1Expertise is generally defined as a “situation” resulting from consultation by a client (CRESAL 1985), in which the individual designated as an expert draws upon knowledge and legitimacy acquired in another social, professional, or academic space (Hassenteufel 2008). According to this definition, being an expert is not a job, and certainly not a profession. The expert’s legitimacy is based on the fact that they primarily exercise another professional activity that has specific skills and knowledge attributed to it. Insofar as expertise is presented as a way to justify public decisions by grounding them in “reason,” making them “not only fair and informed but objective, and excluding any arbitrariness” (Robert 2008),  experts will have greater legitimacy when the knowledge they claim to hold is considered objective and neutral. This is what guarantees that their opinion can be free from any suspicion that they are making personal value judgments or defending their own interests. This conception makes it crucial for experts to be able to connect the ideas and recommendations they produce to skills and knowledge that are universally recognized as belonging to a particular profession, or else to an “authenticated scholarly corpus” that can serve as a reference point (Trépos 1996, 68). In such cases, experts rely on recognized, stable, and thoroughly institutionalized disciplines, as close as possible to Merton’s model of “autonomous sciences” (Merton  1973).
2The rise of organizations whose main goal is to produce expertise runs counter to this conception of expertise and the status of the expert. It has become commonplace for think tanks, government agencies, and even centers attached to universities to use “expert” as a job title. People hired as experts to produce expertise are “professional” experts. A space of expertise has emerged that could be called self-referential, in that it is unconnected to a domain of skills acquired in another professional and/or academic space. I will call this the “space of professionalized expertise,” and the actors who make it up “professional experts.”
3In the English-speaking world, this space of professionalized expertise involves what is generally called policy analysis, which has given rise to the role of the policy analyst or policy expert. In the United States, this terminology became widespread in the 1960s to designate the production of a knowledge based on the evaluation of public policies and aimed at producing recommendations that would be directly useful for political decision-making (DeLeon 1988; Radin 2000). Since then, some have observed the rise of a policy analysis “industry,” initiated by the Johnson administration’s planning and evaluation policies (Drucker 1968; Rich 2004, 48). Such activities have continued to grow. Any organization with some sort of relationship to politics—Congress, the White House, the federal government, think tanks, foundations, consulting firms, and interest and pressure groups—have at least one or two posts with such titles, and often several hundred (DeLeon 1988).
4In some countries, this space of professionalized expertise seems to be increasingly important for situations in which expertise is provided to political decision-makers. Such is the result of my study of individuals in the United States who had been consulted as experts by American decision-makers seeking advice on health insurance and reform proposals (see Box 1). Between the 1930s and 1960s, such experts were either physicians (consulted by conservatives) or Social Security Administration officials (consulted by liberals) (Engel 2002; Lepont 2014). In either case they had links to a well-established professional group, which guaranteed them a professional status external to any potential situation of expertise. My study shows that, from the 1970s onward, the profile of the most consulted experts changed: the vast majority of them now worked for think tanks, foundations, or centers that were attached to universities but whose goal was the production of expertise rather than academic knowledge (Lepont 2016b). When asked about their job, they define themselves as “policy analysts” and, more precisely, “health policy analysts”—not as “researchers” and certainly not as “academics.” They belong to that space of professionalized expertise in which experts have no professional status beyond that of being “experts.”
5This article uses the case of health policy analysts to explore the rise of professional experts in Washington, and the conditions in which they became institutionalized. The central question in this process is about the construction of their legitimacy. If the traditional conceptions of expertise are referred to, the basis of these experts’ legitimacy appears problematic. Unlike their predecessors— physicians and Social Security Administration officials, recognized respectively as possessing medical and bureaucratic legitimacy—they cannot claim any links to an initial profession whose legitimacy is well-established. Nor do they possess scientific legitimacy as generally conceived.  The policy analysis they practice does not position itself as a “pure,” “autonomous” body of knowledge regulated exclusively by the strict rules of an academic discipline. Its goal is to produce socially useful knowledge, and they instead claim to take political, economic, and social constraints into account in order to provide political decision-makers with “politically feasible” recommendations (Wildavsky 1979; DeLeon 1988). The health policy analysts I met with fully recognized that their work is hybrid. As Thomas Medvetz (2010) notes, such work borrows from a number of social roles: from that of the academic scholar, of course, but also from that of the policy aide, the business entrepreneur, and the media specialist. The figure of the researcher is an important component, providing analysts with “an indispensable font of authority, as well as a means of symbolic separation from lobbyists, activists, and political aides” (Medvetz 2010, 561). But this is just a partial borrowing from the academic sphere, as the figure of the researcher must be reconciled with the demands of the other roles the analyst embodies.
6So how did health policy analysts successfully get American decision-makers to recognize their jurisdiction (Abbott 1988) over health insurance issues? I will show how affiliation with a specific, well-established body of knowledge was important in this process. Health policy analysts placed great importance on establishing a particular discipline, that of health services research, and contributed much to this process. Like many other disciplines—for instance, management (Pavis 2008), chemical engineering (Grossetti 2000), marketing (Cochoy 2000), health economics in France (Benamouzig 2005), and many sciences of government (Ihl et al. 2003)—health services research emerged in a “context of application” (Pestre 1997; Robert and Vauchez 2010) to meet a particular social demand. Distinctively, however, and unlike in these other disciplines, health policy analysts did not try to make their own subject scientifically autonomous, in spite of their eagerness to institutionalize it as a discipline. Instead, they continued to claim that it was a practical body of knowledge aimed at resolving public problems, one able above all to meet and adapt to political demands and specific political contexts. The feasibility of the proposals their research produced became the key criterion for justifying any decision. The key issue, which will be explored here, is that this does not seem to have been a hindrance to the recognition of these analysts as experts.
7This seemingly contradictory situation leads us back to recent work on expertise showing that an expert group’s success in gaining recognition and achieving institutionalization does not necessarily depend on the autonomy of the body of knowledge with which it is associated. Gil Eyal (2002) was among the first to develop this hypothesis, using the case of Middle Eastern studies in Israel. These have become institutionalized in spite of their lack of autonomy from the secret services, with which they have always had close links. Antoine Vauchez and Cécile Robert (2010) have demonstrated something similar in the case of European studies, which have been closely entwined with the EU’s bureaucratic elite since the 1950s. Lisa Stampnitzky (2013) has shown how the field of expertise on terrorism gained recognition without any discernible institutionalization of a body of knowledge: the definition of the very object in question, terrorism, is extremely vague and closely dependent on current political events. In his discussion of policy analysts employed in think tanks, Medvetz (2012) shows that their position at the intersection of several fields (politics, economics, the media, academia) offers no barrier to the processes of expertise.  To explain this situation, these authors argue for a reexamination of the distinction between “strong” fields, which have the ability to autonomously define the rules, standards, and knowledge they obey, and “weak” fields, in which the definition of these rules is subject to intervention by other, external social spaces (Vauchez 2011). They show how the permeability of boundaries provides an opportunity for relationships with other spaces, and may constitute a resource for groups that want to establish themselves as experts with legitimate knowledge in a public policy domain. As Vauchez writes, “a weak field is weak in terms of definition, but not necessarily in terms of social effects” (Vauchez 2011). These spaces have been redescribed as “liminal” (Eyal 2002) or, more frequently, “interstitial” (Vauchez 2011; Medvetz 2012; Stampnitzky 2013) fields (or spaces, depending on the author)—ones lying at the intersection of multiple spaces that exert pressure on them and from which they draw resources.
8I believe that this approach is extremely useful for studying the process of institutionalization of American health policy analysts, and particularly their involvement in health services research. The hypothesis I present here is that the success of health services research as an “expert body of knowledge” is precisely the result of its lack of interest in autonomy. As will be seen, health services research is a type of knowledge that provides a scholarly foundation for those who lay claim to it, but that is nonetheless well adapted to the demands of expertise. Clients are thereby more likely to receive a timely response that is well adapted to their needs. The experts themselves find they can considerably lessen the unpleasantness caused by situations where they find themselves in tension with, or transgressing, the norms of their initial professional network—something numerous studies have observed (Memmi 1989; Barthe and Gilbert 2005). Health services research is a way of creating a space of professionalized expertise about health insurance that is free from the constraints of academic standards but maintains some of its legitimacy. Health policy analysts have thereby resolved a tension characteristic of relationships between academic spaces and spaces of expertise.
9Finally, this study examines more generally the rise of these spaces of professionalized expertise based on operational disciplines, and asks about their consequences within public space for forms of knowledge.
10In the first part of the article, I describe how a new generation of experts who could not find a place in academia professionalized themselves as health policy analysts and took up powerful expert positions on health insurance in Washington. In the second part, I explore the project to construct a discipline and a professional group for health services research, which began in the 1980s as a way for health policy analysts to cement and institutionalize their position in the expert world. In the third part, I look at the success of this project for health policy analysts who, in spite of—or perhaps thanks to—their discipline’s low degree of autonomy, found within it the material, financial, and symbolic resources they needed.
Box 1. – Research methods
Based on this definition, I identified seventy-three individuals as the most consulted experts between 1970 and 2010. I did so by constructing a database, using two criteria:
1) The number of times the individual was consulted as an expert by the executive or the legislature, in one of the following situations: participation in Congressional hearings, nomination to the main governmental and Congressional expert commissions on health policy (PPRC, ProPAC, MedPAC, the Congressional Budget Office’s [CBO] Panel of Health Advisers, and the National Advisory Council for Health Care Policy, Research, and Evaluation at the Agency for Health Care Research and Quality), and political appointments in Congress or the federal government.
2) The production of gray literature aimed at formulating tools and programs for health insurance reform, as determined by a review of authors in two specialized expert journals (Health Affairs and the “Perspective” section in the New England Journal of Medicine).
The qualitative component consisted of seventy-eight in-depth interviews, conducted with two thirds of those I identified as the most consulted experts, along with more marginal experts and those who worked with them, including congressional staffers and senior officials. These interviews focused on the actors’ representations, the types of interactions they had with their clients, and their professional and intellectual careers. To prepare for and to complement these, I examined the resumes of the experts and the abundant gray literature they produced, including articles, reports, and memos.
I used a number of these interviews to find out about the history of health services research, which has so far been studied very little, as well as the archives of the Association for Health Services Research during a trip to Washington, and transcripts of two series of interviews with the association’s founding members. The first of these was carried out in 1998 by a historian, Edward Berkowitz, and was commissioned by the Department of Health. Although these interviews are available online, Berkowitz confirms that they have never been used except for a brief report to the department, also available online (Berkowitz 1998).  The second interview series, from 2008, was carried out by Jennifer Muldoon, who has worked at the association since 1992. These interviews have never been published or used, and were made available to me by Muldoon following an interview.
The rise of a new type of expert: Health policy analysts
11In the late 1960s and early 1970s, a new sort of health insurance expert began to emerge in Washington. Unlike those who had dominated the sphere of expertise in the preceding period (Engel 2002), they were not physicians or officials. Most were young economics PhDs who had been trained in neo-classical microeconomics and had specialized in health insurance during their studies or their first job. They initially came to Washington fresh from university, for internships or their first jobs, and did not necessarily intend to stay. But instead of returning to university to become professors, they chose to break with academia and embrace expert careers as health policy analysts. The kind of professional activity they took up was new at the time: they became what I have chosen to call “professional experts.”
12The Nixon and Carter administrations hired them for expert strategic positions on health insurance (Lepont 2014), and in the 1970s they became the leading experts on American health insurance, taking the place of physicians and Social Security Administration officials. 
The children of the Program Planning and Budgeting System
13The rise of these new experts was linked closely to the Program Planning and Budgeting System (PPBS), established by the Johnson administration in the 1960s, and the subsequent growth of the policy analysis sector. Beginning in this period, significant amounts of money were allocated for evaluation studies in areas where they previously did not exist.  The aim was to evaluate social impact relative to the costs and objectives of current and planned programs (ex-post and ex-ante evaluation), and therefore guarantee maximum efficiency in the allocation of resources. Health policy was targeted in particular because the two public insurance schemes, Medicare and Medicaid, were by far the most expensive programs.
14The great amount of importance that the federal state placed on evaluation studies gave rise to what some were already calling the policy analysis “industry.”  Because most of the studies were assigned to nominally independent bodies (Banfield 1980), numerous organizations were created specifically to respond to government tenders.  Universities began offering courses to train people for the new role of policy analyst. The New York Times was already claiming in 1970 that “policy science” was a “new field of study,” citing as evidence the number of new schools and courses with this phrase in their title (Dluhy et al. 1981).
15As a body of knowledge, policy analysis is dominated by microeconomics, from which it draws its main concepts and tools: efficiency, opportunity costs, incentives, and so on (Fourcade 2009; Rhoads 1985).  The rise of policy analysis encouraged the hiring of microeconomists, or at least those trained in microeconomics.
16The emergence of health policy analysts in Washington in the early 1970s—all of them microeconomists—was a direct result of this new situation. Most of those who had written dissertations on health insurance received funding from a program established under Johnson by the Department of Health to encourage microeconomic work in the area.  They were among the first in their universities to adopt this approach. Those who specialized immediately after their studies did so to meet demand by their employers, non-governmental centers of expertise that lacked staff trained in such topics.
Constraints and opportunities in the labor market in the early 1970s
17They may have had PhDs in economics, but young graduates who had specialized in health insurance had relatively little chance of finding university jobs at the start of the 1970s.
We—people like Stuart Altman, Joseph Newhouse, Paul Ginsburg, and Karen Davis—were basically the first cohort of health economists. And none of us thought we’d stay in academia, because we thought what we were doing wouldn’t be recognized there. In fact, many of us were sort of resigned to the fact we would probably never be professors. 
19In this generation, it was relatively common to have difficulty finding jobs as professors: mass enrollment in higher education had led to a sharp rise in the number of PhDs, and universities alone could not absorb this (Gaiger 1993). Furthermore, healthcare was still not recognized as a legitimate object of study in economics departments. Aside from an isolated article by the economist Kenneth J. Arrow (1963), research on the topic had not led to any theoretical advances—unlike education, for instance, which was at the center of the theory of human capital (Teixeira 2000). Microeconomic health research was still seen as applied work—valued little in academia—and was therefore marginal within university research and accorded only a small measure of academic respect. By contrast, the policy analysis industry that developed around the federal government offered young PhDs numerous job opportunities: their microeconomic training was heavily sought after by every organization wanting to position itself in the market for health policy evaluation and expertise. As well as being an open labor market where health policy analysts could easily find work, Washington provided a stimulating environment. As microeconomists interested in questions of allocating (necessarily scarce) resources, they were particularly appealing to American policymakers of the period. Politicians consulted them regularly, and they even received political appointments in the federal government (Lepont 2014). As the interview extracts below show, the feeling of being part of the development of public policy and government action was a motivating factor in these individuals’ decision to stay in Washington as health policy analysts.
Why did I stay there [at Brookings]? Because your research is used for something! It was a much more stimulating environment [than the university]. There were opportunities to testify before Congress, to take part in Congressional committees, and to work for influential organizations like the Institute of Medicine or the Research Review Committee.  I wanted to work for policymakers. Often, research is too separate from any public policy implications. I wanted to translate my research so that it was accessible to policymakers. 
You started your career as a university professor. Why did you choose to join the government? There was a lot of hope at the time in the possibility of reforming the national health insurance system. I was recruited by the cabinet secretary to work in his group. It was the atmosphere at the time and the fact that I was recruited by one of the cabinet that made me leave the university. Otherwise I would probably have stayed. 
21Different factors came together, forming a set of constraints and opportunities that explain why these young PhDs decided to remain in Washington and embrace careers as policy analysts—a role that at the time was still new.
Box 2. – Joseph Newhouse, a young Harvard PhD at the RAND Corporation
Newhouse was exclusively considering an academic career at the end of his PhD in the Harvard economics department in 1969. He did not intend to specialize in health insurance, which he saw as one subject among many that microeconomic theory could be applied to: “I had spent a summer at RAND and I liked it, but I imagined that I would go back to Harvard and get a job as a professor… I never expected to get into health care as deeply as I ended up doing. First of all as an economics student, one was somewhat socialized not to specialize in a particular field like that. The mark of distinction was that one could take microeconomics and apply it to a whole variety of topics.”  On his account, the position at RAND was initially a disappointment: “The year I was going to finish my dissertation, there was a new director sat Harvard who decreed that no students coming out of Harvard would be hired. So I called up RAND and said if they were still interested in me, I was interested in them.” 
RAND offered better, more comfortable research conditions than Harvard. When he was there for the first time, as a PhD student, he had been surprised by the freedom he was given and the material resources available (particularly computing resources). He claimed that his research on the Health Insurance Experiment (HIE) would not have been possible in a university: “I couldn’t have carried that experiment out in a university. It required too much financial and logistical support.” 
Newhouse’s career demonstrates the unfavorable academic context young health economists faced in the late 1960s, caused in part by the limited respect their highly applied domain received. Thinks tanks welcomed them and offered substantial material resources.
Among the health policy analysts I met with, Newhouse was one of those who remained most closely attached to academia. As will be shown, he attempted to “academize” health economics in the early 1980s. He did eventually take up a role at Harvard—not in the economics department but in the Medical School, which valued expertise highly (see below). Newhouse maintained close links to the worlds of politics and expertise. In the 1990s, he was appointed president of a major expert Congressional commission on health insurance.
The sense of belonging to a distinctive professional community
22By opting for careers as policy analysts or policy experts, these individuals took up a distinctive sort of professional activity. They emphasize how different this was from academia. Prima facie, there were ways in which their work differed little from academic work: they had to stay up-to-date and broaden their knowledge by reading books and articles, and they had to produce (or supervise the production of) new knowledge about their field of specialization, health insurance, addressing questions about care access, costs, hospital operations, the behavior of patients toward physicians, and so on. This knowledge activity was distinguished from academic work by its exclusive orientation toward producing solutions to the problems identified, and formulating recommendations for political decision-makers. These experts therefore published memos, policy briefs, and articles in expert journals. They also published books in the 1970s, but with non-academic presses. Much of their time was given to events (meetings, conferences, and lunches) with health sector actors and/or political decision-makers. One of the main expectations of the expert organizations that employed them was that they communicate with the media—something universities also valued, but did not require to the same extent. Similarly, there was far more active demand for providing advice to decision-makers nominated to official positions or, less formally, when they were running for election. 
23The actors I interviewed emphasized the gap between the logics of these two directions: a career as an academic or as a policy analyst. They generally defended the value of the one they had chosen. As one said, “Most academics write for other academics, they’re talking to the academic world. Their life consists of writing papers for other academics.”  Another remarked: “If you only write on the academic model, you might get a Nobel Prize, but you won’t have much influence. And then what? Have you changed anything?”  Using the traditional opposition between expertise and academic knowledge, they defended policy analysis as a body of knowledge directly useful for decision-makers—unlike academic knowledge, which is grand but practically ineffective. Much as Medvetz (2010) observed in his study of think tank policy analysts, the actors themselves experienced the decision to pursue this career as one that contrasted with academia. This choice provided the basis for a feeling of being part of a separate group, made up specifically of individuals invested in the domain of expertise, with a shared vision of their work.
If you’ve worked for a government, you’ve necessarily been interested in the question of making research more useful for decision-making. First of all, experiences like that make you understand who’s got power and who doesn’t. Other questions are important: the way evidence is presented in government is very different from the way it’s presented at a university. I’ve briefed the vice-president, but I’ve never seen anyone present a model. I’ve never met a president or a cabinet secretary who calculates! To understand how politicians make decisions, elected or not, you have to work very close to government.
So the people you meet and who make up your network are academics who are really interested in understanding how politicians take decisions and who want to take part in that battle. They know what’s important or not for the president, for Congress, or for a cabinet secretary. You meet all these people. They haven’t spent their whole career in government. But what connects you, it’s that you’ve been through there and understood the reality of the decision-making process. Anyone who saw Nixon’s office remembers it their whole life! You’re part of a group that knows their experience is different.
So there was this little group I met because I was interested in the impact of policy science on the government, and that’s not something academics typically spend much time on. 
25The feeling of belonging to a specific group is based on the conviction that they shared a distinctive personal interest in public action that set them apart from academia, as well as an unusual grasp of the political world’s rules and standards. This knowledge was based on a shared experience of the political world, which had come from jobs in which they frequently encountered decision-makers. This was accompanied by a shared approach that viewed adaptation to the political world’s particular constraints as imperative—and entirely legitimate if one wanted to be effective at all. They were “professional” experts in the sense that they had no job beyond that of being experts. They absolutely had to occupy situations of expertise regularly if they were to maintain their positions in the organizations that employed them. But health policy analysts did not make up a profession (Freidson 1970) or even a professional group (Demazière and Gadéa 2010) in the late 1970s. There was no professional association or certificate defining its boundaries (Abbott 1988). Instead, as emphasized in the interview quoted above, they made up an informal network of individuals who knew each other and recognized themselves as forming part of a small group of people who did the same work. This informal network was based on links of sociability, constructed by frequenting the same places and institutions (Lepont 2016a). The political events of the early 1980s pushed them to put more structure in place.
Securing their professional position: Betting on health services research
The contradictory position of health policy analysts in the early 1980s
26In the early 1980s, health policy analysts who had acquired strategic positions in the space of expertise during the 1970s found themselves in a contradictory position. On the one hand, their central role was confirmed, particularly by the adoption of major reforms based on the body of knowledge they had promoted.  On the other, their institutional “rear bases”—non-governmental expert organizations—were threatened by a decline in funding. In 1981, the newly elected President Reagan announced he had little interest in the agencies funding health insurance experts, which he felt were mostly wasting federal money.
27This new political context reveals the fragility of the new leading experts’ professional status. They were connected to non-governmental expert organizations that depended on public funding for their survival, and this put them in a distinctively vulnerable position compared to the experts who had preceded them—physicians and Social Security Administration officials—who had professional jobs to fall back on and a salary independent of their expert work with decision-makers.
28The fragility of their professional position was increased by the weakness of their collective structures. Individually, the economists consulted were recognized as experts. And, as we have seen, they had the sense of belonging to a distinct group of health policy experts. But they did not form a coherent group that could easily be identified from outside. Their knowledge was recognized—it had been the basis for major reforms—but it was not enough to grant them existence as a professional group with a particular, immediately recognizable field of intervention. They decided to use the adoption of these reforms as an opportunity to gain such recognition, but this project went nowhere: “We had a hard time, a very hard time, at the beginning pointing to evidence that we had actually made a difference. The first thing that anybody could point to for at least six years was DRGs.” 
29Once again, recognition of a particular body of knowledge was much less of an issue for Social Security Administration officials and physicians, who enjoyed a legitimacy that was attached directly to their professional activity. Physicians were recognized as possessing this because of their medical practice, and officials were recognized as possessing it because of their knowledge of the workings of bureaucracy and the conditions in which programs were put into operation, acquired in the time they had spent in government. By contrast, health policy experts had neither medical nor bureaucratic legitimacy. To secure their current jobs in the long term, they needed recognition for a particular body of knowledge that identified them as a professional group associated with a particular body of knowledge. They consequently needed to rely more heavily on legitimacy of the academic type, based on “scholarly” knowledge associated with a particular discipline. This is shown by their heavy emphasis on publishing, something physicians and Social Security Administration officials did not require of themselves. The resume of a heavily consulted expert might include several books and more than a hundred articles and book chapters. As Medvetz notes, this was the only way for them to distinguish themselves from lobbyists, activists, staffers, and other political aides (2010, 560-1). As we have seen, they did not have the university support: the discipline they had immediately attached themselves to, health economics, was not sufficiently recognized or established in economics departments. Their position in the early 1980s was enviable from the point of view of access to expert strategic roles, but raised concerns about the long-term viability of their jobs.
30In my view, this contradictory situation explains these actors’ desire to join an identifiable professional group, one that had more institutional resources and support from academia and an association with a particular body of knowledge. Two such projects emerged at the same time. The first involved reforming health economics to turn it into a respected sub-discipline of economics by establishing the conditions for its scientific autonomy. This relied on the traditional path to academic professionalization—the one taken by health economists in France, who had also initially been dependent on political and administrative demand. In the United States, the first to move in this direction was Joseph Newhouse, a well-known health economist at the RAND Corporation (see Box 2), who in 1981 founded the first economics journal specialized in the field, the Journal of Health Economics.
31The second project took a very different approach. It was the work of another health economist, Stuart Altman, also an important figure among health policy analysts: under Nixon, he led the first team of health economists who worked directly with the department to develop reforms. His project was to build links with another group of researchers who produced expertise on questions related to health insurance, and who were broadly labelled “health services researchers.” In 1981, Altman founded a professional association, the Association for Health Services Research (AHSR). The aim was to give health services research the status as a discipline that it had lacked up to that point. But, unlike the path taken by health economics, there was no claim to scientific autonomy. Rather, it was a matter of creating a discipline adapted to the demands of expertise.
32The second of these two options proved decisive for health policy analysts in constructing a space of scientific expertise on health insurance and making their positions sustainable in the long term. I will therefore focus on this aspect of the conditions for its emergence. Still, the failure of the first option seems to me particularly interesting for understanding the specific advantages of health services research and its success with health policy analysts, which I will return to below.
The foundation of the Association for Health Services Research
33The term “health services research” was coined in the 1960s by a group of epidemiologists who represented a new approach: clinical epidemiology (Berkowitz 1998). They aimed to increase the level of health in the population not by improving medical science, but by “rationalizing” the care system.  The idea was promoted intensely in Washington as a necessary accompaniment to investment in medical research, and in 1968 these researchers secured the creation of the National Center for Health Services Research (NCHSR) to fund research in pursuit of these goals (Berkowitz 1998). As with many “sciences of government” (Desrosières  2002; Ihl et al. 2003; Rueschemeyer and Skocpol 1996), it was this institutional resource that laid the foundations for the recognition of a specific body of knowledge by providing funding for a journal, Health Services Research, and for the first specialized research institutes, all hosted by departments of medicine. 
34But health services research was not a well-established discipline in the early 1980s. It had emerged in a “context of application,” and was mostly restricted to the administrative sphere and a few isolated centers in medical schools. Like many applied disciplines in their early days (Benamouzig 2005; Pavis 2008), it enjoyed no real academic recognition.  Even within the government, its position was fragile. The NCHSR was effectively marginalized within the Department of Health; it had difficulty justifying the studies it financed to the political world, and faced severe institutional crises on a regular basis (Fox 1976).
35Up until the 1970s, health economists who had leading expert positions in Washington were only very marginally associated with health services research. The two groups claimed to work on different subjects: the health insurance system and economic analyses of physician-patient-insurance relationships (the “health care system”) for the former, and the care system and the organization of medical practices (the “delivery system”) for the latter. It appears that the closer links resulting from Altman’s project were based on a mutual exchange of resources. Clinical epidemiologists could offer economists institutional resources that, even if not yet very stable, were substantial. While health services research had not been widely accepted within academia, it was more firmly established in departments of medicine than health economics was in economics departments. And there existed a dedicated federal agency that, while fragile, nonetheless offered a crucial institutional base. Conversely, economists could offer clinical epidemiologists the political resources they had yet to obtain. In the words of the first president of the AHSR, a clinical epidemiologist, the main goal was “to connect the research community with the policy people who were making decisions.”  Health economists were “indispensable” for this, and this is why they were made spokespeople for the association and were the keynote speakers at the first meetings. 
36But strategic interests alone cannot explain the alliance between economists and clinical epidemiologists, which was also made possible by their relatively close epistemological positions and the fundamental assumptions the two groups shared. The aim of health services research as conceived in the 1960s and 1970s was to improve the “efficiency” of the medical system and to “rationalize” it. It belonged squarely to a context in which the evaluation of public policy was being encouraged. The National Center for Health Services Research (NCHSR) was created in 1968 as a direct result of this context. Economic incentives were supposed to be the primary means for such rationalization. A sort of “analytic compatibility” existed between the two groups of researchers that, as Christopher M. Weible (2008) has argued, encourages such alliances.
37The result was an expansion of the definition of health services research, whose field of investigation was extended from the care system to include the entire healthcare and health insurance system.  But health services research was still presented as an applied, interdisciplinary field, unallied to any specific methodological or theoretical approach and defined instead by its practical goal: to improve the efficiency of the healthcare system (Institute of Medicine 1995).
The work of the Association for Health Services Research: Lobbying and the formation of an operational discipline
38Like Altman, many of the founders of the AHSR led institutions that were threatened by the reduction in public funding. The new association’s main goal was therefore to defend the interests of those who made a living from healthcare expertise by guaranteeing funding in the area would be maintained. The association began lobbying, devoting much of its resources to building links with representatives of Congress. Its aim in the 1980s was to reinforce the NCHSR, which was felt to be on its last legs. As one of the founding members explained, however, the primary difficulty the association faced was health services research’s lack of recognition as a discipline.
And we had trouble defining the field at that point. When you would go to a social gathering of any kind and someone would say, well, what do you do? And you’d say, well, I do health services research. People would look at you for a second and then they’d walk away. They didn’t have any idea what that was—or care. And I think we had a long time there where we had to convince not only other academics, but policymakers that this was, in fact, a field. 
40This problem of recognition also affected many of the individuals whom the association counted as health services researchers but may not have immediately identified themselves as such. There were only a few dozen members of the association when it was created. Its task was to construct a professional community that mostly did not yet exist.  To consolidate health services research, the association’s founders tried to provide it with some of the characteristics of an academic discipline. AHSR took over publication of the existing journal Health Services Research, aiming to increase its circulation. It also created a number of prizes and grants to encourage and reward researchers working on health services research, like the Distinguished Investigator Award (1985), the Alice S. Hersh New Investigator Award (1986), and Article of the Year (1988). It established an annual meeting in 1983 that was meant to announce the discipline’s existence to the outside world. Finally, the same year, the association began to publish a quarterly bulletin for its members, the HSR Report, containing job announcements, information on research grants, and news about the association and its institutional partners, such as the NCHSR and a number of foundations (particularly the Robert Wood Johnson Foundation), who provided funding for the sort of work the association supported.
41In this sense, the association’s founders took a traditional path to constructing a new discipline, one that is encountered in work on both applied and so-called fundamental disciplines (Benamouzig 2005; Grossetti 2000; Pavis 2008). It was a matter of circumscribing the limits of authorized and legitimate knowledge in the field (Abbott 1988), and thereby organizing the closure of access to scholarly positions in the intellectual sphere, and to expert positions advising governing elites (Gouldner 1979). But while the project was a traditional one, it was distinguished from what has typically been observed even for applied disciplines by the lack of any desire to establish distance from other social spheres. In the case of health management and health economics, it was a matter of weakening and, to some extent, obscuring the discipline’s applied dimension. The project of constructing health services research did not require health policy analysts to establish the conditions for autonomizing their discipline. Instead, from the moment they laid its foundations, they affirmed that the discipline’s goal was to assist and advise the government. During the AHSR’s first annual meeting, the association’s president declared that the group’s aim was to encourage cooperation between researchers and decision-makers in the public and private sectors. Other founding members made similar remarks.
We wanted to develop credible policy research that would have more direct influence on political decisions. Just because you publish in journals, it doesn’t mean that the president’s going to behave differently. The association’s main goal was to develop an interface for connecting those who came from policy research and those who came from the government. 
43The association’s founders had a strong desire to keep health services research in a position very close to what Robert and Vauchez (2008), writing about European studies, called a “hybrid space.” Three characteristics of such spaces can be identified: the substantial importance given to practitioners in driving the discipline and producing knowledge; the claim of going beyond disciplinary boundaries; and, for those who founded and are most invested in the discipline, careers marked by professional experience in the very institutions they study. The subject of the same interview pointed out that the distinctive feature shared by all the association’s founding members was their previous work in government, which very clearly set the association apart from any other academic professional group.
When the association was founded, and I was part of that, the policy researchers who had served in government dominated the group… All of us, and especially the director, had worked in Washington for a very long time. It wasn’t the normal group of sociologists or others who had built themselves big reputations in the academic world and were at the top of their disciplines. Instead, the association was rather unusual, and that continued afterwards; the board of directors always included people who had worked for an administration or for the federal government. 
45Furthermore, politicians and leaders in the medical sector were systematically invited to take part in and contribute to the organization’s events, whether it be the annual meeting or other seminars and conferences. They also authored papers for the discipline’s journals, particularly Health Affairs, which was also established in the early 1980s and which is particularly representative of this hybrid space.
46This commitment to health services research matched that of the organizations that employed most health services researchers: expert organizations like think tanks, foundations, consulting firms, and specialized university centers attached to “schools”—medical, public health, business, law, and public policy schools—rather than departments. These centers were part of the university system, but their management primarily demanded knowledge that was directly “useful” and “relevant” for public policy (Fox 1990, 485-6). The practical consequence of this was that the demands and criteria for evaluating the careers of their researchers were often closer to those used in think tanks than traditional academic departments. As shown in the following remarks by a health economist, this was particularly visible in terms of publication expectations:
There are still only a few health economists in academia. Most of them are in public policy schools, public health schools, business schools, or other schools like that.
What is the difference for a health economist between working in academia and working in one of these schools?
Expectations in terms of publications aren’t the same. In academia, if you want to become a well-known professor, you have to publish in journals like the American Economic Review, journals like that. In the other schools, you can do the same thing, but that’s not what’s expected. What is expected, instead, is that you publish in Health Affairs, The New England Journal of Medicine, journals like that. 
48The last two journals cited are expert publications whose explicit mission is to reach an audience of decision-makers.  In the case of European studies, discussed by Robert and Vauchez (2010), their construction as a “hybrid space” seems to be the result of scattered, diluted efforts over the long term. What is striking in the case of health services research is that their position was the result of a conscious and deliberate desire, or strategy, which can be situated at a precise moment in time. This position was successfully maintained in the future.
The success of health services research with health policy analysts
49The project to structure a discipline around the AHSR quickly gained support from those health policy analysts best connected to the political world. Most joined the association in its first years, and even took up roles within it. As one of them summarized: “If you served, you would certainly have been active in the association that became AcademyHealth.”  (The AHSR was renamed AcademyHealth in 1997.)  Numbers subsequently grew, and membership became characteristic of the experts American decision-makers consulted most often on health insurance issues: when I carried out my study in 2010-11, all participants were members of AcademyHealth.Most were members of its board of directors, and a number had served as its president. Their involvement can also be seen in the prizes they received, with a substantial number having received the Distinguished Investigator Award given each year to a health services researcher.
50Such heavy involvement contrasts with their limited presence in the professional associations of other, more generalist, “traditional” academic disciplines. Among the seventy-three most consulted health policy experts in the period 1970-2010, only six were members of the American Economic Association, three of the American Political Science Association, and only one of the American Sociological Association. There was scarcely any more involvement in the International Health Economics Association: only ten health policy experts were members, even though, as we have seen, the association was founded by a well-known health policy analyst.
51How can the success of health services research (HSR) with health policy analysts be explained? I believe there are two factors. First, the growth of HSR as a professional training course in medical and public health schools meant that it became a discipline that could offer stable university positions, even as health economics had not truly been able to find a place in economics departments. Second, HSR’s epistemological choices were adapted to the demands of expertise and so to the professional activity of health policy analysts.
A discipline providing material and financial resources
52The success of HSR with health policy analysts was due first of all to the discipline’s growth, which led to an enormous rise in the number of courses offered in public health and medical schools during the 1980s, 1990s, and 2000s. The number of specialist courses in HSR (which generally began at masters level) went from fewer than a dozen in the 1970s to several hundred in the 2000s.  HSR offered relatively stable university employment opportunities to experts for whom access to university economics departments continues to be very difficult, as shown by the extract above from an interview with the health economist Thomas Rice. 
53The spectacular growth of HSR courses is due primarily to their applied dimension, which became a source of attraction for students beginning in the 1980s, as observed for other disciplines like management (Pavis 2008). One of my AHSR interviewees explained that most new members were students.  HSR courses offer direct training in organization and management roles in the medical sector (in hospitals and insurance). This resonates with developments in the American medical sector since the 1980s. The rise of for-profit insurance and hospitals, mergers and consolidation in the medical industry, and the growth of “managed care” have led to increased demand for new management positions in hospitals, insurance firms, and even public agencies (Starr 1982; Pierru 2007). Health services research had an advantage when it came to these jobs that it exploited very successfully. These jobs are all the more attractive because healthcare is one of the most dynamic sectors of the American economy.  HSR’s highly applied nature resonates with a trend characteristic of academia since the 1980s, which Gibbons et al. (1994) have described—probably wrongly (Pestre 1997)—as the advent of a second mode of scientific production in which practical effectiveness is the key concern.
54Another reason HSR has grown since the 1980s is increasing financial support from foundations and government agencies. This has been obtained mainly through the strategy of the AHSR which, beginning in the late 1980s, reoriented its political lobbying to concentrate on cost-effectiveness research (cost-benefit studies to identify the most effective medical protocols) and variation research (reducing cost variations between institutions by rationalizing how medical facilities are organized). The association proposed a solution to the problem of rising healthcare costs that did not adversely affect either the structure of the insurance system or the incomes of healthcare professionals. Thanks to this argument, which was particularly attractive to political decision-makers wary of the reaction from interest groups in the sector, the association successfully doubled the funding of the NCHSR, which had been renamed the Agency for Health Care Policy and Research (AHCPR) (Gray et al. 2003). In 2009, the association used the same argument to secure $1.1 billion in the American Recovery and Reinvestment Act for “comparative effectiveness research” (another name for cost-effectiveness research). These reforms were accompanied by new institutional and political resources. The AHCPR had been part of the Department of Health, but was moved closer to the cabinet, playing a far more substantial role in the Clinton and Obama reforms than the NCHSR ever had. The final version of the Affordable Care Act of 2010 created three new government centers of expertise, headed by representatives of health services research. 
55Private foundations have also been very receptive to HSR’s position on solving cost problems and, after several decades of withdrawal from health insurance issues, have returned to the subject by lending strong support to HSR (Fox 2010). Once again, health services research provides a good match for the “mode 2” of knowledge production described by Gibbons et al. (1994), which values sciences that can “sell themselves” and attract funding from outside the university. 
56Health services research has grown remarkably as a discipline since the association was created in 1981. The annual meeting had three-hundred participants in 1983, 2,400 in 2005, and more than 4,000 in 2012. This dynamism has been accompanied by growth on the part of the journals in which health services researchers publish. The most spectacular example is Health Affairs: having begun as a quarterly journal with a dozen articles in each issue (a total of 560 pages in 1983), it is now published monthly with around twenty articles in each issue (a total of 2,208 pages in 2015).
A discipline tailored to health policy analysts
57The success of HSR with health policy analysts is also a result of the adaptation of its epistemological position to these experts’ professional activity in think tanks, foundations, some consulting firms, and, as we have seen, university centers that emphasize expertise production rather than that of properly academic knowledge. By emphasizing the production of a body of knowledge useful for political decision-makers, HSR took on both the constraints of expertise and its political dimension. For instance, no one challenged the place of feasibility as the primary criterion in producing recommendations. This relates to one of the main constraints for health policy analysts, who are required by the expert organizations that employ them to justify the “utility” and “impact” of the knowledge they produce. The issue of “feasibility” consequently became prominent in expert discourse. The ideas that circulated in expert forums were developed, discussed, and assessed in the light of this criterion. Any proposition that did not meet it was consigned to the margins of the space of expertise, outside the central forums offering access to the political world. During the period in question, I consistently observed strategies of compromise and adaptation on the part of experts who wanted to remain “in the game” in order to conform to what they perceived as politically in demand and “feasible.” In this sense, the professional ethos of health policy analysts is more one of feasibility than truth, setting it apart from the academic ethos.
58Another political constraint health policy analysts had to contend with was the extreme politicization of American health insurance. In a context of polarization and intense partisan battles, it was impossible to avoid receiving a political label. While very few were members of a party, they were all associated with one camp or another and, to an extent, played the partisan game. This state of affairs might have been embarrassing for someone claiming scientific autonomy, but was welcomed within health services research.
59Furthermore, insofar as the discipline promotes expertise, HSR places far greater value on the work of health policy analysts who have no interest in making it an autonomous discipline. As we have seen, the health policy analysts best connected to the political world were members of the AHSR’s board of directors, sometimes served as the association’s president, were often plenary speakers at the annual meeting, and received the discipline’s most prestigious prizes. Their expert work would never have enabled them to achieve such central positions in a discipline that desired autonomy. HSR allowed them to convert something that is typically little valued or discouraged in academia into a source of prestige. And they could place far greater emphasis on their publications—which, as we have seen, they produced in huge quantities. These appeared almost exclusively with presses and journals with expert audiences but little recognition in academia. 
60In return, health policy analysts allowed the discipline to profit from their close links with the political world. Lobbyists at AcademyHealth told us that they had been able to rely on health policy analysts who had strategic roles in Congress or the federal government. This was the case during Clinton’s reforms in 1993: “During the Clinton reforms, lots of our members were on the task force, either in the White House or in the federal government. That helped a lot!”  The same was true for Obama’s reforms: “I know a lot of people in the administration and the federal government who were close to AcademyHealth and lobbied for it [the creation of a dedicated institute for comparative effective research]: a former member of MedPAC, two White House staffers, and Nancy-Ann DeParle.”  The president of AcademyHealth also mentioned the names of three other members of the association who occupied strategic positions in the Obama administration during the reforms.
Does health services research have enough scientific authority?
61But this closely dependent relationship with the political world raised questions about the authority and legitimacy health services research had obtained as a specific, scientifically valuable discipline. Their status was challenged by those with a more orthodox conception of science. This was true for Newhouse who, as we have seen, had tried to institutionalize health economy on a far more academic model: as he said in an interview, “health services research is a little different in that it’s not really a discipline.”  He argued that the multidisciplinary character of this research, with no precise methodology and a purely operational purpose, does not meet the requirements of a scientific discipline.
62Following this line of thought, we should note the difficulty health services research has faced in policing the scholarly boundaries of the discipline—precisely because of its low level of autonomy, or closure, relative to other social spheres. This has been particularly remarkable since the early 2000s, when Republicans began to include health insurance experts with libertarian views in their teams, who generally did not have PhDs (unlike most health policy analysts) and argued that an unregulated free market was possible and desirable in the health insurance sector. Because this position contradicts many of the fundamental assumptions of health economics (like information imbalance and risk selection), many of the experts I spoke to wanted to exclude libertarians from health services research. But many libertarians were still members of Academy-Health, and still invited as “colleagues” to present their views in the main expert forums. This situation shows that the way the discipline has been positioned, in the service of expertise, makes it impossible to exclude those who now act as the primary experts for one of the nation’s two main parties. From this point of view, it has achieved a degree of “generosity” toward the political space that has not been reached even by European studies, discussed by Robert and Vauchez (2010), or Middle Eastern studies, discussed by Eyal (2002). Without taking on all of its characteristics, the situation resembles the extreme one described by Stampnitzky (2013) in terrorism studies, where definitions of fundamental concepts vary depending on political circumstances, and where there is no stable criterion for membership in or exclusion from the field.
63But if we look at membership numbers and the institutional recognition it has received, it does not seem that these phenomena have hindered the ability of health services research to gain recognition as a discipline by many actors. As Newhouse himself continued: “but still there are a number of people around the country who view that as their professional home.”  What gives a discipline its status is not just the nature of the knowledge and the analytical categories it produces. As Fabienne Pavis (2008) writes, it is also “an academic organization with links to degrees, jobs, and academic careers,” as well as “cultural production with a value in the marketplace for goods and services.” While health services research has chosen an epistemological position that does not correspond to academic standards of autonomization from other social spheres, it appears that its growth in terms of university courses, official recognition, and publications has confirmed its status as a discipline. Health policy analysts can claim that health services research is a discipline, and it can serve as a reference body of knowledge for the expert world. Such research therefore guarantees researchers sufficient scholarly authority with journalists and decision-makers who are seeking technical information, ideas for programs and reforms, and arguments to support their point of view.
64* * *
65In the 1970s, specialists in health insurance affiliated to a number of non-governmental expert organizations made the decision—constrained by the academic labor market—to embrace careers as health policy analysts. They thereby undertook a new sort of professional activity, in which expert activity became a job in its own right, rather than work carried out as part of another job.
66But while they enjoyed regular access to the political world, their status in the early 1980s was still relatively unstable. This was a result of the financial insecurity of the organizations employing them, which were entirely dependent on political decisions on whether to continue funding health insurance research. It was a result, too, of the low level of academic recognition enjoyed by health economics, the body of knowledge they principally laid claim to. I have shown in this article how health policy analysts resolved this problem by investing in a particular discipline, health services research. They constructed this discipline so that it was adapted to their own professional activity, carried out in organizations (think tanks, foundations, consulting firms, higher education schools) whose aim was to produce expertise rather than academic knowledge. It was not just that health services research was defined above all as a practical knowledge aimed at efficiency. Contrary to what has been observed in most other applied disciplines that initially emerged “in an applied context” (Benamouzig 2005; Desrosières  2002; Grossetti 2000; Ihl et al. 2003), when health services research was being formed, the actors most heavily involved did not try to reinforce their autonomy from other social spheres, particularly politics and economics. Instead, they tried to maintain such research as a “hybrid space,” giving practitioners a great deal of room (Robert and Vauchez 2010) and placing value on the feasibility of recommendations as a criterion for evaluating any work produced. On the other hand, they paid little attention to health economics, the discipline they were initially trained in but that, as in France (Benamouzig 2005), had developed according to the model of scientific autonomy. They created a discipline in their own image that allowed them to consolidate a space of professionalized expertise constructed around the non-governmental expert organizations that employed them.
67This hybrid position was not a barrier to the development of university courses that claimed to be part of health services research. Far from it. Health services research has benefited from a marked tendency since the 1980s toward courses that emphasize the “concrete” aspect of the knowledge taught and the opportunities provided (Gibbons et al. 1994; Pavis 2008). The steady growth of the healthcare sector in the American economy only strengthens the appeal of courses that offer training for management roles in hospitals, insurance firms, public administration, and the medical industry. This hybrid position has also encouraged financial support from foundations and government agencies who want to produce directly “useful” knowledge, particularly since leaders in health services research have deliberately chosen to emphasize consensus-based work, accommodating interest groups and political parties. Furthermore, the jobs that health policy analysts involved with the association occupy in government and foundations puts them in a good position to defend the discipline’s interests, primarily through attracting funding for its members.
68Health services research has been characterized over the last four decades by strong growth and a process of specialization accompanied by the construction of a specific academic space in higher education schools (primarily medical and public health schools and, to a lesser degree, public affairs and business schools). As in the case of European studies, which Robert and Vauchez have discussed, there has emerged “a set of institutions and journals where professional models of excellence, criteria of judgment, and specific representations and practices prevail” (2010, 20). This expansion and specialization have been very profitable for health policy analysts, who have more easily found stable posts and increased the security of their professional careers. It has also consolidated health services research as a discipline, particularly for a non-academic audience. As with European studies, however, “far from leading automatically to the autonomization of this academic space, specialization has instead gone hand in hand with the maintenance of substantial heteronomy” (Robert and Vauchez 2010).
69Contrary to the literature on expertise, which typically shows the importance for experts of building links to a body of scientific knowledge that is recognized as autonomous, the position of health services research has not stopped health policy analysts exerting political influence or gaining recognition as legitimate experts on issues relating to the healthcare system and health insurance. This situation resembles that of the fields (or spaces) some authors describe as “liminal” (Eyal 2002) or “interstitial” (Vauchez 2011; Medvetz 2012; Stampnitzky 2013)—that is, located at the intersection of multiple social spaces. For these authors, this position is far from being an obstacle to the institutionalization of a group of experts. Rather, it is an asset, giving them a chance to accumulate resources from these different spaces. An emphasis on useful, “feasible” knowledge, adapted to circumstance and “political demand,” is precisely what gives health services an advantage in its interactions with the political world over academic knowledge, which claims scientific autonomy. Dependence on the political world is a firm advantage in relationships with decision-makers, as clients are more likely to receive a timely response that is well adapted to their needs. In return, these close relations make it easier to defend the interests of the discipline, which partly relies on government agencies’ financial support. The same is true for relations with foundations, which see social impact as a priority. The experts themselves find they can considerably lessen the unpleasantness caused by situations where they find themselves in tension with, or transgressing, the norms of their initial professional network—something numerous studies have observed (Memmi 1989; Barthe and Gilbert 2005). By situating themselves from the outset in a space of expertise that is autonomous from the academic space, and that requires an applied “knowledge-action” to meet political and social demands, experts accept the political aspect of their work and openly play the political game. Because their knowledge is oriented toward expertise, their peers do not view their involvement in the political world and the compromises it demands with suspicion. Rather, these become a source of value.
70It is logical, then, that those who benefit most from this situation—the health policy analysts most connected to the political and/or economic world—are particularly involved in the leadership of health services research (through its professional association), and that these individuals receive the honors distributed by the field (through awards and nominations). By affiliating themselves with the discipline and presenting themselves as eminent figures, the most consulted health policy analysts have found their own sphere of scholarly legitimacy. Health services research gives them scientific credibility without forcing them to accept the heavy constraints of the academic field. Such research gives them a way of consolidating a space of professionalized expertise on health insurance free from the constraints of academic norms, but that still retains some of its characteristics of legitimacy. Health policy analysts have resolved the tension Medvetz (2010, 557) described as internal to the profession of policy expert, caught between “intellectual credibility” and “temporal power.”
71Compared to the research cited above on interstitial spaces, health services research seems to me a case of closure (understood as the ability to protect oneself from incursions by other social spaces) that sits between, on the one hand, Middle Eastern studies in Israel (Eyal 2002) and European studies (Robert and Vauchez 2010; Vauchez 2011), which are characterized by a slightly greater degree of autonomy, and terrorism studies on the other (Stampnitsky 2013), which is marked by complete generosity toward the outside world. Furthermore, the number of actors and institutions involved means the domain of study in question is far broader than that examined here, suggesting similar conclusions could be drawn at different scales. The case’s crucial distinguishing feature, I believe, is that it explains a deliberate approach or strategy on the part of the actors involved, and connects this to changes in public expertise, characterized by the rise of the profession of policy analyst and the institutions that employ them, which form the basis of what I have called the space of professionalized expertise.
72These sites of professionalized expertise—think tanks, foundations, some higher education institutions—have developed considerably in recent years in the United States and Europe. My conclusions invite broader discussion of the consequences for bodies of knowledge in public space brought about by contemporary shifts in their production and diffusion. Such space competes with academic space in its ambition to produce knowledge about society, and it is important to bear in mind that the norms and demands within each are very different. In particular, the priority given to “feasibility” limits the critical dimension of the knowledge produced, as well as the degree of risk taken in the propositions formulated, which must remain within what is felt to be the realm of the possible.
Translator’s note: Unless otherwise stated, all translations of cited foreign language material in this article are our own.
As I have mentioned already, this recognition of the scientific nature of a given sort of knowledge is generally associated with the capacity of those with a stake in producing this knowledge to establish their “autonomy” in setting up its guiding norms, and to have this recognized (Gouldner 1979; Trépos 1996). Autonomy is understood here as the absence of interference from other social spaces, like the political or economic space, in setting the criteria by which knowledge is produced and evaluated (Merton  1973; Bourdieu  1988). While this definition of autonomy often overlooks the reality of scientific work (Pestre 1997), the crucial thing for experts is that they can claim it and that it is socially recognized.
Medvetz discusses policy analysts employed in think tanks across all domains of specialization. He therefore writes about the “space of think tanks.” Given the difficulty of defining the category of “think tanks” (which Medvetz himself recognizes), I prefer to expand my analysis to all individuals who perform a similar function, speaking about the space of expertise (or the space of policy analysis), and including structures that play a role comparable to think tanks. For more on this point see Lepont (2016b).
To my knowledge, this is the only published work on the history of health services research, with the exception of a few documents produced by health services researchers themselves (see especially Ginzberg 1991) and a few older studies (Flook 1973; Fox 1976).
For an explanation of the social and political conditions behind this change, see Lepont (2016a).
Up to that point, such studies had been reserved for defense policies (Porter 1995).
Interview with William Gorham, conducted by Andrew Rich, 1996. Quoted in Rich (2004, 48). See also Drucker (1968).
The number of think tanks began to grow after the Second World War, which Nelson Polsby (1983) sees as the expansion of social programs begun under the New Deal. This growth increased considerably in the 1970s. Several authors estimate that the number of think tanks quadrupled between 1970 and the end of the 1990s (Rich 2004; McGann and Weaver  2006).
In fact, the policy analysis discussed in this article is far from the ideal of policy science formulated by several generations of American political scientists, from Charles Merriam to Harold Lasswell, Charles Lindblom, and Aaron Wildavsky. These political scientists looked poorly on the predominance of economics in the policy analysis practiced in Washington.
More precisely, it was established by the Department of Health, Education and Welfare, which was renamed the Department of Health and Human Services under Carter. I have referred throughout to the “Department of Health” to avoid confusion.
Interview with Uwe Reinhardt (PhD in economics, 1971), professor at Stanford and author of a regular column on health policy in the New York Times, conducted by the author, November 2011. He is considered an iconoclast by his peers for his sometimes relatively critical views on the market.
Interview with Karen Davis (PhD in economics, 1969), president of the Commonwealth Fund, conducted by the author, May 2010. Davis was a member of the Carter administration. Reports by her organization, the Commonwealth Fund (CWF), were among the most influential in Congress in 2009-10.
Interview with Uwe Reinhardt (see above).
Interview with Robert Blendon (PhD in health policy, 1969), professor at the Harvard School of Public Health, conducted by the author, October 2011. Blendon held multiple leadership posts at the Robert Wood Johnson Foundation and in Harvard’s Department of Health Policy and Management. In the political world, he is a specialist in public opinion and surveys on health issues.
Interview with Joseph Newhouse (PhD in economics, 1969), professor of Health Policy and Management at Harvard, conducted by Edward Berkowitz, 1998 (see Box 1).
Interview with Joseph Newhouse, conducted by the author, September 2011.
We see here the different poles of policy analysts’ work described by Medvetz (2010). Note that some aspects of their work have developed over time. Communication with the media has grown more important, and publication focuses increasingly on short policy briefs, with books now a rarity.
Interview with Robert Blendon (see above).
Interview with Uwe Reinhardt (see above).
Interview with Robert Blendon (see above).
These reforms regulated spending in the Medicare public insurance program. The two regulation tools promoted by health economists and adopted by Congress were the Diagnostic Related Group System (DRG), in 1983, and the Resource-Based Relative Value Scale System (RBRVS) in 1989. Health policy analysts benefited substantially from these: for the first time since Medicare was established, the rate of spending increases fell. This provided a source of power, as they were the only ones with the technical skills to use these tools correctly (Brown 1985). The adoption of the DRG and RBRVS was also part of the institutionalization of health economists as experts in health insurance policy, leading to the creation of two Congressional commissions, the Prospective Payment Assessment Commission (ProPAC) and the Physician Payment Review Commission (PPRC), which were run by health economists.
Interview with Gordon DeFriese, director of the Cecil G. Sheps Center for Health Services Research from 1973 to 2000, and a founding member of the Association for Health Services Research (AHSR), conducted by Jennifer Muldoon, 2008. AcademyHealth archives (see Box 1).
This position was based on a criticism of medicine that in many ways resembled the evidence-based medicine movement, which was just beginning to emerge (Marks 1997): the issue was to rationalize medical practices by systematically applying protocols evaluated experimentally and/or using cost-benefit analysis. Clinical epidemiologists were distinctive for their interest in organizational questions rather than treatment (Lepont 2014).
There were five such institutes at the beginning of the 1970s: the Health Services Research and Development Center (Johns Hopkins University School of Hygiene and Public Health); the Health Services Research Center (University of North Carolina, Chapel Hill); the Department of Health Care Systems (Wharton School, University of Pennsylvania); the Center for Health Services and Policy Research (Northwestern University); and the Department of Medical Care Organization (School of Public Health, University of Michigan).
See the interviews conducted by Berkowitz, which are available at http://www.nlm.nih.gov/hmd/nichsr/intro.html (see Box 1).
Interview with Cliff Gaus, founding member of the AHSR, conducted by Jennifer Muldoon, 2008. AcademyHealth archives (see Box 1).
See for instance the definition given by Evelyn Flook (1973) and the Institute of Medicine (1995).
Interview with Gordon DeFriese, conducted by Jennifer Muldoon, 2008 (see above).
Interview with Jennifer Muldoon, an employee at AcademyHealth since 1992, conducted by the author, December 2011.
Interview with Robert Blendon (see above).
Interview with Thomas Rice (PhD in economics, 1982), professor at the UCLA Fielding School of Public Health, conducted by the author, October 2012.
For the NEJM, this was only true for the “Perspective” section.
Interview with Robert Blendon (see above).
This new name, which more closely resembles that of a consulting or lobbying firm than that of a professional academic association, reveals how the group has maintained, and perhaps even reinforced, its hybrid position.
One indication is the list provided by the Association of Schools and Programs of Public Health (ASPPH). This does not include all organizations offering courses connected to health services research, but it includes some of them, representing the fifty schools of public health accredited by a nationally recognized professional committee. Among the training these schools offer are several hundred courses linked directly to health services research, even if they do not use the name. For the association’s website, see https://www.aspph.org/.
Among the health economists I met with in my research, only two belong to economics departments: David Cutler at Harvard and Jonathan Gruber at MIT (who gained tenure long before specializing in health). As already noted, only six are members of the American Economic Association. On the overwhelmingly applied character of health economics, which has been an obstacle to its full integration into economics departments, see Evelyn L. Forget (2004).
Interview with Jennifer Muldoon (see above).
The health sector grew from accounting for 6% of national GDP in 1965 to 17% in 2012 (OECD 2014).
The Patient-Centered Outcomes Research Institute, the Innovation Center, and the Independent Payment Advisory Board.
As noted slightly earlier, I agree with criticisms of the analysis proposed by Gibbons et al., which presents the history of science as a linear historical evolution from “mode 1” to “mode 2” of knowledge production (Pestre 1997). Nonetheless, the opposition between these two modes is interesting to me, insofar as it sheds light on the current valorization—linked to social changes but also, as Dominique Pestre points out, to political ones—of a certain vision of science, for which “mode 2” in a sense provides the ideal model.
Characteristically, the four individuals who published most of their books with university presses are the only four with links to an economics or political science department.
Interview with Jennifer Muldoon (see above).
Interview with Bruce Stuart (PhD in economics, 1970), professor at the University of Maryland’s School of Pharmacy, conducted by the author, October 2011. A health specialist, Nancy-Ann DeParle has held numerous leadership roles in the Clinton and Obama administrations—notably, she was director of the White House’s Office of Health Reform from Obama’s election in 2011, a key position in the reform process—and is also a member of AcademyHealth.
Interview with Joseph Newhouse, conducted by Edward Berkowitz, 1998 (see above).