. 6
( 7)


much. The Left is not as strong in the Netherlands, nor as weak in Italy, as
one might suspect. Despite the presence of Red-Roman political coalitions
in the Netherlands throughout much of the period from 1950 to 1970, the
Ministry of Social Affairs remained controlled by Catholic politicians for
most of the period. Developments in the public pension system following
Drees™s initial emergency provision in 1947 have come largely under the
watch of Christian Democracy, not Social Democracy. As for Italy, despite
its characterization as a Christian Democratic single-party-dominant sys-
tem (Barnes 1990), the Left has not been uninvolved in the area of pension
policy. Communist legislators in Italy have been closely involved in pen-
sion policy making through their positions on the parliamentary commit-
tees responsible for social policy and on the board of directors of the state
pension agency. And as Cazzola (1995, 26“7) remarks, somewhat ruefully,
most social legislation in the 1970s and 1980s was passed in Parliament
with almost unanimous approval of both government and opposition leg-
islators). Finally, Italian unions, like their Dutch counterparts, have been
closely involved in the pension policy-making process during both periods
of neo-corporatist concertation and episodes of more con¬‚icting labor rela-
tions. And while Golden Age Dutch corporatism was primarily a system for
wage control (Hemerijck 1992, chapter 7), in Italy concertation has often
occurred explicitly around the issue of social policy.
So resolving the question of how much impact Social Democracy has
on pension spending is a bit more complicated than it ¬rst appears. Lower
pension spending is clearly associated with a more Social Democratically
oriented polity in the Netherlands, and higher pension spending with
Christian Democratic dominance in Italy. Setting aside these broad char-
acterizations, could it still be the case that the variable political power of
the Left is responsible for higher pension spending in Italy and for lower
expenditures in the Netherlands?
Even taking a more nuanced view of the strength of Social Democratic
actors and institutions in these two countries, our conclusion still must be
that the strength or weakness of the Left is not the key to explaining different
patterns of aggregate pension spending. In neither Italy nor the Netherlands
does the level of left power resources bear primary responsibility for the
policy features that most affect the level of bene¬ts and the number of
Age in the Welfare State

bene¬ciaries, and thus create large differences in pension spending between
the two countries.
The concept of an adequate social minimum, which I argue was crucial
for containing the level of high-end bene¬ts in the Netherlands, seems to
have been supported by a very widespread societal consensus. All political
parties, and employers and crown members as well as labor representatives
on the Socio-Economic Council, agreed on the desirability of a pension
bene¬t linked to an adequate social minimum, even while they disagreed
quite strongly on key features of how the pension system should be admin-
istered. The establishment of bene¬t levels based on a social minimum
adequate to meet living costs cannot, then, be attributed to a “weakness”
of Social Democracy, the more so since the Labor Party was an active par-
ticipant in government and particularly in social affairs, and corporatist
institutions were at their strongest, during the period when the social min-
imum was established.
The failure of the PvdA to completely dominate government and neo-
corporatist policy-making bodies in the Netherlands immediately follow-
ing World War II could conceivably explain why a viable system of private
supplementary pensions remained in place, though it must be noted that
unions were supportive of a private system even if the Social Democratic
Party was not. While PvdA politicians would have preferred to replace the
private system entirely with a public universal bene¬t early in the post-
war period, unions were eager to maintain an important role for private
supplementary bene¬ts on the condition that they be allowed to share in
their administration. The supplementary pension system was also desirable
to unions because it gave them something other than wages to bargain over
in a period of stringent wage controls.
Neither the PvdA nor labor unions in the Netherlands have played an
important role in limiting the number of pension bene¬ciaries, either. Social
Democrats do not appear to have requested at any point a lowering of the
statutory retirement age, but neither do they seem to have been responsible
for setting the retirement age at sixty-¬ve to begin with. Again, this appears
to have been a matter of rather broad societal consensus dating from the
prewar period. Social Democratic actors have pushed for the maintenance of
avenues of early exit from the work force “ for example, disability pensions,
early retirement, and extended unemployment provisions “ but employers
have been equally adamant supporters of such measures. It does not appear,
then, that the variable strength of Social Democracy over the years in the

Old-Age Pensions

Netherlands has been responsible for either setting in place or maintaining
the policies that we have seen lead to low pension spending relative to Italy.
Nor is the Left in Italy the primary architect of policies encouraging
higher pension spending. In the period prior to 1969, left parties and unions
both advocated a universal pension system with a ¬xed social minimum, and
their pressure tactics were aimed primarily at improving low-end pensions.
Neither is there any evidence that the Left or labor unions in any way
stymied the growth of private pensions. High bene¬t levels for high-end
pensions have certainly been supported by some professional associations,
but the confederal unions have pushed for a ceiling on pensionable income,
and there is a clear relationship between left partisanship and sponsor-
ship of pension bills aimed at the most encompassing and lowest-income
groups (Maestri 1987). Category unions such as those representing metal
workers did press for higher bene¬ts for their own sectors, and sometimes
won. But once the confederal unions gained control of the administration
of INPS in 1971, unions increasingly stood in favor of reducing high-end
bene¬ts (Regini and Regonini 1981, 240). Ultimate responsibility for very
high high-end pensions in Italy lies with politicians, primarily politicians of
the DC and the PSI, whose generosity toward key sectors (public employ-
ees, managers, and professionals) resulted in a constant ratcheting up of
The most important provisions determining the number of pension ben-
e¬ciaries in Italy “ the early retirement age and access to seniority pensions
“ were put into place during periods when neither unions nor left parties
had any input into social policy formation (the Fascist period in the case
of the retirement age, and the mid-1950s in the case of seniority pensions
for public sector workers). Can they be held responsible for the failure to
reform these system features? Unions, at least at the confederal level, sup-
ported proposals to raise the retirement age and cut seniority privileges at
least as often as they opposed such measures. Left parties have played a
more ambiguous role. While the rhetoric of left politicians has emphasized
reducing the privileges of public sector workers, PCI/(P)DS politicians have
tended to vote in much the same way as other politicians on major pen-
sion legislation (Regonini 1987, 103; Maestri 1994). It does seem to be the
case that when unions are incorporated into the process of reform, reforms
tend to go more smoothly and parties are less tempted to reach out to their
militant base to gain consensus (Regini and Regonini 1981; Baccaro 2000).
However, there is little evidence to suggest a distinctive Social Democratic

Age in the Welfare State

partisan impact on pension reform that would allow us to attribute Italy™s
high pension spending to the power resources of the Left.
If left power resources do not explain differences in pension spending
between Italy and the Netherlands, what about the power of the elderly?
Perhaps pensioners, who stand to bene¬t most directly from high pen-
sion bene¬ts, are responsible for pushing for the growth of public pension
spending and for opposing any attempts to curtail bene¬ts. To the extent
that the elderly constitute an overwhelming bloc of voters or are organized
into pressure groups that can effectively lobby for pro-elderly legislation,
they may be responsible for the shape and size of a nation™s pension sys-
tem. Of course, other constituencies besides the elderly stand to gain from
generous pensions: one need only think of adult children of the elderly,
who bene¬t by having the burden of care for their parents taken over by
the state, or of current workers, for whom generous pensions constitute a
promise of payment in the future. Nevertheless, there is good reason to
suspect that a politically mobilized and powerful elderly lobby can have an
important impact on pension spending.
The gray power hypothesis predicts that more elderly voters, and
stronger organizations of the elderly, should result in higher pension spend-
ing. Italy has a substantially older population now than does the Nether-
lands, and pensioners in Italy are organized into powerful pensioners™
unions of their own within the national confederations, while Dutch pen-
sioners have had limited success organizing within the union movement. So
one could conclude that higher pension spending in Italy occurs because
the elderly are more powerful political actors there than they are in the
But as with the characterization of the Netherlands as “almost Social
Democratic” and Italy as purely Christian Democratic, this image is not
entirely correct. Until rather recently, both Italy and the Netherlands had
quite young populations by European standards. It is only since the late
1970s that the Italian population has begun to age dramatically. And the
elderly have played an important role in Dutch political life, just as they
have in Italy. Since the immediate postwar period, the elderly have been
an important constituency for Dutch political parties, especially the PvdA,
which still enjoys favor among seniors as the party of Drees and the ¬rst
universal pension. The elderly in the Netherlands have been represented
since the 1950s by interest groups that work as equivalents to the unions
in Italy. These groups became quite active in the 1980s and 1990s and are
regularly consulted on policy matters relating to the elderly. Furthermore,
Old-Age Pensions

the early 1990s saw the emergence and entry into Parliament of several
pensioners™ parties in the Netherlands. These parties were not large, and
they did not maintain their parliamentary seats for long. But they did suc-
ceed in drawing enough voters away from the Christian Democratic Party
to force that party to back off from proposals to cut pensions (Balkenende
2000 interview; Green-Pedersen 2002).
Not only is the difference in the level of political in¬‚uence of the elderly
in Italy and the Netherlands not as great as one might initially assume. In
both countries, key policies that set the stage for future expenditure levels
were already in place well before the elderly became a powerful political
lobby. In the Netherlands, seniors strongly supported the universal pension
introduced by Drees in 1947, but had no involvement in the preservation of
the private occupational pillar. And the establishment of a social minimum
and the retirement age of sixty-¬ve seem not to have been issues of public
debate. In Italy pensioners™ organizations lobbied for better low-end ben-
e¬ts and a universal pension system, but had little success. And they were
involved neither in the establishment of public sector seniority pensions nor
in the setting of the low statutory retirement age. The gray power hypothe-
sis then cannot account for the establishment of the program features most
responsible for high pension spending in Italy and lower spending in the
It might still be the case, however, that the elderly have played a more
important role in maintaining program features that might otherwise have
gone by the wayside. This would accord with Pierson™s (1994) argument
that bene¬ciaries of the welfare state come into their own as important
political actors primarily during periods of attempted retrenchment. There
is some support for this argument on the Dutch side. The elderly in the
Netherlands in the 1980s and 1990s have tended to mobilize reactively
around proposed cuts to their bene¬ts, not proactively to counter such
proposals as has been the case with other groups such as single mothers
or the disabled (Vlek 2000 interview). Dutch elderly organizations have
spoken out against effective cuts in AOW pension bene¬ts resulting from
de-indexation (in the 1980s and early 1990s) and the imposition of taxes on
AOW bene¬ts (in the mid-1990s). However, it was not until pensioners™
parties posed a serious threat to the Christian Democratic Party in the late
1990s that these protests began to see any results. Protests by elderly orga-
nizations successfully derailed an attempt to introduce social security taxes
on AOW bene¬ts in 2000 and have effectively brought to the agenda the
issue of pensioners™ representation on the boards of private pension funds.
Age in the Welfare State

Paradoxically, the record shows that Italy™s very strong pensioners™ unions
have actually taken quite moderate stands on the issue of pension reform,
contrary to the presumption of many commentators that they would oppose
any reforms. Pensioners™ unions were formed after World War II to advance
the interests of former workers, and through the mid-1970s their demands
were focused almost entirely on improving pension provisions and health
care for the elderly. As the pensioners™ unions gained experience and inde-
pendence, their leadership gradually became a source of expertise for the
union confederations in the area of social policy as a whole. And rather than
using their new role to push solely for policies in the short-term self-interest
of current pensioners, the pensioners™ unions began increasingly to adopt
a broader view of their role as defenders of the welfare state. For example,
in 1976, well before the union confederations or political parties had taken
up the call for pension reform, the CISL pensioners™ af¬liate (Federazione
Nazionale Pensionati, or FNP) argued that unless the pension system were
reformed substantially, there was a danger of collapse in the long term (FNP
1992, 12).
Pensioners™ unions supported the failed pension reforms of 1978“9 and
the successful ones of 1992, 1995, and 1996. The yearly uni¬ed policy docu-
ments put forth by the three largest pensioners™ unions in the 1990s express
consistent support for proposals to introduce more equity into the pension
system by limiting seniority pensions and raises for the highest public old-
age pensions. In 1992, the assistant secretary general of the FNP stated, “We
know that the current public pension system is characterized by injustice,
corporativism, waste, and unfair distribution, especially between public and
private sectors; pension spending is ungovernable. . . . The reform of public
pensions can no longer be delayed. It is necessary to rationalize, reorganize,
and overcome imbalances and unfair inequalities” (Noseda 1992, 89).
In 1995, pensioners™ union members voted an overwhelming 91 percent
in favor of the proposed pension reform in the union referendum called by
the confederations to solicit the opinion of the membership (Baccaro 1999,
150). One could argue that the pensioners™ unions supported these reforms
because they placed most of the cost of reform on future pensioners. Indeed,
the end result of the negotiated reforms placed the heaviest burdens on cur-
rent workers with less than eighteen years of accumulated contributions.
However, many provisions did affect current pensioners, most notably min-
imum pension recipients and public servants about to retire on seniority
pensions. Contrary to the expectations of the gray power hypothesis, strong
pensioners™ groups in Italy have in many instances advocated reforms that
Old-Age Pensions

would reduce pension spending, whereas in the Netherlands rather weaker
pensioners™ parties have nevertheless successfully acted to protect pension
bene¬t levels even when other kinds of social bene¬ts were subject to cuts.

Political Competition and Program Structure
The level of the standard replacement rate and the existence of specialized
schemes with very high replacement rates are the key policy features that
account for differences in bene¬t levels between Italy and the Netherlands.
We™ve seen that neither differences in the strength of Social Democratic
actors nor the relative power of elderly lobbies can explain these policy
differences. But the structure of the welfare state and the nature of political
competition do affect these policies in important ways.
In the Netherlands a comprehensively enforced tax system that included
the self-employed as well as employees provided a ¬rm ¬scal basis for a
universal pension system, just as it allowed the development of universal
family allowance and unemployment policies. The Left and the Right were
able to agree on a pension system that covered both employees and the
self-employed, because employees and the self-employed both contributed:
employees through payroll deductions, the self-employed via the tax system
(Ferrera 1993, 169). And as we have seen, the universalistic nature of the
pension system impeded the growth of pension expenditures by establishing
a basic social minimum that prevented claims for bene¬ts expansion based
on criteria of need.
The primarily programmatic mode of political competition in the
Netherlands, combined with the removal of much negotiation over pen-
sions to the private second pillar, has provided little incentive to attempt
to upgrade bene¬ts for small constituencies. This in turn has allowed the
Netherlands to avoid the phenomenon of leapfrogging bene¬ts for different
groups that contributed to ever-increasing expenditures in the Italian case.
At the same time, the universal pension system provides few footholds for
those politicians who might wish to use it to pursue a more particularistic
mode of interest aggregation in the Netherlands. Finally, the ¬‚at-rate struc-
ture of the AOW means that projecting pension outlays is a simple matter.8

8 The legacy of expertise in economic forecasting left by Jan Tinbergen, the ¬rst Nobel
Prize winner in economics and director of the Dutch Central Planning Bureau during the
immediate postwar period, also added to the ease of pension forecasting in the Netherlands.
Comparable expertise did not exist in Italian economics departments at the time.

Age in the Welfare State

Thus it has always been obvious, even to actors who might have advocated
higher pension bene¬ts, what would be the consequences for other social
programs of raising pension bene¬ts. So the universalistic structure of the
Dutch pension system and the programmatic nature of political competi-
tion have reinforced each other and help to explain why the system estab-
lished immediately after World War II did not take on more particularistic
The opposite con¬guration, a tight bundling of particularistic politi-
cal competition and fragmented occupational program structure, is clearly
visible in the case of Italian pensions, where the highly complex and differ-
entiated public pension scheme and a clientelist mode of political competi-
tion reinforced each other in several ways. First, as with family allowances
and unemployment bene¬ts, Christian Democratic politicians™ particular-
istic use of the tax code made it impossible to introduce universal pension
bene¬ts at a guaranteed social minimum level, which could have helped to
contain costs. Second, the eagerness of politicians, particularly of the DC
and the PSI, to provide bene¬ts for key political supporters contributed to
the proliferation of special treatment for public sector employees, which
we have seen increased both the level of bene¬ts and the number of ben-
e¬ciaries in the pension system as a whole. Finally, the fragmentation of
the pension system impeded reform, by setting the stage for competition
between different sectoral groups and by hampering the ability of pension
administrators and other experts to formulate accurate projections. Let us
address each of these points in turn.

Particularism, Taxation and the Failure of Universalism The main
obstacle in Italy to implementing the universalistic pension proposals of
the early postwar period was the tax system. The reports of the D™Aragona
Commission (Commissione per la Riforma della Previdenza Sociale 1948)
and the CNEL reform commission (CNEL 1963a; 1963b) both cite the
impossibility of adequately assessing and collecting pension contributions,
especially among the self-employed, as reasons to continue providing pen-
sion bene¬ts on an occupational basis. Labor unions, too, which professed
support for the idea of universal coverage, feared that any universalization
of the system would be paid for out of increased payroll taxes on employ-
ees, and so repeatedly called for ¬scal reform as a prerequisite of pension
reform (see CISL Consiglio Generale 1950, 13; 1956, 150; 1958, 178; CGIL
Segreteria Generale 1962, 226; CGIL-CISL-UIL Segreteria Interconfed-
erale 1970, 201).
Old-Age Pensions

By the early 1960s, Italy™s pension system was categorized by two types
of funds: those running a surplus, primarily the FPLD and the special funds
for small groups, such as journalists; and those running large de¬cits, pri-
marily the funds for agricultural workers, artisans, and shopkeepers. The
latter funds had been set up in the late 1950s and early 1960s by DC gov-
ernments in order to bene¬t clienteles that were particularly important to
the DC (Regonini 1996, 90). These funds ran large de¬cits because con-
tribution rates were very low and access to bene¬ts was available even for
people with very limited contributory histories. While the state subsidized
these pensions out of general revenues to some degree, in large part it was
the employees™ funds that were asked to make up for shortfalls in a form of
“enforced solidarity” (Ferrera 1993, 262). By the late 1960s, both Con¬nd-
ustria and the unions were complaining about the increased payroll taxes
that had become necessary to support this burden, and union support for
universalizing the pension system waned.
The implementation of the social pension provision between 1965 and
1970 was the ¬nal nail in the cof¬n of the Left™s hope for a universal pension
system with an adequate social minimum. A tripartite agreement in 1964
established this new bene¬t for persons over sixty-¬ve without other means
of support, and it was agreed that bene¬ts should be ¬nanced out of state
revenues deposited in a new Social Fund. It was also agreed that the level of
the bene¬t would be suf¬cient to provide for a “decent” standard of living,
as stipulated in Article 38 of the 1947 Constitution. By 1968, however,
both Con¬ndustria and the labor unions had lodged complaints that state
contributions to the Social Fund were inadequate to cover its costs, and
the fund was instead drawing resources from the FPLD. At this point it
had become clear that social pensions , which constituted the bene¬ts ¬‚oor
in the Italian system, would have to be paid out of employees™ pockets if
they were to be paid at all, and union support for upgrading the level of
the bene¬t to an adequate social minimum faded (Ferrera 1993, 262). As
a result of the state™s inability to ¬nance a universal social minimum out
of general revenues, the safety net that was implemented for retirees in
Italy remained at a level so low as to constitute a meaningless support. This
absence of a social minimum and the continued fragmented, occupational
nature of the Italian pension system contributed, as we have seen, to Italy™s
very high pension expenditures in the 1980s and 1990s.
The reluctance of successive administrations to execute the tax laws con-
tinued even after the de¬nitive defeat of proposals to establish a univer-
sal pension system with an adequate social minimum. Evasion of pension
Age in the Welfare State

contributions for employees continued to be an important issue from the
late 1970s onward. In 1978 the CGIL and the PCI began to complain quite
insistently about the failure of employers to make contributions on behalf
of their employees. This practice contributed to the de¬cits that had begun
to plague even the stronger funds administered by INPS. A 1978 union pro-
posal to make evasion of payments a criminal offense was defeated by the
DC government “ in part because it would have meant absorbing the sys-
tem of occupational injury compensation into INPS, which was controlled
by the unions, and thus would make unavailable a key source of patron-
age for the government (Regini and Regonini 1981, 233). A 1980 study
commissioned by the Labor Inspectorate, INPS, and unions revealed that
in the 10,074 ¬rms studied, there was evasion of payment for 42 percent
of employees (Regonini 1984, 106). A report by an independent watchdog
group noted that evasion of contributions was facilitated by the very small
proportion of the work force of either the Labor Inspectorate or INPS
devoted to enforcement (CENSIS 1983, 237). “Administrative incentives”
for evasion that emanated from the government bureaucracy fostered a lack
of compliance with the tax code (Regonini 1984, 106).
Both the tax laws (failure to adequately tax key clienteles of the DC)
and their implementation (failure to enforce those laws that were in place)
sprang from the particularistic mode of political competition pursued by
politicians of Italy™s leading political parties during the postwar period.
The inability to collect revenues from key economic sectors in turn made
it impossible to universalize the pension system and provide an adequate
social minimum without placing an excessive burden on employees. This
constellation of problems radically altered the Italian Left™s preferences
with regard to universalization and bene¬t levels. And, most importantly
for the future of pension expenditures in Italy, the continued fragmentation
characteristic of Italy™s occupationalist pension regime led to a system in
which reform became nearly impossible for a period of almost thirty years.

The Pension System as a By-product of Clientelism If particularistic
political practices prevented the implementation of a tax system capable
of sustaining a political coalition for universal pensions, clientelism also
had more direct effects on the pension system. In particular, the extreme
generosity of pension provisions for public sector workers, which as we
have seen contributed both to high bene¬t levels (via the absence of a ceil-
ing on bene¬ts) and large numbers of pensioners (via generous seniority

Old-Age Pensions

provisions), can be seen as a direct outgrowth of the competitive strategies
of the DC, PSI, and PCI.
To most observers, Italy™s proliferation of pension provisions, each with
its own bene¬t formula, contribution rate, degree of state subsidy, rules
governing retirement age, years of service required to enter into the plan,
and so on, constitute proof positive that the pension system has been used
as a way to attract support from particular groups in the population (Regini
1981; Regini and Regonini 1981; Ferrera 1984; Paci 1984; Maestri 1994).
So does the practice of staf¬ng pension agencies with party supporters and
of deciding pension claims based on an applicant™s party af¬liation (Regonini
1996, 90“1). This perception is not limited to the scholarly community, of
course. Even the undersecretary of labor, in the midst of the 1982 pension
debate, dispatched a telegram accusing two representatives of the major-
ity coalition of “rampant particularism for pernicious electoral reasons”
(quoted in Regonini 1984, 108).
Christian Democratic politicians advocated extending pension bene¬ts
to the self-employed on very generous terms during the 1950s and 1960s
as part of a strategy to purchase loyalty from these groups. But during the
postwar period, state employees have been the most important target of
clientelist pension legislation, for obvious reasons. As the staff of public
and quasi-public organizations came to be dominated by supporters of the
governing parties, public sector pension bene¬ts took on special cachet
with politicians associated with these parties. Special provisions for pub-
lic sector employees reach back into the Fascist period (Cherubini 1977;
Paci 1984). During the postwar era, however, special provisions for pub-
lic sector workers were defended and extended by parties in government,
as a fundamental aspect of their electoral strategy. Not only have public
sector employees been the recipients of the largest volume of pension leg-
islation during the period 1948“83, but attention to this sector has come
predominantly from DC and PSI lawmakers (parliamentarians and govern-
ment of¬cials; Maestri 1994). The extremely generous pension provisions
enjoyed by public sector employees, provisions that have contributed in
no small way to the development of high pension expenditures in Italy,
can be quite directly attributed to the particularistic competitive strategies
employed by politicians of the ruling parties in Italy during the postwar
What is perhaps less widely appreciated is the extent to which the clien-
telist behavior of the DC (and later PSI) drew parties of the opposition

Age in the Welfare State

into a mode of pension policy making that has exacerbated the expenditure
problem. In the 1950s and 1960s, the Left was opposed to the DC™s prac-
tice of extending pension rights to new groups without making these new
groups responsible for ¬nancing them. However, the Left was too weak
to block this practice (Regini 1981, 122). But even after 1969, when they
had more control over pension legislation, unions opposed moves to uni-
versalize the pension system in part because they feared that a state-run
system would bring their constituencies into the clientelist orbit of the DC,
as had happened earlier with agricultural employees (127). Instead, they
followed a strategy of attempting to upgrade bene¬ts for their constituen-
cies to the level enjoyed by public sector employees (Baccaro 1999). In the
words of Ferrera (1993, 267), “The biggest novelty of [the reform of ] 1969
in a political sense was . . . the enlargement of the spoils system to include
the PCI and unions, opening the way to that ˜assistential grand coalition™
responsible for the profound imbalances that characterized the Italian wel-
fare state in the years to come.” The Left and the unions tried to match
every gain made by public sector workers (and other privileged clients of the
ruling parties), and to the extent that they succeeded, in turn the privileged
clients demanded more privileges. This led to an upward spiral of bene¬t
levels, and a downward spiral of contributory requirements, in the 1970s
and 1980s. These developments are attributable in the ¬rst instance to the
particularistic mode of competition engaged in by the DC and the PSI. But
the strategy of the opposition parties and of labor unions has been condi-
tioned by this mode of competition among the dominant parties, such that
the Left ultimately shares responsibility for the expansion of the pension
sector at the expense of other social spending.

Fragmentation and the Dif¬culty of Reform The ratcheting up of ben-
e¬ts and privileges in the Italian pension system is a by-product of partic-
ularistic political competition, but also of pension system fragmentation.
This dynamic is much less likely to occur in the absence of multiple differ-
ent regimes for different occupational categories, since it would require an
audacious ¬rst-mover to break the status quo of bene¬ts equality. But the
high degree of fragmentation in Italy™s pension system has made it dif¬cult
to achieve reform for other reasons, as well.
First, the pre-existing occupational system contributed to the demise
of universalistic reform proposals in the 1940s“60s. In principle the Ital-
ian Left, particularly the Center-Left, supported a universalistic pension
scheme. But in a context of particularistic behavior by the ruling DC, unions
Old-Age Pensions

had a strong incentive to maintain the pre-existing occupational funds, over
which they had some modicum of control, rather than giving everything
over to a central, DC-run universal scheme (Ferrera 1993). Once again
the interplay of clientelism and fragmentation affects the Left™s preferences
and behavior, with the opposition parties and unions coming to support
particularistic fragmentation under the threat of an even more damaging
particularistic universalism.
Fragmentation further impinges on prospects for reform because it
makes forecasting expenditures more dif¬cult. In one of the earlier pro-
jections of pension expenditures in Italy, the 1963 CNEL report predicted
that pension spending would reach 56 percent of total social spending by
1980, an amount that the commission considered to be too high. But this
projection failed to take into account the potential for explosive growth of
the pension sector inherent in the existing, segmented regime (Franco 2000,
18“19). Even apart from the political dynamics of reform that a highly frag-
mented system generates, it has proved remarkably dif¬cult to arrive at any
agreement about future pension liabilities based solely on economic con-
siderations. At a 2000 conference convened by the research arm of INPS,
one employee of INPS reported that the complexity of the pension system
in Italy was such that the bureaucracy lacked the econometric tools neces-
sary to forecast pension expenditures accurately. In fact, the ¬rst long-term
forecasts of pension expenditures in Italy date from the late 1970s (Franco
2000, 12). Competing studies from the 1980s (Franco and Morcaldo 1986;
Alvaro, Pedull´ , and Ricci 1987; Ministero del Tesoro 1988; INPS 1989)
resulted in widely varying forecasts, with the most optimistic scenarios sug-
gesting little need for reform. Under such circumstances of uncertainty, it
is not surprising that it has been dif¬cult to make a strong case for reform.
The high levels of fragmentation and differentiation in the Italian pension
system, a feature closely allied to the particularistic nature of political com-
petition in that country, have made it dif¬cult to see who bene¬ts and who
loses from the current system, and thus to generate a viable coalition for

This chapter has argued that differences in pension spending, an impor-
tant determinant of the age orientation of social spending in Italy and
the Netherlands, are a product of differences in both the structure of
pension programs and the type of political appeals that politicians make
Age in the Welfare State

to the electorate in these two countries. Contrary to the standard liter-
ature on comparative social policy, I ¬nd that the level of spending on
pensions in Italy and the Netherlands is not reducible to differences in
the power resources of either organized labor and the Left or the elderly
A number of speci¬c policy choices have contributed to high pension
spending in Italy and lower expenditures in the Netherlands. The estab-
lishment of an adequate social minimum, the presence of a second pillar of
private occupational pensions, limits on maximum pension bene¬ts, and a
relatively high retirement age have contained pension costs in the Nether-
lands. At the same time, weak income guarantees for the poorest elderly,
unlimited pension bene¬ts at the high end, and a very low retirement age
have contributed to extremely high pension expenditures in Italy. Analy-
sis of the role of left-leaning and pro-elderly actors in the formation of
these policies reveals that neither the power resources of the Left nor so-
called gray power can explain divergent spending patterns in Italy and the
Netherlands. Rather, these developments provide support for the political-
institutional argument developed in chapter 3.
If a universalistic pension program and programmatic political competi-
tion in the Netherlands serve as a kind of null hypothesis, the development
of pension spending in Italy provides a clear illustration of how the interac-
tion of particularistic political competition and occupationalist program
structure leads to high levels of spending on the elderly. The citizenship-
based pension program in the Netherlands, a program built on the base of
a comprehensive tax system and a politics of programmatic appeals, keeps
spending down because uniform bene¬ts for all citizens make it easy to
calculate the future costs of raising pensions. This universal system also
discourages politicians from using pension bene¬ts as a currency of political
exchange, since it is impossible to raise pensions for one segment of the
electorate alone. This in turn prevents the ratcheting up of pension bene¬ts
for different sectors that proved so expensive in the Italian case. Italy™s
fragmented occupational pension system, by contrast, provides crucial
resources for clientelist politicians, whose opposition to reforming such a
system in turn provides occupationalism with a means of self-propagation.
In addition, the combination of fragmented occupationalism and political
particularism also alters the incentives and behavior even of political
actors whose normal mode of operation might be more programmatic,
such that eventually they, too, come to support the fragmented pension

Old-Age Pensions

system status quo. Both the universalism-programmatic politics dyad and
the occupationalism-particularistic politics dyad are resilient and self-
reinforcing, properties that explain the ability of these political-institutional
characteristics of welfare states to shape the ¬‚ow of bene¬ts to different
constituencies over long periods of time.



This book has sought to elucidate how and why social policies in the rich
democracies vary in the way that they treat older and younger members
of society. Yet in the course of devising a strategy for measuring the age
orientation of social policies, testing alternative theories, and elaborating
mechanisms through the use of case studies, less attention has been paid
to the question of why, after all, the “age” of welfare matters. This ¬nal
chapter, then, explores the implications of the book™s ¬ndings about the age
orientation of welfare states for the well-being of different age groups and
for scholarship about the welfare state.

Age Orientation, Poverty, and Inequality
How does the age orientation of welfare states contribute to the well-being
of different groups in the population? We might think ¬rst of the wel-
fare state™s capacity to reduce the incidence of poverty among children,
working-age adults, or the elderly. It seems reasonable to assume that, other
things being equal, elderly-oriented welfare states would do a better job at
reducing poverty among the elderly than among non-elderly adults and
children. On the other hand, in relatively youth-oriented welfare states,
which in fact merely spend roughly equally on the old and the young, the
poverty reduction due to taxes and transfers should be more equal across age
This proposition is in theory testable using cross-nationally comparable
household-level data on income from the market and from social programs.
Such data are available from the Luxembourg Income Study (LIS) project.
Unfortunately, in practice we can do little more than speculate with these


data. Measuring the amount of poverty reduction that a welfare state carries
out depends on having reliable household-level data on income both before
and after taxes and transfers are taken into account. But recall that opaque
administration of tax systems is a prime resource for clientelist politicians,
who are concentrated in the elderly-oriented welfare states. As a result,
many countries characterized by particularistic political competition either
do not participate in LIS (e.g., Japan, Greece, Portugal) or do not pro-
vide LIS with pretax income data (e.g., Italy, Belgium), making it impos-
sible to estimate the poverty reduction carried out by the welfare states in
these countries. It is of course telling that the problems in the data re¬‚ect
the causal mechanism hypothesized to underlie the very outcomes that
interest us.
Eliminating many of the most elderly-oriented countries from consid-
eration due to a lack of data also makes it harder to examine the effects of
age orientation on poverty even in those countries for which pretax income
data are available, by reducing the already small number of cases. Poverty
and poverty reduction are complex phenomena, and the weight of any one
potential cause (e.g., age orientation) must be considered in light of other
attributes of welfare states and labor markets that may also affect income
distribution and poverty. The amount of poverty reduction carried out by a
welfare state is likely to be related to the amount of poverty before taxes and
transfers, as well as to the size of the welfare state and the level of inequal-
ity generated by the labor market, not to mention the political, societal,
and economic variables that are the root causes of the intermediate-level
phenomena. Without these controls it is impossible to estimate the impact
of age orientation per se. But once we remove Italy, Belgium, and other
countries for which there is no pretax data from the analysis, we have LIS
data to work with for only eleven (mostly youth-oriented) countries “ not
enough to allow for any statistical control.
What we can say, then, about how age orientation affects the ability of
welfare states to reduce poverty in different age groups is limited to what
the raw data tell us from those countries where information is available.
A very low level of aggregate spending seems to limit the quality of out-
comes for the very young regardless of how much or how little is spent on
other age groups. None of the smallest welfare states (the United States,
Australia, Ireland, Canada) is able to substantially reduce poverty levels in
families with children (Fig. 7.1). At the same time, there is a great deal
of variation in how much even these small welfare states reduce poverty

Age in the Welfare State

Poverty reduction: children







1 15
Social expenditure
as a percentage of GDP

Figure 7.1 Welfare state spending and poverty reduction: children. Note: Poverty
reduction is percentage change in poverty rates pre- and post-taxes and transfers.
Poverty rate is percentage of children aged 18 and under and not household heads
living in households with income of less than 50% of size-adjusted median. Sources:
Spending: OECD 2004; poverty: author™s calculations from LIS (Wave IV data).

among elderly and non-elderly adults (Figs. 7.2 and 7.3). Relatively youth-
oriented Ireland does much better for non-elderly adults than any of the
other residualist welfare states, and Australia and Canada do better than the
quite elderly-oriented United States. Size matters “ but at the margin, so
does age orientation.
The age orientation of welfare states may also affect the degree of
inequality among different age groups. Clearly, elderly-oriented welfare
states are not likely to reduce inequality much in any age group, despite rel-
atively generous spending on the elderly. (The occupational programs that
generate high aggregate spending on the elderly are designed to preserve
status differentials, so they tend to reproduce labor market inequalities.)
But we lack the pretax income data that would allow us to see whether
most elderly-oriented welfare states differ systematically (and net of other
factors) from more youth-oriented ones in the way that they mitigate mar-
ket inequalities among different age groups. Another revealing and still
more data-intensive test of the effects of age orientation on inequality
would come from examining inequality within cohorts over time. Does a

Poverty reduction: non--elderly










1 15
Social expenditure
as a percentage of GDP

Figure 7.2 Welfare state spending and poverty reduction: non-elderly adults.
Note: Poverty reduction is percentage change in poverty rates pre- and post-taxes
and transfers. Poverty rate is percentage of adults aged 18“64 living in households
with income of less than 50% of size-adjusted median. Sources: Spending: OECD
2004; poverty: author™s calculations from LIS (Wave IV data).

youth-oriented welfare state reduce inequality over time? If so, does all
youth-oriented spending have this effect, or only some? Is the cost over
a cohort™s lifetime greater or less than if the same degree of reduction
in inequality were achieved through transfers to older age groups alone?
These answers are surely worth knowing, as welfare states in the industri-
alized countries shift their budget allocations across different age groups
and different types of policies (e.g., active vs. passive labor market policies,
services vs. transfers).
The results of elderly-oriented social spending for the capacity of welfare
states to perform arguably their most important function “ protecting vul-
nerable outsiders “ are pernicious. In the account presented here, the quality
of political life emerges as a key determinant of the quality of social bene¬ts
and ultimately social welfare, echoing Marshall™s (1950) linkage of politi-
cal and social citizenship. Where programmatic party competition prevails,
new social programs can come forward to meet the emerging social needs
of adults and children struggling to balance work and caring responsibili-
ties in a changing labor market, and can constrain the otherwise powerful
Age in the Welfare State

Poverty reduction: elderly










Social expenditure
as a percentage of GDP

Figure 7.3 Welfare state spending and poverty reduction: elderly adults. Notes:
Poverty reduction is percentage change in poverty rates pre- and post-taxes and
transfers. Poverty rate is percentage of adults aged 65 and older living in households
with income of less than 50% of size-adjusted median. Sources: Spending: OECD
2004; poverty: author™s calculations from LIS (Wave IV data).

budgetary expansionism of social bene¬ts for protected core workers and
pensioners. In settings where particularism prevails, however, bene¬ts are
concentrated on a relatively small group of privileged, aging insiders, while
the growing mass of outsiders is left to fend for itself. Let us review how
and why this comes to pass.

Program Structure, Political Competition,
and the “Age” of Welfare
This book has argued that the age orientation of welfare is a largely unin-
tended consequence of the structure of social programs and the mode of
political competition in which politicians engage. Early choices about the
structure of social programs are reinforced by the way that politicians use
these programs to compete for votes, and give rise to welfare states that
treat the old and the young quite differently. The causes normally adduced
to explain differences among welfare states “ the economic circumstances
facing welfare states, the political power of Social Democrats or Christian

Democrats, the in¬‚uence of welfare state constituencies such as senior cit-
izens “ tell us rather little about why some countries spend more on their
elderly or their children than others do.
Welfare state programs organized in different ways and put in place dur-
ing a ¬rst critical juncture in the early twentieth century mature into wel-
fare states that privilege different age groups. But in order for welfare state
structures selected at this ¬rst critical juncture to “stick” for long enough to
affect the age orientation of welfare spending in the 1990s, initial choices
about the structure of social programs must be continually reinforced. The
competitive environment within which politicians operated during a sec-
ond critical period after World War II supplies just such a mechanism of
The countries that in the 1980s and 1990s had the most elderly-oriented
social spending “ Greece, Japan, Italy, the United States, Spain, and
Austria “ were welfare states that entered World War II with occupa-
tional social programs and never, or only recently, added a signi¬cant layer
of citizenship-based bene¬ts. Other countries that had occupational wel-
fare states after World War II “ the Netherlands, Germany, and France “
were able to develop more youth-oriented citizenship-based programs, to
“switch tracks,” as I termed it in chapter 3. What accounts for the different
pathways pursued by these two groups of countries following the second
critical juncture? I argue that above all the path to highly elderly-oriented
social policies versus a more moderate age orientation is determined by
the way in which politicians use social bene¬ts and other policies, such as
taxation and labor market policies, to compete with one another.
In countries where politicians competed in an environment of particular-
istic politics and clientelism, fragmented occupational program structures
provided critical resources for politicians, and were thus never abolished.
At the same time, the attachment of clientelist politicians to particularistic
administration of taxes, social security, and labor market regulation made
universalistic social programs unpalatable even for leftist politicians, who
were, in most countries, less inclined to compete along particularistic lines
for reasons of ideology. In countries where political competition occurred
along primarily programmatic lines, it was easier to introduce citizenship-
based programs because politicians were less tempted to undermine tax sys-
tems in order to reward self-employed voters. They were also less tempted
to tailor existing fragmented occupational social programs to appeal to
micro-clienteles, and thus less devoted to the preservation of occupational
program structures.
Age in the Welfare State

The original choice to organize social programs along either citizenship-
based or occupational lines had a lasting in¬‚uence on the age orientation of
social policies in a wide range of industrialized countries. But this choice was
not necessarily a permanent one. At key moments such institutional choices
need to be reaf¬rmed. In the welfare states of Europe, North America, and
the Paci¬c, the predominant mode of political competition in the period
following World War II served as the backdrop against which institutional
structures were either reaf¬rmed or renegotiated. The mode of political
competition, programmatic or particularistic, thus has important conse-
quences for the age orientation of social spending, channeling as it does
the choices politicians make about how to structure and distribute social
welfare bene¬ts.
The development of family allowance policies, unemployment-related
bene¬ts, and old-age pensions in Italy and the Netherlands illustrates how
the structure of social programs and the competitive behavior of politi-
cians interacted to produce an elderly-oriented welfare state in Italy and
more youth-oriented spending in the Netherlands. Italy™s elderly-oriented
welfare state is characterized by fragmented occupational social programs
that do very little for children and working-age adults. At the same time,
Italian pension spending is lavish (even if the distribution of this spending
means that many elderly are left without adequate resources). The relatively
youth-oriented Dutch welfare state has universal citizenship-based bene¬ts
that are quite generous for children and working-age adults, and moder-
ate public pension spending that is supplemented by an equally moderate
private occupational pension system. These very different constellations
of policy outcomes can be explained by the mutually reinforcing dynam-
ics of social program development and political competition in the two
Britain™s victory in World War II ensured that the Beveridgean model of
social protection would become the archetype of the modern welfare state.
Universal citizenship-based programs such as those in the Netherlands
became the stock-in-trade of programmatically oriented politicians seeking
to make their mark by providing public goods, rather than private bene¬ts.
Programmatically inspired political competition in the Netherlands in the
postwar period thus spurred the development of universal social programs
to complement or replace pre-existing occupational ones. Politicians in Italy
also sought to use the programs of the welfare state to generate electoral
support. Much as in the Netherlands, Italy™s moderate Communist Party
hoped to consolidate its working- and middle-class base by offering a vision

of a new, universal, citizenship-based system. But Italian Christian Demo-
cratic (and later Socialist) politicians™ strategic use of the welfare state, tax
system, and public employment service soon made the goal of universal-
ism seem less possible, and less desirable, to the mainstream Communist
When politicians in the Netherlands used the welfare state to “buy”
votes during periods of intense electoral competition, the universalization
of bene¬ts such as family allowances and pensions was one result. But in
Italy, when social programs became the currency of electoral competition,
occupational programs became more fragmented and more entrenched.
The presence of a large group of powerful politicians competing in a clien-
telistic mode explains the persistence of fragmented occupational family
allowances, unemployment bene¬ts, and pensions in Italy during a period
when many other countries of Europe were moving toward a more univer-
salistic conception of the welfare state.
The earliest traces of clientelism followed in this book are to be found
not in the welfare state itself but in the tax systems that underlie social
programs. Clientelist domination of the legislature and the public adminis-
tration in Italy in the early postwar years impinged on the development of a
well-functioning tax system. Politicians and tax collectors looked the other
way as valuable electoral constituencies such as the self-employed failed to
report income or pay taxes on the income they did report. A fragmented
and highly complex tax system eventually rei¬ed many of the special priv-
ileges granted in practice to valuable friends and allies of the Christian
Democratic Party. The result was a tax system in Italy that could not sup-
port, either ¬nancially or politically, the weight of citizenship-based social
Political coalitions advocating universal family allowances, unemploy-
ment bene¬ts, and old-age pensions unraveled in Italy by the 1970s as
it became clear that universalizing social insurance would mean adding
bene¬ts for the self-employed that would be ¬nanced by taxing industrial
workers. Clientelist use of the tax system had made universal social pro-
grams impossible in Italy. In the Netherlands, quite the opposite occurred:
there, a capable ¬scal administration was the precondition for agreements
that extended occupational family allowances, unemployment bene¬ts, and
old-age pensions into a full-¬‚edged safety net entitling the self-employed
and non-employed to the same bene¬ts as employees.
Clientelist use of the tax system ensured that occupational social pro-
grams in Italy could not be replaced. But these programs, which provided
Age in the Welfare State

different levels and types of bene¬ts for different groups of workers, were
also a gold mine for politicians who used particularistic strategies to com-
pete for votes and win elections. Multiple, differentiated bene¬t categories
within a single program “ such as Italy™s scores of separate public pen-
sion funds or multiple different programs providing cash bene¬ts for the
unemployed “ are not just ex post facto evidence that politicians used the
welfare state to target bene¬ts to small groups of voters. The existence of
such fragmentation also made it easier to justify new forms of discretionary
targeting of bene¬ts like a better replacement rate here, or a shorter refer-
ence income period there. As the fragmentation of the Italian welfare state
increased, the very complexity and opacity of what Italian commentators
have come to call “micro-corporativism” in turn protected the politicians
who engaged in it. A thicket of highly specialized provisions has made it
dif¬cult for the public (and sometimes even for policy makers) to know
when changes have occurred, and even harder for them to understand what
the consequences of such changes might be for the public interest. For all
of these reasons, politicians who compete using clientelism have been loath
to see occupational fragmentation overturned, or even reformed.
Occupationalism stuck in Italy for two main reasons. First, clientelist
political competition made it dif¬cult to develop the neutral state capacities,
like strong tax systems or functioning labor exchanges, that are necessary
to make universal social programs politically and ¬nancially viable. Second,
occupational welfare programs themselves provided valuable resources for
clientelist politicians. In the Netherlands, on the other hand, neutral state
capacities provided strong foundations for universal, citizenship-based pro-
grams. These programs, in their transparency and lack of differentiated ben-
e¬ts, both re¬‚ected and encouraged programmatic political competition.
The distinction between occupational and citizenship-based social pro-
grams also affected the demand side of social policy making in important, if
less obvious, ways. Prior to the 1960s, bene¬ts for children and working-
age adults were, in Continental Europe, typically cast as wage supplements.
Family allowances were put in place to make up for wage restraint or
reduced hours; unemployment bene¬ts were extended to those excluded
from the labor market on a long-term basis only in cases of emergency
or restructuring. When such bene¬ts become a part of the apparatus of a
citizenship-based welfare system, though, as they did in the Netherlands
in the 1960s, they take on a different meaning. No longer simply wage
supplements, in the Netherlands family allowances and long-term unem-
ployment bene¬ts grew into full-¬‚edged entitlements, with accompanying

expectations about the appropriate level of the bene¬t.1 In Italy, however,
where family allowances and bene¬ts for the unemployed remained occu-
pational in nature, bene¬t levels were allowed to drift downward as wages
and standards of living grew. Non-indexation of unemployment bene¬ts
and of family allowances in Italy doubly doomed those programs: as long
as wages were rising, few people noticed that the bene¬t levels were falling,
and once the bene¬t had shrunk to insigni¬cance, very few people cared to
spend political capital defending them.
Universal programs for the non-elderly grew quite rapidly in the
Netherlands because, once divorced from wages, these bene¬ts turned into
expensive individual entitlements. At the same time, their undifferentiated
structure meant that if politicians used these programs in an attempt to cap-
ture the votes of any group interested in the program, bene¬t levels would
increase across the board. A raise for one is a raise for all in a citizenship-
based social program. Occupational programs in Italy, on the other hand,
could remain modest and yet still provide valuable currency to politicians.
Even cash rewards too small to be much more than symbolic could be used
to secure votes, as long as there were people who were not getting anything
at all. This was an important reason why the Left came to oppose modest
cash bene¬ts for the youth unemployed in Italy.
If even small youth-oriented bene¬ts can be useful for clientelist politi-
cians in an unsaturated marketplace, large and highly salient bene¬ts such as
old-age pensions take on a life of their own. Differently situated constituen-
cies press for ever better bene¬ts, and the lack of transparency inherent to
fragmented occupational regimes makes it possible for politicians to provide
without invoking the public™s wrath. This situation has made Italy™s pension
system extraordinarily resistant to change “ more resistant, I would argue,
than a policy feedback model based simply on blame-avoiding politicians
would suggest.
In universal, citizenship-based pension systems such as the Netherlands™,
on the other hand, private or supplementary occupational pensions often
act as a release valve for pressures to increase pensions. This safeguard is
necessary since the budgetary consequences of increasing bene¬t levels in
a universal program that provides even a modest income for retirement are
so visible. Universal bene¬ts that by their nature are relatively large “ that

1 Bussemaker (1992) has usefully discussed this transformation with reference to the “indi-
vidualization” “ extension to women as an independent entitlement “ of bene¬ts previously
reserved for male breadwinners.

Age in the Welfare State

is, provide more than temporary or partial income support for large groups
of people “ thus are more resistant to expansionary politicking than either
fragmented occupational bene¬ts or smaller citizenship-based bene¬ts such
as family allowances or unemployment bene¬ts.
Clientelist political competition in Italy locked into place pre-existing
occupational programs, resulting in declining bene¬ts for the non-elderly
and an explosion of pension spending. In the Netherlands, citizenship-based
programs grew on a base of neutral state capacities provided by program-
matic political competition, and in turn reinforced the tendency in that
system to compete along programmatic lines. The development of social
spending in different areas (strong growth in the area of new entitlements
of the non-elderly, and more cautious growth in the larger old-age pension
program) resulted in a relatively youth-oriented welfare state.

Implications for the Study of Welfare State Politics
The argument presented in this book highlights two features of welfare
states that have until now received very little attention in the literature on
comparative social policy: (1) how the structure of welfare state programs
interacts over time with demographic and labor market shifts and (2) the
use that politicians make of such programs in their competitive battles with
one another. But the explanation for why welfare states differ in their age
orientation is perhaps most surprising because it has so little to do with
age. The political power of age-based political actors and the beliefs they
are presumed to carry with them about what is a just distribution across the
life course play far less of a role in determining the “age” of welfare than
one might expect.
At the margins, pressure groups representing both elderly and non-
elderly constituencies (pensioners, families with children, and the unem-
ployed) have certainly had an impact on social policies. And there is no
question that the age orientation of social policies has become a subject
of political contention for demographically de¬ned pressure groups. But
the age orientation of social policies should not be mistaken for a result of
these constituency demands. Rather, policy drift (see Hacker 2004) result-
ing from sticky institutions in a changing environment is responsible for
the outcomes we observe. It creates policies that bene¬t particular groups
and for which these groups now ¬ght, but which are actually the result of
long-term processes having little to do with the demands of welfare state

The interpretation of welfare state age orientations as a product of Chris-
tian Democracy falters on similar ground. Not only is the relationship
between Christian Democratic strength and age orientation inconsistent
(Italy, the Netherlands, Germany, and Ireland all have strong Christian
Democratic political actors, but their age orientations range across the
spectrum from rather youth-oriented to very elderly-oriented). Perhaps
more important, welfare state structures such as status-preserving occupa-
tional social insurance programs and family policies that reinforce patriar-
chal authority “ which accord with Catholic social doctrine and are often
assumed to be a result of the power of actors motivated by it “ are not
necessarily the result of Christian Democratic power. The thoroughness
with which Christian Democracy dominated the state in Italy makes that
country a hard case among the Christian Democratic welfare states for
this argument. But even in Italy, these classically “Christian Democratic”
welfare state structures appear more as spandrels (see Gould and Lewontin
1979), by-products of some other set of evolutionary processes, than as true
outcomes of Christian Democratic actors pursuing ideologically motivated
policy goals.
The power of Social Democracy proves to be an equally poor predic-
tor of the age orientation of welfare states. The egalitarian ideologies of
working-class actors do not necessarily favor age-neutral social spending.
The Left may defend either an aging core work force or younger labor
market outsiders, depending on dynamics internal to the organized labor
movement, on which policy solutions seem possible, and on the tactics of
the opposition. In our cases, we saw the Left™s social policy preferences
change depending on the competitive strategies of the Center-Right. It
would be impossible to pin variation in the age orientation of social spend-
ing to the strength of working-class ideologies and actors without knowing
much more about the political and institutional environment within which
they operate.
As with Christian Democracy and Social Democracy, the preferences of
employers are unlikely to be homogeneous enough across space and time to
account for variation in the age orientation of welfare states. Still, we might
expect employers in small, open economies to create and defend universal-
istic social policies, on the theory that these would lower employers™ direct
nonwage labor costs. This could, over the long run, have the unintended
consequence of creating more youth-oriented social policies in countries
whose economies are highly exposed to international markets. But not all
small, open economies opted for citizenship-based welfare states at our
Age in the Welfare State

¬rst critical juncture around the turn of the century (the Netherlands and
Ireland did not), and even after the second critical juncture, some small,
open economies (e.g., Belgium and Austria) maintained their occupational
welfare states. The presumed preferences of employers within this rela-
tively homogenous set of countries do not then account for the divergences
in welfare state structures that produce varying age orientations.
As this brief summary makes clear, the argument presented in this book
differs in some fundamental ways from traditional studies of comparative
social policy. In the ¬rst place, my argument focuses more on the sup-
ply side of welfare politics than do explanations that appeal to the logic
of constituency demand. While welfare state bene¬ciaries certainly have a
role to play in requesting or acquiescing to certain policy solutions, both
politicians and existing institutions are critical in shaping these demands.
Welfare states cannot be read simply as the revealed or congealed prefer-
ences of powerful constituency groups. But neither can the supply of social
policies be read simply as an expression of politicians™ power to enact ide-
ologically motivated or interest-based programmatic goals. What political
representatives of employers and the working class want, and what they
can get, affects the choice for universalism versus occupationalism at our
¬rst critical juncture. But we have seen that the competitive environment
in which politicians of all ideological stripes ¬nd themselves determines
crucially the kinds of social policies that they will support.
In this view, (some) institutions are the result of purposive action, but
not necessarily of rational choice in the sense of action that pursues opti-
mal policies given a certain set of ideologies or a certain demographic base.
Because (other) institutions constrain the preferences of political actors, it
is not possible to read their policy choices from their ideologies, their con-
stituencies, or their place in the productive structure. This is not simply to
say that political actors seek of¬ce or votes but not policies. Politicians are
policy seekers, but the policies they seek are not necessarily the ones that
inductive logic tells us they should prefer, in part because pre-existing insti-
tutions constrain their choices and make some options seem more appealing
than others.
Still, the unintended consequences of institutional design may look very
much like results of the ideologically motivated policy choices of powerful
political actors. Elderly-oriented welfare states look like the outcome of
successful lobbying by powerful senior interest groups. Age-neutral social
spending looks like a victory for the egalitarian principles of working-
class-based parties. But an important conclusion of this book is that such

appearances can be misleading. While political actors do come to have a
stake in these social policy institutions, they did not invent them. The age
orientation of welfare states is a result of formal policy institutions, kept
rigid by their bundling with informal political institutions and failing to
adjust to long-term processes of demographic and social change.
Deep structural con¬gurations such as these can coincide with and/or
conceal what scholars and lay observers alike more often view as the short-
to medium-term preferences of parties, politicians, or interest groups. This
capacity for masquerade is probably part of what gives some institutions
their resiliency. But if welfare state outcomes are marked by the unintended
consequences of institutional design, then it makes little sense to think about
the development of welfare states over the relatively longue dur´ e of the
postwar period purely in terms of purposive action guided by partisan goals
and ideologies. In fact, observing the process by which both institutions and
interests develop over the entire postwar period, rather than focusing on
the correlation between interests and institutions in recent years, highlights
a less actor-centered alternative set of causes of welfare state consequences
(see Pierson 2004).
The unintended consequences of institutional rigidities undoubtedly
play a larger role in structuring welfare state outcomes than much of the
previous literature has allowed. Policy drift allows old institutions and struc-
tures to generate new outcomes as the context within which they operate
changes. The age orientation of welfare states is an outgrowth of early
choices about welfare state structures, choices that were made without con-
cern for the shape of the labor market, public ¬nance, family structures, or
demographic trends one hundred years hence. Yet as the societal substrate
within which institutions are embedded changed, so too did the effects of
these institutions. It seems likely that other attributes of welfare states that
are also affected by the institutional form of social policies “ attributes
such as aggregate social spending or the extent to which welfare states
decommodify workers “ may also rely more than has often been recog-
nized on policy drift and the unintended consequences of earlier policy
If this is true, then neither the age orientation of welfare states nor some
of these other characteristics of welfare states that interest scholars should
be interpreted purely as offshoots of the standard con¬gurations of ideolog-
ical or power resource variables. To focus on pre-existing institutions and on
the prevailing political rules of the game forces us to consider the resources
that speci¬c contexts of competition confer on (or deny to) politicians,
Age in the Welfare State

as well as these actors™ ideologies and goals. Even when politicians are
ideologically committed to particular policy goals, they may eventually
press for other, sub-optimal policy solutions, as is vividly illustrated in
the studies of Italian policy development in chapters 4 to 6. Left-leaning
political actors in Italy repeatedly chose not to pursue the generous uni-
versal social bene¬ts that they had once advocated because the political
strategies of center and right politicians made other, second-best solutions
preferable. This is not to deny the importance of power resources or of
purposive action on the part of politicians and other policy makers, both
of which have undeniably contributed to the shape of welfare states as we
know them today. But all politicians must do their work within speci¬c con-
texts, only some of which permit them to choose policies that are optimal
from the standpoint of their ideological or organizational commitments.
This ¬nding illuminates an important but often overlooked characteristic
of the roughly one-half of polities in the advanced industrialized countries
where programmatic political competition is not the norm. Particularistic
political competition, even when it is the preferred style of a minority of
politicians, sends out ripples that affect the entire political system. This is
because clientelist behavior on the part of ruling politicians is infectious:
it informs not only the strategies of the clientelists, but also the strategies
and eventually policy preferences of opposition politicians. The clearest
example from this work is the way in which particularistic manipulation
of the tax system in Italy contributed to left politicians™ and union leaders™
decision to abandon the project of building a universalistic welfare state.
When even a small set of powerful politicians begins to shape institu-
tions in a way that optimizes their particularistic competitive strategies, it
forces other politicians to change their strategies and preferences as well.
This gives a ¬rst-mover advantage to politicians who build institutions,
because these institutions reinforce particular competitive strategies. And
it accounts for the persistent bundling of the mode of political competition
and the structure of welfare state programs. Sets of institutions become
path-dependent because of the increasing returns that each generates for
the other.

Possibilities for Institutional Change
The persistence of social policy institutions set in place in the late nineteenth
century and reinforced by different types of political competition creates,
unintentionally, patterns of social spending that generate inequalities in

well-being among different age groups in the late twentieth century. But
are these outcomes set in stone for all time? How does the “stickiness” of
these institutions affect the development of welfare states during the current
period, when policy makers have become more aware of how current demo-
graphic and labor market conditions combine with welfare state institutions
to produce sometimes undesirable spending patterns? What circumstances
might we expect to contribute to changing the long-established age orien-
tations of welfare states?
A new critical juncture could occur, signaling a dramatic shift in age
orientations, if some external force were to generate a change in either the
competitive behavior of politicians or the way social programs are orga-
nized. This would effectively decouple the two, allowing one to change
without requiring that the other give way at the same time. A few candidates
present themselves, especially in Europe, where the collapse of Commu-
nism in the East in 1989, the globalization of trade, and European monetary
union in 1992 have had far-reaching effects on social, economic, and polit-
ical institutions. So far, though, none of these exogenous shocks seems to
have generated enough momentum to shift the age orientation of social
spending meaningfully.
The fall of the Berlin Wall contributed to the overthrow of clientelist
political regimes in Italy and Austria in the early 1990s, as electorates that
had once tightly held their noses and voted for stability against the Commu-
nist threat began to consider other options. But while clientelist politicians
have passed out of of¬ce (temporarily, in many cases), it will take more time
to undo the effects of ¬fty years of particularistic public administration.
Left governments in Italy in the 1990s “ the ¬rst in the postwar period “
were ¬nally able to reform the tax system and lay the groundwork for tax
¬nancing of a variety of new citizenship-based social programs. But whether
these changes will weather alternation with a right that is still largely partic-
ularistic in orientation, and whether they will be enough to turn the ocean
liner that is the Italian pension system and free up resources for a more
age-balanced repertoire of social spending, remain to be seen.
The internationalization of trade might also generate pressure for a shift
in the age orientation of social spending. Growing international competi-
tion has increased employers™ sensitivity to nonwage labor costs, resulting
in a push from some Continental European employers to convert payroll-
¬nanced occupational bene¬ts to citizenship-based programs paid for out of
general revenues. (Rising health care costs and private pension fund insol-
vency have led some large employers in the United States to call for more
Age in the Welfare State

public welfare solutions, as well.) However, at the same time, governments
(particularly subnational governments) are looking for ways to shift costs
onto the social partners (see, e.g., Campbell and Morgan 2005), so the net
effect of these counterpressures may be no movement at all.
A third potential source of pressure for change in domestic institutions,
European monetary union, is probably the most likely candidate to gen-
erate a critical juncture signaling a new path toward substantial changes
in the age orientation of social spending. In particular, the criterion that
government budget de¬cits not exceed 3 percent of GDP generates pres-
sure both for tax reform and for spending cuts. Tax reform, as we have
seen, is a necessary (but not suf¬cient) condition for ¬nancing citizenship-
based social programs that would be more youth-oriented in nature. More
immediately, the Maastricht criteria demand ¬scal restraint even where the
public, unions, or employers might not be sensitive to the costs of social
spending. When the possibility of de¬cit ¬nancing of social welfare dis-
appears, spending begins to look more like a ¬nite pie. Under these new
circumstances welfare state politics are more likely to focus on the alloca-
tion of total social spending among different competing groups, including
different age groups. Indeed, in Italy the reframing of the “crisis of the
welfare state” as a crisis of intergenerational allocation coincided with that
country™s drive to join the monetary union.
In theory, there is no reason that a shift in age orientation must await a
cataclysmic change or a new critical juncture (Streeck and Thelen 2005).
If the age orientations observed in the 1990s are largely a result of how
social policy institutions “¬t” with a society™s demography and labor mar-
ket, new demographic and economic changes could cause a change in age
orientation as well. But societal changes of a magnitude capable of generat-
ing a new age orientation without concomitant institutional changes seem
unlikely to occur in the near future. Even with loosened immigration con-
trols and more generous family policies, populations in the rich democracies
will continue to age. And while the fortress labor markets of Continental
Europe are becoming increasingly open to younger and female workers,
these former labor market outsiders are often integrated on a part-time or
¬‚exible basis without access to the same social rights as the aging core work
force. So social policies are not likely to “drift” their way toward new age
Another potential source of change is a breakdown in the tight cou-
pling between the mode of political competition and the structure of social
programs. If dynamics internal to the welfare state “ some kind of policy

feedback “ were to break down the supportive relationship between clien-
telism and occupationalism, or between programmatic competition and
universal social bene¬ts, so that they no longer generated increasing returns
for one another, we might observe a shift in the age orientation of a welfare
Pierson (1994) and Campbell (2003) posit a type of policy feedback in
which welfare state programs create and de¬ne over time bene¬ciary groups
with the potential to act politically to preserve their bene¬ts when they are
threatened. This constituency-based feedback model leads us to expect that
in youth-oriented welfare states, all other things being equal, we would
see movements of single mothers, youth unemployed, families on assis-
tance bene¬ts, and so on, springing up to protect from retrenchment the
programs that bene¬t them.2 In elderly-oriented welfare regimes, on the
other hand, the absence of meaningful social spending programs geared
toward working-age adults and children should discourage the growth of
policy feedbacks in which groups of relatively youthful bene¬ciaries coa-
lesce to defend “their” programs from cuts. In fact, though, movements
of and/or for the elderly (pensioners™ parties, pensioners™ unions, elderly
advocacy groups) exist in all advanced industrialized countries. And while
the elderly are not responsible for the age orientation of social spending as it
has emerged in the twentieth century, there is good evidence to suggest that
organized gray power may actually help to limit spending on the elderly in
certain contexts (Campbell and Lynch 2000; Anderson and Lynch 2003).
The reform trajectories of welfare states with varying age orientations
do not appear to follow a pure constituency feedback model. It is true that
some of the most youth-oriented countries (Finland and Denmark) have
become even more youth-oriented since 1990, while some of the most
elderly-oriented (Japan and the United States) have moved even further in
that direction. The recent enactment of a prescription drug bene¬t for the
elderly in the United States while over forty million working-age adults and
children go without any health insurance, for example, pushes that country
toward an even more pronounced elderly orientation. However, the Dutch
and Italian case studies suggest that the age orientation of social spending

2 Of course, all other things are not equal, and the likelihood of such groups emerging prob-
ably depends not only on a variety of program characteristics independent of expenditures
(the degree of stigma associated with bene¬ts, the method of ¬nancing, the locus of admin-
istration, etc. “ see Ingram and Schneider 1993; Schneider and Ingram 1993; Soss 1999;
Mettler 2002) “ but also on characteristics of the broader political, economic, and social

Age in the Welfare State

may instead generate a different kind of policy feedback “ one based on
norms rather than voting blocs “ that impacts the politics of welfare state
reform in different ways.
The numerical and organizational strength of the elderly has been
important in the politics of welfare reform in the last decade in elderly-
oriented Italy, as a constituency-based feedback model would lead us to
expect. The elderly in Italy have indeed exercised a powerful constraining
role in pension system reform debates. But it is worth noting that since
the mid-1970s this role has been played predominantly by pensioners as
voters, and not by the pensioners™ parties (insigni¬cant and ephemeral)
or pensioners™ unions (large and signi¬cant, but surprisingly moderate in
their demands; Campbell and Lynch 2000; Anderson and Lynch 2003). Fur-
thermore, we saw in chapter 6 that Italian politicians have developed their
own distinctive interests in maintaining expensive pension provisions, quite
apart from demands arising independently from the elderly. Constituency-
based feedback models do not help us to distinguish in this case between the
possibly contradictory demands of organized interest groups as opposed to
voters, and they discount the importance of politicians™ independent inter-
est in particular programs.
From the point of view of the policy feedback literature, it is more sur-
prising that the elderly in the relatively youth-oriented Dutch welfare state
have also proven to be a powerful political lobby. In 1994, elderly organiza-
tions reacted to the Christian Democratic leadership™s threat to cut pensions
by forming two new pensioners™ parties and placing seven members in Par-
liament. As a result of this political strength, the elderly have been able to
contain cuts to their programs, including exempting pensions from freezes
at a time when other social bene¬ts were subject to austerity measures. In
the ¬eld of labor market policies, similarly, older workers have been much
less affected by retrenchments than have younger ones. At the same time,
sustained mobilization of younger welfare state bene¬ciaries was notably
absent in the 1990s, with the exception of some important protests against
disability insurance reform. How are we to understand this paradox?
Policy feedbacks in Italy and the Netherlands do not appear to depend
just on the nature of social programs, and hence on the size and shape of
the constituency groups that these programs create. Rather, the age pro¬le
of social spending seems to set the parameters for emerging debates over
intergenerational equity. In a pathway that echoes Levy™s (1999) vice-into-
virtue framework more closely than Weaver™s (1986) and Pierson™s (1994)
blame-avoidance scenario, reforms in the Netherlands and Italy are taking

place in the areas where they are arguably most needed, and not simply in
the areas where groups of welfare state bene¬ciaries are weakest.
In Italy since the mid-1990s, the educated elite have increasingly come to
blame the welfare state™s maladies on the “hyper-protected” elderly and on
the groups that in reality or in perception work on their behalf (the splinter
Communist party Rifondazione Comunista, which has assumed the musty
mantle of defender of the existing pension system, and the pensioners™
unions). These are the new villains in a country that many have come
to see as, in the words of former Prime Minister Massimo D™Alema, “a
society organized against the young” (Di Caro 1997). As a result, some
politically vulnerable categories of pensioners, especially women, have seen
cutbacks that affect them immediately, and bene¬ts for future pensioners
have been scaled back. At the same time, efforts to increase protections for
the young and for working-age adults in Italy have seen real results, even
at a time of ¬scal contraction. By contrast, the young in the Netherlands
are the ones cast in the role of the villain. Successful welfare state cuts in
the 1990s targeted “inactive,” “parasitic” working-aged adults, especially
young adults, while pensioners came out ahead.
The age orientation of social policies then seems to create a kind of
ideational policy feedback among elites, structuring how they perceive the
welfare state to administer intergenerational justice, and thus setting new
parameters for discussions about welfare retrenchment and reform. Under
what circumstances ideational feedbacks (a kind of backlash against the
existing age orientation of the welfare state) can take root among mass
publics is a question for further research. But there is reason to think
that ideational feedbacks may be particularly important in shaping political
struggles over the age orientation of a welfare state, where, to borrow strong
words from Albert Hirschman (1977), questions of justice and fairness may
move the passions nearly as much as interest does.


. 6
( 7)