B.J.Pol.S.

34, 693–724 Copyright © 2004 Cambridge University Press

DOI: 10.1017/S0007123404000250 Printed in the United Kingdom

Review Article: Dismantling the Health Care State?

Political Institutions, Public Policies and the
Comparative Politics of Health Reform
JACOB S. HACKER*

This article examines the recent pattern and progress of health care reform in affluent democracies, focusing
in particular on Britain, Canada, Germany, the Netherlands and the United States. Its main contention is that
efforts to reform health care in advanced industrial states have been marked by a paradoxical pattern of ‘reform
without change and change without reform’, in which large-scale structural reforms have had surprisingly modest
effects yet major ground-level shifts have, nonetheless, frequently occurred as a result of decentralized
adjustments to cost control. The main task of the article is to investigate the reasons for and effects of this puzzling
pattern by plumbing the largely unexplored theoretical territory between comparative health policy analysis and
cross-national research on the welfare state. Along the way, the article develops a simple model of the politics
of reform that helps explain cross-national variation in legislative and policy outcomes – particularly outcomes
that occur through decentralized processes of internal policy ‘conversion’ and policy ‘drift’, rather than through
formal legislative reform. It also takes up a number of other intriguing issues raised by recent trends: why, for
example, market reforms are clustered in centralized political and medical frameworks; why these reforms have
generally enhanced state authority rather than market autonomy; why, despite fragmentation, decentralized
political and medical systems shifted towards an expanded government role; and why significant retrenchment
of the public–private structure of health benefits occurred in the United States.

Over the past two decades, structural reform of national health policies has dominated the
political agendas of advanced industrial states as never before. Yet students of the welfare
state – who have intensively examined the post-1970s development of other prominent
realms of social policy – have devoted surprisingly little attention to these crucial disputes,
leaving the field almost entirely to health policy specialists. This is unfortunate, for the
recent progress of health care reform raises important puzzles not just about health policy,
but about the politics and future of the welfare state more generally.
The puzzle at the heart of this article is a paradoxical pattern of policy development that
I describe as ‘reform without change and change without reform’. Government-initiated
structural health reforms, I argue, have not produced much in the way of dramatic
upheavals, much less movement towards a common market-based model. And yet major
health policy shifts do in fact seem to be taking place within mature medical complexes
– and, in some important cases, exposing citizens to significant new costs and risks.
This paradoxical pattern, I suggest, is rooted in the interplay of political decision
procedures and the structure of the medical sector in a climate of budgetary austerity.
Governments that enjoy consolidated authority have sometimes found it possible to enact
major reforms. Yet they have had surprisingly little success in refashioning the highly
hierarchical and embedded medical structures that centralized governance tends to foster.

* Department of Political Science, Yale University. The author thanks Oona Hathaway, Alan Jacobs, Theodore
Marmor, Albert Weale and R. Kent Weaver, as well as several anonymous readers and participants in workshops
at Yale University and the University of Chicago, for comments and support; Nelson Gerew, Gina Kramer and,
especially, Nicole Kazee for indefatigable research assistance; and the William Milton Fund of the Harvard Medical
School and Yale’s Peter Strauss Family Fund and Institution for Social and Policy Studies for financial support.
694 HACKER

Conversely, in more decentralized political structures, rapid or decisive structural policy

change has proved far more elusive. Yet even when political processes foster stalemate,
consequential ground-level shifts may still take place through decentralized processes of
adjustment within existing policy arrangements. As a result, many of the most notable
policy changes of recent years have resulted not from formal reforms, but from the
long-term working out of cost pressures in medical complexes that provide very different
leeway for decentralized adaptations – a process that has created privatization at the
margins in universal public systems but significant regress in the United States. The
character and effect of these adaptations, however, vary markedly across different systems,
both in their aggregate terms and in their implications for the distribution of the burdens
of cost control among citizens.
To buttress this claim, I rely on two sorts of evidence. The first is basic cross-national
statistics on health spending and the characteristics of medical financing. The second is
a more detailed investigation of recent policy developments in five nations: Britain,
Canada, Germany, the Netherlands and the United States. These nations all devote a large
share of gross domestic product (GDP) to health care. They represent a wide range of
medical financing arrangements. And they have all faced considerable cost pressures in
the past two decades. Perhaps most importantly, they have had varying success in enacting
comprehensive reforms, and the changes that have taken place in their medical financing
structures have differed markedly. Finally, these cases each represent one of the four mixes
of political authority and medical financing that form the backbone of my examination of
the varied dynamics of structural reform: (1) centralized political decision making coupled
with centralized medical financing (Britain), (2) centralized decision making coupled with
decentralized financing (the Netherlands), (3) decentralized decision making coupled with
decentralized financing (Germany and the United States), and (4) decentralized decision
making coupled with centralized financing (Canada).
The article begins by exploring the roots and character of the fundamental differences
in these nations’ policies. It then examines the common pressures to control costs that all
these countries have faced. Next, the article considers the politics of structural reform,
outlining a simple framework for explaining the varied course of policy developments
across nations that emphasizes the interaction between structures of governing authority,
on the one hand, and structures of medical financing, on the other. This framework in turn
provides an analytic lens through which I consider when, why and how political and market
actors alter the major institutions of the modern health care state.

UNDERSTANDING AND EXPLAINING HEALTH POLICY VARIATION

Discussions of health policy usually bog down quickly in the sheer complexity of the field.
A recent text, for example, identifies ‘11 structural “parameters” or “dimensions” along
which the systems of various countries differ’.1 The Organization for Economic
Cooperation and Development (OECD) adopts a comparatively parsimonious scheme,
which involves just two key dimensions and eight distinct types, although these types are
then divided across three more continua.2 As Joseph White argues, however, this baffling
1
Francis D. Powell and Albert F. Wessen, eds, Health Care Systems in Transition: An International
Perspective (Thousand Oaks, Calif.: Sage Publications, 1999), p. 12.
2
Organization for Economic Cooperation and Development, The Reform of Health Systems: A Review of
Seventeen OECD Countries (Paris: OECD, 1994).
 Review Article: Dismantling the Health Care State? 695

diversity masks substantial similarity across rich democracies’ health programmes,

virtually all of which share two bedrock characteristics: they cover all citizens, and they
employ measures to contain costs at a high level of aggregation.3 Against this ‘international
standard’, only the United States looks like a conspicuous outlier, its public programmes
covering less than half the population, its overall spending largely unconstrained.
Among nations that uphold the international standard, the usual distinction is between
a ‘national health service’ and ‘national health insurance’. But this division is much less
pure than supposed. The British National Health Service (NHS) and Canadian national
health insurance are in key ways more alike than different, and both are quite distinct from
Germany’s system of non-profit sickness funds, which is also classified as national health
insurance. To capture the similarities, it is helpful to note that both are single-payer
systems, in which the government (in the Canadian case, the provincial governments) pays
for services directly. Germany and most other Continental European nations, by contrast,
have multi-payer systems based on insurance funds that pay for care within a public
regulatory framework. What distinguishes the British and Canadian systems is not so much
financing arrangements as the ownership of medical facilities, which are mostly public in
Britain and mostly private in Canada.
As Figure 1 shows, then, national health programmes divide fairly neatly across two
axes: the number of payers and ownership of facilities. Three of the four resulting
combinations capture the most common system types: a prototypical national health
service (single payer/public ownership), Canadian-style national health insurance (single
payer/mixed ownership), and German-style corporatist health insurance (multiple
payer/mixed ownership).4 Several other distinctions are closely associated with these two
axes. Single-payer systems tend to rely heavily on general tax revenues, whereas
multi-payer systems generally employ payroll-tax financing. Systems with public
ownership of medical facilities usually pay doctors at least partly on a salaried or
per-patient basis, whereas fee-for-service payment is the norm when ownership is mixed.
In public-ownership systems, too, a significantly larger share of spending is usually
governmental than in mixed-ownership systems. Finally, multi-payer systems generally
create the greatest diversity of coverage across citizens and, in some cases, even exclude
from statutory protection wealthier citizens, who are expected to buy into the public system
or insure themselves.
Why do the structures, if not the goals, of national health policies differ so starkly? One
obvious reason is that political parties have differed historically on the proper role of
government in medical care. In cross-national research, a well-supported finding is that
rule by parties of the left, particularly during the formative years of welfare state
development, is associated with more expansive and generous social programmes.5 Leftist
rule is certainly not a necessary condition for universal health care, as it has been adopted
under governments of varying partisan stripes. But it does appear strongly associated with
the establishment of national health services and, more generally, with a diminished role

3
Joseph White, Competing Solutions: American Health Care Proposals and International Experience
(Washington, D.C.: The Brookings Institution, 1995).
4
The fourth logical combination – multiple payer/public ownership – has no clear exemplars, though some
corporatist systems rely more heavily on public ownership than others.
5
See Evelyne Huber and John Stephens, Development and Crisis of the Welfare State: Parties and Politics
in Global Markets (Chicago: University of Chicago Press, 2001). Using Huber and Stephens’s dataset
(http://www.lisproject.org/publications/welfaredata) and OECD expenditure data, the correlation between
1945–75 cumulative left-party governance and the 1975 private share of health spending is ⫺ 0.58.
696 HACKER

Fig. 1. Major types of medical systems

for private insurance and direct consumer payments, which the left has long viewed as
inegalitarian.6
The scope for political leaders to achieve their favoured goals is heavily constrained,
however, by the structure of political institutions, particularly the opportunities for
blocking activity that institutions create for powerful opponents of national health
programmes like the medical profession. As Ellen Immergut has convincingly argued,
opponents of large-scale government entry into the health field have generally been
advantaged when a polity has a large number of ‘veto points’, such as federalism and a
separation of powers between the executive and legislature.7 This no doubt helps explain
why no federal state has adopted a national health service; why across nations the share
6
In the OECD, only Italy’s national health service was not enacted under social-democratic rule.
7
Ellen Immergut, Health Politics: Interests and Institutions in Western Europe (New York: Cambridge
University Press, 1992).
 Review Article: Dismantling the Health Care State? 697

of medical spending financed by government is strongly correlated with the number of

institutional veto points; and why Switzerland, with its strong federalism and tradition of
the use of popular referendums by organized groups, has long been characterized by the
most anemic government role in health policy of all European nations. It is also consistent
with the fact that the United States, which has the most veto-point-ridden polity of any rich
democracy, remains the only advanced industrial state that does not have a broad
framework of public coverage or cost containment and the only one that relies principally
on voluntary employment-based coverage.
Still, with the exception of the United States, all advanced industrial democracies have
adopted some version of the international standard. This suggests that institutional barriers
are better at slowing than halting government’s entry into the medical field. The timing
and sequence of policy interventions, however, may be highly consequential for the form
that national health policies ultimately take. Most countries began to intrude into the
doctor–patient relationship by subsidizing non-governmental insurers, rather than
financing services. These policies created important vested interests in a pluralist financing
structure and reinforced doctors’ preferences for fee-for-service payment. How extensive
and long-lived these arrangements were thus had crucial effects on the types of systems
countries ended up with.8 Countries in which authoritative government action to
consolidate or supplant non-governmental insurance took longer to achieve generally
ended up with more decentralized and costly health financing systems in which private
insurance and finance played a more pivotal role – in part because delay allowed the
formation and enrichment of a formidable collection of private stakeholders, and in part
because sophisticated private care represents such a massive burden for government
budgets to assume.
This is a paradigm example of path dependence, temporal processes in which early
choices create self-reinforcing effects that are inherently difficult to reverse.9 The United
States, again, represents an extreme case: private insurance has, in effect, come to play the
role that public programmes do elsewhere, and this role has proved as difficult to dislodge
as the public foundations of mature welfare states.10 Such path-dependent effects are an
important reason why analysis of the contemporary politics of health reform must take
seriously the constraints created by existing structures of medical finance. Even when
facing similar strains, governments differ greatly in the particular challenges and demands
that they confront and the particular policy tools they have at their disposal, quite apart
from the institutional constraints that electoral and decision-making systems create.

STRAINS FACING MATURE MEDICAL COMPLEXES

Regardless of their structure, national health policies all came under severe pressure in the
1980s.11 The reasons for this shift were both economic and political. In the economic realm,
8
Jacob S. Hacker, ‘The Historical Logic of National Health Insurance: Structure and Sequence in the
Development of British, Canadian, and U.S. Medical Policy’, Studies in American Political Development, 12
(1998), 57–130.
9
Paul Pierson, ‘Increasing Returns, Path Dependence, and the Study of Politics’, American Political Science
Review, 94 (2000), 251–67.
10
Jacob S. Hacker, The Divided Welfare State: The Battle over Public and Private Social Benefits in the United
States (New York: Cambridge University Press, 2002).
11
Paul Pierson, ‘Coping with Permanent Austerity: Welfare State Restructuring in Affluent Democracies’, in
Paul Pierson, ed., The New Politics of the Welfare State (New York: Oxford University Press, 2001), pp. 410–56.
698 HACKER

the 1970s ushered in a marked decline in rates of economic growth. In the political realm,
the period saw the emergence of anti-welfare state political movements, symbolized by
the ascendance of Margaret Thatcher in Britain and Ronald Reagan in the United States.
The effect of these linked developments was exacerbated by strains emerging out of
welfare states themselves: in many nations, core programmes had grown to the point where
avenues for further expansion were limited and the opportunity cost of pursuing them high.
Caught between enduring support for the welfare state and the fiscal demands posed by
these new realities, politicians everywhere faced hard choices, and many countries ran
large deficits to elide tough trade-offs.
As the second most expensive area of the welfare state behind public pension
programmes, health care was scarcely immune from these pressures. Indeed, while public
pensions often presented the greater long-term fiscal threat, the rapid inflation of health
spending was usually the largest and most immediate source of budgetary strain for
countries facing up to the new fiscal order. In one sense, this was nothing new – and had
more to do with the distinctive economics of medical care than with the particularities of
welfare states. For as long as health spending has been recorded, it has tended to grow at
a faster rate than general prices.12 But the persistent problem of ‘excess’ medical inflation
caused by technological change and weakly checked demand suddenly took on new and
pressing urgency in the straitened fiscal circumstances of the period.
Moreover, nearly all nations (again, with the conspicuous exception of the United States)
entered this harsh new era with public programmes in place covering all or most citizens.
Not only, then, was access no longer the central rallying cry, but the cost of medicine had
in most nations been effectively socialized. This is evident in the remarkably high average
share of total medical spending borne by OECD governments in 1980 – some 80 per cent,
excluding the United States. Although health costs are ultimately borne by society
regardless of the source of financing, it makes a profound political difference whether they
are financed by government or the private sector. An old adage of health policy is that every
dollar spent on medical care is a dollar of somebody’s income. In the 1980s, leaders came
to face another unpleasant truth: every dollar spent by government on medical care is a
dollar that cannot be spent on other ends (or, in systems financed by mandatory workplace
contributions, a dollar that crowds out other possible uses of payroll taxes and risks
distorting labour markets).
Many commentators have argued that ‘globalization’ created an additional heavy
burden on welfare states in the 1980s and 1990s. As a wave of studies has demonstrated,
however, the effects of increasing economic integration pale in comparison to the internal
strains that welfare states have faced, and even these real but surprisingly modest effects
have themselves been highly mediated by domestic political and economic institutions.13

12
This is so for least three reasons. First, in some highly labour-intensive and low-technology areas (such as
nursing), medical care presents a classic example of William Baumol’s ‘cost disease of personal services’, wherein
low productivity growth leads to rapidly rising costs relative to high-productivity sectors of the economy. In many
areas of health care, however, a second cause of high inflation dominates: increasing technological sophistication.
Thirdly, and finally, key features of the medical market – particularly uncertainty and risk aversion – empower
agents with an interest in higher spending (that is, doctors) and encourage reliance on third-party insurance, which
blunts the price signals that, in other sectors, help restrain inflation.
13
See, for example, Geoffrey Garret, Partisan Politics in the Global Economy (New York: Cambridge
University Press, 1998); Duane Swank, Global Capital, Political Institutions, and Policy Change in Developed
Welfare States (New York: Cambridge University Press, 2002); and Pierson, ed., The New Politics of the Welfare
State.
 Review Article: Dismantling the Health Care State? 699

All this seems particularly true in the health care sector, where fiscal strains are largely
caused by persistently high medical inflation. Global competition may well have provided
a rationale for assaults on national health programmes. But the distinctive sources of cost
pressure in health care and the internally generated fiscal strains that all social programmes
faced are sufficient to explain why control of health spending became a paramount issue
in the 1980s.
By the early 1980s, then, health care cost containment was a leading item on the political
agenda of all advanced industrial democracies and old political deals came under newly
intense strain. In the past, rapid growth had facilitated a generous framework of
accommodation with medical providers, in which tight control over fees was traded off
against decreased professional resistance to the socialization of finance. In the 1970s and
1980s, this bargain was called into doubt. Rather than ‘states and interests’ in uneasy
co-operation, the pattern became ‘states versus interests’ in the struggle to control costs.14
As Table 1 suggests, the result was a remarkable reversal of past spending trends – one
that carries surprisingly few simple messages about the effectiveness of different types of
medical systems in controlling costs. Indeed, the most striking feature of the comparison
between 1960–80 spending growth and 1980–2000 growth is the universality of decline.
Only Canada comes close to bucking the trend. Among OECD nations as a whole, the
percentage increase in health spending as a share of GDP in the second period is roughly
a quarter of what it was in the first. Two OECD nations, Sweden and Denmark, actually
experienced a decline in health spending as a share of GDP, and a number of countries
saw negligible increases. In the light of past trends, as well as the general impression that
health costs are inherently ‘uncontrollable’, this unambiguous and sustained response
bears emphasis.
How it is that countries once deemed incapable of overcoming perennial institutional
blockages suddenly found vast new reservoirs of power to rein in costs? The statistical
literature on health spending does not provide much guidance on this question. After
controlling for per capita GDP, long known to be the most powerful predictor of national
spending, most studies find a limited role for government policy in explaining variations
in expenditures across nations.15 These studies do call into question some common
assumptions – for example, that population ageing is a prime cause of higher spending
(based on cross-national analyses, it is not) – but their main message is that the effect of
per capita GDP washes out nearly all other influences. What these studies do not do,
however, is explain the apparent structural shift in the spending–GDP relationship around
1980. Why did spending grow so much slower relative to the economy after 1980 than
before?
The most easily identified reason is a major decline in professional power. If one were
to cast the development of national health programmes as an epic drama, the key actors
would be the medical profession and the state. In all industrialized nations, including the
United States, doctors ultimately lost the war – government insurance of some sort came
into being – but in all, professional self-regulation and generous payment were the price
paid to ensure doctors’ submission to state power. By the 1980s, however, political and
business leaders had almost universally come to see the cost of this concordat as too high.

14
David Wilsford, ‘States Facing Interests: Struggles over Health Care Policy in Advanced Industrial
Democracies’, Journal of Health Politics, Policy, and Law, 20 (1995), 571–613.
15
See the review of cross-national spending studies in Uwe Reinhardt, Peter S. Hussey and Gerard F. Anderson,
‘Cross-National Comparisons of Health Systems Using OECD Data, 1999’, Health Affairs, 21 (2002), 169–81.
TABLE 1 Health Spending in Selected OECD Countries, 1960–2000
700

Total Percentage Percentage Public medical Public medical Percentage-

health change in share change in share spending as spending as point
spending of GDP devoted of GDP devoted share of total share of total change in
as share to health to health health health public
of GDP, spending, spending, spending, spending, share, 1980–
HACKER

2000 1960–80 1980–2000 1980 2000 2000

(%) (%) (%) (%) (%) (%)

Australia 8.9% 70.7% 27.1% 63.0% 68.9% ⫹ 5.9

Austria 7.7 76.7 1.3 68.8 69.4 ⫹ 0.6
Belgium 8.6 34.4 72.1
Canada 9.2 31.5 29.6 75.6 70.9 ⫺ 4.7
Denmark 8.3 ⫺ 8.8 87.8 82.5 ⫺ 5.3
Finland 6.7 68.4 4.7 79.0 75.1 ⫺ 3.9
France 9.3 75.8
Germany 10.6 21.8 78.7 75.0 ⫺ 3.7
Iceland 9.3 106.7 50.0 88.2 83.7 ⫺ 4.5
Italy 8.2 73.4
Japan 7.7 113.3 20.3 71.3 77.7 ⫹ 6.4
Netherlands 8.6 14.7 69.4 63.4 ⫺ 6.0
New Zealand 8.0 35.6 88.0 78.0 ⫺ 10.0
Norway 7.6 137.9 10.1 85.1 85.2 ⫹ 0.1
Spain 7.5 260.0 38.9 79.9 71.7 ⫺ 8.2
Sweden 8.4 ⫺ 4.5 92.5 85.0 ⫺ 7.5
Switzerland 10.7 55.1 40.8 55.6
United Kingdom 7.3 43.6 30.4 89.4 80.9 ⫺ 8.5
United States 13.1 74.0 50.6 41.5 44.2 ⫹ 2.7
OECD Mean* 8.7 94.4 23.3 77.2 73.1 ⫺ 4.1

* Excluding Turkey, Ireland, Portugal, Greece and Luxembourg, as well as the five most recent additions to the OECD: Poland, Hungary,
Korea, Mexico and the Czech Republic.
Source: Organization for Economic Cooperation and Development, OECD Health Data 2003 (Paris: OECD, 2003).
 Review Article: Dismantling the Health Care State? 701

Stewards of public and private benefits increasingly conceived of themselves as ‘payers’,

rather than guarantors; and, as payers, their goal was to control what was paid.
Although national embodiments of this shift were distinctive, its general direction was
consistent and its cumulative results profound. In Britain, Thatcher waged a relentless
assault on the British Medical Association’s traditional consultative role. In Germany and
the Netherlands, successive legislative changes strengthened the sickness funds against the
providers and forced doctors to work within increasingly tight limits. In Canada, where
the collegial model of professional self-regulation had reigned supreme, the provinces
moved to cap total spending on physician services. Even amid the antigovernment 1980s,
the US Medicare programme adopted new fee controls. And, of course, American
physicians saw their power wane even further in the 1990s, as health plans contracted
selectively with providers, micromanaged doctors’ decisions and bargained-down fees.
The decline of professional influence is, however, only one manifestation of a broader
move by public authorities to slow the growth of medical spending via new regulatory and
budgetary controls. Even a cursory review of the legislative changes listed in Table 2
conveys both the frequency and increasing stringency of these measures. Medical cost
control has followed a ratchet-like pattern quite different from the zigzag of change and
reversal seen in some other areas of health policy. New state capacities for cost control,
once in place, tend to remain in place and, indeed, grow tighter and more comprehensive
over time. Thus authorities in Canada, Germany and the Netherlands introduced or
strengthened fee controls for physician services, set or tightened budget ceilings for
hospitals, and attempted to put in place relatively fixed budgets for specific sectors or areas
of health spending. The British National Health Service (NHS), financed as it is by general
tax revenues, has long worked within an overall budget constraint, but the NHS budget
was tightened in the early 1980s and new fiscal procedures established in the early 1990s
heightened the visibility of trade-offs across policy priorities and between taxes and
spending. In the United States, not only the two largest public health insurance programmes
– Medicare and Medicaid – but also private health plans moved to rein in provider fees.
For reasons to be discussed, however, US gravitation towards the ‘international standard’
in cost containment yielded decidedly mixed results. Still, the overarching movement
across the five nations was towards increasingly stringent controls.
There is no need to postulate some independent ‘reason of state’ for this tendency,
although the strong autonomy of states in the face of provider resistance is worth remarking
on. Rather, cost-containment efforts followed an eminently political logic in their genesis
and character. The underlying imperative was to restrain government health spending so
as to minimize restrictions on other prized areas of public finance and avoid politically
explosive tax increases. Notable here are the German and Dutch experiences, where rapidly
rising payroll taxes became the key focal point of cost-control efforts. Given widespread
cries by employers that payroll taxes stunted job growth, it was arguably as politically risky
for leaders in these nations to avoid tackling medical inflation as it was for them to take
on providers.
The specific character of cost-containment efforts also had roots in political incentives.
It is striking when one reviews the catalogue of diverse cost-control measures employed
within public programmes how few directly imposed new consumer costs or explicitly
reduced the scope of benefits. This is not to say that consumer spending on private
insurance and services has not increased, nor is it to deny that there are increasingly
important areas of care that fall outside the scope of public protection. As the next section
will emphasize, these forms of policy change are crucial. Yet these changes are for the most
702 HACKER

TABLE 2 Significant Legislative Changes in Health Policy in Five Nations,

1980–2000

Britain
• Series of modest reorganizations accompany a tightened National Health Service budget
(1980–89).
• NHS and Community Care Act 1990 emerges from 1989 White Paper; it envisions the
creation of new purchasing agents to contract with hospitals and the authorization of general
practitioner (GP) ‘fundholding’ under which GPs would be given budgets to purchase
secondary services.
• 1997 White Paper of Labour government proposes the abolition of ‘internal market’ but
the proposed changes retain purchaser–provider split and expand GP fundholding, key
elements of 1990 reform.
• Labour leadership commits itself to reaching spending parity with Continental Europe.

Canada

• Canada Health Act (1984) tightens federal standards for provincial programmes, limiting
role of private finance. Reaffirmed through federal actions in mid-1990s.
• Dramatic long-term decline of federal transfers to support provincial programmes under
Progressive Conservatives (1983–93).
• 1997 Report of National Forum on Health Care, a creation of federal Liberal government,
endorses existing structure and calls for some expansions, to date largely unimplemented.
• Ongoing provincial efforts at hospital restructuring yield change mainly at the margins. A
few provinces consider but do not enact more ambitious reforms.
Germany

• Series of cost-containment acts expand co-payments and tighten budgeting and rate-setting.
• Health Care Reform Act 1988 extends preventive check-ups and home care and increases
co-payments, particularly for pharmaceuticals. Only increased co-payments survive
implementation.
• Health Care Structural Reform Act 1993 makes substantial change to hospital budgeting
(further reformed in 1995 and 1996), creates scheme for equalizing contribution levels
across sickness funds, and allows patients greater choice of funds while further expanding
co-payments for pharmaceuticals.
• Statutory Nursing Care Insurance 1995 expands long-term care coverage.
• Health Care Reorganization Act 1997 lifts some budgetary restrictions imposed in 1992,
maintains and expands patient choice of funds, places new cost-containment responsibilities
on the sickness funds, and imposes greater cost-sharing, though not all cost-sharing
provisions are fully implemented.
• New Social Democratic leadership repeals some of the market-based elements of previous
reforms and promises but postpones further structural reforms.
The Netherlands

• Series of cost-containment acts strengthen central budgeting, introduce prospective budgets

for hospitals and other medical institutions, and create price control system for prescription
drugs.
• 1987 Dekker Report, supported by centre-right government, outlines a major reorganization
to encourage competition among insurers and providers and to change the premium
structure, separating basic from supplementary coverage and eroding firewall between
sickness funds and private insurance.
• Labour government elected in 1989 revises plan, stressing need for broader uniform social
insurance. Series of subsequent revisions foster conflict, leaving many reforms unimple-
mented.
• During the 1990s, policy makers, following a more cautious course, manage to increase
scope for and freedom of private insurance while creating scheme to compensate insurers
covering high risks.
 Review Article: Dismantling the Health Care State? 703

United States

• Medicaid spending is significantly cut in first Reagan budget (1981).

• Passage of two cost-control measures for Medicare: prospective payment in the hospital
sector (1982) and fee schedule for the physician sector (1989).
• Series of coverage expansions under Medicaid through federal mandates on states
(1984–89).
• Medicare Catastrophic Coverage Act expanding Medicare is passed, then repealed
(1988–89).
• Kassebaum–Kennedy (1996) regulations encouraging health insurance portability pass in
wake of failure of Clinton health plan (1994) and aggressive Republican drive to cut social
spending (1995).
• 1997 budget aims to foster Medicare contracts with private health plans while also cutting
payments.
• 1997 budget funds new state plans for low-income children.
• New Republican administration enacts prescription drug benefit for elderly beneficiaries of
Medicare in 2003; it includes measures designed to increase role of private health plans
within the programme.

part not the result of explicit attempts at cost-shifting. To date, the use of cost-sharing has
not increased dramatically in Britain, Canada, the Netherlands or Germany, or within the
US Medicare programme, and the increases that have occurred have been unpopular,
riddled with exceptions and vulnerable to reversal. ‘Delisting’ of benefits once covered by
public programmes is quite rare, and the explicit exclusion of previously covered
populations is essentially non-existent.16 Between 1980 and 2000, only four of the
countries in Table 1 – France, the Netherlands, Switzerland, and the United States – saw
a change in public coverage, and in all four cases, the share of the population covered
increased – on average by 2 percentage points.
Instead of cutting benefits or coverage, cost containment has focused overwhelmingly
on controlling fees and overall spending. The motive is no secret: imposing limits on
providers and then leaving them to cope has proved far more politically attractive than
imposing costs or restrictions on patients. The medical maxim ‘do no harm’ has its
counterpart in political life: ‘Never be seen to impose highly visible costs on large numbers
of voters.’ The politics of cost control is not the happy politics of claiming credit for
things done for constituents; it is the thorny politics of avoiding blame for things done to
them.17
Despite the moderation of medical inflation in the mid-1990s, the United States
continues to stand out as an outlier. Why has the American way of financing medical care
proved so distinctly incapable of reining in medical costs? The strongest hypothesis is that
the fragmentation and opacity of major financing sources and the limited scope of
collective insurance pools have simultaneously muted concern about costs and prevented
public and private authorities from exercising decisive control. For all the differences in
their financing systems, Britain, Canada, Germany and the Netherlands all route a large

16
I say ‘essentially’, because state Medicaid programmes in the United States have reduced coverage in the
past, and are doing so currently, and because the Netherlands in the 1980s enacted legislation that removed a few
relatively small population groups from the mandatory insurance scheme, although these groups were protected
by elaborate new regulations ensuring their access to affordable private health insurance.
17
R. Kent Weaver, ‘The Politics of Blame-Avoidance’, Journal of Public Policy, 6 (1986), 371–98.
704 HACKER

share of health spending through highly visible and encompassing financing pipelines –
principally, the general tax system (Britain and Canada) and the payroll tax system
(Germany and the Netherlands). Of necessity, then, medical costs are highly transparent,
and few obvious avenues exist for public or private payers to shift or obscure them.
In the United States, by contrast, financing not only comes from a huge multiplicity of
sources, but much of it is hidden in the form of tax breaks for private health benefits (which
cost $188 billion in 2004) and reductions in take-home pay.18 And even when payers
undertake concerted attempts to clamp down, as did Medicare in the 1980s and employers
in the 1990s, their universe of control is strictly circumscribed, and their effort likely to
be dissipated by cost-shifting. Moreover, as the next section will discuss, private controls
on spending have resulted in significant shifts of risk and costs from collective
intermediaries on to workers. Such risk-shifting and cost-shifting is much more difficult
when all or nearly all citizens have coverage guarantees that must be altered through
democratic processes.
The barriers to cost containment posed by the fragmentation of American financing
would seem to imply that fiscal concentration is a precondition for cost control, and indeed
‘monopsonistic’ financing in which governments pay for most care has proved quite
effective at controlling spending. The German and Dutch experiences suggest, however,
that more decentralized systems can nonetheless achieve similar levels of budgetary
control through the creation of hard sectoral budgets within which insurers and providers
negotiate. (The success, albeit temporary, of US managed care to control costs in the
mid-1990s also rested largely on sharp reductions in fees paid to providers, although these
efforts proved incapable of controlling the rate of increase of costs over the long run.)
Furthermore, because monopsonistic financing makes it easier to hold politicians
responsible for cost control’s negative effects, governments face a trade-off between
control and accountability, which may help explain why centralized systems do not
demonstrate an even greater advantage.19 Despite these caveats, levels of spending are
typically lower in tax-financed national health services than in systems financed by payroll
taxes. The 1990s Canadian clamp-down suggests that declining national contributions in
fiscally decentralized systems can also be a powerful tool for spending control. (Although
not one of the four cases, a similar process lies behind Sweden’s impressive cost-restraint.)
Here the mechanism would seem to be the capacity of central governments to minimize
blame for subnational cost-containment precipitated by declining contributions from the
centre.20
If blame-avoidance imperatives explain key features of cost-containment policies, a
natural question is whether cost control creates political fallout. The answer is almost
certainly yes, but the evidence is not wholly conclusive. Aggregate levels of public
satisfaction with national health systems do appear to vary in accordance with public per
capita health spending (and, importantly, also with share of total health spending paid for
by the public sector).21 But because public satisfaction is related to public expectations,

18
The $188-plus billion figure comes from John Sheils and Paul Hogan, ‘Cost of Tax-Exempt Health Benefits
in 2004’, Health Affairs, Web Exclusive (25 February 2004).
19
Another notorious problem in estimating the effect of budgetary controls is the classic bugaboo of reverse
causation: countries that adopt controls may do so precisely because they have more difficulty controlling
spending.
20
Or, in the case of Sweden, by central-government limits on local taxation.
21
Carolyn Tuohy, Collen M. Flood and Mark Stabile, ‘How Does Private Finance Affect Public Health Care
Systems? Marshalling the Evidence from OECD Nations’ (unpublished paper, University of Toronto, n.d.).
 Review Article: Dismantling the Health Care State? 705

which are likely to vary across nations, the more relevant indicator is changes in public
satisfaction over time, which, due to the limited number of comparable over-time surveys,
remains difficult to gauge.22 On this more demanding measure, the striking feature of the
scattered available evidence presented in Table 3 is the very sharp decline in satisfaction
that occurred in Canada in the decade after 1988. A smaller but still notable decrease
occurred in Germany between 1988 and 1994, while public views in the United Kingdom
remained relatively stable during the 1990s – with discontent growing sharply, however,
at the end of the decade. The United States presents a mixed picture that hints at increased
opinion polarization: growing support for only minor changes alongside growing support
for a complete overhaul.
Looking at events in individual countries, the conclusion that cost containment has
prompted public backlash is considerably strengthened. In Britain in the late 1980s, for
example, public dissatisfaction fuelled by media stories, provider appeals and mounting
waiting lists forced the NHS to the top of Thatcher’s political agenda. In the United States,
increasing micromanagement of clinical decisions by private health plans prompted a wave
of revulsion against ‘managed care’. In Canada, the decline of federal funds set off an orgy
of recrimination between doctors and provincial governments that fed Canadians’ growing
sense of unease. Public perceptions, however, are mediated by provider strategies and the
dynamics of political competition. Cost control seems most likely to become a topic of
public concern when competition between contending parties centres on the ruling
government’s health care stewardship, when payment disputes are frequent and visible,
and when providers feel sufficiently aggrieved to adopt an ‘outside’ strategy of stoking
public fears, rather than working through established mechanisms of state–professional
bargaining.23
The most straightforward conclusion, however, is that the politics of blame avoidance
cuts both ways. Leaders who fail to act as health care consumes ever larger shares of public
and private budgets risk the wrath of employers and the clienteles of non-health
programmes. But cost containment creates its own potent political risks, raising the spectre
of waiting lists and outmoded facilities, and creating highly public conflict with providers.
In this environment, politicians are understandably attracted to the idea of broader
structural reform of health policies. The language of cost control connotes restraint and
limitations. The watchwords of structural reform, by contrast, are ‘efficiency’ and
‘responsiveness’ – words that suggest the ability to rein in costs without reductions in the
quality or quantity of medical services. Structural reform includes market-based reforms,
but it also encompasses other major systemic shifts designed to control costs while
improving quality and responsiveness. It is what Peter Hall, in the context of economic
policy, terms second-order and third-order policy change: change in the instruments of
policy or the hierarchy of goals that guide their use, rather than simply change in the
stringency with which pre-existing instruments are used.24
In practice, as we shall see, the promise of structural reform has proved exceedingly
difficult to realize. And unlike cost control, the emergence, character and effects of

22
On the general problem of comparing survey results across nations (and other cultural groupings), see Gary
King et al., ‘Enhancing the Validity and Cross-Cultural Comparability of Measurement in Survey Research’,
American Political Science Review, 98 (2004), 191–207.
23
For this reason, the decline of professional influence over policy, which has encouraged popular appeals,
may well be an indirect cause of the increased dissatisfaction voiced in many nations.
24
Peter A. Hall, ‘Policy Paradigms, Social Learning, and the State: The Case of Economic Policymaking in
Britain’, Comparative Politics, 25 (1993), 275–96.
 706

TABLE 3 Public Satisfaction and Health Spending in Four Nations, 1990–2001

Percentage Percentage Percentage

HACKER

saying saying favouring a Health

minor fundamental complete spending as
changes Increase/ changes Increase/ rebuilding of Increase/ share of Increase/
needed (%) Decrease needed (%) Decrease the system (%) Decrease GDP (%) Decrease

Canada
1990 56 38 5 9.0
1994 29 ⫺ 27 59 ⫹ 21 12 ⫹7 9.5 ⫹ 0.5
1998 20 ⫺9 56 ⫺3 23 ⫹ 11 9.1 ⫺ 0.4
2001 21 ⫹1 59 ⫹3 18 ⫺5 9.7 ⫹ 0.6
Germany
1990 41 35 13 8.5
1994 30 ⫺ 11 55 20 11 ⫺2 9.9 1.5
United Kingdom
1990 27 52 17 6.0
1998 35 ⫹8 58 ⫹6 14 ⫺3 6.9 ⫺ 0.2
2001 21 ⫺ 14 60 ⫹2 18 ⫹4 7.6 ⫹ 0.7
United States
1990 10 60 29 11.9
1994 18 ⫹8 53 ⫺7 28 ⫺1 13.2 ⫹ 1.3
1998 17 ⫺1 46 ⫺7 33 5 13.0 ⫺ 0.3
2001 18 ⫹1 51 ⫹5 28 ⫺5 13.9 ⫹ 0.9

Sources: Robert J. Blendon et al., ‘Satisfaction with Health Systems in Ten Nations’, Health Affairs, 9 (1990), 185–92; ‘Who has the Best
Health Care System? A Second Look’, 14 (1995), 220–30; and ‘Inequities in Health Care: A Five-Country Survey’, Health Affairs 21 (2002),
182–91. Spending from Organization for Economic Cooperation and Development, OECD Health Data 2003 (Paris: OECD, 2003).
 Review Article: Dismantling the Health Care State? 707

structural reform have varied greatly across nations To explain this variation requires
exploring the interaction between political and medical systems in a climate of austerity,
and the subtle but fundamental shifts that have occurred without legislative change, as the
effects of cost control have played out within distinctive medical complexes.

THE POLITICS OF STRUCTURAL REFORM

In one form or another, structural reform has emerged on the decision-making agenda of
the five nations under consideration in the past two decades, with Canada only recently
moving towards discussion of the issue. The most ambitious of these proposals was
President Bill Clinton’s plan for universal health insurance via managed competition, but
this initiative failed even to come up in Congress for a vote. The next most sweeping
proposal was Thatcher’s internal market reforms of 1989, which were enacted and in large
part realized. On roughly the same plane of ambition were the so-called Dekker reforms
(named after the head of the committee that formulated them) pursued in the Netherlands
since the late 1980s, which have proved exceedingly difficult to implement. German
leaders have charted a more cautious course, with significantly less comprehensive and
radical structural reforms. The overall Canadian policy framework has remained extremely
stable.
From the foregoing, it should already be clear that interest in reform is not confined to
any single type of medical or political system. The conditions under which structural
reform rises to prominence, however, do seem to differ between more and less centralized
political structures. In the former, the commitment of the ruling government to action
seems both necessary and sufficient to raise reform to the top of the agenda. The question
thus becomes why parties in power seek the goal. In more fragmented settings, the
conditions for the emergence of reform are less clear-cut. Economic downturns appear to
promote interest. Serious momentum towards reform also seems to require cross-party
agreement on the need for change, perhaps because oversized coalitions are required to
achieve results. Thus while structural reforms seem favoured in centralized polities when
ruling parties enjoy a strong position, close party divisions may actually facilitate the rise
of the issue elsewhere.
To begin the analysis, therefore, we need a somewhat more systematic understanding
of the major features of political and health financing institutions. At the risk of papering
over subtle variations, the framework I outline here ruthlessly reduces the complexities of
the terrain to two simple dimensions. In the case of political institutions, I emphasize the
presence or absence of formal veto points, which shapes the incentives and opportunities
of policy makers. In the case of medical financing, I emphasize the centrality of the state
in financing and regulating medical care, which shapes the kinds of challenges policy
makers confront and their ability to implement legislative aims. Figure 2 maps out the
fourfold division that results. Political systems may be either veto-free or veto-ridden;
medical systems may be either hierarchical or decentralized. I have placed the four cases
at different points within the cells to indicate their relative location on these axes.25
Inevitably, some countries fit into this framework less well than others.26 Treating the
25
Canada is the only OECD example of a veto-ridden/hierarchical system, indicating a strong elective affinity
between political centralization and hierarchical organization of the medical sector.
26
Most problematic are Belgium (which has seen the development of federal institutions in the past thirty years
but not in the area of social policy and is thus classified as centralized) and Australia (which has, since the creation
708 HACKER

Fig. 2. Political and medical systems and the prospects for structural reform

Dutch political framework as veto-free, for instance, clearly slights the role of coalition
governments and institutionalized corporatist arrangements in tempering what is, on paper,
a highly centralized political structure. In the present framework, however, unitary states
in which legislative and executive powers are fused are classified as veto-free, and the
Netherlands fits this mould. That said, it is important to note that corporatism does enter
into my assessment and explanation of Dutch policy dynamics via my characterization of
medical financing. To have a decentralized medical system in my framework a nation must
(F’note continued)
of Medibank in 1984, had a hierarchical system of public medical finance but which, unlike Canada, allows private
insurance to duplicate public coverage, has witnessed successive government efforts to encourage private health
insurance, and is thus characterized as veto-ridden/decentralized). With regard to political structure, France might
be seen as ambiguous, because it has an independently elected president. The president, however, shares power
with the prime minister, who can be defeated by a no-confidence vote – the hallmarks of a parliamentary system.
 Review Article: Dismantling the Health Care State? 709

rely significantly on multiple payers, such as sickness funds, commercial insurers and
private employers. These interpenetrating policy institutions are, of course, a key hallmark
of corporatism in the Netherlands and other institutionally similar nations, accounting for
much of the distinctive character of health policy making within them.
It is also important to stress that in focusing on veto points and the centralization of
financing, I am certainly not arguing that these are the only relevant influences on
national paths. (The analysis to come suggests, for example, that electoral rules – and,
in particular, the outsized majorities fostered by plurality elections in Britain and its
former colonies, as opposed to the coalition governments fostered by proportional
representation (PR) – also play an important role.) As we shall see, the countries within
these categories do seem to hang together, but my overarching ambition in tracing these
common patterns is limited to two broad goals: first, to show that the pattern of ‘reform
without change and change without reform’ is robust and general; and secondly, to
suggest that it is rooted in the interplay of financing structures and the rules of the
political game.
Indeed, the principal benefit of the simple two-by-two framework just presented is
that it allows for the formulation of straightforward and verifiable claims about how
different configurations of political and medical authority should be expected to shape
the politics of health reform. The basic expectations are mapped out in Figure 2.
Veto-free/hierarchical regimes are the settings in which structural reforms, once
proposed, are most likely to be enacted and implemented. In veto-free/decentralized
regimes, by contrast, the political system allows legislative change, but the financing
system is likely to become a serious barrier at the stage of implementation. In
veto-ridden/decentralized regimes, it is highly difficult to consolidate authority for major
policy change; yet the fragmentation of the health care system may actually facilitate
smaller-scale reform by leaving room for shifting coalitions of interest. Finally,
veto-ridden/hierarchical regimes – of which, admittedly, Canada is the only OECD
example – may actually be more prone to policy stalemate than veto-ridden/decentral-
ized regimes. This is not only because the political system makes change inherently
difficult, but also because the hierarchical structure of medical authority leaves limited
room for coalitional reshuffling among key stakeholders – and, in particular, limits the
ability of business firms to become key agents of change.
The next two sections examine the dynamics and effects of structural reform within this
framework of expectations, focusing on the five nations under consideration but
occasionally drawing on the experience of other countries to amplify or test key claims.
As will become clear, market-oriented reforms are, perhaps surprisingly, almost entirely
clustered in the veto-free/hierarchical quadrant. Nearly all the nations in this category have
also seen a decline in the government share of health spending – at times quite dramatic.
Yet the specific effects of market-oriented reforms have been relatively modest and, on
the whole, tended to enhance state authority rather than market forces. In the
veto-ridden/decentralized regimes, by contrast, large-scale structural reforms have fared
poorly. Yet these systems nonetheless saw an overall shift towards an expanded public
share of spending in the 1980s and 1990s. Finally, and perhaps most importantly, in all
four nations, the largest changes in ground-level outcomes were the result not of structural
reforms, but of the long-term effects of cost-containment policies as they played out within
distinctive health financing arrangements. These effects, however, varied greatly across
different financing systems, imposing modest and widely shared costs in some and large
and highly concentrated costs in others.
710 HACKER

Fig. 3. Political and medical systems and the public share of health spending

THE VETO-FREE CASES: BRITAIN AND THE NETHERLANDS

When health policy experts speak of a sweeping trend towards market reform, they
typically cite Britain, the Netherlands and Sweden, with Denmark, Finland and New
Zealand often mentioned, too.27 Taking just these six agreed-upon exemplars, all but one
is located in the veto-free/hierarchical quadrant, and all are characterized by relatively
centralized polities. Indeed, given the extreme difficulties that Dutch policy makers have
faced in transforming market reforms into actual policy change, it is arguable that the only
unambiguous cases of realized market reform fall into the veto-free/hierarchical category.
As Figure 3 shows, moreover, all but one of the nations in the veto-free/hierarchical
quadrant display a notable decline in the government share of medical spending: On
average, the public share dropped 5.4 percentage points between 1980 and 1998. By
27
See, for example, Chris Ham and Mats Brommels, ‘Health Care Reform in the Netherlands, Sweden, and
the United Kingdom’, Health Affairs, 13 (1994), 106–19. Germany is occasionally included, though the extremely
limited move towards competition does not even begin to compare to the bolder aspirations elsewhere. See also
Lawrence D. Brown and Volker E. Amelung, ‘Manacled Competition: Market Reforms in German Health Care’,
Health Affairs, 18 (1999), 761–91.
 Review Article: Dismantling the Health Care State? 711

contrast, the public share remained essentially stable, on average, in the veto-free/decen-
tralized quadrant, and rose an average of 1.6 points in the veto-ridden/decentralized
quadrant. Within the OECD, therefore, a significant convergence of the public share has
occurred since 1980, much of it driven by the universal decline in the public fiscal role
in the veto-free/hierarchical regimes.
At first glance, the concentration of market reforms in the veto-free/hierarchical regimes
is at odds with expectations. These are nations, after all, that have demonstrated the greatest
commitment to an expansive government role, and they have generally been the most
successful in restraining health spending. The paradox evaporates, however, once we
consider the capacities and aims that political leaders in these countries had, as well as the
actual character and effect of the reforms that they pursued. First, and most obvious, the
idea of introducing highly limited forms of market freedom simply made sense in nations
where medical care had previously been governed through extremely hierarchical means.
Whereas other nations came under acute fiscal strain at a point when their capacities for
controlling spending were generally underdeveloped, the veto-free/hierarchical regimes
already had in place financing systems and payment forms that embodied substantial
capacities for restraint. To the extent that these nations faced a common challenge, it was
less how to develop the means for control than how to exercise them without backlash or
serious sectoral distortion.
Secondly, and more importantly, leaders in these nations worked within political
institutions that gave them the opportunity to exercise decisive authority. To be sure, the
more consensual policy style in the PR systems differs markedly from the see-saw pattern
of majority-party dominance in Britain and New Zealand. In both settings, however, ruling
governments did not have to cope with the veto points introduced by federalism, a powerful
judiciary, the separation of powers or popular referendums – all features evident
individually or in combination in the more veto-ridden systems. This proved crucial in
Britain and New Zealand, where all evidence suggests that the reforms were initially quite
unpopular. Leaders in veto-free systems did not always have the inclination or room to
pursue structural reforms, but they had the means to do so more often than elsewhere.28
In veto-ridden settings, by contrast, leaders instead generally followed a path of
pre-emptive accommodation – ‘tiptoeing in’ rather than ‘crashing through’. On this score,
it is revealing that the most prominent attempt to use the strategy of one-shot reform within
veto-ridden polities – the Clinton health plan – was distinguished mainly by the glaring
gap between aspirations and accomplishments.
Thirdly, and perhaps least recognized, there was an important policy logic to reformers’
strategies in the veto-free/hierarchical regimes. Health economists typically contrast
regulation and competition.29 Though few are naı̈ve enough to think that competition
requires no regulation, they usually see regulation as limited to policing and enabling
market processes. The comparative record shows that this perspective grossly understates
the extent of central power needed.30 The ability of countries like Britain to pursue market
reforms was not at odds with the hierarchical structure of their systems but deeply

28
On these means, see Geoffrey Garrett, ‘The Politics of Structural Change: Swedish Social Democracy and
Thatcherism in Comparative Perspective’, Comparative Political Studies, 25 (1993), 521–47.
29
For example, Sherry Glied, Michael Sparer and Lawrence Brown, ‘Containing State Health Expenditures:
The Competition vs. Regulation Debate’, American Journal of Public Health, 85 (1995), 1347–9.
30
See James A. Morone, ‘The Ironic Flaw in Health Care Competition: The Politics of Markets’, in Richard
J. Arnould, Robert F. Rich and William D. White, eds, Competitive Approaches to Health Care Reform
(Washington, D.C.: The Urban Institute, 1993), pp. 207–22.
712 HACKER

dependent upon it, and the surest effect of market reforms was to strengthen the state’s
authority. Moreover, these reforms carried with them the seeds of further intervention.
Introduced into hierarchical systems, market mechanisms entailed strong claims for system
improvement, the deployment of visible levers of power and new performance measures.
With these in place, it proved nearly impossible for governments to let go of the instruments
they had deployed.
As the nation where market reforms went furthest, Britain presents the hardest test for
the argument. Early expectations were certainly high. Chris Ham and Mats Brommels
pronounced in 1994: ‘Since Prime Minister Margaret Thatcher published her govern-
ment’s proposal for reforming the National Health Service (NHS), the United Kingdom
has witnessed a near-revolution in health services delivery.’31 On close inspection,
however, Britain is a veritable showcase of the limited and contradictory effects of market
reform. As discussed already, Thatcher turned to reform not as part of an ideological
crusade, but as an attempt to head off a mounting crisis over NHS funding. Once reform
had been promised, however, she was ‘loathe’ merely to spend more and instead presented
a blueprint for structural change, the centerpiece of which was a new internal market in
which regionally based public managers would be given the authority to contract
selectively for services.32 In a third term in power at the time of the campaign,
Conservatives did not fully implement the changes, nor were more radical reforms
considered feasible. Nonetheless, the reforms marked a dramatic break with previous NHS
adjustments.
What they did not mark, however, was a dramatic shift towards markets. If anything,
the reverse was true. Susan Giaimo concludes that ‘the managerial reforms worked in the
opposite direction of devolution to strengthen both the legal authority and institutional
capacity of policymakers for centralized intervention in the NHS.’33 Rudolf Klein
observes: ‘Far from leading to the devolution of decision making – the ultimate logic of
a market system – the Conservative reforms led to increasing centralization.’34 Although
it is tempting to see the limits of competition as a result of implementation gone awry, in
fact the cornerstone of the internal market – new regional purchasers – was inherently
centralizing in its emphasis on greater managerial power and accountability to the central
government. Conservatives saw this as the most effective means of distributing initial
resources and crippling the only other real centre of influence, namely, doctors. But it had
the effect of augmenting central authority. In turn, by selling reform on the basis of
improved outcomes, politicians set themselves up to be judged on the results. Each new
indicator was an invitation to intervention – the ‘bedpan … dropped on a hospital floor’
that the NHS’s creator, Aneurin Bevan, once predicted would henceforth ‘resound in the
Palace of Westminster.’35
As might be expected from the foregoing, moreover, the ground-level changes
introduced by the reform were underwhelming, especially if judged against the rhetoric.

31
Ham and Brommels, ‘Health Care Reform in the Netherlands, Sweden, and the United Kingdom’, p. 108.
32
Alan Jacobs, ‘Seeing Difference: Market Health Reform in Europe’, Journal of Health Politics, Policy, and
Law, 23 (1998), 1–34, p. 19. A second key element was primary care physician ‘fundholding’, under which doctors
would be given new latitude to purchase services.
33
Susan Giaimo, Markets and Medicine: The Politics of Health Care Reform in Britain, Germany, and the
United States (Ann Arbor: University of Michigan Press, 2002).
34
Rudolf Klein, ‘Why Britain is Reorganizing its National Health Service – Yet Again’, Health Affairs, 17
(1998), 111–25, p. 117.
35
Klein, ‘Why Britain is Reorganizing its National Health Service’, p. 117.
 Review Article: Dismantling the Health Care State? 713

Giaimo writes that the fears of critics of the internal market have not been borne out
‘precisely because competition has been so limited’.36 Carolyn Tuohy notes that ‘arguably
little change occurred in the broad balance among state actors, the medical profession, and
private finance.’37 In a comprehensive survey of the relevant micro-level evidence, Julian
Le Grand concludes that ‘[p]erhaps the most striking conclusion to arise from the evidence
is how little overall measurable change there seems to have been.’38
The story of the Dekker reforms in the Netherlands, though in many respects quite
different, bears important similarities to the British saga. There the reform ideal was also
dressed in market garb. Yet in keeping with the Netherlands’ highly decentralized
financing structure, it envisioned not markets within public programmes, but greater
competition between sickness funds and commercial insurers within a framework of
expanded basic risk protection. And more so than in Britain, the Dutch proposals had to
adopt pre-emptive concessions, both to accommodate the coalition partners in power at
the time and to head off the plethora of potential opponents, from insurers and sickness
funds to employers and providers. All the same, the reforms eventually passed were
extremely ambitious, and many experts writing in the early 1990s saw them as the surest
sign of a global trend towards market innovation.
That judgement, however, proved wildly premature, as the reform train ground to a halt
in the implementation stage, blocked by the decentralized medical system that it aimed to
remodel. As Alan Jacobs observed in 1998, ‘Dutch governments have been unable to bring
the Dekker vision to life, leaving the old system largely intact.’39 Two Dutch policy insiders
concluded in 1997: ‘[T]he restructuring process generated growing opposition and, despite
a seemingly realistic time-table of five years for implementation, it was subsequently
abandoned. In many ways, decision-making … lies in limbo.’40 Instead of decentralizing
power to consumers, another insider noted in the mid-1990s, the Dutch reform process
‘provoked a largely unplanned process of creeping estatization’.41 Although one clear
effect of the reforms has been to increase the autonomy of insurers in negotiating fees with
providers, other key elements of the enacted reforms – notably, efforts to level the playing
field between funds and private insurers and the shift of some fund-provided benefits to
a state-run catastrophic insurance programme – were in the direction of enhanced state
power.
What, then, are we to make of the notable decline in the public share of spending seen
in the Netherlands and Britain and, indeed, in all but one of the veto-free/hierarchical
regimes? The fall in the public share of spending suggests that the role of the welfare state
is changing, but this change is almost certainly not caused by market reforms. The British
experience is instructive. The level of private spending and reliance on private insurance
rose most precipitously in the 1980s, before the Thatcher reforms, and they have been
36
Giaimo, Markets and Medicine, p. 221.
37
Carolyn Tuohy, Accidental Logics: The Dynamics of the Health Care Arena in the United States, Britain,
and Canada (New York: Oxford University Press, 1999), p. 198.
38
Julian Le Grand, ‘Competition, Cooperation, or Control? Tales from the British National Health Service’,
Health Affairs, 18 (1999), 27–39, p. 31.
39
Jacobs, ‘Seeing Difference’, p. 15.
40
James Warner Björkman and Kieke G. H. Okma, ‘Restructuring Health Care Systems in the Netherlands:
The Institutional Heritage of Dutch Health Policy Reforms’, in Christa Altenstetter and James Warner Björkman,
eds, Health Policy Reform, National Variations and Globalization (London: Macmillan, 1997), pp. 79–108, at
p. 106.
41
Frederick T. Schut, ‘Health Care Reform in the Netherlands: Balancing Corporatism, Etatism, and Market
Mechanisms’, Journal of Health Politics, Policy, and Law, 20 (1995), 615–52, p. 649.
714 HACKER

relatively stable since. Despite dire stories about the flight of the rich, popular support for
the NHS remains overwhelming, and the reach of private coverage still minimal. All signs
are that the rise of the private spending share in Britain, as in other veto-free/hierarchical
cases, has principally been a by-product of public-sector austerity and only secondarily a
result of formal policy changes. The British experience thus strongly suggests that what
I have termed elsewhere ‘conversion’ and ‘drift’ – that is, changes in levels of protection
within established policy parameters, either through decentralized adaptation (conversion)
or failure to adjust policies to changing circumstances (drift) – has been more consequential
than formal reforms.42
This conclusion is further reinforced by the recent Dutch experience. The Dutch system,
unlike the hierarchical British framework, is based on a public–private partnership in which
lower- and middle-income citizens are required to have publicly overseen coverage while
higher-income citizens must insure themselves through the private sector. Over the past
two decades, the share of financing that comes from private sources has increased. As in
Britain, however, this is almost certainly not because of major explicit alterations in the
scope of public coverage. Indeed, the share of the population covered by the mandatory
public framework (which covers hospital and physician services, pharmaceuticals and
home care) actually increased over the 1980–2000 period. At the same time, attempts at
modestly limiting core public services stalled while a new catastrophic insurance
programme was extended to all citizens, including the well off. As three noted health policy
experts recently concluded, ‘The Netherlands group-based model has persisted in
maintaining a remarkable consistency in the terms and conditions upon which care is
provided across the population.’43 The absence of good historical data on out-of-pocket
spending and private insurance expenditures makes it difficult to pin down the exact cause
of the rise of the private share in the Netherlands, but it appears to reflect three main trends,
all relatively modest in effect: a general increase in cost-sharing, higher growth in private
insurance outlays than in public outlays, and shifts towards pharmaceuticals and other
interventions that have historically entailed user charges.
So far, too, the feared effects of private spending on equity and solidarity have largely
failed to materialize. This is not because private finance cannot have such effects; the US
case strongly indicates the contrary. Nor is it meant to gainsay the real hardships imposed.
But despite the drop in the public share of spending in the veto-free/hierarchical regimes,
as well as in the Netherlands, private finance remains marginal in most clinical areas.44
The exception within the veto-free/hierarchical quadrant is New Zealand, where private
insurance has historically played a much larger role than in Britain. New Zealand, however,

42
On policy drift, see Jacob S. Hacker, ‘Privatizing Risk without Privatizing the Welfare State: The Hidden
Politics of Social Policy Retrenchment in the United States’, American Political Science Review, 98 (2004), 243–60.
See also the discussion of ‘utility drift’ in Douglas Rae, ‘The Limits of Consensual Decision’, American Political
Science Review, 69 (1975), 1270–94. Although rarely acknowledged in current welfare state scholarship, policy
drift was clearly recognized by Hugh Heclo in his classic Modern Social Politics in Britain and Sweden (New Haven,
Conn.: Yale University Press, 1974). Heclo writes of the Swedish Pension Act of 1913 (p. 211): ‘In large part, it
was precisely because this basic framework remained unaltered in the midst of changing circumstances that the
framers’ intentions were unconsciously subverted. As noted throughout this volume, one of the easiest ways to
change a policy is to fail to change a program to accord with the movement of events’. (I am grateful to Kent Weaver
for this citation.) The term ‘conversion’ comes from Kathleen Thelen, ‘How Institutions Evolve: Insights from
Comparative-Historical Analysis’, in James Mahoney and Dietrich Rueschemeyer, eds, Comparative Historical
Analysis in the Social Sciences (Cambridge: Cambridge University Press, 2003), pp. 208–40.
43
Tuohy, Flood and Stabile, ‘How Does Private Finance Affect Public Health Care Systems?’ p. 10.
44
Private finance has, however, historically dominated dentistry, optometry and long-term care.
 Review Article: Dismantling the Health Care State? 715

is atypical. In most nations where the public share has dropped, the strains of austerity have
hit broadly. Cost-sharing has increased, rich patients are jumping queues, but the
fundamental role of public finance and regulation remains.

T H E V E T O - R I D D E N C A S E S : C A N A D A, G E R M A N Y A N D T H E U N I T E D S T A T E S

In contrast with the common reduction in the public share of spending in veto-free/hier-
archical regimes, more fragmented polities have demonstrated much greater diversity of
response. And despite the possible affinity between the more heavily privatized structure
of these nations’ systems and market-based reforms, pro-competitive reform legislation
has faced a rough road. In the United States, the Clinton plan met an inglorious fate, and
the main legislative achievements of the period were increased regulation of health plans
and an expansion of public coverage for the poor. In Germany, reforms promoted greater
choice of sickness fund, but this package was a ‘rather tame strategic contrivance’ justified
by a sprinkling of market rhetoric and ‘so far … unspectacular’ in its effects.45
Judged against the veto-free/hierarchical regimes, one pattern that unites the
fragmented-decentralized regimes is the fairly consistent, and often striking, expansion of
the public share of spending. In contrast with the near-universal decline among
veto-free/hierarchical regimes, the only country in this cell where the public share dropped
is Germany, while Australia saw an increase of almost 6 percentage points and Switzerland
and the United States saw a roughly 3-point increase (in the Swiss case, in just the last
decade). Indeed, if we expand our field of inquiry to all decentralized systems, what stands
out is the degree to which they appear to have been ‘catching up’ with more hierarchical
regimes.
An obvious reason for this is the weaker control over expenditure that public
programmes in these nations generally exercised. Nonetheless, one cannot avoid the
conclusion that a major reason for the shift is the expansion of public programmes. In
Australia and Switzerland, universal statutory coverage was achieved only after 1980. In
Japan, public long-term care insurance was expanded in the 1990s, and the same was true
in Germany. In the United States, public coverage for the poor under Medicaid expanded
significantly in the 1980s and 1990s, though not enough to offset declining private
protection. The US public share of spending rose to nearly 46 per cent in 1990s, and if the
cost of tax breaks for private benefits are included, the share well exceeds half of
spending.46 Given recent challenges to the welfare state, the continued expansion of public
health insurance in decentralized systems is a noteworthy development.
Explicit alterations in public policy are not, however, the only source of change within
mature medical complexes. In two nations in particular, Canada and the United States,
ground-level shifts within the context of relatively stable formal policies have been
pervasive and consequential. In each, legislative stalemate has plagued the debate over
reform, and in each stalemate has opened the door to transformation of policy through the
erosion of public and private insurance, and the decentralized adjustment of private and
subnational policy actors. In the United States, however, these forms of drift and
convergence have been far more rapid and fundamental, reflecting the primary role of
employers. Canada and the United States, in turn, are both distinct from Germany in
45
Brown and Amelung, ‘Manacled Competition’, p. 84.
46
Steffie Woolhandler and David Himmelstein, ‘Paying for National Health Insurance – And Not Getting It’,
Health Affairs, 22 (2002), 88–98.
716 HACKER

that inclusive coalitions have not formed behind structural reforms that remedy these trends
as they have across the Atlantic.
Canada is the appropriate place to begin, because its experience is in key respects
idiosyncratic. The Canadian policy framework relies on the provinces and territories to
insure their citizens via public programmes that meet tight federal guidelines. Private
health insurance is banned from duplicating public coverage, but has a large role in
covering services not routinely covered by public programmes – most notably,
pharmaceuticals. Between 1984 and the late 1990s, a period marked by deep
provincial–federal conflicts and by a dire federal fiscal situation, the federal government
slashed contributions to provincial programmes. The result, understandably, was
acrimony, worsened by fierce disputes at the provincial level between cash-strapped
provinces and providers. Amid the conflict, the legislative changes introduced have been
modest and concentrated at the provincial level.
The immobilism that has marked recent Canadian health politics is an outgrowth of its
singular veto-ridden/hierarchical regime. Canada bears close comparison with the United
States, where the federal Medicare programme for the aged essentially adopted the private
insurance model then prevalent among workers. In Canada, post-war provincial
experimentation and parliamentary government facilitated the creation of a similar
programme for the entire population. Yet in transporting the existing private insurance
model into the public sector, Canada departed sharply from the corporatist framework of
Germany, with its heavy reliance on pre-existing insurance funds. The Canadian path thus
institutionalized a bilateral monopoly between the state and providers, which was
inherently difficult to dislodge, and then overlaid it on a highly decentralized political
structure. The resulting constellation of forces was highly inimical to large-scale reform.
The provinces had exclusive policy jurisdiction but were constrained by the strict terms
of federal transfers. The federal government, meanwhile, could shift blame for cost control
downward by cutting funds. And given the stability of the professional–provincial divide
and lack of an integral employer role, there was no obvious new coalition of interests that
might appeal for change.
The exact effects of the resulting two decades of stalemate are difficult to pin down, but
two are unmistakable. The first is a massive decline in public confidence, as we saw in Table
3. The second is a fairly dramatic drop in the public share of health spending, all the more
notable because it runs sharply against the grain of the near-universal trend towards
expanded public spending in other veto-ridden polities. The stability of the Canadian
framework makes it implausible that this shift is a result of legislative reforms, and indeed
no province has had the latitude or will to enact substantial changes. Instead, the decline
in the public share of spending is a reflection mainly of policy drift. As in the
veto-free/hierarchical regimes, the private sector in Canada is not governed by
system-wide price controls and thus has grown faster than public payments. But another
major cause of rising private spending has been what Tuohy felicitously terms ‘passive
privatization’ caused by rapid growth in spending areas not traditionally covered by
Medicare and the shift of many services on to an outpatient basis.47 Though the exact
distributional consequences of these changes are difficult to assess, their main effects seem
to be to drive up costs for those who utilize uncovered services intensively – namely, the
elderly and chronically ill – and to deter the seeking of uncovered care among the poor.
Thus the highly segmented structure of Canadian coverage – which is shared with the US

47
Tuohy, Accidental Logics, p. 235.
 Review Article: Dismantling the Health Care State? 717

Medicare programme – means that the burdens of drift are more concentrated on the ill
and disadvantaged than in more expansive systems, where the main divide is between the
vast bulk of citizens who seek care within public insurance and the minority who rely on
private coverage or can buy themselves off waiting lists.
The distinctiveness of the Canadian experience is easier to appreciate in comparison
with developments in Germany and the United States, two other nations where federalism
has loomed large. The German case is the less dramatic, but also the more hopeful for those
who believe governments can cope with the strains of the post-1970s order. After more
than a decade of stalemate, two major breakthroughs occurred under the Kohl government.
Neither initiative can be described as market-oriented. Instead, in keeping with the
inclusive coalitional bases upon which they rested, they were uneasy combinations of the
interests of the right (higher cost-sharing), centre (increased reliance on the sickness funds)
and left (greater equality across workers). Practically the only group truly abandoned in
the process was medical professionals, who nonetheless managed to obstruct some reforms
during implementation. In keeping with the pattern elsewhere, the single truly
pro-competitive element of the reforms – greater choice of sickness funds in the context
of increased risk-sharing – was centralizing, because it gave the state the authority to
redistribute finances across funds to equalize risks. Since the reforms, contribution rates
have in fact moved into harmony, and the number of sickness funds has dramatically
dropped by more than half.
If the German reforms were scarcely revolutionary, it still remains to ask how they
happened at all. The short answer is that the shifting and overlapping goals of the payers
– specifically, employers, unions, and the Länder – made reform possible by allowing a
new coalition to form against provider interests and by catalysing co-operation among
disparate factions. Employers and unions were the key players and hence their shared
interests were most important. As in the United States, employers were acutely worried
about the cost of benefits. But because all employers were forced to bear this burden, they
faced strong incentives to organize for the collective good of lower overall spending, rather
than to try to achieve savings through individualized efforts.48 Employers, however, did
desire to shift costs on to patients, and their demands were one reason the reforms included
new co-payments (many of which were later reversed). Yet employers were constrained
in their cost-shifting goals by unions, which jointly managed the system and wanted greater
equality of payroll contribution across sickness funds. Finally, strained by the cost of
reunification, the Länder sought greater federal funds for long-term care, then principally
a state responsibility.49
The importance of reunification in catalysing the 1992 reforms may make the German
experience appear unique. But other aspects of German’s successful reform drive seem
highly useful for grasping the dynamics of reform in veto-ridden/decentralized regimes.
For one, Germany’s cautious but productive path suggests that a decentralized medical
system may well give reformers some advantages in a fragmented polity (or at least temper
their ambitions to the point that they are not tempted to pass legislation that cannot be
implemented, as did overeager Dutch politicians, whose market dreams met the stark
reality of corporatist bargaining once legislation passed). The fragmented overlay of
corporatist health insurance, federal institutions and a bicameral parliament has clearly

48
Giaimo, Markets and Medicine, pp. 172–3.
49
In addition, the Länder had significant influence due to their control over the territorially based second
chamber, which at the time was dominated by the Social Democratic Party.
718 HACKER

thwarted swift change, but also given potential opponents of reform strong guarantees that
their interests will be accommodated in any changes made and created the potential for
new alliances of interests – alliances that were plainly not possible in Canada’s bilateral
interest network.
For another, the limited ground-level changes that have occurred in Germany,
particularly in comparison with the United States, indicate that residualization of health
protection through policy drift and conversion is not inevitable in decentralized systems.
Indeed, Germany’s policy of equalizing contribution levels across the sickness funds
reflected a conscious effort to undo past drift caused by the gradual concentration of
higher-risk workers in specific funds. Although co-payments have shifted some risk and
costs to patients, elaborate protection schemes have mitigated their effects. Moreover, the
extremely broad scope of German coverage has limited the sort of passive privatization
seen in Canada.
A sense of the differential effect of drift and conversion in these nations can be gleaned
from comparative data on out-of-pocket health spending. As Table 4 shows, the first simple
conclusion is that citizens in most advanced industrial democracies were paying a larger
share of national health spending out of pocket at the end of the 1980–2000 period than
at the beginning. Although some of the rise may well have been due to the introduction
of co-payments and other forms of explicit cost-sharing, the evident paucity of such
legislated changes and the large magnitude of the increase witnessed in a number of nations
suggest that most of the shift is an indirect result of patients buying out of public
programmes or purchasing services that are only modestly covered by public programmes,
such as pharmaceuticals.
The second conclusion, however, is that the average increase obscures substantial
variation across different nations and regime types in the degree to which cost control has
precipitated increases in out-of-pocket spending. At the extremes, for example, the share
of health care financed by direct consumer payments increased by more than 7 percentage
points in Italy, while it fell by almost 9 points in the United States, as the share of the
population covered by Medicaid and Medicare expanded. Not surprisingly, given the large
decline in the public share of spending in the veto-free/hierarchical nations, the role of
out-of-pocket spending increased the most in these regimes (3.1 percentage points, on
average). It rose more modestly in the veto-free/decentralized regimes and Canada, and
actually declined in the veto-ridden/decentralized regimes – although the decline is entirely
due to the United States and data are unavailable for Switzerland. Germany, for its part,
saw only a modest increase in out-of-pocket spending (0.3 percentage points), suggesting,
again, that German reforms have produced little of the dislocations seen in Canada or the
United States.
At first glance, the United States might be thought to represent a polar opposite to
Germany – the breakdown lane of reform versus the Autobahn. But, in fact, the institutional
context and reform process, and even some of the achievements, were similar in the two
nations. In both, reform emerged on to the agenda in the early 1990s in response to concerns
about economic decline; in both, employers and state governments, fearful of medical
inflation, were key instigators; and in both, reform was eventually taken up by a moderate
conservative – Kohl in Germany; Bush in the United States – who faced strong pressures
from the left. Had Bush returned to office, there is reason to believe he would have pursued
a cautious course similar in aim to that of Kohl’s (but, of course, vastly different in content)
and reason to believe that modest but meaningful reforms would have passed, just as in
Germany.
TABLE 4 Out-Of-Pocket (OOP) Health Spending in Selected OECD Nations, 1980–2000

OOP spending OOP spending

as a share of as a share of
total health total health Percentage-point
spending, spending, change
Nation/Type 1980 2000 in OOP share

Veto-Free/Hierarchical
Denmark 11.4% 16.4% ⫹ 5.0
Finland 18.4 20.6 ⫹ 2.2
Iceland 11.8 15.2a ⫹ 3.4
Italy 15.7b 22.9 ⫹ 7.2
New Zealand 10.4 15.4 ⫹ 5.0
Norway 14.6c 14.3 ⫺ 0.3
United Kingdom 8.6 11.0d ⫹ 2.4
Mean (weighted by years between observations) ⫹ 3.1
Veto-Free/Decentralized
Austria 14.6e 18.6 ⫹ 4.0
France 11.5c 10.2 ⫺ 1.3
Japan 17.7f 17.1a ⫺ 0.6
Mean (1990–2000, weighted) ⫹ 0.3
Veto-Ridden/Decentralized
Australia 16.1 18.4f ⫹ 2.3
Germany 10.3 10.6 ⫹ 0.3
United States 24.2 15.3 ⫺ 8.9
Mean (weighted) ⫺ 2.2
Veto-Ridden/Hierarchical
Canada 14.7b 16.1a ⫹ 1.4
Mean for all fourteen nations (weighted) ⫹ 1.1

Source: Organization for Economic Cooperation and Development, OECD Health Data 2003 (Paris: OECD, 2003). Data
Review Article: Dismantling the Health Care State?

are not available for Belgium, the Netherlands, Sweden and Switzerland.
a
1999 b1988 c1990 d1996 e1995 f1998.
719
720 HACKER

Indeed, such reforms did pass. The spectacular implosion of the Clinton health plan has
obscured the substantial track-record of smaller scale legislative achievements that marked
the 1980s and 1990s, including a stealth expansion of the state-federal Medicaid
programme for the poor and the creation of a new Children’s Health Insurance Program
(CHIP) in 1997. (More recently, the expansions include the extension of Medicare to
include limited prescription drug coverage for the aged in 2003.) Medicaid, in fact, stands
out as a remarkable oasis of spending and coverage growth amid a parched desert of
hamstrung antipoverty programmes. These expansions gained support from a broad
alliance of conservatives and liberals, and in the case of CHIP, from state leaders eager
for greater flexibility under Medicaid. In comparative perspective, the Medicaid
expansions are also of a piece with a larger pattern of compensatory intervention where
assistance for poor families has been cut: government health benefits facilitate the
movement of former welfare clients into low-wage employment, where private coverage
is rare. As in Germany, therefore, cost containment was coupled with expansions in public
coverage.
There, however, the similarities end. For what stands out in the American experience
is the profound constraints on expansive reform imposed by America’s fragmented polity
and heavily privatized social welfare framework, and the devastating declines in private
social protection that have occurred in the absence of public action. Exhibit A here is the
stunning defeat of the Clinton plan – arguably the most dissected legislative failure in
modern history. Like other pro-competitive proposals, the Clinton plan turned out to be
highly regulatory and complex once the ideal of managed competition had been translated
into legislative details. It is now customary for competition advocates to dismiss the
Clinton plan as a big-government wolf in competitive sheep’s clothing. Yet among
Clinton’s key advisers, the commitment to aspects of pro-competitive reform was in fact
genuine and strong.50 The problem was not insincerity, but inherent barriers to translating
the pro-competitive ideal into politically saleable policy.
The failure of the Clinton plan suggests that the inherited path of policy in the United
States represents an even more substantial barrier to government-led structural reform than
do the well-known hurdles posed by America’s fragmented polity. Business concern about
rising costs, for example, helped push structural reform on to the agenda, as it did in
Germany. Yet, in contrast with their counterparts abroad, American employers were
deeply split along multiple lines by the highly uneven incidence of health costs in
America’s voluntary employment-based system.51 These divisions plagued the Clinton
administration’s efforts and ultimately doomed any chance for employers to play a
constructive and relatively united role. Indeed, the stark difference between US policy
reversals and the more consensual pattern in other veto-ridden/decentralized regimes
cannot be explained without an appreciation of the divisions and barriers created by
America’s distinctive reliance on private insurance. In the end, the Clinton plan was
brought down by much the same political phenomenon that stymied efforts to scale back
public programmes abroad: easily ignited public fears that reform would compromise the
main protection upon which citizens relied – in this case, employment-based insurance.
Although the Clinton plan died, health reform did not – which brings us to

50
Jacob S. Hacker, The Road to Nowhere: The Genesis of President Clinton’s Plan for Health Security
(Princeton, N.J.: Princeton University Press, 1997).
51
Employee Benefits Research Institute (EBRI), Databook on Employee Benefits (Washington, D.C.: EBRI,
1992), Table 6.23.
 Review Article: Dismantling the Health Care State? 721

the final and most striking feature of American policy developments: the extent to which
change occurred independently of the initiative or control of policy makers. The scope of
private control over health benefits is the key reason why. Because policies encouraging
private benefits allow considerable discretion on the part of private actors, they allow
substantial changes within the confines of existing policy through conversion. Further-
more, these non-governmental actors do not have to engage in collective political action
to achieve their ends. If they are able to overcome internal resistance, they can adopt
changes unilaterally.52
And thanks to these largely unilateral changes, the private foundation of the American
system has undergone what can only be described as a radical contraction. From a peak
of more than 80 per cent of Americans, private coverage fell during the 1980s and early
1990s to less than 70 per cent.53 Employment-based protection was the biggest casualty:
between 1979 and 1998, the share of workers who received health insurance coverage from
their own employers fell from 66 per cent to 54 per cent; among the poorest 20 per cent
of workers, it fell from 44 per cent to 26 per cent.54 For more than a decade, the number
of Americans without health insurance has been rising at the rate of about 1 million per
year and now hovers around 44 million. A stunning 75 million Americans, a third of the
non-elderly population, are without insurance at some point in a two-year period, and spells
without insurance are growing longer, with more than half lasting over two years.55 With
employers free to drop coverage, and workers under financial pressure to decline it even
when it is offered, the risk of medical costs is being shifted from insurers and employers
on to workers and their families. For policy makers abroad who wish to increase the role
of private insurance in their medical systems, the continued erosion of American health
insurance sounds a distinctly cautionary note.

R E F O R M W I T H O U T C H A N G E, C H A N G E W I T H O U T R E F O R M

The reform of medical systems has been an important aspect of welfare state restructuring
in recent decades. As constrained budgets have sagged under the weight of rising medical
costs, leaders have found themselves spurred to action, despite the potent risks of
threatening established stakeholders or cherished benefits. Yet once we look past the
feverish talk of reform, the legislative changes undertaken appear at once more modest and
more variable than commonly recognized. As welfare-state scholars have emphasized in
other policy contexts, the dominant pattern of reform over the past two decades is not
radical retrenchment.56 Containing costs rather than cutting benefits has been the major
aim, and these changes have mostly restricted public programmes only at the margins. No
country has seen a contraction of the share of the population entitled to statutory protection
52
On this general point, see Stewart Wood, ‘Labour Market Regimes under Threat? Sources of Continuity in
Germany, Britain, and Sweden’, in Pierson, ed., The New Politics of the Welfare State, pp. 368–409, at p. 374.
53
Health Insurance Association of America (HIAA), Source Book of Health Insurance Data (New York: HIAA,
1996); Kaiser Family Foundation (KFF), The Uninsured: A Primer (Washington D.C.: KFF, 2002).
54
James L. Medoff and Michael Calabrese, ‘The Impact of Labor Market Trends of Health and Pension Benefit
Coverage and Inequality’, Center for National Policy, 2000.
55
Families USA, Going without Health Insurance (Washington, D.C.: Families USA, March 2003); KFF, The
Uninsured.
56
See, for example, Pierson, ‘Coping with Permanent Austerity’, and Dismantling the Welfare State? Reagan,
Thatcher, and the Politics of Retrenchment (New York: Cambridge University Press, 1994); Huber and Stephens,
Development and Crisis of the Welfare State; and Giulano Bonoli, Vic George and Peter Taylor-Gooby, European
Welfare Futures (Cambridge: Polity Press, 2000).
722 HACKER

(although private coverage has been eroded in the United States), and in fact many nations
where coverage was less than universal or left out important benefits have seen expansions
of protection.
Nonetheless, cost containment has imposed visible and unpopular strains in many
nations – from waiting lists and eroding facilities to co-payments and coverage gaps. These
strains have been magnified in popular perception by the persistent acrimony of relations
between the state (and, in decentralized systems, insurers and employers) and medical
professionals. In this environment, leaders have found themselves impelled to take action
not just to cope with fiscal constraints or rising medical spending, but also to avoid blame
for perceived deterioration of strained medical complexes.
My survey of recent developments suggests, however, that the reality of structural
reform has so far lagged markedly behind the rhetoric. In the first place, the market-based
reforms that have attracted the most attention from health policy analysts have scarcely
been universal. Ironically, in fact, it is in nations with the most statist and expansive systems
that market-based reform ideals have been most likely to take root, rather than the
decentralized systems where the idea of competition would seem most congenial. Perhaps
most striking of all, in nearly every case of alleged pro-competitive reform, the largest and
most enduring change was a substantial increase in the power of the state.
More generally, the evidence that I have reviewed suggests that structural reform is not
the crucial catalyst of change we usually assume it to be. Instead, explaining the most
critical shifts – the rise in the private share of spending, the real hardships at the margins
of coverage, and the ongoing privatization of risk in the United States – requires attention
to two powerful sources of internal policy change to which neither health policy experts
nor students of the welfare state have given sufficient emphasis: conversion, or the
decentralized restructuring of policies by actors empowered under them; and drift, or the
failure to update policies to reflect changing circumstances. In a climate characterized by
formidable barriers to expansionary reforms, the scope and effect of welfare states are
crucially evolving despite (and in important respects because of) the hurdles to formal
policy change. New risks are arising with which existing programmes are poorly equipped
to grapple. Gaps in coverage that once seemed tolerable are increasingly hard to close.
Strains on programmes threaten networks of solidarity that once united rich and poor,
healthy and sick. And employers, insurers and even patients – when they have the latitude
– are changing their behaviour and benefits on their own. Because of the semi-sovereign
role of employers in the United States, the effects have been most pronounced, and
inegalitarian, there. But reform without change and change without reform is the dominant
health policy dynamic in advanced industrial states.
That is the crucial message of this article, and it motivates my main explanatory task:
to tease out how and why this contradictory dynamic plays out differently in different
nations. Without insisting on the absolute primacy of one set of factors, I have highlighted
two main sources of variation: whether political systems are veto-free or veto-ridden and
whether medical systems are hierarchical or decentralized. Countries on the same sides of
these twin divides have followed surprisingly similar paths. The veto-free/hierarchical
regimes, for example, have seen a sharp general decline in the state’s financing role, while
the decentralized financing regimes have actually witnessed a general overall increase.
Equally important, reform efforts in veto-ridden and decentralized regimes have not, as
of yet, produced the retrenchment that doomsayers have prophesied. New coalitions have
formed to navigate the shoals of reform, and their efforts have been tempered and
sometimes facilitated by the pluralism of interests in these nations. Even in
 Review Article: Dismantling the Health Care State? 723

the United States, significant expansions of public coverage have come about through this
route.
The picture that emerges, in sum, is one of widespread strains that, at times, have
produced major shifts, but usually without legislative direction. The character of this
non-legislated adjustment is not, however, identical across nations. To the contrary, four
distinct patterns emerge from the analysis, each associated with a distinctive financing
structure: ‘buying out’, ‘passive privatization’, ‘cost sharing’, and ‘risk privatization’. In
hierarchical financing systems, the most prominent form of drift has been buying out in
response to (comparatively effective) restraints on public spending, which have threatened
the timely availability of high-quality care. The Canadian experience is an exception in
the world of hierarchical systems (though it is paralleled by the trajectory of the US
Medicare programme): there, the limited public role at the beginning of the 1980s – which
focused on physician and hospital services, to the exclusion of drugs and other services
– has allowed passive privatization, as spending on services poorly covered by the public
sector has increased.
In more decentralized financing regimes, by contrast, policy drift and conversion have
not generally produced either buying out or passive privatization. Indeed, the trend has
been towards an expanded role for public coverage. Rather, to the extent that cost control
has produced strains, it has resulted mainly in higher consumer cost-sharing. Here, the
United States, as the only nation in which employers have had the latitude to remake health
benefits along the lines they prefer, is the great exception. Only in the United States have
the strains of recent decades produced a true dismantling of key areas of risk protection.
The effect, I have argued, has been a marked privatization of risk, as private coverage has
been eroded and responsibility has shifted from collective risk pools on to families and
workers themselves.
Although I have grouped these diverse trends under the common labels of conversion
and drift, their distributional effects are likely to vary greatly. Buying out, to the extent
that it is limited to the very well off, is likely to exacerbate inequities in access relatively
little – and in fact, if anything, reflects the special dissatisfaction of affluent citizens.
Modest cost-sharing, too, especially when focused on the wealthy, is also likely to have
relatively limited consequences.57 The same cannot be said, however, of passive
privatization and risk privatization, which impose the greatest burden on those with limited
resources. It is notable that among the nations for which adequate opinion data exist, only
in Canada and the United States do lower-income citizens have substantially less
favourable perceptions of the quality of care than do higher-income citizens – in the US
case, dramatically so. In the United Kingdom, in stark contrast, quality is perceived much
more favourably among lower-income citizens, in keeping with the buying-out pattern.58
Researchers would do well to investigate such varied distributional effects more fully in
analyses of health care reform.
Above all, however, the politics of health policy deserves a more prominent place in
our understanding of welfare-state restructuring. Health care reform represents something

57
The available evidence suggests, however, that it also does relatively little to encourage the allocation of
services to those who most need them. Cf. Joseph P. Newhouse, Free for All? Lessons from the Rand Health
Insurance Experiment (Cambridge, Mass.: Harvard University Press, 1996); and Robert G. Evans, Morris L. Barer
and Greg L. Stoddart, ‘Charging Peter to Pay Paul: Accounting for the Financial Effects of User Charges’ (Toronto:
Ontario Premier’s Council on Health, Well-Being, and Social Justice, 1994).
58
Robert J. Blendon et al., ‘Inequities in Health Care: A Five-Country Survey’, Health Affairs, 21 (2002),
182–91.
724 HACKER

of a blind spot in contemporary political analysis, obscured by the emphasis on policy

details in much comparative policy research and by the fascination with classic
income-replacement programmes in much welfare-state scholarship. This article has
suggested that it is a blind spot that contains much of interest for comparative political
analysis. Health care is among the most personal of issues, and this is one reason why it
is also among the most politically volatile. Yet the main message of this article is that
behind the rhetoric and emotion lie powerful and intriguing patterns of policy development
that could deepen our understanding not only of the large corner of social welfare provision
that we call health policy, but also of the modern welfare state and its political future.