbors. The weak correspondence between the age orienta-
tion of social policy regimes and welfare state “worlds” or “families” suggests
that there is an important dimension of variation among different kinds of
welfare states that familiar typologies do not capture.
If standard typologies of welfare state outcomes do not correspond to
the variation we™ve observed, it should not surprise us that the causes of
divergent welfare state characteristics typically cited in the literature also
fail to predict differing age orientations. As the bivariate scatter plots in
Figures 1.3 to 1.5 suggest, neither the demographic structure of a country™s
population, its wealth or “level of development,” nor the overall size of the
welfare state predict consistently how welfare states will allocate resources
to the elderly and non-elderly in their populations. Figure 1.5 does show
an inverse relationship between total social spending and the age orientation









11 14 18

Percent of pop. aged 65+

Figure 1.3 Age orientation and demographic structure. Source: See Fig. 1.1.








10585 17175 24220
GDP per capita

Figure 1.4 Age orientation and “level of development.” Sources: Spending data
from OECD 2004; demographic and GDP per capita data from OECD 2003b.

Age in the Welfare State







12 22 31
Non-health social spending
as a percentage of GDP

Figure 1.5 Age orientation and total welfare state “effort.” Sources: Total non-
health social expenditure data from OECD 2004; demographic data from OECD

of the welfare state (bigger welfare states tend to be less elderly-oriented),
but the presence of two very elderly-oriented outliers makes the relationship
seem much stronger than it might be for the remaining countries. These
data reveal the important point that there are both small ( Japan) and large
(Italy) elderly-oriented welfare states, and both small (Ireland) and large
(Sweden) youth-oriented welfare states. At the same time, classic “power
resources” variables, such as the strength of organized labor, employers™
preferences, and the relative power of Left and Christian Democratic polit-
ical parties, fall short of explaining differences in the age orientation of
welfare states, as we see in chapter 3.
Why don™t classic theories of welfare state development explain these
outcomes? Some scholars have posited that the demographic structure of
a population affects welfare state policies. In particular, the elderly are said
to have distinctive needs and distinctive preferences that drive welfare state
spending (see, e.g., Wilensky 1975; Pampel and Williamson 1989; Thom-
son 1989, 1993). These authors argue that traditional welfare state theories
miss an important set of political actors, the elderly, because they focus
too narrowly on class-based actors. A major aim of this book is to test this

hypothesis about the political in¬‚uence of demographic groups. Can dif-
ferent mixes of welfare bene¬ts for the young and old across countries and
across time be explained by pressure from welfare state constituencies in
the form of age-based lobbies?
The criticism that standard welfare state theories ignore nonclass actors
has merit, but shifting the focus to the role of age-based actors does not
account for diverging welfare state age pro¬les. Two far more important
problems in the comparative welfare state literature need to be addressed
before it can be made to account adequately for the outcome that we are
trying to explain. First, the prevailing view of politicians as largely motivated
by programmatic goals must be revised to take into account nuances in
the varieties of political competition. Second, we must consider how the
institutional environment within which electoral competition takes place
shapes welfare state regimes.

Explaining Variation in Age Orientation: The Argument in Brief
If welfare states vary in surprising ways in their protection of older and
younger age groups in the population, how can we explain this variation?
Why do some welfare states emphasize protection for risks during child-
hood and the working life, while others focus more on covering needs in old
age? This book argues that two types of institutions explain this variation:
the structure of welfare state programs enacted in the early twentieth
century “ occupationalist or citizenship-based “ and the dominant mode
of political competition in a polity, particularistic or programmatic.
First, as we see in chapter 3, the structure of early welfare state programs
affects the populations (labor market “insiders” vs. “outsiders”) that are cov-
ered by public welfare programs. Since these populations take on distinctive
age pro¬les with the development over time of both public and private social
insurance schemes, the choice of which population to cover strongly in¬‚u-
ences the eventual age orientation of social policy regimes. Second, the type
of political competition characteristic of a party system affects the devel-
opment of welfare state programs in the post“World War II period and
determines whether elderly-oriented occupationalist welfare regimes can
“switch tracks” by adding more youth-oriented citizenship-based programs.
The policy studies in chapters 4 through 6 reveal af¬nities between particu-
laristic politics and fragmented occupationalist social insurance regimes that
make program structure and the mode of political competition extremely
dif¬cult to uncouple. In sum, this book argues that patterns of partisan
Age in the Welfare State

competition and social policy structures interact over time to produce
durable, mutually reinforcing constellations of social policies that mature
into either elderly-oriented or more youth-oriented welfare states.

Two Watersheds of Welfare State Formation
At the heart of the distinction between groups of countries with similar age
orientations lie two historical bifurcations in the paths of social policy devel-
opment. The ¬rst split, the basic genetic division between citizenship-based
and occupational regimes, occurred in the late nineteenth and early twen-
tieth centuries, when modern states grappled with new social and political
problems arising from industrialization. A second watershed occurred in the
decades around the Second World War, when most countries with occupa-
tionalist welfare systems considered adopting citizenship-based social pol-
icy regimes, but only a select group actually took concrete steps in this
The initial split between citizenship-based and occupational social wel-
fare regimes had profound consequences for the eventual age orienta-
tion of welfare spending.1 In the countries adopting citizenship-based
regimes (the Scandinavian and British Commonwealth countries), public
welfare provisions developed in the gaps not covered by mutual-aid pro-
grams run by labor unions. State welfare spending supplemented pre-
existing private occupational bene¬ts, and so focused on the risks most
likely to be encountered by people who were not covered by mutual-
ist bene¬ts. In Manow™s (1997) terminology, such welfare regimes “com-
pensated” for the gaps in private coverage, offering bene¬ts for children,
women, and elderly citizens without pensions. Citizenship-based regimes
contained the seeds of programs that would later develop into the main-
stays of youth-oriented welfare states: support for mothers and children,
and comprehensive social assistance for those with weak ties to the labor

1 It should be noted that in practice many welfare states mix citizenship-based and occupa-
tionalist program types. Even prior to World War II, Sweden, for example, had a pension
system that combined a ¬‚at-rate citizenship-based bene¬t with a supplementary contrib-
utory tier offering bene¬ts graded according to occupation. However, throughout this
book I label welfare programs that have a substantial component that is available to citizens
regardless of occupation or contributory history as “citizenship-based,” to distinguish them
from programs in which there is no universal or means-tested entitlement independent of
labor market status.


In those countries that eventually became more elderly-oriented, labor
movements in the late nineteenth century relinquished control over
autonomous forms of social insurance to the state. Public social insurance
programs thus built on the framework of occupational programs that unions
had constructed to bene¬t their own members. This technique of “upgrad-
ing” private occupational social insurance schemes by transforming them
into state-run programs (Manow 1997) resulted in public welfare bene¬ts
that focused almost entirely on the needs of people with close ties to the
labor market. In these states, social protection for groups outside the labor
market remained the province of nonstate actors, primarily families and
charities. Protection for people af¬liated with the core labor market was
provided by the state, setting the stage for elderly-oriented welfare spend-
ing in much of Continental Europe, the United States, and Japan as core
work forces aged dramatically in the 1970s and 1980s. Thus, the structure
of welfare programs initiated in the late nineteenth and early twentieth
centuries laid the groundwork for different types of spending, resulting
in a basic division between youth-oriented universalist and means-tested
welfare states, on the one hand, and more elderly-oriented occupationalist
regimes, on the other.
A second watershed in welfare state development, in the decades around
World War II, introduced further variation into the structure of welfare
state regimes, and hence into the age-orientation of welfare spending
in different countries. During and immediately after the Second World
War, most countries with occupationalist welfare states considered legisla-
tive proposals to introduce substantial elements of the citizenship-based,
Beveridgean model pioneered in victorious Britain (Ferrera 1993). Some
countries succeeded in this agenda, introducing forms of citizenship-based
coverage for children, women, and others with weak ties to the labor market.
Other states, however, did not, and continued on a path toward growing
expenditures on an aging core work force and occupational pensioners, with
minimal coverage for the rest of the population.
How can we account for the persistence of occupationalism in some
countries and the introduction of more youth-oriented citizenship-based
welfare policies in others after World War II? The opposing slopes of this
second watershed are characterized by different modes of political compe-
tition prevalent in different countries. The countries that did not adopt uni-
versal programs in the 1930s through 1960s shared a particularistic mode of
political competition that inhibited the development of substantial univer-
sal welfare programs. As the years passed, highly fragmented social security
Age in the Welfare State

programs continued to provide resources for clientelist politicians and to
obscure the costs of political clientelism, resulting in a self-reinforcing cycle
of particularistic politics, fragmented occupational welfare programs, and
elderly-oriented spending.

Italy and the Netherlands: Contrasting Case Studies
The case studies of Italian and Dutch social policies in chapters 4 through
6 illustrate the “mechanism of reproduction” (see Thelen 1999; Pierson
2000) that has sustained these path-dependent welfare policy outcomes
after World War II. Both Italy and the Netherlands had pure occupational-
ist welfare regimes before World War II, and in both countries after the war
of¬cial reform commissions (the D™Aragona Commission in Italy, the van
Rhijn Commission in the Netherlands) advocated moving to a universal-
ist, citizenship-based system. Other similarities, too, lead us to expect that
the Netherlands and Italy would follow a similar path after the war. Both
countries belong to Esping-Andersen™s (1990) Conservative-Corporatist
world of welfare; in both countries the major expansion of the welfare
state in the postwar period was carried out under coalitions dominated by
Christian Democratic parties; and in both countries labor relations regimes
were characterized by numerically weak unions and sporadic tripartite con-
certation. Yet Italy has a highly elderly-oriented, occupational welfare sys-
tem, whereas the Netherlands is quite youth-oriented and characterized by
a mix of occupational and citizenship-based programs. The Netherlands
succeeded in implementing a number of new universalist welfare programs
after World War II, while Italy, despite repeated attempts to do so, did
not. As a result, the Netherlands entered the 1990s with a far more youth-
oriented welfare system than did Italy.
Why did Italy remain a strongly occupationalist welfare regime, while
the Netherlands adopted many citizenship-based programs? The key to
understanding this difference is the very different ways that political com-
petition has been organized in the two countries for much of the postwar
period. Italian politics has been characterized by an extremely high degree
of political particularism. By contrast, politics in the Netherlands has tended
toward the programmatic end of the spectrum. This difference in the mode
of political competition between Italy and the Netherlands explains why the
Netherlands adopted citizenship-based welfare programs, such as universal
pensions, universal family allowances, and a basic social minimum, while
Italy did not.

Clientelism and occupationalism interacted to prevent Italian social
reformers from introducing the citizenship-based welfare regime envi-
sioned by the D™Aragona Commission in 1947. Politicians associated with
both the Christian Democratic Party and the Italian Left were in¬‚uenced
by the atmosphere of particularistic political competition to block the har-
monization of pension bene¬ts and the introduction of universal public
pensions in Italy “ not just in 1947, but also at least once in every decade
since. Clientelist politicians also resisted the development of neutral state
capacities such as effective taxation, which in turn stymied attempts to intro-
duce universal bene¬ts for children, the unemployed, and the elderly. And
the complexity of occupationalist programs made it dif¬cult for either the
public or policy experts to see the results when politicians offered selective
bene¬ts in return for votes.
If in Italy the combination of fragmented, occupational welfare programs
and particularistic political competition derailed attempts to introduce new
universal social programs after the Second World War, the opposite was
true in the Netherlands. There neutral state capacities such as universal
taxation made it possible to introduce citizenship-based programs fairly
easily. The ability to levy and collect taxes on the self-employed and farm-
ers, in particular, secured labor and the Left™s support for agreements that
introduced universal family allowances and pensions.
The simplicity and transparency of universal programs in turn made it
dif¬cult for politicians to exchange highly targeted bene¬ts for votes. In
fact, once programs were universalized, it became impossible to increase
bene¬ts for one group without increases for all recipients. In the case of
family allowances and unemployment bene¬ts, this led to a gradual esca-
lation of bene¬ts, and when high unemployment hit in the mid-1970s,
costs for these programs soared as the number of bene¬ciaries increased
dramatically. In the case of public pensions, the sheer size of a program
dedicated to providing a decent standard of living for the entire elderly
population, combined with the ease with which future outlays could be
predicted, made many potential advocates of higher pensions think twice
before demanding bene¬t increases. The simplicity and transparency of
citizenship-based social programs tended to increase pressure for spend-
ing in the bene¬t categories that did not provide full income replacement
over a long period “ generally youth-oriented programs, such as family
allowances or unemployment insurance “ and reduced the pressure to
grant large increases in more expensive bene¬t categories such as old-age
Age in the Welfare State

The case studies in chapters 4 through 6 highlight three distinct mech-
anisms by which the structure of welfare state programs and the mode of
political competition combine to affect the age orientation of social policies.
First, the distinction between occupational and citizenship-based welfare
programs determines how politicians can use welfare bene¬ts as tender in
the competition for votes, and thus alters politicians™ preferences about the
level of various types of bene¬ts. Second, the structure of social programs
affects how salient different types of bene¬ts are to potential recipients, and
how visible are the effects of decisions about where to allocate resources.
Finally, the mode of political competition affects the resources available to
politicians and policy makers who might wish to expand particular social
This essentially institutionalist explanation for the variation in the age
orientation of welfare states poses a challenge to the existing literature on
comparative social policy on three fronts. First, the argument presented
here demands that welfare state outcomes be analyzed in relation to other
public policies. In particular, the link between tax systems and welfare ben-
e¬ts is revealed to be a crucial one, which affects both the kind of welfare
bene¬ts that constituencies demand and what politicians can offer to meet
that demand. Second, this book argues that politicians matter for welfare
state outcomes not so much because of their ideological orientations but
because of the way they compete for votes and of¬ce. Finally, this argu-
ment downplays the role of welfare state constituencies in bringing about
the policies that bene¬t them, and asserts instead the causal primacy of
long-term processes and interactions between program structure and politi-
cians™ behavior. In other words, it supports a sharp distinction between
welfare state regimes as the revealed preferences of powerful social groups,
and policies as outcomes of institutionally structured processes of political


Measuring the Age of Welfare

Welfare states clearly work to transfer resources between age groups, not
least through pay-as-you-go old-age pensions, which account for one-¬fth
to one-half of total social spending in most countries of the OECD. But the
elderly in different countries bene¬t to varying extents not only from cross-
national differences in the generosity of pension bene¬ts, but also from
differences in other policy areas, such as housing and health care. Similarly,
working-age adults and children bene¬t from a variety of programs ¬nanced
by the population at large, including education, publicly provided child care,
and income supports.1
The concept of intergenerational justice has prompted a robust theoreti-
cal literature, but little empirical investigation.2 In particular, we know very
little about how social provisions for different age groups vary across wel-
fare state types, across countries, or across time. Because so little is known
about the age-distributive properties of social policies, it is dangerous to
conclude that “the contemporary welfare state in capitalist democracies is
largely a welfare state for the elderly” (Myles 1989). Nor can we be sure that,
as some have argued, a single “sel¬sh generation” that reached adulthood
just after the Second World War has tailored welfare state spending for
its own purposes (Thomson 1993). Without reliable measures of the age
orientation of social policies across nations and over time, it is impossible
to know to what extent contemporary welfare states are biased toward the
elderly of particular generations, toward successive cohorts of the elderly,
or even if they are uniformly biased toward the elderly rather than the

1 This chapter is based substantially on Lynch 2001.
2 See, e.g., Daniels 1988; Johnson, Conrad, and Thomson 1989; and Laslett and Fishkin 1992.

Age in the Welfare State

In this chapter we consider a series of indicators of the age orientation
of social policies. The resulting rankings group countries quite consistently
according to how a variety of social policy instruments “ direct expendi-
tures on social insurance programs, labor market policies, education, and
health care, as well as indirect tax expenditures and housing policies “
allocate resources to different age groups. The most consistently elderly-
oriented welfare states in the sample of OECD countries considered here
are Japan, Italy, Greece, the United States, Spain, and Austria. The most
youth-oriented are the Scandinavian countries, the Netherlands, and the
English-speaking countries other than the United States. A group of Con-
tinental European countries “ Germany, France, Belgium, Luxembourg,
and Portugal “ occupies the middle ground. Because these groupings of
countries are so consistent across policy areas, it is possible to develop a
simple measure of the age orientation of social policy regimes that uses just
a few pieces of readily available aggregate social spending data.
But given that families often share resources across generations, or pur-
chase private insurance that acts as a form of resource transfer across the
life course, is the age orientation of state policies really the form of inter-
generational transfer with which we should be most concerned? And are
aggregate spending measures really the best way to capture the variation in
state policies?
Determining the age orientation of individual social policies can be dif-
¬cult since policies often have effects, and re¬‚ect priorities, other than
those most obviously indicated in statutes. Early retirement provisions in
Italy that allowed female public-sector workers to retire at full pay after
only ¬fteen years of service are a good example. One could interpret these
“retirement” bene¬ts not as a transfer to the elderly, but rather as family pol-
icy camou¬‚aged for a context in which direct subsidies to working mothers
were unacceptable to politically powerful religious forces (Saraceno 1994,
70). Because policies may re¬‚ect hidden priorities of policy makers and
may bene¬t more than one speci¬ed target group, it is risky to draw con-
clusions about who social programs are really intended to help based solely
on spending data, without going deeper into the political struggles behind
the policies™ implementation. This chapter works with aggregate spending
data to sketch a preliminary portrait of policy priorities; case studies in
chapters 4 through 6 ¬‚esh out this sketch.
Welfare transfers taking place within the state sphere are likely to be
closely intertwined with intergenerational transfers that take place within
families and in the context of private markets. Still, state policies toward
Measuring the Age of Welfare

different age groups are important even if they do not re¬‚ect the total out-
put of the state-market-family nexus for particular age groups. The distri-
butional consequences of effecting intergenerational transfers via families,
markets, or the state are not neutral. Welfare states take on distinctly dif-
ferent purposes when redistribution is limited to transfers within families,
rather than between families; and power structures within families are also
likely to re¬‚ect resource ¬‚ows directed by the state.
It is tempting to allow the family to continue to serve as a black box
obscuring the importance of state-sponsored redistribution to different age
groups. Intergenerational ties and resource sharing within families are sup-
posed to be the glue that prevents an explosion of tensions between age
groups similar to the global upheavals of 1968. But the structure of pay-
as-you-go social insurance programs may provide a simpler explanation for
the current quiescence of younger cohorts in the face of elderly-oriented
welfare state spending. When populations and real wages are both grow-
ing, transfers from the young to the old appear to be nothing more than
transfers across the life course “ younger people pay for bene¬ts that they
themselves will receive as they age. Under these circumstances, politiciza-
tion of differences in welfare spending on different age groups is unlikely. As
demographic and economic growth both slow, however, there is pressure
to balance social insurance budgets by increasing contributions now and
cutting bene¬ts in the future. The potential for politicization of con¬‚icts
between age groups over the apportionment of state resources becomes
important under these circumstances, though interpersonal ties between
generations may mitigate the effects somewhat. Again, it is worth investi-
gating the age priorities of state spending because these priorities have a
political impact, even when buffered by the resources of families.

The State of the Art: Work on the Generational Effects of Welfare
Policy to Date
Two main strategies for measuring the generational effects of welfare poli-
cies are in evidence in the existing literature: “generational accounting,”
which emerged in the 1990s as the major form of economic research on
aging and social policy at the macro level, and an older body of sociological
work that sought to evaluate the effects of social policies on the life chances
of different age groups.
Generational accounting models (see, e.g., Kotlikoff and Liebfritz 1998)
evaluate current tax structures and bene¬t patterns to calculate the lifetime
Age in the Welfare State

tax-bene¬t position of speci¬c age cohorts in a given country. Applying a
standard discount rate, these models sum the total remaining lifetime taxes
versus total remaining lifetime bene¬ts in order to arrive at a ¬gure known
as a generational account for a person of a given age. For a person around
retirement age, the generational account will generally be low or negative,
since recent retirees have paid most of the taxes they will pay in their lifetime
and are about to receive a large infusion of bene¬ts in the form of a pension.
Following the same logic, a person at age thirty will tend to have a much
higher generational account: a lifetime of income taxes lies ahead, while the
education bene¬t has already passed and the pension bene¬t is far in the
future. Calculating the generational account for a person born today will
indicate the overall lifetime tax-bene¬t position of a newborn, assuming no
change in tax or bene¬t structures.
Generational accounting provides a useful comparative baseline for
assessing the impact of present tax and transfer programs on different
cohorts, but the highly aggregate nature of the accounts makes interpre-
tation dif¬cult. The combination in one measure of all tax and bene¬t
programs, not just those relevant to social protection, makes it hard to
individuate the effects of welfare policy per se. Furthermore, the use of
discount rates means that accounts for any given age group are highly sen-
sitive to the value of the most proximate tax or bene¬t program. A third
limitation of the generational accounting technique is that accounts for all
age groups assume constant tax and transfer policies. This means that for
the generational accounts to re¬‚ect real aggregate gains (or losses) for a
given age group compared with any other, policies would have to remain
unchanged from the date of birth of the oldest cohort until the date of death
of the youngest. While generational accounts are useful for comparing the
lifetime tax-bene¬t position of newborns across countries were policies to
remain unchanged, they are of little utility (as Kotlikoff and Liebfritz are
careful to point out) in comparing the lifetime accounts of generations that
have actually lived through, or expect to live through, a great deal of policy
The generational accounting framework is concerned with the question
of generations, strictly speaking, not age groups. These concepts are related
but distinct. Public policies may be neutral with respect to generations “
that is, they do not effect signi¬cant transfers between groups of citizens
born at different points in time “ but at the same time are biased toward
a particular age group. A purely contributory pension system, into which
people make payments when they are young and out of which they draw
Measuring the Age of Welfare

bene¬ts when they are old, would fall into this category. Conversely, one
could imagine an age-neutral policy that effects large intergenerational
transfers “ for example, de¬cit spending resulting from a tax cut that is
carefully designed to affect levies on wage income and pension income in
equal measure.
In policy-making circles, generational accounting techniques and claims
about intergenerational justice have come to dominate on those occasions
when the age orientation of social policy regimes is under discussion. But
social policies are not static, and the distribution of resources among dif-
ferent age groups, not among different generations, is often at the heart
of political con¬‚ict over the welfare state. Hence analysis of the age orien-
tation of welfare states should ideally clarify the distribution of resources
across age groups, as well as across generations.
Some important work in this area has been undertaken. O™Higgins (1988)
offered a comparison of the treatment of elders and children in ten OECD
countries, with direct expenditure and some tax data for the period 1960 to
1985. But while this contribution was an important ¬rst step toward the goal
of measuring the age orientation of social policy, a restricted sample size
and highly aggregate spending data limited the analysis. Meyer and Moon
(1988) and Jencks and Torrey (1988) expanded the categories of analysis
beyond the con¬nes of social insurance spending but, as did O™Higgins
(1988) and Pampel (1994), compared the situations of only the elderly and
children, leaving out the middle ground of adulthood, where contemporary
welfare states have had such widely varying success in adjusting to changes in
employment and family patterns. More recently, Castles and Ferrera (1996)
usefully consider the age-distributive effects of the housing/pension policy
complex, but are hampered in the conclusions they can draw by the small
number of cases and the limited set of policies that they discuss. Despite
a growing interest in the relationship between demographic change and
social welfare systems, major lacunae remain in our understanding of how
social policies work to transfer resources across age groups and generations.

Measuring the Age Orientation of Social Policy
The remainder of this chapter is dedicated to evaluating in as much depth
and breadth as possible the distribution of public social policy resources
to different age groups in twenty highly industrialized democracies. Many
different kinds of public policies affect the distribution of public resources
to different age groups. Zoning regulations specifying minimum housing
Age in the Welfare State

lot sizes, state subsidies of credit markets, and policies designed to stimulate
the employment of youth or older job candidates are just a few examples. So
a truly comprehensive measure of how states distribute resources across age
groups would need to consider the totality of policy arenas and instruments
through which states might act to channel resources to different age groups.
This chapter focuses, more modestly, on the distribution of bene¬ts to
different age groups carried out through three key areas of public policy:
direct social expenditures on social insurance bene¬ts, education, and health
care; tax expenditures on welfare-substituting goods; and housing policies.
Only public spending and private spending that is mandated by law (e.g.,
occupational pension schemes in France) are included.
The age categories employed throughout this analysis are elderly and
non-elderly. These categories are rather ungainly as compared with seniors
and children, or labor market participants versus dependents. But they are
useful because public debates so often posit a trade-off between continuing
to support the elderly at a high level and devoting resources to other kinds
of needs in the non-elderly population. The de¬nition of the relevant age
groups is compelled as well by the considerable overlap between the well-
being of children and non-elderly adults, and the scant similarity between
the well-being of seniors and of their children™s and grandchildren™s age
groups. Cross-nationally, poverty rates among seniors, after taking into
account both taxes and social bene¬ts, are not highly correlated with the
same measure for either children (r = .67) or non-elderly adults (r = .59).
However, post-tax, post-transfer poverty rates for children and non-elderly
adults are quite highly correlated (r = .89), with the relationship particularly
strong where poverty is concentrated among families with large numbers
of children.3 Working-age adults and children experience similar risks of
poverty and receive similar degrees of protection from the welfare state,
while the elderly are often in a category all their own. Using elderly and non-
elderly as our basic age categories also responds to the practical demands
of working with social expenditure data. While in most countries social
bene¬ts are paid directly to elderly persons and not to their adult children,
transfers intended for children (e.g., child allowances, day care subsidies,
funds for school fees or books) are always given to the parents and are
considered part of the parent™s income, not the child™s.

3 Poverty rates are the percentage of individuals in each age group living in households with
size-adjusted disposable income (after taxes and transfers) below 50 percent of the country
median. Author™s calculations from Luxembourg Income Study data.

Measuring the Age of Welfare

Table 2.1 Public spending on the elderly, per
person aged 65+, as a percentage of GDP per capita

Australia 37.0
Japan 44.5
Portugal 45.0
Canada 45.2
United Kingdom 46.5
Ireland 46.8
United States 49.5
Norway 58.4
New Zealand 58.8
Spain 60.7
Denmark 61.5
Sweden 62.3
Belgium 63.8
Netherlands 65.6
Finland 70.2
Germany 70.7
Greece 74.0
France 84.1
Italy 90.4
Austria 91.4
Sources: Spending data from OECD 2004; GDP and
demographic data from OECD 2003b.

The Age Orientation of Direct Social Expenditures
The distribution of spending on social programs providing cash bene¬ts
and services for different age groups in the population is the most basic
measure of the age orientation of a country™s social policies, and also the
easiest measure to construct. For a few categories of direct expenditures
(notably, disability bene¬ts, health care, and housing) the age orientation
of spending is not immediately apparent. But a wide range of cash bene¬ts
and social services do have obvious age orientations. Spending on services
for the elderly and pensions for old age and survivors constitutes the core
of elderly-oriented social spending. Spending on unemployment bene¬ts
and active labor market policies, as well as occupational injury and sickness
programs, are the core bene¬ts aimed speci¬cally at working-age adults.
Family allowances, services for families (e.g., child care), and education are
the main items geared toward children.
Tables 2.1 through 2.6 illustrate the relative importance accorded to
elderly-oriented, worker-oriented, and child-oriented social spending in
Age in the Welfare State

Table 2.2 Public spending on unemployment
and active labor market policies, per registered
unemployed, as a percentage of GDP per capita

Sweden 170
Denmark 144
Netherlands 122
Finland 91.3
Belgium 88.7
Norway 82.6
Austria 77.9
Ireland 76.5
Germany 71.0
New Zealand 70.1
France 64.2
Canada 46.1
Portugal 45.5
Australia 43.6
Spain 38.9
Italy 37.5
United Kingdom 35.8
Japan 34.4
Greece 21.0
United States 20.9
Sources: Expenditure data from OECD 2004; GDP
and unemployment data from OECD 2003b.

Table 2.3 Public spending on occupational injury
and sickness programs per member of the civilian
labor force, as a percentage of GDP per capita

Sweden 4.31
New Zealand 3.98
Netherlands 3.79
Ireland 3.26
France 2.27
Germany 1.62
United States 0.65
United Kingdom 0.65
Japan 0.51
Sources: Expenditure data from OECD 2004; GDP
data from OECD 2003b; labor force data from
OECD, Labour Force Statistics.

Measuring the Age of Welfare

Table 2.4 Public spending on cash bene¬ts and
services for families, per person under 15, as a
percentage of GDP per capita

Sweden 22.9
Denmark 19.4
Finland 18.0
Austria 17.7
Norway 16.5
Belgium 13.0
France 12.9
Germany 12.0
United Kingdom 11.8
New Zealand 10.2
Australia 8.92
Netherlands 8.44
Greece 7.16
Ireland 6.47
Italy 5.46
Portugal 4.72
Canada 3.41
Japan 2.43
United States 2.43
Spain 1.64
Sources: Expenditure data from OECD 2004;
demographic and GDP data from OECD 2003b.

different countries. Total expenditures in each category are adjusted for the
size of the bene¬ciary pool and expressed as a percentage of the nation™s per
capita gross domestic product (GDP). These ¬gures are then averaged over
the period 1985“2000.4 Data aggregated across time of course obscure the
dynamics of welfare state change during a period when many regimes were
subject to reform and retrenchment. But period averages remain useful for
visualizing the basic parameters of spending on different age groups.
Per capita direct public expenditures on the elderly (Table 2.1), in the
form of services and pensions for retirees and survivors, vary from a low
of 38 percent of GDP per capita in Australia to a high of 92 percent in
Austria. Generally speaking, the public occupationalist pension systems of
Continental Europe spend the most per person, while the low-spending
welfare states of the English-speaking world and recent developers Japan

4 The OECD Social Expenditure Database (OECD 2004) contains only spotty information
on family policies and active labor market policies prior to 1985.

Age in the Welfare State

Table 2.5 Public education spending, per person
aged 0“20, as a percentage of GDP per capita
(average 1985“2000)

Denmark 33.4
Sweden 31.2

. 1
( 7)