Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pharmaceutical programs and social policy development: comparing Canada, Australia and the UK Boothe, Katherine 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2010_fall_boothe_katherine.pdf [ 1022.05kB ]
JSON: 24-1.0071041.json
JSON-LD: 24-1.0071041-ld.json
RDF/XML (Pretty): 24-1.0071041-rdf.xml
RDF/JSON: 24-1.0071041-rdf.json
Turtle: 24-1.0071041-turtle.txt
N-Triples: 24-1.0071041-rdf-ntriples.txt
Original Record: 24-1.0071041-source.json
Full Text

Full Text

 PHARMACEUTICAL PROGRAMS AND SOCIAL POLICY DEVELOPMENT: COMPARING CANADA, AUSTRALIA AND THE UK   by Katherine Boothe  B.A. (Honours), The University of Alberta, 2004 M.A., The University of British Columbia, 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Political Science)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  JULY 2010     © Katherine Boothe, 2010  ii  Abstract  Canada is the only OECD country that provides broad public health benefits but lacks a universal, nation-wide system for funding prescription drugs. This puzzle cannot be explained by the literature on national health insurance, which suggests that the tendency to consider all health services as a single policy has missed an important source of cross-national variation. How can we explain the lack of a major pharmaceutical program in Canada, in light of the country’s own extensive health system and the experience of almost all other welfare states? More generally, why do some countries adopt universal, comprehensive pharmaceutical programs, while others do not? To answer these questions, the study compares Canada to the UK and Australia using a process-tracing approach, and finds that the range of services in a country’s public health system is determined by the earliest decisions about how to approach policy development. Where institutional, ideological and electoral conditions allowed for large-scale change and all services were introduced simultaneously, countries tended to maintain the full scope of services. But where institutional barriers, ideological dissensus and low issue salience made radical change difficult, health programs were introduced incrementally, and policy development tended to stall after the first priority. Although incrementalism was initially less politically risky, it was also inherently limiting. Barriers to the introduction of services increased over time, and services that were initially lower priorities (such as pharmaceuticals in Canada) were pushed off the public agenda. In investigating this phenomenon, I provide specific mechanisms by which a more limited “path” of policy development becomes “dependent,” and argue that we must consider not only the role of ideas in policy making, but also the role of ideas over time. The study also investigates the implications of the approach to policy development for subsequent policy outcomes. It finds that factors that support the simultaneous adoption of a full range of health services also make it more difficult to retrench these services later on.     iii Table of Contents  Abstract ........................................................................................................... ii Table of Contents............................................................................................. iii List of Tables................................................................................................... vi Acknowledgements.......................................................................................... vii 1 Introduction and theoretical framework ......................................................... 1 Introduction .................................................................................................. 1 The puzzle: variation in public provision of pharmaceuticals....................................................... 1 Research question: the adoption of national pharmaceutical programs ....................................... 3 Variation in benefit schemes: an under-studied area .......................................... 4 Overview of theoretical approach..................................................................... 8 The implications of simultaneous versus incremental reform ....................................................... 8 Choice of priorities and barriers to later stages ............................................................................. 9 Early challenges of pharmaceutical programs............................................................................. 10 Hypotheses ..................................................................................................10 Incrementalism and critical moments for change ....................................................................... 10 Institutions: centralized authority and preferences of veto players ............................................. 12 Ideas: principled and programmatic ........................................................................................... 14 Electoral incentives ...................................................................................................................... 15 Historical variables ...................................................................................................................... 16 Summary .....................................................................................................19 2 Method ......................................................................................................22 Introduction .................................................................................................22 Defining the scope of the project .....................................................................22 Countries chosen ......................................................................................................................... 22 Time period chosen ..................................................................................................................... 24 Methodological approach: process tracing .......................................................28 Terms and concepts ......................................................................................31 Methods for collecting data ............................................................................33 Preparation: secondary literature and published primary documents ........................................ 33 Fieldwork: archival sources and elite interviews.......................................................................... 34 Methods for analyzing data ............................................................................35 Observable implications: linking theory to evidence..........................................37 Simultaneous versus incremental policy development ................................................................ 37 Order of priorities and barriers to later stages............................................................................. 38 Response to early challenges of pharmaceutical programs ......................................................... 39 Conclusion ...................................................................................................40 3 Canada: health insurance “capable of being introduced by several stages” ........41 Introduction .................................................................................................41 Historical overview ...................................................................................................................... 41 The first crucial decision: health insurance in stages .........................................42 Fragmented authority .................................................................................................................. 42  iv Lack of principled ideas about health.......................................................................................... 45 Lack of electoral motivation ........................................................................................................ 47 The second crucial decision: determining the order of priority ...........................49 Resource constraints: the cost of pharmaceuticals ...................................................................... 49 Preferences over priorities: provincial veto players ..................................................................... 51 Barriers to further policy development........................................................................................ 54 An alternative path of pharmaceutical policy: drug prices and patents.................59 Public expectations and the demand for price control ................................................................ 59 The developing consensus on drug prices ................................................................................... 62 The end of federal pharmaceutical proposals: the Drug Price Program, 1972 .......64 Lack of principled ideas ............................................................................................................... 64 Lack of electoral motivation ........................................................................................................ 67 Conclusion ...................................................................................................69 4 Australia: “a high grade service” that requires “progress step by step” .............71 Introduction .................................................................................................71 Historical overview ...................................................................................................................... 72 The first crucial decision: a decisive move to stages ..........................................73 Centralized authority through fiscal means................................................................................. 74 Lack of principled ideas ............................................................................................................... 75 Lack of electoral motivations ....................................................................................................... 78 The incremental decision ............................................................................................................ 81 The second crucial decision: determining the order of priorities.........................83 Resource constraints: the lack of medical manpower.................................................................. 83 Barriers to policy implementation and further development ..............................85 Opposition from organized interests ........................................................................................... 85 Lack of further policy development............................................................................................. 87 Implementing the PBS: the role of public expectations .............................................................. 89 The early challenges of the Pharmaceutical Benefits Scheme ..............................90 Formulary control and early cost concerns ................................................................................. 91 Revised formulary and doctors’ preferences ............................................................................... 93 Patient contributions and public expectations............................................................................. 95 The second step in health policy development, thirty years later .........................98 Centralized authority................................................................................................................. 100 Principled ideas.......................................................................................................................... 101 Electoral motivations ................................................................................................................. 102 Conclusion ................................................................................................. 104 5 The United Kingdom: “something bolder than a mere extension and adaptation of existing services” ....................................................................... 107 Introduction ............................................................................................... 107 Historical overview .................................................................................................................... 108 Choice of simultaneous approach to policy..................................................... 109 Centralized authority................................................................................................................. 110 Principled ideas and electoral motivations ................................................................................ 112 The early challenges of prescription services .................................................. 116 Concerns about the cost of prescription services ....................................................................... 117 Prescription charges, 1952-1956: accounting for public expectations ...................................... 119 Prescription charges, 1964-1968: the role of principled ideas and electoral incentives ............ 124 Conclusion ................................................................................................. 129  v 6 Conclusions .............................................................................................. 131 Introduction ............................................................................................... 131 Summary of evidence .................................................................................. 132 The choice of simultaneous versus incremental reform ............................................................ 132 The choice of priorities for services adoption............................................................................ 135 Barriers to further policy development...................................................................................... 136 Early challenges of pharmaceutical programs........................................................................... 140 Contributions ............................................................................................. 143 Directions for future research ...................................................................... 146 References .................................................................................................... 148    vi  List of Tables Table 2.1 Summary of outcomes and causal factors………………………………………25   vii  Acknowledgements I am very grateful for the support of the many individuals and institutions that made this research possible. I received financial support from the Social Sciences and Humanities Research Council of Canada during three years of graduate studies, and was assisted in my fieldwork by a research assistant position for the Centre for Health Services and Policy Research at the University of British Colombia. I benefited greatly from the enthusiastic involvement of my supervisory committee. Steve Morgan was an invaluable resource on the nuances of pharmaceutical policy and provided important challenges to my interpretation of the research problem. Alan Jacobs’ detailed and insightful comments greatly improved the thesis, and will be a resource for me as I revise and build on it for the next project. Kathryn Harrison has provided encouragement, guidance and the best kind of constructive criticism throughout my graduate program, and is a model for how I hope to advise my own students some day. I could not have asked for a better committee. My fieldwork was made much easier and more enjoyable by my visits to LSE Health, at the London School of Economics, and the Political Science program in the Research School of Social Sciences at the Australian National University. These institutions generously supplied workspace, library access, and a chance to connect with other graduate students and faculty who made me feel welcome and at home. Here at UBC, I was fortunate to participate in the Pharmaceutical Policy Group at the Centre for Health Services and Policy Research at the University of British Colombia, whose members taught me more about current issues in pharmaceuticals and pharmaceutical policy than I could have ever learned on my own, and who kindly provided timely and thoughtful feedback on my work. I am also grateful for the feedback I received from the Department of Political Science’s graduate writing group and Canadian- Comparative Workshop, and particularly for the help of Campbell Sharman in reading the Australia chapter, suggesting resources, and pushing me to clarify my conclusions. Finally, I would like to thank my “thesis support group”: the good friends I made in the political science graduate program at UBC, and my family. My parents provided loving support and helped me navigate the ins and outs of academic life, and my brothers and sister learned not to ask when I was going to finish! Most of all, my husband Kevin has also lived with this project for lo these many years, and was constant in his belief I could do it, even on those days that I was not. This thesis is dedicated to him.  1  1 Introduction and theoretical framework Introduction The puzzle: variation in public provision of pharmaceuticals Pharmaceutical products are a crucial component of modern health care. They may improve quality of life, substitute for more costly and invasive therapies, and occasionally produce major medical breakthroughs, as was the case with insulin, penicillin, and more recently, chemotherapy for cancer treatment and antiretroviral treatments for HIV/AIDS (Quirke 2005). Pharmaceuticals are also an important element of health policy, as they consume an increasing proportion of countries’ GDP and seem set to remain an important cost-driver in the health sector for the foreseeable future.1 Discussions about how best to subsidize drugs and regulate their prices, evaluate the effectiveness of innovations, and ensure reasonable access to medically important and potentially cost-saving therapies take place in both government and academia.2 One question that has not been extensively discussed or studied, however, is why similar countries develop different responses to the problem of pharmaceutical provision in the first place. This question is an overlooked dimension of welfare state development. It has significant implications for those with a normative interest in how governments are involved in providing pharmaceuticals, and on a more general level, for those with an interest in how health policy fits into welfare state regimes, and how these regimes develop in different contexts. In seeking to understand pharmaceutical programs’ development, the research is driven by a particular empirical puzzle: Canada lacks a nationally universal and comprehensive pharmaceutical insurance or benefit program, which is surprising given its provincially-organized but nationally comprehensive public, universal health insurance system. This is all the more  1. An OCED study found that although “[p]harmaceutical expenditure has been rising steadily as a share of GDP since 1970...pharmaceutical expenditure has, on average, maintained its share of total health expenditure in the OECD, close to 15 %” (Jacobzone 20005). 2. For example, in recent years both the US and Australia have undertaken significant reforms to their drug benefit policies. See White House press release (2006) Strengthening Medicare; and Australian Department of Health and Aging (2007) Pharmaceutical Benefit Scheme (PBS) Reforms. Recent Canadian examples of drug plan discussions include a 2003 Institute for Research on Public Policy conference entitled “Toward a National Strategy on Drug Insurance: Challenges and Priorities,” and 2004 issue of HealthCare Papers devoted to pharmacare in Canada (see (Morgan and Willison 2004, 10-20), as well as the reports of two major health reform commissions which contained recommendations for drug benefits or insurance (Commission on the Future of Health Care in Canada 2002; Senate Standing Committee on Social Affairs, Science & Technology 2002) and the progress report on the National Pharmaceuticals Strategy (Federal/Provincial/Territorial Ministerial Task Force 2006).  2 surprising given that universal public drug insurance or benefits exists in all other OECD countries save the United States, Mexico, and Turkey (Jacobzone 2000, 29). Although Canadian provinces offer varying levels of pharmaceutical coverage, these programs are not equivalent to the broad public pharmaceutical programs present in similar welfare states, such as the UK and Australia. This disparity means that explanations of cross-national variations in health policy in general cannot solely explain pharmaceutical policy. Until now, the accepted explanation for the puzzle of Canada’s lack of a broad pharmaceutical program was that it was simply attempted “too late.” It is argued that it was fiscally impossible to develop a comprehensive scheme after the major boom in pharmaceutical innovation in the 1960s because of “constant pharmaceutical cost-escalation” (Morgan and Willison 2004): drugs were becoming more expensive and more widely used, making the initial development of a program much more difficult. Drugs were more costly in the 1960s, but this does not explain why Canada failed to act on pharmaceuticals early on, as did other countries like the UK and Australia. A related argument is that drugs were not seen as an important element of therapy when the earliest discussions of health insurance occurred in Canada. It is true that in the mid-1940s, the number of effective drug therapies was limited: penicillin only reached levels of mass production near the end of World War II (Quirke 2005), and tetracyclines and corticosteroids3 were only widely available for prescription in the early 1950s (Rivett 1998). However, the 1940s were the beginning of a global “therapeutic revolution” in pharmaceuticals, and many countries recognized their present and growing importance at this time by developing policies to ensure patient access to needed drugs. While the limits of pharmaceutical therapy in the 1940s and the escalating cost of pharmaceuticals in the 1960s are undoubtedly factors in Canadian policy non- development, they do not provide a satisfactory explanation alone. They fail to explain why Canada did not develop a program at the same time as similar countries, and cannot account for attempts to expand programs after this important policy window closed. This project is designed to determine the conditions under which countries develop universal- comprehensive pharmaceutical services or benefits.4 It also aims to provide new insight into the  3. Two major new drugs that had a significant impact on therapy: tetracycline is a broad-spectrum antibiotic and corticosteroids are a class of steroid hormones used for a wide range of inflammatory conditions. 4. This refers to a public system for funding pharmaceuticals that provides coverage to the entire population, for a full range of drugs. For more details, please see the “Terms and Concepts” section in chapter two.  3 mechanism by which history matters for health policy variation that may be applicable to other policy areas, particularly those involving the provision of social benefits. Other authors have demonstrated the influence of key moments or turning points on health policy development. For example, Hacker’s (1998) work on national health insurance demonstrates how variation among health policies in Canada, the UK, and the US can be traced to three key historical decisions (Hacker 1998, 83). However, my research points to an earlier decision that is even more basic, which helps explain the later decisions highlighted by other authors, and crucially, provides insight into variation in pharmaceutical policies as well as general health insurance. That key initial decision is whether a country will make a radical change all at once, at a historic window of opportunity, or whether it will build its public health care system piece-by-piece over time. This is a choice in the sense that countries must go in one direction or the other if they are to develop policy, and in the sense that policymakers considered alternatives to the direction actually taken. However, the decision about policy approach is not a completely self-conscious and rational instance of policy deliberation. Rather, it is strongly affected by the variables familiar to students of comparative public policy: the country’s institutional structure, the ideas of policymakers, and the electoral incentives they face. Where institutional, ideological and electoral conditions allow for large-scale change and all health services are introduced simultaneously, countries tend to maintain the full scope of covered services. But where institutional barriers, ideological dissensus and low public salience make radical change difficult, health programs are introduced incrementally. Barriers to the adoption of additional services increase over time, so the sequence of service introduction becomes a critical determinant of the range of health services offered. Thus, in the UK, where the government had centralized authority over health, was committed to the idea of a universal and comprehensive health system, and faced an attentive and supportive electorate, all health services including pharmaceuticals were introduced simultaneously. Canada and Australia lacked two or more of these conditions, and so introduced only one service initially (hospital insurance in Canada, and pharmaceutical insurance in Australia). However, in both these countries, taking the “next step” of health policy development proved much more difficult that policymakers had originally anticipated. Research question: the adoption of national pharmaceutical programs The primary research question concerns program adoption: why do some countries adopt universal, comprehensive pharmaceutical programs, while others do not? A closely related secondary question concerns implementation: how did countries respond to the early challenges  4 of pharmaceutical programs (namely, higher-than predicted usage and expenditure) that threatened their universality and comprehensiveness? As I discuss below, the lack of a universal, comprehensive program of pharmaceutical insurance is a key distinction between Canada and most other welfare states, but in Australia, pharmaceutical benefits were not made truly comprehensive until ten years after their adoption (contrary to the initial program design), when the range of covered drugs was extended while other cost control measures were adopted. In the UK, cost control measures were introduced within five years of the program’s adoption amongst grave concerns about their effect on the program’s scope of coverage. In both Australia and the UK, the first policy response to these challenges occurred before causal factors that are important later in policy development, such as the interventions of the pharmaceutical industry, became influential, so it was possible to examine the causes of early cost control measures without accounting for the more complex factors. I used this as an opportunity to test the causal effect of the conditions of program adoption on early implementation.  I argue that focusing on the fundamental decision between a radical “big bang” approach to health policy and a seemingly less risky incremental approach allows us to see the role of ideas, electoral incentives and institutional complexity in shaping later health policy decisions. This decision, and the decision regarding priorities for an incremental approach to policy development, have not been explicitly examined in previous research. Explaining these decisions provides insight into Canada’s puzzling outlier status with regards to pharmaceutical programs, and lays a foundation for future work addressing variation in long-term program outcomes, as pharmaceutical policies continue to develop and adapt to the challenges of the current day. Variation in benefit schemes: an under-studied area An examination of the causes of cross-national variation in pharmaceutical programs represents a significant gap in the literature. There are two main senses in which this is the case: the causes of variation in benefit schemes are not discussed in depth in the existing literature on pharmaceutical policies, and they cannot be accounted for by existing literature on variation among national health insurance systems. The literature on pharmaceutical policy that is most focused on causal explanation addresses cross-national variation in pharmaceutical regulation (particularly drug approval and pricing), such as Wiktorowicz (2003) and Permanand and Mossialos (2005). These authors use institutional factors and the role of interests, particularly the pharmaceutical industry, to draw conclusions about the differences in pharmaceutical regulation across North America and the EU, and to  5 evaluate the prospects for regulatory harmonization. The regulatory policy this literature focuses on, however, is managed by different agencies than benefit schemes, has different objectives, and is generally characterized by quite different politics. These regulatory policies interact with benefit policies in important ways: for example, when faced with fiscal pressures, policymakers may have to decide between cutting benefits or increasing regulation of industry. However, regulatory policy is by no means synonymous with pharmaceutical benefits policy, and its development cannot explain initial differences in pharmaceutical programs. Drug expenditure literature deals with the causes and consequences of rising expenditures, and strategies for managing expenditures (see for example Morgan 2002; 2005). The latter is especially useful in discussing the theoretical and empirical impacts of various policies that make up pharmaceutical programs: on the demand side, patient co-payments and various efforts to affect physicians’ prescribing behavior, and on the supply side, policies to control which drugs are subsidized as well as the prices or profits of pharmaceutical companies (Jacobzone 2000). This literature is mainly concerned with “how policies work” rather than “how policies come about and why they differ.” Recent works focusing on expenditure management and/or pharmaceutical regulation that involve comparative descriptions of policies include Willison et al (2002), Kanavos (2002) and Mossialos, Mrazek and Walley (2004).  A final relevant approach in the pharmaceutical policy literature evaluates specific programs (in some instances, comparatively) according to a pre-determined set of criteria and goals, for example, access and efficiency (Morgan and Willison 2004), quality of care (Lexchin 2004), fiscal sustainability (Marchildon 2006), and integration with the larger health system (Forgot 2002; Forest 2004). There is also literature that seeks to describe the particular historical context of single systems (such as Sloan 1995). Löfgren (2001) provides an interesting discussion of the evolution of benefit schemes in both Australia and Sweden, but restricts his causal conclusions to predictions about future reforms, rather than system origins. Once again, a causal account of the variation in pharmaceutical programs is missing from this literature. The national health insurance literature is more focused on causal explanations for policy variation, but the incongruence in the Canadian case, which combines universal and comprehensive hospital and medical insurance with the absence of a nation-wide universal pharmaceutical program, demonstrates that it cannot provide a full explanation for variation in pharmaceutical programs. This literature (Hacker 1998; Immergut 1992; Maioni 1998; Tuohy 1999) suggests explanatory variables such as political and economic institutions, past policies, and social forces such as political parties. Many of these works (particularly Hacker) emphasize the  6 path dependent, contingent nature of policy development in this area. This consensus on the relevance of history to policy development directs us to consider more general theories of historical effects for an explanation of variation in pharmaceutical programs (discussed below), but the specific explanations for variation in national health insurance do not provide a satisfactory answer to the puzzle. Hacker (1998) explains variations in public health services or insurance in Canada, the UK and the US by examining three parallel decisions in each country: the degree and form of private insurance allowed to develop; the initial target population of government programs; and “the relative timing of public efforts to bolster the technological sophistication of medicine, on the one hand, and to increase the access of citizens to health care, on the other;” (Hacker 1998, 83) or how technologically advanced and costly health services are before government attempts to insure or subsidize them. Hacker argues Canada, the UK and the US faced similar critical junctures for health policy, and that institutions structure these rare opportunities for major change. The timing and sequence of policy development matters, as past policy choices regarding the three decisions he highlights put significant constraints on the way future health policy developed. My theoretical framework benefits greatly from Hacker’s approach to looking for commonalities in critical junctures across cases, and from his identification of feedback mechanisms in health policy development. However, his argument does not consider pharmaceuticals, and the three decisions he focuses on cannot extend to an explanation for variation in drug insurance. He predicts that if private insurance has time to develop, it will crowd out public insurance. There was no private insurance for pharmaceuticals in the UK or Australia before their public systems were adopted, but private drug insurance was also practically non-existent in Canada until the 1970s, after a national pharmaceutical  program had been rejected a number of times. A broad initial target population is meant to promote the expansion of public insurance, while a specialized or expensive set of initial beneficiaries hinders expansion, but this is also unhelpful, as the UK and Australia provided some limited public drug coverage for workers or veterans, respectively, before universal programs were implemented, and Canada had no categorical programs. According to Hacker’s hypotheses, this should have resulted in broad public drug coverage for the UK, and a program that remained limited to a “clientele that are seen…as outside the mainstream of the economy” in Australia (Hacker 1998, 82). However, there are no predictions for cases lacking any early programs, as did Canada in my research period. Finally, the suggestion that achieving technological sophistication before attempts to expand access to health services blocks public insurance is problematic for pharmaceuticals  7 because pharmaceutical innovation tends to spread quite quickly, so there is little cross-national variation in technological sophistication, at least among advanced industrial countries, and furthermore, the UK was home to the most innovative local pharmaceutical industry at this time, as well as early and generous prescription services. Maioni (1998) compares health policies in Canada and the US and finds the role of a social- democratic party, as shaped by institutions, provides a key explanation for the significant differences in the scope of public health insurance in these two countries. Although I rely on Maioni’s insightful and detailed accounts of actors’ motivations in a number of places in the Canadian case, the full explanation is again not transferable to pharmaceuticals. Although in both Australia and the UK, pharmaceutical programs were introduced under Labour (social democratic) governments, Maioni’s framework relies on the absence of a social democratic party in the US. Canada, the UK and Australia all had social democratic parties at this time, and while British and Australian Labour parties actually formed governments and the Canadian Cooperative Commonwealth Federation/New Democratic Party did not, it is not clear why Canadian parties would be strong enough to put pressure on the federal government with regards to medical and hospital insurance but not pharmaceutical insurance, or why they would be uniquely unconcerned with pharmaceuticals, compared to social democratic parties in other countries. Tuohy (1999) explains variation in health policy initiation and reform in Canada, the UK and the US by examining the way institutions and ideas shape windows of opportunity for major policy change. As detailed below, these concepts provide a basis for my own hypotheses concerning opportunities for simultaneous or incremental reform, but Tuohy’s specific explanation focuses on the 1960s as a starting point for Canadian health insurance and omits discussion of debates in the 1940s, when pharmaceutical insurance was first considered and rejected. Although there is less comparative literature that includes Australia, Gray’s (1995) examination of health insurance and federalism in Australia and Canada provided important historical details to my study. She discusses the introduction of the Australian Pharmaceutical Benefits Scheme, but does not consider the absence of pharmaceutical insurance in Canada, and uses her cases mainly as an illustration of theories of federalism rather than using theory as an explanation for variation. When considering the match between the national health insurance literature and variation in pharmaceutical benefits, we must also account for the fact that although drugs are linked to other medical services, and are currently a major component of both therapy and  8 overall health care costs, the policies are often subject to quite different dynamics. Pharmaceutical programs may or may not be fully integrated into the larger health system: in the UK, drug benefits are part of the National Health Service that came into being in 1948 (Robinson and Dixon 1999), while in Australia, the national Pharmaceutical Benefits Scheme began in 1950, more than twenty years before the current state-administered health insurance scheme (Medicare, started in 1975) (Biggs 2003; 2004). Also, pharmaceuticals have historically been subject to user charges more frequently than other health services or products (currently some form of fees or co-payments is the norm in countries with benefit programs), and pharmaceutical programs have tended to change at different rates and in response to different challenges than the physician/hospital element of the health system. Therefore, an explanation for pharmaceutical variation must extend beyond the existing national health insurance literature, but recognizing the crucial differences between these different components of health policy has the potential to inform and improve theories of both types of policy. Overview of theoretical approach This thesis explains variation in the development of pharmaceutical insurance by a cross- national comparison of policies from the 1940s to the 1970s. The main theoretical contribution is a new insight into how countries select simultaneous versus incremental approaches to policy development, and a demonstration of the implications of this choice for subsequent policy outcomes. The incremental approach to policy development is initially less risky, but it is inherently limiting to the adoption of additional health services, and provides the key to understanding why Canada never developed a nation-wide pharmaceutical program. The thesis provides detailed mechanisms by which paths of health policy development become locked in to a particular trajectory in terms of the services offered, answering criticisms of some studies of path dependence that merely assert that “history matters.” By demonstrating the theoretical leverage gained by applying this approach to pharmaceutical policy, I provide a starting point for generalizing to different countries (to better capture possible effects from different institutional arrangements) and to different policy areas in future research. The implications of simultaneous versus incremental reform I examine countries’ earliest health policy decisions and find that in each case, the national government must choose whether to adopt a wide range of health policies simultaneously, or proceed with health policy development in stages, one service at a time. This choice is a major determinant of whether or not the country adopts a pharmaceutical program, so it is necessary to  9 explain how countries select once approach over the other. The end of World War II constitutes a widely acknowledged critical juncture for welfare state development throughout the industrialized world (see for example Hacker 1998, 81; Lynch 2006, 56). It was a moment of social solidarity in many countries – for instance, Titmuss argues that “the circumstances of the war created an unprecedented sense of social solidarity among the British people, which made them willing to accept a great increase of egalitarian policies and collective state intervention” (quoted in Thane 1996, 210). 1945 was also the beginning of what some analysts have called “the ‘Golden Age’ of postwar capitalism,” (Huber and Stephens 2001, 113) and a time when governments were concerned with providing for returning soldiers and addressing deficiencies in social provision that had started to become problematic in the interwar years and before. My hypotheses concern the types of critical moments for change that occur under different circumstances: I propose that in order to achieve (or indeed attempt) a more radical, simultaneous approach to health policy development, policymakers require the right mix of institutions, principled policy ideas and pressures from public expectations. When one or more of these factors are absent, countries default to a less risky and less politically costly incremental approach to policy development. However, this approach sets the country’s health system on a path that is constrained to a limited range of services, as barriers to the introduction of new services increase over time. Choice of priorities and barriers to later stages If countries take an incremental approach to health policy development (as did both Canada and Australia), there is a second critical decision that affects the adoption of pharmaceutical programs: what service or services are the first priorities for development? Like the choice between simultaneous and incremental approaches to policy development, the order of priorities is affected by institutional and electoral factors. Although it is also affected some factors that are idiosyncratic or external to the theoretical framework, I take Bennett and Elman’s (2006) advice to carefully distinguish between unpredictable or exogenous and explainable factors, and show how these factors interact to affect the outcome. The main determinants of initial priorities may also set up barriers to the adoption of later priorities, as deferred services’ low place on the policy agenda becomes self-reinforcing and realistic perceptions of resource constraints, the development of alternative institutional arrangements, and opportunities for opposing interests to organize more fully may all serve to make the introduction of “later stages” of health policy development less likely over time.  10 Early challenges of pharmaceutical programs The final element of the theoretical framework concerns the way that policymakers respond to the challenge of implementing a universal and comprehensive program in the face of finite government resources. Although this thesis is focused primarily on the politics of program adoption rather than implementation, both the UK and Australia adjusted their pharmaceutical programs shortly after they were adopted, in response to unexpected levels of usage and expenditure. This is significant because measures to control costs potentially affect programs’ scope of coverage, and I hypothesize that these initial responses to the early challenges of pharmaceutical programs are conditioned by the circumstances of program adoption, particularly by the way elite ideas and public expectations about policy problems and solutions become entrenched over time. Hypotheses The remainder of the chapter develops hypotheses to explain the three closely related analytical problems laid out in the theoretical framework: the choice of approach to policy development, the order of priorities for policy development, and the downstream consequences of these choices. To do this, I draw on rich array of literature concerning the effect of institutions, ideas about policy, and pressures from public opinion and politicians’ sense of what makes a policy an electoral winner. There are also important insights from historical institutionalist literature concerning the role of sequence and timing in politics. The chapter concludes with a summary of the hypotheses, organized by analytical problem rather than related literature. Incrementalism and critical moments for change Incrementalism, or policy making in a series of small steps, has long interested both scholars and policymakers. Perhaps the most well-known discussion of incremental policy making is Lindblom’s 1959 article in Public Administration Review, “The ‘Science’ of Muddling Through.” In a 1979 update to the original, Lindblom argues that incrementalism may not necessarily be the most desirable method of policymaking, but that “neither revolution, nor drastic policy change, nor even carefully planned big steps are ordinarily possible” (Lindblom 1979, 517). He goes on to distinguish between incremental politics and incremental policy analysis (does actual policymaking proceed in small steps, and/or does the analysis of options proceed in small, not necessarily rational steps?). I am most concerned with incremental politics, since in the early days of health policy development, policymakers in both Australia and Canada envisioned specific further steps, but chose not to attempt them all at once. Lindblom notes that the “dispersion of veto powers  11 throughout the political system” (1979, 520) can produce incremental politics: this hypothesis is discussed further below in my consideration of fragmented institutional authority. According to Lindblom, we should not expect to see radical, simultaneous policy change, such as I suggest occurred with the introduction of the National Health Service in the UK. Furthermore, he seems to suggest than incremental steps towards policy change may be made indefinitely, noting that these policy changes …do not rock the boat, do not stir up the great antagonisms and paralyzing schisms as do proposals for more drastic change (Lindblom 1979, 520).  If this was the case, countries’ initial choice of approach to policy development should not matter much. In some, rare cases, all health services may be adopted simultaneously, but in the more common instance that countries develop health services incrementally, they should still end up in the same place eventually. As is detailed below, this thesis challenges this notion, suggesting instead that incremental change will stall in predictable ways, in health policy development and perhaps other policy areas as well. This means the choice of approach to policy development becomes much more important. So when does the politics of incrementalism win out over the politics of radical, simultaneous policy development? The agenda-setting literature, particularly work that focuses on the role of “windows of opportunity” (Kingdon 1995), provides important lessons about the process of initiating major policy change. There is a range of literature debating how to identify windows of opportunity or critical junctures (or indeed, whether it is possible to do so before the event) (Mahoney 2000; Thelen 1999), but this thesis conceptualizes these critical moments as points in time when a variety of factors come together to create conditions for policy change. Thus, the goal is to identify the moment where a certain set of conditions occurs, not only the particular decision that occurs in that moment. This is an important distinction, because some critical moments may result in non-decisions: many of the conditions are present, but change does not occur. Similarly, the term “window of opportunity” can cause confusion if it implies that policy entrepreneurs are waiting for an opportunity to act or manipulating conditions to produce a window: this scenario is not included in the present framework. I argue that there may be parallel critical moments across countries, in response to exogenous world-historical events. However, the conditions for change during these critical moments may vary slightly between countries, and uncovering these variations can help explain different policy outcomes. One of the most valuable contributions in this regard is Tuohy’s (1999) discussion of health policy reforms in Canada, the US and the UK: she finds that windows of opportunity (her term) are a product of both institutional factors (consolidated authority over health) and political will,  12 which in my hypotheses I have interpreted as a combination of ideational and electoral motivations. Therefore, I hypothesize that there are three factors that tend to support radical policy change during a critical moment: centralized authority as a result of a country’s institutional structure; politicians’ principled ideas about health policy; and politicians’ electoral incentives to take action. This approach to identifying critical moments allows me to address the effect of federalism on social policy (Pierson 1995), and the effect of electoral incentives and different types of ideas on policy development (Hall 1993; Downs 1957). The other analytical problems suggest an examination of historical institutionalist literature. Institutions: centralized authority and preferences of veto players Institutions are an important factor in the choice of policy approach, and may provide a prerequisite for radical change. Since all three countries included in this study have Westminster parliamentary systems, the main source of institutional variation here is federal versus unitary institutions, which may also be understood in terms of the degree of fragmentation or centralization of institutions. Pierson (1995) suggests that federalism influences social policy through three main institutional effects: it changes the policy preferences and strategies of social actors; may make constituent units significant policy actors; and produces dilemmas of shared authority when constitutional jurisdiction overlaps. Although the empirical chapters provide examples of all three of these dynamics in Australia and Canada, and the later two are particularly important in explaining how Canada’s federal institutions provided opportunities to block a radical approach to policy change, the implications of federalism for health policy in the cases studied here are not as Pierson predicts. His theoretical expectations are that multiple jurisdictions will allow an exit option for business, thus allowing business to oppose welfare state expansion; that a dynamic of competitive state building may promote expansion through policy innovation and emulation; and that joint-decision traps may lead to a tendency for lowest- common-denominator policies (Pierson 1995). However, Harrison (2006) and her collaborators find little evidence of a race to the bottom in the Canadian federation, and neither this expectation nor the expectations of exit by business or competitive state building has bearing on the questions of whether countries take a simultaneous or incremental approach to policy development and whether pharmaceutical insurance develops. Instead, it is more helpful to apply Tsebelis’s (1995) insights about veto players with state or provincial governments considered a “collective actor whose agreement is required for policy change” (Tsebelis 1995). If subnational governments can veto unilateral action by the national government on health, it will be more  13 difficult to achieve the coordination necessary for a radical, simultaneous approach to policy development. Thus, I hypothesize that countries with institutions that centralize authority will be more likely to take a simultaneous approach to policy than countries where institutional authority over health is fragmented. It is important to note, however, than this fragmentation may be more or less formal. Even if states or provinces have constitutional authority over health policy, it must be accompanied by fiscal capacity to take action in this policy area in order for subnational governments to act as effective veto players. Treating subnational governments as veto players brings in a role for their motives, since they must determine whether or not to use their veto over health policy. More generally, this implies that we cannot make claims about the effects of institutions in isolation of the policy preferences of actors at key institutional locations. Harrison (1996) has found that the distribution of costs and benefits of a policy is likely to affect intergovernmental politics, and that “federalism has the potential to create a dynamic of competitive credit claiming in some cases and blame avoidance in others” (Harrison 1996, 20). Interpreting these findings in the context of health insurance, subnational governments may conclude that health insurance or benefits provides an opportunity to claim credit with voters, since it is generally a popular policy when people are asked about it directly. However, if health policy is not a top of mind issue for voters, subnational governments may decide that the financial risks and potential for federal control and credit claiming at their expense outweigh these incentives to act, and they will therefore block radical action by the federal government. This does not preclude provincial governments from taking action on health unilaterally, and this is what a number of Canadian provinces did in between 1945 and 1966. Prior unilateral action would also affect their assessments of whether or not to veto federal action, since a perceived federal “take over” of provincial or state health insurance programs would be a lost credit claiming opportunity, but attempting to maintain a program beyond their financial capacity might push subnational to prefer federal contributions. The nature of federalism also places unique constraints on subnational governments as veto players, however. Since national and subnational governments face an overlapping constituency, a sufficiently motivated national government may bypass provinces to appeal to directly to the shared public, and thus effectively make subnational governments an offer they can’t refuse. States’ or provinces’ institutional veto power still exists, but they may make a political decision that blocking a popular federal policy is simply too costly. I also hypothesize that when authority over health is fragmented, decisions about priorities for policy adoption may be influenced by the preferences of institutional veto players.  14 These actors may also contribute barriers to the adoption of other services over time, as they develop alternative institutional arrangements, for example, provincial governments or professional organizations introduce programs to supply services not offered at the national level. Similar dynamics may occur with regards to organized interests if they have preferences over the selection of priorities, and later, have time to entrench their opposition to further policy development (for example, groups are able to mount a more organized campaign against later services). The most important groups are medical professional associations, because the pharmaceutical industry was only beginning to organize political influence during the research period. However, the strength of medical associations did not differ significantly across countries, and although the potential for organized interests to influence priorities is included in the hypotheses and their interventions are described where relevant, they did not have systematic explanatory power. Ideas: principled and programmatic The second key independent variable in explaining simultaneous versus incremental policy development and the later consequences of this choice is presence or absence of a cohesive, principled idea about health services, in this case that they should be universally available and cover a comprehensive range of services. Tuohy (1999) argues that a window of opportunity for major policy change requires that substantial change in health care policy hold a high priority within the broader agenda of those who command the levels of authority…there must be a commitment to policy change on the part of key political actors – a willingness to elevate the issue above the myriad of others that might be accomplished… (Tuohy 1999).  Thus, a radical approach may require that significant policymakers (prime ministers, cabinet ministers, high level bureaucrats) support the adoption of a policy for principled reasons, although this does not preclude strategic motivations, as discussed below. This would mean that the government in power places high priority on the issue of health services and is willing to expend political capital (particularly vis-à-vis interest groups or other levels of government) to adopt the policy. I hypothesize that this type of principled commitment to policy change will only coalesce around a particular type of idea: what other authors have called “programmatic beliefs” (Blyth 2003) or “issue area doctrines,” (Yee 1996) and which proposes a fairly specific, cohesive solution to a high-priority policy problem. Another helpful source is Hall’s (1993) distinction between “the overarching goals that guide policy in a particular field, the techniques or policy instruments used to attain those goals, and the precise settings of these instruments” (Hall 1993,  15 278). I propose that ideas about overarching goals, such as “the government should work towards a social democratic state” may be necessary but are insufficient to develop the principled commitment for radical policy change. Ideas about goals must be accompanied by a second type of idea about policy instruments, such as “the state should directly provide health services to all citizens.” These instrumental ideas may still be fairly high-level; for example, the influential British Beveridge Report on social services did not detail the way health services should be organized, but rather assumed they would be free, universal, and cover a comprehensive or complete range of services. The crucial characteristic of these ideas is that they provide a cohesive response to a policy problem and a specific motivation for radical action. Electoral incentives Of course politicians’ decisions are also driven by strategic considerations, and these electoral incentives are the third main independent variable included in the research. An important literature in political science deals with politicians’ motivations by focusing on their incentives to avoid blame for imposing costs or taking away benefits (Weaver 1986). Weaver assumes “that most policymakers are motivated in large part by their desire to maximize their prospects for reelection,” (1986, 373) but he points out that voters appear to respond more strongly to losses than gains. This means politicians benefit the most, in electoral terms, from avoiding constituents’ blame for unpopular policies, rather than accepting credit for supporting popular policies. However, the existence of a blame-avoidance dynamic does not negate politicians’ motivations to claim credit and provide goods to constituents. If, as Downs argues in his “economic theory of democracy,” “political parties in a democracy formulate policy strictly as a means of getting votes,” (1957, 137) sometimes this will produce incentives to develop distributive policies, such as health insurance or benefits. As noted above, the end of World War II constituted a “welfare moment” in all three countries, when each government considered some changes to social policy. However, I hypothesize that when there is evidence of public awareness of and support for health policy, policymakers will be more likely to support radical action, such as a simultaneous approach to policy development. When this support is lacking, or more difficult to anticipate, politicians may prefer the less risky incremental approach: a more measured policy response is likely to attract less blame should it fail in the public’s eyes. It should also be noted that there may be both electoral and principled reasons for policymakers to support the adoption universal and comprehensive health policies, and if these two factors motivate the same policy response it may not be possible to disentangle them perfectly. However, we should expect that this would often be  16 the case: “good policies” that promote principled commitment in politicians are frequently beneficial to many people, who recognize their potential benefit and in turn provide strategic electoral motivation for politicians to adopt these policies. Wherever possible I distinguish between electoral incentives that result from public pressure or demands for action, and those that occur primarily because of politicians’ sense that a particular policy will win votes, even if the public is not yet actively demanding it. Electoral incentives may also have a less direct, but still important influence on the way policymakers choose priorities when following an incremental process of health policy development, and on the barriers to further policy development that arise over time. Politicians face many competing pressures when making policies: new health programs may be popular once implemented, but they require tradeoffs against other programs or priorities, such as upgrading infrastructure or paying down the deficit. In public discourse, these tradeoffs are often expressed as health policies being “too expensive” or “fiscally irresponsible,” even though there may be some basic agreement that they are otherwise desirable. I hypothesize in the absence of strong positive motivations to act, politicians’ priorities will be influenced by the perceived financial costs of health programs, regardless the accuracy of cost estimates or countervailing electoral benefits. This means politicians will tend to adopt what is seen as a lower cost service first, and that they will often be reluctant to contemplate subsequent services for reasons of cost. Historical variables I expect health policy development to be constrained by the policy decisions that came before, because of the significant set up costs and potential for entrenched interests from both providers and beneficiaries, and so draw insights from the path dependence literature. Path dependence is an approach to political science based in historical institutionalism, which highlights the historically loaded nature of political development, the importance of timing and sequence, and the effect of small, unpredictable factors on long-term outcomes. It draws our attention to the tendency for inertia in social arrangements and alerts us to the possibility that the causes and/or outcomes we are studying may occur over a long period of time (Pierson 2004). This is particularly relevant for hypotheses about the long-term implications of early policy choices: why does the initial choice of priorities matter, and why do incremental approaches to policy development appear to become “stuck” on first priorities?  One of the earlier and more significant contributions to path dependence theory focuses on the role of increasing returns in political life (Pierson 2000), but others are skeptical about  17 whether increasing returns in fact apply widely (Greener 2005; Ross 2007), or contend that a wider variety of mechanisms for path reproduction must be considered (distinguishing among processes that are self-reinforcing, reactive, or even driven by negative externalities) (Mahoney 2000; Page 2006; Mahoney 2007). I chose to focus on path dependence based on increasing returns as the baseline historical explanation because it represents one of the clearest and most explicit statements of how we should expect path dependence theory to work. It should be noted that at least part of the difficulty with the concept appears to be related to the terminology used. The efficiency-related term “increasing returns” may be more applicable to economics than politics (Pierson 2004), and can cause confusion when authors focus on the most narrow conception of returns as financial or economic returns of a policy (Ross 2007). It is also important to note that there is a better match between theory and evidence if we hypothesize that paths may be constrained, but not determined, by these historical dynamics: paths and policies do change, but we should expect that the parameters of change are strongly affected by self- reinforcing mechanisms.  Pierson defines an increasing returns or self-reinforcing process in politics as one where “the probability of further steps along the same path increases with each move down that path. This is because the relative benefits of the current activity compared with other possible options increases over time” (Pierson 2000 , emphasis in original). Pierson’s work discusses a number of mechanisms by which a process may become self-reinforcing; I focus on three. They are large set up or fixed costs, sequencing effects where certain policies “take up political space” by virtue of coming first, and adaptive expectations, where individuals’ aggregated expectations about which technology will win in the end become a self-fulfilling prophecy (Pierson 2004, 72; Pierson 2000, 254)  I hypothesize that these mechanisms will affect an incremental process of health policy development, where over time, certain factors reinforce existing policies and block deviations such as the adoption of additional services. Large set up costs reinforce existing arrangements because health insurance or benefits require a costly and complex administrative and delivery structure, so there are disincentives to reform this structure in a significant way and to add a new set of administrative structures necessary for additional services. Furthermore, if the national government does not act on certain health services early in the development process, alternative institutional arrangements may arise (from either subnational governments or private insurers), and it will become more difficult to displace other actors’ investments in these arrangements the longer they are in place. Pierson defines sequencing effects as the phenomenon where “some  18 actor [or in this case, policy] achieves a position of influence first, and is able to use that position to consolidate its hold on a particular ‘political space’” (Pierson 2004, 72). In health policy, this effect is expected to occur primarily with regards to the cost of programs. Adopting one service at a time presents an opportunity for policymakers to learn about the real costs of health services, which as noted above, acts as a barrier to adopting later services. Observing higher than expected costs for a first service can act as a powerful disincentive to consider additional policy development, as expensive social policies are often in competition with other electorally-desirable goals, such as limiting budget deficits or preventing tax increases, and contributes to a “collective understanding”  (Pierson 2004, 38) about the limits placed on policy development by fiscal concerns. Essentially, the first health service to be adopted takes up fiscal space in ways that are hard for policymakers to predict in advance. Finally, adaptive expectations shape both policymakers’ and the public’s view of what the health system should do. People base their expectations of what will happen in the future on what has happened in the past, so if additional steps in an incremental process receive little attention after the adoption of the first priority or priorities, voters may come to expect a limited health system. Furthermore, these expectations can restrict policymakers’ deliberations if they come to believe that the problem of health services has been “solved” by the adoption of initial priorities. Researchers must identify and describe the mechanisms that promote stability in the particular path that they are studying (Thelen 1999; Pierson 2004). Although the concept of self- reinforcement provides crucial insights into the identification of mechanisms, it must be used with caution. If it becomes stretched to include anything that makes it difficult to change paths, we have not theorized a specific mechanism of path reproduction, but simply found another way of saying “policies are sticky” or costly to change once implemented. We need specific explanations for why it is difficult to change paths. I suggest that these can be developed with regards to health policy by including more explicitly the effects of reciprocal causality between principled policy ideas and electoral strategy over time. I hypothesize that there is a two-way causal relationship between principled and electoral motivations: early elite expressions of health policy ideas influence public expectations for service by increasing awareness of the potential benefits, which in turn promote greater commitment from policymakers. This mechanism is linked to Pierson’s (1994) insights about the way welfare policies develop “supportive constituencies” of beneficiaries to suggest a new way of conceptualizing the role of the public in social policy development. I expect that even before policies are implemented, campaign promises, party platforms, and other public articulation of  19 ideas about health policy affect voters’ expectations about the type of services to which they are entitled. These ideational influences combine with voters’ own experiences with early policies or non-governmental methods of service provision, resulting in public expectations that can act as an important prompt to policy action, and later, as a barrier to certain types of policy change in terms of the range of services adopted or the parameters of existing policies. Ideas may also have implications for the downstream consequences of early policy choices, especially in combination with electoral incentives. I hypothesize that when policymakers take a simultaneous approach to policy development, their commitment to a cohesive idea of health services will limit the scope for later policy changes that affect universality and comprehensiveness. At least some policymakers will be committed to the idea of universal and comprehensive services for principled reasons, and will oppose changes that they perceive as threatening these principles. As is discussed below, the high level of electoral salience that produced strategic reasons to support a particular type of health policy also means the public will have reasonably well-defined expectations about the type of service they should receive. This makes it politically difficult for policymakers to introduce cost control measures that threaten universality or comprehensiveness.  The corollary of this hypothesis is that in countries where health services are adopted in an incremental, negotiated process, without the pressure to act provided by principled ideas or strong electoral incentives, there should be greater flexibility to introduce cost control measures that affect or potentially affect universality and comprehensiveness. Note that this is not necessarily equivalent to the type of flexibility some scholars have in mind when they refer to a process of incremental adjustment, with policymakers learning over time (Hall 1993). Rather, it refers to the fact that there are fewer ideational constraints on policy reform in countries using an incremental approach to health policy development, since services were not adopted under a national consensus about a particular idea of health policy. Summary This research is based on an empirical puzzle: why has Canada, alone among countries with broad public health insurance, failed to develop universal and comprehensive coverage for pharmaceuticals? This question draws our attention to the fact that there is no satisfactory explanation for cross-national variation in pharmaceutical coverage in either the pharmaceutical policy literature or the national health insurance literature, despite the significant implications of this policy area for patient and population health and the cost and structure of other government  20 health programs. Therefore, the research question is put in general terms: under what conditions do countries develop nationally universal and comprehensive pharmaceutical programs?  To answer the research question, I draw on insights from the comparative public policy and path dependence literatures to produce a new theoretical framework based on two different approaches to policy development, and which explains both their origins and implications for later policies. I have drawn on the comparative literature on institutions, ideas, electoral incentives and historical analysis to formulate hypotheses regard the three linked analytical problems presented above: the choice of approach to policy development, the order of priorities for policy development, and the downstream consequences of these choices. For ease of reference, I summarize below the hypotheses that will be tested in the empirical chapters to follow. The first set applies to all three cases, as I explain countries’ choice of either a simultaneous or incremental approach to policy development. The second set concerns the decision about initial priorities that only applies to countries taking an incremental approach (Canada and Australia), and the third set seeks to explain early adjustment to pharmaceutical programs where they were adopted (in Australia and the UK).  Simultaneous versus incremental policy development Hypothesis 1A: A simultaneous approach to policy development is more likely if the national government has centralized authority over health policy. Hypothesis 1B: A simultaneous approach to policy development is more likely if the national government is committed to a cohesive, principled idea about both the goals and instruments of health policy. Hypothesis 1C: A simultaneous approach to policy development is more likely if the national government faces strong electoral incentives to act on health policy.  Order of priorities and barriers to later stages Hypothesis 2A: When taking an incremental approach to policy development, policymakers will choose priorities based on the preferences of institutional veto players and organized interests, or based on their perceptions of potential opposition from these actors. Hypothesis 2B: When taking an incremental approach to policy development, policymakers will choose priorities based on the tradeoff between the electoral advantages of providing health benefits versus other, electorally desirable goals, and later, may use costs as an electoral justification for avoiding further policy development.  21 Hypothesis 2C: In an incremental approach to policy development, the low position of later priorities on the policy agenda will be self-reinforcing, as they receive less attention from policymakers and the public and do not produce principled or strategic motivations for action.  Response to early challenges of pharmaceutical programs Hypothesis 3: A simultaneous approach to policy development will make existing programs less flexible, so it will be more difficult for policymakers to introduce cost control measures that are perceived to impinge on the universality or comprehensiveness of the program. Conversely, an incremental approach to policy development will make existing program more flexible, so policymakers will be able to introduce cost control measures that may impinge on the universality or comprehensiveness of the program relatively easily.   22 2 Method Introduction This research is designed to find an answer to an empirical puzzle and to address a theoretical question concerning sources of cross-national variation in universal-comprehensive pharmaceutical insurance or benefit policies. The nature of the question dictates both the cases selected and the research methods used. This chapter details how countries and time periods are chosen to achieve variation in both the dependent variable and the intervening causal process hypothesized to produce this variation, and how a process tracing approach is used to uncover the course of policy development in the understudied area of pharmaceutical programs. It then provides background on terms and concepts used in the remainder of the thesis, and discusses how the research methods are applied to both the collection and the analysis of data. The chapter concludes with an overview of the observable implications for each hypothesis introduced in chapter one. Defining the scope of the project Countries chosen Since the research is motivated by a puzzling outcome, the main focus when choosing cases is to get interesting variation on the dependent variable (the development of a universal and comprehensive pharmaceutical program) and to avoid confounding factors that obscure the puzzle. Given a set of hypotheses that predicts a series of steps between causal factors and outcomes, it was also important to get variation in terms of the intermediate steps in the causal process, which I do by selecting one example of a simultaneous approach to policy development and two different variations of an incremental approach (i.e. variation in starting points). Canada, Australia and the UK provide this variation in outcomes and causal process. First, variation on the dependent variable: Canada, of course, lacks a universal-comprehensive pharmaceutical program, although pharmaceutical policies have been the subject of national- level discussion periodically since the 1960s (Royal Commission on Health Services 1964; National Forum on Health 1997; Commission on the Future of Health Care in Canada 2002; Senate Standing Committee on Social Affairs, Science & Technology 2002), and there have been various provincial experiments with pharmaceutical programs, at least one of which achieved  23 universal-comprehensive coverage temporarily.5 Both Australia and the UK have universal- comprehensive pharmaceutical programs, although they differ in their degree of universality and comprehensiveness and the way they achieved these policies: the UK adopted its pharmaceutical program along with all other health services, while Australia adopted a standalone pharmaceutical program decades before other services. The value of choosing Australia and the UK rather than cases with a more similar sequence of policy development is that it emphasizes that pharmaceutical programs may be achieved for different reasons – and that these reasons might not initially have much to do with the specific characteristics of pharmaceuticals, although these certainly become a factor in later policy development and reform. Instead, studying two different processes with similar outcomes in terms of this thesis’s dependant variable suggests that the choice between incremental and simultaneous policy development may have a more general application to other aspects of health policy, and the difference in outcomes between the two countries taking an incremental approach (Canada and Australia) points to the fact that the second step in the causal process, the choice of initial priorities, is also a crucial subject for research. The cases chosen offer an advantage in avoiding confounding factors. Canada, Australia and the UK are all liberal welfare states according to Esping-Anderson’s (1990) typology, which means we expect them to have similar underlying beliefs about social provision. Since other potential cases have welfare states that are structured very differently from Canada’s and are based on different beliefs about state provision it is useful to hold this factor constant. It also deepens the policy puzzle: since Canada’s welfare state generally shares so many characteristics with Australia’s, and to a lesser degree the UK’s, the difference in pharmaceutical programs is all the more puzzling. Also, the cases provide limited institutional variation – all three are Westminster parliamentary systems, and the Canadian and Australian systems are in fact modeled on British institutions in a number of important respects – and this is an advantage as it simplifies the institutional hypotheses and allows me to focus on one aspect of institutional differences: centralized or fragmented authority due to federal arrangements. The lacuna in terms of different sources of fragmented authority provides a direction for future research on other countries, such as those with presidential systems.  5. Saskatchewan implemented a universal-comprehensive drug benefit program in 1975, and in 1993, the program was revised to provide only catastrophic coverage, as was more usual in Canadian provinces (Hickey 2004).  24 All three cases focus on policy making at the national level. In Australia, a 1946 constitutional amendment gave the (national) Commonwealth government exclusive jurisdiction over pharmaceutical benefits, so while Australia’s federal institutions affected program implementation, state governments never attempted to develop their own pharmaceutical programs. In Canada, provinces were significant actors in federal health policy decisions, but did not maintain their own pharmaceutical programs at this time. The major provincial expansion of pharmaceutical programs for seniors did not begin until the early 1970s, after the study period for this project. Finally, the health policy making in Britain was synonymous with policy making in the UK during the study period, as it was not until the devolution of responsibility for health services (among many other policy areas) in 1998 that Scotland and Wales began pursuing independent pharmaceutical programs. Time period chosen The study period begins at the end of World War II, when the national governments in all three countries began considering significant changes in their involvement in health policy. Although health planning efforts predate WWII in each country, the end of the war provided an important impetus for actual policy development. This is also the earliest practical time to consider government involvement in pharmaceutical provision or subsidy, since the immediate post-war period marked the beginning of effective pharmaceutical therapies, with the introduction of penicillin and later different varieties of antibiotics (Quirke 2005; Rivett 1998). The study period ends in the mid-1970s, when pharmaceutical programs in UK and Australia, and hospital and medical insurance in Canada, achieved relative stability. At this time, the Canadian government had rejected a final proposal to extend nation-wide health insurance to cover pharmaceuticals, and pharmaceutical programs were subsequently off the national agenda for more than two decades. The UK and Australia adopted pharmaceutical programs in 1948 and 1950, but Australia’s program did not cover a comprehensive range of drugs until 1960. A longer time period in these two countries, extending beyond the initial decision to adopt the program, is useful because first, it captures the point at which Australia’s program became truly universal and comprehensive and second, it allows for an investigation of a secondary research question about the implications of a simultaneous versus incremental approach to policy development for the way these policies are reformed over time. Since the UK and Australia provide cases of pharmaceutical programs adopted under a simultaneous approach and an incremental approach, respectively, and because they both had the same initial policy response to  25 higher-than-expected program expenditures (the introduction of patient charges for prescriptions), it is possible to isolate the causal effect of the initial policy approach in later reforms before additional explanatory variables, such as the pharmaceutical industry, become major players and complicate the study of pharmaceutical program reforms. The disadvantage of the chosen research period is that it risks overstating Canada’s outlier status with regards to its lack of universal-comprehensive coverage. By 1972, provincial governments were beginning to develop programs to provide drug coverage to senior citizens, and some had programs in place for social assistance recipients even earlier (Grootendorst 2002). Furthermore, while Canadian provinces were gradually expanding public pharmaceutical programs, albeit in a patchwork manner that did not compare to its universal and comprehensive medical and hospital insurance, Australia began a slow and subtle reduction in comprehensiveness, as patient charges increased over time and eventually common and inexpensive prescriptions began falling below the copayment amount, meaning they were entirely unsubsidized. As noted in the conclusion, I plan to extend the research program to capture this important variation in future research.   26 Table 2.1: Summary of outcomes and causal factors Research question: approach to policy development Causal factor Canada Australia UK Centralized authority no fiscal authority centralized yes Cohesive idea  no no yes Electoral incentives no no yes Outcome incremental incremental simultaneous  Research question: choice of priorities Causal factor Canada Australia Electoral tradeoffs  yes – inability to control costs yes – lack of medical manpower Institutional veto players and interests   yes - provinces  no Outcome  hospitals, medical first pharmaceuticals first  Research question: Response to early challenges of pharmaceutical programs Causal factor Australia UK Approach to policy development   incremental  simultaneous Intervening causal factor: level of controversy  patient charges less controversial  patient charges very controversial Outcome  charges increase more quickly; fewer exemptions charges nominal; broad exemptions   Table 2.1 presents the outcomes and causal factors in each country across the study time period, in order to summarize the scope of the project and the variation achieved by this selection of cases. The first step in the causal process is the choice of approach to policy development: does the country select an incremental or a simultaneous approach to health policy? The UK had centralized institutional authority over health, a cohesive idea about how health services should work, and strong electoral incentives to take radical action. It took a simultaneous approach and introduced a full range of health services under the NHS, while Canada and Australia, lacking these factors, chose to adopt one health service at a time.  27  The second step in the causal process, for those countries that took an incremental approach to policy development, was the selection of priorities for policy adoption. This early choice was crucial because over time, barriers to the adoption of additional services developed and policies stalled around the first priorities. I hypothesized in chapter one that we could predict the choice of priorities based on the particular constellation of electoral tradeoffs faced by national policymakers, and the preferences of institutional veto players and organized interests. This was the case in Canada, where the choice of priorities was determined by officials’ assessment that the cost of pharmaceuticals would be uncontrollable and thus electorally inadvisable, and by the strong preference of provincial governments to begin with hospital insurance, which was already being developed at the local level. In Australia, although states held a theoretical constitutional veto, they did not exercise it, and the choice of priorities was entirely determined by federal policymakers’ belief that there were insufficient doctors present in Australia during WWII, and that pharmaceuticals were the most pragmatic choice for a policy that could be introduced during the war.  The final step in the causal process, in the third section of the table, introduces the question of how countries respond to early challenges of their pharmaceutical programs, and tests the hypothesis that the initial choice of policy approach affects the flexibility of programs later on. I find that in Australia, an incremental approach is linked to more easily retrenched pharmaceutical benefits, while in the UK, a simultaneous approach contributed to a significantly higher level of controversy over similar reforms, and less retrenchment in the scope of coverage.  Despite the important advantages of this selection of cases, there are also some drawbacks. The most important concern is that the number of cases and the values the explanatory variables take in these cases result in what King, Keohane and Verba call an indeterminate research design, where there are “more inferences than observations” (King, Keohane, and Verba 1994, 120). As can be seen in Table 2, for each of the two main outcomes (incremental versus simultaneous policy development and the first priority for service adoption), the theoretical framework provides two or three explanatory variables, which are observed in two or three different cases. Table 2 also illustrates a potential problem with multicolinearity, a “situation where we can perfectly predict one explanatory variable from one or more of the remaining explanatory variables” (King, Keohane, and Verba 1994, 122), since in both the cases of incremental policy development, the values of the explanatory variables are identical, which should make it difficult to assess their independent effect. For example, Canada and Australia have similar outcomes on the first research question: they both took an incremental approach to  28 policy development. However, both countries lacked a cohesive, principled idea about health services, and clear electoral incentives to take radical action, so it is not possible using a correlational approach to say whether one or both of these explanatory variables were crucial. The present research is designed to solve this problem using a methodological approach that accounts for the influence of closely related causal factors by observing the mechanisms by which they affect the outcomes of interest, and paying attention to the sequence in which these mechanisms work. Methodological approach: process tracing The emphasis placed on the sequence of policy development in the hypotheses means it is particularly important to differentiate between causal effects and causal mechanisms. When we look for causal effects, we are asking how the presence or absence of different explanatory variables affects the dependent variable, for example, whether the absence of a centralized authority over health is linked to an incremental approach to health policy development. Causal mechanisms, on the other hand, describe the processes by which explanatory variables produce causal effects, for example, the ways in which fragmented authority over health prevents a national government from acting unilaterally, and requires often contentious negotiations with subnational governments, increasing the likelihood of a less radical, incremental approach to policy. Both these types of causal factors are important, but, as Bennett and George (1997) point out, different methods are “better adapted” to the study of effects versus mechanisms. The small number of cases in the present research does not allow for direct inferences about causal effects: the explanatory variables and outcomes are correlated as the theoretical framework predicts, but there are insufficient cases relative to the number of explanatory variables to provide convincing evidence for the causal relationships. However, a careful sequencing of decisions and events in a policy process allows for a verification of causal effects, while also on uncovering causal mechanisms. This strategy process tracing is often more fruitful for small-N, qualitative research (Bennett and George 1997; Hall 2006; Hall 2003; George and Bennett 2005), as it allows the researcher to “unpack” the different variables involved in what might otherwise be an indeterminate causal relationship. Process tracing is a method that Hall suggests is particularly suited to dealing with “theories…based on path dependence or strategic interaction” (2003). It “offers the possibility of identifying different causal paths that lead to a similar outcome in different cases,” (George and Bennett 2005, 215) which is important given even an initial scan of British and Australian health  29 policy landscapes suggests different paths to universal pharmaceutical programs. Instead of “assessing the ability of a theory to predict outcomes,” process tracing assesses a theory’s ability to “predict the intervening causal process” implied by the hypothesis (George and Bennett 2005). In other words, process tracing evaluates evidence for a hypothesized set of causal mechanisms, which can be identified by gathering detailed, chronological accounts of the policy process in different cases, and then be analyzed using the theoretical framework proposed in chapter one. For example, in Canada the decision to approach health policy incrementally, one service at a time, was established over a number of years between 1945 and 1950. By collecting health planning reports commissioned by the federal government, records of federal-provincial conferences, departmental memoranda and minutes of cabinet meetings, I am able to reconstruct a process where the initial expectation of federal experts that health services would be introduced simultaneously was sidelined by provincial opposition, resulting in a number of new, more incremental proposals for policy development. The preference for introducing one service at a time was reinforced by continued interventions by provincial governments, as well as dissension about the value of public health insurance from within the federal cabinet, where politicians and even the Prime Minister did not see health policy as a priority for either principled or electoral reasons, and so favoured a cautious, incremental approach. Having the details of these debates about the most desirable and feasible way to approach policy development provides better evidence for my hypotheses than simply observing that fragmented authority and an incremental approach occur together in this case, and indeed, it would be very difficult to measure variables such as the presence of a cohesive, principled idea or electoral motivations to act on health policy without collecting this type of evidence. Process tracing allows a researcher to build specific accounts of how causal relationships work in a given case, and undertaking theoretically guided process tracing in a number of cases increases the power of this method to identify common mechanisms across time and jurisdictions. Process tracing also solves a major challenge of studying the role of ideas, one of the key causal factors in the project’s theoretical framework. Other authors have pointed to the risk in historical institutionalist literature of treating ideas as a residual category (Blyth 1997): anything that cannot be clearly explained by institutional structures or pressure from various actors’ self- interested behavior is attributed to the often amorphous category of ideas. In this project, it is particularly important to be able to distinguish principled ideas and beliefs from policymakers’ strategic, or electorally motivated beliefs or preferences. When a politician supports a particular policy, how can the researcher tell whether it is because she has principled reasons to think it is a  30 worthwhile policy, or because she is responding to public pressure or anticipating an electoral reward for supporting the policy? As discussed in chapter one, there is no reason to believe that these two types of motivations are mutually exclusive, particularly in a policy area like health insurance and benefits that involves the distribution of goods to voters. However, the methodological approach used and type of data collected for this research provides good opportunities for distinguishing between principled ideas and electoral motivations, because archival documents provide records of both public and private policy discussions, and the sequence of different types of ideas can be determined. Statements of principled ideas may be more common in public fora, such as election platforms or manifestos and speeches in parliament. Process tracing allows for a comparison of when these public statements of principled ideas were made, relative to both private statements and the collection of information about electoral incentives, such as the publication of opinion polls. Furthermore, when politicians in discuss policy in more private settings, such as cabinet meetings, departmental memos, or discussions, they have an opportunity to be more frank about electoral motivations. If they still reference principled ideas, there is more evidence for the fact that their support for a policy is genuinely motivated by “good policy” thinking. The UK case provides an example of how this technique can disentangle principled ideas and electoral motivations for supporting a radical approach to policy development. The British Labour party made radical action on social services part of its platform by 1937 (Attlee 1937, 176, 192; Bealy 1970, 17). This predates many of the interwar reports on a broad public health service (Klein 1995, 2). These reports provided a principled idea about how a health service should work, which in turn helped form public opinion, so that when the Beveridge Report was published in 1942 it became, as Jacobs argues, “a lightning rod, serving as a focus and indisputable symbol of existing public attitudes” (1993, 113). This is an example of elite ideas prompting public expectations, and the wartime coalition government responded to these electoral motivations when they agreed on the 1944 White Paper, which set out a plan for a national health service. The principled ideas of politicians doubtless played a role in the White Paper as well, but their influence can be most clearly distinguished by examining the 1946 decision by a new Labour majority government to adopt a more radical version of the NHS than contemplated by the White Paper. In terms of the timing of the decision, both a clear, cohesive idea (that health services should be universal in coverage, comprehensive in scope, and free at the point of use) and electoral motivations (broad support for the Beveridge Report) existed. However, the candour with which Labour politicians supported a radical NHS for principled  31 reasons in cabinet suggests that it was not a purely electoral decision. There would be no electoral incentive to reference these ideas in a setting that was presumed to be permanently out of the public view unless the speaker was committed to them for principled reasons. Thus, principled ideas in the UK appear to have existed before widespread public support and were expressed in forums where electoral motivations were moot. A process tracing approach can make these distinctions by focusing on the sequence and the nature of the evidence it considers. Terms and concepts The previous chapter describes the theoretical framework, including explanatory variables, but it remains to detail the scope of the dependent variable (universal, comprehensive pharmaceutical programs) and some of the more specialized pharmaceutical policy concepts that will be used throughout the thesis. Since the three countries studied use different terms to describe their pharmaceutical programs, it is important to clarify the basic similarities between them. I use “pharmaceutical program” as shorthand for any public provision of prescription drugs, through direct benefits, insurance, or reimbursement mechanisms. In the UK, the public pharmaceutical program is referred to as “prescription services” and is part of the National Health Service; Australia refers to pharmaceutical benefits or the Pharmaceutical Benefits Scheme. Canadian policy debates tend to refer to programs as pharmaceutical or drug insurance, or occasionally “pharmacare.” In all cases, the government ensures the patient pays only a portion, if any, of the price for most drugs at the pharmacy. Pharmaceutical programs vary on a number of dimensions, but the most significant difference between Canada and most other OCED countries is the scope of coverage: who is covered by pharmaceutical programs, and under what conditions and to what extent are they covered. Despite the fact that issues of coverage appear to underlie discussions of Canada’s difference in both descriptive cross-national reports (Jacobzone 2000) and in evaluative policy prescriptions (for example, Flood 2002), the precise nature of this difference is not explicitly described, and the literature lacks a set of agreed-upon measures for scope of coverage. This is a significant omission, because the scope of coverage affects both individuals’ access to medication and the way drug therapy interacts with other parts of the health care system, and because it provides a clear and relatively simple basis for comparison in a complex policy field. I classify the scope of coverage as more or less universal and comprehensive. These terms refer to different concepts – who is covered and how much coverage they have – but in the  32 policies examined in this thesis, they vary together as early universal programs did not greatly limit the extent of coverage they provided. The main concern is to determine the extent to which the pharmaceutical program is designed to serve as the primary provider of prescription drugs for the majority of the population. This is in contrast to programs that seek only to provide benefits to certain segments of the population, to act as an insurer/subsidizer of last resort, or function as a supplement to private insurance schemes. However, determining the scope of coverage at the time when pharmaceutical programs were first adopted is fairly straightforward. Policymakers often stated their intentions for the scope of coverage when a new policy was adopted, and initially programs did not tend to distinguish among beneficiaries (they provided the same level of coverage to different populations) and had very simple structures of patient contributions or limitations on the types of drugs that were subsidized, so it is clear what depth of coverage was being provided. For example, the prescription drug coverage in the UK was and still is universal and completely comprehensive: all residents are eligible for coverage, and virtually all drugs are eligible for subsidy. The Pharmaceutical Benefits Scheme in Australia was initially not comprehensive: it covered all residents, but only paid for a limited list of “life-saving and disease preventing drugs.” Thus, until the list was expanded in 1960, Australia had a less comprehensive scope of coverage than the UK.  Some pharmaceutical policy terms that were briefly defined in chapter one also require greater elaboration. A formulary is a list of drugs that are subsidized under a pharmaceutical insurance or benefit scheme. During the first years of the research period, many pharmaceutical preparations were still compounded to order by chemists, rather than being mass-manufactured, and so a formulary might contain the actual formulae to be used for various preparations, as well as the names of manufactured drugs. This was a major policy issue in Australia, where the Australian branch of the British Medical Association objected to the government’s proposal to create a formulary that was comprehensive in scope, but aimed to provide precise formulae for medicines that were prepared in pharmacies by chemists. Doctors saw this as a constraint on their clinical freedom and judgment, since they would not be able to prescribe individualized compounds when they saw fit.6 It should also be noted that not all pharmaceutical programs operate with a formulary: in the UK, for example, all drugs approved for marketing are assumed to be eligible for subsidy, although in 1984 a small “Selected List” was designed to function as a  6. BMA pamphlet to hospital patients, 1945. National Archives of Australia (hereafter cited as NAA). A571 1943/1812 Part 1.  33 negative formulary, delimiting those drugs that were not eligible for subsidy under NHS prescription services. Patient charges or copayments are another important pharmaceutical policy tool discussed in this thesis. They constitute a portion of the cost of the subsidized drug that is charged to the patient, and are significant because at higher levels, they may lessen the comprehensiveness of a public pharmaceutical program’s coverage, if patients are required to pay for a large portion of their drug costs out of pocket or using private insurance. Although these charges may be determined in many different ways, from a percentage of the cost of the drug to a tiered level of payments based on the drug’s therapeutic class, during the research period both Australia and the UK used a flat charge, so patients contributed the same amount for every prescription. Copayment schemes may also differentiate among beneficiaries, with lower rates for seniors, low- income persons, or patients with certain chronic conditions. Although the UK introduced a series of exemptions from charges (based on income) during the study period, Australia did not introduce its concessional category of beneficiaries until 1983. Methods for collecting data Preparation: secondary literature and published primary documents The first stage of research was a search for literature related to the historical development of each county’s health system. There was generally limited literature on the history of pharmaceutical programs, as distinct from the broader health system, although the separation of Australia’s PBS from public health insurance meant that there was some literature here (Sloan 1995; Hunter 1965; Hunter 1964). This literature search allowed for the development of a preliminary timeline for policy in each country, listing the start (and end, if applicable) of various policies related to pharmaceuticals, as well as potentially significant related events, such as the publication of major reports on the health system. The preliminary timeline was made as detailed as possible, with the expectation that various policy events or issues would drop off when further research determined they were not significant. In addition to secondary literature, I conducted a search for published primary documents, including reports published by government agencies, parliamentary committees, NGOs and professional associations. I also collected political materials, primarily political party manifestos or election platforms and published Cabinet diaries produced by politicians who held ministerial positions in health or finance during times of major policy development. These materials allowed  34 me to add detail to the preliminary timelines and to begin to identify important themes in different periods of policy development in each country.  Public opinion polls were another important type of published document used in this research. During the study period, the Gallup Poll operated in all three countries, in conjunction with a national polling organization (the Canadian, British and Australian Institutes of Public Opinion). In the UK, the results of relevant polls were complied from a secondary source (Jacobs 1993) which reports extensively on those Gallup polls that reference health policy or provide evidence about the salience of health to voters. For Canada and Australia, I used a Gallup Poll news service, published about twice a week, which reported on the results of current polls. I examined the annual indices for these polls and included any polls referencing health or pharmaceutical policy, as well as polls that asked a “most important problem” question (such as, “what do you think is the most important task facing the government after the election?”). This sample of polls provided evidence about both the popularity of health insurance or benefit schemes when voters were questioned about them directly, and how important health was compared to other issues in voters’ minds at different points in time. Fieldwork: archival sources and elite interviews Archival sources were an important component of the research, since previous historical research on each country’s health system had not tended to focus on pharmaceutical policy. I accessed primary documents related to the creation of health insurance schemes in all three countries starting at the end of World War II, as well as materials related to various policy changes up to the mid- to late-1970. My selection of materials was guided by the preliminary timelines I developed, as well as the initial identification of themes from published primary documents. The main goals of archival research for this project were to verify and refine the chronological narrative of policy development in Canada, Australia and the UK and to uncover the rationale behind past policy choices. This means that when the research identified a significant step in the sequence of policy development, such as the selection of a priority service during an incremental approach to policy development, I searched for additional information on why this step was taken and what alternative options were considered and discarded. My strategy for selecting documents included searching for materials related to events in my preliminary timelines, and then systematically searching the records of relevant government agencies, departments, and commissioned external research bodies (such as royal commissions). In many  35 cases, I was able to access more than one account of a given event (for example, the introduction of patient contributions in the UK from the perspective of both the Department of Health and the Treasury). I also paid particular attention to major policy documents as they underwent revision (for example, multiple copies of a Canadian Department of Health and Welfare proposal regarding pharmaceutical insurance, from first draft to Cabinet Memorandum). This strategy allows me to draw conclusions about policy rationales with a reasonably high degree of confidence. It follows Trachtenberg’s suggestions for uncovering what key policy makers were thinking: starting at the top (in my cases, with the Prime Minister’s Office and central agencies) and working your way down, and checking what policy makers say “in a wide variety of contexts, over a considerable period of time” (Trachtenberg 2006). As noted above, comparing internal departmental memos with correspondence to other departments and central agencies, and with public documents or published primary material helps distinguish between principled and strategic motivations for action, and gives a sense of what actors believed to be key policy problems and valid solutions. Although Booth and Glynn caution that when dealing with government archives, researchers must be aware that they are sometime “incomplete, and possibly misleading,” and that there is an inevitable “self-justificatory element in Cabinet, departmental, and other political papers, including memoirs” (Booth and Glynn 1979, 315), it is unlikely that a wide variety of sources with different interests and goals will omit the same information or provide similarly biased accounts, so the broad scope of archival research acts as a self-correcting mechanism.  The research also included a two semi-structured interviews with a former policymaker and academic whose experience with health or pharmaceutical policy extended to the study period. Understandably, there were few available interviewees for the time period studied, but the insights from these two interviews were nonetheless very helpful and are included in the empirical chapters. Methods for analyzing data Once archival and interview field notes were complied, the first method of analysis was to refine each country’s preliminary timeline (supplemented by secondary and published primary sources). This allowed me to drop events or resources that did not appear to have influenced policy, highlight events that were consistently identified as crucial, and add detail to significant processes. Potential turning points were identified based on the responses of interviewees and my assessment of the archival record (for example, events that generated a high volume of policy  36 discussion in more than one department or agency were good candidates for turning points). I wrote summary narratives for each country to complement the timelines and begin drawing comparisons of sequence across cases. The second step in analysis was a process of “open coding” of all documentary notes. Emerson et al (1995) recommend that the researcher initially code field notes without pre- established concepts, instead simply generating and naming concepts and themes. This allowed for the identification reoccurring policy issues and debates in each country. For example, what were consistent sources of conflict or concern in health policy planning? How were pharmaceuticals linked to other aspects of health policy? I then created a list of questions based on these initial themes and the observable implications of my hypotheses, detailed below. These questions were used as the basis for a second round of more focused coding, where data related to each question was extracted and sorted into a table, providing an overview of evidence for each observable implication or theme from a variety of sources. For example, hypothesis 1A is that centralized authority is required for a simultaneous approach to policy development. The observable implications of this hypothesis are that the national government has constitutional and fiscal authority over health policy, and that it does not consult with subnational governments on its approach to policy. For each country, I tabulated constitutional references to health (where they existed), degree of fiscal autonomy of subnational governments, and any references in archival documents or published accounts to national policymakers considering subnational perspectives and preferences when deliberating on policy approaches, and especially any instances of national policymakers modifying their plans or proposals in response to subnational input. Analyzing a variety of data related to a single hypothesis allowed for more nuanced and complete causal inferences. In the example of hypothesis 1A, this method demonstrated that although Australian states had formal constitutional authority over all aspects of health except quarantine before 1946, their capacity to develop new programs was greatly constrained by their fiscal subordination to the Commonwealth government, and there is no evidence that Commonwealth policymakers consulted state governments or considered opposition from them when choosing their approach to health policy development. Therefore, I was able to classify Australia as having relatively centralized authority over health, despite the first appearances of its federal institutions.  37 Observable implications: linking theory to evidence  Simultaneous versus incremental policy development Hypothesis 1A (centralized authority):  It should be noted that there are a variety of ways authority may be fragmented: a separation of powers system, such as the United States, is characterized by fragmented policy making authority, and a coalition or minority government in some parliamentary systems may also fragment authority. However, the current project is focused on Westminster parliamentary systems, which are known for their centralization of authority in the executive (Moe and Caldwell 1994). Thus, my measures of centralized authority are focused on the federal or unity characteristics of each country. I determine whether subnational governments have constitutional and fiscal authority over health policy by examining constitutional documents, intergovernmental fiscal arrangements, and secondary literature concerning the balance of power between levels of government. Hypothesis 1B (principled ideas): I measure the presence of cohesive, specific ideas using a variety of actors’ ideational statements. If there is a high degree of principled commitment to the idea of universal and comprehensive health services, it will be the subject of public rhetoric (in election platforms and speeches and speeches to the legislature), but will also be present in more private discussions (archival records of cabinet meetings, internal memoranda and correspondence). I determine whether health reform is a high priority by looking for specific mentions of health (as opposed to social policy in general) in political parties’ platforms and election materials. Additionally, I look for sources of cohesive ideas in major health reports or commissions of the time, evaluating the degree to which they appear in the policy discussions noted above, and the variety of rival reports or commissions that existed at the same time. These reports can provide a public and influential statement of principled ideas, and may be more influential in the absence of competing expert viewpoints. Hypothesis 1C (electoral motivations): I separate out strategic and principled motivations by collecting available information about public opinion and issue salience from polls and newspaper articles, and documenting the timing of “good policy motives” such as the commissioning of reports and internal discussion of health reform in comparison to these strategic pressures. If politicians are committed to a health policy idea for strategic reasons, I expect that health insurance or benefits will rank highly when citizens are polled on the “most important problem” facing the country or their family; that major newspapers will include positive coverage of potential health policies and/or coverage focused on the problems caused by the lack of health insurance or benefits; and that politicians will reflect on the strategic benefits of  38 supporting a policy in private deliberations, such as cabinet discussions and departmental memoranda. Order of priorities and barriers to later stages Hypothesis 2A (preferences of veto players): If institutional veto players and organized interests affect the choice of initial priorities, I expect to find evidence of consultation with other levels of government and/or professional groups, in intergovernmental conferences and archival records of meetings and correspondence between the cabinet minister(s) involved and group leadership. I also expect to find private statements of national policymakers’ expectations about opposition from other levels of government or professional groups: even if these actors do not launch a concerted campaign in favour of one priority or in opposition to another, policymakers’ strategic calculations about the political capital necessary to implement different priorities will be affected by their expectations of other actors’ preferences. If other levels of government and professional groups present a barrier to the adoption of later priorities, I expect to observe the same indicators, and also to find evidence that these actors have developed alternative arrangements to fill the gaps in national health policy. For example, if the second stage to be implemented is medical insurance, provincial governments and medical associations may have developed limited provincial medical plans, or expanded the scope of physician-sponsored private insurance, as was the case in Canada. This would provide incentives to block the adoption of later stages at the national level, as a national program would compete with or supersede these alternative arrangements. Of course, it is possible that institutional veto players and organized interests would share the national government’s goals and priorities for health services, in which case I expect to find evidence of consultation but not conflict or adjustment of national priorities. If these actors do not present a barrier to the introduction of later priorities, I expect to observe that they have absented themselves from the health policy field and failed to develop alternative institutional arrangements. Hypothesis 2B (electoral tradeoffs): If a consideration of the electoral tradeoffs of adopting certain services affect the selection of initial priorities for health policy development, I expect discussion to focus on the cost of services versus other desirable goals. I should observe discussion of the costs or resources necessary to implement the program before it is adopted (particularly in private discussions such as cabinet meetings), a high level of involvement from the Treasury or Department of Finance in policy development, and potentially some comparison of the expected  39 costs of different programs. If costs are a barrier to the introduction of later services, I expect to observe many of the same indicators repeated at a later period, in addition to reflections on the actual cost of services already implemented. If costs and electoral tradeoffs present an increasing barrier over time, policymakers will cite their experience with higher-than-expected costs of initial service priorities when they discuss the cost of implementing new services. Hypothesis 2C (self-reinforcing characteristics): I hypothesize that feedback between limited elite discussion and low public expectations will help keep services that are initially lower priorities off the policy agenda, and there are obvious difficulties with observing a hypothesized non-event. However, I expect to observe a difference in the way policymakers discuss later priorities during the selection of initial priorities and in later years, after the adoption of the first priority. Initially, lower priorities should be mentioned as part of a sequence of health service development, but later they may drop out of the discussion altogether, or their possible adoption may be contrasted with a need to maintain or reform existing services. I expect that media coverage of health policy will focus on existing services rather than “gaps” caused by the non- adoption of lower priorities. It may not be possible to measure public opinion about lower priorities if these services are not explicitly mentioned in public opinion polls, but existing data will be reviewed to confirm that this is the case. Response to early challenges of pharmaceutical programs Hypothesis 3 (limits on policy change with a simultaneous approach and flexibility with incremental policy development): If the original ideas about universal and comprehensive services act as a constraint on the introduction of cost control measures, I expect that both private and public discussions of changes will contrast the need for cost control with the principles on which the policy was founded. This may be particularly evident in forums such as cabinet meetings and interdepartmental correspondence when officials and ministers from Health and Treasury departments debate the issue. I also expect to find newspaper coverage of the proposed policy change, and possibly a negative public reaction measured in opinion polls or letters to members. If incremental health policy development allows for greater flexibility when responding to the unintended consequences of pharmaceutical programs, I expect to observe little or no reference to principles of universality and comprehensiveness when policymakers discuss adjustments to the program; less frequent private debate about the electoral fallout of introducing cost control measures, compared to simultaneous countries; and less intra-party or intra-governmental debate about the desirability of making changes (for example, between cabinet ministers and  40 backbenchers, or ministries of health and finance). I also expect the policy changes to have low salience with the public, so I should observe only limited newspaper coverage, particularly negative coverage, and little or no evidence of a negative public reaction, which makes this hypothesis easily falsifiable. Conclusion King, Keohane and Verba (1994), in a chapter entitled “The Science in Social Science,” note that “science at its best is a social enterprise.” They call for inferences based on empirical information and reached according to a set of rules, uncertain conclusions, and public procedures: “explicit, codified, and public methods to generate and analyze data whose reliability can therefore be assessed” (all emphasis original). Although these authors’ recommendations are mainly directed towards correlational research, their appeal for transparent methods applies to all researchers. In this chapter, I have detailed how the scope of my project relates to the research questions, how my overall research method fits the theoretical basis of the project, and how my specific research procedures attempt to correct for potential biases and gaps in the data, and allow me to reach valid conclusions. The remainder of the thesis will present evidence from the three cases, and demonstrate the theoretical conclusions that result.   41 3 Canada: health insurance “capable of being introduced by several stages”7 Introduction Canada’s lack of public coverage for pharmaceuticals is unexpected, given its universal, nation-wide public hospital and medical insurance, and the experience of all other countries with similarly broad public health systems. This chapter argues that Canada’s lack of a universal, nation-wide pharmaceutical insurance program is a result of first, the decision in the late 1940s to develop health insurance incrementally, and second, the decision in the 1950s to relegate pharmaceutical insurance to a low place on the agenda for health policy development. In the immediate postwar period, Canada lacked the conditions that make a radical, simultaneous approach to health policy more likely (centralized authority, cohesive, principled ideas and strong electoral motivation), and the choice of incremental policy development proved limiting. After the first priority for health policy, hospital insurance, was in place, barriers to the adoption of additional services increased. Pharmaceuticals’ low place on the policy agenda became self- reinforcing, as they did not benefit from the combination of politicians’ principled, ideational commitment and public support that my theory predicts is necessary for later policy adoption, and they remain an important gap in Canada’s public provision for health services today. This chapter explains Canada’s decision to attempt health insurance in stages and investigates how priorities for policy development were determined and why barriers to the adoption of later services (particularly pharmaceuticals) increased over time. It then explains the origins of Canada’s separate path of pharmaceutical policy, with its focus on tools to control drug prices starting in the late 1950s. The final section highlights a policy event that has not been the subject of previous scholarly investigation, the 1972 federal proposal to develop a national pharmaceutical insurance scheme, and demonstrates how the failure of this proposal is linked to both the longstanding barriers to the adoption of later stages of health policy that arise in an incremental process, and to the narrow conception of drug price control arising from earlier pharmaceutical management policies. Historical overview Canada’s path to public health insurance was slow and difficult. Although the federal government began developing plans for health insurance towards the end of World War II, and presented its first proposals to the provinces in 1945, Canada did not achieve nation-wide  7. Canada. 1945. Dominion-provincial conference on post-war reconstruction: Plenary session. Ottawa: King’s Printer.  42 hospital insurance until 1957. Medical insurance followed even later, with a federal-provincial agreement in 1966, and gradual provincial implementation between 1966 and 1972. Despite the inclusion of pharmaceutical insurance in the original federal proposal and repeated calls for its development from various bureaucratic and research bodies, this component of health insurance has never been implemented. In 1958, the federal government began contemplating options to control drug prices, which culminated with major changes to the Patent Act in 1969, but this policy was pursued on a surprisingly separate track from health insurance deliberations. In 1972, the federal cabinet rejected a policy proposal that attempted to link a national drug insurance program to better control over drug prices. The first crucial decision: health insurance in stages To explain Canada’s puzzling lack of pharmaceutical coverage, we must turn to its first proposals for health insurance. The idea that health services should be implemented “in stages” was first suggested in the 1945 Green Book (the Health, Welfare and Labour Reference Book for Dominion-Provincial Conference on Reconstruction), which called for a comprehensive national plan but specified it should be “capable of being introduced in any province by several stages” (Canada 1945). The choice to implement health policies incrementally had far-reaching implications for the way Canadian health insurance developed, and is a crucial factor in explaining the lack of pharmaceutical benefits. But why did the federal government decide to proceed this way? Although Canada lacked the relatively centralized, unitary control over health policy present in Britain, the British Beveridge report on universal and comprehensive social services made a significant impression on some Canadian experts (Marsh 1943) and some of the first studies and proposals from the Department of National Health and Welfare (DHW) emphasized the benefits of a program covering a comprehensive range of services (Marsh 1943; Canada. Advisory Committee on Health Insurance 1943). I argue that a lack of commitment to a cohesive idea of universal and comprehensive health services in the federal government, combined with the complications posed by Canada’s federal institutions and the low salience of the issue with voters, made the incremental introduction of health insurance a seeming inevitability to most policymakers. Fragmented authority  Federal institutions potentially allow subnational governments to block a radical approach to health policy, which requires an extraordinary degree of intergovernmental coordination and consensus. In Canada, health care is constitutionally a provincial responsibility. Provincial  43 governments have tended to be protective of their jurisdiction, at least with regards to policies that allow them to claim credit with voters, such as health benefits (Harrison 1996; Cairns 1977). Although federal and provincial governments can and do cooperate on policy initiatives, levels of cooperation ebb and flow, and during the late 1940s they were not high: Touhy refers to “the inauspicious climate of federal-provincial relations that prevailed in the immediate postwar period,” and says this made it “impossible to mobilize the authority to adopt a federal-provincial health insurance plan” (Tuohy 1999, 44). The federal fragmentation of authority in Canada is accentuated, compared to Australia, by the unique position of Quebec and greater limitations on the power of the federal government to spend in areas of provincial jurisdiction. Unlike any Australian states, Quebec is protective of its jurisdiction for reasons of cultural and linguistic distinctiveness. Although the Canadian federal government can and does spend money on programs (such as health) that are not within its constitutional jurisdiction, Smiley argues that its ability to do so was limited by the “unchallenged right” of provincial governments “to participate or not in shared programmes,” that provinces retain “primary administrative responsibilities,” and that at this time the constitution was interpreted so that “the federal Parliament is precluded from contributing to a particular provincial activity from the proceeds of a federal levy made for that purpose” (Smiley 1962, 62). This can be contrasted with the constitutional situation in Australia, where the Commonwealth Parliament was empowered to “grant financial assistance to any State on such terms and conditions as the Parliament thinks fit,” (Australia 2003, Sec 51 xxiiiA). Smiley also points out that conditional grants, where the Canadian federal government provided funds in an area of provincial jurisdiction in return for provincial programs meeting certain federal conditions, were condemned as “an inherently unsatisfactory device” by the 1940 report of the Royal Commission on Dominion-Provincial Relations (the Rowell-Sirois Report) (Smiley 1963, 214). In fact, the exigencies of Canadian federalism appear to have ruled out a simultaneous approach to policy development quite early. The 1945 Green Book envisaged a complete health program provinces would “have to take, in its entirety, and in a fixed order, within a certain time limit,” although health services could be introduced one at a time.8 This proposal failed after being linked to tax rental agreements (where provinces were to give up powers of direct taxation in return for a fixed payment from the federal government) that the provincial governments  8. Department of National Health and Welfare Memo 22 December 1949. Library and Archives Canada (hereafter cited as LAC). RG 29 Vol 1061 File 500-3-4 pt 1: “Health Insurance Studies – Health Insurance Proposals 1949-1950.”  44 would not accept (Maioni 1998; Taylor 1987). In 1946 the Cabinet Working Committee on Health Insurance recommended that “further consideration of the second stage [anything beyond the initial planning grants to provinces] of the Health Insurance Proposals be deferred pending the outcome of the reports from the provinces regarding planning and organization.”9 In 1949, the DHW was asked “to arrange the various features of an over-all Health Insurance program into related parts which might be treated as separate units for introduction at different times.”10 In response, the Department wrote a memo for the federal government’s Interdepartmental Working Committee on Health Insurance. This memo outlined two additional options with more flexibility for provincial governments, who were presumably wary of a major new health program with federal direction and provincial funds. The first option was to propose a program provinces could take up as they liked, and the second was to offer a program starting with one basic service (either general practitioner services or hospital care), with the rest “following in a related order of priority, possibly within a fixed period of time.”11 When health insurance proposals were discussed at the 1955 Federal-Provincial Conference, the Prime Minister’s opening statement demonstrated both deference to provincial governments and acceptance of the principle of incremental policy development: he noted that the federal government would not “wish to be party to a plan for health insurance which would require a constitutional change or federal interference in matters which are essentially of provincial concern,” and solicited provincial input “as to the order of priority of the various services” (Canada 1955).  As described in more detail below, hospitals were eventually selected as the first priority for policy development, with medical insurance to cover doctors’ visits outside of hospitals as a close second. However, provincial preferences for a staged approach to health insurance persisted even after nation-wide hospital insurance was adopted in 1957. At a federal-provincial conference in 1964, British Columbia recommended that “a National Health Services Programme be designed in such a way so as to permit the step-by-step implementation, and that each province should be free to introduce phases of the programme as it sees fit.”12 A survey of  9. Meeting of the Cabinet Committee on Dominion-Provincial Relations: Report of the Working Committee on Health Insurance, 4 January 1946. LAC. Brooke Claxton fonds. MG 32-B5  Vol 138 File: “Dom-Prov Conferences DP-2 Committees and Meetings.” 10. Minutes of the Second Meeting of the Interdepartmental Working Committee on Health Insurance, 9 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt 1. 11. DHW Memo 22 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt 1 12. Statement of Principle advocated by British Columbia Respecting Preliminary Observations of the Report of the Royal Commission on Health Services, volume I, prepared for meeting 20-21 July  45 provincial departments of health in 1965 noted that “[t]he majority of the provinces stated that their ultimate aim was complete health services available to everyone, but stressed the importance of priorities, phasing and timing, which would depend to a considerable extent on federal financial participation.”13 Although the institutional barriers to simultaneous policy development were considerable, it is possible that a high level of commitment to a principled idea about health services, or a high degree of electoral motivation on the part of the federal government could have overcome them. If federal policymakers saw the adoption of a radical new health program as integral to the government’s platform or the dictates of “good policy making,” or there was sufficient public attention to and support for the idea of universal comprehensive health services, the federal government would be motivated to spend the necessary financial and political capital to placate provincial governments. It is typically true that every veto player has its price, as can be seen later on with the federal government’s concessions to provinces in order to achieve nation-wide medical insurance, but some combination of principled ideas and strong electoral incentives is required before policymakers are willing to pay that price. Lack of principled ideas about health Post-war Britain provides an example of the extraordinary level of principled ideational commitment that is necessary to achieve the simultaneous implementation of a broad range of health services, but ideas about health care were much more divided in Canada. To understand the lack of principled ideas in Canada, we must consider the impetus for the first attempts to introduce health insurance. In Britain, Labour’s post-war victory resulted in the party and its ideology of social reform winning its first majority government. In Canada, the Liberal party under Prime Minister Mackenzie King had been in power for almost a decade, and had included health insurance in its platform since 1919 (Boychuk 2008). However, the main reason that the Liberals promised action on social security, including health insurance, in 1944 was because of electoral pressure from the Co-operative Commonwealth Federation (CCF), a social democratic  1964. LAC. RG 29 Vol 1133 File 504-5-11 pt 1: “Health Insurance – Federal-Provincial Conferences 1964.” 13. Consolidated Report of views expressed by the provinces on health services, 1965. LAC. RG 29 Vol 1133 File 504-5-12: “Health Insurance – Federal-Provincial Conferences 1965.” Note that an interviewee involved in policy making expressed the belief that this report only captured the views of professionals in the health departments at this time, and was out of touch with politicians who were much more wary of health insurance. Author’s interview with Tom Kent, Kingston, 11 February 2008.  46 party that was gaining power at both the provincial and federal level (Hacker 1998; Maioni 1998). Pressure from the CCF meant the Liberals were forced to act on health insurance, but action was a political compromise rather than an ideological imperative for the party, and this favoured the slow, staged introduction of actual policy. The lack of consensus on health policy within the Liberal party is well documented in the memoirs of Paul Martin Sr., who was appointed Minister of National Health and Welfare in December 1946. He discusses his difficulty in getting cabinet to approve public health and hospital improvement grants to provinces after the first federal health insurance proposals failed in 1945, and his concerns that these grants would not lead to an insurance plan as he hoped:  I knew that some of the most powerful voices in cabinet did not share my desire to move quickly towards a national insurance plan. Although the party had proclaimed its support for such a scheme on many occasions, I had my work cut out to keep it fully committed to proceeding towards this objective (Martin 1985).  This contrasts with principled ideas about comprehensive health services evident in the British Labour cabinet at this time, where Health Minister Anuerin Bevan told cabinet, “As I see it, the undertaking to provide all people with all kinds of health care…creates an entirely new situation and calls for something bolder than a mere extension and adaptation of existing services,”14 and Klein notes the “remarkable…shared assumption that the health service should be both free and comprehensive,” both within the Labour government and the Conservative opposition (Klein 1995, 24, emphasis added). In Canada, Martin struggled to get support from Prime Minister Mackenzie King, and from King’s successor, Louis St. Laurent. Beyond concerns about difficult federal-provincial negotiations or expense, it seems that St. Laurent simply did not perceive government health insurance to be a good policy idea. Instead, he favoured the expansion of voluntary insurance through existing private plans (Maioni 1998). The Conservative premier of Ontario, Leslie Frost, pushed for hospital insurance’s inclusion on the 1950 Federal-Provincial Conference agenda, and Martin reports that, “St. Laurent was taken aback. He did not believe in health insurance and was amazed that a Conservative premier would openly confess that he was for it” (Martin 1985). This lack of support, especially at the highest level, made it difficult for the DHW to keep health insurance on the agenda. A 1950 memo from Martin’s deputy expresses the hope that “we can keep this whole matter [of health insurance] a live issue,” and advises preparing health  14 Memorandum: Proposals for a National Health Service, Minister of Health, 13 December 1945. The National Archives (UK) (hereafter cited as TNA). CAB 129/5.  47 insurance materials for the upcoming federal-provincial conference despite St. Laurent’s desire to avoid the problem.15 After the 1953 election, Martin was faced with a cabinet where “most ministers supported voluntary health insurance and opposed government involvement” (Martin 1985). Although there was more support for the idea of broad government-sponsored health insurance in caucus, Martin says, “the division of opinion made it obvious that I would never get a combined hospital and medical plan into operation, so I opted for hospital insurance as the easier route” (Martin 1985). Lack of electoral motivation The lack of cabinet support for a bold health policy is at least partly explained by the low salience of health insurance among the Canadian public at this time. There is a two-way relationship between elite ideas about health policy and public expectations for services: generally, public expectations develop in reaction to some government proposals or debate about policy, but since neither federal nor provincial governments outside Saskatchewan were providing clear proposals, there was little evidence of public enthusiasm and therefore no clear electoral motivation at the national level for the Liberal government to act radically or quickly. In his memoirs, Martin reports the results of a Gallup poll mid-1947, saying that the public wanted more funds for research, hospitals and free clinics, but “National health insurance unfortunately received scant support” (Martin 1985, 45). Martin’s references to them indicate that politicians were cognizant of public opinion reports, so it is helpful to refer to the weekly news service Gallup published on its polls. These reveal high levels of support for a national health plan between 1942 and 1945, and in 1949 when the question was asked again, but this was when respondents were questioned directly about their support for health insurance.16 When Canadians were asked variations of a “most important problem” question (a standard measure of salience that provides information on the unprompted top-of-mind issues) between 1945 and 1951 the top answers were jobs, taxes, prices or price control, housing, or threat of war. Health appeared on the “most important problem” list for the first time in July 1949: 10 percent of Canadians thought “social  15. G.D.W. Cameron to Paul Martin, 27 November 1950. LAC. RG 29 Vol 1061 File 500-3-4 pt. 2. 16. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 8 April 1942; 22 May 1943; 8 April 1944; 13 July 1949. The question was: “If the government should start a national health plan, would you be willing to pay a small part of your, (or your family’s) income every month so that you and your family would receive medical and hospital care whenever you needed it?” Support ranged from 69 percent willing in 1943 to 80 percent in 1944 and 1949. Note that by the mid-1950s Canadians were polled separately on hospital and medical plans, and were more supportive of government hospital insurance. See CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 18 April 1956; 21 April 1956.  48 security; old age pensions; hospitalization; national health plans; etc.” were the “most important job the government will have to deal with” after the election. This was behind housing (19 percent) and “employment problems; avoiding depression” (15 percent).17 Health did not appear among the listed problems again until 1953, when “health and hospitalization” were listed separately from social security measures and ranked fourth in August (3 percent, compared to 26 percent who thought taxation was the top problem) and sixth in November (4 percent, compared to 21 percent who were again most concerned with taxation, reduced taxes, lower budget and economy).18 In addition to the lack of pressing public concern about health insurance demonstrated by the “most important problem” questions, there were other indicators that health was not a top- of-mind issue for the public when early decisions about how to approach health policy were being made. In 1943, when the Beveridge Report was generating astonishingly high levels of public attention and support in the UK (Jacobs 1993, 113), only one in four Canadians could recognize the phrase “the Beveridge Report.”19 Those who could identify it were supportive of the idea of implementing a similar plan in Canada, but the Gallup news service noted that, despite the wide publicity given the report in Canada, some Canadians had a confused idea of the subject matter contained in the Beveridge proposals. This includes the man who, when asked by the Gallup interviewer whether he had heard of the Beveridge plan said: “Yes, but one quart a day is enough for me.”20  This confusion illustrates that the cohesive idea of universal and comprehensive health services contained in the Beveridge Report did not enjoy similar prominence in Canada, and I argue that there was no domestic equivalent until the Report of the Royal Commission on Health Services (the Hall Commission) was published in 1964. Furthermore, even the compromise health policy proposals that appeared in the late 1940s failed to draw the public’s attention. In 1948, only 52 percent of Canadians had “heard or read of the new health services plan recently announced by the government in Ottawa,”21 referring to the National Health Grants, a five-year program providing funds to provinces “for the extension and strengthening of the country’s preventive and  17. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 6 July 1949. 18. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 1 August 1953; 14 November 1953. 19. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 6 February 1943. 20. Ibid. 21. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 24 July 1948.  49 treatment facilities,” including hospital construction.22 Of the 52 percent who had heard of the plan, “about four in ten…were not able to name any specific feature of it,” leading Gallup to conclude that despite the “sweeping” nature of the plan, “there are large sections of the Canadian voting public on which the proposals have made little or no impression.”23 No impression or attention from the public meant there was no clear incentive for the federal government to take radical action, given the opposition from the provincial governments. This meant that health policy would follow a less risky, and ultimately less comprehensive, incremental approach. The second crucial decision: determining the order of priority Once it became clear (as early as 1946) that implementing the various aspects of health insurance simultaneously or in a closely timed series of steps would not be possible, policymakers and politicians turned to the question of the order in which services should be introduced. As theory predicts, there were unpredictable, or at least inexplicable within this framework, elements to the choice of priorities, particularly in how Canadian policymakers viewed the potential costs of health programs and the tradeoffs they posed with other electorally-desirable goals. There was extensive debate in the DHW about the order of priorities, which resulted in an early federal decision to take pharmaceuticals off the table, but the decision to introduce hospital insurance first was influenced by provincial input. The end result was a pragmatic decision based on concerns about the resources needed to implement the program, the need to negotiate with provinces and avoid confrontation with the medical profession, and the need to respond to evolving public expectations about health insurance. Resource constraints: the cost of pharmaceuticals Since the first proposals were developed after World War II, the cost of health services was a constant concern within the DHW, the federal cabinet, and in intergovernmental discussions, despite the fact that Canada’s economy recovered more quickly than anticipated after the war, and the years between 1950 and 1956 were marked by relatively high economic growth (Perry 1989). Klein (1995) notes that in Britain there was a surprising lack of discussion of the financial implications of the NHS during its creation, but in Canada cost estimates were  22. Health Services Organization: Memo on Canada’s National Health Services. LAC. RG 29 Vol 2576 File 181-6-2: “Reference files of the Dominion Council of Health.” 23. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 24 July 1948.  50 prepared for each new proposal, and financial concerns greatly influenced policymakers’ discussions. Cost concerns stalled consideration of pharmaceutical benefits very early. Sometimes electoral tradeoffs are assessed in terms of the overall costs of programs, but this was not the case in this instance: in the late 1940s and early 1950s, pharmaceuticals were clearly less expensive overall than hospital or medical care. Instead, it was policymakers’ conclusions about the rate of cost increases that made a pharmaceutical program undesirable. Drugs were included as part of the “later stages” of the 1945 proposals (Canada 1945), but when the DHW reconsidered the order of priority for services in preparation for the 1950 Federal-Provincial Conference, officials recommended leaving pharmaceuticals off the agenda entirely, because  … All the experience to date indicates that it is almost impossible to control the costs in such services, and, until something has been developed in this connection, we do not feel capable of making any suggestions at all as to cost (emphasis added).24  Why Canadian officials were more pessimistic about the possibility of controlling the costs of pharmaceutical benefits than other health services, or why they were more pessimistic than policymakers in other countries, is not clear. By 1949, higher-than-expected costs of prescription services were becoming an issue in the UK, but Canadian officials did not explicitly cite British experience at this time. 25 Australian policymakers were more concerned about pharmaceutical costs than the British initially, but they focused on designing tools that would allow some measure of cost control. However, Tom Kent, the architect of Liberal health policy in the 1960s, notes that at that time, drugs were seen as more difficult to ration than doctor’s visits, and that it was easier to “want too much” in terms of pharmaceutical products, and it seems likely that this thinking played a role at this earlier juncture as well.26 Whatever the reason for Canadian policymakers’ reluctance to grapple with the financial implications of pharmaceuticals, it was persistent. In 1955, a meeting of federal and provincial  24. Health Insurance brief, 7 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt.1. 25. However, Canadian policymakers were likely aware of the cost overruns in the British NHS. Martin (1985) cites the general secretary of the Canadian Medical Association’s warnings about following the British path of “socialized medicine” around 1948, and both the Department of National Health and Welfare and the CMA undertook studies of the NHS in 1949. See LAC. RG 29 Vol 1111, File 504-2-4-PT2: “Department of Health and Welfare – Health Insurance Studies/Canadian Medical Association.” 26. Author’s interview with Tom Kent, Kingston, 11 February 2008. This insight certainly matches the early concerns of British and Australian policymakers with overprescribing, seen as a key policy problem when pharmaceutical programs were introduced.  51 deputy ministers of health concluded that pharmaceutical benefits were “not considered to be feasible at this stage…except for the necessary drugs which would be provided as part of the in- patient treatment services under a hospital care program.”27 Then, after the first stage of health services (hospital insurance) had been implemented in 1957, another round of priority discussions again dismissed pharmaceuticals. In 1963, the federal Departmental Group to Study Health Insurance suggested “that in view of the difficulties inherent in the control of costs and in light of the availability of drugs provided in hospitals, that pharmaceutical benefits might be excluded from any Canadian medical care program,” although the group also considered options for “acquiring drugs at a reduced rate,” such as bulk purchasing.28 The decision to place a low priority on pharmaceutical insurance is a result of the somewhat inexplicable way cost concerns manifested themselves in the DHW in the late 1940s and 1950s, although as will be discussed below, federal bureaucrats took a different position on drug costs in the early 1970s, a time when the greater availability of new drugs might lead us to expect bureaucrats to be even more wary of costs. The order of priorities that won the day – hospital and diagnostic services, with medical services following after a lag of nine years – may seem “logical,” since these are clearly the larger programs, and particularly in the case of hospital care, are more likely to impose costs the individual is unable to deal with herself. However, it is important to remember that this “logical” sequence was not universally followed in other countries taking an incremental approach, as the Australian case will demonstrate, and also that Canadian policymakers at this time were making an explicit decision to take pharmaceuticals off the agenda, rather than simply arguing other services were comparatively more urgent. As will be demonstrated in the later part of the chapter, this decision had a long term effect on politicians’ ideas about the desirability and feasibility of a pharmaceutical program when the possibility of additional health services was discussed in the early 1970s. Preferences over priorities: provincial veto players The selection of priorities that won out was strongly influenced by the preferences of provincial governments, as predicted in my hypothesis about the role of institutional veto players in an incremental policy process. The conditions the federal government placed on a possible  27. Draft Report to the Chairman of the Preparatory Committee for the Federal-Provincial conference 1955 on a Personal Health Care Program. LAC. RG 29 Vol 1132 File 504-5-6 pt.1: “Health Insurance – Dominion-Provincial Conference – 1955.” 28. Meeting of the Departmental Group to Study Health Insurance, 27 March 1963. LAC. RG 29 Vol 1129 File 504-4-15 pt 1: “Comments on Hall Commission, 1963-1966.”  52 health insurance agreement – that a majority of the provinces representing a majority of the population must reach a tax rental agreement with the federal government prior to any funding for health insurance – meant that Ontario and Quebec had an effective veto over policy development. Saskatchewan’s success with hospital insurance also suggests that smaller provinces could have an influence through a demonstration effect, although this is perhaps not as strong as a veto. Provincial preferences over policy were a key determinant of the order of priority, and these preferences in turn were shaped by the interaction of resource constraints, principled ideas and public expectations about health insurance. Saskatchewan’s smaller population meant it could not single-handedly block a federal- provincial agreement – and indeed, given that the Saskatchewan government preferred more federal involvement rather than a continuation of the status quo, this would not have been a helpful strategy. Saskatchewan’s success as a “first-mover” on heath insurance did influence how the federal government thought about priorities, however. In 1949, the DHW noted that there were already two provincial hospital plans launched in expectation of federal support (Saskatchewan’s compulsory program was implemented in 1946, and British Columbia began a similar program in 1948), and that it would be a shame to let them lapse by not providing stable funding.29 In fact, there is a two-way causal relationship between Saskatchewan’s implementation of hospital insurance and the advent of federal support. Saskatchewan launched its hospital insurance program in the expectation that federal cost-sharing would follow shortly (Taylor 1987; Boychuk 2008). Boychuk argues that “the Saskatchewan government certainly would not have proceeded with the hospital insurance plan had it realized that no federal cost-sharing would be forthcoming for more than a decade” (2008). The province’s decision to begin with hospital insurance was also likely influenced by a commissioned survey of provincial health needs in 1944 (Maioni 1998) as well as concerns about doctor shortages (Taylor 1987), but its expectation that the federal government would begin the staged introduction of health policies with hospital insurance was not unreasonable, given that this was the default position of federal bureaucrats prior to their in-depth consideration of the issue.30 Thus, the provincial government began with hospital insurance at least in part based on  29. Health Insurance brief, 7 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt.1. 30. Draft minutes of the first meeting of the Interdepartmental Working Committee on Health Insurance, 30 November 1949. LAC RG 29 Vol 1061 File 500-3-4 pt. 1. In considering the order of priorities, Health and Welfare officials note that the American Health Association suggested beginning with general practitioner services, but “up to this point consideration had been in terms of introducing hospital services first.”  53 its expectations of federal support, and later, the federal decision about where to begin supporting health insurance was influenced by Saskatchewan’s success in the hospital field. Taylor (1987) and Boychuk (2008) argue that Ontario’s preferences were instrumental in the decision to begin with hospital insurance, and this is supported by Ontario’s veto position in the quest for a federal-provincial agreement on health insurance and tax rental arrangements. The federal government was well aware of Ontario’s preferences: a 1955 memo to Health Minister Martin notes that hospital insurance “was the only practical possibility at this time so far as Ontario was concerned,”31 and this position was repeated forcefully by the Ontario premier in more formal federal-provincial settings (Canada 1955). Boychuk (2008) argues that Ontario’s preferences for hospital insurance over other priorities was shaped by the existence of an extensive system of private hospital insurance in that province in the early 1950s. Instead of crowding out the development of public insurance, as Hacker (1998) suggests, pre-existing private benefits solved potential problems of administration and increased public acceptance of the principle of collective insurance (Boychuk 2008). Taylor suggests that Ontario’s desire to act on hospital insurance was also motivated by the limits of voluntary insurance that had appeared by 1953, when about one-third of the population still “did not have available a satisfactory system of budgeting their unpredictable hospital costs” (Taylor 1987). Furthermore, Taylor notes a change in public expectations with regards to insurance: “in 1945 only a small proportion of the population had any direct experience with [hospital insurance], now almost everyone was aware of it,” but coverage was far from universal, and even those with coverage faced significant additional out-of-pocket costs (Taylor 1987). Certainly, once proposals for hospital insurance began to be discussed publicly, they were very popular with Canadians.32 Ontario’s insistence at the 1955 Federal-Provincial Conference that “at this time, our proposals relate to hospital and diagnostic services only,” appears to have trumped federal deliberations about the costs and benefits of different orders of priority. Instead, it was settled between the federal and provincial governments that “the introduction of a hospital care scheme would cause less disruption of the existing order of things.”33  31. Memo to Minister on Health Insurance and Ontario, 5 August 1955. LAC. RG 29 Vol 1132 File 504-5-6 pt.1. 32 CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 18 April 1956; 21 April 1956. The headlines for these releases were “Big majority like the idea of government hospital plan,” and “Wish for government hospital plan increases.” 33. Memo December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt. 1. Although the choice of priority was the subject of much deliberation and intergovernmental wrangling in the early 1950s, this 1949  54 Barriers to further policy development In all these discussions, medical insurance was the expected next step of policy development. However, the struggle to implement nation-wide medical insurance in the mid 1960s was in many ways more difficult than hospital insurance, demonstrating the way barriers to policy development increased over time. Here, the key issue was the development of alternative institutional arrangements to provide coverage for physicians’ services. Medical insurance faced steeper opposition from provincial governments, who by this time had the opportunity to implement their own health insurance solutions, with the attendant investment in a particular system of administration and delivery that might be challenged if the federal government attempted to “take over.” This made provinces more likely to use their veto power, and meant the federal government was able to put significantly fewer conditions on its funding of medical insurance than it had tied to hospital insurance funding nine years earlier.34 Opposition from organized medicine had also increased over time, as non-profit, physician-sponsored medical insurance schemes expanded their coverage across the country. Although public medical insurance was able to surmount this barrier, as discussed below, doctors’ stronger opposition to medical insurance meant the federal government was reluctant to provoke them with further health insurance proposals, such as the addition of pharmaceutical coverage. I have argued that certain conditions make a country more likely to take a radical, simultaneous approach to health policy development. Perhaps reflecting the more significant change to existing policy that they represent, the adoption of later priorities in countries that took an incremental approach requires similar conditions. Nation-wide medical insurance was implemented when a key ideational and electoral factors came together, and the federal government was motivated to overcome the problem of fragmented authority over health and find an agreement with provincial governments, only some of which supported the program. In the early 1960s, objective financial conditions did not present a serious constraint to further policy development. The economic situation between 1962 and 1966 was particularly good: one economic historian calls these years “almost a golden age for the economy” (Perry 1989), and perennial pressure to extend social policies from the New Democratic Party (the  memo demonstrates hospital insurance was considered a good option by some federal bureaucrats quite early. 34. Memorandum to Cabinet, “Measures to lower the unit cost of prescription drugs including a drug benefit program [Pharmacare- handwritten]”, 2 February 1972. LAC. RG2 Vol 6397 File 120-72: “Canada – Cabinet documents”. This memo notes the greater flexibility of the Medical Care Act compared to the Hospital and Diagnostic Services Act.  55 CCF’s successor in federal politics) gained increased leverage given the Liberal government’s minority position in parliament (Hacker 1998; Maioni 1998; Taylor 1987). Perhaps more importantly, the 1964 Report of the Royal Commission on Health Services, or Hall Commission, provided a cohesive idea of a universal and comprehensive health care system (Hacker 1998; Maioni 1998; Taylor 1987; Boychuk 2008). Like the Beveridge Report in the UK twenty years earlier, this report was both high profile and specific. It motivated government by presenting principled ideas about the value of medical insurance alongside recommendations for the necessary types of policy instruments. There was also an ideological shift within the Liberal party itself, as Prime Minister Louis St. Laurent was replaced by Lester B. Pearson and a cadre of progressive cabinet ministers dedicated to the principle of public insurance (Maioni 1998; Taylor 1987). As Taylor notes, “for the first time, there was genuine enthusiasm in the Prime Minister’s Office,” (Taylor 1987), and in Canada’s executive-dominated policy process, this was a key change. Furthermore, by the 1960s the Liberal government appeared to have greater electoral incentives to adopt this second priority for health insurance. Taylor argues that public expectations had been conditioned by previous insurance policies at the federal and provincial levels to the point that medical insurance was a “natural, normal expectation” (1987). Although at this time private medical insurance was extensive, it did not provide comprehensive financial coverage and there were significant gaps in which populations could obtain coverage (Maioni 1998; Department of National Health and Welfare (Research Division) 1954), which may have prompted greater public interest in government insurance. However, as was the case with hospital insurance, the strongest evidence of public expectations comes in response to direct questions about medical insurance, after federal proposals were made public. Health insurance continued to be popular when people were questioned about it directly: a 1963 poll found that most Canadians had heard of medicare (which has been in the Liberal’s recent election platform) and that of these, “more than 8 in 10…believe that a satisfactory programme between government, patients and doctors could be worked out.”35 Existing public opinion summaries do not show a major change in the salience of health insurance that might motivate government action. Unemployment remained the most important problem in Canadian’s minds when they were asked for an unprompted list of their concerns.36  35. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 27 March 1963. 36. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 12 June 1963; 8 January 1964; 5 January 1966.  56 The fact that health failed to move up on this important measure of salience may be related to the peculiarities of Canadian public opinion at this time: in 1966, Gallup reported that “Canadians seem to have something of a fixation about unemployment. They still name it as our chief problem, despite the fact that the ratio of those out of work is lower than it has been for years.”37 It is also possible that the Liberal party simply decided that health insurance was a more tractable credit-claiming opportunity than further efforts on unemployment, given that successes in that area did not appear to alleviate public concern. Whatever the reason, Liberal campaign materials and internal documents from the early to mid 1960s demonstrate a new level of priority placed on medical insurance as well as a new type of strategic thinking about the payoffs of the issue and how they might be achieved. Liberal pamphlets from the 1962 election promised “Health Care as Needed,” and proposed coverage for medical services, with “other services” such as pharmaceuticals to follow later.38 Other election materials informed voters that they “shouldn’t have to worry about heavy medical costs.”39 Although the Liberals lost that election, they came to power in 1963 and the extension of health insurance remained an important promise. In a campaign memo in July 1965, Tom Kent advised Prime Minister Pearson that,  The theme of the campaign should be what we have done and what we want to do if we have a chance to govern. What we have done means more employment, pensions, flag, etc. – a short list of the best. What we will do means medicare and a few other things in a shorter list.40  At this same time, the federal government was preparing for a federal-provincial conference that would discuss the extension of health insurance. Kent’s comments to Pearson illustrate the degree to which the government anticipated that medical insurance would be an important electoral tool: he said, “we are prepared in effect to appeal to the public over the heads of the Provinces” on this matter,41 thus limiting provincial governments’ effective veto power over the extension of health insurance.  37. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 5 January 1966. 38. Liberal campaign pamphlet, “Health Care as Needed: The Answer,” 1961. LAC. RG 33-78, Vol 51, File: “Conference of Ministers of Health.” 39. Liberal campaign pamphlet, “Take a Stand for Tomorrow: Vote Liberal,” 1961. LAC. RG 33- 78, Vol 51. 40. Kent to Pearson, 29 July 1965 (Personal & Secret). Queen’s University Archives. Thomas Worral Kent Papers. 41. Kent to Pearson, 18 July 1965 (Secret - Federal-Provincial Conference: Strategy). Queen’s University Archives. Thomas Worral Kent Papers.  57 The adoption of medical insurance in Canada nine years after hospital insurance demonstrates that sometimes, the first step in an incremental approach to policy development acts as a primer for public expectations, and a promise of more to come. However, differences in how medical and pharmaceutical insurance were discussed in Canada, as well as the different nature of the services themselves, make it fairly clear when a later priority may be able to clear the barriers to later policy development, and when it is likely to fall off the agenda. Between the first health insurance proposals in 1945 and the adoption of hospital insurance in 1957, hospital and medical insurance were discussed together, whereas pharmaceutical insurance was rarely mentioned after early conclusions about its uncontrollable costs. Therefore, medical insurance and pharmaceutical insurance were not equally plausible as a second step for health policy development. Moreover, there is some evidence that pharmaceuticals were viewed as a good rather than a service, like hospital and medical care, demonstrated by Kent’s comment (cited above) that it was easier to “want too much” in terms of pharmaceutical products.42 Pharmaceutical insurance’s position as a service subject to uncontrollable costs may have uniquely disadvantaged it in Canada, beyond the more general predictions that the adoption of later services will be more difficult. The conditions necessary for the adoption of medical insurance (an ability to bypass institutional fragmentation, principled ideas, and electoral motivation stemming from a high profile policy issue) were not present for pharmaceuticals, and much of the failure of pharmaceutical insurance can be attributed to the legacy of pharmaceuticals’ longstanding place near the bottom of the list of health insurance priorities. Politicians’ ideas about the nature of the problem presented by both health insurance and pharmaceuticals became more restricted over time, and the public’s expectations adapted to the lack of elite ideas on comprehensive health services, influencing opportunities for electoral motivations to develop. The discourse of staged health insurance places limits on public expectations: the top priority or priorities are discussed in public as part of election campaigns, newspaper coverage and parliamentary debate, but later priorities by definition receive less attention, and this results in a cycle of limited discussion, little public attention, and little policy action – feeding back into low levels of salience. Internal DHW documents indicate that pharmaceutical insurance was subject to greater cost concerns than other services (focused on their perceived potential for uncontrollable growth rather than  42. Author’s interview with Tom Kent, Kingston, 11 February 2008.  58 absolute expenditures), and was repeatedly removed from consideration for this reason.43 This low priority position became self-reinforcing: governments’ public debate and deliberation on hospital and medical insurance was crucial to the development of public expectations for services and therefore electoral incentives to act, and there was no chance for this to occur for pharmaceuticals. Prior to government action on health insurance, there was little spontaneous public demand for policy: around 1947, Paul Martin Sr. notes that he “realized that the public demand was a weak leg on which to stand. Canadians supported health initiatives in a general way, but they had no precise idea of the preliminaries or the consequences” (Martin 1985). The situation was very different around 1965, where nearly two decades of government promises, intergovernmental negotiations, and provincial experiments had produced, a “natural, normal expectation,” of government medical insurance (Taylor 1987). In contrast, public expectations with regards to pharmaceutical insurance are unclear. Although newspaper coverage and government debates at this time demonstrate the public was concerned about high drug prices,44 there is limited evidence of whether there was any clear preference for or debate about price control versus public insurance. However, since pharmaceutical insurance was rarely debated outside the closed circuit of the DHW and the federal cabinet, and until the 1970s the public had very limited experience with either private drug insurance or provincial drug benefits, there was little basis for the formation of public expectations about drugs that were analogous to the development of expectations for hospital and medical insurance over time. Drugs were not entirely absent from the public agenda, however. In the late 1950s and throughout the 1960s, Canadian publics and governments became very concerned about the high prices of prescription drugs. It is curious that this issue was not linked to the problem of public drug insurance, but it nonetheless affected the consideration of later proposals for a federal pharmaceutical program.  43. Health Insurance brief, 7 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt.1; Draft Report to the Chairman of the Preparatory Committee for the Federal-Provincial conference 1955 on a Personal Health Care Program. LAC. RG 29 Vol 1132 File 504-5-6 pt.1. 44. Globe & Mail 16 December 1955, p.39; 6 October 1960, p.3; 24 January 1961, p.25; see also Cabinet memo 2 February 1972. LAC. RG 29 Vol 1549 File 1006-5-2: “Office of the Deputy Minister, Memoranda to Cabinet 1969-1972, volume 3”; Cabinet memo 24 November 1972. LAC. RG 29 Vol 1572, File 1016-1-2: “Office of the Deputy Minister of Health and Welfare – Health Programs Branch – Pharmacare (Registry files 1971-1975).”  59 An alternative path of pharmaceutical policy: drug prices and patents Although federal pharmaceutical management policies in Canada have their origins in concerns about monopolistic pricing of drugs rather than a need to control the cost of a benefit program (as was the case in Australia and the UK), they were subject to some similar dynamics. In this case, public expectations and the attendant electoral incentives they produced actually helped ensure the issue’s place on the government’s agenda. Federal investigation of drug prices began in 1958 with an internal report by the Director of Investigation and Research, Combines Investigation Act, called the “Green Book” (not to be confused with the federal government’s 1945 health insurance proposals of the same name), which was prompted by “a large number of complaints that prices of drugs in Canada were exorbitant,” and “informal complaints about the high cost of drugs” (Director of Investigations and Research 1961). Between 1958 and 1969, drug prices were the subject of at least four more government inquiries, both internal and public.45 The problem was identified as drug patents, which produced a monopoly situation and high prices. The solution was a series of changes to patent law and tariffs on drugs, the centerpiece of which was a new allowance for compulsory licensing of imported pharmaceutical products. Compulsory licensing creates a legal requirement for a drug patent holder to let other firms produce its drug before the patent expires, in return for a specified level of royalties from the new generic product. The new compulsory licensing provisions passed in 1969 reduced the patent life from seventeen years to six or seven (and arguably the effective patent life to zero, given that most of that time is pre-market, while the drug is undergoing trials) and introduced competition between branded and generic medicines (Eden 1989). This had a significant impact on drug prices (Gorecki 1981, xii), but an unintended consequence was to restrict policymakers’ views of pharmaceutical policies in a way that made it very difficult for them to consider later proposals for pharmaceutical insurance. Public expectations and the demand for price control The federal government’s decision to introduce policies aimed at controlling drug prices appears to stem from demands for action from interested individuals and the public. Although the Gallup news service releases from the late 1950s and early 1960s do not report questions about pharmaceuticals, complaints about drug prices were provided as the main rationale for the  45. The studies were by the Restrictive Trade Practices Commission (Report Concerning the Manufacture, Distribution and Sale of Drugs, 1963), the Interdepartmental Committee on Drugs in 1964, the Royal Commission on Health Services, (Hall Commission, 1964), and the Special Committee of the House of Commons on Drug Costs and Prices (Harley Commission 1966/1967).  60 first Canadian investigation (Director of Investigations and Research 1961), and there is evidence that the issue enjoyed reasonable salience at this time. Public concern about drug prices was based on high profile international investigations of drug prices and Canadians’ own experience with rising costs at the pharmacy. Concern (both among the public and policymakers) was also sparked by interventions from interested parties such as the Canadian Congress of Labour, and the unpredictable but highly consequential interventions of an individual Canadian generic manufacturer. Although there is little evidence to suggest that the Canadian public expected government drug insurance, they were clearly concerned about drug prices. In 1960, an Ontario parliamentary committee noted that, “for some reason the public has attached great importance to the spread in drug prices.”46 This concern was fueled by a high profile investigation into monopolistic drug pricing in the US by Senator Estes Kefauver, begun shortly after the first Canadian study began inside government in 1958.47 One Canadian report noted that discussion of the profits of manufacturers had been widespread “particularly since the publicity following hearings before a Committee of the United States Senate” (Director of Investigations and Research 1961). Public concern was also likely influenced by the increase in so-called “wonder drugs.” Although sulfonamide drugs and penicillin had been available since the 1930s and mid 1940s, respectively, the rate of discovery was increasing: 1949 saw the introduction of corticosteroids, 1953 the first classes of tranquilizers, and in 1958 the first synthetic, patentable antibiotics to deal with the problem of drug-resistant infections (Director of Investigations and Research 1961; Mann 2004). Canadians were certainly experiencing increases in their out-of-pocket drug costs, although this was likely a combination of increased drug use and higher prices for new products. The DHW found that “each Canadian in 1961 spent nearly twice as much in retail drug stores on prescribed drugs as he spent in 1953,” although in 1961 “payments for prescribed drugs sold by retail drugstores made up about seven per cent of total expenditures on personal health care, as they have for many years” (1963). Finally, public concern was exacerbated by interventions from interested individuals and groups. In 1955, the Globe & Mail newspaper reported that the Canadian Congress of Labour, which played an ongoing role with the CCF in pressuring the Liberal party to act on hospital and  46. Globe & Mail, 6 October 1960, p3. 47. Formal hearings before the Senate Anti-Trust and Monopoly Committee began in 1959, and resulted in the Kefauver-Harris Drug Control Act of 1962.  61 medical insurance (Maioni 1998), had asked the government to look into drug prices. According to the labour group, “drug prices are too high and it appears that someone is gouging the public.” Prime Minister St. Laurent responded that complaints should be directed to the Restrictive Trade Practices Commission, an arm’s length body of the Department of Justice.48 The public statements of Jules Gilbert, the owner of an eponymous generic drug manufacturing company in Toronto, about problems with drug pricing also appear to have been an important impetus to early government investigation of drug prices.49 Gilbert’s campaign to expand the market for his lower-priced generic drugs had a unpredicable effect on policy development, in that theory has difficulty predicting the potential impact of one particularly motivated individual. However, it appears that this effect was significant. In a 1961 profile of Gilbert by Pierre Berton, published in the Toronto Daily Star, Berton argues that If any single man can be said to have focused public attention on high drug prices in Canada, it is Gilbert…For the past two years he had been regularly in the headlines, appearing before commissions and panels, issuing statements, making speeches and (until members of the pharmaceutical industry refused to appear with him) taking part in consumer panels.50  Gilbert’s argument was that “there isn’t a valid drug patent written in Canada,”51 and furthermore, the Food and Drug Directorate disadvantaged generic manufacturers such as himself with overly strict regulation of manufacturing plants and products. In 1966, Gilbert wrote to the chairman of the Interdepartmental Advisory Board on Standards for Pharmaceutical Manufacturers with regards to an unsatisfactory plant inspection and alleged that the Food and Drug Directorate, “under its present method of operation, is designed…to be the basic cause for maintenance of high drug prices and to be the future cause for increasing drug prices” by blocking competition.52 The file on Gilbert held by Library and Archives Canada contains almost a decade’s worth of similar correspondence and news coverage. Although Gilbert’s actions were undoubtedly self-interested – as Berton reports, he was “the first to admit…that this particular  48. Globe and Mail, 16 December 1955, p.39. 49. Lang (1974, 28) suggests this connection, quoting an out-of-print booklet that states “It was partly the publicity caused by Gilbert’s statements to the press and to anyone else who would listen, which forced the Canadian Parliament to initiate its own investigation…” (B. Stephenson, The Sugar Coated Pill, Pagent Productions Ltd, Toronto 1967, 29). Although the Green Book and RTPC studies do not mention Gilbert, the archival evidence discussed below demonstrates that he was a very active contributor to the debate, and a dogged participant in later public commissions related to drug prices. 50. Pierre Berton, “Jules Gilbert’s profitable crusade for cutting drug prices,” Toronto Daily Star, 1 November 1961. LAC. RG 27 File 13-27: “Jules R. Gilbert Limited, Toronto Ontario 1965-1966.” 51. Ibid. 52. Gilbert to Dr. H.S. Showalter re: Plant Inspection Report, 8 July 1966. LAC. RG 27 File 13-27.  62 crusade has also been a profitable one for him”53 – it also alerted the Canadian public and policymakers to the issue of high drug prices in Canada in a way that provided strong electoral motivation for policy action and prompted a particular type of policy response, given Gilbert’s focus on the role of patents in causing high prices. The developing consensus on drug prices The Green Book was submitted to the Restrictive Trade Practices Commission (RTPC) in 1961 with a request that the Commission conduct public hearings and report on the manufacture, distribution and sale of drugs in Canada. This was an important step, because the Green Book findings were not made public until the request for further investigation went to the RTPC.54 The choice of venue for this investigation was also important in shaping the government’s understanding of the problem and the eventual choice of policy solutions. The Prime Minister’s statement in 1955 that the Canadian Congress of Labour should direct its complaints about drug prices to the RTPC was almost certainly an instance of buck-passing, however, any federal response to public expectations for action on drug prices was also constrained by the limits of federal jurisdiction over pharmaceuticals. Major change only occurred where it was possible for the federal government to exercise centralized authority over the policy area. The Food and Drugs Act (1953) and the Narcotics Control Act (1961) were the only pieces of federal legislation that directly mentioned drugs: the former was focused on the safety of drugs, while the later was concerned with trade controls (Department of National Health and Welfare (Research and Statistics Division) 1963).  Although drug prices were the subject of “considerable concern” for both federal and provincial levels of government throughout the 1960s (Gorecki 1981), the federal government’s constitutional jurisdiction over trade and commerce, specifically the Combines Investigation Act (1952), contributed to the fact that initial investigations were conducted in an arena that allowed unilateral federal action. Following the 1961 Green Book, government investigations tended to be more public and to encompass a broader range of federal government departments. The RTPC investigation began holding public hearings in 1961, and in 1962 a Special Parliamentary Committee on Drug Costs and Prices (later known as the Harley Committee) was formed. In late 1964, an Interdepartmental Committee on Drugs was established to provide advice from the Departments of Justice, Finance, and Health and Welfare,  53. Berton, 1 November 1961. LAC. RG 27 File 13-27. 54. Globe & Mail, 4 February 1961, p2. The “Green Book” report was eventually published as Appendix Q to the RTPC’s final report in 1963.  63 among others. Also in 1964, the Hall Commission released its report, which included a recommendation that the federal government contribute 50 percent of the cost of a Prescription Drug Benefit programme to be introduced by the provinces, but focused most of its discussion of pharmaceuticals on a series of recommendations for drug price control, including the removal of federal tax on prescription drugs, the study of a voluntary price restraint program by industry, and changes to patents and tariffs (Royal Commission on Health Services 1964). By 1965, these various reports and investigations had produced a consensus that “Canadian drug prices [were] among the highest in the world,”55 and that drug manufacturers “taking undue advantage of the tariff structure,” along with the current rules around drug patents, were to blame (Director of Investigations and Research 1961; Lang 1974). Lang (1974) argues that this consensus, particularly among bureaucrats, was an important policy driver:  The initial investigations into drug costs in Canada had been begun by the civil service in 1958. By 1965, the same civil servants had developed a self-induced commitment to lower the cost of drugs and to defeat what they considered to be an arrogant industry (Lang 1974).  As part of the strong, centralized executive characteristic of Canada’s parliamentary system, the civil service is an important policy actor. However, as will be demonstrated below with regards to the demise of the 1972 proposal for national pharmaceutical insurance, bureaucratic enthusiasm is not sufficient to introduce a major new program or pass a significant policy change. The difference between the 1972 program and these earlier amendments to the Patent Act was that, in 1969, the bureaucrats were able to get politicians onside, presumably because there were external, electoral incentives for government action on drug prices. Furthermore, the Patent Act amendments were a fundamentally different type of policy than the subsequent proposals for pharmaceutical insurance: they required limited outlay of federal resources, could be enacted without consultation with provincial governments, and although they required confrontation with one major organized interest, the pharmaceutical industry, this actor enjoyed far less influence and prestige than the Canadian Medical Association, or industry groups in other countries like  55. R. B. Bryce to Health Minister Judy LaMarsh, 16 March 1965, “Possible action on drug prices.” LAC. RG 33-78 Vol 51 File: Conference of Ministers of Health and Related Correspondence (Fed- Prov).  64 the UK.56 These lower barriers were thus surmounted by the combination of causal ideas (that patents produced high prices) and electoral motivation to act on drug prices. The end of federal pharmaceutical proposals: the Drug Price Program, 1972 Both the barriers to further policy development presented by cost concerns and anticipated opposition from provincial governments and medical associations, and the lack of principled ideas or electoral motivation to overcome these barriers, are illustrated in a single policy incident: the failure of the last internal federal government proposals for pharmaceutical insurance, presented to cabinet by the DHW between 1971 and 1972. Despite the significant changes to patent laws, concerns about high drug prices persisted (Lang 1974). In 1971, the Minister of Consumer and Corporate Affairs and the Minister of Health proposed a Drug Price Program that would include the extension of medicare (as nation-wide health insurance was known) to cover prescription drugs.57 The proposal called for a national formulary that would list which drugs were covered and their prices, created by an expert group,58 and drug benefits that could be phased in for different age groups starting with seniors, although the proposal’s authors argued in favour of a universal program. It anticipated that the cost of the program would be shared with provinces, and it was suggested that the program might not even require new legislation, if it was possible to adjust some of the provisions of the existing Medical Care Act.59 Lack of principled ideas In a 1971 cabinet meeting to consider this proposal, Prime Minister Pierre Trudeau noted the provinces did not like being “forced into Medicare”, and would “undoubtedly object to the proposed extension of the scheme to drugs.”60 The president of the Treasury added that “provinces should be given time to increase the effectiveness of the present Medicare scheme,  56. Lang (1974) contains a detailed discussion of the relative weakness of the Canadian pharmaceutical industry at this time: it was not a major part of the economy, and according to a federal report, “the great majority [of Canadian-controlled firms]…are small in size, and mostly involved in the production and distribution of proprietary medicines, household remedies and sundry drugs,” and, for all intents and purposes, the Canadian pharmaceutical market was simply an extension of the US market (Department of National Health and Welfare (Research and Statistics Division) 1963). 57. The Drug Price Program, 23 September 1971. LAC. RG 2 Vol 6381 Series A-5-a: “Canada – Cabinet documents.” 58 Report of meeting of the Cabinet Committee on Social Policy, 7 February 1972. LAC. RG2 Volume 6397 File 145-72. 59 Memorandum to Cabinet: Measures to lower the unit cost of prescription drugs including a drug benefit program [Pharmacare- handwritten], 2 February 1972. LAC. RG2 Volume 6397 File 120-72. 60. The Drug Price Program, 23 September 1971. LAC. RG 2 Vol 6381 .  65 before any significant additions were made to it.”61 At this time, cabinet also anticipated opposition from organized interests, saying that the “drug lobby would learn of the interdepartmental studies [of drug insurance], and would react violently against them,” and that the inclusion of prescription drugs in health insurance would “only exacerbate” the medical profession’s dissatisfaction with the scheme.62 Since the proposal never left the confines of cabinet, the validity of these concerns were not tested, but they acted as effective barriers to policy development. Concerns about governments’ capacity to deal with the costs of a pharmaceutical program also posed a barrier to policy development, despite the fact that the economy was still reasonably strong at this point. Perry (1989) notes that “except for the onset of serious inflation, [1967-1973] might well be regarded as a continuation of the previous very prosperous years,” and the rate of GDP growth increased between 1970 and 1973, only beginning to decrease in response to the energy crisis in 1973-1974 (Perry 1989). However, the Prime Minister said he did not wish to extend medicare to drugs “because of the considerable expenditures involved and the difficulty of getting the provinces to pay their share,” although he added that if the minister of health could show “the great majority of provinces wanted and were willing to pay for such service the question might be raised again.”63 Later, the Cabinet Committee on Social Policy noted that in principle it supported “the provision of a prescription Drug Insurance Benefit for Canadians when budgetary conditions permit” (emphasis added).64 However, various ministers thought pharmaceutical insurance should be avoided because “the government’s first priority should be to restore public confidence in its economic policies” (and pharmacare would detract from this priority)65, and that “pharmacare would be the beginning of a very expensive program which would undermine the confidence of the middle-income groups in the government’s ability to control the budget.”66 Cabinet’s focus on the financial barriers to the program, as well as the potential opposition from provinces, the medical profession, and the pharmaceutical industry demonstrate that pharmaceutical insurance was not an issue supported by any specific principled ideas on the  61. Ibid. 62. Ibid. 63. Ibid. 64. Memorandum to Cabinet: “Measures to lower the unit cost of prescription drugs including a drug benefit program [Pharmacare],” 8 February 1972. LAC. RG 2 Vol 6397 File 145-72. 65. “Measures to lower the unit price of drugs including a drug benefit program,” 23 March 1972. LAC. RG 2 Vol 6395 Series A-5-a. 66. “Measures to lower the unit price of drugs including a drug benefit program,” 30 March 1972. LAC. RG 2 Vol 6395 Series A-5-a.  66 part of politicians. This can be further illustrated by contrasting ideas within the DHW, where the key policy proposals were developed, with the ideas expressed by ministers in cabinet discussions of these proposals. The bureaucratic authors of the proposals clearly saw them as a principled policy choice that would not only reduce drug prices, but also fill a gap in the present provision for health services and rationalize the use of existing public services. A draft memo entitled “Some Social Reasons for Pharmacare” argues that “society has come to think of health care as being part of a total system and as a result has recognized that an important segment of the health care system is not presently being covered by an insurance program,” and furthermore, that “[i]t does not make much sense to pay a physician under Medicare to examine and prescribe for his patient if the patient is unable to benefit” because the prescription is unaffordable.67 They recommended that benefits be introduced on a universal basis, as the federal government must be in a monopsony position as a drug purchaser in order have a bargaining advantage with regards to drug prices.68  These ideas about the importance of pharmaceutical insurance, particularly as a way to lower the social cost of pharmaceuticals, contrast with the position of cabinet ministers, who did not even consider the department’s recommendation for a universal program, and seemed most concerned with containing the cost of pharmaceuticals to the government. On the recommendation of the Minister of Health, John Munro, they focused on a “staged program” that would provide drug coverage to the elderly and eventually expand to cover children and other groups.69 The result was drug insurance proposals were not debated as a principled extension of medicare, but rather as one of a number of unrelated options under consideration for assisting elderly Canadians. One minister expressed the opinion that “[i]f anything were to be done for older people, it should be simple and dramatic,” such as “a once-and-for-all increase to $100 for the OAS” (Old Age Security program).70 The issue was deferred, and there is no record of cabinet discussing the pharmacare proposals after March 1972. Thus, earlier discussions of the proposal focused on costs and potential opposition; later discussions focused on the proposal as targeted  67. “DRAFT – Some Social Reasons for Pharmacare” and “Arguments for Pharmacare.” LAC.  RG 29 Vol 1572 File 1016-1-2. 68. Memorandum to Cabinet, 2 February 1972. LAC. RG 29 Vol 1549 File 1006-5-1. 69. “Measures to lower the unit cost of prescription drugs including a drug benefit program,” 23 March 1972. RG 2 Vol 6395 Series A-5-a. This is another unpredictable decision that had a significant impact on the subsequent policy discussion: it is not clear why the Health Minister chose to disregard the intent of his departmental advisors in taking this approach. 70. “Measures to lower the unit price of drugs including a drug benefit program,” 30 March 1972. LAC. RG 2 Vol 6395 Series A-5-a.  67 assistance to seniors. At no time was the proposal debated by cabinet as a measure to lower drug prices and extend universal and comprehensive health insurance, and thus there was no opportunity for the development of principled ideas about the benefits of a federal drug program that might allow it to overcome the barriers to its development. DHW attempts to frame pharmaceutical insurance as a tool for price control failed, and I argue that this failure is a legacy of politicians’ entrenched ideas about the nature of both the drug price and drug insurance problems. As we have seen, a consensus that patents caused high drug prices in Canada had developed over a number of years and through a number of different research efforts. This allowed for strong action in this policy area, but it also made it difficult for politicians to conceptualize the drug price issue in any other way: in cabinet discussions of the 1972 Drug Price Program, the problem of drug prices was not even mentioned by politicians. The efforts of DHW officials to link drug prices to public insurance were unsuccessful, as politicians, including their own Minister, only interpreted the proposals as a potential benefits program that had historically been dismissed for cost reasons. Lack of electoral motivation  As noted above, there is limited evidence with regards to public expectations for drug coverage, but the existing evidence suggests there were fewer opportunities for public expectations to develop. A 1972 cabinet memo arguing for the Drug Price Program notes that federal departments “have received and continue to receive many letters from the public complaining about the high cost of prescription drugs and many requests that a drug insurance program similar to Medicare be made available.”71 However, the same memo goes on to discuss strategies for the implementation of a pharmaceutical program and says that since the federal government is not in a position to act unilaterally, it could “wait…for provincial and public pressures to build up,” or actively encourage these pressures in hopes of igniting a desire for intergovernmental cooperation on the issue.72 This suggests that proponents of pharmaceutical insurance recognized the potential for public opinion to aid policy development, but that the necessary pressure did not yet exist. Furthermore, most provinces did not begin to introduce targeted public drug benefits (for seniors and social assistance recipients) until the early 1970s (Grootendorst 2002), so Canadians’ first experience with public insurance for drugs was both late and limited to a relatively small  71. Memorandum to Cabinet: “Measures to lower the unit cost of prescription drugs including a drug benefit program [Pharmacare],” 8 February 1972. LAC. RG 2 Vol 6397 File 145-72. 72. Ibid.  68 portion of the population. Private insurance was also limited: a 1963 study of prescription drugs in Canada reported that, “insurance against expenditures for prescribed drugs became available in Canada only recently, in a few prototype schemes” (Department of National Health and Welfare (Research and Statistics Division) 1963). Eight years later, the situation remained much the same: a report of the Canadian Pharmaceutical Association noted that private, third-party prescription drug insurance programs “are not believed to cover a significant portion of the population” (Commission on Pharmaceutical Services 1971). Certainly the campaign promises of political parties, and policy agendas of governments, never alluded to pharmaceutical insurance as anything other than a vaguely distant possibility. Thus, although it is possible that the public was beginning to develop expectations about drug insurance based on a perceived “gap” in the now-comprehensive public hospital and medical insurance they enjoyed, there is less evidence for this kind of public pressure than there was for hospital and medical insurance, where governments had a clear sense of having promised insurance, and a need to fulfill those promises for electoral reasons. After 1972, there was a lull in federal efforts towards drug prices or insurance. Compulsory licensing was the key element of federal government pharmaceutical management policies for twenty-five years, until it was repealed as part of the North America Free Trade Agreement Canada signed with the US and Mexico in 1994. Drug insurance was effectively off the agenda until the National Forum on Health recommended universal, first-dollar pharmaceutical insurance in 1997, but the link DHW bureaucrats had attempted to forge between price control and drug insurance in 1972 would become an important theme in this later period of Canadian drug policy, although usually not in the direction they had intended. The mantra “no pharmacare without cost control; no cost control without pharmacare” would become what many actors saw as a false paradox of drug insurance policy, but one that acted as a barrier to policy development for years to come, as policymakers wrestled with the perception that drug insurance would be too expensive to the government, given the lack of control over drug expenditure, but that it was not possible to enact effective brakes on drug prices or to promote truly cost- effective prescribing, which could lower the cost of drugs to society, unless the government was the main purchaser of pharmaceuticals.73  73. Author’s interviews with Judith Maxwell, Ottawa, 22 October 2008, and Abby Hoffman, Ottawa, 16 October 2008.  69 Conclusion This chapter demonstrates that Canada’s lack of a nation-wide pharmaceutical insurance program has its origins in the federal government’s decisions, first, to develop health insurance incrementally, rather than introduce all services simultaneously and second, to make pharmaceutical insurance a low priority from a very early point. The story of Canadian pharmaceutical policy highlights the interaction between institutions and ideas in determining the scope of policy development. In the 1940s and 1950s, the fragmented authority over health produced by the federal system, the lack of commitment to the idea of universal, comprehensive public health insurance, and the low salience of health policy with voters resulted in the federal government choosing a more cautious and incremental approach to health policy. This staged approach, in turn, placed limits on the opportunities to develop additional health services later on. Pharmaceuticals were initially a low priority because of policymakers’ concerns that their costs would be uncontrollable, and hospital insurance was selected as the first priority based on provincial preferences and a desire to avoid confrontation with the medical profession. After the implementation of hospital insurance, a second opportunity for medical insurance opened based on a new Liberal government’s principled ideas and electoral incentives from both the Canadian public and the NDP, which was aided by the fact that hospital and medical insurance were both initially high priorities for policy development. However, after the adoption of these initial priorities, the federal government was very reluctant to consider further extensions to public health insurance, citing the likely opposition from provincial governments and organized interests as well as resource constraints. Because it had always been a low priority for a “later stage,” there were no promises and little public discourse about pharmaceutical insurance, and hence no public pressure that might prompt further policy development. Although concerns about and proposals for public drug insurance were mostly absent in Canada, public concern about drug prices was an important policy driver in the 1960s. Prompted by both individuals’ experiences with rising costs at the pharmacy and an exogenous or unpredictable set of outside influences (American investigations of the problem and strategic interventions from a particularly persistent businessman), a series of government reports reached a consensus about the problem of high drug prices (patents) and the solution (significant new limits to patent life through compulsory licensing). These conclusions were in part a function of where the initial studies took place: the Restrictive Trade Practices Commission represented an area of centralized federal authority over pharmaceutical policy, and was able to produce a relatively low-cost regulatory solution that got politicians onside.  70 Policymakers’ ideas about the nature of pharmaceutical policy become entrenched over time, as we saw with the federal cabinet’s discussion of the 1972 Drug Price Program. The understanding of pharmaceutical insurance as both subject to uncontrollable costs, and quite separate from the issue of patents and drug price control, meant that politicians were not willing or able to discuss that proposal on its own terms, as an attempt to use universal public drug insurance as a tool for price control and fairness. Canada ended the study period with the main components of its public health insurance system in place: there was universal hospital and medical insurance in each province, but pharmaceutical insurance was firmly off the federal government’s agenda. Instead, federal pharmaceutical policy was limited to the regulation of new drugs, and a policy of compulsory licensing that, however imperfect, resulted in low Canadian drug prices, compared to the United States, for the next two decades. The 1970s saw the introduction of both limited public drug programs by provincial governments and a wider variety of private options from non-profit and commercial insurers. Although this did not result in a comparable degree of pharmaceutical coverage to national programs in the UK and Australia, it did influence more recent attempts to develop nation-wide pharmaceutical insurance in Canada, and raises new questions about the prospects of pharmaceutical policy in the present day.   71 4 Australia: “a high grade service” that requires “progress step by step”74 Introduction Like Canada, Australia took a staged approach to the development of health insurance, but the sequence, and hence outcomes, of its policy development were very different. Australia’s first step was to introduce a system of pharmaceutical benefits in 1944, and the federal government intended to follow this with hospital benefits and later comprehensive medical insurance. The implementation of the Pharmaceutical Benefits Scheme (PBS), however, proved unexpectedly difficult, and it did not operate until a new government modified the scheme in 1950. There was no further development of government health insurance in Australia until the 1970s, when a public system of hospital and medical insurance was proposed; it was finally adopted in 1975. Australia today looks similar to the UK, with a universal public health insurance system that covers doctors, hospitals and drugs, although there is a much greater role for private insurance in Australian hospital and medical insurance than in the UK’s National Health Service (private insurance for pharmaceuticals is prohibited in Australia). Up to the mid-1970s, however, health policy in Australia looked much more like Canada, with policy development that stalled after the implementation of a first priority. This chapter explains why Australia chose an incremental approach to health policy development and began with pharmaceutical benefits. It then discusses the difficulties of implementing the PBS, and why policy implementation stalled after the adoption of the first priority. The next section turns to Australia’s response to the early challenges of its Pharmaceutical Benefits Scheme, when it became apparent that the limited list of covered drugs was producing perverse incentives to prescribe expensive drugs and costs increases were becoming a problem. I study the first policy response, which was to impose a charge on prescriptions, in tandem with an expansion of the PBS formulary, and compare this with a similar policy change in the UK, where the government also imposed patient charges to control costs, but did not have the option of expanding the formulary as NHS prescription services always covered a comprehensive range of drugs. The final section presents some preliminary hypotheses about why Australia finally adopted public hospital and medical insurance, more than twenty years after the first step of health insurance, pharmaceutical benefits, was implemented.  74 Treasury Memorandum, January 1944. National Archives of Australia. A571, 1943/4513.  72 Historical overview The Australian Labor Party took power in 1941, after almost nine years of non-Labor government at the federal level. Between 1942 and 1943, the Commonwealth (national) government decided to begin the staged introduction of health benefits before the end of the war, and to start with pharmaceutical benefits. Although the Commonwealth government lacked constitutional jurisdiction over health, it passed the first Pharmaceutical Benefits legislation in 1944 in hopes of gaining greater powers through a constitutional amendment. Later that year, the Commonwealth government held a referendum on such a constitutional amendment, which would have given it a new jurisdiction over health.75 The referendum failed, and this provoked a constitutional challenge of the Pharmaceutical Benefits legislation in the High Court (spearheaded by the medical association). Early in 1946, the PBS was ruled unconstitutional. Later in 1946, there was another, more narrow, constitutional referendum: this time, the Commonwealth government gained powers over health benefits, with some limits. Specifically, the amendment gave the Commonwealth government authority over the provision of maternity allowances, widows’ pensions, child endowment, unemployment, pharmaceutical, sickness and hospital benefits, medical and dental services (but not so as to authorize any form of civil conscription), benefits to students and family allowances (Australia 2003, Sec 51 xxiiiA).  This was understood to cover benefits to individuals, rather than control over hospital administration or the regulation of private medical practice, which remained with the states owing to the phrase preventing “any form of civil conscription,” which was intended to block any possibility of a national service (Mendelson 1999). In 1947, the government passed new Pharmaceutical Benefits legislation, but since doctors refused to participate, the scheme did not operate. Negotiations between the government and the medical association continued until 1949, when part of the legislation was again declared unconstitutional. Shortly thereafter, Labor lost power, and the Liberal Party formed a coalition government with the Country Party. In 1950, the Liberal government modified the PBS to make it acceptable to doctors, and the scheme finally began operating. The first major change to the program came in 1960, when the Liberal government expanded the formulary and introduced patient charges for prescriptions in an effort to control the costs of the PBS.  75. The 1944 Referendum on Post-War Reconstruction and Democratic Rights was known as the “14 Powers” referendum. If passed, it would have given the Commonwealth government jurisdiction over topics including prices, monopolies, the rehabilitation of ex-servicemen, national health and pensions for five years after the end of the war, and provided constitutional guarantees of freedom of speech and religion  73 After the successful implementation of the PBS in 1950, public health insurance policy was stalled for more than twenty years – until the next election of a national Labor government, in 1972. The Whitlam Labor government introduced a public system of medical insurance called Medibank along with grants to states that allowed for free treatment in public hospitals, but this program was short-lived. A Liberal government was elected in 1976, and despite a campaign promise to maintain Medibank (Gray 1991, 147; Whitlam 1985, 349), had significantly shrunk the system by 1982. Australia achieved public medical and hospital insurance on a more permanent basis in 1984, after the election of the Hawke Labor government. It should be noted, however, that the new Medicare system retained an important role for private insurance, unlike the Australian PBS, the British NHS and the Canadian hospital and medical insurance system. The first crucial decision: a decisive move to stages Like Canada, Australia’s decision to proceed incrementally was taken quickly, but unlike Canada, there was no lengthy debate about the order of priorities subsequent to the stages decision. In fact, the decision to begin with pharmaceutical benefits appears to have been taken at about the same time as the decision to proceed in stages, between 1942 and 1943. Given that Australia is a federation and, until 1946, state governments had constitutional authority over all aspects of health save quarantine, theory predicts that fragmented authority should have been a major barrier to a radical approach to health policy in Australia. However, as detailed below, there is compelling evidence that federalism did not actually act as a barrier in Australia, suggesting that the analyst must be alert to both formal constitutional divisions of authority and functional financial capacity, as well as contextual factors, such as the way Quebec’s unique place in the Canadian federation emphasized fragmented authority. Furthermore, Australia demonstrates that in this case, it is more helpful to look for evidence of the mechanisms by which fragmented authority affects the approach to policy development (such as the inclusion of subnational governments in policy deliberations) rather than simply identifying a correlation between the institutions of federalism and the outcome of an incremental approach. I find that the most important barriers to the simultaneous development of health insurance in Australia were the lack of a principled, cohesive idea about how health services should function, and the lack of strong electoral incentives to act. This resulted in the government’s concern about resource constraints taking precedence in policy discussions, and meant that the simultaneous introduction of comprehensive health services was not seriously considered. This section explains why the lack of principled ideas and electoral motivation, rather than fragmented  74 authority over health, were the main barriers to simultaneous policy development, and then discusses how the federal Treasury came to the decision to proceed with the incremental adoption of health benefits.  Centralized authority through fiscal means Although Australian states had jurisdiction over most aspects of health, opposition from state governments was not the main barrier to the simultaneous development of health policies in Australia. Labor did not consider federalism a significant barrier to broader social services: they simply planned to request a constitutional amendment to provide them with the necessary jurisdiction, unlike the federal government in Canada which was very conscious of a need to negotiate with provincial governments. Although in constitutional terms the Commonwealth government should have accounted for the needs and preferences of state governments when considering options for health policy, there is a lack of evidence that they did account for states. As will be detailed later in the chapter, federalism became a major factor when the Commonwealth government attempted to implement pharmaceutical benefits, but federal institutions did not have a significant impact on actors’ deliberations or choices when selecting an approach to policy development or a first priority. The unexpectedly minor impact of federalism in Australia’s choice of policy approach is explained mainly by the weak fiscal position of Australian states vis-à-vis the Commonwealth government. In 1942, the federal government took over the income tax field in return for fixed grants that provided a lower level of revenue that states had previously received from their own taxation of income, and unlike Canada, Australian states never regained this tax room, making them highly dependent on federal grants and loans (Matthews and Jay 1972). Although the states retained constitutional jurisdiction over health, they lacked fiscal resources, and were constrained by the Commonwealth government’s “unfettered grants power” (Matthews and Jay 1972) under Section 96 of the Constitution, which states that “until the [Commonwealth] Parliament otherwise provides, the Parliament may grant financial assistance to any State on such terms and conditions as the Parliament thinks fit,” (Australia 2003, Sec 51 xxiiiA). This meant states were not in a strong position to turn down federal assistance health programs. This can be contrasted with the Canadian situation, where provincial governments were more assertive and retained an effective veto over health policy: war time tax rental agreements ended in 1947, and provinces regained some powers of direct taxation (Bélanger 2001) As noted in chapter three, the power of  75 the Canadian federal government to spend in areas of provincial jurisdiction was also more restricted (Smiley 1962, 62), and more likely to be resisted by Quebec. Another element of the Commonwealth’s government lack of concern about securing the cooperation of state governments (relative to the federal government in Canada) was its greater confidence in its ability to achieve a constitutional amendment for powers over a range of social services. Unlike Canada, Australia had a constitutional amending formula at this time, and Commonwealth governments benefited from a more relaxed attitude about constitutional amendment. Livingston finds that  in Australia there is no great reverence for the written constitution…No party need hesitate to alter it simply because it is the constitution; the only concern need be whether sufficient votes can be collected to effect the alteration (Livingston 1956).  Attempting constitutional amendment was simply another policy tool for federal governments in Australia, whereas it was an extraordinary measure for Canadian governments.76 Furthermore, the federal government had already imposed a uniform income tax (taking over the states’ role in the income tax field, as noted above) over the objections of the states, and this major expansion of federal fiscal capacity had been sanctioned by the High Court, which ruled that the enabling legislation was “valid not merely under the defence power from the duration of the war but also under the normal powers of the Commonwealth in times of peace” (Matthews and Jay 1972). Thus, it appears that the Commonwealth government had the institutional authority for a more radical approach to health policy. Why was Australian health policy developed incrementally? Lack of principled ideas The primary barrier to the simultaneous development of hospital, medical and pharmaceutical programs in the early 1940s was the lack of principled ideas about health policy, specifically, in the Australian Labor Party (ALP). I argue that because the government lacked a cohesive and galvanizing idea about how health services should work (such as the Beveridge Report in the UK) and because the ALP had a fairly pragmatic ideology at this time, a radical approach to health policy was never seriously considered. This can be seen in Labor’s manifestos and election speeches, which even after the introduction of pharmaceutical benefits legislation do  76. To date, there have been forty-four attempts to amend the Australian constitution, eight of them successful. In contrast, there were three changes to the division of federal and provincial powers in Canada’s British North America Act, 1867 before the constitution was repatriated in 1982. After this time there have been a number of successful amendments that apply to only one province, but both attempts at major, national constitutional change have failed.  76 not place a high priority on health policy. The 1946 Labor policy speech mentioned the passage of the Pharmaceutical Benefits Act and new hospital grants to states without comment (McAllister and Moore 1991, 27), while the 1949 speech simply alludes to the fact that legislated benefits were not functioning, saying that “The people are now dependent on the good-will of members of … [the British Medical] Association,” who were refusing to comply with the legislation (McAllister and Moore 1991, 37). The decision to turn responsibility for making health policy decisions over to Treasury, where pragmatic financial concerns were paramount, provides further evidence that the Labor government lacked principled ideas about health policy. As is discussed in chapter five, comprehensive health services were an integral part of the British Labour party’s platform because their inclusion was supported by a prominent report (the Beveridge Report) that presented a cohesive idea about the shape health services should take, and prompted a high degree of public support. No such authoritative ideas were influential in Australia. Roskam points out “that Australia did not have anything like the 1942 Beveridge Report on social insurance in the United Kingdom, the basis for creation of the British welfare state, could be cited as evidence of a lack of policy development on both the Coalition and Labor sides of politics in Australia” (2001, 279). He goes on to argue that there was nonetheless a longstanding debate about welfare policy in Australia, and that “Australian governments of both persuasions were a great deal more practical,” than the British Labour government with its embrace of the comprehensive and expensive Beveridge proposals (Roskam 2001, 279). However, I argue that it was precisely this lack of a cohesive idea about the value of a comprehensive program, and the inherent pragmatism of the ALP, that prevented the adoption of a radical, simultaneous approach to health policy development in the 1940s. Instead of a specific focus on health that might be prompted by a particularly salient report, Labor focused on broader social services, and health was only incidental to these plans (Gillespie 1991). The main support for the idea of comprehensive health services came from a body outside government: the National Health and Medical Research Council (NHMRC), an independent forum for consultation between Commonwealth and state health departments and the medical profession. However, this idea was contested, rather than cohesive: the NHMRC’s position was only one part of more than twenty years of debate about the proper approach to health policy at the federal level. Public health benefits do fit Labor’s general “belief that there should be freedom from want for all our people” (Chifley election speech 1946, quoted in McAllister and Moore 1991, 27). However, the Australian Labor party was a less ideological party than its British counterpart,  77 with more in common with the Canadian Liberal Party in terms of a left-centre orientation that was strongly conditioned by practical concerns. In 1943, Labor Prime Minister John Curtin proposed a National Welfare Scheme and mentioned health as a subject which “embraces many items such as medical, hospital and dental services, and children and maternal welfare,” but went on to say that It is impracticable in war-time to devise and introduce a comprehensive scheme for all these services…but investigations are proceeding, and it is anticipated that it will be possible to give effect to some or all of these services either wholly or in part when the inquires have been completed (quoted in Crowley 1973).  This suggests that early social policies, including pharmaceutical benefits, were viewed by the government as part of a larger effort towards a welfare state. This view is reinforced by the statement of the Minister for Labour and National Service, E.J. Holloway on the introduction of the first Pharmaceutical Benefits legislation in 1944: he called the bill “a further installment of the complete plan of social welfare services promised by the Treasurer (Mr. Chifley) last year, when he introduced the National Welfare Fund Bill” (quoted in Crowley 1973, 100). However, even these more principled plans for long term policy development were situated in a basically pragmatic outlook, which acted as a barrier to the adoption of a simultaneous approach to health policy. Commentators have contrasted the Australian Labor Party’s relative lack of ideological content at this time with the vigor and focus of the British Labour party. McMullin says that in 1934, John Curtin “took over the leadership of a party well and truly in the doldrums,” and that the party’s “most glaring weakness – in sharp contrast with the British Labour Party – was the ALP’s intellectual bankruptcy…intellectuals were distrusted as flighty and unreliable” (McMullin 1991). Certainly there are some differences between the two parties in their early statements of principles. In 1937, Clement Attlee, who was to become the first Labour Prime Minister to preside over a majority government in Britain, wrote that Labour “has its general objective – the establishment of the Socialist Commonwealth.” This had a practical element – Attlee went on to say that Labour  “must take the first steps for its [the Socialist Commonwealth’s] realization, but at the same time it has to deal with immediate and pressing evils which call for remedy” (Attlee 1937, 166). However, it is still a more radical commitment to socialism than is expressed by the ALP, for whom the general 1921 objective of “the socialisation of industry, production and distribution and exchange” was tempered by the Blackburn Declaration “which added crucial qualifications indicating that socialisation was appropriate only when targeted at specifically  78 exploitative industries but not in cases where the instruments of production were used in a ‘socially useful’ manner and ‘without exploitation’” (Beilharz 1994, 58). Johnson also points out that “both the Curtin and Chifley governments believed that sections of private enterprise had an important role to play in post-war economic development” and argues that “once the continued existence of capitalism is accepted a given, there are only limited and contingent options for improving the standard of living of the mass of the population … ‘Pragmatism’ becomes the order of the day” (Johnson 1989, 16, 6). Her conclusions about Labor’s lack of radical, principled ideas is supported by the report of the Canberra Times newspaper upon the election of the Curtin Labor government in August 1943. The editors opine that Labor’s “appeal to the electorate was based on three main points – firstly, that Labor has the capacity to govern; secondly, that Labor has saved Australia from her direct peril [in regards to the war effort]; and thirdly, that Labor will not use power to embark on a programme of socialisation” (quoted in Crowley 1973, 89, emphasis added). The newspaper also reported on the promises of H.V. Evatt (soon to be Attorney-General and Minister for External Affairs), who “emphasised two things to the electors. First, he pictured Labor as the protector of the middle class, and second, he gave assurance that private enterprise would be called upon to play its full role in post-war Australia” (quoted in Crowley 1973, 89). In Australia in the 1940s, there was no cohesive, principled idea about how health services should be structured: health planning bodies such as the NHMRC were excluded from the main track of government decision making, and never produced a clear, prominent statement of policy. The Labor government, in turn, wished to expand social services but placed no special emphasis or priority on health policy, and its practical brand of ideology did not support a radical approach to its policy goals. This was one major element of the choice to take an incremental approach to policy develop; a second was the lack of obvious electoral incentives to take radical action. Lack of electoral motivations The lack of a single, prominent report on health policy or services also meant there was no real rallying point for public opinion on health policy. During this time, there was very little public opinion polling in Australia on the importance or even popularity of health insurance with voters, which is perhaps itself an indicator of the issue’s low place on the public agenda. The earliest polls on health from this period, in 1943, questioned voters about their preferred method of financing social services, including medical care (from taxation or a weekly household contribution), but not where they ranked health or social service amongst other public issues, or  79 whether they liked the idea of government involvement in health services at all.77 In the fall of 1943, Gallup reported that “Three out of four Australians would favor a national medical service, irrespective of whether it were free, or cost each family 2 [shillings] a week,”78 but about ten months later, found that “the proposed non-contributory national hospital and medical service is right out of step with public opinion,” since 60 percent of voters wanted to require contributions for medical services.79 In 1944, Australian voters rejected a constitutional referendum that would have greatly increased the Commonwealth government’s powers over health and thirteen other areas. The preference for contributory social benefits, rather than free services such as the PBS, was maintained until 1947, when a slight majority (53 percent) of Australians preferred that social services be financed “from ordinary taxes.”80 However, Australians also maintained a suspicion of high rates of taxation for social services. In 1945, Gallup found that “The Federal Government’s plan of post-war social services is not supported by the people, if it means maintaining the present taxation levels,”81 and in 1948, voters were asked whether they would prefer increased social services (the type of increase was not specified) or reduced taxes, and about 70 percent preferred lower taxes.82 Therefore, it appears that even after the Labor government’s attempts to implement the PBS were well underway, social services were not a high priority for voters. Perhaps the most striking contrast with the UK in terms of electoral incentives for action is that opinion on health policy in Australia remained divided along party lines for almost the entire study period. In the UK, both major political parties supported the concept of a free, universal and comprehensive public health service. The original plan for a broad health service was made by the wartime coalition government of both the Labour and Conservative parties, and when a new Labour government proposed the revised plan for the NHS, the Conservative Party quickly  77. Gallup Poll (Australia), May-June 1943, Polls 124-131, “National Insurance by Contribution Popular;” August-September 1943, Polls 141-152, “National Insurance by Contributions Favored.” 78. Gallup Poll (Australia), October 1943, Polls 153-161, “Support for National Medical Service.” 79. Gallup Poll (Australia), July 1944, Polls 205-212, “Public wants medical service contributory.” 80. Gallup Poll (Australia), March 1947, Poll 416-425, “Big change in opinion of financing social services”. However, the first part of the decade was marked by a strong preference for contributory benefits: see Gallup Polls (Australia), December 1944-January 1945, Polls 241-248, “Social Benefits Should be On Contributory Basis;” November 1945, Polls 304-313, “Australians in Poll Insist on Contributory Social Services Scheme;” and September 1946, Polls 382-397, “Contributory Social Services Wanted.” 81. Gallup Poll (Australia), May 1945, Polls 264-271, “Present Taxes Too Heavy for social services.” 53 percent of Australians opposed maintaining the current level of taxation after the war. 82. Gallup Poll (Australia), February-March 1948, Polls 487-491, “Lower taxes preferred to more social services”.  80 got on board.83 Therefore, there was no policy alternative of “no public health services” being offered, and British polling from the time indicates that the principles of the NHS were both important to and popular with a large majority of the entire electorate, not a particular partisan group (Jacobs 1993, 113). However, in Australia, compulsory public health services were the exclusive policy domain of the Labor party, and tended not to elicit high levels of support from conservative Liberal-Country Party voters. This is evident late in 1945, shortly after the High Court ruling against the constitutionality of the PBS. Gallup questioned voters about the benefits that the Commonwealth Government should provide, and while both Labor and non-Labor voters were in favor of benefits that had been in place for some time, such as old-age pensions, widows’ pensions, and child endowments, they were divided about pharmaceutical benefits. Overall, 64 percent were in favor of the Commonwealth government providing free medicines, but this represented about 75 percent of Labor voters and only 50 percent of non-Labor.84 This partisan division continued in May 1948, when voters were asked if they favored or opposed the government’s plan for free medicine – the as yet unimplemented PBS. Gallup reported that “Public opinion is unsettled…a bare majority of 51 percent is in favor, 33 percent opposed, and 16 percent undecided.”85 However, a breakdown by voting intention provides more information: 70 percent of interviewees who planned to vote for Labor in the next election were in favour of the free medicines scheme, while only 28 percent of Liberal-CP voters were in favor. This split is repeated in later polls on the PBS and other health policies,86 which suggests that even though these policies might appeal to existing Labor voters, they did not have the broad public appeal necessary to motivate radical action by the government. There is imperfect information available on where health issues ranked on Labor and Liberal voters’ agendas, but the fact that free services were actually unpopular with a significant portion of the electorate likely made it less tempting for politicians to tackle.  83. The Conservative party’s 1945 conference included a motion welcoming “the Government’s decision to make a comprehensive National Health Service available to all” (Craig 1982). 84. Gallup Poll (Australia), December 1945-January 1946, Polls 314-326, “Federal Social Services Approved.” 85. Gallup Poll (Australia), May 1948, Polls 511-528, “Public uncertainty about free medicines” 86. Gallup Poll (Australia), February-March 1949, Polls 569-578, “Medical & Dental Plan”; May- June 1949, Polls 590-599, “Public still wants free medicine, but opposition grows”; March-May 1969, Polls 2105-2118, “Keep medical funds voluntary”.  81 The incremental decision The preceding two sections have shown that in the 1940s, the ALP embraced a less radical version of Labor principles than its British counterpart and gave clear indications that they preferred an incremental approach to social policy development, and that the party lacked clear electoral incentives to take more radical action: voters prioritized tax cuts over social service expansion, and health policies only appealed to decided Labor voters. However, Labor did wish to begin implementing new social security measures during the war, perhaps to take advantage of the “elastic” properties of the wartime defence powers under Section 51 (vi) of the Australian constitution, which beginning in World War I expanded beyond purely military matters to social and economic powers for the Commonwealth government as well (Gilbert 1980, 316). Chifley argued that if the government did not begin on cash benefit programs during the war “all sorts of excuses will be found when the War ends for not passing them,” (quoted in Gray 1991). Some pension programs for invalids and the elderly were already underway (McAllister and Moore 1991): Labor expanded these, but also initiated novel benefits for health and unemployment (Matthews and Jay 1972). The government’s desire to implement new policies, amidst its concerns about the availability of resources to do so, helps explain the Curtin government’s decision to bypass the ambitious and often internally conflicted health planning bodies and give Treasury the responsibility for health proposals (Gillespie 1991). It was the federal Treasury, rather than the Department of Health or any of the preexisting health planning bodies, which made the decision to proceed incrementally and begin with pharmaceuticals.  This meant that “control of costs began to dominate official thinking” regarding the initial design of the PBS (Gillespie 1991), and is further evidence of the lack of principled, programmatic ideas with regards to health policy development in Australia at this time. As noted above, when Labor took power in September 1941, the National Health and Medical Research Council was already planning a new, centralized system for public health. Labor initially seemed supportive of the NHMRC’s approach to health policy, but this approach was incongruent with the party’s focus on general social security and its concerns about the country’s existing medical capacity. In 1941 the new Commonwealth Health Minister, E.J. Holloway, promised there would be no changes towards a salaried medical service until after the war, owing to the absence of so many Australian doctors with the armed services (Gray 1991; Gillespie 1991). In October 1942, Chifley, acting in his capacity as Treasurer, requested a report  82 from the bipartisan Joint Parliamentary Committee on Social Security (JPCSS)87 on health services “with particular reference to such measures as it might be possible to introduce during the period of the war” (Gray 1991; Gillespie 1991). The Commonwealth Director-General of Health, J.H.L. Cumpston, responded the following month with a “radical scheme for immediate action,” based on previous NHMRC reports and calling for the gradual introduction of a salaried medical service, beginning in remote areas (Gillespie 1991). This plan was echoed in the JPCSS’ formal Sixth Interim Report in July 1943, but archival records indicate that by this time the government had already bypassed the health planning bodies and begun consultations for the eventual Pharmaceutical Benefits Scheme with the Pharmacy Guild.88 In December 1942, one month after Cumpston’s report on the wartime development of health services, Cabinet met to discuss social security. Gillespie (1991) notes that “the government’s plans for the health service remained sketchy” and there was a greater focus on unemployment and sickness benefits, and widow’s pensions. Although Chifley’s cabinet submission included Cumpston’s cost estimates for his plan, it also included “a proposal for pharmaceutical and dental benefits, contradicting the direct service emphasis on the NHMRC’s proposals” (Gillespie 1991). This was the first indication that the Labor government intended to proceed with the staged introduction of services, rather than the gradual introduction of salaried medical services by area, as the NHMRC recommended, but the need to proceed in stages became a common theme in the next decade of health policy development. For example, a 1944 memorandum notes that “The Treasury has announced that the Government intends to deal progressively with the provision of a high grade service under which the public can, at the public cost, obtain all necessary medicines, a hospital service…and ultimately, the introduction of a system under which medical services…also will be available to every citizen at the public expense,” and also that “These measures for their full practical application require progress step by step, and the Government has decided that the first step, which will be immediately introduced, is the provision of a service under which necessary medicines will be made available to every citizen” (emphasis added).89 In the JPCSS’ next formal report, in July 1943, “recommended strongly against the early adoption of a comprehensive scheme” (Gray 1991), but this principle was already accepted in government: the recommendation came seven months after Cabinet’s  87. The JPCSS was created by the non-Labor government in July 1941 but relied heavily on NHMRC ideas and analysis (Gillespie 1991) 88. Valuations, 24 May 1943 – 8 June 1943. NAA. A571, 1943/4513. “Pharmaceutical Benefits Bill - Development of legislation, 1943-1944.” 89. Memorandum, January 1944. NAA. A571, 1943/4513. It  83 approval, in January 1943, of a cash benefits programme for pharmaceuticals, hospitals, and tuberculosis care (Gillespie 1991). Labor’s lack of principled ideas about health policy and lack of compelling electoral motivations to take a particular type of action on health allowed Treasury to take over health planning in the face of slow and contested deliberations by the NHMRC and JPCSS, and resulted in a timely and pragmatic solution to the government’s desire to move forward on social benefits during the war. In the next section, I turn to Treasury’s reasons for selecting pharmaceutical benefits as the first priority for the staged introduction of services. The second crucial decision: determining the order of priorities In Canada, the decision to proceed with health policy development in stages was taken in 1945, but debates about which services should be implemented first continued in some forums until 1955. In Australia, the decision about where to begin was taken about the same time as the decision to proceed in stages. Beginning with pharmaceutical benefits is not intuitive: they are not the most obviously important service, and this decision has been something of a puzzle in Australian health policy. However, archival records indicate that, in keeping with the government’s concerns about the lack of resources to implement social security measures, the decision to begin with pharmaceuticals was a pragmatic one. An unusually high proportion of Australian doctors were overseas during the war, and the government did not believe there would be sufficient medical personnel for any health services besides pharmaceutical benefits until after the war. However, the choice of priorities was also supported by compliant state governments and the federal government’s expectation that pharmaceutical benefits would be less controversial with the medical profession. As we shall see, the government was incorrect on this last point, which meant that the PBS did not function until a Liberal-Country coalition government was elected in 1949 and was able to negotiate new terms with the Australian branch of the British Medical Association (BMA) for the scheme’s implementation. Resource constraints: the lack of medical manpower During World War II, more than one-third of Australian doctors were on active service (Crowley 1973), and military dominance of medical mobilization resulted in “major gaps in civilian services” (Gillespie 1991). This had a significant impact on the federal government’s priorities for health services. A 1944 Treasury memorandum explained that pharmaceutical benefits were meant to be part of a comprehensive health scheme eventually, but alone, these services “will not involve any significant additional drain on professional man power and it is this  84 feature which enables the introduction of the scheme before the end of the war.”90 Another Treasury document reiterates that the government was considering medical benefits but “owing to the absence in the fighting services of a substantial proportion of medical men, it is expected that it will not be possible to introduce a scheme of free medical services until after the war.”91 A survey of pharmacy facilities between 1943 and 1944 concluded that they were “completely, indeed more than, adequate as regards facilities for a government scheme of pharmacy benefits,” since “mortality among dispensers has been negligible since war broke out – pharmacists have joined the Forces, but their shops have not been closed.”92 Although this suggests that the choice of pharmaceuticals as a first priority was dependent on fairly idiosyncratic factors, the Australian government’s response to these particular resource constraints is consistent with what we would expect from a government motivated by a pragmatic desire to implement some policy change, without guidance from principled policy preferences. Government documents emphasize that pharmaceutical benefits were a practical rather than principled decision. A draft speech for the second reading of the Pharmaceutical Benefits Bill in the House of Representatives notes that other health priorities such as hospital construction and tuberculosis treatment were important, but that they were “subject to matters of higher policy, such as manpower, etc.” The Minister went on to say, “I am quite sure that the average citizen is not likely to object to having free medicines simply because he is not able, at present, to get free medical advice.”93 A caucus meeting in 1944 repeats this point, with one member arguing that pharmaceuticals were going first because they were practical, not because they were most important: “I will admit that the provision of a medical service would have a more profound effect on the public health and on the axiom of first things first it might be argued that free medicines could wait.”94 A government pamphlet published in 1945 also deals with this issue, asking whether the money intended for the PBS should be spent on other health services. The publication answers,  The only justification for such a question would be some evidence that certain important health services were being denied to the public merely because there was no money to finance such schemes. The reason why pharmaceutical benefits have been introduced before other urgent  90. Memorandum, January 1944. NAA. A571, 1943/4513. 91. Social Security in Australia memorandum. NAA. A571, 1943/4513. 92. Pharmacy manpower surveys. NAA. A571, 1943/4513. 93. Second reading speech, Pharmaceutical Benefits Bill (House, draft). NAA. A571, 1943/4513. 94. Caucus meeting notes. NAA. A571, 1943/4513.  85 measures is merely a matter of practicability since war conditions do not hamper its introduction, on the contrary it assists the war effort (emphasis added).95  Therefore, pharmaceutical benefits do not seem to have been introduced out of any sense of urgency on the part of the government to ensure that citizens had access to free drugs. Rather, pharmaceutical benefits were a starting point for other policies. For a government anxious to implement some new social security measures “before the disruption of demobilization,” (Gillespie 1991) but wary of straining existing medical resources, pharmaceutical benefits became an “obvious” first choice of priorities. Pharmaceutical benefits and a short-lived hospital benefits scheme that followed after the end of the war were not strongly contested by state governments, but encountered greater opposition than expected from physicians. Barriers to policy implementation and further development Once adopted, early health services in Canada and the UK did not face any extraordinary difficulties in implementation; that is, they functioned fairly soon after the policies were adopted without much additional political negotiation. This was not the case in Australia, where the medical association was strongly opposed to the PBS and was able to use the institutional arrangements of federalism and simple non-compliance to prevent the scheme from functioning for the first five years of its existence: the legislation was in place, and was modified twice in the face of constitutional challenges, but almost no patients received free prescriptions under its auspices. This period had a significant impact on the form the PBS took when it was finally operational, and so it is useful to consider the opposition and government’s reaction in more detail. The circumstances of the PBS’ adoption suggest that electoral incentives to maintain benefit programs can arise surprisingly quickly, even in the absence of a functioning program and also help explain why there was no further policy development. Opposition from organized interests Unlike Canada, state governments in Australia had little or no input into the selection of initial priorities for health policy development. The federal government’s lack of constitutional jurisdiction over health did temporarily stymie its attempts to implement the PBS, but this was not a reflection of state preferences over priorities: at this time, “even the most conservative state governments were reluctant to challenge the Commonwealth’s powers” (Gillespie 1991). Instead, the failure of the 1944 referendum provided an opportunity for the BMA to initiate a  95. Government Q&A’s on the PBS, 1945. NAA. A571/1943/1812 Part 1. “Pharmaceutical Benefits, 1938-1945.”  86 constitutional challenge through the Attorney-General of Victoria, a key success in the medical association’s long battle against the PBS. Gillespie argues that one of the reasons pharmaceutical benefits and hospital benefits were attractive first and second priorities was because “both (it appeared) could be implemented immediately without waiting for the long and cumbersome negotiations with the medical profession” (Gillespie 1991). This was a reasonable expectation: to this point, the medical profession’s main concern was that there be no changes to medical service until after the war, as Health Minister Holloway promised in 1941, so benefit schemes that subsidized drugs or hospital beds without touching on the contentious issue of physician remuneration seemed promising. Furthermore, physicians already participated in the limited Repatriation Pharmaceutical Benefits Scheme, which provided free medicines to veterans on a similar basis as the proposed PBS (Sloan 1995). However, the government did not predict that the BMA would view the PBS as the wedge towards socialized medicine and would oppose it “with a furor and effectiveness which decided its fate” (Hunter 1965, 412). The government’s strong expectation of cooperation can be seen in its decision not to consult with doctors until December 1943, six months after successful consultations with the pharmacists, and to present the scheme as a fait accompli - a choice which did not endear government planners to the BMA’s Federal Council (Hunter 1965, 415; Gillespie 1991). Therefore, doctors’ opposition did not affect the initial selection of priorities for health services development: the Commonwealth government had made significant progress in planning the PBS by the time this opposition became apparent, and continued to try to implement the scheme despite the opposition. The BMA opposed the introduction of the PBS and, once it was passed, prevented the scheme from functioning by simply refusing to use the government prescription forms necessary for patients to obtain free drugs at the pharmacy. The doctors’ most overt objection was to the proposed Commonwealth Formulary, or list of subsidized drugs. Doctors argued that being constrained to even a broad formulary would compromise their patients’ care, and that they would only cooperate with the government’s Formulary Committee when it agreed to the principle that “where a drug is ordered which is not within the formulary the patient shall nevertheless be supplied with such drug without cost.”96  96. BMA to Health Minister, 19 February 1944. NAA. A571, 1943/4513.  87 However, these concerns overlaid a more basic fear of major changes to the medical profession and the introduction of “socialized” or “nationalized” medical services. At the first reading of the bill in 1944, Liberal member of parliament and BMA champion Sir Earle Page commented that “this provision may be the thin end of the wedge of nationalization of the medical profession…My own strong view is that the nationalization of the medical profession would cause a deterioration of the caliber of its future personnel” (quoted in Crowley 1973). In a 1945 article in the conservative Argus newspaper, a doctor argued that “[i]n addition to the formulary, the individual practitioner was indignant because he saw in the Pharmaceutical Benefits Act a threat to nationalize the medical profession.”97 The BMA fought the PBS in private meetings with the Minister,98 and in public, where it issued newspaper statements to argue that the formulary would constitute a cap on total pharmaceutical expenditure,99 and published pamphlets to explain to patients that doctors “are not, in their patient’s interest, prepared to drift into a groove bounded by the narrow limits of a formulary,” particularly since, “in the minds of the Minister and the Government, a method of pricing is more important than the patient’s welfare.”100 The Labor government was not successful in countering this campaign; in fact Gillespie (1991) argues that “the Chifley government…proved singularly inept in mobilizing public support” for the PBS, and this meant the BMA was able to prevent the implementation of this first priority for five years.  Lack of further policy development In Canada, policy development stalled after the introduction of medical insurance in 1966, despite the fact the Liberal party remained in office for all but one of the next eighteen years. In Australia, however, a change in government was a decisive factor in blocking further health policy development. The position of the Liberal Party under Prime Minister Robert Menzies can be understood as a significant lack of principled ideas concerning comprehensive and universal health policy. According to Gray (1991), “Non-Labor favoured a policy of minimum government  97. “Objections to Free Medicine: bad for patients,” Argus, 8 June 1945. NAA. A571 1943/1812 Part 1. 98. Transcript of meeting, Minister and BMA, 21April 1947. NAA. A571 1943/1812 Part 2. “Social Services in Australia – Pharmaceutical Benefits. 1946-1947;” Transcript of meeting, Minister and BMA, 2 July 1948. NAA. A571 1943/1812 Part 3. “Social Services in Australia – Pharmacy Benefits. 1948-1947.” 99. BMA statement, 10 August 1944. NAA. A1928 781/4 Section 2. “Pharmacists and Pharmacy – Pharmaceutical Benefits Act – General. 1944.” 100. BMA pamphlet to hospital patients, 1945. NAA. A571 1943/1812 Part 1.  88 intervention in welfare areas,” and this is demonstrated in the Liberal party platforms. Instead of first-dollar coverage, they emphasized contributory insurance measures: the 1945 Liberal platform called for “the encouragement of supplementary voluntary schemes in addition to government schemes” for social security (White 1978), indicating a strong role of private insurance. This platform focused on public health initiatives and stated the party’s opposition to “the nationalization of the medical profession and service” (White 1978). Menzies’ 1946 election speech again calls for contributory social security and makes no mention of health, while his speech in 1949 calls for preventative health programs, saying “it is a grave error to treat the problem of a national medical health service as if it meant nothing more than the making of monetary payments to citizens through the Treasury” (McAllister and Moore 1991). Although Labor lacked the principled ideas and electoral motivation necessary for the simultaneous introduction of a wide range of health benefits, its platform was clearly more aligned with a public, first-dollar approach to health insurance. Liberal principles point away from public health benefits, and the government demonstrated these principles in 1952 by restructuring Labor hospital subsidies to states, albeit in a way that initially had limited impact on most patients.101 After the implementation of the Pharmaceutical Benefits Scheme, there was no further development of public health insurance until 1972, when a Labor government again came to power at the national level. We have seen that doctors were clearly opposed to public health benefits of any kind, and the BMA had expended considerable time and resources fighting the PBS, although they were prepared to be mollified by a Liberal government that was much more in tune with their interests. Given these factors, the lack of further health policy development after 1950 is expected, and the real puzzle is to explain the Liberal decision to implement a modified PBS. I argue that this decision is an example of how quickly public expectations of service can take hold, a conclusion which is supported by the  101. The 1946 Labor Hospital scheme provided federal funds to state governments conditional on states removing the means test for free public wards, and introducing a per diem subsidy for private patients (Gray 1991). Some states were initially reluctant, but unlike provincial governments in Canada, no one state had an absolute veto power over an intergovernmental agreement: the federal government “made clear that the reluctance of individual states to enter the [hospital benefits] scheme would not prevent implementation in those more co-operative” (Gillespie 1991). When the Liberal-Country coalition took power in 1950, they modified the hospital benefits scheme in keeping with their principled ideas about the importance of private insurance and self-help, by reintroducing the means test for public hospital wards, and making per diem hospital grants directly to individuals conditional them holding approved private insurance. Although these changes were important in creating a larger market for private insurance, Gillespie (1991) argues that “[n]either provision…represented a major shift from the practices in force under [Labor PM] Chifley,” or had an immediate effect on patients.  89 evidence contradicting any alternative explanations for the PBS’s implementation, such as support for the scheme from within government or from a major interest group. Implementing the PBS: the role of public expectations When the Liberals took power in 1949, the PBS had been on the books for five years, but lack of cooperation from doctors meant very few people were receiving benefits. Although the Liberals criticized the Pharmaceutical Benefits Act when it was introduced in 1944 (Crowley 1973), once the scheme was (nominally) in place, there seems to have been a different attitude, reflecting the difficulty of removing outright an existing benefit (Pierson 1994). In the 1946 election speech where he called for contributory health insurance, Menzies also promised that “no person who enjoys any social service payment today will have it taken away or reduced by a Liberal Government” (McAllister and Moore 1991). The turning point came in 1948, when the Liberals added pharmaceuticals to their party platform: the health section called for “the free provision of certain specific drugs vital to the preservation of life (such as insulin)” (White 1978). This hinted at the Liberal’s eventual restriction of the PBS formulary, but nonetheless demonstrates that some level of pharmaceutical benefits was understood to be desirable, or at least inevitable. Gillespie’s (1991) findings supports this: he notes that the Liberal-Country coalition “accepted the fact that the Commonwealth could not withdraw from the finance of health care,” although they did believe that administration could be private if government subsidized private not-for-profit insurance. Roskam also notes that, when Menzies became Prime Minister in 1949, he “endorsed the central elements of welfare policies adopted by Labor in the 1940s, and he recognised that first the Great Depression, and then war, had changed community expectations about economic and welfare policies” (2001, 278). By 1950, the PBS had become a reasonably salient issue: it had been the subject of two High Court cases (one only a year previous) and a successful constitutional referendum. Furthermore, opinion polls demonstrate that the public was also reasonably attached to the idea of a public pharmaceutical program. In March 1950, when the Liberal’s scheme for a less comprehensive PBS was first introduced, Gallup reports that “at this stage, the public is not keen on the idea.”102 The poll asked whether people favored a scheme where all medicines were free (like the original PBS), a scheme were only expensive medicines were free (as was proposed by the new Liberal government), or if they preferred no free medicines. Overall, 43 percent of respondents favored  102. Gallup Poll (Australia), March-April 1950, Polls 662-676, “Public not keen on Page medical plan”  90 “all medicines free”, and 37 percent preferred that only expensive medicines were free. As expected, there was some difference along party lines, with 57 percent of Labor voters supporting the “all free” scheme, versus 33 percent of Liberal-CP voters. Crucially, however, only 15 percent of respondents overall favored “no free medicines,”103 suggesting a new set of expectations based on government’s policy promises. A few months later, this dropped to only 11 percent of respondents preferring no free medicines,104 so taking some sort of action on a pharmaceutical program was clearly an electoral winner across party lines. The Liberals recognized the significance of having the PBS already in place, if not operating, even when they criticized Labor’s approach to health policy. In 1948, an opposition member asked whether the government was “sincere in introducing a “free medicine” scheme for the benefit of the public, or does it believe that the inauguration of such a scheme will pave the way for the nationalization of all medical services?” and in the same speech deplored the fact that citizens were paying taxes for free medicines that they were not receiving.105 This type of rhetoric set the stage for the early Liberal implementation of the PBS when they took power, although the scheme that began operation in September 1950 differed in some important ways from the original Labor design. In the next section, I argue the Liberal government’s initial design was shaped its ideological position opposing comprehensive health services in principle and the need to secure doctors’ cooperation, and that its later choice of cost control measures was influenced by the need to manage public expectations of service. The early challenges of the Pharmaceutical Benefits Scheme The final element of the theoretical framework addresses the question of how policymakers initially respond to the challenges of implementing pharmaceutical programs, where cost increases quickly became an important policy issue. In Australia, there was much greater consideration of the costs of pharmaceutical benefits than there were for prescription services or other elements of the NHS in the UK. The reason for this is not entirely clear, but as I argued in chapter three, it is possible that cost concerns become more important when there is a dearth of principled ideas about a policy area. The PBS was designed with one main tool for cost control: formulary control, or management of the list of drugs available for subsidy. Although the Labor government stressed the importance of a formulary in its initial design for the PBS, the scheme as implemented by the Liberals in 1950 had a much more restrictive formulary. Once the  103. Ibid. 104. Gallup Poll (Australia), June-July 1950, Polls 690-699, “Page medicine plan hasn’t “caught on”.” 105. Mr. Harrison, Hansard, 17 June 1948. A571 1943/1812 Part 3.  91 unintended consequences of this scheme for the way doctors prescribed and for the cost of the program became clear, patient contributions were added. These were introduced in 1960 in combination with a broadened formulary and have increased regularly since then. Formulary control was initially a response to cost concerns prompted by international experience with pharmaceutical benefits programs, and the Liberal government’s decision to tighten the formulary considerably from the original Labor plan was determined by doctors’ preferences, who preferred that public insurance only account for a limited portion of their prescribing. The choice ten years later to implement patient contributions for prescriptions was also driven by cost concerns, given the predictable explosion of expenditures under the original scheme, but here public expectations played a role. The government introduced patient charges, which I argue was less politically risky in Australia than the UK because of the legacy of limited commitment to comprehensive health services. However, the government also aimed to mollify the public by combining new charges with a wider formulary, an opportunity for compensation that was not available in the UK. Formulary control and early cost concerns Control over the formulary has been an important tool for pharmaceutical management in Australia since Labor’s earliest plans for pharmaceutical benefits. Conflict between the government and the medical profession over the formulary was a key determinant of when and how the PBS was eventually implemented, and over time, the power to select which drugs would be subsidized, and under what conditions they would be subsidized, became a key strategy in Australian policymakers’ struggles to control overall costs of the PBS. This contrasts with the situation in the UK, where aside from a fairly insignificant Selected List (introduced in 1984), any drug approved for sale in the UK is potentially available for subsidy. Labor policymakers did aim to cover a broad range of drugs, but this goal was tempered by the experience of New Zealand, which introduced a program of pharmaceutical benefits without a formulary in 1938, and by the mid-1940s, was struggling with much higher costs than anticipated.106 In 1944, the government reminded the BMA that it was “paying for the entire scheme and therefore must maintain a certain amount of control over what it will pay for…Experience in other countries suggests that, without the use of a formulary, the cost per  106. Notes on a Visit to New Zealand, January 1945. NAA. A571, 1943/1812 Part 1.  92 script tends to rise year by year.”107 The government believed that a formulary was necessary to maintain “the greatest economy in [drugs’] usage” during the war, as well as to lower administrative costs and provide a quick and easy method of prescribing, since it would standardize prescriptions that were often still compounded for each individual patient.108 The government also noted, in a 1943 memo, that “[t]he public, under war time conditions, will more readily accept limitations – such as a restriction of dispensing to a formulary.”109 Labor saw a formulary as necessary to control public drug expenditure, but it became a sticking point with doctors. Indeed, a Cabinet submission in 1947 concluded, “the whole controversy in relation to pharmaceutical benefits revolves around the matter of the use or non- use of the Formulary.”110 The BMA was unsatisfied with Labor’s promise that “[t]he formulary will be as comprehensive as the medical professional desires, [and] no restriction will be placed on the number or variety of drugs and compounds used except where it is necessary to avoid redundancy or to preserve efficiency,”111 and would only accept a formulary as long as drugs prescribed outside it would also be subsidized, effectively rendering the list meaningless. Doctors’ preference was for a limited list,112 so they could retain complete autonomy over compounded prescriptions: those medicines that chemists mixed in their shops according to the doctor’s instructions, rather than manufactured medicines that were mass-produced and shipped in their final form. If a formulary was sufficiently comprehensive, it would provide formulas for most commonly used drug preparations and prevent doctors from prescribing individualized compounds as they saw fit. A limited list would mean many prescriptions would fall outside of government purview and control, in terms of preparation and cost. The Liberal party supported the BMA on this position (Gillespie 1991), and when they took power in 1950 they implemented a PBS with a new system of formulary control that responded to doctors’ preferences. Essentially, physicians’ concerns about the formulary appear to have been grounded in their desire to retain the maximum autonomy over prescribing, whether that was by the government covering any  107. Notes for BMA Conference, 28 January 1944. NAA. A571, 1943/4513. See also Social Security in Australia memorandum. NAA. A571, 1943/4513. 108. Valuations, 24 May 1943 – 8 June 1943. NAA. A571 1943/4513. 109. The advantages of establishing a pharmaceutical benefits scheme during war time, memorandum, 1943. NAA. A571, 1943/1812 Part 2. 110. Cabinet submission, 19 September 1947. NAA. A1658, 813/1/1 Part 1. “Pharmaceutical Services – Pharmaceutical Benefits Act – General – Policy. 1949-1952.” 111. Minister of Health to physician, 9 August 1944. NAA. A1928, 781/4 Section 2. 112. Cabinet memorandum, 24 September 1947. NAA. A27000, 1005C. “Pharmaceutical Benefits Act 1947.”  93 script a physician wrote (even if it was compounded according to individual instructions), or by limiting government prescriptions to a small number of costly manufactured medicines. Revised formulary and doctors’ preferences Although the secondary literature emphasizes the similarities between the original Labor PBS and the PBS eventually implemented by the Liberals (Hunter 1965, 413), at the time of their design policymakers regarded the two schemes as “fundamentally different.”113 The restricted formulary proposed by the Liberals differed from the Labor plan and allowed them to secure the BMA’s support and finally get the PBS underway. Although polls in 1950 indicated a slight public preference for the original scheme where all medicines were to be provided free,114 the public had never actually experienced a functioning pharmaceutical program, and this may have lowered expectations somewhat. In February 1951, Gallup interviewers reminded respondents that under the new PBS “pensioners now get all medicines free, and that the general public gets the expensive life-saving drugs free,” and then asked “Do you think that’s enough, or should all medicines be free?” Overall, 51 percent of Australians thought the Liberal plan was “enough,” which breaks down to 59 percent of Liberal-CP voters and 40 percent of Labor voters.115 The original Labor formulary was to “constitute all reliable combinations of medicines covering a whole range of the practice of medicine.”116 In notes prepared prior to the second reading of the Pharmaceutical Benefits Act in 1944, the government reiterated its “very definite intention…that when the Commonwealth Formulary is published it will be one of the most comprehensive and up-to-date formularies in the world,”117 and again in 1945, that “the Commonwealth Pharmaceutical Formulary is far more comprehensive than any formulary previously published in Australia or, as far as we know, the world.”118 The PBS formulary implemented by the Liberal government was quite different: it was limited to “life-saving and disease-preventing drugs” (Sloan 1995), and was aimed at “prescribing  113. Keith Attiwill, “Report from Public Relations,”Gilseal News (Pharmacuetical Guild newsletter), 16 October 1952. NAA. A1658, 811/1/1 Part 1. “National Health Scheme – Pharmaceutical Benefits – General. 1946-1953.” See also Cabinet submission (Sir Earle Page), 27 August 1952. NAA. A571, 1950/596. “Pharmaceutical Benefits – New Scheme. 1950-1953.” 114. Gallup Poll (Australia), March-April 1950, Polls 662-676; June-July 1950, Polls 690-699. 115 Gallup Poll (Australia), February-March 1951, Polls 744-755, “Page medicine plan goes far enough, most say”. 116. Memorandum, January 1944. NAA. A571 1943/4513. 117. Pharmaceutical Benefits Bill: Notes on general policy. NAA. A571, 1943/4513. 118. Chifley to Senator Lamp, Tasmania, 3 May 1945. NAA. A571, 1943/1812 Part 1.  94 for serious disease and not minor aliments.”119 Liberal politicians emphasized the differences between the two schemes when they were discussed in cabinet: for example, the Treasurer compared the old formulary which covered a “wide range” of drugs to the new list which could cover “life-saving and costly drugs – about 50 in all – which do not require compounding by the chemists.”120 Sir Earle Page, the Liberal Minister of Health, argued that the broad Labor formulary “makes it very difficult to limit the volume of medicine prescribed and breeds the expectation of a free bottle of medicine on every visit to the doctor,” while the new scheme with its limited formulary was “comparatively simple and safeguards against extravagant waste of drugs.”121 In 1950, the Commonwealth Department of Health predicted that the new limited formulary would cover about 20 percent of prescriptions.122 This satisfied the Liberal party’s ideological desire to avoid any hint of “nationalization” in its provision of social benefits, and the government’s concern about controlling costs. It also ensured doctors’ cooperation, since it matched their desire to have freedom to prescribe outside the formulary: many types of drugs were not listed and therefore available only on private prescriptions. However, the limited formulary had some unintended (if not entirely unpredictable) consequences with regards to costs. In 1950 a departmental committee predicted that limiting the formulary would “attract constant pressure for change and expansion,” and would encourage the overuse of expensive drugs,123 since patients would pressure doctors (and perhaps doctors would prefer) to prescribe a more expensive, listed drug when a cheaper unlisted drug would be equally effective or appropriate. The Treasurer made many similar warnings (Gillespie 1991), and this was in fact what occurred. In 1960 rising costs forced the Liberal government to reconsider their program of formulary control and to introduce a new tool for cost control: patient contributions.  119. National Health – Pharmaceutical Benefits – Background Press Package, 11 August 1959. NAA. M2568/25. “Personal Papers of Prime Minister Holt: Miscellaneous papers 1959-1961.” 120. Treasury submission, 19 June 1950. NAA. A571, 1950/596. 121. Earle Page, “Department of Health – Pharmaceutical Benefits,” Cabinet submission 22 June 1950. NAA. A1658 813/1/1 Part 1. 122. Major principles of the proposed National Health Scheme. NAA. A4933. “NHS – National Health Services 1950 – Committee – Department of Health – Pharmaceutical Benefits. 1950.” 123. Major principles of the proposed National Health Scheme. NAA. A4933.  95 Patient contributions and public expectations Cabinet documents show that concerns about the rising costs of the PBS were present as early as 1952.124 Government’s initial response was to further control the formulary, by “limiting the use of several of the more costly drugs to certain specified diseases [and] further revision of the list of drugs available for prescription.”125 However, this additional formulary control was not sufficient to rein in costs, and in 1959, the Liberal government recognized the need for more radical options. In early July, cabinet met “to review developments in connection with Pharmaceutical Benefits, particularly the rapid increase in costs, and to suggest measures which might be considered.”126 Cabinet debated closing the formulary to expensive new drugs, but this was deemed too difficult. Other suggestions were to broaden the list of free drugs in hopes of encouraging doctors to be more economical in their prescribing, to (further) restrict use of expensive drugs to certain diseases, and to impose a patient contribution: on this last option, the Health Minister’s submission noted, “It has the obvious political disadvantage, but is the one sure way to control over-prescribing.”127 On July 13, the Prime Minister considered a proposal to introduce a fee for prescriptions, and at the same time broaden the formulary, which was the option selected for implementation in 1960.128 The different focus of private government deliberations versus public policy announcements demonstrates that the government made a conscious attempt to frame these changes to the PBS so as not to thwart public expectations of service. The Prime Minister’s memo (written by his staff in preparation for an upcoming cabinet meeting) focuses on the introduction of the fee, while the later public materials announcing the two policy changes focus almost exclusively on the broader formulary. The Prime Minister’s memo discusses the problems with relying on doctors to adopt more economical prescribing habits and concludes, it seems clear that the only way to provide any effective brake is by the charging of a fee. This would be more acceptable to the public if it associated with a wider range of drugs being available under the scheme, and the evidence in the submission suggests that the latter would be no burden on the Commonwealth (emphasis added).129   124. Cabinet minute 31 March 1952 (Decision 360). NAA. A1658, 813/1/1 Part 1. 125. Cabinet minute 31 March 1952 (Decision 360). NAA. A1658, 813/1/1 Part 1. The PBAC still limits the use of certain drugs to specified conditions as a measure of pharmaceutical management today. 126. Cabinet submission, 2 July 1959. NAA. A1658, 813/1/2 Part 1. “Pharmaceutical Benefits Act – General – Policy – Review commenced June 1958.” 127. Cabinet submission, 2 July 1959. NAA. A1658, 813/1/2 Part 1. “ 128. Notes on Cabinet Submission No. 252. 13 July 1959. NAA. A5619, C523. 129. Notes on Cabinet Submission No. 252. 13 July 1959. NAA. A5619, C523.  96 The Liberal government’s strategy appears to be a clear example of what Pierson calls “strategies of compensation,” which recognizes that “offering something positive to the victims of retrenchment policies will diminish prospects for heated opposition” (Pierson 1994, 23). Offering a broader range of subsidized drugs as compensation for the introduction of a prescription charge was a particularly effective strategy because the characteristics of pharmaceutical benefits allowed it to avoid the two main drawbacks Pierson points to for strategies of compensation: cost and policy irrationalities, or the potential of creating new problems through attempts to compensate for retrenchment (Pierson 1994, 24). As the quote above suggests, expanding the formulary to cover inexpensive and previously excluded drugs was unlikely to increase of the cost of the program so much that it would outweigh the revenue generated by a charge for every prescription. Furthermore, an expanded formulary was potentially simpler to administer, since the same procedures would now apply for most drugs, and would avoid certain perverse incentives built into the original program, such as prescribing costly drugs in place of equally effective and cheaper options, when the government paid for only the costly drugs. The government’s internal consideration of the merits of adding a prescription charge and expanding the formulary, above, can be contrasted with the public announcement of the change, in February 1960, which focused on formulary issues. It stated that “the Government regarded the new scheme confidently as an economical method of extending to the community as a whole the hitherto limited application of the pharmaceutical benefits legislation,”130 referring to the fact that the 1950 scheme included a broad formulary for pensioners and a limited list for the general population. The Minister said the five-shilling (about 50 cent) fee was necessary to stabilize the ever-increasing cost of the previously free scheme – a consideration which the more thoughtful taxpayer would readily appreciate. More important, however it made possible the expansion of the former list of free medicines to a greatly widened area of prescribing…Drugs would be available in some form or other under the new scheme for the treatment of virtually every illness or disability encountered by doctors (emphasis added).131  Therefore, the deliberations that led to the policy change focused on a patient contribution as the only “effective brake” on expenditure, with a wider formulary as a measure to make the change more politically palatable, while the government’s public announcements of the new scheme framed it mainly as a formulary expansion, “made possible” by the fee, in recognition  130. Press statement by Dr. [Donald] Cameron [Minister of Health] re: New Pharmaceutical Benefits, 16 February 1960. NAA. A463, 1956/1905. 131. Press statement by Dr. [Donald] Cameron [Minister of Health] re: New Pharmaceutical Benefits, 16 February 1960. NAA. A463, 1956/1905.  97 that they would be dealing with the public’s decade long experience with having at least some prescriptions for free. The policy change was not entirely unopposed: some cabinet ministers were concerned that the fee might “impose great hardship on persons suffering from diseases or conditions requiring long term drug therapy,”132 and pharmacists initially said they would only collect the fee if compelled by legislation, since as the Guild President complained, “Dispensing at Government rates is a dead loss…Take away our private prescriptions and we fold up, unless, of course, the Government is prepared to pay for drugs at private rates.”133 Interestingly, however, the framing effort appears to have been successful as far as the public was concerned. The archival records of the policy change only record the complaints of pharmacists and some doctors who were concerned that they might face government pressure to prescribe inexpensive drugs more often,134 and no records of public outcry in newspapers or letters to ministers, as exist in the UK after the introduction of prescription charges. When cost control measures were first introduced, the Liberal government was able to forestall potential electoral backlash using a strategy of compensation, but this option was only available once: the program was already universal, so after coverage was expanded to all drugs, there were limited options for expansion. The fact that subsequent cost control policies also encountered limited opposition suggests that the lack of “founding principles” in Australia made this easier. As PBS expenditures continued to rise in later years, government returned to patient contributions as an effective policy option with limited concerns about political risk. In 1965, the fee was described to a policymaker in another country as both a “brake” on pharmaceutical expenditure and as a policy “in line with the principle of “self-help” upon which the Australian National Health Scheme is based.”135 In 1970, the fee was doubled (to AUS$1), and in a cabinet meeting discussing the change, the Department of Health justified the fee because patient contributions had not increased since they were introduced in 1960, while the annual costs of the PBS continued to rise, and because the fee should be considered in reference to the increase in incomes since 1960.136 Neither this cabinet decision, nor a similar move to double the patient  132. Cabinet Minute (Ad Hoc Committee) – Decision No. 388 HOC, 17 August 1959. NAA. A1658, 813/1/2 Part 1. 133. Melbourne Herald 21 September 1959. NAA. A1658, 813/1/2 Part 1. 134. Ibid. 135. Reply to information request from Senator, Jersey, Channel Islands, 1965. NAA. A1658, 813/1/1/1 Part 3. “Pharmaceutical Services – Pharmaceutical Benefits Act – General – Policy. 1954- 1965.” 136. Cabinet Minute, 21 July 1971. NAA. A5908/192. A5908/192. “Pharmaceutical Benefits – Patient contribution – Decision 259 (M). 1971.”  98 contribution again in 1975,137 records concern about the political or electoral implications of increasing the fee. This contrasts with the situation in the UK, where for many years every increase in the patient contribution was the subject of much political angst.138 In Australia, the Liberal government was sensitive to the possibility of thwarting public expectations of a free service in 1960, and countered it by linking the new fee to a wider formulary. They were aided by both the opportunity to use a strategy of compensation, and the fact that patient contributions never as politically charged as they were in UK, because of the lack of principled ideas that fostered clear public expectations of a certain type of service – in the UK that included “free” services, to the detriment of later cost-controllers. No Australian government, and particularly no non-Labor government, had ever promised free health services the way British Labour did,139 and Australians always had to pay for some prescriptions, since before 1960 some common drugs were not included on the formulary. This made patient contributions an easier policy tool, and as is discussed further in chapter give, made it possible for Australian fees to increase at a faster rate and to a much higher level that the nominal charges present in British prescription services. The second step in health policy development, thirty years later This thesis argues that an incremental approach makes the adoption of other health services, after the implementation of the first priority, much more difficult and less likely. In Canada, pharmaceuticals’ low place on the policy agenda was self-reinforcing, and health policy development with regards to the addition of pharmaceuticals, homecare, or other additional services remains stalled, despite various proposals for reform in the 1990s and early 2000s. In Australia, however, a new Labor government was able to overcome the barriers to later policy  137. Minister of Health to Prime Minister, 13 June 1975. NAA. A5931, CL640. “Pharmaceutical benefits scheme – Policy. 1973-1975.” 138. See for example Cabinet item: Consideration of proposed cuts in public expenditure including the introduction of prescription charges. October 14, 1949. TNA. CAB 134/220/34, “Internal financial situation. EPC(49)111”; Meeting of Ministers on Prescription Charges: Meeting 1. 15 November 1956. TNA. CAB 130/121; William Ross to James Callaghan re: need to abolish prescription charges, 22 October 1964. TNA. BN 72/175, “Prescription charges. 1956, 1964”; Ministry of Health briefing note on Prescription Charges, 30 May 1968. TNA. CAB 151/136, “Prescription charges. 1968.” 139. The bald statements by Chifley in 1946 and 1949 saying which legislation had been passed with regards to health (McAllister and Moore 1991) can be contrasted with British Labour’s more expansive promises to “vigorously develop the health services” (in 1935), and to provide a “new National Health Service [where] there should be health centres where the people may get the best that modern science can offer, more and better hospitals, and proper conditions for our doctors and nurses” (1945). The 1950 election manifesto reminded voters that “Labour has honoured the pledge it made in 1945 to make social security a birthright of every citizen. Today destitution has been banished. The best medical care is available to everybody in the land” (Craig 1975).  99 development with the introduction of public medical insurance and free public hospitals in 1975, and then reintroduce it permanently in 1984 after a period of Liberal government retrenchment. Although the research question focuses on explaining the presence or absence of pharmaceutical programs, the argument that an incremental approach precluded the development of pharmaceutical insurance in Canada naturally raises a question of why Australia was able to introduce Medibank/Medicare so long after the adoption of the Pharmaceutical Benefits Scheme. This development suggests that while incremental policy development makes later steps more difficult, it does not make it impossible, and that major change in health policy can still occur. An initial reading of the literature on the introduction of public medical insurance in Australia suggests that taking these later steps was somewhat analogous to the decision to take a radical approach to health policy initially, in that it requires a combination of institutional authority, principled ideas, and electoral motivations to overcome barriers to additional services that develop over time. In Canada, the gap between the introduction of the first priority for policy development (hospital insurance) and the second (medical insurance) was only nine years, and the two services were closely linked in policy discussions about the order of priorities during the 1950s. The adoption of nation-wide medical insurance still required a fortuitous combination of principled ideas and electoral incentives in order for the federal government to overcome the barriers to its implementation and strike a deal with provincial governments. In Australia, the barriers to the introduction of additional health services was arguably much higher: when Medibank, Labor’s public medical insurance program, was first introduced in 1972, it had been more than two decades since the adoption of the PBS, and in that time, states had developed alternative institutional arrangements for hospital care, private health insurance had flourished, and a continuous period of Liberal government, with its principled preference for private, voluntary health insurance, had prevented the development of any influential, government sponsored reports or commissions that could provide competing ideas about universal public insurance. However, I suggest that when the Whitlam Labor government was elected in 1972, after twenty- three years in opposition, ideological and electoral conditions allowed it to clear these barriers, at least temporarily. Although a Liberal government was elected in 1975 and slowly retracted the new system, the original policy prompted sufficient public attention and support that when a new Labor government was elected in 1983, it was able to implement public medical and hospital insurance on a permanent basis. The evidence presented below to support these hypotheses is, at this point, based only on published accounts from those involved in this policy incident, the  100 secondary literature, and some public opinion data, but it does suggest that the three factors that support a radical policy approach at the beginning of a process of policy development (centralized authority, principled ideas and electoral incentives) may have a wider application to the difficult process of adopting additional services in countries that take an incremental approach to health policy development. Centralized authority As was the case with pharmaceutical benefits in 1945 (before the constitutional amendment), the Commonwealth government did not have centralized authority over health in a constitutional sense, and states maintained different systems for public hospital services. However, the Commonwealth government retained the financial supremacy discussed earlier. Gray argues that during the process of adopting compulsory public health insurance, “the role of the states was primarily reactive... The centralization of power made it relatively easy for successive federal governments to impose their will on what is, essentially, a set of financial arrangements” (1991, 132). Labor leader and later Prime Minister Gough Whitlam certainly did not see Australia’s federal institutions as a barrier to the policy change he wished to introduce. He argued that “In Australia, if a significant function is not financed by the national Government it will be unfairly financed, inadequately financed, or not financed at all” (Whitlam 1985, 3). Changes to medical insurance did not require input from the states, but ensuring universal free access to public hospitals required new funding arrangements between the Commonwealth government and the states, and these were complicated by the fact that at this time four of the six states had non-Labor governments. However, Gray (1991) argues that once the Commonwealth government had negotiated agreements on public hospitals with its co-partisan governments in South Australia and Tasmania, there were greater electoral and financial incentives for the remaining states to fall in line. The flow of financial benefits to South Australia and Tasmania was now apparent, and this likely influenced Queensland, which had long provided free public hospital care without receiving Commonwealth aid. Furthermore, the Commonwealth Minister for Social Security “expressed his opinion that New South Wales, Western Australia and Victoria were sensitive to the political costs of remaining outside the program and were now “quite keen” to proceed towards agreement” (Gray 1991, 141), since these “Liberal states could be seen to be withholding benefits from citizens” (Gray 1991, 146). Although its lack of exclusive constitutional authority over health meant the Commonwealth government had to negotiate with state governments to implement the hospital portion of its  101 insurance plan, its superior financial resources and ability to negotiate with more cooperative states first (and hence provoke electoral incentives for agreement in other states) meant that federal institutions were not major barriers to program adoption. Interestingly, the real difficulty for the original Medibank proposals was the lack of centralized authority within the national government, since Australia has a powerful elected Senate that was at this time controlled by the Liberal party. The Liberal Senate was able to delay the passage of Medibank legislation considerably, and it was only “ten weeks after Medibank came into operation in all states [that] the Fraser Liberal-Country Party government was elected to office” (Gray 1991, 147). This type of delay is much less likely to occur in Canada or UK, which have unelected and therefore less powerful Senates, but it confirms the expectation in the literature that fragmented authority may affect opportunities for significant change in a variety of institutional settings. Principled ideas The lack of completely centralized authority somewhat hindered the adoption of additional health services in Australia, but its effect was mitigated by financial and electoral incentives for states to cooperate. The second factor that supports radical change, a principled, cohesive idea about health policy, was a key policy driver. A radical expansion of public health insurance fit with current Labor ideology, since it “provided all Australians with adequate health treatment as a social right, rather than as a function of their income” (in line with a 1972 promise to “promote equality”) (Whitlam 1985, 337, 12), and reintroduced free treatment in public hospitals according “with traditional Labor party policy” (Gray 1991, 134). More importantly, Whitlam found a specific idea about the shape this expansion should take in the work of two Melbourne University economists. Whitlam records in his memoirs that, Although well aware of the inadequacies of the existing health insurance system, I was yet to develop a viable policy alternative on behalf of the ALP. The solution came in 1967 when Cass asked me to his home to meet John Deeble and Dick Scotton of the Institute of Applied Economic and Social Research at Melbourne University…Deeble and Scotton were preparing an alternative health insurance program which built upon the criticisms, identical to my own, that they had developed of the existing system. Medibank was conceived that night” (Whitlam 1985, 335).   In Scotton’s own account of that meeting, he recalls that although he and Dr. Deeble had been doing research on the “provision, cost and use of health services” which “prompted skepticism about the social outcomes of the voluntary [health insurance] system,” in 1967 their “ideas were still quite tentative” (Scotton and MacDonald 1993, 21, 24). However, they put them into writing at Whitlam’s urging, and by February 1968, presented a plan for compulsory health  102 insurance to Liberal government health officials. Scotton and Deeble’s plan was rejected by the government as impractical, but it joined a number of published articles “which Whitlam lost no opportunity to cite,” in bringing “the inadequacies of the voluntary insurance scheme into the limelight” (Scotton and MacDonald 1993, 24) and promoting the universal, compulsory alternative. Scotton and MacDonald note that in 1968, “the Pandora’s box which had contained public debate on health insurance was now well and truly open” (Scotton and MacDonald 1993, 25), and the greater salience of this cohesive, principled idea of how health insurance should work helped fuel electoral incentives for reform – the third important factor for radical change. Electoral motivations The electoral motivation to attempt radical health policy reforms developed in a number of ways. First, there was simply the worsening objective condition of the voluntary scheme: “by the mid-1960s the limitations of Australia’s voluntary health insurance scheme were starting to be felt. The financial growth of the [private] health insurance funds contrasted with the growing dissatisfaction with rising contributions and gaps in coverage” (Scotton and MacDonald 1993, 19). Whitlam also saw a practical social policy as an important electoral tool. He felt that Labor needed to become a more programmatic party, and upon taking leadership, “set out to formulate activist, reformist and thoroughly documented policies on all areas of significance,” including health (Scotton and MacDonald 1993, 19). Whitlam writes that Labor’s electoral defeat in 1963 convinced me that the Party was providing inadequate alternatives to the policies of the Menzies [Liberal] Government. New and attractive policy programs, meeting the aspirations of both the Party and the electorate, had to be developed if the Party was to regain government” (Whitlam 1985, 4).  Health reform was an attractive policy program because it clearly distinguished Labor from existing Liberal policies, fit with party ideology, as noted above, and involved a clear distribution of benefits to voters. After exposure to Scotton and Deeble’s ideas for compulsory health insurance helped focus Whitlam’s priorities on this topic, health reform assumed increasing salience in the late 1960s and early 1970s. Scotton and MacDonald note that after universal health insurance was adopted as Labor policy, it “rapidly assumed increasing prominence as an issue” in the 1969 election (Scotton and MacDonald 1993, 32). The policy was a “key aspect” of Whitlam’s election campaign policy speech, “in which heavy emphasis was placed on the high level of medical benefits, universal availability of free public hospital care and the statement that the cost would be lower than the existing scheme for ‘four out of every five families’” (Scotton and MacDonald 1993, 33). Although Labor lost this election, and rhetoric about compulsory  103 health insurance was less prominent in the 1972 election campaign, “ ‘free medical services’ were identified by 46.3 percent of respondents – more than any other item – as the most important single issue in a public opinion poll conducted in September 1972” (Scotton and MacDonald 1993, 51, citing Stubbs 1989). Compulsory insurance also became more popular during this time. A March 1969 poll found that most voters preferred to keep the existing voluntary insurance system rather than to increase income tax by 8 percent to fund a free medical and hospital program.140 However, in August 1969, Gallup reported that “The ALP [Australian Labor Party] promises most likely to swing voters from the Liberal-CP to ALP are free medical services and pensions without a means test,” and 58 percent of all voters favored a revised proposal for a free medical service with a 1.25 percent levy on taxable income.141 Whitlam saw public health insurance as a potential policy “winner,” in terms of electoral support, and its increasing importance to, and popularity with voters seemed to confirm this assessment. There is further evidence for the electoral motivations involved in universal public health insurance after it was implemented in 1975. A Liberal government was elected in December 1975, following a constitutional crisis where the Liberal Senate refused to pass supply bills and the Governor-General dismissed the Whitlam government. However, the new Fraser government promised to maintain Medibank (Gray 1991, 147), despite the fact that its “members were strongly opposed ideologically to Medibank, and its broad election policy involved a commitment to massive reduction of government expenditure” (Scotton and MacDonald 1993, 235). It appears that the same electoral dynamic that drove the Menzies Liberal government to implement the PBS in 1950 resulted in the Fraser government to promise to maintain Medibank, and its attempt to only undertake hidden, incremental retrenchment. According to Scotton and MacDonald, “Fraser’s position [on Medibank] was a logical response to evidence of the growing popularity of Medibank shown in public opinion polls,” (Scotton and MacDonald 1993, 236). Although the compulsory element was removed in 1976 and public benefits were gradually reduced, the popularity of the original scheme is evidenced by how easily it was reinstated when a  140. Gallup Poll (Australia), March-May 1969, Polls 2105-2118, “Keep medical funds voluntary”. Only 35 percent of voters overall preferred increased tax and a free system; 63 percent of Liberal-CP voters favored the voluntary system, compared to 47 percent of Labor voters. 141. Gallup Poll (Australia), August-September 1969, Polls 2132-2142, “Some ALP planks could be net losers”. There was an increase in the popularity of the compulsory option with both Liberal-CP voters (51 percent in favor) and Labor voters (68 percent in favor).  104 Labor government took power again in 1983. Whitlam sums up the electoral significance of Medibank as follows: Medibank…had been one of the decisive issues in the great campaign of 1969. It was crucial to our victory in 1972. Its rejection by the Senate Opposition contributed significantly to victory in May 1974…Fraser acknowledged its popularity by his specific promise to retain Medibank in 1975. Hawke won in March 1983 with a simple one-line undertaking: ‘We will restore Medibank’” (Whitlam 1985, 349).   Thus, the somewhat fragmented authority over hospital insurance was not a significant barrier at this time, and the presence of a cohesive, principled idea about health policy along with strengthening electoral incentives were key to the adoption of additional public health insurance policies in Australia. The design of these services was shaped by Australia’s long history of private insurance, as demonstrated by their less comprehensive nature than pharmaceutical benefits, but it was nonetheless a significant change in health policy. Conclusion Australia’s path of health policy development reinforces the crucial effect of sequencing: as was the case in Canada, the choice of initial priorities for the introduction of health services shaped policy outcomes for decades, with the first priority, pharmaceutical benefits, remaining the only public health insurance program for more than twenty years. In the case of Australia, later priorities were delayed because of a change in government, which may be interpreted as a more pronouncedversion of the lack of principled elite ideas or electoral motivations that stymied later priorities in Canada. Australia’s incremental development of health policy demonstrates again the importance of principled ideas in determining the choice between staged or simultaneous policy implementation: here the lack of a cohesive, principled idea within the Australian Labor party and Canadian Liberal party stand in clear contrast to the British Labour party. The Australian case also acts as a caution against taking a monolithic view of federalism’s effects: although Australia, like Canada, is a federation, this was not a key factor in the decision to proceed in stages. Australian states did not exercise a veto over the decision about how to proceed with health policy development.  A novel finding of this research was the reason for Australia’s choice of pharmaceuticals as the first priority for health policy development. It is not an obvious choice: it is generally acknowledged that hospital and medical services have a larger or at least prior impact on health, and they certainly were more important services at this time. However, Australian policymakers  105 were driven by pragmatic concerns about the availability of medical personnel during World War II, and conversely, the desire to introduce new social benefits before the end of the war. Pharmaceuticals appeared to be “low hanging fruit” to federal politicians in the early 1940s, although they quickly found that although state governments complied as expected, opposition from the medical association was much more problematic than anticipated.  Although pharmaceuticals were selected as a priority and the first federal legislation was passed in 1945, the Pharmaceutical Benefits Scheme was not in operation until 1950. The struggle over the implementation of the PBS confirms the importance of doctors’ preferences in shaping the eventual design of the program, as other authors have pointed out. Perhaps more importantly, it also demonstrates the role for public expectations in health policy implementation. The PBS was eventually implemented by a new governing party that had very different ideas about the proper role for government in providing social welfare, and although it was modified to better reflect the preferences of doctors, it was still a program that had been vigorously opposed by the medical association for five years. However, the PBS had also been a salient issue during these years, and as early as 1948, the Liberal party recognized the potential popularity and perhaps inevitability of the promise of “free” drugs, and incorporated pharmaceutical benefits into its platform. Thus, when the Liberals took power in 1950, they had a previous commitment to pharmaceutical benefits that ensured its implementation, even as they quietly dismantled grants to states for public hospitals and effectively stopped the development of additional public health insurance for twenty-three years.  The PBS implemented by the Liberal government in 1950 had one major difference from the original Labor scheme: a more limited formulary, where subsidized drugs were restricted to a list of “life saving and disease preventing” medicines that did not require compounding by chemists. This limited formulary was important in two respects. First, it allowed the Liberal government to secure doctors’ cooperation with the scheme. It responded to the BMA’s longstanding argument that a government formulary covering all commonly prescribed drugs would restrict doctors’ ability to prescribe individualized compounds as they saw fit, and the Liberal’s preference for voluntary private insurance versus public compulsory programs reassured doctors that participating in the PBS would not be the beginning of a wider program of “socialized medicine.” The second reason that the limited formulary was important was that it provided an opportunity to use a compensatory policy in 1960, when higher than expected expenditures forced the government to introduce cost control measures to the PBS.  106  In the UK, the same type of cost control policy, the introduction of patient charges, caused outcry both within the governing Labor party and from the public. The change was decried as going against the founding principles of the NHS, which promised a free, universal, and comprehensive service. In Australia, there were no such principles to reference. The PBS was not adopted as part of a cohesive, principled plan for health services, and was in fact implemented by a government that was otherwise opposed to the expansion of public health insurance. However, the program was salient and popular enough that the Liberal government felt electoral incentives to implement it in 1950, and expressed concerns that implementing patient charges would be an unpopular move in 1960. It would be interesting to attempt to measure the differences in rhetoric and opposition to prescription charges in Australia and the UK in the absence of an opportunity for compensation by the Australian government, as the best way to measure the effect of early principled ideas on program flexibility. However, Australia’s limited formulary meant that policymakers could combine patient charges with formulary expansion, so patients paid a small fee for all prescriptions rather than receiving listed drugs for free, and paying full price for others. Politicians explicitly discussed the benefits of this strategy in terms of managing public opposition, and it appears to have been successful, as charges were introduced with little controversy. The fact that Australia has been able to continue to increase its patient charges without controversy, despite the fact that no further compensatory strategies were available, while in the UK increases are much more controversial, despite a broader system of exemptions and lower fees overall, suggests that early principled ideas may play a longer term role in affecting program flexibility, but this question requires further research beyond the limits of this project’s time period.  Finally, a preliminary examination of Australia’s much later adoption of additional health services, with public hospital and medical insurance introduced on a permanent basis in 1984, demonstrates that certain factors may allow governments to overcome the barriers to additional services in an incremental process when three conditions exist: centralized authority, cohesive, principled ideas, and electoral motivation. Of course, these are the same conditions that supported a simultaneous approach to health policy development at earlier in the policy process, so there may be important similarities between opportunities for radical change at different points in policy development. This hypothesis requires additional cross-national data for testing, but suggests that the conditions set out here for radical versus incremental policy development may have a quite general application.  107 5 The United Kingdom: “something bolder than a mere extension and adaptation of existing services”142 Introduction The United Kingdom differs from Canada in that it offers nation-wide, universal and comprehensive prescription services. It also differs from Australia, because it adopted prescription services as part of a “big bang” approach to health policy development: all services were introduced simultaneously under the National Health Service in 1948. Thus, the UK represents an alternate path to pharmaceutical coverage: it may be achieved as an early priority in an incremental approach, as was the case in Australia, or as part of the simultaneous introduction of health services. The chapter is organized as follows: first, I explain the choice of simultaneous policy development with reference to the centralized authority over health policy in the UK, the high degree of principled commitment to the idea of comprehensive and universal health services by the Labour government, and the strong electoral motivation for the government to act on this idea. Next, I examine the how the UK responded to the early challenges of its generous prescription services, which were subject to significant new cost concerns very soon after adoption. I study the first policy response, which was to impose a charge on prescriptions, and which continued to be a contentious policy issue for the remainder of the study period. This section also demonstrates the distinctiveness of pharmaceutical politics relative to other health services. The cost of prescription services rose more quickly than the cost of the NHS overall, and initially seemed to politicians to present a less painful venue for cost-cutting measures: some of the increase in cost could be linked to misuse of the system, and policymakers were very concerned with the possibility of overprescribing at this time. Furthermore, although the programs in the UK and Australia were structured quite differently and implemented under different circumstances, prescription services in both countries were subject to the same types of pressures and challenges shortly after their introduction. Initially, there was some similarity in the two countries’ response to these challenges, as both sought to introduce patient charges or copayments for drugs. However, in the UK, policymakers’ scope for introducing this type of cost control were more limited, as a consequence of the way prescription services were introduced and the type of public expectations they produced.  142. Labour Health Minister Aneurin Bevan, Memorandum: Proposals for a National Health Service, Minister of Health, 13 December 1945. The National Archives (UK). CAB 129/5.  108 As noted in the methods chapter, I discuss the United Kingdom as a unitary actor with regards to health policy, but since devolution in 1997, regional governments in Scotland and Wales have made significant divergent decisions with regards to prescription services, particularly with the abolition or planned abolition of prescription charges.  However, since the current project covers the early 1940s to the mid 1970s, the findings apply to the entire United Kingdom. Historical overview Although there were discussions of health policy reform in the UK prior to World War II, the earliest and most influential report espousing a universal and comprehensive health service was the report on Social Insurance and Allied Services by Sir William Beveridge, published in December 1942. This report did not mention pharmaceuticals, but called for “a comprehensive medical service for every citizen, covering all treatment and every form of disability” (Great Britain. Inter-departmental Committee on Social Insurance and Allied Services 1942). In February 1944, the wartime coalition government (consisting of Conservative, Labour and Liberal members) tabled a White Paper (A National Health Service) that represented a compromise position on health policy reform, although it maintained the principles of universality and comprehensiveness. Before any policy was adopted, however, Labour won a landslide majority government in the general election of July 1945, and proceeded to prepare legislation for health reform that modified some of the compromise positions of the 1944 White Paper, particularly with regards to hospital administration. The National Health Service legislation was passed in 1946, and services were implemented in 1948. Prescription charges were proposed the following year in response to higher-than anticipated costs for pharmaceuticals and concerns about the cost of the NHS overall. The first prescription charges were implemented in 1952, abolished in 1964, and reinstated in 1968. They continued to be a major source of political controversy until the late 1970s, when a small annual increase to the charges was made automatic (Wordsworth et al. 1998, 196). Although charges still occasionally reach the public agenda and prompt renewed controversy, for the most part they are less salient today, perhaps in response to what one interviewee noted was policymakers’ general sense that “if you kick this particular sleeping dog, you are going to be in trouble.”143 This echoes  143. Author’s interview with Dr. Rudolph Klein, academic and British health policy expert, London, 1 May 2008.  109 a much earlier conclusion by a Treasury official, who reviewed a report on prescription charges in 1966 and observed, “I imagine the question of charges will always be political dynamite.”144 During the 1950s and 1960s, the UK also began to experiment with drug price control, introducing the first Voluntary Price Regulation Scheme (later called the Pharmaceutical Price Regulation Scheme, an agreement with the pharmaceutical industry) in 1957. However, as was the case in the Australia chapter, I omit a detailed discussion of this change to pharmaceutical policy. Adjustments to the VPRS occurred throughout the remaining study period and into the current day, but unlike prescription charges, they did not have a direct or immediate effect on the universality or comprehensiveness of the program. They are also unsuitable for testing the causal effect of initial program conditions on policy implementation and reform, since a new causal factor (the influence of the pharmaceutical industry) became a major player in this policy choice. Choice of simultaneous approach to policy The National Health Service (NHS) was a groundbreaking achievement in welfare state development, providing universal coverage for a comprehensive range of health services that was “free at the point of use.” Although there were important elements of continuity with previous policies,145 it represented a significant innovation in terms of the population covered, the range of services included, and the mechanism for coverage, moving from an insurance principle to a nationalized service. There is a range of excellent scholarly literature that seeks to explain why the UK was able to introduce this radical policy change (see for example Hacker 1998; Klein 1995; Pater 1981). As discussed in chapter one, I base my explanation on Touhy’s (1999) argument about windows of opportunity that occur only when “the political system…provide[s] a consolidated base of authority for political action,” and when there is “an extraordinary mobilization of political authority and will” (Tuohy 1999). I look for evidence of centralized authority in institutional terms, a principled idea on the part of policymakers, and strong electoral incentives, and contrast the UK situation with the absence of one or all of these factors in Canada and Australia.  144. Letter from O. L. Williams, 1966. The National Archives (UK) (hereafter cited as TNA). T 227/2522 “Treasury history of prescription services: social services aspects. 24 February 1966 – 31 May 1967.” 145. Hacker (1998) in particular emphasizes links to the National Health Insurance Act of 1911, discussed below. Furthermore, an early NHS planning document notes that establishing a universal and comprehensive health service is not “a question of a wholly new service, but of one with many roots already well-established” (Ministry of Health 1944)  110 This section also demonstrates that the specific politics of pharmaceuticals were not a factor in their inclusion in the NHS. Instead, they were part of a broad effort to provide comprehensive health services, and the particular challenges of prescription services only became apparent to British policymakers after the NHS had been introduced.  Centralized authority Unlike Canada, policymakers in the UK did not have to contend with other levels of government as veto players, as the unitary system ensured that the national government had the final say over health policy development. The Australian case demonstrated that this can be a financial, as well as jurisdictional, dynamic: the Australian Commonwealth government had strong constitutional powers with regards to grants to states, and state governments’ weak financial position meant they were unable to challenge the national government’s health policies effectively, despite having fairly clear constitutional authority over health. Besides the centralization of authority provided by a unitary system, however, there was the centralization of power inherent in the UK’s Westminster parliamentary system, where responsible government and strong party discipline provides cabinet with a high degree of policy control (Weaver and Rockman 1993). This effect was heightened in the UK after 1945, when a “landslide upset” election (Jacobs 1993) moved the country from the wartime coalition government to a Labour majority government. The importance of centralized authority to significant health policy reforms is evident in decisions that preceded the adoption of the NHS: Hacker (1998) finds that the supremacy of the UK’s national parliament played an important role in the passage of its first major piece of health insurance legislation, the National Health Insurance (NHI) Act of 1911, which provided limited health insurance for workers though not for their dependents, as well as the later NHS. He notes that the NHI was possible in part because “Britain has a national administrative state that…was fully capable of taking on significant administrative tasks and faced no serious challenge from local governments” (Hacker 1998, 88), conditions that were not present in Canada (or the US, as he also argues). The role of centralized authority becomes more apparent when comparing the language of the Canadian federal government in the early days of health policy development with the UK’s Ministry of Health. In 1946, Canada’s Cabinet Working Committee on Health Insurance recommended that, “further consideration of the second stage of the Health Insurance Proposals  111 be deferred pending the outcome of the reports from the provinces regarding planning and organization.”146 This deference to other levels of government was not present in the UK: the Ministry’s 1944 White Paper recognized the risks of a single, national health service when it discussed “the danger of over-organisation,” but argued that a nationalized service was necessary because “medical resources must be better marshaled for the full and equal service of the public, and this must involve organisation – with public responsibility behind it” (Ministry of Health 1944, 8). A year later, the Minister of Health noted that when the nationalization of hospitals was approved there were some concerns about “the risk of losing from the health service the benefits of local interest and local knowledge in day-to-day administration,”147 but it was determined that the benefits of a single national service outweighed these risks, and local governments were not in a position to fight back. Perhaps the most compelling evidence for the role of centralized authority is the fact that during the design and adoption of the NHS, national policymakers recognized that certain elements of its organization would be contrary to the interests of local authorities, but were able to easily overcome this and to introduce the type of program favored by the central government. Most of this controversy focused on the nationalization of hospitals, which were previously the responsibility of local authorities. The first design for the NHS was put forward in a 1944 White Paper by the wartime coalition government. Klein notes that this plan involved “compromise proposals designed to reconcile multiple and conflicting objectives….[which] left most of the actors involved feeling dissatisfied” (Klein 1995). For example, the White Paper called for joint authorities (made up of counties and county boroughs) to be responsible for “securing all the hospital and consultant services covered by it” where previously this had been the responsibilities of smaller local authorities (Ministry of Health 1944, 11). Although “the local government lobby had averted the danger of nationalization, …[it] was left dissatisfied by the prospect of the new joint authorities actually taking over control of their hospitals” (Klein 1995). When Labour took power they abandoned some of the compromise positions of the 1944 White Paper and nationalized hospitals, which “flew against the interests of local authorities,” but was supported by the powerful consultants, or specialists physicians who staffed these hospitals  146. Meeting of the Cabinet committee on Dominion-Provincial Relations: Report of the Working Committee on Health Insurance, 4 January 1946. Library and Archives Canada. Brooke Claxton fonds. MG 32-B5  Vol 138 File: “Dom-Prov Conferences DP-2 Committees and Meetings.” 147. Minister of Health, Memorandum: Proposals for a National Health Service, 13 December 1945. TNA. CAB 129/5.  112 (Tuohy 1999). In 1946, the Conservative Health Spokesman told the House of Commons that although his party had the same goals for a national comprehensive health service as Labour, they could have achieved it more efficiently “if Bevan [the Labour Minister of Health] had not been determined to sweep away the voluntary hospitals, to weaken the structure of local government by removing one of its most important and vital responsibilities…” (Powell 1994, 144, citing Hansard, April 30). It is interesting to note that the capacity to take a more centralized and unilateral approach to health policy existed before Labour came to power, but the coalition government was not willing to act on it, suggesting the importance of a new ideational and electoral commitment to health services. Principled ideas and electoral motivations Although centralized authority was a key prerequisite for the simultaneous development of health services, enabling the national government to take decisive action on its preferred plan, it was also important that the government have a principled commitment to the idea of comprehensive, universal services, and electoral motivation to act. There was a mixture of strategic and principled reasons for the post-war government to place a high priority on the adoption of the NHS, but the key element of this decision from the point of view of prescription services is that it occurred in a situation where the comprehensiveness of health services (i.e. range of services provided) was taken for granted by all actors. Thus, although there were controversies over issues such as the method of hospital administration or doctors’ remuneration, even before Labour took power in 1945 there was what Klein calls a “remarkable…shared assumption that the health service should be both free and comprehensive” (Klein 1995, 24). This meant that the earliest archival records of pharmaceutical policy discussions, from 1943 and 1944, focus on questions of how prescription services were to be delivered, not whether they should be included or where they fell in the order of priority for services, as was the case in Canada and Australia. For example, in April 1943, an internal Ministry of Health memo discussed a system for dispensing prescriptions and the need to educate doctors about excessive prescribing, and later that summer, the Health Minister discussed pharmacists’ participation in the service with the Pharmaceutical Society of Great Britain. A Ministry of Health memo in March 1944 provided estimates of the drugs bill and suggested that it would “be worth while to set up some considerable machinery to ensure that we get what we are paying for.”148  148. Memorandum, March 1944. MH 77/120 TNA.“Formation of Policy under National Health Service Act/ Pharmaceutical services. 1944 March-1946 November.”  113 The release of the Beveridge Report in 1942 provided a powerful, cohesive idea for health policy development in the UK that set out both the goals for a national welfare state and the type of health policy instruments that might help achieve them. A review of previous health services reports in the 1944 White Paper notes that “the principles most frequently reoccurring in the presentation of plans for the future developments are the following: (1) that there should be made available to every individual in the community whatever type of medical care and treatment he may need; (2) that the scheme of services should be a fully integrated scheme” (Ministry of Health 1944, 76). The Beveridge Report took these principles as a basis for its recommendations, and described “comprehensive health and rehabilitation services” as one of three key assumptions (along with children’s allowances and the maintenance of employment) for its proposed program of social security and “freedom from want” (Great Britain. Inter-departmental Committee on Social Insurance and Allied Services 1942, 158). The Beveridge Report was both very prominent and broadly supported by the British public: Jacobs reports that a special Gallup poll conducted the month of its release found that “Fully 95 percent of the public had heard about the Beveridge Report,” and “88 percent of respondents favored its implementation” (Jacobs 1993, 113). This high degree of public support and, as Jacobs (1993) argues, government’s attentiveness to public opinion polls, provided incentives for government action on health policy, and the government announced in February 1943 that it accepted Beveridge’s assumption “that a comprehensive national health service, for all purposes and for all people, would be established” (Ministry of Health 1944, 76). However, the coalition government had some difficulty agreeing on the precise form this action should take. The 1944 White Paper refers to a consensus on the need for broad public health services, but says, “There is much agreement of what the aim should be, if not on the method of achieving it”(Ministry of Health 1944, 8). This disagreement on methods became moot after the 1945 election: Klein argues that Labour’s landslide win meant that, “The fragile consensus which had constrained Labour Ministers in the Coalition Government appeared to be shattered. The way was open for the politics of ideology to take over from the politics of compromise” (Klein 1995). Certainly the Labour election manifesto emphasized health, saying, “the best health services should be available to all. Money must no longer be the passport to the best treatment” (Craig 1975). Although the Conservative manifesto also promised “a comprehensive health service covering the whole range of medical treatment” (Craig 1975), Jacobs notes the perception that the Conservatives were “lukewarm” on the recommendations of the Beveridge Report and cites  114 analysis of polls and election results which “suggests that Labour was uniquely identified with social welfare reform and greater state involvement” (Jacobs 1993, 169). The literature provides slightly different interpretations of the electorally motivated or principled nature of Labour’s actions. Some authors emphasize the public’s expectations for services stemming from dissatisfaction with the old limited National Health Insurance arrangements, which promoted “widespread popular support for free and universal health care” (Hacker 1998, 93), and the prominence and widespread support for the Beveridge Report (Jacobs 1993, 113). Jacobs argues that the 1945 election “provided indisputable evidence of public support for innovative legislation; this perception created nearly irresistible incentives for policymakers to pursue reform” (Jacobs 1993, 167). In 1948, the Gallup Poll of Britain noted that “The Health Service is scheduled to start July 15 next, and [pollsters] ascertained that over six in every ten adult Britons were sufficiently interested in the service to know this date,” and 61% reported that they felt the new health service was a “good thing,”149 so there is clear evidence of electoral motivation for bold action on health policy. However, there is also evidence of a principled component to Labour’s commitment. In 1937, future Prime Minister Clement Attlee wrote about the Labour Party’s preference for radical action, saying, “The Labour Government will not dissipate its strength when returned to power by dealing only with minor matters. It will proceed at once with major measures while its mandate is fresh” (Attlee 1937, 176). He added that, in terms of priorities for action, “Labour does not intend to delay the introduction of measures calculated to effect an immediate improvement of a far-reaching character in the social services” (Attlee 1937, 192). A review of Labour thought by Bealy highlights the party’s longstanding commitment to “Universality in social services,” which “was affirmed [in the 1930s] in the repetition of the demand for a National Health Service, free secondary education for all up to the age of sixteen, and most importantly, the institution of a campaign for family allowances…” (Bealy 1970, 17). Tuohy also attributes more importance to Labour’s principled policy preferences: “Notwithstanding the fact that the 62% majority of seats [in Parliament] had been achieved with less than a majority (47.8%) of the popular vote, the Labour government interpreted the results as providing it with a broad mandate for change” (Tuohy 1999, 39). Other authors emphasize the values and charisma of the Labour Minister of Health, Anuerin Bevan, who was a powerful cabinet member and key decision maker in health policy (Klein 1995, 13; Webster 2002, 13), and  149. BIPO/Gallup Poll, 17 May 1948.  115 argue that for Bevan “and for many others”, the idea of a free health service “represented the embodiments of a pure Socialist ideal”(Ryan 1973, 219). Besides the commitment demonstrated in election manifestos, there is also evidence of commitment in more private forums. In a 1945 Memorandum to Cabinet, Bevan argued for major reforms: “As I see it, the undertaking to provide all people with all kinds of health care…creates an entirely new situation and calls for something bolder than a mere extension and adaptation of existing services.”150 Bevan further demonstrated his commitment to the NHS’ founding principles in 1949 by opposing the new legislative powers to impose charges for certain services,151 and in 1951 by resigning from cabinet temporarily when prescription charges were first introduced, arguing that this represented “the beginning of an avalanche” eroding NHS principles (Ryan 1973, 225). It should also be noted that Labour was, in part, elected in 1945 because of their ideas, since the public found them more credible on health policy than the Conservative party (Jacobs 1993, 169). Furthermore, Labour ideas about health policy predated voters’ ideas: they were in favour of a universal, comprehensive system before the Beveridge Report ignited public opinion. This provides evidence that principled ideas were important prior to electoral motivations. Although it is not possible to say whether Labour would have taken radical action without electoral motivations, the two factors appear to reinforce one another in important ways.  In the 1940s, the UK had three factors that were completely or partially missing in Canada and Australia. The government had centralized authority over health, principled policy preferences with regards to a previously established idea of universal and comprehensive services, and there was strong public support for this idea. This along with a legacy of previous health insurance promoted “the continued growth in the British public’s acceptance of, and demand for, state involvement in the delivery and finance of medical care” (Hacker 1998, 92, emphasis added). These ideational and electoral pressures for the introduction of a broad public health service, combined with institutional centralization, resulted in the simultaneous adoption of hospital, medical, pharmaceutical, and other auxiliary services (notably dental and ophthalmic) in 1946, and their implementation in 1948. However, it was not until these services had been in operation for a year or so that policymakers began to recognize the specific challenges posed by  150. Memorandum: Proposals for a National Health Service, Minister of Health, 13 December 1945. TNA. CAB 129/5. 151. Chancellor of the Exchequer, Memorandum EPC(49)111: Consideration of proposed cuts in public expenditure including the introduction of prescription charges. 14 October 1949. TNA. CAB 134/220/34, “Internal financial situation.”  See also Ryan (1973, 224)  116 prescription services and the need to devise management strategies that might impinge on the NHS principles of “free,” universal and comprehensive services. The early challenges of prescription services Pharmaceuticals were not mentioned specifically in the Beveridge Report, despite its emphasis on the need for a complete range of health services and its receipt of submissions from both the Pharmaceutical Society of Great Britain and the National Pharmaceutical Union. Pharmaceuticals appear briefly in the 1944 White Paper, which states that arrangements for dispensing will be discussed “with appropriate pharmaceutical bodies” (Ministry of Health 1944, 37). However, policymakers were clearly planning to include pharmaceuticals. Archival records show extensive discussions of which health professionals would be allowed to dispense prescriptions in which types of settings during 1943 and 1946;152 consideration of pharmacists’ remuneration in 1947;153 and a somewhat surprising amount of debate over the issue of medicine bottles in 1947: due to a shortage of glass bottles (and their expense), pharmacists wanted patients to buy or supply their own bottles for prescriptions, while the Ministry of Health was concerned that this would make the service look “rather squalid to the middle class patient.”154 By 1943, policymakers took the comprehensiveness of the NHS for granted, and so the inclusion of pharmaceuticals was also taken for granted, but there was still some effort prior to the implementation of the NHS to consider the specific problems of administering a prescription service. As the remainder of the chapter will argue, the distinctiveness of these problems became clear to policymakers almost as soon as the service was implemented, leading to conflict over the need to control costs while maintaining the principles of the original program. In the Australian case study, I examine an early implementation decision, to introduce patient charges and significantly expand the formulary in 1960, because it had an important effect on the scope of coverage. Although policymakers’ main motivation was to introduce better measures for cost control given the higher-than-expected expenditures in the program’s first decade, these policy changes also made the Australian PBS a comprehensive program for the first  152. Memorandum M.U. Wilson to Dr. Whiting, 19 March 1943. TNA. MH 77/120; General Medical and Pharmaceutical Services Regulations, National Health Service Bill 1946. TNA. MH 137/75 “Legislation and Regulations/General medical and pharmaceutical services: correspondence and draft regulation 1946-1947.” 153. Memorandum 16 April 1947 (preparation to discuss new Health Service with pharmacists). TNA. MH 135/576 “Joint Committee on a National Pharmaceutical Service: correspondence/negotiations with pharmacists 1947-1953.” 154. Agenda of meeting of the Joint Committee on a National Pharmaceutical Service, 17 June 1947. TNA. MH 135/576.  117 time. Here, I examine a parallel policy decision in the UK: the introduction of patient charges in 1952, and their subsequent increase, abolition and reinstatement during the 1950s and 1960s. Because of the different circumstances around the adoption of UK prescription services, there was no opportunity to cushion the introduction of charges by expanding the range of subsidized drugs at the same time, as was the case in Australia. This, combined with the different rhetoric that accompanied the NHS’ adoption, made UK prescription charges much more politically contentious than charges in the Australian PBS, and placed different limits on the types of policy changes that politicians could introduce. Concerns about the cost of prescription services Although prescription services were introduced as part of a broad health scheme, they raised distinct political problems. Initially, any prescription drugs and many appliances and bandages were provided free of charge, and there were no policies for pharmaceutical management, but escalating costs quickly demonstrated that this was a mistake. The problem was not entirely, or even primarily, related to policy design, however: during the later 1940s until the late 1960s, pharmaceuticals were undergoing what is often termed a “therapeutic revolution,” as the mass- screening of chemical entities for therapeutic properties resulted in a large number of new, effective drugs, and drug companies began patenting synthetic versions of these naturally- occurring substances (Morgan 2008). Costs would have escalated even in the absence of a universal program (as was the case in Canada), but the original design of NHS prescription services did not account for these problems. In the first nine years the NHS was in operation, the total cost of prescriptions doubled (Ministry of Health (UK) Central Health Services Review 1959, 24). Concerns about the overall cost of the NHS led to the early introduction of prescription charges as a strategy for recouping revenue and limiting demand. Despite the fact that the low fees, and later the introduction of widespread exemptions, meant that prescription charges were never particularly lucrative (Klein 1995, 35), and the fact that in 1959 the government’s Hinchliffe Committee on the Cost of Prescribing expressly recommended against their continued use in 1959,155 charges persisted and increased throughout the study period, always accompanied by some degree of political controversy. Klein (1995) notes that there was strikingly little analysis of the financial implications of the NHS before its adoption, but there was at least some consideration of the cost of drugs. In 1944,  155. The committee concluded that “the present prescription charge is a tax which, beside stimulating the wrong incentives, has proved disappointing financially” (Ministry of Health (UK) Central Health Services Review 1959, 11).  118 a Ministry of Health memo estimated that the drug bill would be “somewhere about £10, 000, 000 and it will, therefore, be worth while to set up some considerable machinery to ensure that we get what we are paying for” (in terms of the quality of drugs).156 The following year, Health Minister Anuerin Bevan told cabinet that although “only very approximate estimates are possible of the cost to public funds of the comprehensive health service…in the early years of the service the annual expenditure might be £145 million,”157 so drugs would account for only about 7 percent of this estimate: in 2009 pounds, the total estimated cost of the NHS in 1945 was about £4.6 billion, while the drugs bill was estimated at £325 million.158 However, in December 1948, after less than a year in operation, Bevan reported to cabinet on cost overruns in the NHS, where the total gross cost was currently £225 million (£6.2 billion in 2009 pounds), saying, “That, then is the cost of social innovation. The justification of the cost will depend upon how far we get full value for our money.”159 Although the proportion of drug costs in this higher total was not reported for 1948, the gross cost of prescriptions dispensed in 1949/50 was £31, 674, 000 (or £853, 893, 592 in 2009 pounds, more than three times the budget predicted in 1944, and about 14 percent of the total cost of the NHS the previous year), (Ministry of Health (UK) Central Health Services Review 1959, 24), so it appears that the Ministry of Health was not particularly successful at estimating NHS costs at this time, even when it attempted to do so, and the actual cost of prescription services was likely unexpected. Since the limited National Health Insurance program of 1911 covered prescription drugs for workers, we might expect that policymakers would have learned from this previous experience. The 1959 Hinchliffe Committee report did refer to the influence of the NHI drug coverage when it considered current problems with the cost of prescribing, saying that under the NHI, “it was found necessary at an early stage to take administrative action to control costs” (Ministry of Health (UK) Central Health Services Review 1959, 15). These actions included prohibitions on doctors prescribing “non-drug preparations,” educational materials in which prescribers were “urged to prescribe with economy,” and financial penalties for individual doctors who were found to be prescribing excessively (Ministry of Health (UK) Central Health Services Review  156. Memorandum 31 March 1944. TNA. MH 77/120 157. Memorandum: Proposals for a National Health Service, Minister of Health, 13 December 1945. TNA. CAB 129/5. 158. All currency conversions use the UK GDP deflator series available from the International Monetary Fund’s International Financial Statistics, available at GDP deflators for 1944-1947 constructed from historical inflation rates available at 159. Minister of Health, Memorandum: Cost overruns, 13 December 1948. TNA. Cab 129/31.  119 1959, 15). Although prior to World War II, the prescribing environment was quite different as there were very few effective drugs (Rivett 1998, 55), it is curious that the lessons of the earlier restricted plan for prescription services were not applied more systematically to the design of NHS prescription services. The apparently limited learning over the course of these two policies provides additional evidence for the importance of principled ideas, since the desire to adopt free, universal and comprehensive health services, including pharmaceutical coverage, may have trumped previous experience with the problems of pharmaceutical expenditure increases. Moreover, even those lessons noted by the Hinchliffe Committee all emphasized physician’s prescribing behavior as the key to cost control. Politicians were reluctant to impose regulatory limits on physician’s clinical autonomy in this regard,160 and the result was that NHS prescription services were introduced without any policies in place to control expenditure. Shortly after the NHS was adopted, however, the government introduced patient charges for prescriptions as a method of cost control. As will be discussed below, physicians’ behavior turned out to be a second major policy challenge, after the initial shock of higher than expected drug usage and rising prices, as doctors proved quite adept at avoiding or minimizing prescription charges on behalf of their patients. As the Hinchliffe Committee reported, the charge is regarded by patients as a tax and…it stimulates all the instincts which taxes usually arouse in ordinary people, viz. to avoid paying more than they can help and to obtain as much as possible for what they have to pay. At the same time the doctors’ representatives have told the Committee that they regard the charge as a tax on illness and old age (Ministry of Health (UK) Central Health Services Review 1959, 89). Prescription charges, 1952-1956: accounting for public expectations I hypothesized that a simultaneous approach to health policy development would make programs less flexible once adopted, since they were introduced under conditions of principled ideational commitment from politicians, and high levels of attention and support from the electorate. This means that any change to an existing program that threatens its key characteristics (in the NHS, a comprehensive range of free and universal services) should prompt backlash from those politicians who had strong preferences for the policy and, perhaps more importantly, from the public who had developed strong expectations about the type of services to which they were entitled. The introduction of prescription charges in the UK provides evidence to support this hypothesis, especially when considered alongside Australia’s different experience  160. The Hinchliffe Committee’s insistence that “no restriction should be imposed on the doctor’s right to prescribe whatever drugs he considers to be proper and necessary for his patients” (Ministry of Health (UK) Central Health Services Review 1959, 7) is typical, despite longstanding efforts to educate physicians about economical prescribing.  120 with the same type of policy change. Although they were clearly in opposition to the NHS principle of free services and had the potential to impinge on the comprehensiveness of prescription drug coverage, charges were introduced successfully. It is clear that prescription charges were adopted in response to financial pressures, but it is less clear whether policymakers saw charges primarily as a revenue-raising opportunity, to offset the high cost of prescription services, or as a method of reducing demand by discouraging “excessive” resort to doctors and medicines, or as fulfilling financial requirements through some other mechanism. A Treasury report on the history of prescription charges, completed in 1968, notes a range of possible reasons for charges, from “general disinflationary measures” to “concern over increase in drug bill” to “prevention of “abuse”.” However, it finds that charges were not a particularly effective response to any of these problems, and goes on to say that “It never seems to have been clear precisely what officials believed to be the situation at any time. The drug bill was large and was rising, but whether this was caused by ‘abuse’ was never properly investigated.”161 This makes it hard to know what the real objectives were, but even if they were as publicly stated, “then it seems that the relationship between stated objectives and action to secure objectives was too tenuous.” The authors of the Treasury report conclude that It seems as if the Treasury, having assumed the charges were good without further analysis, concentrated merely on tactics, i.e. waiting, on each occasion, till it was thought that ministerial opinion was ripe for a further attempt [to get charges]. Thus miscalculations took place on each occasion.162  Some justifications for the charge, such as the prevention of abuse, might be thought of as a principled policy motivation, since it implies policy reforms in order to make the original universal and comprehensive system sustainable. However, regardless of the primary motivation behind the introduction of charges, they clearly clashed with the principle of a “free” service, and this was the main source of public backlash and governmental concern. Implementing this change aroused much greater opposition than the same policy did in Australia, and prescription charges continued to be a politically controversial issue in the UK for much longer than in Australia, while remaining significantly lower and more narrowly applied. Prescription charges had been on the policy agenda in at least some parts of government since the NHS was implemented. The 1968 Treasury report notes that in 1948, “The size of the estimate [for prescription charges] alarmed Treasury officials so much that they began almost at  161. Treasury history of prescription charges: S.S. aspects. 24 February 1966 – 31 May 1967. TNA .T 227/2522. Emphasis added. 162. Ibid.  121 once to propose a prescription charge.”163 Although policymakers saw the introduction of charges as financially necessary and a less painful option than alternatives such as cutting back services,164 they were still quite concerned with the public’s reaction, given the continued popularity of the free, universal NHS: in 1950, the Gallup Poll of Britain reported that 71 percent of respondents thought that the nationalization of medical services had been good.165 When prescription charges were first proposed in the House of Commons in 1949, it was as part of the Labour government’s proposals to deal with the difficult postwar economic situation. Prime Minister Clement Attlee proposed prescription charges as the last of a list of expenditure reduction items, but still linked the measure to “good policy” motives of preventing abuse of the system rather than economic necessity, saying “The purpose [of the charges] is to reduce excessive and, in some cases, unnecessary resort to doctors and chemists…resultant saving will contribute about £10 million, although this is not the primary purpose of the charge…”166 However, the Treasury report notes that although Attlee justified the charge in light of evidence of excessive prescribing, “this ‘evidence’ is not available in Treasury papers.”167 When prescription charges were debated in cabinet in 1950, policymakers were also attentive to the potential for backlash. The Chancellor of the Exchequer argued that if charges were to be implemented, government must immediately try to get the medical professions (and particularly doctors) onside since “they would be quick to see the fruitful possibilities of working up an agitation, among the public no less than among themselves.”168 The Chancellor’s concerns proved to be well-founded, as doctors began to find ways to subvert the charges almost as soon as they were implemented by a Conservative government in 1952. The charges were originally levied at one shilling per prescription form, equal to about £1.13 in 2009 pounds and which at  163. Ibid. 164. Chancellor of the Exchequer, Memorandum: Ceiling on expenditure, 29 March 1950. TNA. CAB 129/39. See also Minutes of meeting with BMA, 31 October 1956. TNA. MH 135.79. Here, the Conservative minister of health (Dame Enid) explained that “When it was decided that money should be raised from the Health Service, it was felt that the least harmful way would be from the Drug bill.” The BMA rejected this explanation and noted that it “considered the whole matter most distasteful.” 165. BIPO/Gallup Poll, 13 November 1950. 166. Prime Minister Clement Attlee “The Economic Situation (Government Proposals).” 24 October 1949. H.C. Debates 1948-1949 Vol 468, c.1016-1019. 167. Treasury history of prescription charges. 24 February 1966 – 31 May 1967. TNA .T 227/2522. 168. Chancellor of the Exchequer, Memorandum: NHS: charge for prescription, 26 January 1950. TNA. CAB 129/38.  122 that time was about equal to the cost of two loaves of bread,169 but if a patient required multiple prescriptions at one doctor’s visit, these were typically written on one form. A Ministry of Health circular explaining charges to doctors said “Doctors are asked not to increase either the number of prescriptions on one form or the quantities ordered in one prescription,”170 but it appears that this direction was not taken to heart by physicians. A letter from Boots Drugs Co. to the Ministry in June of 1952 complained that in order to minimize charges, doctors were prescribing potentially dangerous quantities of drugs in a single prescription, and putting all prescriptions for a family on a single form so it was not clear who should get which medicine.171 Perhaps in part because of doctors’ avoidance, and perhaps because excessive prescribing was not as severe problem as policymakers assumed, Treasury found that after the introduction of the charge in 1952, the reductions in the number of prescriptions were “nowhere near the 10% reduction apparently expected.”172 This led to a change in how prescription charges were administered. In 1956, a Conservative government moved from a shilling charge per prescription form to a shilling charge per item, effectively increasing the charge for people taking more than one drug. Interestingly, the Conservative government used the same language of financial necessity as previous Labour governments when justifying this change, demonstrating the widespread acceptance of the idea of a free and comprehensive service. In 1956, the Minister of Health wrote to chemists to say that, “this alteration…was made with great reluctance, and only because it is regarded as unavoidable in the present financial and economic situation.”173 In a cabinet meeting to discuss the change, the Minister of Health raised concerns that the charge per item would cause hardship and suggested that the BMA could mitigate this with “generous prescribing in particular cases,” such as a larger prescription of insulin for diabetics. Others noted that, “the Local Medical Committees could confidentially be instructed to disregard heavy prescribing in cases of chronic sufferers. It could then be stated that doctors would not be penalized for generous prescribing in suitable cases.”174 It seems that ministers were looking for a way around their own policy in certain cases, and again, this became a source of concern later on when doctors acted on it.  169. In 1950, a loaf of bread cost about one shilling, and was no longer rationed. Joyce Williams, personal communication, 3 December 2009. 170. Executive Council circular, 1952. TNA. MH 135/76. 171. Boots Drugs Co. to Minister of Health, 18 June 1952. TNA. MH 135/77. 172. Treasury history of prescription charges. 24 February 1966 – 31 May 1967. TNA .T 227/2522. 173. Minister of Health to National Pharmaceutical Union, October 1956. TNA. MH 135/79. 174. Meeting of Ministers of Prescription Charges: Meeting 1. 15 November 1956. TNA. CAB 130/121.  123 Ministers were also concerned with more public methods of blame avoidance, saying “It might…be easier to meet criticism of the increased charge if some concession in the health field could be announced at the same time, for instance, the provision of local authority chiropody services for old people...”175 Despite these plans, and a policy allowing low-income patients to apply to the National Assistance Board for a refund of their prescription charges, increased charges still caused public consternation. In November 1956, a constituent wrote to her member of parliament to say that once the new charges were in place, her prescriptions would cost 5 shillings each time “yet how can I possibly afford to pay 5/- every two weeks. I have three young children and our income is considerably small. Do you think you have given this scheme fair consideration…”176 This file, which was maintained by the Ministry of Health concerning the amendment regulations for the new charge, contained a number of similar complaints from other constituents, so the public appears to have been concerned even before the change was implemented. A Conservative government doubled prescription charges in 1961, and this again was the subject of much policy discussion, despite the fact that the cost of the average prescription had increased to the point that doubling the charge still resulted in patients paying a slightly lower proportion of the cost of each prescription than they did when charges were first introduced in 1952.177 Government’s justification for the increase focused on the rising cost of prescription services and the need to maintain sufficient budget room for other priorities, such as hospital construction. A draft statement prepared for the Minister of Health stated, “The Government are determined to continue their policy of developing the Health Services and to carry through a long-term programme of modernizing our hospitals. These objectives would be endangered if the cost of the Service to the Exchequer were allowed to go on increasing at so high a rate.”178 An internal memo argued that, “it is the duty of the Government to do everything it can to limit the cost of the Pharmaceutical Service to that reasonably required for effective treatment.”179 The Conservative government’s somewhat weaker commitment to NHS principles is demonstrated in  175. Meeting of Ministers of Prescription Charges: Meeting 1. 15 November 1956. TNA. CAB 130/121. 176. Mrs. Richards to Member of Parliament, 22 November 1956. TNA. MH 123/290. 177. The average cost of a prescription in 1949/50 was 3s.1d and by 1960/61 had increased to 7s.3d. If the cost of prescriptions in 1949/50 was similar to 1952, patients initially paid about 32 percent of the cost of a prescription, whereas the two shilling charge in 1961 only represented about 27 percent of the average prescription cost. See Undated memo “Increased charges for drugs and appliances.” TNA. MH 135/73 “Drugs and appliances. General. New prescription charges 1961 – policy.” 178. Undated Draft Statement (by Minister of Health on increase in charges) TNA. MH 135/73. 179. Undated memo “Increased charges for drugs and appliances”. TNA. MH 135/73.  124 a question-and-answer memo on prescription charges that argued, “The Government adheres to the principle that a patient under the National Health Service shall be supplied with all necessary drugs, but they must be paid for and it is reasonable that patients should pay a reasonable share,”180 contrary to the original principle of a “free” service. The government’s rationale was rejected by a pharmacists’ group who wrote to the Ministry of Health that “We have always held the view that the levy [patient charges] system is socially unjustifiable.”181 Prescription charges, 1964-1968: the role of principled ideas and electoral incentives The ongoing political relevance of prescription charges is demonstrated by the Labour party’s push to abolish them, which was part of the party platform starting in 1955.182 It was not an election-winning issue, but the fact that this fairly specific aspect of the NHS received individual attention in the platforms is significant. In their 1955 election manifesto, Labour promised that, “In order to restore a free Health Service, we shall abolish all charges, including those on teeth, spectacles and prescriptions.” The Conservatives countered by emphasizing hospital building and preventative medicine, and saying “we rank [these priorities] higher than free wigs or free aspirins” and won the 1955 election (Craig 1975, 198), but the Labour promise persisted in manifestos in 1959 and 1964. Labour won the 1964 election with a bare majority and moved quickly to fulfill its promise about prescription charges. In a letter to Prime Minister Harold Wilson just a month after the election, First Secretary of State George Brown said “I am worried by the omission from the Budget Speech of any reference to the date for the abolition of Health Service prescription charges. I think it matters that we should take the greatest credit for this and proclaim it at an early date, and it ought to operate from the earliest date practicable.”183 In his reply, Wilson agreed with this assessment. This exchange shows Labour was aware of credit-claiming opportunity and potential electoral payoffs for abolishing prescription charges, but there is also evidence that some politicians, in particular Health Minister Kenneth Robinson, may have wished to get rid of charges because of a principled commitment to the idea of free health services. A Treasury memo notes that, “The Minister (Mr. Kenneth Robinson) has personally  180. Q&A memo “Prescription Charges” TNA. MH 135/73, emphasis added. 181. National Pharmaceutical Union to R. F. Tyras (MH), 2 February 1961. TNA. MH 135/73 182. This is despite the fact charges were first proposed by the Labour government of Clement Attlee in 1949: the Attlee government passed legislation that gave government the power to impose charges, but prescription charges were not introduced until a Conservative government under Winston Churchill came to power in 1952. 183. George Brown to Harold Wilson, 12 November 1964. TNA. T 227/2285 “National Health Service. Abolition of Medical Prescription Charges 1964-1965.” Emphasis added.  125 pledged abolition of prescription charges, and has attacked the charges on many occasions.”184 Robinson wrote to the Chancellor the Exchequer, James Callaghan, in November 1964 to argue against postponing the abolition of charges until spring 1965, as this “would have forced the chronic sick and the elderly to endure another full winter of the charges and exposed us to the accusation of dragging our feet.”185 As will be seen below, Robinson’s principled ideas about the need to abolish charges were also in evidence during his vigorous but ultimately unsuccessful attempt to prevent his own government from reinstating the charges four years later. Both principled ideas and strategic considerations were evident in internal government debates about the feasibility of abolishing charges: a Treasury memo written very shortly after the 1964 election cautions that “the total cost of the Health Service is growing fairly rapidly, and the abolition of prescription charges will add significantly to the total cost.” In order to deal with this, the memo’s authors suggest “Unpalatable measures to control or pay for this expenditure might be easier to present at the same time as the abolition of prescription charges rather than at a later date.”186 This acknowledgement of both the financial implications of abolishing prescription charges and a potential need to compensate for them suggests that by 1964, prescription charges had taken on a special political meaning that made it more electorally expedient (or ideologically acceptable, for some politicians) to address these charges even if it meant other, perhaps less visible, cost control measures in other parts of the health service. During the controversial reintroduction of charges, Cabinet minister Richard Crossman recalls that he was against abolishing prescription charges in 1964 (as was the Chancellor of the Exchequer and Health Minister Kenneth Robinson himself, initially) and that “the decision was taken by Harold Wilson, who argued that we’d made the pledge in the election manifesto and we couldn’t let the electors down” (Crossman 1975, 646), pointing again to charges’ electoral significance. Prescription charges were reinstated in 1968 by the same Labour government that had abolished them four years earlier. The internal party debate and public dismay over this change demonstrates the conflict and electoral tradeoffs between the need to control health spending and demonstrate economic responsibility to the voters, and the principled ideas and public expectations about NHS services that were pushing politicians they other way. This can be compared to the evolution of patient copayments for drugs in Australia, where after some initial  184. Treasury memo “Free Prescriptions,” (undated). TNA. T 227/2285. 185. Kenneth Robinson to James Callaghan, 12 November 1964. TNA. T 227/2285. 186. Treasury memorandum re: “Abolition of Prescription Charges: Minister of Health’s Letter of 19th October 1964”, 21 October 1964. TNA. T 227/2285.  126 concerns about the public reaction that were cushioned by expanding the formulary, copayments became a fairly uncontroversial element of pharmaceutical benefits. In Australia, increases to copayments were made with little debate and public attention, and there was never a move to abolish charges by either major party. A ministers’ briefing note prepared by the UK Ministry of Health in 1968 calls the reintroduction of prescription charges “a regrettable necessity forced upon the government by the economic situation,” and goes on to compare the reintroduction of charges to other cost-cutting measures such as (a) cutting out some element of the health service altogether; (b) cutting the hospital building programme…Of these, (a) would be repugnant, [and] (b) would involve cutting back a long- term programme of planning and construction which has taken years to build up and would take correspondingly longer to restore.187  Here, policymakers made two justifications for prescription charges, given the financial imperative to reduce health spending: first, they implied a hierarchy of NHS principles, arguing it was better to impinge on the principle of free services than to limit the range of services provided, and second, in this case prescription services were a lower priority than other elements of the NHS. This view is supported in archival materials from the Treasury concerning prescription charges: in December 1967, Health Minister Robinson wrote to Roy Jenkins, the new Chancellor of the Exchequer, to describe the difficulty he was in given his personal association with the abolition of charges, and his conclusion that hospital building should also not be cut, as the necessary savings “could only be achieved, if at all, by wrecking the entire hospital building programme.”188 Robinson further elaborates his opposition to prescription charges in a memo written at the same time, raising concerns about the hardship they would cause, despite proposals for a system of exemptions, and about the political repercussions of such a move, saying, I can see no advantage whatever in reducing Health Estimates (assuming they must be reduced) by a method which would cause the maximum political dissension, and probably disaffection, which would yield an uncertain revenue and certainly far less than the alternative I have suggested [an increase in the general NHS contribution]; which would pose considerable administrative problems and require extra staff; and which would breach a public pledge by the Prime Minister and others at the time of devaluation, by specifically taxing the sick.189   187. Ministry of Health, Ministers’ Briefing Note on Prescription Charges, 30 May 1968. TNA. CAB 151/136 “Prescription charges. 1968.” 188. Kenneth Robinson to Roy Jenkins, 18 December 1967. TNA. T 227/2652 “Reintroduction of Medical Prescription Charges. 1968. 13 December 1967 – 29 February 1968 (Part A).” 189. Minister of Health, Memorandum re: Reintroduction of Prescription Charges, 18 December 1967. TNA. T 227/2652. Emphasis added.  127 Interestingly, proponents of the prescription charge within Treasury also seemed aware of the potential electoral consequences and toll on Labour party unity posed by this policy, but were more sanguine about the benefits outweighing the costs. In December 1967, a Treasury official wrote that the “reintroduction of prescription charges would by the quickest, simplest and most economically effective course. Here again it is politically the most difficult, but the most rewarding in terms of public confidence,” perhaps in reference to the improvement in the economic situation.190 The issue of prescription charges and a new system for exempting the chronically ill from these charges continued to be a major source of discussion within government and in the House of Commons throughout the summer and fall of 1968.191 Excerpts from Hansard from this period reveal that the Labour government was attacked by its own backbenchers, who chastised the government for “seeking to tax the sick,” and emphasized that “the reintroduction of prescription charges will cause bitter disappointment among many people in the Labour movement” and “hits at the very basis and principle on which the [National Health] Service was established.”192 Prime Minister Wilson assured the House that the charges were only being reintroduced “after the most serious consideration and with the utmost reluctance” and argued that “it would not have been possible to do what was necessary to get the economy right without making some cuts in the social services, which account for a very high proportion…of total Government expenditure” and that maintaining funds for the hospital building programme was a priority.193 Although Wilson referred repeatedly to the proposed scheme of exemptions from charges, whereby “something like one-half of the prescriptions dispensed in this country will be the subject of the exemption categories I listed…the charges on the stamp will be on those who, in the main,  190. Treasury official letter (Rampton), 13 December 1967. TNA. T 227/2652. 191. Internal government discussions are contained in three additional files maintained by Treasury, as well as others from the Ministry of Health. See for example T 227/2655 “Reintroduction of Medical Prescription Charges. 1968. 1 March 1968 – 11 April 1968 (Part B);” T 227/2656 “Reintroduction…17 April 1968 – 10 June 1968 (Part C);” T 227/2273 “Reintroduction…17 June 1968 – 20 September 1968 ) Part D;” and PREM 13/2805 “National Health: Date from which abolition of charges should become effective; loss to Exchequer; reintroduction of charges, 1964- 1968;” and MH 149/1085 “NHS Prescriptions Numbers and Costs: Reports to Minister 1967-1971.” 192. Misters Thorpe, Winnick and Pavitt, 16-17 January 1968. H.C. Debates 1967-1968. Vol. 756 c.1578-1814. 193. Harold Wilson on “Public Expenditure,” 16-17 January 1968. H.C. Debates 1967-1968. Vol. 756 c.1578-1814.  128 are in work and who can afford it,”194 members did not appear to put much stock in this scheme. Similar doubts were expressed in private by Wilson’s own Health Minister. Early in January 1968, Prime Minister Wilson announced the plan for exemptions to cabinet, and Crossman notes that, “a reply had come from Kenneth Robinson that no formal exemption was administratively possible. Of course I recognized at once this was a political tactic…to prevent any increase of prescription charges” (Crossman 1975, 637). When recounting the next week’s cabinet meetings, Crossman refers to the resistance to prescription charges in the Labour backbenches and says that when the meeting discussed charges “despite all my efforts Kenneth Robinson had produced a lamentably inadequate paper. It was clear that he was trying to make as many difficulties as he possibly could” (Crossman 1975, 644).  The events of the 1964 demonstrate that there were both ideational reasons for the government in power to abolish charges, and electoral incentives to do so. In 1968, these two factors were still at work. They were trumped by the financial imperatives of an increasing expensive health service during a difficult economic time, which created countervailing electoral incentives,195 and charges were reinstated. However, in comparison with Australia, UK prescription services still appear more strongly conditioned by the policy preferences and public expectations generated by the radical circumstances of their adoption. Although both countries reformed their pharmaceutical policies to control costs, the British reforms imposed lesser burdens on patients in terms of the extent of charges and their impact on the program’s comprehensiveness. If we take the counterfactual of existing UK prescription charges to be a system of charges like Australia’s, rather than no charges at all, the British policy is a much more moderate system of pharmaceutical management. Charges were relatively low: only 20p per item in 1971, which is about £2.12 in 2009 pounds. In contrast, in March 1971 the Australia copayment was $0.50 per prescription, or $4.59 in 2009 Australian dollars, and the rate increased to $1 or $8.66 (2009) in December of that year.196 Charges were also applied much less broadly than in Australia: the system of exemptions from NHS charges in the UK began when charges were reinstated in 1968, and has expanded to cover new groups since that time. In  194. Ibid. 195. Persistent balance of payment problems forced the government to devalue the sterling in 1967 (Booth 1995), which decreases the domestic standard of living and increases the burden of a country’s foreign currency debts. 196. Australian price conversions use a CPI deflator constructed from the historical index available from the Australian Bureau of Statistics at  129 contrast, in Australia a system of lowered copayments for concessional categories of beneficiaries (based on income and health status) was only put in place in 1983. Conclusion The UK adopted free, universal prescription services in 1948 as part of its comprehensive National Health Service. This radical policy change was possible because of the national government’s centralized authority over health policy, and because of politicians’ high degree of ideational and electoral motivation to act. The Labour party had a longstanding principled commitment to the idea about universal-comprehensive health services provided by the Beveridge Report, and this commitment was bolstered by the high degree of public knowledge of and support for the plan. Although prescription services were included because of the commitment to a full range of public services, rather than on their intrinsic merits, after they were adopted they raised a separate set of concerns with regards to cost and possible abuse, and they were subject to reform before any other aspects of the NHS. Faced with higher than expected costs for prescription services and for the NHS overall, governments determined that prescription charges were the least painful and best first measure for controlling expenditure. However, the difficulty of implementing these charges in the UK compared to Australia demonstrates that the conditions under which a program is adopted (in this case, commitment to comprehensive universal free services) place limits on the type of changes that are politically feasible. In the UK, the government’s own backbenchers (and at one point, its health minister) condemned charges as a “tax on the sick,” and a repudiation of founding NHS principles,197 and the abolition and reinstatement of charges were significant policy events during the 1960s. In Australia, there were no such founding principles to reference, and it was possible to link the new charges to “a greatly widened area of prescribing,”198 owing to the original limited PBS formulary. Five years later, an Australian policymaker was able to describe the charges as “in line with the principle of “self- help” upon which the Australian National Health Scheme is based,”199 despite the marked differences in goals between the universal and comprehensive PBS and the government- subsidized system of private insurance for hospital and medical insurance. Therefore, although  197. H.C. Debates 1967-1968. Vol 756 c.1578-1814; Minister of Health, 18 December 1967. TNA. T 227/2652. 198. Press statement by Dr. Cameron re: New Pharmaceutical Benefits, 16 February 1960. NAA. A 463, 1956/1905. 199. Reply to information request from Senator, Jersey, Channel Islands, 1965. National Archives of Australia. A1658, 813/1/1/1 Part 3.  130 the UK and Australia both adopted similarly universal and comprehensive pharmaceutical programs in the post-war period, the different paths of adoption – simultaneous in the UK, and incremental in Australia – had longer term implications for the way these programs were reformed in later years.   131 6 Conclusions Introduction Canada’s lack of a nation-wide, universal and comprehensive public pharmaceutical program is puzzling in light of the experience of almost all other established welfare states, and its own broad public system for hospital and medical insurance. This empirical puzzle points to a more general research problem: previously, little was known about sources of variation in countries’ pharmaceutical programs, and about the conditions under which countries developed universal, comprehensive programs. This was a significant gap given the importance of drugs to therapy and their position as cost drivers in the health system. The combination of these two issues suggested that a new approach to health policy research, capable of considering the various components of a health system as potentially separate entities with different politics and causal forces, was necessary.  This research addressed the general question of how countries developed, or failed to develop, pharmaceutical programs by focusing on a key decision in the immediate post-WWII period: is health policy development approached simultaneously, with a radical introduction of all services at once, or does it proceed incrementally, with one service implemented at a time? The research finds that certain institutional, ideational and electoral conditions allow for more radical policy change, in this case, the simultaneous adoption of a wide range of health services. Lacking centralized authority, principled ideas and electoral motivation, policy change may still occur. However, it is of a very different character than the radical, simultaneous approach. This less risky, incremental approach of adopting one health service at a time favours the introduction of initial priorities, but presents barriers to later priorities, as they represent deviations from the “path” generated by a country’s first decisions about health services. Somewhat counter-intuitively, I find that there is greater flexibility within the programs adopted in an incremental process, at least in the early days of policy implementation, than there is in simultaneously adopted health programs. The high levels of principled and strategic commitment to the idea of broad public health services that allows for their simultaneous adoption also constrains attempts to control costs in ways that might limit the universality or comprehensiveness of coverage. Thus, an incremental approach affects policy development in at least two ways. In terms of broad health policy outcomes, it restricts the range of services that are introduced, so policy development is constrained to the path of the service or services that were early priorities. In terms of changes to a specific pharmaceutical policy, the ideational and  132 electoral conditions of policy adoption affect the parameters of change that are feasible later on, so an incremental approach generates more flexibility to reform an existing program in ways that limit its comprehensiveness and universality.  The remainder of the chapter reviews the research question of cross-national variation in pharmaceutical programs in terms of the three related analytical problems: why do countries take an incremental versus a simultaneous approach to policy development? If countries take an incremental approach, why do they chose certain priorities for service adoption, and what barriers arise to the adoption of additional services over time? Finally, how are countries’ responses to the early challenges of program implementation affected by their initial approach to policy development? The chapter summarizes the evidence for each of the hypotheses presented in chapter one. It then returns to the broader lessons of the project for the way scholars think about incrementalism, and about the way elite ideas and public expectations work over time to shape opportunities for policy change. The chapter ends with a discussion of policy developments since 1970 and directions for future research. Summary of evidence  The choice of simultaneous versus incremental reform  When countries are initially considering the creation of public health care programs, a simultaneous approach to policy development is the more radical option: it requires a greater initial outlay of financial resources and political capital, more coordination, and entails a greater risk: it aims for a much larger program than more limited incremental approaches. Three factors are important in determining whether a country takes a simultaneous or an incremental approach: centralized authority over health, a cohesive, principled idea on the part of policy elites about the way the health system should work, and voters who are both in favour of public health policies and place a high level of importance on them. When all three were present, as was the case in the UK, the country took a simultaneous approach. When this was not the case, as in Canada and Australia, countries defaulted to the less risky incremental approach, implementing one service at a time.  In terms of centralized authority, there is a clear contrast between Canada, which took an incremental approach, and the UK, which took a simultaneous approach to policy development. The national government in Canada initially proposed a comprehensive health system, but when these proposals were rejected by provincial governments, the federal cabinet was very deferential to provincial preferences, requesting plans from the provinces with regards to the organization  133 and adoption of health services200 and ensuring that the Department of National Health and Welfare’s next proposals allowed maximum flexibility for provinces, by providing an option for provinces to take up health services in stages as they saw fit.201 Perhaps most telling is Prime Minister Louis St. Laurent’s position at the 1955 Federal-Provincial Conference where he stated that the federal government would not want to interfere with areas of provincial jurisdiction or introduce constitutional change.  In the UK, this deference to subnational governments was clearly lacking – which is unsurprising given their lack of constitutional authority under the British unitary system. The 1944 British White Paper (A National Health Service) argued that a nationalized service was necessary because “medical resources must be better marshaled for the full and equal service of the public, and this must involve organisation – with public responsibility behind it” (Ministry of Health 1944, 8). By 1946, a new Labour government, enjoying the advantages of a majority in a centralized Westminster parliamentary system, was able to nationalize the hospital system over the objections of local authorities and implement its plan for a comprehensive health service (Tuohy 1999). Although this aspect of institutional centralization of authority was important, Australia demonstrates that when considering fragmentation and centralization of authority over health, it is useful to look beyond the wording of constitutional responsibilities and consider the practical functioning of fiscal federalism and amendment rules. State governments in Australia had constitutional authority over health, but federalism per se was not the primary barrier to a simultaneous approach to policy development. The Commonwealth government did not take state’s preferences into account when planning an approach to policy development, mainly because of the weak fiscal position of Australian states vis-à-vis the Commonwealth government (Matthews and Jay 1972), and because of the Commonwealth’s greater confidence in its ability to achieve a constitutional amendment for powers over a range of social services (Livingston 1956). Instead, Australia’s choice of an incremental approach to policy development was much more dependent on the reciprocal relationship between principled ideas and electoral incentives, factors that also shaped the choice of policy approach in Canada and the UK.  In the UK, policy elites held a cohesive, principled idea about the way a health service should function, and these once these ideas had been introduced to the public, they were  200. Meeting of the Cabinet Committee on Dominion-Provincial Relations: Report of the Working Committee on Health Insurance, 4 January 1946. LAC. Brooke Claxton fonds. MG 32-B5  Vol 138 File: “Dom-Prov Conferences DP-2 Committees and Meetings.” 201. DHW Memo 22 December 1949. LAC RG 29 Vol 1061 File 500-3-4 pt 1.  134 supported by voters that placed a high degree of importance on the issue of broad public health services. Both these factors were lacking in Canada and Australia, leading politicians to default to a less risky incremental approach. I discuss these two factors together because they have a reciprocal relationship: public support is more likely to develop in the presence of a prominent elite idea, which then feeds back into politicians’ priorities for policymaking. This relationship is particularly important over time, as will be discussed below in terms of barriers to the adoption of additional services in an incremental process. The British Labour Party had a longstanding commitment to public health services: for example, its 1935 election manifesto promised that a Labour government would “vigorously develop the health services” (Craig 1975). The promise of health services gained prominence on policy agendas with the publication of the report of the Inter-departmental Committee on Social Insurance and Allied Services Report, or the Beveridge Report, in 1942. This high-profile document called for the introduction of a comprehensive range of free health services on a universal basis. Its remarkable popularity and salience with the British public (Jacobs 1993) encouraged the wartime coalition government to reach a compromise on a national health service in 1944 (Ministry of Health 1944), and emboldened the new Labour majority government in 1945 to abandon certain compromise positions and implement an even more radical new health service between 1946 and 1948. A broad public health service was a prior element of Labour platforms, but as Jacobs argues, the 1945 election “provided indisputable evidence of public support for innovative legislation; this perception created nearly irresistible incentives for policymakers to pursue reform” (Jacobs 1993, 167). In Canada and Australia, in contrast, there was no motivating idea comparable to the Beveridge Report, no principled commitment to an idea of health services on the part of pragmatic Canadian Liberal and Australian Labor parties, and consequently no elite rhetoric to foster the strong and immediate public demand for health policy, as there was in the UK. The federal cabinet in Canada between 1946 and 1953 was divided on the question of public health insurance, and the two Prime Ministers during that time remained skeptical about a broad government program. This lack of principled ideas on the part of elites meant they preferred an incremental approach, and the corresponding lack of attention to health from the Canadian public meant that politicians were not provided with electoral motivation to act more decisively. Canadian Health Minister Paul Martin Sr. noted the results of a Gallup poll in mid-1947, saying that “National health insurance unfortunately received scant support” (Martin 1985, 45) and by 1953 had concluded that, “the division of opinion [in cabinet] made it obvious that I would never  135 get a combined hospital and medical plan into operation, so I opted for hospital insurance as the easier route” (Martin 1985). In Australia, there was a similar lack of principled ideas from the governing party: the Australian Labor Party was generally pragmatic, rather than principled, in its policy positions (Johnson 1989, 16, 6). Once in government, Labor planned to introduce some new social benefits, but placed no particular priority on health: main responsibility for health policy was in fact marginalized in a consultative forum for Commonwealth and State health departments. Thus, there were no principled ideas to galvanize the public, and limited new social benefits including health benefits were not a clear electoral winner (as was the case in the UK) owing to the consistent division of public opinion along party lines. Gallup polls from 1945 to 1949 demonstrate that compulsory government health insurance was the exclusive policy domain of the Labor party, and tended not to elicit high levels of support from conservative Liberal- Country Party voters.202 The UK took a simultaneous approach to health policy development because it had the institutional prerequisites, centralized authority over health, as well as a principled idea about health policy on the part of elites, and the support of an attentive public. Canada and Australia lacked these factors and took an incremental approach, but crucially, both initially planned to implement a full range of services by proceeding in stages. This proved much more difficult than early policy designs anticipated, and in both countries, the first priority for policy development determined the scope of the health system for some time.  The choice of priorities for services adoption  In countries taking an incremental approach, health policy development tended to stall after the adoption of the first service. This is the second analytical problem: how do countries choose their first priority for service adoption, and what barriers arise to the adoption of additional services later on? Why are services that were initially discussed as “later priorities” less likely to be adopted? Priorities were determined by the preferences of institutional veto players, and a calculation of electoral trade-offs: what type of opposition certain services would encounter (including potential opposition from organized interests), and how much they would cost. The Canadian and particularly the Australian cases demonstrate that there is some role for case- specific factors in the choice of priorities. For example, a new finding of this research was that a lack of medical manpower, owing to an absence of a large proportion of Australian doctors with  202. Gallup Poll (Australia), December 1945-January 1946, Polls 314-326, “Federal Social Services Approved.”  136 the armed forces during WWII, was a key factor in Treasury’s decision to adopt a program of pharmaceutical benefits first.203 However, the choice of priorities is ultimately predictable based on the logic of beginning with “low-hanging fruit” in an incremental process, in the sense that the politicians wished to start with policies that either had lower material costs or would require less political capital to introduce. Both Canada and Australia felt some action on health was required: in Canada, this was because of electoral pressures on the federal Liberal government from the social democratic Co-operative Commonwealth Federation party (Hacker 1998; Maioni 1998) and from early provincial adopters of hospital insurance such as Saskatchewan.204 In Australia, the pressures were internal, from a Commonwealth government determined to act on a new social benefit “before the disruption of demobilization,” (Gillespie 1991) and before the more centralized wartime defence powers came to an end. In both counties, the choice of first priorities was dictated by the service that appeared to be the simplest to implement. In Canada, federal bureaucrats actually removed pharmaceuticals from the policy agenda quite early, concluding in 1950 that, “All the experience to date indicates that it is almost impossible to control the costs in such services,” and so they should not be considered further.205 After this determination was made at the federal level, provincial preferences for hospital insurance became the key factor, although federal bureaucrats discussed medical insurance as a closely-related second priority.206 In Australia, as noted above, the choice was determined by what the Commonwealth government thought was feasible with wartime resources and manpower. Barriers to further policy development Barriers to the adoption of additional services after a first priority is in place include alternative institutional arrangements, realistic perceptions of resource constraints, and ideational barriers from the reciprocal relationship between elite ideas and public expectations. These barriers developed over time because of certain characteristics of public policy that are predicted by the path dependence literature (Pierson 2000; Arthur 1994): large set up costs, sequencing effects, and adaptive expectations. All these factors reinforce existing services and block deviations such as the adoption of additional services.  203. Memorandum, January 1944. NAA. A571, 1943/4513; Social Security in Australia memorandum. NAA. A571, 1943/4513. 204. Health Insurance brief, 7 December 1949. LAC. RG 29 Vol 1061 File 500-3-4 pt.1. 205. Ibid. 206. Draft minutes of the first meeting of the Interdepartmental Working Committee on Health Insurance, 30 November 1949. LAC RG 29 Vol 1061 File 500-3-4 pt. 1.  137 Large set up costs refer to the costly and complex administrative structures required for public health services, which are difficult to change substantially once in place. One aspect of this is the national government’s own cost of setting up a program, but the cost to other actors of setting up alternative programs is also particularly important: when national governments do not take early action, alternative institutional arrangements for the delivery of health services tend to arise. Other levels of government, private industry, or organized interests such as the medical profession may provide services, and it becomes more difficult to displace other actors’ investments in these arrangements the longer they are in place. Sequencing effects occur with regards to the cost of programs. In the specific case of incremental health policy, adopting one service at a time presents an opportunity for policymakers to learn about the real costs of health services. The first service has the advantage of being adopted under a certain degree of fiscal ignorance, or at least uncertainty, that does not exist for additional services. Instead, policymakers have had an opportunity to develop ideas (which may be more or less accurate) about the cost of health services and the kind of tradeoffs they require with other desirable policy goals. Observing higher than expected costs for a first service can act as a powerful disincentive to consider additional policy development, as the first service has taken up “fiscal space,” perhaps in reality but certainly in policymaker’s collective understanding of how expensive health policy is. Finally, adaptive expectations shape both policymakers’ and the public’s view of what the health system should do, as “projections about future aggregate use patterns [or the services included in a health system] lead individuals to adapt their actions in ways that help make those expectations come true” (Pierson 2000, 254). This is a matter of “moving the goalposts” by which the public and politicians evaluate health policy, as elites’ principled ideas about health services and the public’s expectations influence one another: little elite discussion of later services means public expectations are less likely to develop, and this lack of electoral motivation feeds back into policymaker’s priorities. The relationship between elite ideas and public expectations is an important factor earlier in the process, when countries chose between a simultaneous and an incremental approach to policy development. It becomes even more important after some health service or services have been adopted, as these services then act as a template for what a health system “should” do. In an incremental process of policy development, early public promises for a comprehensive system or additional services are typically vague, in keeping with the lack of principled ideas, and the best evidence of plans for additional services come from internal policy documents. If additional steps receive little attention after the adoption of the first priority, voters  138 may come to expect a limited health system. Policymakers’ deliberations may also become more restricted if they come to believe that the problem of health services has been “solved” by the adoption of initial priorities. Even if additional services represent significant gaps in public health care, many politicians appear to develop conceptions of health care where the management and cost of these existing services become the totality of the debate, an instance of ideas “direct[ing] actors’ attention in the course of decision making” (Jacobs 2009, 253, emphasis in original). In Canada, some of these mechanisms that reinforce a limited system were evident after the adoption of hospital insurance in 1957, and made subsequent adoption of medical insurance more difficult, even though the two services had been planned together and were implemented less than a decade apart. By the 1960s, provincial governments and private insurance funds sponsored by the medical profession had begun to fill the gap in health insurance. The investment in start-up costs incurred by these other actors meant they were reluctant to cede control over their programs to the federal government, and they provided steeper opposition to federal plans for nation-wide medical insurance than was the case with hospital insurance. This barrier was overcome by a new statement of principled ideas about health insurance from the Royal Commission on Health Services, or Hall Commission, and an ideological shift within the governing Liberal party, which was accompanied by growing public support for the new medical insurance proposals.207 This combination of factors meant that the federal government was temporarily willing to overcome the objections of some provincial governments and respond to new demands from within the government’s ranks and from the national electorate. However, this change in conditions did not extend to additional services such as pharmaceuticals: discussion was limited to the adoption of public medical insurance. The barriers to the adoption of additional services theorized in chapter one were again in evidence in the early 1970s, when the last federal proposal for a nation-wide, universal pharmaceutical program was rejected in cabinet. The rejection of the Drug Price Program, proposed by the Department of National Health and Welfare between 1971 and 1972, provides a clear example of the restriction of politicians’ understanding of a problem over time and the reciprocal relationship between elite ideas and public expectations.  In the early 1960s, the federal government had embarked on a program to curb drug prices through patent law that was surprisingly separate from health insurance policy, and this produced a consensus among federal politicians that the “problem” of pharmaceuticals was restricted to prices and patents. This consensus about the problem definition  207. CIPO/Gallup Poll of Canada, Public Opinion News Service Release, 27 March 1963.  139 limited how policymakers though about the issue later on, and meant that cabinet rejected the 1972 proposal without considering the authors’ arguments concerning its implications for controlling the drug prices or addressing a social need.208 Even though the proposal framed pharmacare as an extension of public health insurance, it was not debated in cabinet on these terms, it seems because the “problem” of health insurance had been limited to hospital and medical insurance, while the “problem” of pharmaceuticals was limited to prices and patents. Instead, cabinet focused on financial barriers to the proposed pharmacare program, arguing “pharmacare would be the beginning of a very expensive program which would undermine the confidence of the middle-income groups in the government’s ability to control the budget.”209 These financial barriers had been reinforced since pharmaceuticals were first declared “too expensive” by Canadian policymakers in 1950, both by real increases in drug prices and usage patterns, and by the repeated tendency to focus on these costs rather than available information about cost management strategies, from other countries or from the DHW’s own analysis. At this time, alternative institutional arrangements for drug insurance had not yet developed, but cabinet members still cited potential opposition from provincial governments, the medical profession, and the pharmaceutical industry as reasons not to proceed.210 Federal politicians also lacked electoral motivations to seriously consider a pharmaceutical program. Although the Canadian public was concerned about drug price control in the 1960s,211 they did not have any experience with public or private drug insurance, which potentially could have raised issues about the quality of services or even the “right” of individuals to get assistance with out-of-pocket drug costs, and made federal drug insurance more salient. Perhaps more importantly, the Canadian public also did not have any experience with their politicians talking about drug insurance. Since at least 1950, the service had been repeatedly pushed off policy agendas without public discussion: it did not feature in campaign platforms or speeches, was a smaller and mostly overlooked element of major reports on health insurance such as the Hall Commission, and was curiously omitted from other federal reports on drug prices and patents. This meant there was no basis for a demand for pharmaceutical insurance to develop.  208. Memorandum to Cabinet: Measures to lower the unit cost of prescription drugs including a drug benefit program [Pharmacare- handwritten], 2 February 1972. LAC. RG2 Volume 6397 File 120-72; “DRAFT – Some Social Reasons for Pharmacare” and “Arguments for Pharmacare.” LAC.  RG 29 Vol 1572 File 1016-1-2. 209. “Measures to lower the unit price of drugs including a drug benefit program,” 30 March 1972. LAC. RG 2 Vol 6395 Series A-5-a. 210. The Drug Price Program, 23 September 1971. LAC. RG 2 Vol 6381 . 211. Globe & Mail, 6 October 1960, p3.  140 This contrasts with the situation in Australia between 1949 and 1950, when the feedback between previous elite ideas and public expectations forced the implementation of the Pharmaceutical Benefits Scheme (PBS), counter to the general preferences of the newly elected Liberal government. In Australia, the primary barrier to the development of additional services after the implementation of the PBS was a lack of principled ideas or electoral motivation: in 1950, a Liberal-Country coalition government was elected with a “policy of minimum government intervention in welfare areas,” (Gray 1991), presumably by the same Liberal-Country voters who were skeptical when polled about compulsory government health insurance that year.212 The puzzling outcome here is why the PBS was implemented at all. After five years of constitutional wrangling and obstruction from the medical profession, the election of a new government that was much less inclined to favor public health services or benefits might have provided an opportunity to abandon the previous Labor government’s plans for the PBS. Instead, the implementation of the PBS demonstrates how quickly public expectations can arise in response to prompting from elites (in this case, from a government policy), and the way this relationship between elite ideas and public expectations may also be used to pass policies in unlikely circumstances. Although the Labor government had been unsuccessful in implementing the PBS, the fact that legislation had been on the books for five years and subject to high profile court challenges and a constitutional referendum meant that the public was well aware that the government had made provisions to supply them with free drugs – they just were not receiving the benefit. The Liberal party adjusted its platform to respond to these public expectations. In 1948, Liberals added pharmaceuticals to their party platform: the health section called for “the free provision of certain specific drugs vital to the preservation of life (such as insulin)” (White 1978), while still maintaining a commitment to voluntary medical and hospital insurance. When the Liberals were elected, they implemented the PBS, albeit in a revised format that was to have important implications for later attempts to reform the program. Early challenges of pharmaceutical programs The final analytical problem addressed in this thesis was the way that countries respond to the early challenges of pharmaceutical programs, given how quickly both the UK and  212. Gallup Poll (Australia), February-March 1949, Polls 569-578, “Medical & Dental Plan”; May- June 1949, Polls 590-599, “Public still wants free medicine, but opposition grows”.  141 Australia were confronted with rising expenditures and a changing pharmaceutical market. I found that taking an incremental versus a simultaneous approach to policy development also affects policy implementation, as can be seen in the first reforms to pharmaceutical programs in both the countries that adopted them.  The UK and Australia faced similar challenges with regards to the cost of pharmaceutical programs, which were mainly the result of the pharmaceutical market rather than the policy design in either country. The mid 1940s to late 1960s marked the “therapeutic revolution” in drug development (Morgan 2008), with a rapid increase in the number of therapeutically effective drugs and hence both use of drugs and drug prices. The UK and Australia also chose similar policy tools to address these challenges: both initially focused on introducing patient changes in an effort to limit demand and bolster revenues, rather than attempting to regulate physicians’ prescribing or pharmaceutical industry pricing. However, the public response and policy outcomes in the two countries were quite different, and these differences can be linked to the circumstances of program adoption. In the UK, the principled elite ideas and high levels of public support and attentiveness that made a simultaneous approach to policy development possible also made it very difficult to implement cost management policies. In Australia, the lack of principled ideas on the part of the government and the relatively lower public expectations that accompanied an incremental process of policy development made the PBS more flexible. The adjustments to the scheme as implemented by the Liberal-Country government provided an opportunity to compensate the public for charges initially, and eventually allowed for more significant retrenchment than was possible in British prescription services.  In the UK, the Treasury proposed patient charges for prescriptions as early as 1948,213 and charges were debated in cabinet in 1950. At this time, politicians noted the potential for backlash from both physicians and patients,214 given the emphasis placed on the original NHS principle of a universal free service. Politicians’ concern and public consternation continued when charges were increased by a Conservative government in 1956 and 1961,215 and the Labour party made the abolition of prescription charges part of its platform in 1955 (Craig 1975, 198). Upon forming the government in 1964, the Labour party moved quickly to fulfill its promise to  213. Treasury history of prescription charges: S.S. aspects. 24 February 1966 – 31 May 1967. TNA .T 227/2522 214. Chancellor of the Exchequer, Memorandum: NHS: charge for prescription, 26 January 1950. TNA. CAB 129/38. 215. Meeting of Ministers of Prescription Charges: Meeting 1. 15 November 1956. TNA. CAB 130/121; Mrs. Richards to Member of Parliament, 22 November 1956. MH 123/290; Undated Draft Statement (by Minister of Health on increase in charges) TNA. MH 135/73.  142 abolish charges, only to reinstate them four years later. At this time, as at earlier points, the unpopularity of prescription charges was trumped by a perceived economic necessity of reducing NHS expenditures and avoiding cuts to other parts of the service.216 However, the controversy that prescription charges evoked meant that they were abolished for a time, something that was never considered in Australia, and also kept UK charges quite low in comparison to Australia, so they were not an effective source of revenue or likely constraint on usage. Furthermore, widespread exemptions from charges were introduced in 1968, designed to exclude “something like one-half of the prescriptions dispensed in this country”217 from any fees. In Australia, a concessional fee was not introduced until 1983, and it still required the more narrowly defined concessional beneficiaries to pay something for their prescriptions.  In Australia, patient charges were introduced somewhat later, in 1960, and were much less controversial. In part, this was because of a particular opportunity to compensate patients for the new charges: when the Liberal government implemented in PBS in 1950, they did so with a restricted formulary that only covered certain “life-saving and disease-preventing” drugs. When the government needed to address rising PBS expenditures later on, they concluded that it would be possible to introduce a fee, which “would be more acceptable to the public if it associated with a wider range of drugs being available under the scheme, and the evidence…suggests that the latter would be no burden on the Commonwealth.”218 The combination of charges and an expanded list of covered drugs meant that although patients faced a new fee for prescriptions, all prescriptions were now available at this subsidized rate, whereas previously some drugs had been free to the patient, while other drugs had to be bought out-of-pocket at full price. Thus PBS charges were introduced without evidence of electoral backlash, and perhaps more interestingly from a theoretical standpoint, the public indifference to charges continued when they were doubled in 1971 and again in 1975: there is no evidence of voter protest, such as the file of angry constituent letters maintained by the British Ministry of Health, and in fact Australian cabinet discussions of the increases did not even anticipate controversy.219 This was the case even though there were no further opportunities for compensation after the one-time broadening of the formulary in 1960. The incremental approach that resulted from the lack of principled ideas and  216. Harold Wilson on “Public Expenditure,” 16-17 January 1968. H.C. Debates 1967-1968. Vol. 756 c.1578-1814. 217. Ibid. 218. Notes on Cabinet Submission No. 252. 13 July 1959. NAA. A5619, C523. 219. Cabinet Minute, 21 July 1971. NAA. A5908/192. A5908/192. “Pharmaceutical Benefits – Patient contribution – Decision 259 (M). 1971”; Minister of Health to Prime Minister, 13 June 1975. NAA. A5931, CL640.  143 electoral motivations concerning health services allowed for less controversial and eventually more significant patient charges in Australia, compared to the UK where even nominal charges were a matter of principle both within the Labour caucus and the voting public. As will be discussed below, this raises new questions about the long-term implications of the choice between incremental and simultaneous policy development. Contributions This thesis makes an empirical contribution by explaining an understudied policy puzzle: why Canada did not introduce pharmaceutical insurance to complement its nation-wide, universal public hospital and medical insurance, as did almost every other developed welfare state. In doing so, it challenged two versions of the accepted wisdom among Canadian experts. First, that when health insurance was first debated in the 1940s and 1950s, drugs were not therapeutically very important: policymakers did not expect that they would be an important or costly part of health services in future and so excluded them. However, when pharmaceuticals are considered in comparative perspective, it is clear that other countries recognized drugs’ importance at this time, which was the beginning of a very innovative period in the development of therapeutically significant drugs. Instead, my research found that Canadian policymakers considered pharmaceuticals, but were worried that growth in drug expenditure would be uncontrollable. Second, an alternative type of historical explanation suggests that we focus on the timing of policy choices relative to technological development, arguing that when health insurance was finally implemented in Canada with the introduction of Medicare in 1966, drugs were already too expensive, because of the aforementioned therapeutic revolution. The expense of drugs near the end of the therapeutic revolution was certainly a factor in Canadian policymakers’ deliberations, but my archival research points to the fact that these concerns were present much earlier, in 1950. Additionally, the research highlights a policy episode in Canada that has not previously been the subject of scholarly investigation: the rejection of the 1972 Drug Price Program. This episode allows me to provide a more detailed explanation of the origins and evolution of Canadian politicians’ concerns about pharmaceuticals, and why bureaucrats were not able to convince them to take additional steps to develop a pharmaceutical program. The research also makes empirical contributions with regards to Australian policy development, where the Treasury discussions of the reasons for beginning with pharmaceutical benefits in  144 1943, and the differences between the initial Labor PBS formulary, and the list of drugs eventually covered by the functioning Liberal scheme in 1950, have not received prior attention. In theoretical terms, this research contributes to the historical institutionalist and national health insurance literatures by providing a more nuanced explanation of the origins and implications of a critical moment in health policy that I expect to have more general application to a variety of questions of social policy development. The research provides guidelines to predict the approach the policy development a country will take, based on institutional, ideational and electoral conditions, and explains why this choice of policy approach is so important. Two key theoretical contributions of the research are first, the challenge to the notion that incrementalism is just another way, and sometimes the only possible way, of achieving policy reform. The research demonstrates that in health policy development, the dynamics of incremental approach result in a stalled policy process, as barriers to the adoption of additional services increase as time goes by. Second is the argument we must consider not only the role of ideas in policy making, but also the role of ideas over time, as elite ideas about the particular problems and solutions in a given policy area become entrenched, and influence public expectations to either maintain existing policies or create demand for change. This research finds that it matters whether you decide to adopt big reforms in one package or in stages, and then the sequence matters. In countries taking an incremental approach to policy development, where you start tends to be where you end up. This is because instead of continuing indefinitely, in a series of uncontroversial steps (Lindblom 1979, 520), incremental policy processes stall in predictable ways. The mechanisms by which policy stalls are suggested by the path dependence literature: this thesis focused on the role of large set up costs, sequencing effects, and especially adaptive expectations as applied to health policy, and responds to criticisms about the vague nature of some path dependence studies by providing specific mechanisms by which a particular “path” of health policy becomes “dependent.” Adopting additional services becomes more difficult when subnational governments or organized interests have time to develop alternative institutional arrangements for the delivery of additional health services, because their investment in set up costs means they will provide greater opposition to national programs. Additional services are also more difficult in the face of politician’s enhanced knowledge (and concern) about health care costs following the adoption of the first service, and when elite ideas and public expectations work to maintain the low level of discussion and awareness of services that were initially lower-priority. The new insight is that these mechanisms are likely to prevent the type of policy development that incrementalism has taught us to expect.  145 This occurs because even small additional steps that were part of the initial policy design become, over time, deviations from the path of policy. This is not to say the incrementalism prevents all later policy change: indeed, the adoption of medical insurance in Canada, nine years after hospital insurance, and the adoption of public hospital and medical insurance in Australia more than twenty years after pharmaceutical benefits, demonstrate that this is not the case. My findings suggest that these types of changes to policy only occur under specific conditions, however. As an incremental process produces barriers to change over time, small additional steps that might not have been radical earlier in the process of policy development come to require a combination of centralized institutional authority, or a temporary lessening of institutional fragmentation, new principled ideas on the part of policy elites, and a corresponding increase in public attention and support. The insight regarding the role of ideas over time is closely connected to this argument about the implications of incrementalism, since the reciprocal relationship between elite ideas and public expectations as policy develops is a key factor in preventing additional steps in an incremental process (such as the adoption of additional health services) and in determining the flexibility of existing policy (how services that were successfully implemented are reformed). Both principled ideas on the part of elites, and electoral motivations from an attentive and supportive public are necessary for countries to undertake a radical, simultaneous approach to policy development. However, for many policy areas including health, public expectations are unlikely to develop in the absence of some prompting from the elite, and even a clear, principled idea about how a policy should work can rarely be sustained on the agenda in the absence of some attention and support from voters. Furthermore, an incremental process, particularly one where the steps are not attempted in quick succession, means there are more opportunities for policy elites to develop “blind spots” about a policy area, and reject attempts to address lower priorities because they do not conform to previous expectations about the nature of the problem and plausible or necessary solutions. As Jacobs argues, over time policymakers develop mental models that “channel their reasoning toward certain causal probabilities and obscure others from view” (Jacobs 2009, 273). Thus, the reciprocal relationship between the elite ideas and public expectations becomes more important over time, and is a key element in the self-reinforcing nature of low priorities in an incremental process of policy development. I expect this dynamic to be present in a variety of policy areas, which suggests directions for future research on the bounded rationality of both policymakers and voters, and further investigation of the ways elites and publics construct policy problems.  146 Directions for future research The research period ends in the early 1970s, after the rejection of the final federal government proposal for nation-wide pharmaceutical insurance in Canada, and after the first reforms to pharmaceutical programs in the UK and Australia had been implemented. What happened next? Did Canada maintain its outlier status with regards to pharmaceutical programs? Did subsequent reforms in the UK and Australia conform to expectations about the relative flexibility of programs adopted under an incremental process, versus a simultaneous approach? How different are the countries now? Here I provide some preliminary answers and suggest a research program that will address them more fully. Canada still lacks a nation-wide, universal program for prescription drugs, but various provincial drug programs began to be introduced in the mid-1970s, and have developed in sometimes strikingly different directions. About 50 percent of Canadian pharmaceuticals are funded privately, although the division between public and private payment for pharmaceuticals varies by province. In the UK, retrenchment in prescription services appears minimal, despite the introduction of a broader range of cost management policies. Australia, on the other hand, implemented a more aggressive program of price control, formulary management, and increases to patient charges that are currently at a level where many common, inexpensive prescriptions fall below the subsidy level, and so are paid for by patients out of pocket. This convergence and divergence over time in the pharmaceutical programs of three similar welfare states raises a distinct theoretical question: how do early decisions about policy development, particularly the choice to either attempt radical reform or to proceed incrementally, constrain and shape later policy change? The next stage of the research will pick up on the findings concerning the outcomes of the earliest reform attempts in the UK and Australia and examine the implications of incremental versus simultaneous policy development over the long term, by comparing the two countries from the 1970s to the present day. It will also include an investigation of provincial drug programs in Canada, against the backdrop of previous national attempts at policy development. The research will ask how the more limited provincial policies were chosen, and how they affect the prospects for change at the national level.  A further step in the research program is to apply the argument to the process of health policy development in other countries, including those with welfare states that differ from Canada, the UK and Australia more markedly. This would require taking into account the longer histories of government involvement in health benefits or insurance, in some cases, and thus a greater influence from past policies. In more general terms, this project anticipates a  147 research program focused on identifying policy processes that are subject to an incremental versus simultaneous choice, or where there a possibility of a middle option, between radical action and doing nothing. The factors that make a policy path dependent are expected to apply widely, and so research should ask whether other policies that can be easily disaggregated (for example, child care policy, or perhaps some broad environmental policies) are also subject to stalled development processes when approached incrementally. It should consider whether we observe more radical policy development in the presence of centralized authority, principled ideas and electoral motivation, either at the beginning of the policy process, or later, when development has already stalled. In this way, we can develop a more complete understanding of policy development that acknowledges the realities of the process – small steps may often be necessary – while providing a theoretically satisfying explanation of why this is the case, and addressing the predictable elements of policy success, failure, and change over time.    148 References  Archival references Archival institutions are identified in the footnotes using the abbreviations provided. Under each institution, I list the record groups cited.  Library and Archives Canada (LAC)  RG 2: Privy Council Office  RG 29: Department of National Health and Welfare  MG 32-B5: Brooke Claxton fonds  RG 33-78: Royal Commissions – Royal Commission on Health Services  RG 27: Central registry files  Queen’s University Archives  Thomas Worral Kent Papers.  National Archives of Australia (NAA)  A432: Attorney-General's Department, Central Office.  A463 and A5619: Prime Minister's Department.  A571: Department of the Treasury, Central Office.  A1361: Commonwealth Office of Education, Central Office.  A1658 and A1928: Department of Health, Central Office.  A2700: Secretary to Cabinet/Cabinet Secretariat (Curtin, Forde and Chifley Ministries).  A3300: Australian Legation, United States of America [Washington].  A4933: Secretary to Cabinet/Cabinet Secretariat (Fourth and Fifth Menzies Ministries).  A4940: Secretary to Cabinet/Cabinet Secretariat (Menzies and Holt Ministries).  A5908: Cabinet Office (McMahon Ministry).  A5915 and A5931: Cabinet Office (Whitlam Ministries).  A5954: Department of Defence, Central Office   149 M450: Personal Papers of Prime Minister McMahon.  M1274: Personal Papers of Prime Minister Fraser.  M2568: Personal Papers of Prime Minister Holt.  The National Archives of the United Kingdom (TNA)  CAB: Cabinet office  PREM: Records of the Prime Minister’s Office.  MH: Ministry of Health  T: Treasury  Public opinion polls  Canadian Institute of Public Opinion/Gallup Poll of Canada, Public Opinion News Service Release • 8 April 1942 • 22 May 1943 • 8 April 1944 • 24 July 1948 • 6 July 1949 • 13 July 1949 • 1 August 1953 • 14 November 1953 • 18 April 1956 • 21 April 1956 • 27 March 1963 • 12 June 1963 • 8 January 1964 • 5 January 1966  Gallup Poll (Australia) Public Opinion News Service Release • May-June 1943, Polls 124-131 • August-September 1943, Polls 141-152 • October 1943, Polls 153-161 • July 1944, Polls 205-212 • December 1944-January 1945, Polls 241-248 • May 1945, Polls 264-271 • November 1945, Polls 304-313 • December 1945-January 1946, Polls 314-326  • September 1946, Polls 382-397 • March 1947, Poll 416-425 • February-March 1948, Polls 487-491 • May 1948, Polls 511-528 • February-March 1949, Polls 569-578 • May-June 1949, Polls 590-599 • March-April 1950, Polls 662-676 • June-July 1950, Polls 690-699 • February-March 1951, Polls 744-755 • March-May 1969, Polls 2105-2118 • August-September 1969, Polls 2132- 2142  British Institute of Public Opinion/Gallup Poll, 17 May 1948. British Institute of Public Opinion/Gallup Poll, 13 November 1950.  Hansard (United Kingdom)  Prime Minister Clement Attlee “The Economic Situation (Government Proposals).” 24 October 1949. H.C. Debates 1948-1949 Vol 468, c.1016-1019.   150 Misters Thorpe, Winnick and Pavitt, 16-17 January 1968. H.C. Debates 1967-1968. Vol. 756 c.1578-1814.  Harold Wilson on “Public Expenditure,” 16-17 January 1968. H.C. Debates 1967-1968. Vol. 756 c.1578-1814.   Published references  Arthur, W. Brian. 1994. Increasing returns and path dependence in the economy. Ann Arbor: University of Michigan Press.  Attlee, C. R. 1937. The Labour party in perspective. London: Victor Gollancz Ltd.  Australia. 2003. Commonwealth of Australia Constitution Act.  Bealy, Frank. 1970. The social and political thought of the British Labour party. London: Weidenfeld and Nicolson.  Beilharz, Peter. 1994. Transforming labor: Labor tradition and the labor decade in Australia. Cambridge: Cambridge University Press.  Bélanger, Claude. Canadian federalism, the tax rental agreements of the period of 1941-1962 and fiscal federalism from 1962 to 1977. Marianopolis College: Studies on the Canadian Constitution and Canadian Federalism, 2001, (accessed 4 November 2009).  Bennett, Andrew, and Alexander L. George. 1997. Process tracing in case study research. Paper presented at the MacArthur Foundation Workshop on Case Study Methods, Belfer Center for Science and International Affairs (BCSIA), Harvard University, October 17-19 ed., (accessed May 8, 2007).  Bennett, Andrew, and Colin Elman. 2006. Complex causal relations and case study methods: The example of path dependence. Political Analysis 14, (3) (July 1): 250-67.  Biggs, Amanda. 2004. Medicare - background brief. Parliament of Australian - Parliamentary Library, (accessed March 15, 2007).  ———. 2003. The pharmaceutical benefits scheme - an overview. Parliament of Australia - Parliamentary Library, (accessed February 25 2007).  Blyth, Mark M. 2003. Structures do not come with an instruction sheet: Interests, ideas and progress in political science. Perspectives on Politics 10, (4) (December): 695-706.  ———. 1997. Review: "any more bright ideas?" the ideational turn of comparative political economy. Comparative Politics 29, (2) (Jan.): 229-50.   151 Booth, Alan, ed. 1995. British economic development since 1945. Manchester: Manchester University Press.  Booth, Alan, and Sean Glynn. 1979. The public records and recent British economic historiography. The Economic History Review 32, (3) (August 1979): 303-15.  Boychuk, Gerard. 2008. National health insurance in the United States and Canada: Race, territory and the roots of difference. Washington, D.C.: Georgetown University Press.  Cairns, Alan C. 1977. The governments and societies of Canadian federalism. Canadian Journal of Political Science 10, (4) (Dec.): 695-725.  Canada. 1955. Federal-provincial conference 1955: Plenary session October 3. Ottawa: Queen's Printer.  ———. 1945. Dominion-provincial conference on post-war reconstruction: Plenary session. Ottawa: King's Printer.  Canada. Advisory Committee on Health Insurance. 1943. Health insurance; report. Presented on March 16, 1943 [Final report of the Heagerty Advisory Committee on Health Insurance]. Ottawa: King's Printer.  Commission on Pharmaceutical Services. 1971. Pharmacy in a new age: Report of the commission of pharmaceutical services. Toronto: Canadian Pharmaceutical Association.  Commission on the Future of Health Care in Canada. 2002. Commission on the future of health care in Canada: Final report (Romanow report). Ottawa: The Commission.  Commonwealth Department of Health and Ageing. 2007. Pharmaceutical benefit scheme (PBS) reforms, m (accessed April 25 2007).  Craig, Fred W. S. 1982. Conservative and Labour party conference decisions, 1945-1981. Chichester, West Sussex: Parliamentary Research Services.  ———. 1975. British general election manifestos, 1900-1974. Revised and enlarged ed. London: Macmillan.  Crossman, R. H. S. 1975. The diaries of a cabinet minister. London: Hamilton and Cape.  Crowley, F. K. 1973. Modern Australia in documents: 1939-1970. Vol. 2. Melbourne: Wren.  Department of National Health and Welfare (Research and Statistics Division). 1963. Provision, distribution and cost of drugs in Canada: Report to the Royal Commission on Health Services. Ottawa: Queen's Printer.  Department of National Health and Welfare (Research Division). 1954. Voluntary medical care insurance: A study of non-profit plans in Canada. Ottawa: Published by the authority of the Hon. Paul Martin, Minister of National Health and Welfare.  152  Director of Investigations and Research. 1961. Material collected for submission to the RTPC in the course of an inquiry under section 42 of the combines investigation act, relating to the manufacture, distribution and sale of drugs ("Green Book"). Ottawa: Department of Justice, Appendix Q to RTPC Report.  Downs, Anthony. 1957. An economic theory of democracy. New York: Harper and Row.  Eden, Lorraine. 1989. Compulsory licensing and Canadian pharmaceutical policy. In International business in Canada: Strategies for management, ed. Alan M. Rugman, 245-267. Scarborough, Ontario: Prentice-Hall Canada.  Esping-Anderson, Gosta. 1990. The three worlds of welfare capitalism. Cambridge: Polity Press in association with Basil Blackwell.  Federal/Provincial/Territorial Ministerial Task Force. 2006. National pharmaceuticals strategy progress report. Health Canada, dgps/pdf/pubs/2006-nps-snpp/2006-nps-snpp_e.pdf (accessed May 9, 2007).  Flood, Colleen M. 2002. Prescriptions from down under: Can Canada import Australia’s pharmaceutical benefits scheme? IRPP conference: Toward A national strategy on drug insurance: Challenges and priorities. Draft conference paper, (accessed January 28, 2007).  Forest, P. G. 2004. To build a wooden horse...integrating drugs into the public health system. HealthcarePapers 4, (3): 22-6.  Forget, Claude E. 2002. The Quebec experience: Lessons to be learned. IRPP conference: Toward A national strategy on drug insurance: Challenges and priorities. Draft conference paper, (accessed January 28, 2007).  George, Alexander L., and Andrew Bennett. 2005. Case studies and theory development in the social sciences. BCSIA studies in international security. Cambridge, Mass.: MIT Press.  Gilbert, Christopher D. 1980. "There will be wars and rumours of wars": A comparison of the treatment of defence and emergency powers in the federal constitutions of Australia and Canada. Osgood Hall Law Journal 18, (3): 307-35.  Gillespie, James A. 1991. The price of health: Australian governments and medical politics 1910-1960. Cambridge: Cambridge University Press.  Gorecki, Paul K. Regulating the price of prescription drugs in Canada: Compulsory licensing, product selection, and government reimbursement programmes. Economic Council of Canada, 1981.  Gray, Gwendolyn. 1991. Federalism and health policy: The development of health systems in Canada and Australia. Toronto: University of Toronto Press.   153 Great Britain. Inter-departmental Committee on Social Insurance and Allied Services. 1942. Social insurance and allied services :Report (Beveridge Report). Papers by command 6404. London: H. M. Stationery Office.  Greener, Ian. 2005. The potential of path dependence in political studies. Politics 25: 62-72.  Grootendorst, Paul. 2002. Beneficiary cost sharing under Canadian provincial prescription drug benefit programs: History and assessment. The Canadian Journal of Clinical Pharmacology 9, (2) (Summer): 79-99.  Hacker, Jacob S. 1998. 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, (Spring): 57-130.  Hall, Peter A. 2006. Systematic process analysis: When and how to use it. European Management Review 3: 24-31.  ———. 2003. Aligning ontology and methodology in comparative research. In Comparative historical analysis in the social sciences, eds. James Mahoney, Dietrich Rueschemeyer. Cambridge: Cambridge University Press.  ———. 1993. Policy paradigms, social learning, and the state: The case of economic policymaking in Britain. Comparative Politics 25, (3) (Apr.): 275-96.  Harrison, Kathryn, ed. 2006. Racing to the bottom? Provincial interdependence in the Canadian federation. Vancouver: UBC Press.  ———. 1996. Passing the buck: Federalism, and Canadian environmental policy. Vancouver: UBC Press.  Hickey, Daniel J. 2004. The evolution of public drug benefits in Saskatchewan: 1945-2002. University of Regina: Saskatchewan Institute of Public Policy.  Huber, Evelyne, and John D. Stephens. 2001. Development and crisis of the welfare state. Chicago and London: University of Chicago Press.  Hunter, Thelma. 1965. Pharmaceutical benefits legislation, 1944-50. Economic Record 41, (95) (09/24): 412-25.  Hunter, Thelma A. 1964. Phamaceutical benefits legislation (1955-50): Doctors, government and the constitution. Paper presented at the Sixth APSA Conference, Sydney.  Immergut, Ellen M. 1992. The rules of the game: The logic of health policy-making in France, Switzerland, and Sweden. In Structuring politics: Historical institutionalism in comparative analysis., eds. Sven Steinmo, Kathleen Thelen and Frank Longstreth. New York: Cambridge University Press.  Jacobs, Lawrence R. 1993. The health of nations: Public opinion and the making of American and British health policy. Ithaca and London: Cornell University Press.   154 Jacobs, Alan M. 2009. How do ideas matter?: Mental models and attention in German pension politics. Comparative Political Studies 42, (2) (February 1): 252-79.  Jacobzone, S. 2000. Pharmaceutical policies in OECD countries: Reconciling social and industrial goals. OECD Publishing, 40.  Johnson, Carol. 1989. The labor legacy: Curtin, Chifley, Whitlam, Hawke. Sydney: Allen & Unwin.  Kanavos, Panos. 2002. Pharmaceutical regulation in Europe. IRRP conference: Toward a national strategy on drug insurance: Challenges and priorities. Draft conference paper, (accessed January 28, 2007).  King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing social inquiry: Scientific inference in qualitative research. Princeton, N.J.: Princeton University Press.  Kingdon, John. 1995. Agendas, alternatives and public policies. 2nd ed. New York: Harper Collins.  Klein, Rudolf. 1995. The new politics of the NHS. 3rd ed. New York: Longman.  Lang, Ronald William. 1974. The politics of drugs : A comparative pressure-group study of the Canadian Pharmaceutical Manufacturers Association and the Association of the British Pharmaceutical Industry, 1930-1970. Saxon house studies. Farnborough, Hants.; Lexington, Mass.: Saxon House; Lexington Books.  Lexchin, J. 2004. Stop the merry-go-round. HealthcarePapers 4, (3): 29-34.  Lindblom, Charles. 1979. Still muddling, not yet through. Public Administration Review 39, (6): 517- 26.  Livingston, William S. 1956. Federalism and constitutional change. Oxford: Clarendon.  Lofgren, Hans. 2001. Pharmaceutical benefits in Australia and Sweden: Welfare policy and the cost of prescription drugs. Australian Journal of Social Issues 36, (3) (Aug): 207-20.  Lynch, Julia. 2006. Age in the welfare state. Cambridge: Cambridge University Press.  Mahoney, James. 2000. Path dependence in historical sociology. Theory and Society 29, (4): 507-48.  ———. 2007. Qualitative methodology and comparative politics. Comparative Political Studies 40, (2): 122-144.  Maioni, Antonia. 1998. Parting at the crossroads : The emergence of health insurance in the United States and Canada. Princeton studies in American politics. Princeton, NJ: Princeton University Press.  Mann, John. 2004. Life saving drugs: The elusive magic bullet. Cambridge: Royal Society of Chemistry.  Marchildon, Greg. 2006. Federal pharmacare: Prescription for an ailing federation? Inroads Canadian Journal of Opinion Winter/Spring, (18): 94-108.  155  Marsh, Leonard Charles. 1943. Report on social security for Canada, prepared for the advisory committee on reconstruction. Ottawa: King’s Printer.  Martin, Paul. 1985. A very public life: So many worlds. Vol. 2. Toronto: Deneau.  Matthews, R. L., and W. R. C. Jay. 1972. Federal finance: Intergovernmental financial relations in Australia since federation. Melbourne: Nelson.  McAllister, Ian, and Rhonda Moore. 1991. Party strategy and change: Australian electoral speeches since 1946. Sydney: Longman Cheshire.  McMullin, Ross. 1991. The light on the hill: The Australian Labor Party, 1891-1991. Oxford: Oxford University Press.  Mendelson, Danuta. 1999. Devaluation of a constitutional guarantee: The history of section 51(XXIIIA) of the Commonwealth Constitution. Melbourne University Law Review 14, (accessed 30 March 2010).  Ministry of Health. 1944. A national health service (“White Paper”). London: Great Britain, Parliamentary Papers, House of Commons, Sessional Papers Vol 1943/44-8, CMND 6502.  Ministry of Health (UK) Central Health Services Review. 1959. Final report of the committee on the cost of prescribing (“Hinchcliffe Committee”). London: HMSO.  Moe, Terry M., and Michael Caldwell. 1994. The institutional foundations of democratic government: A comparison of presidential and parliamentary systems. Journal of Institutional and Theoretical Economics 150, (1): 171-95.  Morgan, Steve. The determinants of prescription drug expenditure...and what to do about them. Vancouver: UBC Centre for Health Services and Policy Research, 2008.  ———. 2005. Booming prescription drug expenditure: A population-based analysis of age dynamics. Medical Care 43, (10) (Oct): 996-1008.  ———. 2002. Peeling the onion: What drives pharmaceutical expenditures in Canada? IRPP conference: Toward a national strategy on drug insurance: Challenges and priorities. Draft conference paper, (accessed January 28, 2007).  ——— and D. J. Willison. 2004. Post-Romanow pharmacare: Last-dollar first...first-dollar lost? HealthcarePapers 4, (3): 10-20.  Mossialos, Elias, Monique Mrazek, and Tom Walley, eds. 2004. Regulating pharmaceuticals in Europe : Striving for efficiency, equity, and quality. European Observatory on Health Systems and Policies series. Maidenhead, Berkshire: Open University Press.  National Forum on Health. 1997. Canada health action: Building on the legacy. Ottawa: Health Canada, Directions for a Pharmaceutical Policy in Canada,  156 sss/pubs/renewal-renouv/1997-nfoh-fnss-v2/legacy_heritage7_e.html (accessed July 12, 2007).  Page, Scott E. 2006. Path dependence. Quarterly Journal of Political Science 1: 87-115.  Pater, John E. 1981. The making of the National Health Service. King's fund historical series. London: King Edward's Hospital Fund for London.  Permanand, Govin, and Elias Mossialos. 2005. Constitutional asymmetry and pharmaceutical policy-making in the European Union. Journal of European Public Policy 12, (4) (08//): 687- 709.  Perry, J. Harvey. 1989. A fiscal history of Canada: The postwar years. Canadian tax paper no. 85. Toronto: Canadian Tax Foundation.  Pierson, Paul. 2004. Politics in time: History, institutions, and social analysis. Princeton, N.J.: Princeton University Press. ———. 2000. Increasing returns, path dependence, and the study of politics. American Political Science Review 94: 251-67.  ———. 1995. Fragmented welfare states: Federal institutions and the development of social policy. Governance 8, (4): 449-78.  ———. 1994. Dismantling the welfare state? : Reagan, Thatcher, and the politics of retrenchment. Cambridge studies in comparative politics. Cambridge, England ; New York: Cambridge University Press.  Powell, Martin. 1994. The changing blueprints of the British NHS: The White Papers of 1944 and 1989. Health Care Analysis 2: 111-7.  Quirke, Viviane. 2005. From alkaloids to gene therapy: A brief history of drug discovery in the 20th century. In Making medicines: A brief history of pharmacy and pharmaceuticals, ed. Stuart Anderson, 177-201. London, UK; Chicago: Pharmaceutical Press.  Rivett, Geoffrey. 1998. From cradle to grave: Fifty years of the NHS. London: King's Fund.  Robinson, Ray, and Anna Dixon. 1999. Health systems in transition: United Kingdom. Copenhagen: European Observatory on Health Care Systems, (accessed February 12,2007).  Roskam, John. 2001. Liberalism and social welfare. In Liberalism and the Australian federation., ed. J. R. Nethercote, 267-286. Sydney: Federation Press.  Ross, Fiona. 2007. Questioning path dependence theory: The case of the British NHS. Policy and Politics 35, (4): 591-610.  Royal Commission on Health Services. 1964. Royal Commission on Health Services: 1964: Volume I (Hall Commission). Ottawa: Queen's Printer.   157 Ryan, Michael. 1973. Free at time of use; a study of charges and costs in the National Health Service. Social Policy & Administration 7, (3) (09/24): 219-34.  Scotton, R. B., and C. R. MacDonald. 1993. The making of Medibank. Kensington, NSW: School of Health Services Management, University of New South Wales.  Senate Standing Committee on Social Affairs, Science & Technology. 2002. The health of Canadians, the federal role, vol. six : Recommendations for reform, final report on the state of the health care system in Canada (Kirby Report) Ottawa the Standing Committee.  Sloan, Clyde. 1995. A history of the Pharmaceutical Benefits Scheme, 1947-1992. Canberra: Australian Government Public Service.  Smiley, D. V., ed. 1963. The Rowell-Sirois report: An abridgement of book I of the Royal Commission Report on Dominion-Provincial Relations. Toronto: McClelland and Stewart Ltd.  Smiley, D. V. 1962. The Rowell-Sirois report, provincial autonomy, and post-war Canadian federalism. The Canadian Journal of Economics and Political Science 28, (1): 54-69.  Taylor, Malcolm G. 1987. Health insurance and Canadian public policy : The seven decisions that created the Canadian health insurance system and their outcomes. Canadian public administration series. 2nd ed. Toronto: Institute of Public Administration of Canada.  Thane, Pat. 1996. Foundations of the welfare state. 2nd Ed. ed. London and New York: Longman.  Thelen, K. 1999. Historical institutionalism in comparative politics. Annual Review of Political Science 2: 369-404.  Trachtenberg, Marc. 2006. The craft of international history: A guide to method. Princeton, NJ: Princeton University Press.  Tsebelis, George. 1995. Decision making in political systems: Veto players in presidentialism, parliamentarism, multicameralism and multipartyism. British Journal of Political Science 25, (3).  Tuohy, Carolyn J. 1999. Accidental logics : The dynamics of change in the health care arena in the United States, Britain, and Canada. New York: Oxford University Press.  Weaver, R. Kent. 1986. The politics of blame avoidance. Journal of Public Policy 6, (4): 371-98.  ———, and Bert A. Rockman, eds. 1993. Do institutions matter? Government capabilities in the United States and abroad. Washington D.C.: Brookings Institution.  Webster, Charles. 2002. The National Health Service: A political history. New ed. Oxford: Oxford University Press.  White House press release. 2006. Strengthening medicare,, (accessed April 25 2007).  White, D. M. 1978. The philosophy of the Australian Liberal party. Richmond, Victoria: Hutchinson.  158  Whitlam, Gough. 1985. The Whitlam government: 1972-1975. Ringwood, Victoria: Penguin.  Wiktorowicz, Mary E. 2003. Emergent patterns in the regulation of pharmaceuticals: Institutions and interests in the United States, Canada, Britain, and France. Journal of Health Politics, Policy & Law 28, (4): 615.  Willison, Donald, Mary E. Wiktorowicz, Paul Grootendorst, Bernie O’Brien, Mitchell Levine, Raisa Deber, and Jeremiah Hurley. 2002. International trends in managing drug plans: Lessons for canada. IRPP conference: Toward a national strategy on drug insurance: Challenges and priorities. Draft conference paper, (accessed January 28, 2007).  Wordsworth, Sarah, Mandy Ryan, Cam Donaldson, and Anthony Scott. 1998. New Labour, new charges?: Will charges curb costs and raise more revenue for the NHS? New Economy 5, (4).  Yee, Albert S. 1996. The causal effects of ideas on policies. International Organization 50, (1) (Winter): 69-108. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items