Document Details
Conference Paper
Publication Year:

Spanish HTML
Federalism and Health-care Policy

FEDERALISM AND HEALTH POLICY: FIVE CASES Cost Containment in Federal States CONCLUSIONS Chapter One: Federalism and Health-Care Policy: An Introduction Keith Banting and Stan Corbett Governments everywhere are wrestling with health policy.1 They must balance the needs and expectations of citizens, the demands of health-care professionals and the pressures on public budgets; and everywhere the trade-offs are becoming more difficult. In federal countries, these challenges are met through political institutions that require the participation and cooperation of two levels of government in the design and redesign of health policy, adding another layer of complexity to the management of health policies. This study examines the ways in which different federal systems manage the tensions inherent in multi-level governance, and the implications of federalism for the nature of health programs. The analysis is based on a comparative study of the experience of five federations: Australia, Belgium, Canada, Germany, and the United States, which are examined in detail in the chapters that follow. However, the implications of the study extend beyond these five classic federations. Indeed, the historic distinction between federal and unitary states is blurring at the edges as traditionally unitary systems such as Britain and Italy experiment with new forms of regional government, and a wide range of countries devolve important policy responsibilities to a diverse range of regional assemblies and administration. Although important differences remain between federations and devolved systems of governance, many of the political dynamics and policy issues generated by these differing forms of multi-level governance are remarkably similar, and a closer study of health care in federal states therefore has broader lessons.2 A comparative approach establishes a much stronger basis for assessing the policy consequences of distinctive political institutions. Too often, analysts and commentators casually assume that the inability of their own political leaders to solve pressing policy problems must reflect flaws in the political institutions and processes of their country. Most often, however, other countries are wrestling with remarkably similar problems, usually with equally mixed results. A comparative perspective can help identify what is truly distinctive about a country, and provide a much stronger starting point for attempting to assess the policy consequences of any particular configuration of political institutions. This study employs a variety of methodologies to highlight these relationships. It draws on insights from the wider literature on the influence of political structures; it presents quantitative comparisons of federal and non-federal states; and it utilizes a “most similar systems” approach to comparative analysis. This last technique involves the comparison of countries that share many similarities — federal institutions, relatively affluent economies, and common pressures on their health care systems — in order to pinpoint the influence of variation in the nature of their federal institutions. In all of the federations surveyed here, both levels of government are involved in shaping health policy. There are, however, important differences in the form of their federal arrangements and in the design of their health-care systems. It is the interactions between these distinctive political institutions and health care that lie at the heart of this study. This introductory chapter provides an overview of the issues and seeks to draw out the patterns that emerge from the country chapters for several critical policy issues. The first section briefly summarizes the common health-care agenda confronted by governments across the Organisation for Economic Co-operation and Development (OECD). The second section sets the context for the analysis by examining the wider literature on the implications of federalism and asking whether federal and non-federal countries differ systematically in their broad approach to health policy. The third section shifts the focus more directly onto the five federations explored in this book, describing their federal institutions and the ways in which decisions about health policy are made. The fourth section then examines the implications of their federal structures for two basic policy challenges: citizens’ access to health services and the effectiveness of strategies deployed to contain the growth of health-care expenditures. A final section then pulls the threads of the discussion together. As we shall see, these threads are quite strong. While federalism is clearly compatible with a wide range of health care systems, the presence of federal institutions does seem to influence the balance between the public and private sectors in the provision of health care. In addition, differences in intergovernmental arrangements within federations have potent implications for such issues as the equality of health services enjoyed by citizens across the country as a whole, and the capacity of governments to manage health care systems efficiently. A COMMON AGENDA Health policy is a salient issue in virtually every OECD country, and the governments of these countries — federal and non-federal alike — confront a remarkably similar health agenda. In part, this common agenda reflects similar policy goals. In its 1999 World Health Report, the World Health Organization set out a list of six core health goals: improving the health status of the population; reducing health inequalities; enhancing responsiveness to legitimate expectations; increasing efficiency of the health-care delivery systems; protecting individuals, families, and communities from significant financial loss as a result of health problems; and enhancing fairness in the financing and delivery of health care.3 Different countries, including the five examined here, have developed very different approaches to health care, with different public/private mixes, different relationships among service providers, and different delivery systems. Nevertheless, few, if any, of those involved in health policy in these countries would find much to disagree with in this core list of goals. The shared agenda in health care also reflects the fact that OECD countries face a common set of pressures: aging populations, rapid technological change, rising health expenditures, changing understandings of the determinants of health, and a more nuanced appreciation of the relationship between health care and health outcomes. Moreover, in most countries, governments must respond to these pressures in the context of fiscal constraint. Although the public sector has moved into surplus in some countries, international economic competitiveness and domestic political resistance to high levels of taxation mean that few governments can manage the challenges in the health sector by simply opening their wallets wider. Common goals, common pressures and common constraints have generated a shared agenda. Governments throughout the OECD debate a remarkably similar set of issues: the provision of quality health care; cost containment through a mix of price controls, volume controls and/or global budgeting; new incentive structures for health professionals through changes in payment systems; shifts in the public/private mix in health services; responses to rapidly escalating pharmaceutical costs; the need for innovation and flexibility in delivery systems; provision for long-term care and support for the frail elderly; and greater accountability to, and empowerment of, the public. These are the familiar points of discussion among health policy specialists across western, democratic countries. The issue at the heart of this study is whether federal institutions matter to the ways in which our five countries respond. DOES FEDERALISM MAKE A DIFFERENCE? There is a substantial body of literature that explores the extent to which political institutions influence the types of policies that emerge at the end of the day. The broad conclusion that flows from this new “institutionalist” approach is that political institutions alone are never determinative. They interact with other factors shaping policy choices, and it is therefore difficult to identify simple relationships of the sort “federal institutions lead to X.” Nevertheless, the structures of government are seldom completely neutral. They make some outcomes easier than others, and therefore influence the capacity of political agents to act, their perceptions of realistic policy alternatives, their strategic options, and their preferences.4 Studies have suggested that, in combination with other factors, the configuration of political institutions can matter to such things as the size of the public sector, the redistributive role of government, the level of coherence across programs, the interregional distribution of benefits across the country, the level of innovation and flexibility in the policy system, the capacity of governments to resist powerful interests, and so on. Given the contingent nature of the impact of political institutions, one should not expect to find powerful patterns between such broad categories as federal versus non-federal countries. For example, in their basic approach to the welfare state, the five countries examined here fall into three or more of the regime clusters proposed by Esping-Andersen and extended by others.5 Nor does a more consistent pattern emerge when attention is narrowed to the health-care system. As we shall see in greater detail below, our five countries have developed very different regimes of health care. Belgium and Germany have comprehensive health-care systems in which policy is defined through corporatist systems of decision-making and services are managed and delivered locally through networks of social funds. Australia and Canada have developed comprehensive public health insurance programs which are managed by public agencies. And the United States has developed a unique mixed system in which public programs cover the elderly, the disabled and many low-income families, while the rest of the population relies on private coverage and a significant minority lack coverage. The basic structure of health care is clearly rooted in other factors, such as the political culture and the dominant political coalitions in each country, with the particular configuration of political institutions playing a decidedly secondary role.6 Federal institutions do seem to have implications, however, for the size of the public role in the health sector. This pattern is consistent with a substantial body of research that concludes that federalism and decentralization tended to constrain the expansion of the welfare state during the twentieth century. A large number of studies have concluded that federalism and decentralization create several types of barriers that constrain an expansive and redistributive welfare state: by increasing the number of sites of political representation, federalism multiplies the number of veto points at which action can be delayed, diluted, or defeated; by creating separate regional jurisdictions, federalism generates interregional economic competition as state/provincial governments compete for private capital which can exit for other regions with more hospitable fiscal regimes. These arguments have been advanced in a large number of studies of individual federations.7 Although the completeness of the evidence deployed in studies of individual countries has been criticized,8 the argument gains strong additional support from a substantial body of quantitative, cross-national studies of the determinants of social spending. These studies have identified a variety of factors that help explain differences in the expansiveness of the welfare state across OECD nations, including the openness of the economy, the strength of organized labour, and the ideological orientation of dominant political parties. This literature has increasingly focused on the structure of political institutions, and findings repeatedly indicate that, other things being equal, the dispersion of policy-making authority through federalism, decentralization, and other forms of institutional fragmentation is negatively associated with social expenditures as a proportion of gross domestic product (GDP).9 Moreover, a recent contribution to this literature finds that decentralization has more powerful (negative) effects on social welfare effort even than other institutional variables, such as the level of corporatism in decision-making, the nature of the electoral system or the presence of a presidential system of government.10 In addition, new research that approaches the issue from the other side of the ledger by examining the private share of overall social expenditures in OECD nations comes to similar conclusions.11 What about the health-care sector in particular? As Table 1 indicates, total health expenditure, including both public and private spending, actually tends to be higher in federal than non-federal states. However, when attention focuses on the balance between the public and private sector in health care, the pattern that emerges is consistent with the conclusions of the cross-national literature. Table 1, which examines the balance between public and private expenditures on health in 1998, indicates that public spending represented a smaller percentage of total health expenditures in federal states than in non-federal ones (e.g., an average of 70.3 percent versus 79.1 percent). This result is influenced to some extent by the United States, where public spending represents an especially small proportion of total health expenditures, but it does not disappear when the United States is excluded from the calculations: public spending averages 74.6 percent of total health expenditures in all OECD federal states minus the United States, versus the 79.1 percent in non-federal OECD states. Table 2 puts this difference in a longer term perspective by tracking the data from 1960 to 1998. The ratios at the bottom of the table suggest that the difference between federal and non-federal states is a long-standing one. However, they also suggest that the difference has been decreasing over time. The data on the rates of change provide some clues to this partial convergence: the public share of total health spending grew more rapidly in federal states during the 1960s, narrowing to some extent the gap generated by the earlier expansion of the public role in non-federal states; and then federal states did not constrain the public share during the 1980s and 1990s as sharply as did non-federal states. One interpretation of this pattern is that the complexity of decision-making in federal states means that they tend to react more slowly to new conditions than do non-federal states. If so, one might expect the pattern of convergence to slow, as federal states develop stronger instruments of cost containment over time. These differences in spending patterns are intriguing, and we return to some of the issues they pose below in our discussion of the challenges that federal states face in coping with cost pressures in the health sector. However, these spending differences represent only incremental variations on the basic patterns of health care in OECD countries, and the simple contrast between federal and non-federal regimes only takes one so far. Certainly, as Table 3 confirms, there is no reason to believe that institutional differences make an appreciable difference to the overall health of populations within states at broadly comparable stages of economic development. A fuller and more nuanced understanding of the interactions between federalism and health care requires an examination of the experience of specific federal countries. FEDERALISM AND HEALTH POLICY: FIVE CASES A closer look at federal institutions highlights that there is no single model of federalism. Indeed, in the words of one analyst, “perhaps the most intellectually endearing quality of federalism” is that there is “an eclectic array of federal political models … from which other states may draw lessons and experience as they choose.”12 The five federations examined in this study differ considerably in the role of federal and state or provincial governments in health care, the ways in which decisions about health policy are made, and the mechanisms for coordinating relations between different orders of government. It is hardly surprising, therefore, that federalism does not influence health policy similarly in all countries. There is a myriad of ways in which one can compare federal institutions.13 A common approach is to assess the level of centralization between the central and state/provincial governments by measuring the proportion of public expenditures or revenues flowing through each level of government. This fiscal approach to federalism has important advantages beyond mere quantitative precision. Money is power, and command over resources does influence the political balance between levels of government in federations. However, finances represent an incomplete measure of the policy role of different levels of governments. Given the focus in this study, a better measure is the role of different levels of government in defining the basic framework of health policy. Are the core features of a health-care regime set centrally or regionally? This question cannot be answered in precise, quantifiable terms, but it is the heart of the matter. If the basic elements of the health-care system are set centrally, then whether actual expenditures are made locally is a secondary question. In this chapter, therefore, two questions are critical. First, how specific is the country-wide framework for health-care policy? Does it resolve most of the important decisions, or does it leave substantial scope for regional variation? Second, how is the central framework determined? Is it set unilaterally by the central government or do state/provincial governments have an important role in its determination as well? In answering these questions, we draw heavily on the country chapters that make up the heart of this volume, and readers are referred to those chapters for fuller details. Because our analysis draws on two distinct criteria — the comprehensiveness of the central framework and the process of its determination — no simple rank order of centralization is possible. In presenting our five cases, we start with the countries with the most comprehensive and detailed central frameworks. In the case of countries with comparably extensive, common frameworks, we first examine countries in which the central government has the greatest independence in setting that framework. Belgium Although Belgium was established in 1830 as a unitary state, four major waves of constitutional reform in 1970, 1980, 1988, and 1993 created a decentralized federal state, with substantial powers exercised by the new Community and Regional governments. Moreover, a new wave of constitutional reform was launched in 2001. Despite this dramatic restructuring of state structures, however, primary responsibility for social security and health care remains lodged in the central government. The federal government retains exclusive responsibility for health insurance, and sets the framework within which local sickness funds function. In addition, Brussels establishes framework legislation for health-care institutions such as hospitals, setting regulations for planning, accreditation, staffing, equipment, advanced technology and the designation of academic hospitals. According to the calculations in the chapter by Johan de Cock, approximately 97 percent of total health-care expenditures remain in areas of federal jurisdiction. The Communities do have a role in the construction and internal organization of hospitals, and for the organization and management of home care and nursing homes.14 However, they are constrained by federal norms or financing policies in many of these areas and must communicate their decisions — including those pertaining to individual cases — to the central authorities who are responsible for ensuring that central norms are respected. Only in public health and medical education is Community authority dominant. Although power remains highly centralized in health care, policy-making within the federal government is subject to the elaborate consultative mechanisms of corporatist governance that characterize the Belgian political system. Three dimensions of concertation are relevant. Major policy decisions require the consent of federal representatives from both the French- and Flemish-speaking groups. In addition, although the Regional and Community governments have no formal role in the federal legislature, an elaborate set of rules and regulations in the health sector mandate notification and consultation between levels of government. Finally, consultative mechanisms incorporate the social partners and the medical professions into the policy process. As a result, changes in health programs require high levels of consensus among both linguistic blocs and social groups. Despite this consultative tradition, the highly centralized nature of health policy in Belgium has come under powerful pressure as a result of the growth of Flemish nationalism. Flemish nationalists have demanded significant decentralization of key elements of social security, including health insurance, in order to gain more direct political control over their social future and to reduce the inter-communal transfers inherent in the currently centralized system, demands which have been articulated by the new Flemish parliament. However, determined resistance from the Wallonian community to any further decentralization has blocked such changes and, for the moment at least, Belgian health care represents a centralized corner of a decentralized, binational state. Germany As in the case of Belgium, federalism plays a secondary role in health policy in Germany. As the chapter by Dietmar Wassener makes clear, the framework for health programs is set, with few exceptions, through federal legislation that applies evenly across the country. However, the basic features of health policy are defined through highly corporatist processes that incorporate not only representatives of the social partners — such as employers and employees, health professionals and social funds — but also both levels of government. The Länder governments are directly involved in the federal legislative process through the Bundersrat, the upper level of the federal parliament, and establishing the parameters of health policy therefore requires a high level of intergovernmental consensus. The legislative framework established in this way defines the core features of the statutory system of health insurance, which covers over 90 percent of the population. The framework sets minimum standards for health services; the principles governing contribution rates, including maximum rates; and maximum budgets for the hospital sector, ambulatory care, and pharmaceuticals. Within this common framework, the specific design and delivery of health services are highly decentralized, being the responsibility of close to 600 independent social funds. These funds are managed by representatives of employers and employees, and the services they provide are financed by contributions from employers and employees rather than tax revenues from the state. Within the national parameters, the funds decide on the precise health services they will cover, set their contribution rates, and negotiate contracts with associations representing doctors and hospitals. As a result, the direct delivery responsibilities of governments are limited. The Länder governments do have responsibility for public health promotion and for financing the capital costs of hospitals. This last provision produces a dual source of financing for hospitals, with the Länder providing for capital costs and the social funds providing the bulk of operating costs, a division that occasionally produces conflict between the two. Nevertheless, in the parlance of new public management, the primary function of government in health care is steering, not rowing. In such a system, it is hard to define precisely where leadership ultimately comes from. Decisions to reform the statutory health insurance system are taken at the national level, and are normally led by the federal government, usually in concert with state governments controlled by the same party. However, their proposals tend to be modified during negotiations with representatives of the funds and provider associations, and during passage through the upper house. Indeed, the process requires such a high level of intergovernmental and social consensus that critics, including Wassener, complain of an institutionalized rigidity and inflexibility that constrains policy-making in a rapidly changing world.15 Australia Federalism is more central to the politics of health care in Australia, but once again the federal government plays the leading role in setting the basic parameters of health policy, as Linda Hancock’s chapter demonstrates. A 1946 constitutional amendment extended the Commonwealth’s powers to include laws on pharmaceutical, sickness and hospital benefits, and medical and dental services. As a result, Medicare, the national health program established in 1984, is delivered in two parts, one purely federal and the other federal/state in design. The Commonwealth government provides directly for access to doctors, pharmaceuticals, and nursing homes under the Medical Benefits and Pharmaceutical Benefits Schemes, which are administered by the Commonwealth’s Health Insurance Commission and operate on similar terms across the entire country. In contrast, access to public hospital care is established through bilateral Commonwealth/state agreements, which are renegotiated every five years. However, the Commonwealth government exerts considerable influence here as well, and hospital care preserves the characteristics of a common “national” service. The primary instruments of federal influence are Special Purpose Payments (SPPs), grants to states that are subject to conditions aimed at ensuring compliance with national goals. The main SPP in the hospital sector, known as the Health Care Grant, is ringed with highly detailed requirements; for example, all public hospitals in the states are expected to comply with performance targets set by the Commonwealth Department for Health and Family Services, which constantly audits to ensure that the targets are met. During the 1980s, the reliance on the conditional SPPs did shift as a result of changes in the government in power in Canberra, with Labour governments favouring conditional grants and Liberal/National coalitions tipping the balance more toward unconditional block grants. Since then, however, the role of SPPs has been reinstated as a continuing source of central influence on hospital care, and even conservative administrations have shown little inclination to relax conditions attached to federal social program funding. In addition, as the chapter on Australia by Linda Hancock indicates, the Commonwealth government also influences health services through other initiatives that cover government services as a whole, such as its National Competition Policy. Although the parameters of health policy therefore tend to be set centrally, state governments have opportunities to influence the federal framework through a complex set of intergovernmental bodies that characterize Australian federalism. The Commonwealth Grants Commission, which is jointly appointed by the two levels of government, plays a critical role in managing intergovernmental conflicts over financial issues. In addition, The Premiers’ Conferences, ministerial conferences, the Loans Council and, in recent years, the Council of Australian Governments (COAG) represent sites of intergovernmental negotiation in which important initiatives are developed jointly and intergovernmental conflicts managed. The Commonwealth government tends to provide the leadership within these bodies, but the network of institutions does provide for regular debate and coordination between levels of government. Although these mechanisms of intergovernmental coordination are not as powerful as in the German case, they do ensure that many social programs are guided by national, as well as more narrowly federal objectives. United States The United States resembles Australia in relying on separate federal and state delivery of different components of the public health care-system. However, the US represents a more starkly bipolar case, combining both highly centralized and highly decentralized programs. Medicare, the largest public program, is a purely federal program, with few intergovernmental aspects. It covers virtually all people over age 65 and about five million disabled individuals under age 65; and it represents about two-thirds of the total public health-care expenditures in the country. Policies concerning Medicare are determined by the federal Congress and the program is administered across the country by a federal agency. Throughout its life, Medicare has been sustained by strong public support, powerful bureaucratic champions, and protective congressional committees. As a result, this largest pillar of the public health insurance system operates as a single program across the country as a whole. In contrast, the second, smaller pillar of the public system operates in a decentralized fashion. Medicaid, a health program targeted mainly at the welfare poor, and the State Child Health Insurance Program, a new initiative directed at children from low-income families, are federal-state programs supported by federal conditional grants and delivered by state governments. In Medicaid, broad federal guidelines determine general eligibility and coverage standards, but leave considerable room for states to tailor their programs to local conditions and preferences. In addition, the federal administration has the authority to grant waivers from some regulations to individual states to provide for experimentation in program design. As we will see in greater detail below, the result is that state programs vary considerably in eligibility requirements, service coverage, utilization limits, provider payment policies, reliance on managed care and spending per recipient. Moreover, the State Children’s Health Insurance Program provides even more flexibility than does Medicaid. In contrast to most other federal systems, the mediation of federal-state conflicts in the United States does not flow through formal intergovernmental mechanisms. Although the courts can and do play a role, conflicts over the terms and conditions of federal support flow into national politics, with state governments bringing pressure to bear on Congress, especially the Senate where these issues tend to be resolved. Although members of the Senate normally protect the interests of their state in battles over such issues as the funding formula for federal-state programs, they are independent political agents and do not necessarily agree with or speak for the state governor and administration in matters of general health policy. Thus, in the final analysis, it is the central government that resolves intergovernmental tensions in the federal-state components of the system. Canada Health policy in Canada constitutes the most decentralized of the five systems examined here. Health insurance and health services generally fall within provincial jurisdiction, and the first steps toward universal health insurance took place at the provincial level, with the province of Saskatchewan playing a leading role. Unlike Australia and the United States, the federal government does not provide any significant portion of health coverage directly to citizens as a whole.16 Federal influence has been exerted through financial transfers to provincial governments, which facilitated the extension of provincial innovations across the country as a whole and the establishment of a pan-Canadian approach to Medicare during the postwar years. However, the politics of the Canadian federation ensured from the outset that the conditions attached to federal transfers were less specific than in other federations; and the shift from conditional grants to block-funding for health care in 1977 largely eliminated day-to-day federal scrutiny of specific provincial decisions. As Antonia Maioni’s chapter highlights, Canadian health care is best thought of as a series of provincial health insurance systems operating within broad federal parameters. The federal legislation, the Canada Health Act, specifies that provincial insurance plans receiving federal funding must reflect five principles: they must provide universal coverage; they must cover all “medically necessary” services; they must be publicly administered; coverage must be portable outside the province; and accessibility must not be limited by user fees or extra-billing by physicians, both of which are prohibited by the Act. Within these parameters, provinces shape health policy and delivery systems as they see fit. Provincial governments define the “medically necessary” services that are actually covered, and some minor differences have emerged across the country. Provinces also have responsibility for the delivery process, and larger organizational differences have developed here. Provinces regulate hospitals, clinics, nursing homes, and other health institutions; they negotiate fee schedules with doctors and other health professionals; they set global budgets for hospitals; and they have the final responsibility for the costs of health care. In this context, it is not surprising that provincial governments have been taking the lead in the restructuring of health-care delivery in Canada, and that — as we shall see below — there are growing differences in the governance and delivery mechanisms in the health-care sectors across the country. The tensions between federal parameters and provincial responsibility were intensified in the 1980s and 1990s as federal contributions to provincial health budgets were cut, especially in the 1995 federal budget. At that point, what had been a long-standing intergovernmental tension flared up into a full-fledged political warfare between the two levels of government. Unfortunately, Canada had few powerful intergovernmental mechanisms to help manage the conflict. Unlike Germany, provincial governments have no role in the federal legislature; unlike Belgium, there were no formal requirements for advance notification and consultation; and unlike Australia, there were no standing intergovernmental institutions or expert commissions to coordinate elements of the relationship. Intergovernmental negotiations in Canada operate through an informal assemblage of committees at the level of officials, ministers, and prime ministers. In the aftermath of the 1995 cuts, the federal government and all of the provinces except Quebec reached a compact known as the Social Union Framework Agreement, which provides a modest level of structure for these processes. But by the standards of several other federations, the intergovernmental structures in Canada remain weak compared to the intensity of the divisions. The Overall Patterns It is striking that in all of our federations, health care operates within a broad policy framework which sets core features of the system for the country as a whole. Health care involves the provision of highly personal services to individuals in diverse settings, a circumstance that has led many theorists to suggest a decentralized approach, and the state/provincial level has important roles in all of these countries. Even in the most centralized of the five federations examined here, the federated units have significant responsibilities for health institutions such as hospitals and clinics. But it appears that in no federation is health policy a purely regional responsibility. This is true for other countries not examined in detail here, such as Switzerland, often considered the most decentralized of federations.17 Nevertheless, as we have seen, federations clearly differ considerably in both the comprehensiveness of the federal framework, and in the ways in which decisions about that framework are made. Setting policy parameters is a highly centralized and corporatist process in Belgium and Germany, although program management and delivery proceeds on a decentralized basis through networks of social funds. Australia and the United States are middle-level cases. In both countries, the central government has full responsibility for important components of the health-care insurance, delivering the program directly to citizens. Both countries also rely on shared-cost programming for other components of the system, but Australia establishes more complex conditionality for its transfers in such programs. The state governments are more directly represented in the process of defining health policy than in the United States. Finally, Canada is the case among these five in which the common framework is most limited. The federal government delivers no significant component of the health-care system directly to citizens, and the principles associated with the Canada Health and Social Transfer (CHST) are quite general, leaving most of the big policy decisions to provincial governments.18 The balance between central and regional governments is constantly evolving in federations. Pressures for further decentralization exist in virtually all of these countries, and decentralist steps have been taken in some places. However, the trend is not uniform. In some cases, pressures for cost containment, which are discussed more fully below, are generating centralizing dynamics. Germany is one country in which federal legislation has intervened more extensively than in the past in efforts to contain health expenditures. Australia also settled into a more centralized model, after brief experiments with greater decentralization in the 1980s. In other cases, pressures for decentralization have had limited impact. In Canada, the federal government’s financial contribution dropped during the 1980s and 1990s, arguably weakening the political legitimacy of its role in the system. However, the policy parameters embedded in the Canada Health Act were not relaxed, and the federal government has recently reinstated financial contributions cut in the mid-1990s. Moreover, where there has been decentralization in health care, it has tended to be less extensive than in other policy sectors. In Belgium, the establishment of a federal system did see the transfer of limited responsibilities for health services to the new Community and Regional governments; but social security, including health insurance, remains highly centralized in comparison with most other important policy sectors, and more recent demands for decentralization advanced by the Flemish Community have been blocked. In a similar vein, the federal government in Canada accepted significant decentralization in social assistance and labour market programs, but has resisted pressures for a similar shift in health care. The decision in the United States to decentralize significant control over Medicaid is therefore something of an exception in these five countries. Even here, however, it is worth noting that although social assistance was shifted to a block-grant mechanism, a similar proposal for Medicaid was vetoed by President Clinton. In many ways, therefore, the continued role of central governments is a striking pattern in these countries. Health care seems to retain a special political sensitivity that constrains pressures for decentralization. THE IMPACT OF FEDERALISM ON HEALTH CARE POLICY What then is the impact of these differences in federal institutions on health-care policy? The case studies in this volume give very different answers to this question. At one extreme, analysts from countries with powerful central frameworks and consensual decision processes tend to assign the importance of federalism to a secondary status. Johan de Cock concludes that “the impact of federal state structures on health policy (in Belgium) is still quite limited;” and Dietmar Wassener argues that “federalism will continue to play a secondary role in shaping German health care.”19 At the other extreme, Antonia Maioni concludes her study of Canada by noting that “federalism is a defining feature of the Canadian health care model.”20 Not surprisingly, judgements about the two intermediate cases, Australia and the United States, are more qualified, and tend to focus less on the implications of federalism for the basic characteristics of the health-care system and more on the efficient management of the system. In Australia, Linda Hancock points to federal obstacles to efficiency and reform, citing a recent commission report that recommended “where practicable, it is best to avoid multiple levels of government involvement in the first place.”21 But in the case of the United States, David Colby is also restrained about the importance of federalism, arguing that “our lack of rationality in program development does not lie in our federal system, but in our party system and government.”22 A closer comparison of the experiences revealed in the country chapters, however, does throw a slightly sharper light on the influence of federal institutions on health policy. This chapter draws out those comparisons by exploring issues embedded in two distinct agendas that in combination define the core tension in health-care politics in OECD nations: access to health care on one side and budgetary planning on the other. In addressing the access agenda in a federal context, we concentrate on the extent to which citizens in all regions of a country receive comparable levels of health services. In addressing the planning agenda, we explore the capacity of these five countries to pursue strategies to constrain the growth of health-care expenditures. Access to Health Care: Social Citizenship and Regional Diversity Every federal state must establish a balance between two social values: a commitment to social citizenship, to be achieved through a common set of public services for all citizens across the entire country; and respect for regional communities and cultures, to be achieved through decentralized decision-making and significant scope for diversity in public services at the state/provincial level. The debate over this balance is an ongoing one in all federations. The discourse varies from one country to another, and in practice discussion can quickly become embroiled in amazingly technical issues of intergovernmental finance and complex points of constitutional interpretation. But the underlying question is both simple and profound. Which community should be paramount in the definition of social benefits: the “national” community of all citizens on one hand; or regional communities defined by state/provincial boundaries on the other?23 There is no single answer to this question. The appropriate response will vary from federation to federation, depending on the nature of political identities and the conceptions of community embedded in its culture. As used in this context, the concept of social citizenship is not restricted to universal programs provided to each and every citizen. Selective or targeted programs are also relevant if they function similarly across the country. The issue is whether citizens in similar economic and social situations are treated equally, irrespective of where they live in the country, or whether the public benefits to which a citizen is entitled also depends significantly on the region in which he/she resides. Does a sick baby in one region have access to the same level of care on similar terms and conditions as a sick baby in another region of the same country? The balance between common benefits and regional diversity in federations is influenced by two key instruments: the strength of the federal policy framework established for health policy, and the strength of interregional financial transfers. As noted in the previous section, a common framework can be established in two ways. In some federations, important health programs are designed and delivered directly to citizens by the federal government, as in the case of Medicare in the United States and the Medical Benefits and Pharmaceutical Benefits schemes in Australia. These programs operate on a country-wide basis, providing all citizens with common benefits for an important component of the health-care system. A second approach is central legislation that sets policy parameters within which other agencies design and deliver health programs. In Belgium and Germany, such legislation sets the framework for social funds which administer health insurance. In other federations, federal legislation establishes parameters for state/provincial governments, as in the case of Medicaid in the United States, hospital services in Australia, and health care generally in Canada. The second instrument critical to the agenda of social citizenship is interregional transfers. The case for a powerful system of interregional transfers in a federation lies in the conviction that citizens in all parts of a country should be entitled to comparable benefits and services without having to pay significantly different taxes.24 Richer regions in any country enjoy the virtuous circle of fewer social needs and greater revenue capacity; poorer regions confront a vicious circle of greater social needs and weaker revenue capacity. Sustaining a common or even comparable benefit/tax regime in such circumstances inevitably requires some form of interregional transfer. Such transfers also allow federations to minimize the danger that regional differences in tax and benefit levels will begin to influence migration of both capital and individuals across the country, helping to avoid the much discussed twin dangers of “capital flight” and “welfare magnets.” In federal states, interretional transfers take two forms. In programs delivered directly by the central government, the transfers are implicit rather than explicit, resulting from the differential impact of common benefits and taxes across regions of uneven economic strength. Such transfers tend to be hidden, but they are no less real for their opaque nature. In the case of programs delivered by other authorities, whether social funds or state/provincial governments, transfers are more explicit. In Germany, health insurance is funded through contributions levied by the social funds themselves, and reducing variation in the benefit/contribution package across plans has led to the development of a major inter-fund redistribution scheme, known as the risk equalization mechanism (REM), which is well analyzed in the chapter by Wassener. Although the REM was not designed as an explicitly interregional transfer mechanism, it does have the effect of shifting resources among regions of the country. In addition, massive transfers from the west to east Germany have been required to create comparable standards in public services, including health care, across the old divide. In countries in which major health programs are delivered by state/provincial governments, net interregional transfers are embedded in formal transfer mechanisms. In some cases, as occurs in a limited way in the case of Medicaid in the United States, redistribution is built into the funding formula for the program. In other countries, interregional redistribution flows through a separate program, as in the Belgian “national solidarity” grant, the Canadian equalization program, the inter-Länder transfers in Germany, and the system of adjustments to intergovernmental transfers in Australia. The politics of interregional transfers are becoming increasingly controversial in virtually all federal states, as the various chapters in this book make clear. Interestingly, the form of redistribution does not seem to affect the intensity of political debate. One might expect the less visible, implicit transfers embedded in centrally delivered programs to be less politically contested; and certainly few people seem to care, for example, about the interregional transfers embedded in the Pharmaceuticals Scheme in Australia. However, the insulating effects of implicitness are hardly perfect. In the United States, the formula governing payments for health maintenance organizations (HMOs) under Medicare sparked a political battle between the rural and urban states which the Senate had to resolve. More dramatically, Flemish nationalists in Belgium have mounted a powerful political challenge to the transfers implicit in the federal government’s social security and health insurance programs. In this case, the opaque nature of the transfers probably created more opportunities for radical nationalist politicians to make inflammatory statements about their size.25 Explicit transfers to other governments or social funds can also attract political heat. Once again, however, there appears to be no neat correlation between the form of transfer and the level of controversy. Embedding interregional redistribution in the general funding formula for health programs was received with relative calm in the case of Medicaid in the United States, perhaps because no separate equalization programs exists in that country.26 In Canada, however, richer provinces traditionally fight hard against a differential formula in the federal transfer for health care, insisting that interregional redistribution should be limited to the separate equalization program. In Germany, both forms of transfer have generated recent challenges. In many ways, the commitment to interregional solidarity is strongest in Germany, and the Basic Law empowers the federal government to act to ensure “the establishment of equal living conditions throughout the federal territory.” 27 Despite this commitment, political challenges have emerged in recent years. The risk equalization mechanism, which is strictly speaking an inter-fund transfer rather than an inter-Lander transfer, has been challenged legally by some health funds and politically by some Länder governments.28 In addition, the richer Länder in the south have launched legal challenges to the general-purpose, tax-financed inter-Länder transfer scheme, complaining about the recipient regions in the north. In the end, the politics of interregional redistribution seem rooted less in the form of the transfer, and more in the underlying level of political support for notions of solidarity and social citizenship. Federal countries differ in their tolerance of regional variations in tax and benefit packages, reflecting different levels of commitment to the equality of individual citizens on the one hand and respect for cultural differences, regional autonomy, and decentralization on the other.29 Among our five federations, the strongest levels of interregional redistribution to support health care seem to be found in Belgium, Germany, and Australia. Canada seems to fall into an intermediate category. The Canadian constitution includes a commitment to an equalization program to support less affluent regions, but the funding formula is not as powerful as in these other federations. It is the United States, however, that defines the other end of the spectrum. There is no separate program for equalizing the fiscal capacity of state governments, and states receive relatively limited fiscal assistance from the central government, constraining the capacity of poorer states to provide average levels of public services or to introduce innovative programs on their own.30 In combination, the specificity of the central framework and the strength of interregional redistribution set the structural underpinnings of the balance between social citizenship and regional diversity in the definition of health care. The patterns across the five federations are summarized in Figure 1. Belgium, Germany, and Australia comprise one group, characterized by strong common frameworks and strong interregional redistribution. The United States and Canada represent contrasting cases. The United States has an intermediate framework but a low level of interregional redistribution, whereas Canada has the leanest common framework but a middle level of interregional redistribution. These structures define the real policy room available for distinctive regional or local approaches to health care. How that policy room is used in practice depends on a much wider range of factors: the extent of cultural and political differences across the country; differences in the relative strength of stakeholders; and so on. A fuller analysis of the determinants of health-policy choices at the regional level goes beyond the scope of this study. However, the case studies in this volume do shed light on the extent of regional variation in health benefits that does result from the interaction of national frameworks, interregional redistribution, and distinctive regional societies. Interregional Variation in Health Services. Figure 1 also summarizes the overall extent of interregional variation in the health care systems in each of our five federations. Not surprisingly, a common standard of health benefits across the country seems strongest in Belgium, Germany, and Australia. In the case of Australia, for example, there are no significant regional variations (except for the case of the Northern Territories) in such dimensions as the number of hospital beds per 1,000 population or per capita use of medical services, and the primary geographic inequalities in access to health services tend to be between urban and rural areas within each region.31 Moreover, the basic structure of health-care policy and delivery has become more uniform across the country with the expansion of the federal role under Medicare after 1984. Fifteen years ago, there were substantial differences between states in hospitalization rates, costs and public expenditures…. Most of these have disappeared. The high spending states have all pared health outlays at the same time that previously low spending ones have raised them. There were equally large structural differences within state systems. Queensland and Tasmania were traditionally “public” states, Victoria a “private” one with New South Wales having the most complex interweaving of the two. Much of this has also gone. Membership of private insurance reflected the same systemic diversity…. There is now no significant difference in coverage between the states and it would be very surprising if there were one.32 The impact of differences in the strength of common frameworks and interregional transfers is also highlighted by the contrast with Canada and the United States. The Canadian package of federal principles embedded in the Canada Health Act and equalization between rich and poor provinces produces a common approach to eligibility and a relatively common package of health services for Canadians across the country, as the tables reported in Antonia Maioni’s chapter indicate. Health expenditures in poorer provinces represent a significantly higher proportion of provincial GDP than in richer provinces, something that would be highly unlikely without interregional transfers. Within these common parameters, however, the Canadian system leaves considerable scope for provincial variation; and different provincial approaches to restructuring and expenditure restraint are generating progressively larger differences in governance, management; and health service delivery. As a team of leading commentators remarked, “With the exception perhaps of Quebec, over the past twenty-five years the provincial health care systems have shared not only the five principles of Medicare but also similar delivery and management structures. In the coming years they may resemble each other only in sharing the principles of Medicare….. These divergent paths will challenge the concept of a “national” system, if such a conception ever existed.”33 As we have seen, the US represents a bipolar case. Medicare establishes a common approach to public health services for elderly and disabled Americans, but the combination of a weak framework and weak interregional transfers in the area of health care for poor children and families means that their protections are subject to marked regional disparities. Differences in eligibility meant that in 1994 Medicaid beneficiaries, as a proportion of the low-income population, varied from a high of 79 percent in Vermont to lows of 30 percent in Nevada, 36 percent in South Dakota, and 39 percent in Florida.34 Variation in service levels are also clearly implicit in the differences in average payments per recipient of Medicaid services, which in 1998 ranged from $8,961 in New York to $2,386 in California, a difference which cannot be explained away simply as a reflection of different health costs in these two affluent states.35 Federalism and Access to Health Care: The Overall Pattern. The strength of the commitment to comparable access to health services is a striking feature of these federations. Despite the underlying importance of diversity embedded in the logic of federalism, these five federations have organized themselves so as to constrain interregional variation in the access to public health services enjoyed by citizens across the country. This is true both of multi-nation federations such as Belgium and Canada and federations such as Australia and Germany in which cultural diversity is less regionally concentrated. There are undoubtedly important regional variations in many federations, as evidenced by state differences in Medicaid in the United States and the increasingly distinctive delivery systems across Canada. Nevertheless, it would appear that there is limited scope for pervasive differences in health services in contemporary federal systems, an issue to which we return at the end. Cost Containment in Federal States Health-care systems throughout the OECD world have been under cost pressures during the last 20 years, and governments have debated a wide array of policy instruments they hope will slow the seemingly relentless growth of health expenditures. The choices have ranged from relatively blunt instruments such as global caps on expenditures for particular programs, to changes in the mix of services provided, to more complex instruments designed to change the incentives facing citizens and service providers.36 The approach to cost containment adopted in any individual country is shaped by the structure of its health-care system. As Tuohy has argued, the basic structure of a health-care system creates its own internal logic, which governs the way in which it responds to external pressures. 37The structure also determines the sorts of levers that governments can hope to use most readily to constrain expenditures: the single-payer system in Canada presents different levers than the public/private mix in the United States. Changing the structure of a health-care system and creating completely new levers requires the mobilization of substantial political will by government leaders. Such major interventions are rare. The attempt by the Thatcher government in the UK to introduce “internal markets” in the National Health Service and the reform effort of the Clinton administration in the US represented two such initiatives, and both fell short of their champions’ initial aspirations, dramatically so in case of Clinton’s proposed reforms. In the main, therefore, governments tend to rely on the mechanisms that the health care system makes available to them. Federal institutions do add an additional layer of complexity to cost containment, and it is striking that federal states do seem to have greater difficulties in containing cost pressures. As Table 4 indicates, federal states consistently devoted a larger portion of their GDP to health expenditures than did non-federal states throughout the 1960–98 period (9.9 percent versus 7.8 percent in 1998). This is true whether or not the United States, which has committed a larger portion of its GDP to health than any other OECD country since 1970, is included; excluding the United States only lowers the average for federal states to 9.3 percent. The rate of increase in health expenditures as a percent of GDP has also been somewhat higher in federal states than in non-federal states, a pattern that began in the 1980s and continued into the 1990s. While health costs in federal states increased by an average of 30.3 percent between 1980 and 1998 (25.7 percent when the United States is excluded), the average increase in non-federal states was only 16.4 percent. Table 5 indicates that the higher rate of increase also held for public spending on health. The increase in federal states was over 33 percent, a rate that is reduced only to about 32 percent by the exclusion of the US. Over the same time period, the increase in non-federal states was a mere 3.3 percent. Why should federal states have greater difficulty in containing cost pressures? It has often been argued that attempts to contain health spending in one area simply shift the pressures elsewhere in the system, much as when a balloon is squeezed at one end and expands at the other. Federal systems may be more prone to cost-shifting in two directions. First, as we saw earlier, the private health sector tends to play a larger role in federations, increasing the opportunities for cost-shifting between public and private funders. Second, the participation of two levels of government in shaping public health programs increases the chances that cost containment will involve cost-shifting between governments. The simplest form of this dynamic occurs when a central government reduces its transfers to state or provincial governments, without simultaneously easing conditions attached to the funding, as in the case of unfunded federal mandates in the United States, or the practice of the Canadian government of reducing its transfers to provinces while continuing to enforce the principles established in the Canada Health Act. This logic would seem to suggest that efforts to constrain public health spending in federations will be easiest when control over the key policy instruments is effectively lodged at one level of government, whether at the federal or the state/provincial level. In cases when control over the key levels of cost containment is divided, the prospects for cost containment would seem to depend heavily on the effectiveness of mechanisms of intergovernmental coordination. These dynamics can be illustrated by an examination of several of the countries considered in this study. As de Cock notes in his chapter on Belgium, the concentration of control over the fiscal levers in health policy in the central government virtually eliminates the scope for cost-shifting between levels of government. In the case of Germany, while service delivery by funds represents a very decentralized system, control over the key policy levers is concentrated at the national level, and the decision system requires a high level of intergovernmental consensus. In this context, federalism represented no barrier to the sweeping effort at cost containment in the Structural Health Reform Act of 1993. The legislation contained both a massive short-run effort to stabilize costs and a longer term strategy of structural reforms designed to alter the underlying dynamics within the health-care system. This strategy included significant changes in the system of hospital remuneration, a redefinition of the division of labour between general practitioners and specialists, controls on the number of doctors and their regional distribution, changes in reimbursement for drugs, greater competition among health funds for clients, and the risk-equalization mechanism designed to level the playing field on which they would compete. Although there was debate about the relative effectiveness of the various components of the 1993 legislation, it clearly represented a substantial package. Because the legislation affected the Länders’ budgetary role in the hospital sector, the legislation had to be adopted in the Bundesrat as well as the lower house in parliament. As a result, it required an intergovernmental consensus as well as one between the social partners. Such approval is not automatic. In 2000, the federal government proposed another Statutory Health Insurance Reform Act, which included a comprehensive maximum budget, not just for specific areas, but for the system as a whole. The proposal had to be dropped in response to opposition in the Bundesrat. Nevertheless, the combination of central responsibility for framework legislation and a powerful mechanism of intergovernmental coordination has proved consistent with broad action on cost containment. At the other end of the spectrum, the highly decentralized Canadian model also provides the capacity for powerful cost containment. Admittedly, the Canadian system provides considerable scope for burden-shifting, as noted above. The federal government shifted the potential risks inherent in health-cost pressures to provincial governments in the late 1970s when it moved from open-ended cost-sharing to block-funding. During the 1980s and 1990s, Ottawa shifted actual costs, by reducing the block-fund transfer. In addition, the conditions attached to the federal Canada Health Act preclude certain options at the provincial level, especially those involving user fees and co-payments. Within those constraints, however, the provincial governments have the advantage of a single-payer model, which provides powerful levers for cost containment. Provincial governments limited their own financial obligations by capping hospital and then physician services budgets, and by closing many hospitals and restructuring others. In a single-payer system, service providers were left with nowhere to shift their costs, and they were drawn into expanded and intensified negotiations with provincial governments and with the regional structures established in most provinces.38 This highly concentrated power proved relatively effective in containing costs. During the period between 1992 and 1997, when Canadian governments were struggling to eliminate their substantial deficits, public spending on health care declined an average of 2 percent each year. Although private health spending did grow somewhat in this period, the dominance of the public sector in core health services ensured a substantial slowing of overall health expenditures in Canada in that period. The consolidation of power at one level clearly does create the capacity for governments to squeeze the system when needed. Federations in which spending responsibilities are more evenly divided between the two levels of government would seem to face larger challenges in ensuring that cost-containment efforts do not degenerate completely into cost-shifting. This challenge is illustrated most vividly in the United States, in which the multiplicity of payers in both the public and private sectors makes integrated cost-containment strategies impossible. CONCLUSIONS In keeping with other studies of the role of political institutions, the conclusions that flow from this study suggest that federal institutions on their own are never determinative. At the broadest level, federalism is clearly compatible with a wide range of health-care systems: large public roles and small public roles; corporatist systems relying on social partners to deliver benefits; systems relying on management through government agencies. On average, the public share of health-spending in federations is somewhat smaller, a pattern that echoes findings in the larger literature on the welfare state. But, in general, the simple distinction between federal and non-federal systems does not take one very far. Other factors, including the clash of economic interests and political ideologies as well as the norms and values embedded in the underlying culture leave powerful imprints on the health-care systems that emerge in different countries. Nevertheless, the structure of political institutions does have an important role in shaping the ways in which competing interests and groups engage in the struggle to shape health policy. From this perspective, federalism matters a great deal. But how it matters depends on the particular structure of federal institutions, and the ways in which they are rooted in the wider political environment. Laying bare these complex linkages between federal institutions and health policy requires careful examination of different federal countries, such as those appearing in the chapters of this book. The broad patterns that emerge from these country comparisons are striking. First, in none of these federation is health policy a purely regional responsibility. Federations vary enormously in the responsibilities of the central and regional governments in health policy, the role of regional governments in shaping the health policies of the central government, the nature of fiscal relations between the two levels of government, and the mechanisms for coordinating their programs. However, the central government plays a role in all of these systems, and decentralist pressures have had less impact on the balance between central and regional governments in this policy sector than in many others. The political sensitivity of health care seems to ensure that the politics of health policy resonate across the country as a whole, even in systems otherwise marked by highly regionalized policy-making in other sectors. Moreover, the particular structure of federal institutions and norms in each country has important implications for key features of the health-care system. As we have seen, federalism seems to matter to at least two distinct agendas at the heart of health-care politics: the agenda of social citizenship and the agenda of rational planning. The division of labour between levels of government and the nature of interregional fiscal relations have powerful implications for the distribution of health services among citizens across the country as a whole. Whether a sick child in one region has access to the same level of treatment on comparable terms and conditions as a sick child at the other end of the country depends heavily on the often arcane details of federal institutions, norms, and fiscal relations. The extent to which federal states have succeeded in establishing interregional evenness in health services is striking. In this area, federal states look much like non-federal ones, suggesting perhaps that democratic politics is less tolerant of interregional inequality in health care than in other forms of inequality. “It is through the territorial units they live in,” Sidney Tarrow reminds us, that citizens “organize their relations with the state, reconcile or fight out conflicts of interest, and attempt to adapt politically to wider social pressures.”39 Other dimensions of inequality are not represented as directly in political life. Similarly, the structure of federalism seems to matter to the capacity of governments to manage health-care systems effectively, ensuring that resources are used wisely and that systems respond smoothly to change. None of the linkages is simple. But none of them is unimportant. Hence the enduring debate about the central role of the institutional framework in the politics of health care in federal states. FIGURE 1 Interregional Variation in Health Care: Instruments and Outcomes Country Instruments Outcomes Specificity of Policy Framework Interregional Transfers Interregional Differences in Health Care Systems Belgium high high low Germany high high low Australia high high low United States medium low medium Canada low medium medium NOTES 1. This chapter benefited from the research assistance of Vincent Melillo. We were also helped by comments from Jacob Hacker and participants in a policy seminar series at Brandeis University where an earlier version was presented in May 2001. 2. For a study that draws on the experience of several federal states in order to anticipate the issues involved in the devolution of health services to regional authorities in Italy, see George France, ed., Federalismo, Regionalismo e Standard Sanitari Nazionali: Quattro Paesi, Quattro Approcci (Milano: Dott. A. Giuffrè Editore, 2001). 3. World Health Organization, The World Health Report 1999 (Geneva: World Health Organization, 1999). 4. For discussions of this approach, see E. Immergut, “The Theoretical Core of the New Institutionalism,” Politics & Society 26,1 (1998):5-34; F.W. Scharpf, “Institutions in Comparative Policy Research,” Comparative Political Studies 33, 6/7 (2000):762-90, and Games Real Actors Play: Actor-Centered Institutionalism in Policy Research (Boulder: Westview Press, 1999); P. Pierson, “Increasing Returns, Path Dependence and the Study of Politics,” American Political Science Review 94,2 (2000):251-67; and K. Thelen and S. Steinmo, “Historical Institutionalism in Comparative Politics,” in Structuring Politics: Historical Institutionalism in Comparative Analysis, ed. S. Steinmo, K. Thelen and F. Longstreth (Cambridge: Cambridge University Press, 1992). For applications to health policy, see J. 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 Policy Development 12 (1998):57-130; E. Immergut, Health Politics: Interests and Institutions in Western Europe (Cambridge: Cambridge University Press, 1992); and A. Maioni, Parting at the Crossroads: The Emergence of Health Insurance in the United States and Canada (Princeton, NJ: Princeton University Press, 1998). 5. G. Esping-Andersen, The Three Worlds of Welfare Capitalism (Princeton, N.J.: Princeton University Press, 1990); and F. Castles, “Australia’s Strategy for Coping with External Vulnerability,” in The Comparative History of Public Policy, ed. F. Castles (Cambridge: Cambridge University Press, 1989). 6. It is true that none of the five federations has established an integrated national health service on the British or Swedish model, in which the state directly delivers health services to citizens through medical professionals who are employed by the state and hospitals and other institutions owned by the state. Establishing such a comprehensive role for the state in a federation would be particularly challenging, and it is perhaps not surprising that these federal states rely on public insurance systems. I am indebted to Jacob Hacker for this point. 7. This literature has long roots. For early examples, see H. Laski, “The Obsolescence of Federalism,” reprinted in The People, Politics and the Politician, ed. A. Christensen and E. Kirkpatrick (New York: Holt 1941); and A.H. Birch, Federalism, Finance and Social Legislation in Canada, Australia and the United States (Oxford: Clarendon Press, 1955). For an excellent review of more recent literature, see P. Pierson, “Fragmented Welfare States: Federal Institutions and the Development of Social Policy,” Governance 8,4 (1955):449-78. 8. A. Noel, “Is Decentralization Conservative?: Federalism and the Contemporary Debate on the Canadian Welfare State,” in Stretching the Federation: The Art of the State in Canada, ed. Robert Young (Kingston: Institute of Intergovernmental Relations, Queen’s University 1999). Noel surveys a number of country studies, but does not consider the large cross-national quantitative literature on the subject. 9. For a sampling of these studies, see E. Huber, C. Ragin and J. Stephens, “Social Democracy, Christian Democracy, Constitutional Structure and the Welfare State,” American Journal of Sociology 99,3 (1993):711-49; A. Hicks and J. Misra, “Political Resources and Growth of Welfare in Affluent Capitalist Democracies, 1960-1982,” American Journal of Sociology 99,3 (1993): 668-710; A. Hicks and D. Swank, “Politics, Institutions and Welfare Spending in Industrialized Democracies, 1960-82,” American Political Science Review 86,3 (1992): 658-74; M. Crepaz, “Inclusion versus Exclusion: Political Institutions and Welfare Expenditures, Comparative Politics 31 (1998): 61-80; and D. Cameron, “The Expansion of the Public Economy: A Comparative Analysis,” American Political Science Review 72 (1978): 1243-61. 10. D. Swank, “Political Institutions and Welfare State Restructuring: The Impact of Institutions on Social Policy Change in Developed Democracies,” in The New Politics of the Welfare State, ed. P. Pierson (Oxford: Oxford University Press 2001). 11. See Jacob Hacker, The Divided Welfare State: The Battle over Public and Private Social Benefits in the United States (Cambridge: Cambridge University Press, forthcoming). Hacker draws on new OECD data on net social expenditures, which includes both public and private expenditures on central social programs, in assessing the impact of institutional factors on the balance between public and private action across countries. 12. L. Brown-John, “Other Models of Federal Systems Reviewed,” in Federal-Type Solutions and European Integration, ed. L. Brown-John (Lanham: University Press of America, 1995), p. 195. 13. See especially R. Watts, Comparing Federal Systems. 2d ed. (Kingston and Montreal: Institute of Intergovernmental Relations and McGill-Queen’s University Press, 1999). 14. Belgian federalism is characterized by considerable asymmetry, and these responsibilities are managed by the Regional government in the French-speaking community. 15. See also Fritz Scharpf, “The Joint-Decision Trap: Lessons from German Federalism and European Integration, Public Administration 66 (1988):239-78. 16. The only exceptions are the direct federal responsibility for health care in the northern territories and for specific classes of people, such as Aboriginal peoples, members of the armed forces, and inmates in federal prisons. 17. Under the Swiss constitution, health insurance is a federal responsibility. Under a major reform in 1996, federal health insurance legislation established a universal obligation on all residents of the country to insure themselves with one of the approximately 120 companies offering health insurance, and specified the package of services that companies must include in the basic health insurance scheme. Companies remain free to set their premiums, and differences do exist both within and among cantons. However, the federal legislation establishes a risk equalization mechanism designed to offset skimming by companies, and the federal-cantonal transfers compensate for the impact of insurance costs on low-income residents. On the post-1996 system, see E. Thuerl, “Some Aspects of the Reform of the Health Care Systems in Austria, Germany and Switzerland,” Health Care Analysis 17, (1999):331-54. For fuller discussions of the pre-1996 system, see OECD, The Reform of Health Care Systems: A Review of Seventeen OECD Countries (Paris: OECD, 1994), ch. 20; and P. Lehmann, F. Gutzwiller and J. Martin, “The Swiss Health Care System: The Paradox of Ungovernability and Efficacy,” in Success and Crisis in National Health Care Systems, ed. M. Field (New York: Routledge 1989). 18. For a different ranking of the extent of decentralization in the health-care sector in federal states, see P.G. Forest and K. Bergeron, “Les politiques de réforme du système de santé dans cinq fédérations: une analyse de travaux scientifiques récents,”in Federalism and Sub-National Policies, ed. L. Imbeau (forthcoming). In assessing the level of decentralization, Forest and Bergeron appear to focus primarily on federal-state programs, and give much less attention to programs delivered directly by central governments across the country as a whole. This has substantial implications. In the case of the United States, for example, they focus on Medicaid and seem to ignore Medicare. This leads to the conclusion that health care is more decentralized in the United States than in Canada. Unfortunately, the assessment seems to ignore most of the public action in health care in the United States. 19. J. de Cock, “Federalism and the Belgian Health are System,” in this volume; D. Wassener, “The German Health-Care System,” in this volume. 20. A. Maioni, “Federalism and Health Care in Canada” in this volume, p….. In a similar vein, Jacob Hacker concludes in his comparative study that “the most distinctive aspect of the Canadian welfare state … is the prominent role that federalism has played in its development,” Hacker, “The Historical Logic of National Health Insurance,” p. 96. 21. L. Hancock, “Australian Intergovernmental Relations and Health,” in this volume. 22. D. Colby, “Federal-State Relations in United States Health Policy,” in this volume. 23. The use of quotation marks around “national” community denotes the ambiguity of the language of nationhood in some federations. In multi-nation federations such as Canada and Belgium, political identities are multiple and overlapping, and the phrase “national” can be ambiguous. In the Canadian case, for example, Québéçois political leaders — both federalist and separatist — claim that Quebec is a “nation” and that another term is required for the Canada-wide community of all citizens. 24. Watts, Comparing Federal Systems, p. 50. 25. For details, see K. Banting, “Social Citzenship and the Multicultural Welfare State,” in Citizenship, Diversity and Pluralism: Canadian and Comparative Perspectives, ed. A.C. Cairns, J.C. Courtney, P. MacKinnon, H.J. Michelamnn and D. Smith (Montreal and Kingston: McGill-Queen’s University Press, 1999). 26. In contrast, the uneven impact on different states of the financial formula associated with the shift to block-funding in social welfare did generate intense controversy in Congress. 27. Article 72.2 of the Basic Law for the Federal Republic of Germany. The Basic Law had originally authorized action to preserve “the uniformity of living conditions,” but was recently amended to provide somewhat more flexibility in the expectations implicit in the constitutional language. 28. We are indebted to Deitmar Wassener for this supplementary information. 29. See Watts, Comparing Federal Systems, pp. 50-53. 30. The ranking of the strength of interregional redistribution in different federations is consistent with the analyses to be found in Watts, Comparing Federal Systems; R. Bird, Federal Finance in Comparative Perspective (Toronto: Canadian Tax Foundation, 1986); and L.S. Wilson, “Lessons for Canada from Other Federal Systems,” in Equalization: Its Contribution to Canada’s Economic and Fiscal Progress, ed. R. Boadway and P. Hobson (Kingston: John Deutsch Institute for the Study of Economic Policy, Queen’s University, 1998). 31. See the data in Australian Institute of Health and Welfare, Australia’s Health 2000: The Seventeenth Biennial Health Report of the Austrian Institute of Health and Welfare (Canberra: Australian Institute of Health and Welfare, 2000), such as Tables 5.22, S30, S31 and S50. With the exception of the Northern Territory, variation on most dimensions tends to be less than +/- 10 percent of the national average. The report gives far more attention to the urban/rural divide than to inter-state differences. 32. J. Debble, as quoted in G. Gray, “Access to Medical Care under Strain: New Pressures in Canada and Australia,” Journal of Health Politics and Law 23,6 (1998):905-47. 33. J. Hurley, J. Lomas, and V. Bhatia, “When Tinkering is Not Enough: Provincial Reform to Manage Health Care Resources,” Canadian Public Administration 37,3 (1994):514 . The authors anticipate that greater provincial variation in management and delivery will inevitably generate challenges to the basic principles of Medicare as well. 34. D. Liska and A. Salganicoff. Medicaid Expenditures and Beneficiaries: National and State Profiles and Trends, 1988-1994, 2d ed. Report of the Kaiser Commission on the Future of Medicaid, November 1996, Table 6. See also M. Sparer, Medicaid and the Limits of State Health Reform (Philadelphia: Temple University Press, 1996). 35. US Department of Health and Human Services, Health Care Financing Administration, A Profile of Medicaid: Chart Book 2000 (Washington, DC: Government Printing Office, 2001), Figure 2.8. 36. For a sophisticated comparative study of the complex factors driving health costs, see OECD, New Directions in Health Care Policy (Paris: OECD, 1995). 37. C. Hughes Tuohy, Accidental Logics: The Dynamics of Change in the Health Care Arena in the United States, Britain, and Canada (New York: Oxford University Press, 1999). 38 Ibid., p. 249. 39. Sidney Tarrow, “Introduction,” in Territorial Politics in Industrial Nations, ed. Sidney Tarrow, Peter J. Katzenstein and Luigi Granziano (New York: Praeger, 1978), p. 1.