. 3
( 7)


that substantially mix occupationalist and citizenship-based programs have
intermediate age orientations as well.

1 Information on bene¬ts structures from MISSOC 1995; Palme 1990, 77; and Wennemo
1994, 99.

Theories and Hypotheses

Program Structure
Mixed Occupational
Most Elderly-Oriented

Age Orientation

Least Elderly-Oriented


Figure 3.3 Age orientation and welfare state program structure.

The First Critical Juncture
How can we understand the development of this connection between occu-
pationalism and pro-elderly spending, and vice versa? The late nineteenth
and early twentieth centuries marked the beginning of a rapid phase of
welfare state development in the countries of Europe, North America, and
the British Commonwealth. There were two fundamentally different ways
that welfare states grew up: along either citizenship or occupational lines.
Accounting for the success of different organizational models in different
countries is a task that has been undertaken elsewhere (see, esp., Baldwin
1990; Ferrera 1993; Manow 1997). Regardless of the precise process that is
thought to generate it, these analyses highlight the same outcome: different
kinds of welfare state programs that cover very different kinds of people.
We can think about the different populations covered by occupationalist

Age in the Welfare State

and citizenship-based regimes in terms of the distinction between labor
market “insiders” and labor market “outsiders.”
In citizenship-based systems, the core of the welfare state is made up of
programs designed to complement, rather than replace, bene¬ts provided
by mutual associations for their members (policies such as old-age and
invalidity pensions, unemployment insurance, or health insurance; Manow
1997). As a result, state welfare programs in citizenship-based regimes cover
labor market outsiders: people with weak ties to unions and mutual organi-
zations, such as children, abandoned mothers, or the indigent. For exam-
ple, in the United Kingdom in 1910, outdoor relief (income maintenance
programs for the poor of all ages) and noncontributory old-age pensions
accounted for 84 percent of all public social welfare spending, while social
insurance programs such as those for occupational accidents and disease
accounted for only 16 percent, and there were no public provisions for
unemployment insurance or occupational pensions (data from Ritter 1983).
These work-related programs were instead provided by nonstate actors,
namely, Britain™s friendly societies.
By contrast, in occupationalist regimes the state takes over from mutu-
alist organizations the job of protecting people with tight links to unions
and the labor market: people with long-term, stable employment in insured
sectors of the economy, as well as retired workers, people in between jobs,
and people with job-related health problems. In other words, state welfare
provision focuses on social insurance programs for labor market insiders. In
Germany in 1910, for example, 52 percent of the labor force was enrolled
in public occupational pensions, 51 percent in public occupationally linked
health insurance, and 87 percent in public occupational injury insurance
(data from Flora and Alber 1981). But basic social assistance and poverty
alleviation were relegated to religious charities, municipalities, and, above
all, families. The task of caring for labor market outsiders in occupationalist
regimes falls to nonstate actors.
If citizenship-based regimes originated by providing state protection for
labor market outsiders, and occupational regimes for labor market insiders,
how does this affect the age orientation of these different types of welfare
states today? To answer this question, we need to understand how the age
composition of these insider and outsider groups changes as welfare states
and private insurance markets matured over the course of the twentieth
In a schematic citizenship-based regime, bene¬ciaries of state-provided
welfare programs are primarily labor market outsiders: people who do not
Theories and Hypotheses

receive bene¬ts through union schemes, friendly societies, and the like.
Because the supply of private old-age insurance was quite limited in the
early twentieth century, most former workers fall into this outsider category
and rely on the state (and their families) to protect them from poverty in old
age. Thus, the majority of the labor market outsiders covered by the public
welfare programs in citizenship-based regimes are elderly people who have
ceased working. As a result, in the early twentieth century citizenship-based
regimes tended to be quite elderly-oriented.
As markets for occupational pension insurance matured, however, more
and more of the elderly became covered under employment-based pen-
sions. One of the things that advanced industrial capitalism does rather well
is look after retired workers “ perhaps, as Myles (1989) argues, because it is
in the interest of employers to move older people out of manufacturing jobs
to make way for more productive younger workers. In any event, as more
and more elderly people gain access to private employment-based pensions,
fewer and fewer of them are left as outsiders to be cared for by the state.
Once most of the elderly poor are removed from the outsider group, though,
citizenship-based regimes become much more youth-oriented. The pool of
labor market outsiders who constitute these regimes™ core clients becomes
dominated by young people: children, the long-term unemployed, single-
parent families, and the like. As a result, by the latter part of the twentieth
century, citizenship-based welfare regimes were quite youth-oriented.
The reverse transformation occurs in occupationalist welfare regimes.
Unlike in citizenship-based regimes, these countries began the twentieth
century with fairly youth-oriented welfare states. Social programs covered
labor market insiders, leaving outsiders to be cared for by their families and
by private and/or religious charities. But in the early twentieth century, as
noted above, there were relatively few old people in the insider category
that receives state protection, since employment-related old-age pensions
were not yet fully developed. As employment-related pensions “ which in
occupationalist welfare states are publicly provided “ grew, retired people
became insiders. This means that occupationalist systems grew increasingly
elderly-oriented over the course of the twentieth century as pension sys-
tems matured. In addition, the aging of the protected core work force has
meant that in occupationalist systems, even public programs such as unem-
ployment insurance or disability pensions that should bene¬t working-age
people tend to be skewed toward older workers.
Occupationalist programs lead to more elderly spending, and
citizenship-based programs lead to more spending on young people, despite
Age in the Welfare State

the fact that in the early twentieth century these different types of systems
had opposite age orientations. This is because as pension systems and labor
markets develop and mature, the age structure of labor market insider and
outsider groups changes. In interaction with dynamic changes in markets
for labor and insurance, static welfare state institutions create a pattern of
social policy spending that matures over time to result in the age orienta-
tions we observed in the 1990s.

The Second Critical Juncture
But how static are these welfare state institutions, really? Welfare state insti-
tutions are often characterized by policy feedback mechanisms that make
them rather sticky (Weaver 1986; Pierson 1994). But they can and do change
under certain circumstances. We can think about the ¬rst critical juncture,
the initial choice between occupationalist and citizenship-based regimes in
the early twentieth century, as setting countries off on one of two tracks
of welfare state development. But an opportunity to switch tracks occurred
in the period around the Great Depression and World War II. The 1930s
and 1940s were a time of great institutional ¬‚uidity, when many advanced
industrialized countries had an opportunity to re-evaluate and rebuild their
welfare programs after the disasters of the 1920s and 1930s. Wartime con-
ditions aggravated social problems, while in many countries a drive for
national unity fostered during World War II contributed to a new push
for national social programs. In addition, both public and private insur-
ance programs had been bankrupted throughout much of Europe because
of runaway in¬‚ation and wartime destruction of property. This presented
occupationalist welfare states, in particular, with a prime opportunity to
experiment with new forms of social protection.
In most European countries with occupational welfare states, govern-
ments commissioned of¬cial studies to investigate the feasibility of intro-
ducing universal, citizenship-based welfare programs along the lines of the
Beveridge Plan in Britain. Such inquiries occurred in France, Belgium, the
Netherlands, Austria, Germany, and Italy in the period between 1945 and
1948 (Ferrera 1993). Under the in¬‚uence of the International Labour Orga-
nization (ILO) and policy lessons diffusing from Britain, some countries that
before World War II had had purely occupational welfare systems adopted
citizenship-based programs, beginning the process of switching tracks. But
an equal number of occupationalist regimes stayed the course, despite the
conclusions of government advisory panels that encouraged the adoption
Theories and Hypotheses

of British-style universal bene¬ts. How can we explain this divergence in
national trajectories after World War II, a divergence that, as we have seen,
would have profound consequences for the age pro¬le of welfare regimes?

Programmatic versus Particularistic Political Competition
The opposing slopes of this second watershed, the division between coun-
tries that stayed on the occupationalist track and those that began the
switch toward the citizenship-based track, are characterized by different
modes of political competition. The switchers were all countries where
programmatic political competition prevailed, while the countries that did
not adopt citizenship-based welfare programs in the 1930s through 1960s
shared a particularistic mode of political competition that inhibited the
development of substantial universal welfare programs. Clientelist poli-
tics and occupationalist welfare programs reinforced each other through a
host of mechanisms, as we shall see. But ¬rst, let me clarify the distinction
between the two modes of political competition.
The mode of political competition varies along a continuum ranging
from programmatic to particularistic. Programmatic competition occurs when
politicians and parties vie for votes by promising to enact policies that they
argue will bene¬t society at large. This type of political competition is char-
acterized by the relatively low degree of selectivity of the bene¬ciary groups
(e.g., entire classes, rather than particular industries, neighborhoods, or eth-
nic groups). To the extent that policies are designed to bene¬t somewhat
selective groups (e.g., the working class) rather than the public at large,
they are justi¬ed with reference to coherent political ideologies. A vari-
ety of different labels “ responsible party government, universalism “ have
been attached to this phenomenon. In Shefter™s seminal work (1994, chap-
ter 2), policy orientation is the polar opposite of patronage orientation.
I reject this label because patronage-oriented politicians are no less con-
cerned with policies; they simply care about policies for distributive rather
than programmatic reasons.
At the other end of the spectrum of competitive strategies lie partic-
ularistic political practices ranging from log rolling, constituency service,
and intensive interest group involvement in policy making to an out-and-
out exchange of bene¬ts for votes. Particularism occurs when politicians
offer tangible bene¬ts to selective groups of voters in return for their votes.
Alternative labels for this phenomenon include clientelism and patronage
(and sometimes patron-clientelism). There are subtle distinctions between
Age in the Welfare State

patronage and clientelism, usefully discussed in Piattoni (2001). I use the
terms particularism and clientelism interchangeably to denote behaviors
that meet the de¬nition offered above. Particularistic politics may or may
not be justi¬ed rhetorically with reference to political ideologies or the
common good. A politician operating in this environment might offer, for
example, to introduce favorable public pension legislation affecting workers
in a single industry, in the expectation that the bene¬ciaries of the proposed
policies would reward the politician with their votes.
Measuring the mode of political competition in a polity presents a num-
ber of challenges. Both programmatic and particularistic modes of political
exchange operate in all polities, so what we are really trying to measure is the
rough balance of the two. But determining the nature of political appeals is
not always straightforward. Piattoni (2001, 6 n. 9) notes, “The most striking
feature of mass clientelism is that, in an effort to truly reach the masses, it
often works through fairly impersonal means, such as the passage of laws
or implementation of measures that favor entire categories of persons.”
The wide variety of practices included under the umbrella term make it
dif¬cult to establish the extent of particularism in any comprehensive way.
And the opprobrium with which many of these practices are viewed makes it
dif¬cult to obtain reliable information about the extent to which they occur
in any given polity. Still, there is wide agreement about the degree to which
political life is dominated by particularistic versus programmatic parties and
politicians in the different countries of the OECD. Specialists most often
classify Austria, Belgium, Greece, Italy, Japan, Spain, and the United States
as possessing distinctively particularistic styles of politics and policy making.
France and Portugal occupy a middle ground, while political competition
in the remaining countries of Northwestern Europe, Canada, and the
Antipodes is primarily programmatic.2 Corruption rankings such as those
summarized for the World Bank in Kaufmann, Kraay, and Mastruzzi (2003)
can be combined with other measures of clientelism to triangulate in a
situation of imperfect measurement. Corruption rankings generally concur
with the intensity of patronage politics noted in case studies. Greece, Italy,
Japan, and Belgium suffer from both high levels of corruption and strongly
particularistic politics, whereas Austria, the United States, and Spain score
somewhat lower on both counts. Some scales report some problems with
corruption in Portugal and France, while the remaining countries of

2 See, e.g., Lyrintzis 1984; Cazorla 1992; Shefter 1994; Cotta 2000; Kitschelt 2000; Garc´a
and Karakatsanis 2001; and Hopkin 2001.

Theories and Hypotheses

Northwestern Europe, Canada, Australia, and New Zealand are apparently
corruption-free zones. The World Bank corruption ranking index does
not measure the same precise phenomenon as do the case-based qualitative
characterizations of the mode of political competition, but the consonance
of the two measures of closely related phenomena should give us some con-
¬dence that the characterizations drawn from the case studies are reliable.
Let us now return to the link between occupationalism and particularism.
That such a link exists is clear: each of the countries in the right-hand col-
umn of Figure 3.3, that is, the countries with occupationalist welfare states,
also appear on our list of predominantly clientelist polities, and none of the
countries in which particularistic politics is the norm has anything other
than an occupationalist welfare state. There are several reasons why this
correspondence occurs. At the most intuitively basic level, occupational
programs provide crucial resources for particularistic politicians. Occupa-
tionally based social insurance programs plainly lend themselves far more
than do universal programs to the kind of ¬ne-grained targeting of incen-
tives on which particularistic political competition thrives (Skocpol 1992;
Shefter 1994). Clientelist politicians thus tend to expand welfare policies in
ways that enhance their fragmentation and to oppose proposals that would
harmonize existing programs.
The complexity of occupational programs vis-` -vis a single, uniform
universal bene¬t also redounds to the bene¬t of politicians who engage in
particularistic targeting of bene¬ts for votes. Multiple, fragmented bene¬t
schemes geared toward different kinds of workers and their dependents
generate compartmentalized sets of winners and losers. The costs for society
at large of any individual deal with a small slice of the electorate are very hard
to see. The aggregate effect of a profusion of micro-legislation targeting
bene¬ts toward different groups can confound even professional budgetary
analysts, making it dif¬cult to reach consensus on the “real” impact of social
policy or of proposed reforms.
The reverse is true for universalistic systems. When a raise for one is
a raise for all, programmatically oriented politicians may gain by offering
bene¬t increases to broad swathes of the population, but particularistic
politicians, who live by targeting, bene¬t little. Uniform bene¬ts are also
transparent, meaning that both bene¬ts and costs are highly visible. Cost
control is easier to obtain in the big-ticket areas, such as pensions, because
even defenders of generous bene¬ts must confront spending projections
that are easy to calculate and generally agreed on. Universal systems offer
neither the targeting nor the opacity that appeal to particularistic politicians.
Age in the Welfare State

At a deeper, structural level, particularism and occupational programs
are linked because particularistic politicians are unlikely to support the
development of neutral state capacities. Neutral state capacities clearly
undermine the politics of selective bene¬ts, but without them, universal
social policies are nearly impossible to enact. Tax systems are a prime exam-
ple. Clientelist administration of tax systems often results in ineffective tax-
ation, or outright exemption from taxation, of the self-employed in indus-
try, services, and agriculture. In the absence of effective taxation of these
groups, however, social programs must be ¬nanced by industrial workers
and public employees. This makes it harder, both politically and from a ¬scal
standpoint, to expand occupational social programs to cover new groups of
bene¬ciaries among the self-employed. Where the self-employed are taxed
effectively, on the other hand, it is possible to extend occupational pro-
grams to cover the entire population without invoking the ire of industrial
workers and the Left. Neutral labor exchanges and nonpreferential adminis-
tration of social bene¬ts, too, are unlikely to be developed by particularistic
politicians. In their absence, nonclientelist politicans in the system refrain
from supporting nominally universalist social programs that are likely to be
administered selectively.
The shift in political preferences that occurs in the context of particu-
laristic competition is the most subtle, but perhaps the most important, of
the mechanisms linking such competition to occupational, and ultimately
elderly-oriented, social programs. As we have seen, particularistic political
competition tends to undercut support for universal social bene¬ts even
among those politicians and parties who would be ideologically inclined to support
them. In particular, ¬nancing general social bene¬ts through payroll
taxes often strains the solidaristic impulses of the Left, while a state
administration that is colonized by a clientelist opposition may lead the
Left to prefer occupational programs where unions retain some control
over administration.
One might well ask whether clientelism and occupational social poli-
cies could be jointly caused by some third factor. A large small-¬rm or
small-farm sector, weakness of the Left, fears of Communist takeover in
the postwar period leading to tolerance for dirty politics, or a societal pref-
erence for collectivism versus individualism in both politics and policy are
all more or less plausible candidates. Each seems vulnerable to criticism
that it does not ¬t all country cases (e.g., there are many small ¬rms and
small farms in the Netherlands, but that country has had a pristine and very
effective tax administration for centuries) or that it might be contradicted
Theories and Hypotheses

by another equally plausible explanation (should the Left be weak or strong
to encourage clientelism and/or occupationalism?).
Another study linking the large literatures on clientelism and the origins
of solidaristic universal social policies would surely be welcome, but this is
not that book. I stop by noting that in countries where particularistic politi-
cal competition prevails, dominant politicians have an interest in preserving
or extending occupational fragmentation; that in these countries the basic
infrastructure of government necessary for constructing universalistic wel-
fare programs tends to be underdeveloped; and that political support for
universal programs is further undermined by clientelist administration of
taxes, labor exchanges, and social bene¬ts. Under such circumstances, it
hardly seems necessary to ask where clientelism comes from; what matters
is that clientelism and occupationalism reinforce each other so strongly,
and in so many ways, during the period after World War II when the new
global consensus was for universalism. In the presence of particularistic
political competition, countries that entered into the second critical junc-
ture with occupationalist social programs are thus unlikely to introduce the
new citizenship-based programs that would allow their welfare states to
develop in a more youth-oriented direction.
To sum up, the two great divides in welfare state formation represented
by our two critical junctures condition the eventual age orientation of social
policies. The ¬rst divide, the split in the early twentieth century between
occupational and citizenship-based regimes, creates divergent age orienta-
tions as welfare programs mature: occupationalist regimes become elderly-
oriented, while citizenship-based ones are more neutral with respect to age.
The second great divide occurs around World War II. Some occupation-
alist countries reduce the elderly orientation of their welfare spending by
becoming mixed systems, adding universal programs to their base of occu-
pational programs. Others remain pure occupational systems and continue
to develop highly elderly-oriented spending patterns. This second divide is
both facilitated and reinforced by the mode of political competition, pro-
grammatic or particularistic, that prevails in these countries.

The next three chapters examine the development of family allowance,
unemployment, and old-age pension bene¬ts in Italy and the Netherlands
through the early 1990s. These Dutch and Italian case studies ¬‚esh out
the mechanisms behind the main claim of this book, that the dominant
Age in the Welfare State

mode of political competition is crucial for the eventual age orientation of
social spending. The key conceptual link is the recognition that the mode
of political competition reinforces choices about the structure of welfare
programs. Joining this intuition to our understanding of how different pro-
gram structures mature over time to produce different age orientations, it
becomes possible to visualize, as in Figure 3.2, a tree-like set of branching
pathways by which speci¬c age orientations in social policy emerge.
Italy and the Netherlands both lie on the right-hand path in Figure 3.2,
indicating that both had occupational welfare states prior to World War
II but followed different trajectories toward their respective age orienta-
tions in the postwar period. Of necessity, then, the case material in this
book focuses on how universalism and a moderate age orientation may or
may not develop during this later period, not how they persist on the left-
hand pathway, among those countries that had developed non-occupational,
citizenship-based welfare states already in the early twentieth century. The
medium-N analysis presented in this chapter helps to make the case that
these results generalize to both main branches of the tree.3 Restricting the
case studies in this book to Italy and the Netherlands helps us to hone in on
the precise mechanisms that allow for institutional change or persistence
and that are so important for the story of how age orientation develops.
Italy and the Netherlands are an appealing, though little-used, pair of
country cases for the analysis of welfare states. Esping-Andersen (1990)
classi¬es both as Conservative-Corporatist welfare regimes in structure,
even though the Netherlands™ very generous unemployment bene¬ts gen-
erated an almost Social Democratic decommodi¬cation score by 1985.
More gendered analyses of social policy would see further similarities in
both welfare states™ male-breadwinner orientation, which re¬‚ects a shared
origin in Red-Roman (Social Democratic“Christian Democratic) collabo-
ration in the construction of the welfare state through the 1970s. On more
social-structural dimensions, too, Italy and the Netherlands share impor-
tant similarities. Small businesses and the agricultural sector have long been
mainstays of both economies. And while in neither country was the labor
movement moribund, in neither did it come to full force. Unionization
rates are moderate in both countries, and both Dutch and Italian policy
concertation has been episodic during the period since World War II.

3 Further statistical analysis of both cross-sectional and time-series relationships among the
potential causal variables outlined here would also be useful, though care must be taken to
model the path-dependent aspects of these relationships correctly.

Theories and Hypotheses

Yet social policies in Italy and the Netherlands had very different age pro-
¬les by the 1990s. Chapters 4 through 6 explore the institutional features
and the political battles that lie behind the persistence of an elderly social
policy orientation in occupationalist Italy, and the growth of more youth-
oriented policies in the universalist welfare state that developed in the
Netherlands after World War II. Four distinct mechanisms of persistence
emerge. First, the distinction between occupational and citizenship-based
welfare programs alters the costs and bene¬ts to politicians of expanding
programs in different ways. Second, program structure affects the salience
to the public of different types of bene¬ts, and thus the ability of polit-
ical actors to mobilize voters around the expansion of particular welfare
state programs. Third, the way that social programs are structured affects
the degree of transparency surrounding political decisions about spend-
ing, which rewards particularistic and programmatic politicians unevenly.
Finally, the ability of clientelist politicians to use the welfare state for patron-
age affects the preferences of other political actors in ways that lead to the
preservation of occupationalist systems where political competition is pri-
marily particularistic. The effects of program structure and the mode of
political competition work on youth-oriented and elderly-oriented pro-
grams in different ways, resulting in a tight bundling of elderly-oriented
welfare spending with occupationalist regimes and particularistic politics,
and of youth orientation with citizenship-based regimes.


Family Allowances

This chapter traces the development of spending on family allowances
in Italy and the Netherlands from the end of the Second World War
through the early 1990s. Family allowance spending grew dramatically
in the Netherlands during the postwar period, contributing to its youth-
oriented social policy regime, while in Italy the opposite occurred. A focus
of this chapter is the strategic behavior of politicians working within polit-
ical parties, behavior that interacts with the structure of family allowance
programs in Italy and the Netherlands to determine spending outcomes.
The way that family allowance programs are structured “ along universalist
lines in the Netherlands and occupationally based in Italy “ is in turn an
outgrowth of the competitive strategies of politicians. At the same time,
the structure of family allowance programs sets the parameters for future
growth of bene¬ts by altering both politicians™ and potential constituencies™
perceptions of the bene¬ts to be gained by either increasing bene¬t levels
across the board or extending family allowances to new constituencies in a
piecemeal fashion.
Family allowances are an important indicator of the age orientation of
social policies because they are usually the largest public expenditure item
for families with children, even in countries where the state provides things
such as day care and other services for families (Gauthier 1996). Other kinds
of bene¬ts for families (day care, parental leave, health care, educational
subsidies) could in principle siphon resources away from family allowances.
But the most recent OECD social expenditure data do not reveal a zero-sum
relationship between spending on cash bene¬ts and services for families
(OECD 2004). For example, France and Denmark in 2001 both spent the
equivalent of 1.5 percent of their GDP on cash bene¬ts for families “ near
the OECD average of 1.4 percent. But France spent the least of any OECD
Family Allowances

country on services for families (0.004 percent of GDP), while Denmark
spent the most (2.3 percent of GDP).
Family allowances are an especially good indicator of policy attention to
children, because they involve direct expenditures that are politically salient.
As Pampel and Adams (1992, 527) argue, family allowances are “a direct
and overt expression of a nation™s institutional commitment to families with
children.” The fact that allowances are direct cash transfers, rather than tax
expenditures, makes them “politically subject to close legislative scrutiny,”
and thus a good “measure of the depth and effectiveness of political support
for children and their parents.” So while family allowances certainly are
not the only kind of social welfare spending directed toward families with
children, they are probably the best single indicator of the politics of support
for families with children.
It may come as a surprise to those familiar with contemporary wel-
fare state classi¬cations that “residualist” Italy (Titmuss 1974) began the
post“World War II period with one of the most generous family allowance
programs in the developed world. Family allowances in Italy expanded to
cover an increasing share of the total population until the mid-1970s; but
aggregate and per child spending on families began to decline dramatically
starting in the mid-1960s. Family spending in the Netherlands shows an
opposite trend. In the decades immediately following the Second World
War, spending on family allowances in the Netherlands was quite low, but
in the early 1960s family allowances became a universal social insurance
bene¬t and both aggregate and per child spending on families began to
rise. The 1970s to mid-1980s marked an important expansion of family
spending in the Netherlands on both a per capita and an aggregate basis,
even as other programs such as unemployment insurance, disability, and
old-age pensions succumbed to austerity measures (see Fig. 4.1).1
What drives the expansion of family allowances in the Netherlands, and
their spectacular contraction in Italy, during the post“World War II period?
This chapter argues that these dynamics are best understood as a product
of the interaction between the structure of family allowance programs,
on the one hand, and the behavior of politicians, who use these programs
in order to build constituencies and win or maintain of¬ce, on the other.

1 While this chapter is about family allowances, the spending in Figure 4.1 re¬‚ects other
expenditures for families as well. The cross-nationally comparable OECD social expenditure
data on which this ¬gure is based do not allow us to track family allowances and other family
spending separately prior to 1980.








1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989

Itlay Netherlands

Figure 4.1 Public spending on families, per person aged 0“14, as a percentage of GDP per capita, Italy and the Netherlands.
Sources: Expenditure data from Varley 1986; OECD 2004. GDP and demographic data from OECD 2003b. Data de¬nitions
differ between the two spending series. The ¬gures from Varley 1986 are adjusted by the average difference between the two
series for 1980“5, when both measures are available.
Family Allowances

The strategic behavior of politicians both determines and is determined
by the structure of family allowance programs. The structure of family
allowance programs affects the behavior of politicians, but it also affects the
preferences of potential constituencies, who may choose to place a higher
or lower priority on family allowance bene¬ts versus other kinds of public
policies depending on the characteristics of these bene¬ts. This inter-
twining of program structure, politicians™ behavior, and the preferences of
potential welfare state constituencies explains divergent spending on family
allowances in Italy and the Netherlands. In so doing, it also helps to explain
the age orientation of the social policy regimes in these two countries.

Explanations for the Level of Spending on Family Allowances

Partisan Effects on Family Allowance Spending
Scholarship on the welfare state has long recognized the importance of
partisan effects on social welfare spending. These effects are hypothesized
to occur either because class-based parties translate the power resources of
particular classes into policy preferences or because parties are themselves
carriers of ideologies and values that affect the policy preferences of legisla-
tors and government of¬cials. It is not surprising, then, that many compar-
ative studies of family allowance policies have posited that the strength of
partisan actors, particularly Social Democratic and Christian Democratic
parties and politicians, is a key determinant of policy outcomes. Evidence
for this proposition is mixed, however.
Wennemo™s (1994) study of cash and tax allowances for families in eigh-
teen OECD countries during the period 1947 through 1985, for example,
¬nds that both left and Christian Democratic power in government make
a difference for the level of family bene¬ts. Both left and Christian Demo-
cratic parties are associated with levels of spending on family allowances
higher than those observed in countries where Liberal or Conservative
political forces have dominated. The political effects she observes may be
moderated by excluded variables, though, as Wilensky (1990) and Pampel
and Adams (1992) in fact ¬nd. Demographic variables such as female labor
force participation and the percentage of elderly in the population have
been found to be important predictors of family policy outcomes (Wilensky
1990; Pampel and Adams 1992), as have political variables such as the in¬‚u-
ence of pressure groups including women™s, family, elderly, technocratic,
and employer lobbies (Wilensky 1975; Aldous, Dumon, and Johnson 1980;
Age in the Welfare State

Wilensky 1990; Skocpol 1992; Pedersen 1993; Pierson 1994). Still other
research points to the importance of expert opinion (Gauthier 1993), party
system competitiveness (van Kersbergen 1995), and institutional variables
such as neo-corporatist concertation (Wilensky 1990; Pampel and Adams
1992) or the degree of government centralization (Koven and Michel 1990).
Leaving aside some of these more esoteric considerations, we can still
seek independent con¬rmation of Wennemo™s hypotheses that partisan pol-
itics and demographic constituency groups are important predictors of fam-
ily allowance spending. Pampel and Adams (1992), in an eighteen-country
study spanning the period 1950“80, do ¬nd some support for the notion
that partisan politics matters. They do not evaluate directly the effects of
Christian Democratic Party strength on family allowance spending but do
¬nd a substantial effect on the level of family allowance spending relative to
other social spending produced by the percentage of the population that is
Catholic. Left power resources “ left party dominance of government and,
more particularly, corporatist institutions “ also seem to have an important
impact on family allowance spending.
Van Kersbergen (1995) also considers the effect of both Christian Demo-
cratic and Social Democratic political forces on family policies. Despite his
clearly articulated thesis that Christian Democratic governance results in
stronger bene¬ts for families with children, van Kersbergen™s evidence for
an independent effect of either Christian Democracy or Social Democracy
on family allowances per se is rather weak. Net replacement rates of major
income maintenance schemes for families with children versus single indi-
viduals show no systematic differences between Christian Democratic and
other regimes. Van Kersbergen ¬nds that during the 1960s the combined
effects of wages and taxes do seem to result in higher disposable incomes
for families with children living in Christian Democratic countries, but
the results do not hold for the 1980s. Cash transfers as a percentage of
the gross wage (at average production worker levels) for the (brief) period
1972“6 again offer partial support for van Kersbergen™s thesis, but Sweden,
New Zealand, and Norway cluster with the Christian Democratic countries
on this measure. Neither do differences in post-tax, post-transfer income as
a percentage of gross wages for different family types correspond to differ-
ent political regimes: highly Christian Democratic Italy falls in with Japan
and the United States in offering the least amount of additional support for
families with children.
There is evidence that the working-class base and the egalitarian, redis-
tributive thrust of Social Democratic political parties results in higher (and
Family Allowances

more egalitarian) welfare state spending in general. Likewise, there seems
little doubt that cross-class Christian Democratic parties, with their ide-
ological heritage of Christian social doctrine, also contribute to large but
less egalitarian welfare states. However, the evidence for partisan effects on
family allowance spending per se is quite mixed.

Demographic Constituency Arguments
Might focusing on demographically de¬ned constituency groups provide
a useful corrective to standard “power resources” or partisanship-based
hypotheses about the sources of variation in family policies? Wennemo
(1994) theorizes that while family allowance bene¬t levels tend to decrease
over time when parties other than left or Christian Democratic are dom-
inant, in general family allowance programs are relatively resilient to
roll-back because they affect such broad constituencies. In other words,
Wennemo ¬nds that partisan politics explains the structure and level of
family allowances, while an electoral argument based on the large size of
the natural constituency of families with children helps to account for the
fate of family allowance programs over time.
Similarly, the papers collected in Aldous et al. (1980) emphasize the
importance of lobby organizations for large families for the development
of family policies. While the authors of these papers ¬nd that even very
numerous and well-organized family lobbies probably cannot affect policy
on their own (they argue that what labor and especially employers want is far
more important), family organizations do serve an important role in agenda-
setting. Pedersen (1993), too, ¬nds that employers, not the pronatalist and
family lobbies in France, were ultimately responsible for the passage of
generous family allowance policies. But she stresses that the ¬t between the
ideological stances of family lobbies and pronatalist organizations, on the
one hand, and the needs of employer groups, on the other, was crucial for
promoting generous family allowance policies. Family lobbies cannot be
held responsible for developments in family policies, but neither can they
be ignored.
Perhaps more surprisingly, several studies have found support for the
hypothesis that large numbers of elderly voters increase the level of
spending on families with children. Wilensky (1990) explains this phe-
nomenon with resort to the assertion that older people, having been through
child-rearing themselves, understand how dif¬cult it is and thus support
public policies that make life easier for parents with children. Pampel and
Age in the Welfare State

Adams (1992) envision a different motivation for the “altruistic” spending
effects of large numbers of elderly voters. While normal pluralist assump-
tions would link larger elderly populations to a stronger emphasis on
pension spending, Pampel and Adams hypothesize that shrinking youth
cohorts may prompt elderly voters (and policy makers) to recognize the
interdependency of generations: without adequate supports for child rear-
ing, younger cohorts will fail to produce the number of new labor market
entrants that is required to maintain solvent social security systems for the
Demographically de¬ned voting blocs and/or lobbies “ groups of families
with children and/or the elderly “ are thus possible rivals to partisan actors
in determining levels of spending on family bene¬ts. And while, as Pierson
(1994) predicts, the impact of such groups may be more pronounced during
periods of attempted retrenchment, plainly there is also a case to be made
that demographically based lobbies have an impact, sometimes indirect, on
the construction of family allowance programs.

An Alternative View: Program Structure and Political Competition
Determine Family Allowance Spending
Both partisanship-based and demographically based hypotheses about the
development of family allowance policies focus their attention on forces
largely exogenous to the welfare state: political parties (and their class
and/or ideological bases), and political pressure groups (and the demo-
graphic trends that create them). While these external forces are surely
important, one important lesson from recent literature on welfare state
retrenchment is that processes endogenous to the welfare state can have an
important impact on future developments.2

2 Even during periods of welfare state construction and consolidation, such endogenous forces
may limit the potential for family policies to expand. For example, developments in other
policy areas with large ¬scal demands (e.g., old-age pensions) may constrain the growth of
family allowances. Conversely, granting family allowances to groups of bene¬ciaries de¬ned
by their very status within the welfare state (e.g., pensioners or people receiving unemploy-
ment bene¬ts) could increase the potential for growth in family spending. So the develop-
ment of family policies may be subject to constituency effects that arise out of the welfare
state programs themselves, rather than out of any objective demographic trends. To cite
another example, the family lobby, one imagines, would be far more politically important
where there were family bene¬ts to lobby for, an administrative structure to talk to, and
perhaps even organizational resources to be gained from the system of family allowances
itself, as in France and Belgium.

Family Allowances

One such endogenous determinant of the development of family
allowance spending is the administrative structure of the programs. Univer-
salist programs allow for a different pathway of development than do occu-
pationally based programs. The former can be expanded only by increasing
bene¬t levels, either across the board or by raising the income or wealth
limits imposed by means tests. The latter expand in a patchwork fashion,
increasing the number of bene¬ciaries independently of the level of bene¬ts.
These structural features of family allowance programs condition politi-
cians™ and constituencies™ preferences regarding family allowance bene¬ts.
In interaction with the competitive strategies of politicians, different pro-
gram structures result in very different spending patterns.
When family allowances are universal bene¬ts, they become less useful
to politicians who might wish to use them to make particularistic appeals
to speci¬c groups of voters. Thus, where political competition is partic-
ularistic rather than programmatic, family allowance programs will tend
to shrink in relevance once they are universalized. Where political com-
petition is programmatic, universal family allowances are likely to grow,
particularly if family allowances have become delinked from their origi-
nal meaning as wage supplements and reconceptualized as entitlements of
At the same time, the patchwork expansion of bene¬ts that occurs in
occupational systems means that for particularistic politicians, the level
of family bene¬ts is less important than the number of new bene¬ciaries.
So in occupational systems, the numerical expansion of family allowance
bene¬ciaries may actually drive down both bene¬t levels and aggregate
spending. The continuing linkage in occupational systems between wages
and bene¬ts reinforces this downward trend in the level of bene¬ts by
perpetuating the view of family allowances as a supplement to wages, rather
than an independent entitlement. As long as wages are rising, most potential
constituencies of family allowance programs will perceive these payments
as unimportant relative to either wages or other bene¬ts (such as pensions)
that provide closer to full income replacement.
The cross-national comparative literature on family policy agrees that
family allowance spending is likely to be highest when there are left-
ist and/or Christian Democratic parties in power (Wilensky 1990; van
Kersbergen 1995) and when there is concertation between the social part-
ners in conjunction with a high percentage of elderly people in the pop-
ulation (Wilensky 1990; Pampel and Adams 1992). But while both Italy
and the Netherlands after World War II experienced almost uninterrupted
Age in the Welfare State

participation of Christian Democratic parties in government and a sig-
ni¬cant policy-making role for left parties, they ultimately achieved very
different levels of public support for families.
While both Italy and the Netherlands have undergone periods of intense
concertation, neither growth in expenditures on family allowances nor the
implementation of policies that would promote such growth coincide with
the peaks in intensity of neo-corporatist policy concertation in either coun-
try. Italy has a wealth of older voters, and in the mid-1970s a minor recovery
in the level of family allowance spending was in fact driven by pensioners
eager to enjoy the dependent spouse allowance. But the Netherlands, which
already by the early 1970s spent more on family allowances than did Italy,
has one of the younger populations in Europe. Neither the partisan poli-
tics prediction nor the demographic constituency prediction accounts for
the reversal of fortunes in Dutch and Italian family allowance policies over
the course of the post“World War II period. A closer look at the cases is
needed to account for divergent trajectories in the level of family allowance

Family Allowances in the Netherlands
The ¬rst comprehensive legislation on family allowances in the Netherlands
appeared in 1962, when the General Family Allowances Act (Algemene
Kinderbijslagwet) guaranteed child bene¬ts for all residents of the Nether-
lands starting from the third child. Family allowances were fully indexed to
the cost of living in 1964. In 1980, separate legislation for the ¬rst and
second children of public servants, wage earners, and low-income self-
employed was merged with the laws governing family allowances for the
general population. This rati¬ed formally what had existed already in prac-
tice since 1962, a uniform family allowance bene¬t available to all residents
of the Netherlands.
To understand why Dutch family allowances got off to a slow start rela-
tive to family allowances in Italy, it is useful to understand why the family
allowances system was among the last social programs in the Netherlands to
become universalized. This would be easy to understand if family allowances
were, as is sometimes assumed, primarily pronatalist measures. The high
postwar birth rate in the Netherlands would then explain a lack of interest
in family allowances. But family allowances cannot be properly understood
as pronatalist policies in the Netherlands. The brake on universalization
before the early 1960s was political, rather than demographic, in nature.
Family Allowances

A combination of opposition from employers and the Protestant Reform
Party (Anti-Revolutionaire Partij, or ARP) within the Labor Foundation
and rather weak incentives to expand family allowances to the excluded
Catholic constituency of nonpoor self-employed account for the delay in
universalization. The breakthrough in 1962 came about because of a gen-
eralized trend toward a more universalistic conception of social rights in
Dutch society and increased electoral competition that made expansion of
family allowances to the self-employed a valuable electoral tool for both
Catholic and Protestant politicians. The existence of a tax system capa-
ble of taxing the self-employed made it possible to formulate this policy
expansion as a universalization of bene¬ts.

From the Beginnings to Universalism
The history of family allowances in the Netherlands extends back to the
¬rst years of the twentieth century, and it is worth investigating the process
through which the 1962 system of universal coverage emerged from the
¬rst collectively bargained agreements for textile workers and provisions
for municipal workers.3 The papal encyclical Rerum novarum, published in
1891, advocated a “family wage,” and Dutch clergy quickly adopted the
term. In 1906 the city of Amsterdam introduced a wage supplement for
family heads, and confessional groups in the Netherlands shifted the focus
of their social justice rhetoric from the concept of a minimum, just wage to
a family wage. The year 1912 saw the introduction of the ¬rst national-level
family allowances, for postal workers and teachers, and in 1918 Catholic
parliamentarians passed a resolution calling for the government to consider
family size in setting wages for all government workers.
During the interwar period, the national government granted family
allowances to all civil servants, and family allowances were introduced for
employees of some local governments. Private family allowance funds were
also developing “ in 1919, Catholic textile mill owners signed the ¬rst
collectively bargained contract including family allowances “ but by and
large employers supported the idea that the state should bear most of the
costs (Akkerman 1998). In 1920 the Catholic Minister of Labor announced
a plan to establish a government fund for child allowances, and in 1921
the Dutch government, at the urging of Catholic trade union leaders,

3 The historical reconstruction in this section draws largely from van Praag 1977; Damsma
1994; Rigter et al. 1995; and Akkerman 1998.

Age in the Welfare State

commissioned a study of the feasibility of a public family allowance sys-
tem (Damsma 1994).
Socialist unions and the dominant Liberal wing of the Dutch feminist
movement were opposed to the idea of a family wage, supporting instead
the idea of a just minimum wage for all workers that would be high enough
to support a family™s needs regardless of the actual number of children
present. But Catholic support for family allowances was strong, in part
because it dovetailed so neatly with the Dutch Church™s strong pronatalist
stance during that period.4 By the 1930s, Socialist opposition to family
allowances waned as it became clear that confessional parties and unions
were gaining the upper hand among both Catholic and Socialist voters as
“defenders of the family.”
By 1937, some 146 collectively bargained labor agreements included
family allowances (van Praag 1977), and in 1938 the Family Allowances
Act (Kinderbijslagswet) was passed. This act provided allowances for wage
earners only, based on an insurance principle: the level of the bene¬t
depended on the premium paid, which in turn depended on the wage
level. The impetus for the 1938 bill came from Catholic legislators, but
by this time the Socialist movement had accepted the idea of a general fam-
ily allowance scheme covering male workers. (It is interesting to note that
in the same year the Socialist Party and unions abandoned their advocacy
of a state-¬nanced old-age pension system, believing that population aging
would make public pensions prohibitively expensive [van Praag 1977].) The
only principled opposition to the Family Allowances Act came from Liberal
parliamentarians (Akkerman 1998), who objected to state provision of any
kind. Nevertheless, implementation of the act was delayed due to wartime
exigencies, and it did not come into force until the Nazi occupation gov-
ernment™s reform of the social welfare system (van der Valk 1991).
After the Second World War, legislation in 1946 expanded on the 1938
act by introducing a temporary family allowance measure to cover the ¬rst
and second children of wage earners. This temporary expedient was explic-
itly linked to incomes policies, in that it sought to compensate employees
for wage controls in the face of rising living costs. In 1948 the Pensioners™
Family Allowance Act (Kinderbijslagwet Rentetrekkers) introduced child

4 This strong pronatalist position may seem unwarranted, given exceptionally high population
growth in the Netherlands relative to the rest of Europe (Goddijn 1975). It is important to
keep in mind that pronatalism in the Netherlands during this period is essentially a result
of the minority position of Dutch Catholics.

Family Allowances

bene¬ts for old-age and invalidity pensioners, and in 1951 another tem-
porary family allowances act (Noodwet Kinderbijslagwet voor Kleine
Zelfstandigen) introduced bene¬ts for the ¬rst and second children of low-
income self-employed families.
In 1962 separate programs for public sector workers and the self-
employed were abolished and replaced by a quartet of provisions that
together covered the entire population. The General Family Allowances
Act (Algemene Kinderbijslagswet) of that year provided ¬‚at-rate allowances
for all residents of the Netherlands from the third child, with the amount of
the bene¬t augmented for each additional child. The Wage-Earners Fam-
ily Allowances Act (Kinderbijslagswet Loontrekkers), the Self-Employed
Persons Family Allowances Act (Kinderbijslagswet Zelfstanigden), and the
1963 Government Personnel Family Support Act (Kindertoeslagregeling
voor Overheidspersoneel) provided additional bene¬ts for the ¬rst two chil-
dren of, respectively, employees, the self-employed with modest incomes,
and public servants.
The history of Dutch family allowances up to 1962“3 shows a clear pro-
gression in the direction of citizenship-based entitlement. By the 1930s,
Dutch feminists™ vision of a carer™s wage had been defeated, as had the
Socialists™ goal of a family policy de¬ned by a minimum wage for all work-
ers suf¬cient to maintain a family, plus collectively provided services for
mothers. Instead, the idea of a family wage for family heads, introduced in
Rerum novarum and favored by confessional parties and unions, prevailed.
Family allowances were introduced as part of the wage package for
working men with dependent children. This type of bene¬t was gradu-
ally extended from a select group of civil servants and private employees
to include, ¬rst, all wage earners, then social insurance pensioners and
low-income self-employed persons, and ¬nally in 1962 all residents of the
Netherlands. This universalization of bene¬ts never occurred in Italy. But
why did it take so long in the Netherlands, despite the popularity of these
programs and the in¬‚uential report of the van Rhijn Commission, which
advocated universal family allowances?
Gauthier (1993) suggests that the expansion of family allowances intro-
duced as pronatalist policies may stall in the absence of a credible threat
of population decline. Fertility rates were indeed high in the Netherlands
during the immediate postwar period “ so high that the Dutch government
actively encouraged emigration in order to counter feared overpopulation.
And it is true that Catholic concerns about their minority status and the
pernicious in¬‚uence of “Neo-Malthusian” birth control advocates formed a
Age in the Welfare State

strong impetus for family allowances during the interwar period. But Dutch
Catholics continued to place a high priority on reproduction even in the
postwar period, a priority re¬‚ected by their extraordinarily high fertility
rates (van Poppel 1985). At the same time, Catholic families were among
the least likely to receive non-universal family allowances in the 1940s and
1950s, since they were concentrated in agricultural self-employment rather
than in industrial or civil service jobs (Stoffelsma and Oosterhaven 1989).
So despite high levels of fertility in the Dutch territory as a whole, Catholic
politicians were strong advocates of expanding the family allowance system
as quickly as possible “ both to encourage higher birth rates among Catholic
self-employed families and to capture electoral support from expansion of
family allowances to this constituency.
Compared with family policies in France, Belgium, or Sweden, coun-
tries that adopted (quite different) family policies for explicitly pronatalist
reasons, Dutch family allowances were considered a component of wage
policies (Rigter, van den Bosch, van der Veen, and Hemerijck 1995, 222).
Old-age and invalidity pensioners became eligible for family allowances in
1948, but by virtue of their status as former workers entitled to a deferred
wage. Social Democratic unions continued to think of family allowances as
an integral component of the wage as late as 1957. Although the Minister of
Social Welfare advocated in 1955 a universal tax-¬nanced family allowances
scheme, and the Socialist trade union confederation (Nederlands Verbond
de Vakverenigingen, or NVV) generally agreed with this recommenda-
tion, they made it clear that family allowances were not to be classed with
other “welfare” bene¬ts. A report from an NVV congress in October 1957
declared that tax ¬nancing of social insurance was an important goal since
it would provide a means of balancing the incomes of “non- and semi-
productive groups” with “a reasonable income (including children™s allowance
and after tax deduction) for the productive groups”5 (NVV 1958). Family
allowances were linked primarily to ideas about a just wage, and secondarily
to ideas of poverty alleviation “ not to increasing birth rates.
Still, we might reasonably presume that Catholic politicians had an
objective interest in promoting expansion of bene¬ts in order to reach
their large constituency of self-employed families with numerous children.
So it makes sense to ask to what extent these political actors were actually
capable of and/or desirous of extending family allowances to the entire
population. Catholics controlled the social welfare ministry through most

5 Emphasis added.

Family Allowances

of the 1950s and 1960s. And their Social Democratic coalition partners
certainly had no principled objection to universal social programs, having
supported universal old-age pensions in 1947. Union publications indicate
that the NVV supported proposals for both universal family allowances
(NVV 1955) and universal widows™ and orphans™ bene¬ts (NVV 1957),
though expressing a preference that such universalistic programs be
¬nanced through general revenues rather than employer and/or employee
contributions. The ARP generally opposed expansionary social policy, but
was not represented in government until 1952. So introducing universal
family allowances should have been an easy task for Catholic politicians.
What, then, accounts for the delay?
Cox (1993) assigns primary responsibility for the delay in expanding
social welfare bene¬ts in the Netherlands to the in¬‚uence of confes-
sional organizations within corporatist decision-making bodies. According
to Cox, while the Catholic Party supported universalizing social insurance
bene¬ts, confessional employers™ and labor groups opposed it and launched
“counter-attacks” from within the Labor Foundation (111). In particular,
the ARP™s objections to universalization had an important impact on policy
via the in¬‚uence of Protestant representatives in the Labor Foundation.
This in¬‚uence stalled the expansion and universalization of social policies
through the 1950s, despite the fact that a Social Democratic party was in
control of government and the relevant ministries.
Many observers of Dutch social policy argue that growing seculariza-
tion and depillarization of the electorate in the mid-1960s account for
the explosion of new rights and social spending during this period (de
Swaan 1988; Cox 1993; van Zanden 1998). According to these analysts,
both Catholic and Protestant parties and organizations began a “populist
campaign” (Cox 1993, 168“9) using welfare bene¬ts as tender in a context
of declining electoral margins. The ARP, which had traditionally been a
party of ¬scal conservatism, softened its objection to government spend-
ing under the in¬‚uence of a booming economy and pressure from self-
employed groups, which were threatening to exit the weakening party. Once
the Social Democratic Labor Party (Partij van de Arbeid, or PvdA) began
to make serious inroads into the Catholic vote in the early 1960s, Catholic
politicians as well began looking for concrete bene¬ts to offer in exchange
for electoral support. Intense competition with both the PvdA and smaller
left-Catholic parties increased the in¬‚uence of the left current within the
Catholic People™s Party (Katholieke Volkspartij, or KVP), which was cer-
tainly not averse to expanding social bene¬ts even if that meant increasing
Age in the Welfare State

the state™s role in social provision. Catholic unions were also radicalizing in
response to competition from Social Democratic unions and began to sup-
port universal social bene¬ts in response to their popularity with the public
(Cox 1999).
However, the move to introduce universal family allowances in fact pre-
ceded depillarization and increasing electoral competition by some ten
years (Balkenende 2000 interview; van der Veen 2000 interview). Political
parties could not have changed their policy commitments quickly enough
in response to depillarization to account for such a major change in the
family allowance system as early as 1962 (Cuyvers 2000 interview). The
growth of the welfare state in general, and of family allowances in par-
ticular, may be better understood in terms of the ideological maturation
of Dutch policy makers than in terms of increasing electoral competi-
tion resulting from depillarization. In the judgment of Christian Demo-
cratic Appeal party leader Jan Peter Balkenende, “[KVP social policy
makers] Veldkamp and Klomp´ were left by conviction, not by strategy”
(Balkenende 2000 interview). Roebroek (1992), too, cites a growing consen-
sus among confessional politicians on the desirability of state intervention
as the key motivator of welfare state expansion in the 1960s, adding that of
course explosive economic growth was a necessary precondition for such a
If increased electoral competitiveness in the 1960s cannot explain the
drive for welfare state expansion and the adoption of universalistic welfare
policies, neither did it dampen the interest of confessional parties in expand-
ing social welfare spending. However, whether or not increasing electoral
competition was the key to welfare state expansion, it still would remain to
be shown why expansion of the family allowance program took the form of
universalization. Here the electoral argument may be useful in augmenting
the rather vague argument prevalent in the secondary literature about the
ideological transformation of Dutch political culture in the 1960s.
With the breakdown of pillarization in the Netherlands, when voters™
automatic ties with the traditional religious subcultures weakened, the lead-
ing parties found themselves with a new problem: how to attract voters
who might decide to vote for a party other than the one af¬liated with
their religious denomination. For both the Protestant and Catholic parties,
the most consistent source of electoral strength was also the one that was
excluded from family allowances under the old occupational system. This
group of voters resided in the families of the Protestant and rural Catholic
self-employed. Extending family allowances to the self-employed, which in
Family Allowances

the Dutch case was accomplished by universalizing the family allowance
system, covered this precise segment of crucial voters and constitutes a
motive for expansion in a universalist direction.
The universalistic solution was possible because of the comprehensive
Dutch tax system, which allowed citizenship-based bene¬ts to be paid for
out of general revenues. After the introduction of the income tax in 1917,
there was little or no discussion of the administration of the income tax
system in the Netherlands (van der Veen 2000 interview). By all accounts,
it was a system that worked well and was implemented in a bureaucrati-
cally rational manner (see Ferrera 1993). The van Rhijn Commission had
proposed a universal family bene¬t to be ¬nanced out of general revenues,
but in the immediate postwar period Drees opposed the idea of extending
family allowances to the self-employed on the grounds that it would be
too expensive (Rigter et al. 1995, 222). By the late 1950s, however, both
the Socio-Economic Council (Sociaal-Economische Raad, or SER) and a
majority of parliamentarians were eager to integrate family allowances for
the self-employed with legislation for wage earners. In fact, when Minister
of Social Welfare van Rooy proposed in 1958 a law instituting a universal
family allowance system for the third child and up, he was forced to step
down because politicians in the lower house of Parliament did not ¬nd his
plan generous enough. Parliamentarians argued that given the feasibility
of funding the system out of general revenues, there was no reason not to
extend bene¬ts to ¬rst and second children as well (Rigter et al. 1995, 257).
The construction of a universal Dutch family allowance system based on
a general tax base6 was a key development. The strategic maneuvering of
Catholic and Protestant parties to appeal to the rural self-employed created
a demand for a universal bene¬t, and effective taxation of the self-employed
made it possible for politicians to please these important constituencies as
well as their urban supporters. On the one hand, politicians could offer the
concrete bene¬t of improved family allowances to the self-employed. At
the same time, the tax-¬nanced nature of the system ensured the support of
unions and employers, who in general favored universal citizenship-based
social insurance but needed to be reassured that the cost of universalizing
family allowances would not result in increased social insurance contri-
butions. This compromise had, as we shall see, important consequences
for the development of family allowance spending in the 1970s through

6 Note however that until 1980, only bene¬ts for the self-employed were paid out of general
revenues; wage earners continued to fund their bene¬ts via employer contributions.

Age in the Welfare State

1990s. It also provides a useful contrast with the Italian case, where a weak
tax base prevented the implementation of a universalistic family allowance

From Universalism to Entitlement
Following on the universalization of the family allowance system in 1962,
expenditures in the Netherlands showed a steep increase from 1963 to 1968.
The mid-1960s were a phase of major growth in many areas of social pol-
icy spending, initiated by Catholic ministers Veldkamp and Klomp´ and e
strongly supported by the Catholic and Social Democratic unions. This
major expansion of social welfare spending coincided with the depillariza-
tion of the Dutch electorate, beginning around 1965 and characterized by
a very strong increase in competition between the major parties for lower-
and middle-class voters. Between 1963 and 1972, the KVP lost half of
its electorate, mostly to the Labor Party and the Radical Party (Politieke
Partij Radicalen, or PPR, a left-wing offshoot of the KVP; Irving 1979,
201, 224“5). By the late 1960s, many of the Catholic Party™s core voters had
abandoned the KVP in order to support smaller parties like D™66, the PPR,
and DS ™79 (Daalder 1987). Given the intense electoral competition of the
mid- to late-1960s, it is not surprising that spending on family allowances
increased just as the Catholic Party found itself most in need of incentives
with which to retain a key constituency: families with young children.
It is far from clear, however, that increasing electoral competition was
solely responsible for the rise in family allowance spending that began in
the mid-1960s. The abolition of the income ceiling for contributions (in
1963 for employees, 1964 for self-employed) and the indexation of bene¬ts
to the cost of living (1964) surely also resulted in higher spending. And even
as the electoral climate became less favorable for family policy expansion
in the 1970s, expenditures continued to rise.
By 1975, when the Catholic and Protestant parties began the process of
merging to form the Christian Democratic Appeal (Christen Democratisch
Appel, or CDA), the support base of the Christian parties was substantially
older than the average voter (Bakvis 1981, 81), so an expansion of family
allowances would hardly have constituted a tremendous appeal. The merger
of the confessional parties effectively marginalized the left-wing elements
that had pushed for higher social spending in the late 1960s (van Zanden
1998, 71), and between 1974 and 1981 CDA voters moved to the right of the
political spectrum (Daalder 1987). This diminution of the core constituency
Family Allowances

of the family allowance program within the Christian Democratic parties
might account for their willingness to go along with de-indexation of family
allowances in 1972, a situation that continued until 1981. But even after de-
indexation, family allowance spending did not decline. In fact, there was
no signi¬cant decline in family allowance spending until 1984“6, when the
level of family allowances was brie¬‚y frozen as part of the general austerity
policies of the era.
Aggregate family allowance spending held steady through the 1970s,
and from 1980 onward, the general trend was again one of strong growth
in spending. Why does spending on family allowances start to rise again
just when recipients are farthest from the core of the Christian Democratic
electoral project, and when family allowances are less popular among voters
than other kinds of welfare bene¬ts (Palomba 1995)? While the Dutch
Family Council and the Dutch Association of Housewives were active
in pressing for disability bene¬ts for housewives (Cox 1993, 164), these
organizations do not seem to have been particularly concerned with family
allowances (Cuyvers 2000 interview). Neither do labor unions seem to have
pressed very hard for bene¬t increases. The combined union Urgency Pro-
gram for 1973 objected to the idea of freezing bene¬ts for the second child
under the wage earner™s program and called for a re-evaluation of the whole
family allowance system (NVV 1973) “ a call that is repeated in the Urgency
Program for 1975 (NVV 1975). But it is not clear from these documents
what unions wanted from a potential reorganization of the system, and van
Berkel and Hindriks™s (1991) survey of the relationship between unions and
social insurance claimants makes no mention of the children™s allowance
The rise in family allowance spending in the 1980s can be attributed to
three factors. First, the general climate of austerity may have led to demands
for increased bene¬t levels. Second, the 1980 consolidation of Dutch family
allowance legislation into a single law involved a change in the mode of
¬nancing the system. Whereas prior to 1980 the family allowance fund
for employed workers was ¬nanced by contributions from employers and
employees, after 1980 the entire system, including bene¬ts for employees,
became entirely ¬nanced out of general revenues. According to Balkenende
(2000 interview), popular pressure for increases in the family allowance
bene¬t arose immediately upon this change in ¬nancing, since payments
into the system had effectively been made invisible.
Finally, and perhaps most importantly, changing the structure of the fam-
ily allowance system also changed societal expectations about the function
Age in the Welfare State

that allowances should serve. Bussemaker (1998) argues that the feminist
movement™s drive to obtain equal rights for women under social security law
in the 1970s drove up social spending in general. Individuals increasingly
became entitled to bene¬ts that were once meant to support a family head
plus dependents, but the amount of bene¬ts was not adjusted downward
to account for this new trend toward “individualization” of social rights.
Indeed, the public campaign to “save” family allowances in 1974“5 was
spearheaded by left-wing women™s groups seeking to protect bene¬ts for
single mothers (Vlek 1997; Cuyvers 2000 interview). Women™s organiza-
tions pressed for increases in the level of family allowances in order to allow
single mothers to continue to provide care in the home, without recourse
to income from employment (Nederlandse Vrouwenbeweging 1976; Molin
The universalization of the family allowance system in 1962 thus had
important consequences for expenditures in later years. The 1920s and
1930s marked a process of de¬ning family allowances socially as a wage
supplement, as we have seen. The occupational system installed in 1938
and continued in the 1940s and 1950s reinforced the notion that family
allowances were primarily an antidote to wage controls and not an inde-
pendent entitlement of parenthood. Once family allowances were univer-
salized, however, they became removed from the realm of workers™ bene¬ts
and came to be seen as an entitlement of citizenship. No longer linked to
wages as a top-off, family allowances in the Netherlands after the 1960s
came to be regarded as an entitlement that should enable a family head,
male or female, employed or not, to support children.
The Dutch family allowance system got off to a slow start after World
War II, relative to Italy™s generous bene¬ts in the immediate postwar period.
The drive to expand the family allowance bene¬t in the late 1950s and early
1960s may have been a result of increasing electoral competition and a
concomitant desire on the part of Catholic and Protestant parties to secure
the votes of the self-employed, as many scholars argue, or it may have
simply re¬‚ected a leftward shift in the ideological orientation of the Dutch
polity at large. Regardless of the motive for expanding the family allowance
system, it is clear that the opportunity for creating a universal, citizenship-
based plan system in 1962 arose from the capacity to fund bene¬ts for the
self-employed via a system of effective taxation.
Once the family allowance system became universalized, the possibility
for further use of the system to provide selective incentives to consolidate

Family Allowances

electoral gains among particular constituencies became more limited. But
the growth of family allowance spending continued through the 1990s, even
in the face of general austerity measures. The move from an occupational
conception of family allowances to a universal, citizenship-based system
changed societal expectations about the nature and function of those ben-
e¬ts. In the process, it created new sources of pressure to increase family
bene¬ts, which resulted in continued spending growth.

Family Allowances in Italy
Family allowance bene¬ts in Italy have followed a path opposite to the
Dutch trajectory. Italian family policy in general, and family allowances in
particular, were by the 1990s widely considered to be utterly inadequate.
Public debate blamed the lack of transfers and services for families for
high levels of child poverty and the world™s lowest fertility rates. But family
allowances in Italy have not always been the “Cinderella” of social pol-
icy, as one in¬‚uential tract of the late 1970s termed them (Gorrieri 1972).
In the 1950s and early 1960s, family allowances were quite generous by
international standards, and although they were not available to all citi-
zens, progress toward universalization seemed well under way. Why did
the expansion of Italian family allowances stall in the late 1960s? What can
explain the pattern of rapid growth and then equally rapid retrenchment in
the postwar period?
The strategic use of family allowances by party actors, in interaction
with the occupational structure of the family allowance system, explains
the developmental trajectory of family allowances in Italy. The motive for
expanding the scope of family allowance bene¬ts in Italy was the desire
of politicians to reach out with selective bene¬ts to particular groups in
the electorate. Unlike in the Netherlands, however, where electoral incen-
tives led to a universalization of family bene¬ts, in Italy the opportunity to
introduce a universal system of bene¬ts was missing because of the under-
developed tax system, itself a by-product of electoral clientelism. As a result,
expansion of the family allowance system in Italy took place along patch-
work lines, within an occupational structure that continued to feed and
be fed by particularistic politics. The continuing occupationalism of the
system also set a brake on the growth of family bene¬ts by affecting the
priorities of potential constituents. Family allowance funds were siphoned
off to pay for other social priorities, especially pensions, in a move that the

Age in the Welfare State

unions and the Left agreed to quite willingly. As long as family bene¬ts
remained linked conceptually to the wage system, and wages were indexed
but family allowances were not, family allowances came to seem less and
less important, and less and less worth defending.

General Historical Overview
Until the reforms of the mid-1980s, the basic outlines and principles
of Italian family allowances all derived from developments during the
Fascist period. Family allowances were ¬rst introduced in Italy during the
Fascist period as a way to compensate employees with children for reduced
wages resulting from reduced hours. After the Second World War, fam-
ily allowances were rapidly re-established as a core feature of the welfare
state and reached a peak of generosity during the mid-1950s and early
1960s. At this point, Italian family allowances, despite being available only
to industrial workers and small farmers, supported a wide variety of fam-
ily dependents (spouses, parents, and siblings, in addition to children and
grandchildren) and were admired both within Italy and by international
Family allowances were gradually extended to cover new groups of con-
stituents, until by the mid-1970s they covered nearly 80 percent of children
under eighteen in Italy. However, starting in the mid-1960s, the allowances
underwent a dramatic devaluation, and by the late 1970s Italian family
allowances were among the least generous in Europe. By the 1970s, sur-
pluses in the family allowance fund were routinely diverted to pay for pen-
sions, and family allowances dwindled to insigni¬cance despite the favorable
ratio of contributions to payouts.
The mid-1980s brought a series of reforms, including the introduction
of means testing and the graduation of bene¬ts according to family size and
income. However, even after these reforms, family allowances remained
restricted to needy families of dependent employees or the self-employed
in agriculture, and did not become a citizenship-based form of assistance.
The reforms of the 1980s, motivated by the desire to reduce spending
while targeting resources more effectively to needy families, substantially
decreased both the number of recipients and aggregate expenditures on
family allowances. Not until the mid-1990s did there emerge an open debate
on the condition of families in Italy™s welfare state, a debate that has spurred
in recent years a slight increase in aggregate spending on families over 1980
Family Allowances

The Fascist Period
After some initial experiments during World War I that vanished with the
1920s, family allowances got their true start in Italy during the Fascist
period. With economic crisis and unemployment looming, the corporatist
employer association Con¬ndustria and the Fascist labor unions signed an
agreement in 1934 to reduce working hours in industry from forty-eight to
forty per week, in an attempt to reduce unemployment. Workers agreed to
concomitant reductions in salaries, while those workers with two or more
dependent children received a family allowance to make up for some of the
lost wages. The family allowances were to be funded by joint contributions
from employers and employees. This interconfederal agreement, strongly
advocated by Con¬ndustria, was transformed into legislation in 1936, but
without the linkage to reduced hours. By this time the Fascist pronatalist
agenda was in full swing (the campaign to increase birth rates began in
1927, and in 1931 the government-sponsored Inquiry into the State of
the Family came out), and government policies linked the papal encyclical
Rerum novarum™s call for family wages to the demographic campaign.
In 1937 family allowances were extended from industry and some other
sectors to agricultural employees and the self-employed with incomes under
a certain threshold. Allowances for children were graduated to give greater
per-child bene¬ts to larger families. In 1940 the family allowance scheme
was placed under government authority with the creation of the Single Fund
for Family Allowances (Cassa Unica per Assegni Familiari, or CUAF) within
the National Institute for Social Insurance (Istituto Nazionale per la Prev-
idenza Sociale, or INPS). Allowances were extended to dependent spouses
and parents in addition to children.
Three main forces drove expansion of family allowances during the
Fascist period: poverty, pronatalism, and patronage. Family allowances
were a way to combat poverty by supplementing the wages of low-income
earners. Large families, in particular, were subject to poverty because wages
were so low, and family allowances helped to prop up consumption at the
lower wage levels. The demographic impulse was of obvious importance
as well. Despite the fact that Italy had one of the highest birth rates in
Europe at the time, the Fascist government was concerned about Italian
birth rates for a variety of reasons, including the impact of the French
debate over pronatalist policies, widespread emigration of men, the impor-
tance of human resources for development in a capital-scarce environment,
and the in¬‚uence of Church doctrine favoring high fertility (De Grazia
Age in the Welfare State

1992, 25). In the end, family allowances did little either to alleviate poverty
or to promote more births. But they served other purposes, which ensured
the survival of the system. Family allowances were a form of patronage for
middle-class cadres, who formed the backbone of support for Mussolini™s
regime. Family allowances were much more generous for this group than
for ordinary workers, despite the fact that state employees had smaller fam-
ilies and higher incomes than the country™s working classes.
During the Second World War, family allowances lost much of their
purchasing power due to declining resources and high in¬‚ation. In 1944
the child allowances returned to a ¬‚at-rate sum for each child, regardless of
birth order, and the family allowances were supplemented by an additional
cost of living allowance. During the war the state brie¬‚y took over ¬nancing
of the allowances, but the main outlines of Italy™s family allowance system
had been set in place with the 1940 legislation, and after the war the old
¬nancing system returned.

The Golden Age of Italian Family Allowances: Postwar through 1964
During the immediate postwar period, family allowances rapidly regained
the purchasing power they had lost to in¬‚ation, by 1951 surpassing the
highest real value they had achieved during the peak of the Fascist demo-
graphic campaign. While numerous observers (see, e.g., Sabbadini 1985;
Campanini 1993; Silvestrini 1994) argue that poor bene¬ts for Italian fami-
lies in the contemporary period are a result of politicians™ desire to avoid
any perceived continuity with hated Fascist pronatalism “ a claim echoed
in Valiente™s (1996) work on Spain “ the development of family allowances
is clearly more complicated than that.
Progressive allowances for larger families, a key feature of pronatalist
policies, had been removed already in 1944, and after the war progressive
allowances were not reinstated. So what remained of the Fascist family
allowance system after the war was simply a wage supplement that, by
all accounts, was extremely popular. The system showed no sign of being
dogged by the shadow of Fascism. It just kept growing and growing. From
1946 until 1955, when family allowances were automatically indexed to the
cost of living, there were sixteen legislated increases to the family allowance.
Contributions went up in like measure (despite the massimale, a ceiling on
the total wage bill above which employers did not have to pay contributions,
in place until 1974), so that through the 1950s and 1960s payouts rarely
exceeded contributions received by the Family Allowance Fund.
Family Allowances

Family allowances arrived at this golden age in the 1950s and 1960s
because they were uncontroversial. The Church liked family allowances,
because while it generally opposed the idea of state intervention in the
family sphere, family allowances were a cash bene¬t distributed to the
(usually male) main earner in a household, a format relatively unthreat-
ening to family autonomy and, within that, patriarchal authority. And after
all, the family wage had been the Pope™s idea to begin with. The Christian
Democratic Party liked family allowances, in part because the Church liked
them, and in part because the socially oriented current that was dominant
within the party through the mid-1950s believed that supporting families
with children, especially low-income families, was the right thing to do.
Con¬ndustria, the large employer™s association, liked family allowances,
because they kept wages down and because the existence of the massimale
meant that the biggest northern employers who had the largest wage bills
and who dominated Con¬ndustria paid the least, proportionally.
Unions liked family allowances, because they were a substantial compo-
nent of the wage (around 20 percent of the average industrial wage in the
1950s and 1960s), and the ¬‚at-rate format made the allowances even more
helpful for low-income workers. In fact, according to Franco and Sartor
(1994, 86), during this period family allowances were the social insurance
program with which unions were most concerned. The Left, interested in
social justice and in spurring the economic development of the south, liked
family allowances because workers in the south, whose wages and living
costs both tended to be lower than in the north, bene¬ted substantially. On
aggregate the family allowance system was an effective means of transferring
resources from the rich, less fecund north to the poor, proli¬c south. Finally,
Italy™s family allowances were well regarded by social policy experts as an
ef¬cient system delivering tangible bene¬ts, something in which the Italian
government could take some pride (see Masini 1953; Pasi 1956). In short,
there was no substantial opposition to maintaining the family allowance
system as it had been laid out in the 1940s, while increasing bene¬ts as it
became necessary and/or possible.
Extending family allowances beyond the original constituency covered
in the 1940 law proved trickier. Some extensions did occur during the early
postwar period. In 1947 and 1948, artisans and tobacco workers came under
the INPS family allowance scheme, and in 1952 all public employees came
to be covered by a separate public family allowances program. In 1958 the
regular family allowance scheme was extended to ¬shermen. Beginning
in the early 1960s, there were regular proposals by Christian Democratic


. 3
( 7)