COPING WITH PERMANENT AUSTERITY:

WELFARE STATE RESTRUCTURING IN AFFLUENT DEMOCRACIES

Paul Pierson
July 30th, 1999
This essay will appear as the concluding chapter for Paul Pierson, ed., The New
Politics of the Welfare State (Oxford: OUP), forthcoming. Thanks to Andrew Karch and Effi
Tomaras for research assistance. I have benefited from many conversations with participants
in the New Politics project and the European Forum at the EUI. I owe special thanks to John
Myles and Ann Orloff, although I suppose it would be unfair to hold them responsible for the
end result.
1
The welfare states of the affluent democracies now stand at the center of political discussion
and social conflict. Analysts frequently portray these conflicts as fundamental struggles
between supporters and opponents of the basic principles of the post-war social contract.
They often emphasize that the politics of social policy are played out against the backdrop of a
transformed global economy that has undercut the social and economic foundations of the
welfare state. While containing elements of truth, such portrayals distort crucial
characteristics of the contemporary politics of the welfare state. Changes in the global
economy are important, but it is primarily social and economic transformations occurring
within affluent democracies that produce pressures on mature welfare states. At the same
time, support for the welfare state remains widespread almost everywhere. In most countries,
there is little sign that the basic commitments to a mixed economy of welfare face a
fundamental political challenge. Nor is there much evidence of convergence towards some
neo-liberal orthodoxy.
Yet the chapters in this volume have also stressed that welfare states are undergoing quite
significant changes. In this conclusion, I argue that the contemporary politics of the welfare
state takes shape against a backdrop of both intense pressures for austerity and enduring
popularity. In this context, even strong supporters of the welfare state may come to
acknowledge the need for adjustment, and even severe critics may need to accept the political
realities of continuing popular enthusiasm for social provision. Thus in most of the affluent
democracies, the politics of social policy centers on the renegotiation, restructuring, and
modernization of the terms of the post-war social contract rather than its dismantling. The
crucial issue is whether particular national settings facilitate the emergence of such a centrist
reform effort, and if so, on what terms.
The argument proceeds in three stages. In the first, I outline a basic framework for studying
the politics of reform in a context of permanent austerity. In the second, I discuss two
complications: the need to incorporate different dimensions of social policy reform and the
need to recognize three quite distinct configurations of welfare state politics among the
affluent democracies. In the third, I apply these arguments to analyze the politics of
restructuring in the liberal, social democratic, and conservative worlds of welfare capitalism.
I. An Initial Framework for Studying the Politics of Permanent Austerity
1
Despite some disagreement concerning the main sources of pressure on mature welfare states,
the chapters in Part I all conclude that the welfare state now faces a context of essentially
permanent austerity. Changes in the global economy, the sharp slowdown in economic
growth, the maturation of governmental commitments, and population ageing all generate
considerable fiscal stress. There is little reason to expect these pressures to diminish over the
next few decades. If anything, they are likely to intensify.
Underlining the severe pressures confronting mature welfare states does not, however, imply
that the expected result is a collapse or radical retrenchment of national welfare states. Major
policy reform is a political process, dependent on the mobilization of political resources
sufficient to overcome organized opponents and other barriers to change. The welfare states
opponents have found it very difficult to generate and sustain this kind of political
mobilization.

1
This section draws on material presented in Pierson 1998.
2
I have developed this argument elsewhere (Pierson 1994; Pierson 1996) and present only a
condensed version here. The sources of the welfare states political strength are diverse, but
are of two basic types: the electoral incentives associated with programs which retain broad
and deep popular support and the institutional stickiness which further constrains the
possibilities for policy reform. Together, these features have created tremendous resilience in
the face of two decades of welfare state crisis (Stephens, Huber and Ray 1999).
Electoral Incentives. In market democracies, voters play a crucial role. Implementing and
sustaining policy reforms over time generally requires electoral vindication. Voters, however,
remain strongly attached to the welfare state. The broad public support consistently revealed
in polls stems from several sources. As the welfare state has expanded, so has the size of its
constituencies. As Peter Flora has noted, including recipients of [pensions,] unemployment
benefits and social assistance and the persons employed in education, health and the social
services, in many countries today almost of the electorate receive transfer or work income
from the welfare state (Flora 1989, p. 154). Figure 1 and Table 1 offer a very rough
indication of the political transformation associated with welfare state expansion. Figure 1
shows the dramatic impact of welfare state expansion from roughly 1950 to 1980. It tracks a
conservative measure of welfare state employment, plus a very conservative estimate of
transfer recipients, as a share of the total electorate for three representative countries. From
the beginning of the post-war expansion (1960) to the rough end of the maturation period
(1980) these ratios increase markedly in all cases. Table 1 provides snapshots from the mid-
1990s for the same countries. Again measured on a conservative basis, one can see both that
the core constituencies of the welfare state are very large, and that there are marked variations
across countries.
[Figure 1 and Table 1 About Here]
These crude indicators considerably underestimate the scale of the welfare states reach into
contemporary political life. Many voters who do not currently receive social benefits expect
that they may at some point in their lives. Or they may be in the same household with
someone who receives benefits from, or is employed by, the welfare state. In most countries,
universal programs, especially health care but also other social services, generate widespread
support. Furthermore, the welfare state retains considerable legitimacy as a source of social
stability and guarantor of basic rights of citizenship. Popular support generally appears to
extend well beyond the confines of narrow economic self-interest.
Support for the welfare state is intense as well as broad. Intensity of preference matters
because it is associated with higher rates of political mobilization and with voters actual
choices at election time. The intensity of support for the welfare state stems from two factors.
First, while the benefits of retrenchment for welfare state opponents are generally diffuse and
often uncertain, the large core constituencies for the welfare state have a concentrated interest
in the maintenance of social provision. Huge segments of the electorates of advanced
industrial societies rely on the welfare state for a large share of their income. It is one of the
few basic axioms of political science that concentrated interests will generally be advantaged
over diffuse ones.
The second source of intensity stems from the fact that welfare state supporters are in the
position of fighting to sustain already existing benefits. Students of electoral behavior and
3
political psychology have found that voters exhibit a negativity bias -- they react more
intensely to potential losses than to commensurate potential gains. Thus, the welfare states
electoral base is not only enormous, but primed to punish politicians for unpopular initiatives.
Institutional Stickiness. Those seeking policy reform must confront not only the potential
opposition of voters and program beneficiaries but the stickiness of existing policy
arrangements. By stickiness I have in mind two features of developed polities that reinforce
the electoral obstacles to radical reform: formal and informal institutional veto points, and
path dependent processes, which in many cases tend to lock existing policy arrangements
into place. Each of these characteristics pushes reform agendas in the direction of incremental
adjustments to existing arrangements.
The basic point about veto points is straightforward and clearly spelled out in Bonolis
contribution to this volume (see more broadly Scharpf 1986; Tsebelis 1995). Most political
systems make policy reform dependent on more than a simple 51% majority, allowing
minorities (including in some cases quite small ones) opportunities to block reforms.
Examples of such institutional arrangements include federalism, a strong judiciary,
bicameralism, use of referenda, requirements of super-majorities, and coalition-based
governments.
2
Difficult as it is to create a majority coalition in favor of restructuring the
welfare state, even that may not be enough. The multiplication of veto points can hamstring
efforts at policy change, frustrating even an ambitious and aggressive reform coalition such as
the Republican congressional majority in the United States after 1994, or an enduring
conservative coalition such as the one governing Germany from 1982 to 1998.
A second major source of stickiness, path dependence, is more complex. Because path
dependence is often invoked as an explanation without further explication, some elaboration is
necessary.
3
Certain courses of political development, once initiated, are hard to reverse. It is
not just that institutional veto points may make a reversal of course difficult. Individual and
organizational adaptations to previous arrangements may also make reversal unattractive.
Recent work on path-dependence has emphasized the ways in which initial social outcomes
concerning institutional, organizational or policy design -- even suboptimal ones -- can become
self-reinforcing over time (Krasner 1989; North 1990). These initial choices encourage the
emergence of elaborate social and economic networks, greatly increasing the cost of adopting
once-possible alternatives and therefore inhibiting exit. Major social arrangements have major
social consequences. Individuals make important commitments in response to government
actions. These commitments, in turn, may vastly increase the disruption caused by
institutional reforms, effectively "locking-in" previous decisions. As a result, as Douglass
North has emphasized, change in well-institutionalized polities is typically incremental.
Research on technological change has revealed some of the circumstances conducive to path-
dependence (David 1985; Arthur 1994). The crucial factor is the presence of increasing
returns, which encourages actors to focus on a single alternative and to continue movement
down a particular path once initial steps are taken. Large set-up or fixed costs are likely to
create increasing returns to further investment in a given technology, providing individuals

2
Veto points may also be policy-specific. The Canada Pension Plan (CPP), for example, cannot be reformed
without a large supermajority among provincial governments.
3
See also the discussions in the chapters by Myles and Pierson and Wood. A more detailed discussion can be
found in Pierson 1997.
4
with a strong incentive to identify and stick with a single option. Substantial learning effects
connected to the operation of complex systems provide an additional source of increasing
returns. Coordination effects (or network externalities) occur when the individual receives
increased benefits from a particular activity if others also adopt the same option. Finally,
adaptive expectations occur when individuals feel a need to "pick the right horse" because
options that fail to win broad acceptance will have drawbacks later on. Under these
conditions, individual expectations about usage patterns may become self-fulfilling.
As North has argued, all of these arguments can be extended from studies of technological
change to other social processes, particularly to the development of institutions. In contexts
of complex social interdependence, new institutions or policies often entail high fixed or start-
up costs, may involve considerable learning effects, and generate coordination effects and
adaptive expectations. Established institutions generate powerful inducements that reinforce
their own stability and further development. "In short", North concludes, "the interdependent
web of an institutional matrix produces massive increasing returns", making path-dependence
a common feature of institutional evolution (North 1990, p. 95).
Over time, as social actors make commitments based on existing institutions, the cost of "exit"
rises. Learning from past events may lead actors to act differently in launching new initiatives.
Recapturing ground in previously institutionalized fields of activity, however, will often be
quite difficult. Actors do not inherit a blank slate that they can remake at will when their
preferences change or the balance of power shifts. Instead, they find that the dead weight of
previous institutional choices seriously limits their room to maneuver.
Because this point is so often misconstrued, it should be stressed that the claim is not that path
dependence freezes existing arrangements in place. Change continues, but it is bounded
change. North (1990, pp. 98-99) summarizes the key point well: At every step along the
way there [are] choices political and economic that provide real alternatives. Path
dependence is a way to narrow conceptually the choice set and link decision making through
time. It is not a story of inevitability in which the past neatly predicts the future.
One of the major themes of this volume is that contemporary welfare states, and the politics
that surrounds them, strongly reflect these path dependent effects. As Myles and Pierson
argue, old-age pension systems provide a powerful example. Most countries operate pensions
on a pay-as-you-go basis: current workers pay contributions that finance the previous
generations retirement. Once they have been in place for a long time, pay-as-you-go systems
may face incremental cutbacks and adjustments, but they are highly resistant to radical reform.
Shifting to private, funded arrangements would place an untenable burden on current workers,
requiring them to finance the previous generations retirement while simultaneously saving for
their own. Even partial privatization has generally proven possible only in the relatively few
countries lacking extensive and mature pay-as-you-go systems.
Similar if less severe path dependent effects are likely in areas of social policy where complex
sets of institutions and organizations have co-evolved over extended periods. The chapters
by Giaimo, Manow, and Wood all emphasize this type of dynamic. In health care provision
and many aspects of labor market systems, social actors need to coordinate their activities and
they invest resources in line with the incentive structures of their existing environment(see also
Giaimo and Manow 1999; Hacker 1998). This probably helps to explain why employers, for
instance, have often been more half-hearted and internally divided over policy reform than
many theories of political economy might have anticipated (Thelen 1998).
5
Both the popularity of the welfare state and the prevalence of stickiness must be at the
center of an investigation of restructuring. The essential point is that welfare states face
severe strains and they retain deep reservoirs of political support. For political analysts, the
central question can thus be put as follows: What happens when the irresistible forces of post-
industrialism meet the immovable object of the welfare state? Acknowledging the strength of
both sides of this collision generates several implications for investigating the politics of
reform.
There are strong grounds for skepticism about the prospect for any radical revision of the
welfare state in most countries. Almost nowhere have politicians been able to assemble and
sustain majority coalitions for a far-reaching contraction of social policy (Stephens, Huber and
Ray 1999). The reasons have already been outlined. The broad scale of public support, the
intensity of preferences among program recipients, the extent to which a variety of actors
(including employers) have adapted to the existing contours of the social market economy, and
the institutional arrangements which favor defenders of the status quo make a frontal assault
on the welfare state politically suicidal in most countries.
Yet the chapters in Part I of this volume also suggest that pressures associated with post-
industrialism, intensified in some respects by globalization, have rendered the maintenance of
the status quo an increasingly unrealistic option. Continuing low growth coupled with the
challenges of creating service sector employment, population aging, and the overcommitments
of existing policies are already generating intense pressures. Tax levels strain public tolerance.
Payroll tax burdens and their possible adverse impact on employment and wages create
tensions within the traditional support coalitions of the welfare state, particularly in the ranks
of private sector unions (Visser and Hemerijck 1997). Barring an extremely unlikely return to
an era of high economic growth, fiscal pressures on welfare states are certain to intensify.
While tax increases may contribute to closing the gap between commitments and resources, it
is difficult to imagine that in many European countries changes in revenues alone could be
sufficient to maintain fiscal equilibrium. Thus even strong supporters of the welfare state
increasingly acknowledge that sustaining basic arrangements will require significant reforms.
It is a context of permanent austerity.
The prospect of permanent austerity transforms political conflicts over the restructuring of
social policy. Welfare state conflict is often portrayed as a clash between those wedded to
the status quo and those eager to dismantle basic social protections. In countries where
aggressive advocates of neo-liberalism have been in power, such as New Zealand and until
recently the United Kingdom, this has not been too inaccurate a portrayal. Yet in a climate
where social trends make pressures on budgets intense and unrelenting, political cleavages are
likely to become more complex. Those advocating restructuring will include many who wish
to preserve and modernize key elements of the social contract, but seek to do so in a manner
which does not create unsustainable budgetary burdens, contributes to economic performance,
and gives emerging social demands some chance of competing for public attention and
resources with well-established ones. In the current climate, restructuring must be
distinguished from retrenchment or dismantling.
My central contention so far is that neither the alternatives of standing pat or dismantling are
likely to prove viable in most countries. Instead, as in many aspects of politics, we should
expect strong pressures to move towards more centrist -- and therefore more incrementalist --
6
responses. Those seeking to generate significant cost reductions while modernizing particular
aspects of social provision will generally hold the balance of political power. In Claus Offes
words, the objective for those wielding electoral power in most countries will be smooth
consolidation (Offe 1991).
A useful initial framework for fleshing out this claim is a simple version of the pivotal politics
argument suggested in recent studies of American politics (Krehbiel 1998; Brady and Volden
1998). In any collective choice situation where policy preferences can be arrayed on a single
continuum, there is a pivotal actor whose vote determines whether an initiative moves forward
or is blocked. This pivotal voter is likely to wield disproportionate power, and policy
outcomes should generally gravitate toward that location. Pivotal voters need not be median
voters. A great deal turns on the institutional environment governing choice. Depending on
the significance of multiple vetoes and supermajoritarian systems, the pivotal voter will
generally be closer to the status quo than the median voter -- often much closer. In the United
States, for instance, if a President is prepared to veto legislation, the pivotal voter in the
legislature is not the one generating a bare majority, but the one producing the two-thirds
majority required for a veto override.
4
In practice, the political vulnerability of those seeking to modify popular social welfare
programs is such that they will often seek relatively broad consensus on reform rather than a
minimum winning coalition. Broader consensus legitimates the claim that policy change is
necessary, and intended to sustain rather than gut the program under review; it thus provides
essential political cover (Weaver 1986). The desire to make reform durable generally points in
the same direction. Especially with large, complex, and deeply institutionalized programs like
health care and pensions, social actors place a high value on predictability and continuity in
policy. Reform is not enough; powerful interests seek reasonable assurance that the new
policies can be sustained. This again encourages the assembly of grand coalitions. Such
coalitions, often informal and issue-specific, extend the range of actors with a stake in the
reform outcome, and increase confidence that the next election will not overturn the new
initiatives.
In other words, the pivotal actor in practice will generally be closer to the status quo than the
actor identified by a formal analysis of the institutional preconditions for minimum winning
coalitions. In a wide range of countries, the coalitions engaged in the restructuring of welfare
states have been far broader than minimum size, incorporating key interest organizations as
well as political parties outside the current government. Such a broadening purchases the
increased legitimacy and potential durability of enacted reforms. Most often, the price is a
more incremental adjustment than would have been (theoretically) possible with a smaller
coalition.
A final, related factor in pushing the pivot towards the status quo is the possibility that the
governments capable of enacting reform at a limited political cost will be those possessing the
greatest credibility with voters on the issue (Ross 1998). Following a Nixon goes to China
logic, it will often be those governments with reputations of support for the welfare state that
have the greatest room to maneuver. Yet the very factors that produce such credibility (past

4
More precisely, it is the voter producing a two-thirds majority in the chamber where that voters preferences
are closest to the status quo.
7
commitments, ideological orientations, and the nature of a partys core constituencies) make it
unlikely that the favored party will use its enhanced maneuverability to dismantle established
social policies. Again, political incentives point towards more moderate, modernizing reforms.
Of course, students of comparative public policy have long recognized the role of positive-sum
bargaining among crucial organized interests and political parties. In most advanced industrial
democracies the new economic and fiscal environment has transformed, but not undermined
the conditions for consensus-oriented policy making based on political exchange. What has
changed is the currency for such exchanges. Traditionally, labors contribution to consensus
was wage restraint. As Rhodes argues, even following the demise of Keynesianism this
contribution remains important. Yet reformers of the welfare state also require credibility and
legitimacy, particularly in the eyes of voters. Left and center-left parties, and-or trade unions,
generally need to be brought into reform coalitions to make the restructuring of welfare states
politically sustainable.
The implications of the discussion so far can be seen in Figure 2. Depicting policy reforms on
a continuum from the status quo (q) to a full-fledged neo-liberal agenda of radical
retrenchment (r), one would expect the median voter (m) to be a considerable distance from
(r). The need to surmount institutional veto points pushes the government reform agenda
back to (v), and the desire to gain legitimacy-enhancing and stability-inducing consensus
promotes a further move to (c). My argument is thus that in most of the affluent democracies
the viable reform space will be, at a maximum, in the region (c)-(v). Reform thus entails a
substantial shift from (q), but it is a long way from (r).
[Figure 2 about here]
II. Two Major Complications
This is of course an extremely stylized treatment of reform politics. A more satisfying account
would need to complicate the analysis in two crucial respects. First, one needs to develop a
more nuanced conceptualization of the reform agenda, or, in social science terms, the
dependent variable. Second, one needs to consider the distinctive reform dynamics of
different welfare state regimes. Each of these complications represents an important part of
the current agenda for research on the politics of the welfare state.
Three Dimensions of Welfare State Restructuring. One of the striking features of current
comparative research on the welfare state is the lack of consensus on outcomes. How much,
and in what ways, have welfare states changed since the end of the post-war boom? The
authors in this volume generally depict most reforms in most countries as incremental rather
than radical, and focused on restructuring rather than straightforward dismantling. However,
some have argued that the degree of cutbacks has been more severe (Clayton and Pontusson
1998). Even among those who see the overall degree of change as fairly limited, there may be
little agreement about how to characterize the nature and scope of change cross-nationally.
It is difficult to exaggerate the obstacle this dissensus creates for comparative research. As
Kitschelt notes in his chapter, it is impossible to seriously evaluate competing explanations
when there is no agreement about the pattern of outcomes to be explained. Thus it is
important to ask why researchers have so much trouble with the dependent variable.
8
The problem lies partly in the concept of the welfare state itself, partly in data limitations, and
partly in limitations of current theorizing about welfare state change. "The welfare state" is
generally taken to cover those aspects of government policy designed to protect against
particular risks shared by broad segments of society. Standard features, not necessarily
present in all countries, would include: protection against loss of earnings due to
unemployment, sickness, disability, or old age; guaranteed access to health care; support for
households with many children or an absent parent, and a variety of social services -- child
care, elder care, etc. -- meant to assist households in balancing multiple activities which may
overtax their own resources.
Needless to say, this list covers an extremely wide range of government activity. Furthermore,
the trend in scholarship has been to broaden the already extensive domain of the subject
matter. Recent analyses have advocated more attention to public/private interplay (e.g. Shalev
1996, Howard 1997) and to the interfaces between the public sector, the market, and the
household (Orloff 1993; Esping-Andersen 1999; OConnor, Orloff and Shaver 1999). In this
volume, Schwartz makes a similar appeal for extending attention to systems of social
protection built into the regulatory arrangements governing particular economic sectors.
There is little doubt that this broadening has had salutary effects, forcing attention to
dimensions of social life that had previously received scant attention, and illuminating the
extent to which welfare states are nested in a set of broader institutional arrangements.
Yet as the concept of the welfare state, or welfare regime, "stretches", it becomes inevitable
that quite distinct processes and outcomes will be joined together under the umbrella of a
single master variable (Collier 199x). This in turn spreads confusion in two ways. First, it
fuels a process where analysts discussing what has happened to "the welfare state" find
themselves talking past each other because each is concerned with distinct dimensions.
Second, it makes efforts to develop summary measures of what has happened extraordinarily
difficult. The complexity of this multi-faceted concept cuts against our attempts to generate
the relatively parsimonious measures of outcomes that make a serious enterprise of
comparative explanation possible.
A second problem stems from data limitations. Even if we agree about the outcomes that we
are interested in measuring, how do we carry out the measurements? This problem has
become more acute as -- again for quite good reasons -- analysts have criticized simple efforts
to characterize outcomes through indicators of public or social expenditure. Following
Esping-Andersen's lead, there has been a broad recognition that many of the theoretically-
relevant outcomes of welfare state change will simply not be captured by expenditure data.
Indeed, I have argued elsewhere (Pierson 1994) that this is especially true in the current
environment. There is every reason to believe that policymakers will seek systematically to
engineer changes that produce their major expenditure implications only at a later point in
time. In short, there is probably no substitute for investigations that pay attention to fairly
detailed dimensions of policy change, including attempts to map their (perhaps uncertain)
long-term implications. Rigorously applying consistent criteria to even a small subset of the
affluent democracies is a time-consuming and expertise-taxing enterprise.
5
Carrying out such

5
Two excellent examples are Alber 1998 and Lindbom 1999. Alber covers three countries, although he limits
himself largely to transfer programs. Lindbom covers one country.
9
research for the affluent democracies as a whole would require the efforts of a large and well-
funded team. So far no one has carried it out.
6
The final element of the "dependent variable" problem stems from limitations of theory. It is
this issue that I want to explore in more detail here. One of the hidden premises of much
recent writing on the welfare state has been a return to the simple dichotomy of "more" vs.
"less." Implicitly, change is measured along a single continuum, stretching from the intact (or
even expanding) welfare state on one end to the seriously eroded or dismantled on the other.
Yet here again Esping-Andersen's core insights retain force. In a context where actors have
complex motives, and the dependent variable is so heterogeneous, attempts to reduce change
to a single dimension will be counterproductive.
Instead, starting from the perspectives of prominent actors in the reform process, we can think
about change along three dimensions: recommodification, cost containment, and
recalibration.
7
Each constitutes a potentially important dimension of welfare state
restructuring. Any effort to focus on only one will necessarily distort what the process of
restructuring is about. For any welfare state, or any particular welfare state program, we can
fruitfully think about the extent to which reform agendas and policy outcomes involve change
along each of these dimensions.
Recommodification. Decommodification, in Esping-Andersen's influential formulation,
occurs when a service is rendered as a matter of right, and when a person can maintain a
livelihood without reliance on the market" (1990, pp. 21-22). For Esping-Andersen, the
centrality of decommodification stems from his earlier reflections on the construction of the
Social Democratic welfare regime (Esping-Andersen 1985). He argued that this regime was
fashioned through the efforts of a highly-mobilized and well-organized working class, which
sought to use political power to overcome its vulnerabilities in the labor market.
Recommodification essentially involves the effort to reverse that process -- to restrict the
alternatives to participation in the labor market, either by tightening eligibility or cutting
benefits.
Particularly for those who see the current era as defined by the rise of business power,
recommodification stands as the key dimension for an investigation of welfare state
restructuring. The transformation of the welfare state, from this perspective, is primarily
about dismantling those apsects of the welfare state that shelter of workers from market
pressures, forcing them to accept jobs on employers' terms. This formulation has proven
particularly central to analyses of welfare state reform produced by those who come to the
topic from previous work on industrial relations. For these scholars, the shifting balance of
power between employers and unions stands at the center of political analysis (Clayton and
Pontusson 1999; Swenson 1991).
Recommodification clearly represents an important dimension of welfare state change. Yet it
is increasingly evident that the basic logic of decommodification outlined in Esping-Andersen's
Three Worlds was at least somewhat misleading even for the period of welfare state expansion

6
The voluminous Scharpf/Schmidt project probably comes closest, but it does not attempt to systematically
measure policy outcomes at the level of programs across the countries in their study.
7
I am grateful to Jonathan Zeitlin for suggesting the term recalibration.
10
during the post-war period.
8
It suggests an image of welfare states foisted on capitalists. The
problem with this line of thinking is not that it takes capitalism too seriously, but that it fails to
take capitalism seriously enough. These are, after all, economies where investment depends
on the capacity of capitalists to earn profits, and where the need to induce investment
therefore confers substantial political power (Lindblom 1977). The question thus arises: how
would programs seriously damaging to economic performance thrive so extensively for so
long?
As Manow stresses in his chapter (see also the contributions of Rhodes, Swank, and
Huber/Stephens) the attitude of private firms to the expansion of the welfare state cannot be
reduced to one of recalcitrant opposition to decommodification. To argue that employers
have been enthusiastic builders of the welfare state would be revisionism run amok. Yet
particular elements have been enormously appealing to particular employer interests,
facilitating rather than impeding their core strategies (Mares 1998). Moreover, in many other
respects employers have adjusted to welfare state arrangements and policymakers have
accommodated the welfare state to employers -- over extended periods of time. As Soskice
(1999) has argued, particular types of firms are likely to thrive in particular institutional
settings. Thus there is often a strong co-evolutionary aspect to the intersection between
varieties of capitalism and systems of social provision.
This is not to suggest that commodification has not been a relevant dimension in recent
struggles over the welfare state. For particular actors, in particular countries, with respect to
particular programs, this dimension has been highly salient. "Work incentives" have been the
focus of concern in many cases, ranging from the reform of Dutch disability pensions to the
abolition of AFDC in the United States. Yet there remains considerable variation in the extent
to which welfare state reform has focused on improving work incentives. All of the studies in
Section IV emphasized that reform is often not primarily about recommodification. As Wood
argues, even actors who one might think would make this a high priority -- e.g. German
employers -- have often had more pressing concerns. The task for analysts then becomes
threefold: to identify the conditions under which a focus on recommodification is significant,
to establish the degree of change along that dimension, and to explain the observed patterns.
At the same time, one must avoid the temptation to reduce the discussion of welfare state
restructuring to this single aspect.
Cost-containment. In his powerful critique of expenditure-based analysis of welfare state
variations, Esping-Andersen observed that spending levels were essentially derivative of, and
often not a good proxy for, other outcomes (such as decommodification, poverty relief, or
status maintenance) which actors valued. During the period of welfare state expansion, he
argued pointedly, "it is difficult to imagine that anyone struggled for spending per se" (Esping-
Andersen 1990, p. 21).
In the current climate, however, people do fight against spending per se. Indeed, this is a
defining characteristic of the era of austerity. As I argued in chapter two, a range of
endogenous pressures, including the shift from manufacturing to services, demographic and
household change, and the maturation of governmental commitments are placing inexorable

8
As Ann Orloff has stressed to me, commodifications connotation of diminished autonomy or choice is also
problematic when applied to women, for whom the shift from unpaid household work to participation in the
paid labor market may enhance rather than diminish autonomy.
11
demands on government budgets. Deficit reduction is a high priority for many of the countries
that have joined or seek to join the European Monetary Union. In many contexts, powerful
actors are concerned first and foremost with the implications of reform for levels of
government expenditure. Of course, the imposition of austerity may become a vehicle for the
pursuit of other ambitions, but often the principal focus is cost containment itself.
With the exception of Wilensky's important work (Wilensky 1981), social scientists have
generally treated social expenditure and taxation as two distinct realms of research. In reality,
of course, they are two sides of the same coin, inextricably linked. The need to finance public
spending in the current environment often becomes the most powerful constraint on existing
social policy arrangements. Employers and financial interests worry about high tax levels
because of their potential impact on profits -- an impact that may be felt through a number of
distinct channels. In many countries, for instance, a major preoccupation is the fear that high
fixed labour costs, generated in part by payroll tax rates, are seriously impeding the ability of
employers to hire low-skilled workers.
Of course, the other political channel for opposition to taxes runs through the electorate. As
the welfare state has expanded, its financial underpinnings have shifted, necessarily, to heavy
reliance on funds from middle-income households. Higher taxes, combined with slower
growth of real incomes, have generated popular discontent. In most countries, politicians
must also cope with the downside of pay-as-you-go financing. As Myles and Pierson argue,
the current era marks the reversal of the favorable political dynamic that accompanied the
phase-in of pay-as-you-go pension systems. During the "golden age" politicians could make
generous promises while deferring the cost (i.e., high payroll tax rates). Today's politicians,
rather than being in a position to claim credit for new initiatives, act primarily as the bill-
collectors for yesterday's promises. Although voters almost everywhere retain widespread
allegiance to public social provision, these sentiments are now intermingled with stiff
resistance to significant tax increases (Bonoli, George and Taylor-Gooby, forthcoming).
Governments face the unenviable task of reconciling strong tendencies for higher outlays with
the potential for voter backlashes and the possibility that new taxes will damage economic
performance. At the same time, most countries face tighter constraints on their ability to run
deficits (most obviously in countries subject to the convergence criteria of EMU). In this
climate, cost-containment itself emerges as a top priority. Again, the questions for analysts are
to identify the circumstances where a focus on cost-containment becomes prominent, to
establish the degree to which that goal is achieved, and to explain the patterns identified.
Recalibration. Still, the agenda confronting contemporary welfare states cannot be reduced to
cost-containment plus commodification. By recalibration I mean reforms which seek to make
contemporary welfare states more consistent with contemporary goals and demands for social
provision. Two different types of recalibration should be distinguished. Rationalization
involves the modification of programs in line with new ideas about how to achieve established
goals. Updating concerns efforts to adapt to changing societal demands and norms -- e.g.
changes in the household, the life course, the nature of the labour market, or the age
composition of societies. Rationalization includes attempts to correct obviously "incentive-
incompatible programs" or cases of overshoot which may become evident over time, especially
if external circumstances change in a way which greatly changes the functioning of programs.
Examples might include disability programs in the Netherlands, or public service pensions in
France and Italy. This kind of modernization would also include reforms of service systems,
12
including health care, designed to improve the efficiency of provision or the responsiveness of
such systems to consumer needs and demands.
The reform of Sickness Pay in Sweden can serve as an example of rationalization. By the mid-
1980s, absenteeism rates in Swedish workplaces began to reach extremely high levels. There
was fairly broad agreement that sick pay programs were being abused that these programs
were helping to produce high rates of absence from work rather than, as intended, providing
protection against an important social risk. Spending rose sharply, from under 2% of GDP in
the early 1980s to almost 3% of GDP in 1988 (Benner, Vad and Schludi 1999, Graph 23). The
result was a consensus partial and not without conflict -- on a set of reforms to correct these
tendencies. Replacement rates were cut, and waiting days added. Revenue-neutral changes
were introduced to increase the incentives of employers to monitor use of sick pay provisions.
Both the pattern of reforms and their timing suggest that the dominant concern was to restore
sick pay to its originally-intended role rather than to worsen the terms on which genuinely-sick
workers could choose to be absent from work.
This is tricky territory analytically. Since in the current context these reforms will often be
designed to save money, how do we distinguish the impact of new ideas about how to do
things, or efforts to recalibrate errant programs, from simple cutbacks in provision? Yet clearly
we do not want to smuggle all modifications of programs into a framework of "assaults on the
welfare state." Over time actors will sometimes discover that particular programs do not work
as intended, or they will determine that there may be better ways of achieving their goals. In
such a context, they will push for changes. Thus recent reforms in Swedish health care
provision have produced considerable reorganization and resulted in public employment
reductions, particularly among the low-skilled. From one vantage point (Clayton and
Pontusson 1998) this represents a fundamental assault on public sector workers. From
another, however, it represents an attempt to increase productivity in the public sector in order
to provide high quality and flexible services to consumers at a politically-sustainable cost. The
equation of public sector reform with simple retrenchment or the rollback of the state is highly
questionable without evidence that cost-savings have reduced service quality.
Updating involves the modification of existing programs, or the initiation of new ones, in
response to newly-recognized social needs. A defining characteristic of the current era is the
co-existence of social conditions which are in many ways "new" with welfare states which are
in many respects decidedly "old" (Esping-Andersen 1999). As has been discussed, mature
welfare states will tend to be characterized by various forms of stickiness. In addition, many
current policy outcomes are the lagged effects of decisions taken decades ago. There is thus
often a considerable "mismatch" between emerging social risks and shared understandings of
appropriate targets for state intervention, on the one hand, and the existing array of
government social policies on the other. The various aspects of this mismatch constitute an
important dimension of restructuring agendas in contemporary welfare states.
Problematic as the concept of recalibration might be, it is more problematic to reduce the
nature of current welfare state restructuring to simple cost containment plus
recommodification. A variety of initiatives, dealing with issues such as gender equity and
social exclusion, simply cannot be squeezed into these other categories. Consider some of the
most striking initiatives of recent years: the introduction of private but essentially
mandatory superannuation in Australia (C. Pierson 1998); the establishment of long-term care
insurance in Germany (Gotting 199x); the enactment or expansion in many countries of a
13
range of subsidies or "contribution credits" for unpaid caring work (Daly 1997); and the
marked expansion of various initiatives to make work pay by subsidzing the terms on which
workers enter the low-paid labour market (Myles and Pierson 1997). None of these initiatives
can be incorporated into a simple vision of expanding markets (commodification) and
contracting states (cost-containment).
Introducing multiple dimensions of welfare state restructuring confers several very
considerable analytical advantages. First, breaking down the very broad category of "change"
into discrete dimensions sharpens our capacity to discern distinct patterns in outcomes. One
can identify striking variations both across policy arenas and across countries. Particular
sectors of the welfare state tend to be much more preoccupied with one or another dimension
of welfare state restructuring. The differences are clearly highlighted in this volume's "policy
domain" chapters (Part III). Wood shows that in the case of labor market reform the primary
agendas are recommodification and recalibration, in differing and quite interesting
configurations. By contrast, cost-containment is a much less significant motivation for policy
reform. In the case of health care and pensions, on the other hand, cost containment is the
issue in most countries, though flanked in some cases by efforts at modernization. Only
rarely will recommodification provide the primary lens for analyzing the character of reform
in these core welfare state sectors. To take a third example, in the case of family policies, the
main pressure all welfare states face is the need to adapt social policy arrangements to the
radically transformed interfaces between market, state, and households.
Distinct dimensions of welfare state restructuring will be of varied salience not only across
program areas, but across countries and over time. For instance, acute fiscal crises or the dash
to meet EMU's convergence criteria obviously place a premium on cost-containment efforts.
More generally, I will argue in the next section that the three different "worlds" of
contemporary welfare capitalism give rise to distinct policy agendas and, in part as a result,
distinct competing political coalitions and reform dynamics. It has been in the liberal welfare
states that a focus on recommodification has been most pronounced. These already highly
commodified welfare states have become more so -- especially in Britain, New Zealand, and
the United States. By contrast, recalibration and cost-containment have been more central to
the policy agenda in Continental welfare states, while cost-containment has been the principal
issue in the Social Democratic welfare states of Scandinavia. A focus on multiple reform
dimensions thus allows the analyst to highlight the distinct problem loads of different welfare
states.
Finally, disaggregating these reform dimensions increases our capacity to make sense of
relevant political processes, facilitating more nuanced accounts of actor interests and political
activity. Different actors may be concerned about distinct dimensions of reform; particular
actors may have multiple but partly conflicting objectives. Furthermore, because of these
multiple priorities there may be unsuspected opportunities for issue-linkage and negotiated
change in which different actors trade off lesser concerns for greater ones. As noted in the
chapters by Bonoli, Rhodes, and Myles and Pierson, one of the striking observed outcomes in
some configurations is unexpected coalitions based on quid pro quos. One can place this in
the framework outlined in the first part of this chapter. Recognizing the multi-dimensionality
of reform is crucial in a context where many actors agree on the need for change but have
different interests and priorities. Analysts elsewhere (Levy 1999; Myles and Pierson 1997)
have stressed the centrality of compromise and the search for positive-sum trade-offs under
14
conditions where broad political coalitions are necessary and straightforward retrenchment is
politically difficult to sell.
Three Distinct Welfare State Regimes. The second necessary complication of my initial
framework stems from the need to recognize the existence of very distinct political and policy
configurations distinct regimes -- within the ranks of the affluent democracies. Up until now
I have spoken of essentially a single logic of welfare state reform. Just as we have to
distinguish different reform dimensions, however, we also need to recognize the existence of
quite different settings for the emerging politics of restructuring.
9
Esping-Andersens typology of three worlds of welfare capitalism liberal, conservative, and
social democratic has been enormously influential. At the same time, it has been subjected
to frequent criticisms, mostly stressing that particular countries are poorly categorized, or
arguing that other dimensions of variation are neglected. Any typology covering the complex
realities of a large number of countries is vulnerable to this kind of challenge. Yet there are
excellent theoretical and methodological reasons to organize the explanation of variance in
outcomes around an analysis of distinct regimes rather than lumping all OECD countries into a
single pool in which one scrutinizes variance along a single, continuous range of
independent variables.
To talk of a regime Esping-Andersen maintains, is to denote the fact that in the relation
between state and economy a complex of legal and organizational features are systematically
interwoven (1990, p. 2). What was most compelling in Esping-Andersens analysis of
modern welfare states was his insistence that welfare states be seen as part of complex
historical configurations.
10
The three worlds did not result from more or less of a few
discrete, independent master variables, but from the interactive and cumulative effects of a
number of interdependent causal factors. Variables may have a particular impact only when
accompanied by a set of additional factors. This perspective has strong affinities with Ragins
concept of complex conjunctural causation (Ragin 1987). As Shalev has put it, the
argument is that welfare regimes should be seen as a limited number of qualitatively different
configurations with distinctive historical roots. Causally, the claim is that countries cluster
on policy because they cluster on politics (Shalev 1999, p. 13, emphasis in original). In this
framework, it makes no sense to argue for a linear relationship between discrete independent
variables and discrete dependent variables, e.g., that any discrete increment of Catholicism or
absolutism ought to yield a discrete and uniform increment in pension corporativism (ibid).
11

9
In case it is not obvious, I should acknowledge that many of the arguments in Pierson 1994 and Pierson 1996,
especially those which deemphasize the contemporary role of organized labor, suffer from precisely this defect.
This criticism is effectively advanced in Visser and Hemmerijck 1997.
10
In this respect, the analysis runs directly parallel to a very prominent theme in comparative political
economy, emphasizing the varieties of contemporary capitalism (Kitschelt et al 1999; Hall 1999; Soskice
1999). In Soskices influential account, there are strong interlocking complementarities between different
parts of the institutional framework. Each system depends on the other systems to function effectively (1999,
p. 109). Spurred in part by Esping-Andersens work, a number of researchers have begun to explore the
connections between production systems and welfare state regimes (see Ebbinghaus and Manow 1998 and
Huber and Stephens, Manow, Rhodes, and Wood in this volume).
11
Esping-Andersen himself employed such regressions, but as Shalev argues persuasively, there is an obvious
mismatch between Esping-Andersens claims and his methods The regression approach treats both
policy and politics as continuous variables It is hard to exaggerate the fundamental incompatibility between
[multiple regression] and Esping-Andersens regime approach In his hands [multiple regression] was
simply a blunt instrument for tapping gross differences between groups of countries that could have been
15
The arguments about path dependence discussed earlier in this chapter and elsewhere in this
volume (Myles and Pierson; Wood), as well as the contributions which focus on linkages
among regime features (Manow, Huber and Stephens) point in the same direction.

Path
dependent processes are very likely to be prevalent in contexts where a set of organizations
and institutions develop together over extended periods of time, reinforcing each other
through processes of mutual adaptation and competitive selection. Institutions and
organizational actors that constitute a poor fit are less likely to survive over time. Such
processes foster the emergence of quite distinct configurations, containing many elements
which make sense in the context of the others. To take a prominent example (Huber and
Stephens, forthcoming) a number of factors operated together to create the Social
Democratic service state. In particular, the rapid entry of women into the workforce, in a
context marked by social democratic party and union strength, fueled demand for supportive
social services. Efforts to meet that demand through public sector expansion both created
additional opportunities for womens employment and strengthened political forces pushing
for further expansions.
Different welfare state configurations are the products of complex conjunctural causation,
with multiple factors working together over extended periods of time to generate dramatically
different outcomes. There is no theoretical justification for arguing that a ten percent shift in
the value on one variable or another will have a simple, direct, linear effect on outcomes
across all cases. To clarify, let me discuss one example: the impact of political institutions on
the prospects for reform in the current period. Most discussion this issue has been at the level
of welfare states in general. Pierson and Weaver (1993) stressed that there was no clear
theoretical basis for believing that increased institutional fragmentation necessarily made
retrenchment more difficult. The concentration of authority also concentrated accountability,
which might lead to difficulties in pursuing unpopular policies. This accountability/blame
avoidance dynamic is important for understanding why even unified governments will often be
cautious, for explaining the strategies that such governments employ, and for highlighting the
possibility that aggressive action may lead to electoral backlash.
By now, however, the evidence would seem to show pretty clearly that on balance the
concentration of political authority is an asset for those seeking reform.
12
As Huber and
Stephens argue, the experiences of the two pure Westminster cases, New Zealand and the
United Kingdom, are particularly striking. New Zealands National government was able to
move aggressively on multiple fronts. The British Conservatives gradually learned to shift
from a strategy of frontal assaults to one of relentless, low profile adjustments, which whittled
away important elements of the welfare state over time. By contrast, veto points in other
countries have generated significant, sometimes overwhelming, obstacles to radical change.

conveyed by the use of tables and charts without the implication of constant linear effects across countries
(Shalev 1999, p. 13).
12
Again though, Swanks chapter suggests that while fragmented institutions might impede retrenchment
efforts, they are also likely to slow welfare state expansion in the first place. This slowdown occurs both
because of institutional fragmentations direct role in blocking reform and because of its indirect negative
effects on social solidarity through the promotion of interest heterogeneity. This is in keeping with a general
presumption in much institutional research, namely, that the effects of institutions will generally be multiple
and cross-cutting rather than simply direct and unidimensional. A different way to put this is that we need to
distinguish between the short-term and long-term causal effects of institutional fragmentation. For a
discussion of this point see Shalev 1998.
16
Yet it is no accident that both these examples of major retrenchment in the absence of formal
institutional veto points concern liberal welfare states, where social actors are disorganized
and popular support for the welfare state is (while still extensive) more limited than elsewhere.
It seems highly unlikely that the same value on this particular independent variable would
have similar consequences in a configuration where popular support for the welfare state was
broader and/or the power resources of labor were greater a conclusion which Bonolis
dicussion of the French reform experience supports. The freedom to produce radical reform
stems not just from high concentration of political authority, but from that factor combined
with a number of other features of a particular configuration.
In analyzing the dynamics of social policy restructuring, one needs to attend to both the
particular scale and shape of welfare states and particular political contexts. With respect to
the latter, the scope of popular support for social provision, the connections between social
provision and systems of economic production, and the relationship between the
electoral/partisan arena and systems of interest intermediation are especially critical. In short,
we need to recognize the existence of distinct worlds of welfare capitalism.
Even if one accepts the basic case for treating cases holistically, as configurations rather than
compilations of variables, one can nonetheless challenge Esping-Andersens specification of
regimes. A number of reasonable objections have appeared, and for particular purposes one
might prefer a different typology. Yet I am struck by the extent to which other analysts,
including those studying broader shifts in political economy, have gravitated towards similar
demarcations (Iversen and Wren 1998; Kitschelt et al 1999; Scharpf 1997). My view is that
this reflects a reasonably tight fit between particular welfare state configurations and particular
political configurations. This is indeed what one should expect if Esping-Andersen was
correct in arguing that each regime type emerges from a particular political milieu.
As Walter Korpi has argued, in evaluating the heuristic merits of typologies, the analysis of
causes and consequences of welfare states should be in the foreground (Korpi 1999, p. ).
Ultimately, the crucial issue is whether the distinction among regimes provides leverage for
explaining important variations across the relevant cases. In Part III of this essay, I argue that
it does although not always in the manner suggested by Esping-Andersens own discussions.
When combined with the disaggregation of reform dimensions outlined above, one can begin
to make sense of variations both across regimes and among cases within each regime.
III. Restructuring the Three Worlds of Welfare.
What follows is a preliminary attempt to outline the distinct politics of welfare state reform in
the three worlds, or regimes, of liberal, social democratic, and continental welfare states.
13
I
argue that the basic framework developed in Part I is helpful for making sense of what is
happening in each of these three worlds. Each world, however, is composed not only of
particular types of welfare states, but also of distinct political settings. Thus the agendas for

13
A number of the cases sometimes considered in such comparisons are excluded from this analysis. Greece,
Ireland, Portugal and Spain are late developing welfare states, which were still very much in the process of
welfare state construction at the end of the golden age. In my view this context raises quite different issues
from those cases where affluence had been achieved and welfare states were already close to maturity by the
mid-1970s. Switzerland and Japan represent distinct configurations that do not fit easily into any of the three
regimes.
17
welfare state restructuring and the dominant political coalitions will vary. Furthermore, we
can expect additional variation across cases within each world, and one of the major tasks for
analysts should be to identify and explain that variation.
By distinguishing among regimes, and disaggregating different dimensions of restructuring,
one can identify patterns that would not be evident in a more unified analysis which tried to
explain a single outcome (e.g., retrenchment) over the whole set of cases. In these respects,
I follow Esping-Andersens regime typology (although utilizing different outcome
dimensions). On the other hand, our accounts of the political dynamics in the three regimes
are quite different.
14
The goal is to make the investigation complicated enough to capture and
account for crucial elements of diversity, but not so complicated that it becomes impossible to
identify general patterns.
The Liberal Regime. The liberal cases include Australia, Canada, Great Britain, New Zealand,
and the United States.
15
The most politically salient features of the welfare state constellations
in these countries include the following (data on most of these features can be found in the
Huber/Stephens chapter in this volume). Taxes, and spending, have remained low by
international standards. Public sector service employment is also low. Many transfer
programs are income-tested, although the range of coverage varies from very narrow (the
United States) to quite broad (Australia). In part as a consequence of the failure of the
welfare state to meet demands for social provision, private sector activity in pensions and
social services such as child care (as well as health care in the United States and New Zealand)
is extensive. In many cases, tax expenditures subsidize private provision for the upper middle
class (e.g. Howard 1997). Finally, these welfare state arrangements operate in the context of
liberal market economies. There is no overlap between the world of liberal welfare states and
the world of "organized market economies." Thus there are numerous linkages, explored
below, between liberal welfare state arrangements and the liberal or disorganized model of
capitalism.
The political constellations of these cases also share a number of common features. First, and
most crucial is the weakness of encompassing interest organizations. By comparative
standards these are cases where organized labor has modest political capacities. Again, there
is considerable variation among these countries, ranging from Australia at one end to the
United States at the other. It is not only that labor is weak in these cases; the capacities of
employers for collective action are also limited. In Soskices words, companies have little
capacity to coordinate their activities collectively. Their inability to act collectively means that

14
This is not an issue that Esping-Andersen has pursued in detail. In Three Worlds he identified distinct
political cleavages in each regime. In Postindustrial Economies he seems to maintain, although without much
elaboration, that path dependence and median voters will prevent major policy change. The current
analysis points to quite different political cleavages in the three worlds, and argues that while path dependence
and pivotal voters channel reform, they are unlikely to prevent it.
15
Castles and Mitchell (1994) have persuasively argued that the two Antipode countries should be seen as a
distinct "wage-earner" model, in which protectionism combined with intensive labor market regulation to
produce relatively egalitarian outcomes without extensive formal welfare states. However, the wage-earner
model, grounded in protectionism, came under acute pressures from a changing international economy. For
these countries globalization has clearly mattered enormously. This alternative model began to break down in
the mid-1970s, forcing a gradual shift towards broadly liberal arrangements. Australia, in particular, remains
somewhat distinct from other liberal cases in the continuing role of politically-mediated wage bargaining. The
differences are less pronounced than they were two decades ago, however, especially following the recent
period of Liberal-National governance.
18
they cannot combine to negotiate discretionary framework solutions with the state (1999, p.
110). In short, with the partial exception of Australia, not only does labor lack the power to
veto change, but the capacity of these systems to pursue negotiated reform through systems of
organized interest intermediation is very low. Unlike the case in other welfare state regimes,
policy changes must be executed almost exculsively through electoral and partisan politics.
Several crucial features of the electoral/partisan environments thus deserve special emphasis.
First, until New Zealand's recent reform, none of these cases employed proportional
representation. Instead, "first-past-the-post" electoral systems have been the norm.
16
Thus all
of these countries have had strong tendencies towards two (or two-and-a-half) party systems.
With the exception of the United States, where division between Congress and the President
has become the norm, these systems tended to produce single-party governments. They
varied, however, in the extent to which political institutions provided checks on these
governments. New Zealand (pre-1996) and Great Britain constitute pure "Westminster"
models of "elected dictatorship." By contrast, the federal systems of Australia, Canada, and
the United States create additional veto points, the severity and nature of which varies across
countries, issue area, and electoral context.
Furthermore, there is little question that the scope of popular support for public social
provision tends to be more conditional in these cases, although again with considerable
variation (Svalfors 1997). One can see here how multiple features of a particular
configuration point in the same direction. Because these welfare states are relatively small, the
"core" support group for social provision is also relatively small. Reliance on means-testing
may divide those who benefit from many taxpayers. The political clout of labor unions, a
traditional bastion of support, is also relatively modest. The institutionalization of (often state-
subsidized) market alternatives weakens middle class attachments to public provision. High
levels of inequality are also associated, in many countries, with large class biases in electoral
turnout. Low turnout among the economically vulnerable further diminishes their already
limited political influence.
To be absolutely clear, the claim is not that a majority of voters tend to oppose the welfare
state in these countries; nowhere is this true. Compared to the social democratic and
continental countries, however, pivotal voters are likely to possess weaker attachments to
social provision, and to be more susceptible to alternative political appeals (such as the
demand for tax cuts). Liberal welfare states thus provide the greatest potential for parties to
reconcile political success and a relatively aggressive, even openly hostile stance vis--vis
significant components of the welfare state. Under particular conditions, pivotal voters may
be within reach, and in some of these countries labor can be essentially excluded from playing
a significant role. Thus the political conflict between advocates of moderate restructuring and
radical retrenchment is more equally-matched than it is in the other two regimes.
This is reflected in the pattern of outcomes. By comparative standards, quite radical cutbacks
have been achieved in New Zealand, and, to a lesser extent, in Great Britain (Stephens, Huber,
and Ray 1999, Castles and Pierson 199x). Canada and Australia, on the other hand, have
pursued a distinctive course, also marked by efforts at cost-containment and commodification,

16
Australia employs a preferential voting system. Because voters second choice matters, this will generally
induce parties to moderate their policy stance.
19
but balanced by serious efforts to protect the most vulnerable. In the United States, change
has been more limited, and contains elements of both tendencies.
Although the variations across countries both in policy outcomes and in inequality trends are
quite striking, these outcomes are not sufficiently stable to allow firm conclusions about
national trajectories. Australias current National-Liberal government is seeking to erode
many of the provisions that marked the accords reached between the Australian Labour
governments and the trade unions between 198x and 199x. Under the current British Labour
government, important new policy initiatives emulate some of the Canadian and Australian
strategies which combine a strengthening of work incentives with compensatory policies for
those affected (Hills 1998). These recent developments suggest that partisan control of
government may be most important in the liberal regime. This is indeed what we should
expect given the relatively narrow base of welfare state support in these countries, and the
dominance of the electoral/partisan arena for political action.
A defining characteristic of restructuring in all the liberal welfare states has been the priority
placed on recommodification. Indeed, both the emphasis placed on commodification and the
degree of change has clearly been greater than in the other welfare state regimes. In all the
liberal welfare states, programs providing transfer payments to those of working age but out
of the labor force unemployment benefit and social assistance --have faced major cutbacks.
Eligibility rules have been tightened, and benefit levels have been reduced significantly.
Coverage rates (in Canada and the United States) and benefit levels (everywhere) for the
unemployed have fallen sharply. Most dramatically, in Canada and the United States, national
commitments to social assistance have been severely weakened. Provincial and state
authorities have often moved aggressively to push the poor off of social assistance roles and
into the workforce.
At first glance, this characteristic of liberal regime reform is puzzling. Why would these
systems, already the most commodified in the OECD, push so much more aggressively in
this direction? Part of the answer lies in the political weakness of those who might resist
commodification. Equally important, however, is the connection between income support and
the labor market in political economies where wage flexibility is treated as the principal buffer
against high unemployment. The deteriorating market position of low-skilled workers has
confronted policymakers everywhere with difficult trade-offs (Iversen and Wren 1998; Scharpf
1997). Consistent with the basic workings of a liberal, disorganized market economy, all of
these liberal countries have implicitly accepted that the new environment requires larger wage
differentials, and, in particular, deteriorating relative wage conditions among the low-skilled.
Yet this deterioration can only be carried out if exit options (reservation wages) are not
attractive. That is, the push for (downward) flexibility among the low-skilled implies a
hardening of conditions for income support to those out of work. Improving work
incentives has thus been a common frame of reform, justifying stricter eligibility, benefit cuts,
or even outright abolition of programs.
Many comparative analysts view the liberal welfare states primarily as an analytical or
normative foil for the real welfare states of continental Europe and Scandinavia. For these
analysts, the policy changes just described are taken to define the essential outcomes for these
cases. From this perspective, the United States is taken to typify the liberal world, and the
welfare reform which abolished Aid to Families with Dependent Children is taken as the
20
paradigmatic policy change in the U.S. Social policy dismantling and recommodification are
portrayed as the logical political destination for liberal welfare states in the new era.
The actual patterns, however, are both more complex and more interesting. All the liberal
welfare states have shared in the shift towards recommodification, albeit at different paces and
to different degrees. Yet there has been considerable variation in the extent to which this
transition has been subsidized or buffered. The main political debate in the liberal welfare
states has not focused on whether or not low-skilled workers should be in the labor market,
but on the terms under which their participation should take place. The key issue is the extent
to which commodification should be subsidized. Those advocating subsidization have
accepted demands for reform in social assistance and unemployment benefit, but have sought
to make work pay by supplementing poverty-level wages available in an unregulated market
with various forms of targeted social provision. Arguments in favor of such provisions
emphasize both social equity concerns and the need to foster the development of human
capital. Many of these supplements are targeted on groups that are considered particularly
vulnerable and/or particularly deserving of support, especially households with children.
The particular package of supplements varies, but may include some or all of the following:
improved child allowances and access to affordable child care, reductions in social insurance
contributions for the low-paid, an increased minimum wage
17
, expanded access to public
health insurance (in the United States), and the expansion or introduction of tax-based wage
supplements. The last provision has been particularly important. All the liberal countries have
introduced significant wage subsidies, operated through systems of taxation, for at least some
of the working poor usually families with children. Most dramatic has been the major
expansion of the Earned Income Tax Credit in the United States and of the Working Families
Tax Credit (formerly Family Credit), in Great Britain. In Australia, the universal child benefit
was means-tested (although 60% of the population remain eligible), but an expanded
Additional Family Payment provided significant additional cash on a per child basis to working
families.
Some of these subsidization initiatives could be dismissed as cosmetic embellishments to harsh
exercises in retrenchment. Yet the overall scale of these mediating efforts is difficult to ignore.
In the United States, for instance, while there were considerable cutbacks in AFDC and
Unemployment Insurance even before the 1996 welfare reform, overall federal spending on
low-income groups rose significantly in the late 1980s and the first half of the 1990s (Weaver
1998). By the time AFDC was transformed and largely handed over to the states, the federal
government was spending twice as much per year on the rapidly-expanding EITC as it spent
on AFDC (Myles and Pierson 1997). Along with modest increases in the minimum wage,
EITC expansions since 1989 have contributed to a large real increase in earnings for single
employed mothers with two children.
Cost containment has also been a major political priority in all these countries. Deficit
reduction has been a dominant theme, and it has been common for center and center-left
parties (e.g., Blair, Clinton, and the Australian Labor government) to commit themselves to
policies of not raising taxes. This outcome is striking, given the relatively low levels of public
expenditure in these countries, and the absence of anything like the Maastricht criteria to

17
In the UK the Labour government has introduced a minimum wage for the first time, and set its level
significantly closer to the median wage than is the case in the United States.
21
generate an external constraint. Although in the cases of the three economies dependent on
export of primary goods (Australia, Canada, and New Zealand) one could point to external
pressures for budget balance, it is difficult to make this argument for Great Britain or,
especially, the United States. Within the liberal regimes, the more right-leaning position of the
pivotal voter appears to facilitate a focus on expenditure (and tax) restraint, along with deficit
reduction, even in the absence of an institutional forcing mechanism such as EMU.
18
Cost containment efforts, however, have not on the whole generated the kind of radical
rollbacks popularly associated with neo-liberalism. Indeed, it is hard to discern a distinct
liberal response in the big spending areas of health and pensions. Unlike the case of income
maintenance for the able-bodied, there is no clear policy imperative. In both health care and
pensions, outcomes appear to be heavily constrained by the particular policy arrangements
established before cost containment demands became paramount. Where there were well-
established, deeply embedded policy frameworks in these areas continuity has been more
evident than change. Thus in health care, Australia, Britain, and Canada have all made
modifications of their pre-existing systems in an attempt to cut costs, but without challenging
basic rights to quality care and in contexts where overall expenditures were increasing. In the
United States, the major policy initiative was a failed attempt to dramatically expand health
care coverage, albeit as part of a cost containment strategy. New Zealand again stands out as
a divergent case. Only there was a major effort undertaken to rollback the contours of state
provision for health care.
As Giaimos chapter demonstrates, there is no logic of liberal market economies that generates
a strong push for market expansion and state contraction in the area of health care. On the
contrary, a good case can be made for the opposite proposition: a strong state role in health
care will generally serve the interests of most employers. The Clinton administrations reform
initiative was based on this presumption. Its failure arguably reflected the incapacity of a
disorganized polity to facilitate the negotiation of acceptable terms (Judis 1995) and the path
dependent accumulation of vested private interests in the policy status quo (Hacker 1998)
rather than a more fundamental conflict of interest between employers and consumers.
It is also clear that there has been no convergence in pension systems. Canada and the United
States have modified their mature systems, emphasizing cost containment, with Canada
introducing a significant element of funding. Not locked-in by past commitments, the UK and
Australia have embarked on new (but very distinct) paths of funded provision. New Zealand
has been unable to proceed in any direction. As Myles and Pierson show, path dependent
effects in pensions are very strong. At the same time, however, these path dependent qualities
largely reflect the congealed liberalism of an earlier period. It is no accident that liberal
countries were less likely to have mature, earnings-related pay-go systems by the end of the
post-war boom. Thus the substantial expansion of funding in Canada and Australia, and
private (but heavily tax-subsidized) funding in the United Kingdom reflected not only the
unmet demand for earnings-related provision, but also the availability of fiscal space and the
relative sophistication of private financial markets made possible by an earlier history of
liberalism. Within a regime analysis, it is the American Social Security system (like Britains

18
Again, the regime approach may help us to think about multiple paths to particular outcomes. Italy and the
United States may both pursue vigorous deficit reduction, but this may reflect dramatically different political
dynamics.
22
NHS in the case of health care) that emerges as anomalous. What is revealing is the so far
modest transformation of this essentially Bismarckian element of the American welfare state.
As in the case of recommodification efforts, the Canadian and Australian experiences of cost
containment demonstrate their distinct path of restructuring. While programs have been cut,
significant efforts have been made to protect the most vulnerable. In each country, means-
testing from the top has been a dominant strategy for generating savings, and some of the
savings have been redistributed to low-income groups (Castles 1996; Myles 1996). This quid
pro quo has facilitated the maintenance of a reasonably broad, centrist consensus on
restructuring efforts. In a particularly dramatic example, a clawback of pension benefits for
high-income seniors in Canada has been accompanied by a major expansion of minimum
pension benefits. This shift has been so successful that it has allowed low-spending Canada to
sharply reduce poverty rates for the elderly, which now approximate Scandinavian levels
(Myles 1999).
One can see the differences within the evolving liberal world most starkly in measures of
income inequality and poverty. Poverty rates have risen massively in the United Kingdom and
New Zealand far more than in any other OECD country. They have risen much more
moderately in the United States but from a higher baseline. On the other hand, at least
through the early 1990s there was little change in poverty rates in Canada and Australia.
Although social welfare policy does not account for all of the differences across these cases it
has been quite important. Policy interventions that might have sheltered the poor from
shouldering the costs of adjustment were actively dismantled in New Zealand and, until
recently, Great Britain. In Australia and Canada (and much more tepidly in the United States)
such protections were in many cases expanded to offset growing inequality in market incomes.
What explains this variation straightforward and severe cost containment and
recommodification in New Zealand and the UK, compensated restructuring in Australia and
Canada, with the United States as an intermediate case? New Zealand and the UK stand out
as cases where there was little inclination or necessity for compromise in the pursuit of
neoliberal goals. As has been discussed, the broad contours of the liberal regime type make
such an agenda at least conceivable, while it is clearly beyond the reach of any likely political
coalition elsewhere. In addition, however, two political conditions have appeared essential for
such an aggressive coalition to achieve significant policy successes: first, a series of economic
setbacks which discredited the policy status quo and its supporters; and second, a partisan-
institutional configuration that translated an electoral plurality into a governing majority and
allowed that majority to operate essentially unhindered (Huber and Stephens this volume;
Schwartz and Rhodes 1999).
As Schwartz and Rhodes argue, the aggressive neo-liberal agendas pursued in both Britain and
New Zealand grew in part out of severe economic dislocations which underscored the need for
major reforms. In both cases, the neo-liberal projects were reactions to failed efforts to
intensify policy strategies developed in the 1960s and early 1970s Keynesianism, incomes
policy and expanded public spending in Britain, protectionism and state-subsidized production
in New Zealand. The dramatic collapse of these projects lended credence to the neoliberal
agenda, summarized in Thatchers memorable phrase that there is no alternative. At the
same time, the loss of credibility and divisions generated by the failed projects demoralized and
disorganized neoliberalisms political opponents.
23
A second, equally critical condition is a partisan-institutional setting that shifts the pivotal
voter well to the right. I have argued that in general we can expect pivotal voters to be closer
to the welfare state status quo than the median voter. Yet not only is that median voter likely
to be comparatively far to the right in the liberal configuration, but the pivotal voter may
actually be to the right of the median. As noted by Huber and Stephens, conservative parties
in both Britain and New Zealand exploited Westminster, first-past-the-post electoral systems
and a divided opposition to translate a plurality of votes into comfortable parliamentary
majorities. During their eighteen consecutive years in office, for example, the British
Conservatives never received more than 44% of the vote. In New Zealand, parliamentary
majorities in the four elections between 1984 and 1993 were based on electoral pluralities of
43, 48, 48 and 35 percent.
Even under such circumstances governments face constraints or run the risk of backlash at the
polls. The Thatcher government was less aggressive in attacking the welfare state than other
aspects of the post-war order, and particularly cautious with respect to popular middle class
programs like the National Health Service (Pierson 1994). New Zealand's National Party
attempted to "crash through or crash." Ultimately it crashed, as a popular referendum in 1993
scrapped the Westminster model, which had permitted the government to ignore the median
voter. Nonetheless, compared with more typical settings, this partisan-institutional
configuration provided governments with greatly increased latitude to pursue unpopular
policies.
Unlike New Zealand and Great Britain, neither Australia nor Canada had institutional
arrangements that gave national governments unfettered control over the reform process. And
unlike the United States, these governments did have well-organized, relatively encompassing
interlocutors with whom to negotiate and implement a more consensual process of
restructuring. In Australia, the crucial mechanism was a series of accords between the unions
and the Australian Labour Party, based on a relatively high (by liberal standards) capacity for
concertation and a shared appreciation for the political risks associated with failure (Rhodes
and Schwartz 1999). We lack a convincing and detailed account of the political dynamics in
Canada, but part of the explanation must be the manner in which a decentralized federal
structure encouraged negotiation on the contours of adjustment between a series of national
governments and powerful (and politically diverse) provincial premiers.
The Social Democratic Regime. The Scandinavian or Social Democratic cases include
Denmark, Norway, Sweden, and (more marginally) Finland. The basic characteristics of this
welfare state configuration are well-known: generous transfer programs covering a wide-
range of risks with high replacement rates; public social services which, by comparative
standards, are extremely extensive; a set or supportive family and labor market policies which
generate very high rates of labor force participation for both men and women. The wide scope
of public provision has, of course, also filled spaces which elsewhere where occupied by
market arrangements. The private provision of social services, as well as pensions, remains
extremely limited.
The political constellation in these countries includes multiple features that create very strong
pro-welfare state coalitions. Union density is high, and the unions have maintained strong
links with electorally powerful social democratic parties. Public sector unions are also
enormous by comparative standards. Women are highly mobilized and committed to the
social service state (which is both the primary source of womens employment and provides
24
crucial support for womens labor force participation). Not surprisingly given these elements
of the political setting, public opinion remains strongly supportive of the key features of the
social democratic welfare state (Svallfors 1993; Svallfors 1997, Goul-Andersen 199x,
Rothstein 1998b). The formal political systems of these countries contain few checks on
parliamentary majorities, but the frequency of coalition governments makes political authority
less concentrated than in the Westminster setting. At the same time, governments in these
systems operate in a context of well-established and encompassing employer and union
organizations, which represent both possible partners in dialogue and potentially formidable
political opponents in cases where dialogue is absent or fails.
The basic results of welfare state restructuring in the social democratic cases reflect this
context of very broad support for public social provision. In the words of Eitrheim and
Kuhnle, politicians who blow the trumpet of neo-liberal economics too loudly, and those who
blow the tune that the welfare state is an evil invention, have a hard time surviving in Nordic
politics(Eitrheim and Kuhnle 1999). As Huber and Stephens persuasively argue (see also
Stephens 1996), the immediate challenge confronting this regime has been the adjustment to a
context of increased unemployment. The social democratic welfare states were in fiscal
equilibrium with 3% unemployment, but given their generous transfer programs they ran into
trouble when unemployment rose sharply. Once the increase was recognized as permanent,
which occurred at different times in different countries, a combination of benefit cuts,
economizing reforms in social services (especially health care) and tax increases were
introduced to restore equilibrium. Cross-national variation in the timing and degree of
restructuring largely reflects the timing and severity of economic crisis, and the degree of
adjustment required. Thus, the most painful reform process occurred in Finland, and changes
have been most modest in Norway. A strong argument can be made that this shift has largely
been accomplished in all these cases while leaving a modestly less generous version of the pre-
existing welfare state structures intact (Stephens 1996; Goul-Andersen 199x). The primary
task is to understand this general adjustment path for the social democratic regime, rather than
cross-national variation inside it.
Before directly outlining the basic reform dynamics, it is necessary to address the orthodox
interpretation of contemporary social democratic welfare state politics, which focuses on
cleavages between the exposed and sheltered (or traded and non-traded) sectors. Esping-
Andersen (1990;1996) identifies this as the predominant line of cleavage in social democratic
countries. This view has been elaborated in several analyses by scholars working from a
perspective shaped by the investigation of collective bargaining systems (Swenson 1991;
Clayton and Pontusson 1998; Iversen 1998). In this volume, Schwartz offers a nuanced
version of this sectoral argument, stressing the fact that both workers and employers in the
traded sector may continue to support significant aspects of social provision.
Yet the evidence that this potential cleavage drives the politics of welfare state reform is, to be
generous, extremely limited. Even in the social democratic countries where large-scale public
sectors might be expected to generate such effects, strong advocates like Clayton and
Pontusson are forced to acknowledge that the empirical support is modest.
19
In his recent

19
And even this modest evidence has been subjected to sharp criticism (Lindbom 1999). In Sweden, the case
considered by Clayton and Pontusson, there is little sign that social service employment has been targeted at
the expense of transfer programs. In some cases, it appears that services have actually been expanded. In
others, notably health care, there have been drops in employment but these appear to be linked to efficiency
gains rather than declining quality of service.
25
effort to determine if hypothesized cleavages over the welfare state are actually evident in
public opinion Svallfors (1997, p. 292) concludes that " public- versus private-sector
employment does not seem to constitute a particularly important fault line. Differences are
seldom statistically significant, and attitudinal differences point in different directions on
different indices, sometimes showing public-sector employees to be more in favour of
redistribution and small income differences than private-sector employees, sometimes the
opposite.
20
None of the issue area chapters in this volume find evidence of such an alliance
of tradeable interests; nor do the chapters by Manow and Rhodes, which concentrate
specifically on the contribution of economic actors to the formation of reform coalitions.
Why is there so little evidence for a claim that seems both credible and has received
considerable scholarly support? The answers say a good deal about the current social
democratic regime, and therefore help to provide a firmer basis for analyzing its contemporary
politics. First, the core assumption that productive workers will rebel against
unproductive public sector ones is based on a doubtful interpretation of the role of public
sector workers in the social democratic economies. Although often left implicit, explications of
the tradeable/non-tradeable cleavage bear strong affinities to more extreme neo-classical
treatments of the public sector which treat it as a domain dominated by rent-seeking. From
this perspective, a modestly-paid public sector worker engaged in, say, the care of alcoholics
or disabled children is esentially a parasite on the productive elements of society. Over time
one would expect these productive elements to repudiate the extraction of rents and seek to
rollback the public sector.
As Huber and Stephens (and, perhaps somewhat inconsistently, Esping-Andersen 1999) argue,
there is a strong case to be made for the beneficial economic effects of large-scale public
service provision in the context of the broader social democratic regime. Many of these
services focus on the development of human capital. Active labor market policy, training, and
education provide crucial support for skill development, which in turn has made a key
contribution to the high standards of living in the social democratic countries. At least as
significant, however, is the success of this regime in supporting the efforts of many women to
combine caring responsibilities, including child-rearing, with active participation in economic
life beyond the household. Both the direct expansion of economic activity related to these
policies, and the contribution that they make to fertility, represent considerable social assets.
All of this is missed by standard analyses of the cleavage between exposed and sheltered
workers. For accounting purposes, economists often treat the outputs of public services as
equivalent to the inputs, but we need to clearly distinguish our inability or lack of interest in
measuring productivity/output from the lack of productivity/output. The evidence in public
opinion continues to show that most voters in the social democratic countries perceive public
social services as important contributors to their quality of life.
There is a second problem with the exposed/sheltered cleavage argument. Ironically for an
analysis based around classes or class fragments, the focus on cross-class alliances is strangely

20
He adds a modest qualifier for the social democratic cases: some support is found for the hypothesis about
the sector cleavage being most important in the social democratic regime. It is only in Norway and Sweden
that sector appears to have any importance at all when explaining (re)distributive attitudes, but even here
differences are not very substantial."
26
abstracted from social context.
21
Private sector workers, and their public sector counterparts,
are treated as isolated monads. Yet we know that the labor markets in this regime are highly
stratified by gender, with male workers overwhelmingly preponderant in the private sector and
women concentrated in the public sector. In short, this hypothesized political cleavage is
likely to run right down the middle of many households. If a private sector worker is married
to a public sector worker, and they send their children to publicly-financed child care, then
what is the likely alliance pattern? Here is one clear advantage of bringing gender and
household structure to bear on political economy approaches. It shows why some of the basic
arguments about cross class coalitions in welfare state restructuring may be far weaker than
expected.
It is worth stressing that my arguments do not deny the relevance of the public/private divide
with respect to wage bargaining processes. The division of the share of economic output
going to labor has a zero-sum quality, and bargaining over this share is organized by unions
structured along sectoral lines. Thus the public/private cleavage is likely to be significant, and
it is understandable that this cleavage occupies the attention of industrial relations scholars.
What is in dispute is the extent to which this will carry over to the political determination of
the social wage in a context where the electoral realm constitutes an important domain of
interest articulation. Here, evidence for the activation of a cross-class, private-sector-based
coalition remains meager. Tendencies for wage-bargaining cleavages to spill over into welfare
state politics are partially if not totally offset by the strong reliance of almost all Swedish
households on public sector services. These services either directly provide employment or
indirectly provide the means to juggle demands of work and family.
If the political agenda is not one of undercutting public sector workers, what, then, have been
the pressing restructuring issues in the social democratic countries? Overwhelmingly, they
focus on cost-containment and the recalibration of programs within an essentially fixed budget.
By contrast, recommodification has generally been a less important priority. This reflects the
fact that while social democracy has traditionally been strongly de-commodifying, it has
done so in a manner that produces very high rates of labor force participation and
employment. Pushing (or pulling) people into the labor market has always been part of the
social democratic model. What have been distinctive are the quite favorable terms on which
this occurred. Although what recent Swedish reformers have called the work-line (everyone
who can be working should be working) has been strengthened in all these countries, these
changes have generally accentuated what had always been high expectations about work effort
within the social democratic model. Much of the reform agenda has focused on rationalizing
programs which, either because of new labor market conditions or because of program flaws,
had gone astray from broadly accepted goals. As noted earlier, sickness pay in Sweden
constitutes a clear example.
In most respects, however, the social democratic countries have had less need to focus on
issues of recalibration certainly when compared to the conservative regime which I will
discuss in a moment. In particular, there has been relatively little need for what I have termed
updating the adjustment of old welfare states to new needs and demands. It should be
emphasized that social democratic social policies already generate outcomes that the
conservative continental countries can only dream of achieving. The social democratic regime

21
In this respect as well, the focus on cross-class alliances reflects the strange marriage of post-Marxism and
neo-classical economics.
27
sustains very high rates of labor force participation for both men and women; it supports
comparatively high fertility levels; it has extensive experience with active labor market policies
designed to restrict exclusion and enhance skills; and it contains none of the clientelistic
remnants that constitute such a costly problem on much of the European continent.
If the social democratic welfare states have succeeded in restoring their fiscal equilibrium over
the medium-term, their biggest long-term problem is reconciling the need for continuing cost-
containment with the maintenance of solidarity around the welfare state. As Rothstein
(1998a) has forcefully argued, the political foundations of the social democratic welfare state
have depended both on its compatibility with acceptable economic performance and on the
ability to generate high quality services for middle and upper middle income households.
Sustaining the allegiance of these crucial voters, whose economic resources might provide an
exit option, requires two difficult balancing acts. First, increasing heterogeneity in
consumption tastes requires public sector services that provide meaningful choice, but without
generating "creaming or tipping effects which lead to class polarization in service
consumption. Second, high quality services must be sustained in a context where taxes are
already high but Baumol effects which accompany most social service provision generate
continuing pressures for new spending. It is these dilemmas, rather than an effort to erode the
position of public sector workers as part of a general liberalization, which have largely dictated
public sector service reform in the social democratic countries.
The Conservative Regime. This third cluster of countries, variously termed conservative,
continental, Christian Democratic, or Bismarckian, includes Austria, Belgium, France,
Germany, Italy, and the Netherlands. The welfare state configurations of these countries
generally include the following features: high levels of spending, strongly weighted towards
transfers in general and pensions expenditure in particular; limited development of public
sector employment; strong support for early labor market exit through early retirement
schemes or unemployment and disability programs (Ebbinghaus 1999); high levels of payroll
tax financing; and explicit and implicit family policies (including tax structures, the absence of
child care arrangements, and weak support for service sector job growth) which discourage
womens labor force participation. These features are linked to a status-preserving
Bismarckian benefit structure, and, in many cases, to occupationally segregated (and often
highly variable) pension and sickness funds. Unions and employers often play an important
institutional role in the administration of these funds. In some cases (e.g. Italy), Conservative
welfare states have also provided a strong basis for clientelistic welfare state politics (see
Ferrera 1996).
The political constellations associated with these welfare state regimes are varied. Yet some
generalizations are possible, and it should be stressed that in combination these generalizations
suggest a political configuration quite different from those appearing in the other welfare
regimes. Although not all these countries (e.g. France) could be defined as corporatist,
systems of organized interest intermediation are generally important: across Europe, four out
of five workers receive wages that reflect the outcome of a process of collective bargaining
(Golden, Wallerstein and Lange 1999, p. 204). In many cases, such as France, organized
labors role in social insurance confers a degree of influence which it might not be able to
achieve through wage bargaining systems (Bonoli and Palier 1996). These systems differ,
however, in the extent to which the social partners are willing and able to negotiate over broad
issues of relevance to the welfare state. At the same time, due to the modest role of public
28
social services, public sector unions are smaller and politically weaker than in the social
democratic countries.
The electoral systems of these countries tend to generate multiple veto points. The prevalence
of proportional representation (although restricted in some countries) makes single-party
governance rare. Federalism (Germany, and increasingly Belgium) and the institutionalized
role of the social partners reinforce tendencies towards negotiated reform. A final feature of
the political institutional setting deserving emphasis is that all of these countries are members
of the European Monetary Union and subject to the Maastricht criteria. Given the fiscal
circumstances discussed below, this membership may be of considerable importance.
As Kitschelt argues, party systems and public opinion in these countries have also helped to
consolidate political support for the welfare state. Social Democratic parties are far weaker
than in Scandinavia but generally influential. The center-right in many of these countries have
been dominated by Christian Democratic parties, which played leading roles in constructing
these welfare state models (Kersbergen 1995). Despite including factions that are now more
critical of extensive social provision, these parties contain significant elements committed to
the basic contours of the social contract. Given the scale of these welfare states, it is not
surprising that public opinion remains broadly supportive (Svallfors 1997). There is, however,
probably less enthusiasm than in Scandinavia, and greater ambivalence where the inequities
associated with clientilistic elements of social provision have been most in evidence (Ferrera
1997).
Although they might disagree on the root sources of pressure, there is a fairly wide consensus
among commentators that these continental countries face the greatest demands for
adjustment in their core welfare state arrangements (Esping-Andersen 1996; Esping-Andersen
1999; Scharpf 1997). Esping-Andersen has provided an elegant diagnosis of the interlinked
problems: high and rising costs associated with generous pensions and subsidized labor
market exit; low womens labor force participation; disturbing rates of unemployment among
the unskilled (linked to the high fixed labor costs associated with heavy payroll taxes); and low
fertility rates.
As in the other regimes, there are powerful complementarities among these elements. In the
current economic and social climate, however, these complementarities appear to work in a
pernicious direction. High social insurance charges, especially to finance pensions,
simultaneously block development of private service employment (by creating high wage
floors) and public service employment (by generating fiscal overload). In turn, the
unemployment and stagnation of labor force participation rates undercuts the revenue base for
pension systems and creates political pressures to subsidize labor market exit through social
insurance programs, especially early retirement and disability schemes. All of these countries
responded to initial employment problems by easing entry to various tax-financed paths out of
the labor market (Kohli et al. 1991, Ebbinghaus 1999). Employment stagnation and increased
reliance on transfers in turn precipitate further increases in payroll taxes. Compounding the
difficulties, the blockage of public and private service development jeopardizes the capacity of
women to combine labor force participation and child-rearing. This contributes to declining
fertility rates, which further imperil the long-term fiscal equilibrium of these systems. Thus the
problem is not just that these countries face major pressures for adjustment, but that the
possibilities for a bad equilibrium of low employment and low fertility seem very real. In
29
Esping-Andersens gloomy synopsis, these systems find themselves locked into a self-
reinforcing negative spiral (1996, p. 68).
The countries in the continental regime face two quite distinct but potentially complementary
reform agendas. First, there is a need to expand employment opportunities, which in the
postindustrial world means service employment. This aspiration confronts a series of
interconnected obstacles. Several impediments hinder womens efforts to combine paid-work
and child-rearing (Gornick et al 1998). Already very high rates of taxation render the social
democratic path of social service employment expansion (which would also facilitate the
movement of women into other forms of employment) implausible. High fixed social charges
and reservation wages limit private service sector growth, especially for low productivity (and
hence low pay) services (Scharpf 1997).
The second agenda is to achieve cost-containment in the main social insurance programs --
pensions, health care, and disability. The continental systems are transfer-heavy, and in their
most extreme form (e.g. Italy, but to a lesser extent France and Germany) approach the status
of pensioner states. The combination of demographics and existing commitments generates
projections of an even more severe shift in that direction. In most of these countries, health
care has also been partly financed from social insurance contributions. These high levels of
taxation strain public tolerance while crowding out fiscal space that might allow some social
service expansion. As just noted, high social insurance contributions also raise fixed labor
costs for firms, undercutting prospects for private sector job growth. Successful pursuit of
cost containment would make an important contribution to the task of expanding service
employment.
Thus the continental welfare states face an imposing list of reform demands, centering on cost
containment and on a series of recalibrating adjustments to diminish reliance on payroll taxes,
expand employment, and facilitate the efforts of women to combine paid employment with
caring responsibilities. Needless to say, this reform agenda confronts formidable political
challenges. The major social insurance programs create powerful vested interests. Serious
efforts to reduce expenditures by cutting benefits or tightening eligibility can threaten very
broad voting blocs. It is these types of reforms that prompted the large public demonstrations
in France, Italy, Germany, and the Netherlands mentioned at the outset of this chapter. Nor is
it difficult to point to governments (Jupp in France, several coalitions in the Netherlands,
Berlusconi in Italy) which subsequently paid large electoral costs for unpopular retrenchment
initiatives. In addition, as shown in Woods discussion of Germany, efforts to open the labor
market for low-paid work may confront challenges from the social partners. It should be
underscored that the resistance come not just from unions, but often from employers as well.
Having adapted to many aspects of the high wage-high skill regime and benefited from the
welfare states subsidization of cost-shifting efforts, employers too may resist or only half-
heartedly support labor market reforms.
Esping-Andersen (1990) has depicted the new cleavage in the conservative regime as one
dividing insiders from outsiders. As in the investigation of the social democratic welfare
regime, however, a simple political economy approach is likely to err by exaggerating the
political significance of cleavages that run right through the middle of households.
Outsiders (e.g., unemployed women and young men) are often economically dependent on
insiders, and this may in fact intensify resistance to anything threatening the existing social
welfare and employment regime. Indeed, Esping-Andersens recent acknowledgment (1999,
30
p. xxx) that political mobilization of outsiders was unlikely probably underlies his extremely
pessimistic tone about these countries.
Yet there is a vigorous politics of restructuring emerging in the continental regime, captured
by the concept (if not all the details) of Blair and Schroeders recent declaration of a new
middle (neue Mitte). The initial framework outlined in Part I presented the underpinnings
for this political depiction. The basic claim made there that welfare states generate both very
strong support and face very severe pressures applies especially clearly to the continental
regime. The dominant political struggle in the continental countries is not between insiders and
outsiders. Instead, it is between those who seek to create the foundations for a modernized,
fiscally and politically viable system of social provision and their opponents, who generally
advocate either the preservation of the status quo (standing pat) or a more radical program
of liberalization (crashing through).
Thus the central political issue in continental Europe during the past two decades has been
whether or not particular countries could develop the capacity to initiate and sustain such a
new middle coalition of welfare state restructuring. I begin by characterizing the pattern of
cross-national outcomes at the end of the 1990s, before turning to the issue of explanation.
Regarding outcomes, the consensus view appears to be that the Netherlands represents a clear
case of successful reform (the so-called Dutch miracle), with Italy deserving the second
prize. By contrast, Belgium, France and Germany represent the frozen welfare state
outcome, showing so far very modest capacities for restructuring.
22
Particular emphasis is
given to the contrast between Dutch success and German failure (Hemmerijck and Manow
1998). Finally, Austria is often left to one side in these comparisons; its economic
performance, and in particular its success in restraining unemployment, has so far generated
only modest pressures for restructuring.
As I have argued earlier, establishing the pattern of reform outcomes across countries and
over time is an extremely difficult task. Doing so systematically for these cases is well beyond
the scope of this essay. Rather than attempting it, I wish to focus briefly on two revealing
errors in the consensus judgment. On the one hand, it is overly generous in depicting the
achievements of reform coalitions in the Netherlands and Italy; on the other, it exaggerates the
extent to which the German welfare state has been locked into a politics of the status quo.
Derided as economic basket cases until recently, both Italy (Ferrera and Gualmini 1999;
Baccaro and Locke 199x) and the Netherlands (Visser and Hemmerijck 1997) have since
initiated quite substantial restructuring efforts. To deflate the depictions of miracles is not
to deny the significance of the reforms carried out in each case. Consider one telling statistic.
Policies in place prior to the Italian reforms of 1992, 1996 and 1997 had been projected to
generate public pension spending of over 23% of GDP by 2035; current projections estimate
that spending will peak at just under 16% (from the current level of about 14%) in 2030
before beginning to decline (Ferrara and Gualmini 1991, Table 1). In the Netherlands, policies
of labor market reform and negotiated wage restraint have generated a big increase in
employment, especially part-time employment. There have been vigorous efforts, so far with

22
See especially the country chapters in the Scharpf/Schmidt project and, for an explicitly comparative
assessment, Ebbinghaus and Hassel 1999. Levys (1999) account of the French case has a more optimistic
tone, but he acknowledges that his review of policy changes is highly selective, designed to illuminate a
particular style of reform rather than to systematically assess the overall trajectory of policy change.
31
uncertain results, to diminish reliance on disability and early retirement exit paths from
employment, including a restructuring of disability programs that greatly enhanced the states
oversight role.
The comparative assessment of reform performance is heavily dependent upon the analysts
choice of time frame. Too often, it is also based on an implicit assumption that the countries
being compared are starting from the same point. Yet if one takes the Dutch and Italian cases
from 1973, the date usually adopted to demarcate the end of the Golden Age, then ones
evaluation diverges from the consensus view. In both cases, conditions first dramatically
deteriorated as policymakers resorted to a variety of short-term expedients, initially including
further programmatic expansion, and later reliance on heavy borrowing and subsidized labor
market exit. Arguably, the main achievement of the reform initativbes beginning in 1982 in the
Netherlands and in the early 1990s in Italy has been to partially undo the policy damage done
in prior years. Despite its jobs miracle, Dutch rates of labor force participation are still
below the German level. Much of the heavily-touted expansion of womens participation
appears to involve very part-time jobs. If one measures the number of hours worked in the
paid labor market, rates for Dutch women remain at the very bottom end for European Union
countries (Daly 1999). Italy, despite notable reforms, continues to have the highest level of
pension spending in the affluent world, a very heavy debt load, and alarmingly low rates of
womens labor force participation and fertility. Evaluating these reform experiences requires
the perspective famously attributed to Samuel Johnson, who, upon witnessing a circus dog
walking on its hind legs, remarked that it is not done well, but one is impressed that it is done
at all.
Adopting a common baseline also suggests a kinder interpretation of the German case, which
is widely regarded as an instance of failed adjustment (Manow and Seils 1999). As Wood
argues, efforts to reform German labor markets roughly fit the conventional story. In a
political system marked by veto points, efforts at reform have generally failed to overcome the
resistance of organized interests committed to a deeply embedded labor market regime. Other
aspects of the German welfare state, however, have seen rather considerable adjustments,
particularly oriented around the goal of cost containment. Giaimo demonstrates that while
health care reform came slowly and with difficulty it did come. By the early 1990s, an
informal coalition involving the two major parties as well as unions and employers was able to
impose significant new cost containment on providers, along with modest user fees. Two
major quid pro quos, each an important instance of recalibration, made this painful
restructuring acceptable. The reforms reduced inequities among sickness funds (a long-term
goal of the unions) and allocated a considerable share of reform savings to a new program of
long-term care insurance.
In the case of pensions, there have been a series of major policy changes since the mid-1970s,
all of which strived to control costs. Each successful reform involved extensive negotiations
between the major parties and the social partners.
23
As Alber (1998) shows , the upshot of all
these efforts was not to reduce actual levels of pension spending or pensioner benefits from
previous levels, but to essentially eliminate an extensive expansion of pension costs which had
been built into previous pension policy. Pension benefits are roughly 30% lower than they

23
In at least one case, employers and unions made a joint submission of their position to the Commission
charged with developing the reform.
32
would have been if the policies in place in the mid-1970s had not been repeatedly modified.
24
Approximately half of the savings came from policies of the SPD-FDP government in place
until 1982, and about half can be attributed to the Kohl governments which succeeded it.
Efforts at cost containment have also taken place in other programs, although Alber shows
that the worst-off (e.g. families dependent on social assistance) have been treated relatively
gently.
German restructuring has been discounted not only because in some areas it has been
unimpressive (especially labor market reform), but because it has been undramatic.
Furthermore, the magnitude of change has been partially obscured by the massive shock of
German unification, which generated a severe fiscal disequilibrium. Under these exceptional
adverse circumstances (which, unlike the case of the Netherlands and Italy were largely
unavoidable), however, Germany has arguably done quite well. Germany has managed to
keep social expenditures share of GDP virtually level without sharply reducing benefits for
the poor while at the same time introducing some important recalibration of core welfare state
practices.
Thus in my view, the Dutch, Italian and German cases should all be classified as cases of
moderately significant restructuring. Each combines a focus on cost containment with some
efforts at recalibration. The reform paths differed, with the Netherlands and Italy first creating
large problem loads, prompting economic crises, which then triggered the formation of
political coalitions with a clear commitment to sweeping reforms. In Germany, reform has
been more continuous and incremental. All these cases contrast with the French and Belgian
experiences, where restructuring initiatives have generally been half-hearted.
In all three of the cases of substantial reform, the crucial political dynamic has been the
construction of a new middle coalition. The negative dynamic associated with these welfare
state regimes prompt calls for fundamental reform. Yet the political configurations are such
that serious reform generally requires careful negotiation among all, or at least most, major
political actors not only to surmount veto points, but in order to provide legitimacy for the
imposition of losses and assure successful implementation. In the Netherlands and Italy this
process has been apparent. In each case, broad centrist coalitions of varying shades have
succeeded in convincing key actors, including voters, of the necessity of reform and have
introduced major changes. Crucially, in each case vigorous efforts were made to provide
reassurances that the ambition of reformers was to save, not destroy, the welfare state. In
each case, reforms of social policy concentrated on costly elements (pensions in Italy,
disability programs in the Netherlands) that were both damaging to efficiency and normatively
difficult to justify. Given the fiscal difficulties facing each country, opportunities for quid pro
quos were limited, but in each country coveted enhancements of underdeveloped aspects of
social provision rendered reforms more palatable. Finally, in each case efforts were made to
support the most vulnerable; so far, considerable cost-containment has not been accompanied
by significant increases in poverty.
In Germany as well, successful reform reflects the employment of new middle methods of
elite negotiation and mutual legitimation. Both major parties, along with the social partners,

24
The German case thus illustrates the magnitude of loss imposition that is possible in pensions policy if
policymakers pursue long-term goals, gain broad elite consensus, and do not markedly cut the real benefits of
current recipients.
33
have been active participants in sustained dialogue over major structural reforms. What was
different in Germany was the informal, intermittent and compartmentalized character of the
process. In contrast with the Dutch and Italian cases, reform took place on a policy-by-policy
basis rather than through a single, self-described and well-publicized campaign of overarching
reform.
Thus the central issue for analysts of the conservative regime is to establish the factors
determining the viability of such a new political middle.
25
At this stage, we can probably only
begin the process of developing a reasonably convincing set of hypotheses. One can usefully
distinguish between factors influencing the demand for a negotiated political response and
those factors influencing the capacity of the system to generate that response. On the demand
side are, especially, conditions that convince significant actors that standing pat or crashing
through are not viable options. Marginalizing those who advocate crashing through will
often be easier. Given prominent features of the conservative regime, most important political
elements are likely to be skeptical about the desirability of full-fledged liberalization in any
event. Even for those who might consider such an outcome desirable in principle, the clear
signals of public opinion, or the evident realities of veto points, may be sufficient to generate
skepticism about crashing through. There may also be a need for political learning, however,
through failed attempts at unilateral liberalization. Berlusconis negative example appears to
have played such a role in Italy, for instance.
Perhaps more critical is the determination by key political actors, especially in unions and left
parties, that standing pat is not a viable option. These actors may potentially (although not
always) have relatively long time horizons, which allows them to pay attention to issues of
policy sustainability (e.g., of old-age pension systems). Crucially, as Ross (1998) and Levy
(1999) have emphasized, they are also politically positioned to sell reforms as both necessary
and intended to revitalize the social contract rather than dismantle it.
26
What convinces these actors of the need for reform? While a sense of policy imperative is
always socially constructed, objective circumstances are likely to play a critical role. Well-
established and highly valued policies, often the fruits of long political labor, are unlikely to be
jettisoned easily. In the words of Maurizio Ferrera, actors must come to believe that there is
no status quo
27
that existing arrangements simply cannot be sustained. This is likely to
involve both social learning (about what the effects of particular policies will be) and
political learning (about the probabilities of achieving alternative policies). Social learning
may be driven by focusing events (such as dreadful performance on key indicators) or financial
crisis (such as a run on the currency). A more gradual process of learning, in which trial-and-
error leads to the discrediting of alternative strategies, may also be crucial (Visser and

25
Aspects of the following analysis have some relevance for cases in the other regimes as well. Yet as
discussed in the conclusion, those configurations vary because they generally face considerably more modest
pressures for restructuring, and, in the case of the liberal countries (with the partial exception of Australia),
because they have a much lower capacity for negotiated adjustment.
26
Typical is the rhetoric accompanying German Chancellor Schroeders signing of the third way, neue Mitte
document with Tony Blair, which called for major welfare state restructuring including cost containment and
labor market reform. In the words of a close aide, there is no other way If we dont take steps to reform
the welfare state, it will collapse. Another adviser stressed that we must gradually convince people there is
no alternative. Financial Times, June 9
th
, 1999, pp. 1, 9.
27
Personal communication.
34
Hemmerijck 1997). Both these processes may be at work, with a period of cumulative
learning punctuated by a focusing event that triggers the initiation of reform.
Political learning is likely to involve a similar trial-and-error process as actors explore the
viability of alternative strategies. Here a plausible case can be made that political learning is
most likely to lead to a new middle outcome where left parties and unions are moderately-
strong. Moderately-strong actors possess resources that are significant enough to place them
at the bargaining table, but not so substantial that they can be confident that other actors (a
right of center government or employers) will not try to go it alone if they attempt to
exercise a veto.
28
This, for instance, was clearly the case in both the Netherlands (Hemmerick
and Visser 1997) and Italy (Ferrera and Gualmini, 1999), where decisions of left and
center-left actors to negotiate and on what terms occurred in the shadow of possible
exclusion (Scharpf 1998). In the Netherlands and Germany center-right parties have been
integral participants in reform efforts. Thus while Levy and Ross rightly argue that left
participation in these restructuring processes is politically critical, they are wrong to portray
these initiatives as essentially projects of the political left.
Yet if predispositions for reform among key actors are important, so are conditions which
make the supply of such restructuring initiatives possible. Success requires that the relevant
social actors and political parties have both the desire and capacity to reach agreements and to
make those agreements stick. A great deal of the traditional work on corporatism, which
considered similar problems in the field of industrial relations, is relevant here. Two factors,
however, appear crucial. The first factor is the quality of links between the social actors and
the policymaking process, which facilitate the effective implementation of agreements. The
second is the ability of elites to maintain sufficient membership support for unpopular
initiatives. As Kitschelt argues in this volume, this is an issue not just for unions but for
political parties as well. Much is likely to turn on the extent to which potential negotiators
find their organizations vulnerable to poaching from competitive organizations (or
competing factions within their organization) who could exploit the opportunities created by
efforts to impose losses.
29
At first glance, this list of preconditions seems to imply that the prospects for negotiated
reform would be bleak. After all, the conclusion of much of the traditional literature on
corporatism was that breakdown was likely in the absence of relatively high levels of
organizational concentration. Yet in fact, as Rhodes shows, there has been a rather striking
sprouting of social pacts in recent years, even (or especially) in settings that might not have
been considered fertile soil. In the current climate, those seeking a negotiated solution possess
two significant advantages. First, in line with the analysis offered in Figure 1, they are likely to
straddle the political space occupied by the pivotal voter. This position gives these elites

28
Of course, the continental countries, except France, generally score in the middle range on most scales of left
power resources.
29
Both France and Belgium fall short in these respects. And unlike the case in the liberal regimes, the capacity
of these two countries to concentrate political authority was not sufficient to generate major reforms. As
Bonoli notes, French political authority is concentrated. In Belgium, governments ruling by decree were able
for some time to exploit a similar concentration of formal authority (Scharpf 1999). In both cases, however,
efforts to impose rather than negotiate reforms ultimately floundered in the face of widespread political
opposition and the inability to implement key changes without consent. This underscores my earlier
suggestion that the effects of concentrated political authority will be different within the conservative regime
than they are in the liberal one.
35
considerable leverage, especially, as Kitschelt argues, if they are willing to trade off some
erosion of membership support for sustained influence over policy.
The second advantage, as Jonah Levy has eloquently argued, is the possibility for turning vice
into virtue. In all these welfare states, Levy notes, there are opportunities to target
inequities within the welfare system that are simultaneously a source of either economic
inefficiency or substantial public spending [I]nherited welfare vices can be manipulated so
as to soften or even obviate the supposedly ineluctable tradeoff between efficiency and equity
(Levy 1999, p. 240). In the terms employed in this essay, restructuring packages need not be
limited to cost containment and recommodification. They may also include substantial efforts
at recalibration. As has been noted, the need for updating in many of these systems is evident,
which may facilitate legitimation while generating crucial fiscal resources. When orchestrated
through a broad, centrist coalition (formal or informal), which has credibility in claiming that
the goal is restructuring rather than dismantling and that there are no viable alternatives, this
can be a genuinely formidable political formula. In practice, the Dutch, German and Italian
cases all reveal a modest but very real possibility for the negotiated restructuring of
conservative welfare states.
The argument presented in this section is summarized in Table 2. The three worlds of welfare
capitalism differ in critical respects, which generate quite different political dynamics of
welfare state restructuring. Most crucially, the different regimes vary in both the scope of
support for the welfare state and in the scale and nature of adjustment pressures. In the liberal
welfare states, adjustment pressures have been moderate, but so has been support for the
existing welfare state. In the social democratic welfare states popular support is widespread
and adjustment pressures have also been moderate. In the conservative welfare states, popular
support of the existing regime is strong but so are pressures for adjustment.
[Table 2 About Here]
These differences in the depth of popular support and the scale and nature of pressures in turn
generates different political alignments, different plausible reform trajectories, and (somewhat
more speculatively) different critical variables in the restructuring process. In short, there is
not a single new politics of the welfare state, but different politics in different
connfigurations. In the liberal world reform focuses on cost-containment and
recommodification, with the crucial divide between those advocating thorough-going
neoliberal retrenchment and those seeking a more consensual and compensatory solution. In a
context where support for social provision is relatively thin and the electoral/legislative arena
plays a dominant role, partisan control of government and the degree to which institutions
concentrate political authority become critical factors. In the social democratic world, the
focus of reform is on cost-containment and recalibrations, which aim at rationalizing programs
to enhance performance in achieving established goals. On the whole, reform has been
negotiated, consensual and incremental. In the conservative world, reform has centered on
cost-containment and recalibration, with a heavy emphasis on updating old programs to
meet new demands. Neoliberal retrenchment is not a viable option, so politics centers on
efforts to construct a viable reform coalition. The structure of interest intermediation and
party systems in particular, whether they allow reformers to overcome their fears of
poaching are crucial.
36
Two final observations about this summary statement underscore key features of the
contemporary politics of social policy. The first, as indicated in Table 1, is that a critical factor
in structuring reform in all three configurations is the severity of economic pressures. Intense
disequilbria have a major impact on reform dynamics. These disequilbria may or may not stem
primarily from globalization but it is important to stress that it is not politics alone that
determines how or when countries undertake significant reform initiatives. By implication, the
future fate of mature welfare states is likely to be dependent upon the economic performance
of the particular countries to which they are joined. This volume has offered grounds for
skepticism about the deleterious economic effects of extensive social provision. However, if
these negative effects are as severe and persistent as some critics allege, more wide-ranging
reforms seem likely.
The second observation is that while reform agendas vary quite substantially across regime
types, all of them place a priority on cost-containment. This shared emphasis reflects the onset
of permanent austerity. Welfare states are not being dismantled. Efforts to achieve
recalibration can generate interesting innovations and even extensions of social provision. Yet
everywhere, such adjustments occur in a context where the control of public expenditure is a
central if not dominant consideration. The core structures of most welfare states are not in
jeopardy. Nonetheless, the contemporary climate remains a harsh one for efforts to improve
social provision for the vulnerable, or to address newly-recognized risks.
37
TABLE 1: THREE WORLDS OF WELFARE STATE REFORM
LIBERAL SOCIAL
DEMOCRATIC
CONSERVATIVE
Political Support
for Welfare State
Moderate High High
Adjustment
Pressures
Moderate Moderate High
Reform Agenda Recommodification/
Cost-Containment
Cost-Containment/
Recalibration
(rationalization)
Cost-Containment/
Recalibration
(updating)
Line of Conflict Neo-liberal
retrenchment vs.
Compensated
commodification
No dominant
cleavage.
Negotiated,
incremental
adjustment
Stand Pat vs.
Negotiated Reform
Distinct key
Variables
Severity of
economic pressures;
Concentration of
political authority
Severity of
economic pressures
Severity of
economic pressures;
Vulnerability of
centrist reform
organizations to
poaching