Bricolage or Entrepreneurship? Lessons from the Creation of the European Centre for Disease Prevention and Control

Abstract

This paper belongs to the field of theories of the Policy process and contributes to the literature on the multiple streams approach (MSA), by investigating bricolage as an alternative type of agency in agenda‐setting and policy formulation. The bricoleur frames conditions as a problem that can and must be fixed and emerges as the one who looks for a solution. For the bricoleur outcome goals, or rather, the choice of a particular outcome is less important than the process goal. Therefore, the bricoleur selects policy ideas depending on the properties they display and combines them to create the best “fit” to couple the streams. Bricolage relates to the organizational properties of certain actors, in this case the European Commission. Evidence comes from the case of the European Centre for Disease Prevention and Control (ECDC) and its origins, created in the aftermath of the severe acute respiratory syndrome crisis. I will demonstrate that the crisis did not trigger the creation of the ECDC but rather was framed as a salient issue to be resolved by the Commission. The process of policy formulation that followed resulted in a bespoke solution. As such, this paper contributes to the development of the MSA but also to the literature on ideational politics and European agencies.

Introduction

Cross‐border health crises have attracted a lot of attention among the general public in the last few years, most recently with the Ebola virus. The European Union (EU) has a long history of health crises (such as the Bovine spongiform encephalopathy or “mad cow” disease). Most pertinently perhaps, the creation of the European Centre for Disease Prevention and Control (ECDC) is often portrayed as the functional response to the severe acute respiratory syndrome (SARS) crisis that occurred between November 2002 and July 2003 (Greer 2012; Greer and Löblová 2016; McKee, Atun, and Coker 2008). This perspective on the importance of the crisis in the creation of the agency is often used to explain why such a short period was spent between the European Commission drafting a proposal to create the ECDC in July 2003 and the creation in situ of the agency in May 2005. The ECDC hence looks like a mundane case of agency creation: a crisis triggered a response; this response was a new agency. This idea that the time “has come” for an agency to be created is a recurring theme in the official narratives of agency creation as well as in the literature on the emergence of agencies that regulate or distribute information used by regulators (Alam 2007; Maggetti 2013). The ECDC, in particular, is charged with bringing expertise to the table, thus reducing uncertainty and allowing an evidence‐based regulation of risk.

The crisis‐followed‐by‐agency‐creation causal mechanism seems plausible. However, a minimum standard for causality is that the effect comes after the cause. But in the field of disease prevention and control in Europe, we can observe that some features of the ECDC were pre‐existent to SARS; notably a network of epidemiological surveillance, now an integral part of the ECDC, existed since the early 1990s. Moreover, the creation of a European agency dedicated to disease control has been disputed among experts and has been a topic of disagreement between European institutions since the late 1990s. At a minimum, the creation of an agency as response to the crisis has to be put in its context of conflict and contestation of the various institutional choices aired at the time. In this paper, I challenge the conventional wisdom of agency creation as response to the crisis and open the black box of the organizational and political processes of creation dynamics. As shown by Moe (2005), institutional emergence is a political process where power is created, distributed, and re‐arranged in the form of a precise set of organizational features.

To answer questions about the causes of institutional creation, it is necessary to go beyond the crisis as a single explanatory variable and open our grand‐angle on the politics of agenda‐setting and policy formulation. The multiple streams approach (MSA) seems a suitable, sophisticated theoretical lens to analyze how the agenda for the creation of the ECDC was set and how the policy idea of an agency emerged. In European Studies, the MSA has been fruitfully applied to the study of agenda‐setting in the EU (Ackrill, Kay, and Zahariadis 2013; Herweg 2015; Zahariadis 2008)—adapting a model explaining agenda‐setting in the United States (Kingdon [1984] 2003). But subsequent developments of the field have highlighted that the MSA can be applied to the entire policy formulation process (Blankenau 2001).

The model breaks down causality across three paths or streams: policy, politics, and problems. This is already more sophisticated than the crisis‐response model. In addition, the model presents three necessary conditions for change (Herweg, Huß, and Zohlnhöfer 2015):

• (1)Each stream must be “ripe”, meaning that there is a perceived problem, that the policy stream comprises at least one viable alternative and that policymakers embrace a proposal.
• (2)A policy window opens following a change in the politics or the problem stream.
• (3)A policy entrepreneur couples politics, policy, and problem streams.

Using the MSA brings agency back into the study of institutional emergence: policy entrepreneurs have policy solutions which they try to link to changes in the preferences of policymakers or to the emergence of new problems. However, in the case of the ECDC, there is no figure that emerges as a policy entrepreneur. Some ideas were championed by different actors; however, rather than one idea emerging, bits and pieces of different ideas seem to have been recombined in a proposal of the European Commission. How can organizational change be explained without a policy entrepreneur? Can the MSA be refined to present coupling the streams as something other than the result of policy‐entrepreneurs? How can we explain that ideas, sometimes presented as contradictory, may be identified in the institutional features of the ECDC?

This paper makes two conceptual moves. First, this paper presents a critical approach to the MSA and looks for an alternative to the policy entrepreneur. I will define a new type of agency based on the search for solving a problem rather than the advancement of a solution. Second, this paper presents a different relationship between policy ideas and the agent; as such it contributes to the development of the MSA and to the literature on the creation of European agencies.

The recent literature on the MSA (Ackrill, Kay, and Zahariadis 2013; Cairney and Jones 2015) shows that there is an appetite for refinements of the model: “there is a need for a study to specify MSA's theoretical benchmarks and hypotheses to clearly identify the potential and limits of conceptual stretching” (Jones et al. 2015, 29). This paper contributes to refining our understanding of coupling the streams and suggests a critical approach to agency in the MSA.

Agency in the MSA is usually incarnated in the figure of the policy entrepreneur. Walker (1974, 113) defines policy entrepreneurship as the advancement of solutions that are tied closely with the maintenance of entrepreneurial needs and interests. Zahariadis (2008, 18) presents the policy entrepreneur as the proactive element of the MSA capable of developing strategies to promote her solutions. In the MSA, policy entrepreneurs are successful when their efforts for entrepreneurship lead to coupling the streams in accordance to their preferences. Policy entrepreneurship is thus defined as efforts made by the policy entrepreneur to promote solutions consistent with her needs and interest. From these definitions, I infer that the actions of the policy entrepreneur are a solution‐driven process from softening‐up to coupling the streams. Nevertheless, Zahariadis (2003, 73) identifies two types of coupling: “when windows open in the politics stream, coupling is likely to be doctrinal (finding a problem to a given solution),” and “when policy windows open the problem stream, coupling is likely to be consequential (finding a solution to a given problem).” While doctrinal coupling suits the traditional assumptions on the activities of the policy entrepreneur (a solution‐driven process or an outcome process), consequential coupling underlines that coupling can be seen as a problem‐driven process. Therefore, let us turn to the literature on the problem stream that is eloquent on the notion of focusing events and how they lead to policy windows. This notion, first introduced by Kingdon ([1984] 2003), was then adapted by Birkland (1998a; 1998b) to study the effect of accidents and natural disasters as drivers of change. In the literature, new research venues appeared recently, inspired by the financial and budgetary crisis, such as Saurugger and Terpan (2015), refining the MSA to demonstrate that the stronger the crisis, the more important the change. While the literature informs us on drivers and scope of change, there is room for studying a type of agency in the MSA that is defined as a problem‐driven process rather than a solution‐driven one.

I thus suggest the introduction of a new type of agency in the MSA. My argument is that coupling is not necessarily the result of the efforts of a policy entrepreneur, but rather arises out of a different type of agency: bricolage, in which the one who couples the streams combines different policy ideas to formulate a bespoke solution to a problem.

The bricoleur as a type of agent distinct from the policy entrepreneur does not have clear or fixed preferences for one solution over another. Her choice is thus dictated by a consequential search for fit (Zahariadis 2003, 73). Put bluntly, the bricoleur looks for policy ideas that would be useful in crafting a new solution: policy ideas as a resource to create solutions. Here enters the second theme: how does the bricoleur select these policy ideas? Why are bricoleurs drawn to some ideas rather than others? The emergence of ideas in the MSA is likened to a process of natural selection (Kingdon [1984] 2003) in which the “survival” of ideas is arbitrated by two criteria: value acceptability (compatibility with the values of the policy network) and technical feasibility (to what extent can the idea be translated into the real world?). These criteria are relevant to understand what ideas are at the disposal of the bricoleur and why they survived the natural selection process, but do not offer explanatory leverage on why the bricoleur selects some ideas rather than other ones. On what does the bricoleur base her judgement? I suggest two criteria that the bricoleur uses when she arbitrates different policy ideas.

• The first one is an important element of the MSA: the ripeness of the politics stream (Herweg, Huß, and Zohlnhöfer 2015). The bricoleur selects policy ideas to which policymakers are ripe to; otherwise, the policy solution would not pass the decision‐making stage.

• The second one is more innovative and suggests that the bricoleur builds on the initiatives developed by the policy community. This is the criterion of increasing returns: the bricoleur avoids the costs of exiting initiatives and even capitalizes on projects already developed by the policy community. Therefore, the bricoleur is likely to use ideas in which the policy community has invested time and efforts.

Following these criteria, the bricoleur creates a bespoke policy solution from different ideas presenting advantages in terms of ripeness of the politics stream and/or in terms of increasing returns.

This paper proceeds as follows: after entering bricolage and how it can contribute to the enrichment of the MSA, I first draw on process tracing to present the context in each stream, up to the SARS crisis. The first stream is the policy stream. Different ideas navigate in the policy stream: how do they stand the criterion of increasing returns? Evidence comes from the debate that animated the policy community by analyzing articles in different publications of practitioners. This paper then investigates the variations in the politics stream and analyze how the respective positions of the Council of the European Union and the European Parliament (EP) have shown appetite for different solutions, how the European Commission has evolved from opponent to promotor of the creation of the ECDC and how this has affected the recombination of ideas. In order to do so, official positions of the institutions have been identified from the Official Journal of the European Union and the archives of the different institutions. The problem stream is eventually analyzed in terms of perception, also drawing from archives and speeches. The second methodological step is to demonstrate the lack of fit of the figure of the policy entrepreneur in a part that reflects on the window of opportunity and agency and on how the European Commission's framing of the SARS crisis led to coupling by bricolage. Eventually, building on the analysis of the three streams, this paper focuses on the policy formulation part of bricolage, as well as the decision‐making process and delivers a review of the concept in conclusion.

The Relationship Between Bricolage and Change

Entering Bricolage

Bricolage is a concept that presents a mode of scientific thoughts. In his book The Savage Mind, first published in 1962, Levi‐Strauss (1988, 12) introduced the bricoleur as one who “addresses himself to a collection of oddments left over from human endeavours”. While Levi‐Strauss looked at the concept from an anthropological point of view, the concept of bricolage resonated in different disciplines of social science. In public policy, the concept captures epistemological strategies and rationalities, under the name of Epistemological Bricolage (Freeman 2007). The concept is identified as a self‐learning process, the act of piecing together knowledge, as one “acquires and assembles tools and material as he or she goes.” (Freeman 2007, 486; see also Carstensen 2011). Freeman is interested in understanding how knowledge is formed and thus identified dynamics of learning. However, if the concept of policy change is questioned, identifying the type of “learning” only gives a partial answer. The concept of bricolage in public policy needs to be refined to unlock its potential in explaining policy change.

Levi‐Strauss used the French word of Bricolage as an analogy to underline an oblique strategy. In his classic bricolage piece of 1962, the word is mainly used to describe some extraneous, oblique movements in ball games, as well as snooker or racing. The use of the word bricolage and bricoleur is nowadays very different. The bricoleur is one in a domestic environment such as a workshop, using whatever is at hand to perform a form of craftsmanship.

Now, imagine a bricoleur trying to fix a table, running out of screws. She is deprived of a solution to the problem and needs a new solution. However, she is bounded to what is available in the workshop; she looks for alternatives and decides to use a nail instead of screw. Let us now imagine that the nail is too short for the hole of the screw. The bricoleur looks again around her and finds a chip of wood. She has an idea: by sticking the chip of food at the bottom of the hole, she can use the nail. The bricoleur has thus fixed the table but rather than using “by the book” solutions, she combined different solutions to make it fit to the situation. The solution is not a type of modus operandi: the bricoleur will not start using nails and wood chips instead of screws. Rather, this solution is a devious mean used to keep the process of fixing the table on going.

Let us now return to Levi‐Strauss to generate theoretical leverage on bricolage as a form of policy change. Applied to public policy, the bricoleur is one who cares about finding a solution rather than using a specific solution. Bricolage is thus a form of agency that is problem‐solving oriented and characterized by:

1. Contextual conjectures: the bricoleur creates structures by means of events (Levi‐Strauss 1988, 15). The bricoleur realizes there is a problem after an event occurs. The bricoleur's agency lays in the choice to ignore the problem or to solve it. Once the choice made to solve the problem, the bricoleur looks for the means to come up with a solution.

2. A process goal: the bricoleur uses devious means (Levi‐Strauss 1988, 11) to create a bespoke solution to a specific problem. Outcome goals, or rather, the choice of a particular outcome, is less important than the process goal.

3. Bounded rationality and scarce resources that cannot be expanded: the elements used by the bricoleur are “pre‐constrained” (Levi‐Strauss 1988, 12), there is a finite amount of policy ideas at hand.

4. As a result, the bricoleur produces a policy solution created from the “new arrangement of elements” (Levi‐Strauss 1988, 13).

These four characteristics of bricolage will be subsequently leveraged to understand the creation of the ECDC and more generally how the MSA can work when, instead of an entrepreneur, there is a bricoleur.

The Bricoleur inside the MSA

The bricoleur is a type of agent that is hybrid in the sense that she is active in both the problem and the policy stream. I develop below how the foundations of the MSA, contingency, and ambiguity are compatible with this type of agency.

Rüb (2016, 56) posits that two phenomena are contingent in the MSA: the political entrepreneur and the window of opportunity. This paper ought to add a third phenomenon, the bricoleur. Contingency is the assumption that everything that is could be different; I relate bricolage as background contingency (Rüb 2016) which predicates that agency is achieved in a context laced with coincidences and surprises. The bricoleur acts by means of events, but this explains only partially her agency; she makes the choice to focus on a problem and, therefore, is a contingent element following events.

The bricoleur borrows from the problem broker who promotes a particular problem, or “define(s) conditions as problems” (Knaggård 2015, 452). The bricoleur thus participates actively in opening a problem window, which in an ambiguous context “create(s) meaning for the policymakers” (Zahariadis 2008, 16). However, unlike the problem broker, the bricoleur's agency is not only a matter of creating meaning in the problem stream: she finds the imperative for policy change and consequentially engages on a search for solutions.

The bricoleur is not attached to one solution in particular, rather her attention is on the resolution of the problem and the means are secondary. This means that unlike the policy entrepreneur, the bricoleur does not look for a way of maintaining her needs and interests but rather defines her own role as one who takes responsibility for solving the problem. The bricoleur is the problem‐solving driven agent that emerges to take responsibility for finding a solution.

I rely on the concept of consequential search for fit and problem windows to define bricolage as the process of finding a solution to a problem. The bricoleur is established as one who frames the problem and actively participates in opening a problem window. Zahariadis describes the consequences of a problem window as follows: “a problem window triggers a search with a problem already in mind, however vaguely it may be defined. […] Consequently, the process begins with a search for clues about appropriate solutions to an already existing problem” (Zahariadis 2003, 73). The search for an “appropriate” solution that “fits” is a theme that has been understudied in the MSA literature; nevertheless, Ackrill and Kay (2011) have developed a model of coupling that answers some of the interrogations raised by the consequential search for fit. However, this paper will suggest a departure from the direction taken by Ackrill and Kay.

Ackrill and Kay posit that changes in the political or problem streams may signal to policymakers to select from known proposals (Ackrill and Kay 2011, 77) and suggest that both selling an idea and selecting an idea are entrepreneurial. The concept of policy entrepreneurship is thus stretched: policy entrepreneurs do not necessarily have a pet policy solution, they can pick‐it up as a change in the politics or problem stream appears. I argue that the concept of policy entrepreneurs proves too limited when agents do not shape preferences for a solution but rather look for the best fit. Here then enters bricolage. The bricoleur has a process goal and as such cannot be likened to a form of policy entrepreneur that is inherently driven by an outcome goal. Rather than selecting ideas depending on preferences, the bricoleur creates by recombining ideas and formulating a bespoke solution. It is thus a type of agency that uses unexploited aspects of the full theoretical leverage of the consequential search for fit.

Inside the MSA the bricoleur is an agent who has an incidence on both content and content perception and who is active in two of the streams. In the problem stream, the bricoleur opens a window by framing conditions as a problem that she can and must be fix; the bricoleur thus sets her goal to engage on a path of consequential search for fit in the policy stream. Only then does the bricoleur modify the content to produce a “fit” solution. This bricolage leads to an oblique way of coupling the streams: rather than a solution being joined to a new problem, a bespoke solution is created to accommodate the definition of the problem.

Nevertheless, the search for “fit” does not only depend of the way the problem was framed but also of the ripeness of the politics stream and of the elements that the bricoleur finds in the policy stream. Hereby I address how the bricoleur creates a bespoke solution.

Formulating Policy in an Oblique Way

Once the bricoleur has framed conditions as a problem, the search for fit depends on two variables: first, the ripeness of the politics stream and, second, increasing returns of policy ideas. These criteria determine how the bricoleur creates a bespoke solution. As for formulating policy, bricolage is an oblique way to create a new solution by selecting and recombining ideas depending on their properties, but also depending on how they are perceived by decision makers.

Ripeness of Policy‐Makers

Building upon Kingdon's idea of receptivity ([1984] 2003), Herweg, Huß, and Zohlnhöfer (2015) contributed to the enrichment of the MSA by bringing the concept of “ripeness” of the stream which describes how the stream becomes ripe to new policy ideas. The bricoleur then pays attention to select elements of ideas which the policymakers are ripe to. With bricolage, it is even possible to piece together a solution that will comfort dissonant preferences among policymakers. This is a type of contingency defined by Rüb (2016) as action contingency in the sense that it arises from the interaction of individuals of groups competing for power.

Increasing Returns of Policy Ideas

Zahariadis in his definition of consequential search for fit posits that an “immediate action [such as bricolage] is motivated more by the need to avoid higher costs rather than the need to reap more benefits” (Zahariadis 2003, 72). This echoes Pierson's (2000, 252) conceptualization of increasing returns: “the relative benefits of the current activity compared with other possible options increase over time” (Pierson 2000, 252). The bricoleur's choice is to avoid the costs of exiting initiatives already developed in the field and, therefore, policy ideas with increasing returns are always bound to the case under consideration. The cost‐efficiency analysis performed by the bricoleur is thus not only about the cost it represents for herself but also about the cost for the policy community.

Ideas represent a cost for the policy community when policies exist as initiatives. They thus do not bare any cost when ideas are “on the shelf”: they have been simply exposed, most of the time in written form and bare virtually no cost but the time the author spent on shaping and communicating the idea. Ideas represent a cost and, therefore, produce increasing returns in two cases: The applicability of the idea has been successfully accepted within the policy community and developed as a turn‐key solution. It involves efforts of research and policy design. These initiatives exist only on paper but the policy community has bared the cost of developing, designing, and preparing a policy solution directly exploitable by decision makers. The idea has been partially or fully translated into the real world and as such bares financial costs and/or mobilizes part of the policy community. It thus has a structuring value for the policy community, mobilizes resources and may be seen as a structure that inhibits organization members from seeing a need for change (Kelman 2005, 27);

The criterion of increasing returns tells us that the bricoleur avoids the costs of exiting initiatives and even capitalizes on projects that have been developed by the policy community. Therefore, the bricoleur is likely to use ideas in which the policy community has invested time and efforts. As such it is also a matter of action contingency (Rüb 2016) in the sense that it arises from the interaction of different individuals or groups within the policy community.

As developed in the Introduction section, the criteria for idea selection in bricolage are distinct from the criteria of survival of ideas. Nevertheless, I ought to contrast this distinction: while the two processes are a matter of emergence of policy ideas, the survival of ideas is an incremental process, whereas bricolage is an immediate selection. The two processes are not antithetic, rather the policy ideas that a bricoleur will take under consideration are at her disposal because these ideas survived in the primeval soup. The survival of ideas is a matter of contingency, while the two criteria of bricolage are a matter of agency and explains the judgement of the bricoleur when she considers different ideas.

Looking at each of the criteria, they add a new layer to the survival criteria.Ripeness of policymakers is about how much policymaker are receptive to the idea, while this receptiveness might depend on the value acceptability and technical feasibility of the idea, the focus remains on the relationship between an idea and the policymakers. Increasing returns is not a measure of how much an idea is accepted in the policy community or how much its technical feasibility has been proven, but a criterion based on the efforts made to make an idea acceptable, feasible or even to realize the idea.

Bricolage as a Mode of Coupling the Streams

Because of increasing returns, it can be assumed that the bricoleur tries to avoid the cost of exiting from initiatives developed by the policy community. I hypothesize that the bricoleur will incorporate elements of an idea with no increasing returns to the solution only because policymakers are ripe to it.

Coupling and formulating a policy are for the bricoleur simultaneous and intertwined processes. The bricoleur is thus an agent that couples the streams by formulating a bespoke solution. The bricoleur creates a policy solution that is wary of contextual elements in both the politics and the policy stream. She formulates a policy solution that “fits” and in doing so couples the streams.

The hypothesis underpinned by bricolage as a mode of coupling the streams is the following: after framing conditions as a problem, a bricoleur pieces together a solution in order to solve the problem. It is cost‐effective for the bricoleur to use policy ideas demonstrating increasing returns; nevertheless, the bricoleur is also wary of the ripeness of policymakers. As such the bricoleur has the possibility to combine different policy ideas, selected because they present the advantage of increasing returns or because they resonate with policymakers' preferences. Bricolage results in a bespoke solution that, in a context framed as urgent, is then swiftly adopted.

The Context: Streams of Disease Prevention and Control in Europe

The Policy Stream: Policy Ideas and Their Increasing Returns

The earliest traceable idea of organization of disease prevention and control in Europe is the “Charter Group.” It has been briefly mentioned by Greer (2012, 1009), on the political science side and in public health publications by Krause (2008), MacLehose, McKee, and Weinberg (2002) and by Newton, Grimaud, and Weinberg (1999), in the latter cases as members of this “Charter group.” The most precise academic source on the origins of the Charter Group is a 1998 lecture given in Washington by Chris Bartlett, the then‐Director of the British Communicable Disease Surveillance Centre (CDSC) who shared paternity of the Charter group with Gijs Elzinga from the Rijksinstituut voor Volksgezondheid en Milieuhygiene (RIVM), the Dutch National Surveillance Centre. They convened experts from each of the then 12 EU member states, as well as the heads of institutions, with responsibility for national surveillance, to a meeting at CDSC London in December 1993. This was the first meeting of what would become the “Charter Group”: a network of public health experts who would draw on national resources to achieve common surveillance in Europe (Bartlett 1998).

The raison d'être of the Charter Group was to actively flesh‐out coordination of epidemiological surveillance between existing national centers of disease control. In the mid‐1990s, the Charter Group had been developing jointly agreed standards for disease surveillance via the prioritization of infectious diseases (Newton, Grimaud, and Weinberg 1999); and, as soon as September 1995 they published a monthly peer‐reviewed scientific journal of epidemiologic surveillance “Eurosurveillance” and developed a high‐level program for Intervention Epidemic Training (EPIET), training public health doctors and epidemiologists to the same methods, standards and ethos (Bartlett 1998). At this point, the proposal was already producing increasing returns: it had been successfully accepted within the policy community and involved efforts of both research and policy design. The idea had then been translated into the real world and mobilized as part of the policy community. It thus had a structuring value for the policy community, mobilized resources and may, in the future, inhibit its members from seeing a need for change.

The Charter Group's network approach was politically endorsed in September 1998 with the creation of A Network for the Epidemiological Surveillance and Control of Communicable Diseases in the Community established by a decision of the EP and the Council of the European Union. ^B However, decisions are nonbinding instruments, here used in order to facilitate the work of the Charter group and provide limited funding rather than to create a new instrument. The decision lists epidemiological surveillance and prevention, two elements of the self‐defined mission of the Charter group, and was first and foremost a list of guidelines on desirable developments of the networks. This decision underlines that increasing returns started to be more important because of the financial costs and because the network mobilized efforts from the experts of the policy community as well as European decision makers.

The primeval soup became thicker as another proposal emerged: in September 1998, the International Board of Scientific Advisors (a group mainly comprised of micro‐biologists and researchers) met in Paris and manifested their support in favor of a European Centre for Infectious Disease (ECID) (Butler 1998). The idea was also supported by “several scientific organizations, including the European Society of Clinical Microbiology and Infectious Diseases” (Butler 1998). The proposal was championed by Michel Tibayrenc, Director of the Centres d'Etudes sur le Polymorphisme des Mico‐organismes in Montpellier, France and suggested the creation of “scientific board” based on the existing US Center for Disease Control (CDC). In this perspective, the ECID would be created bearing in mind that “health policy remaining under the sovereignty of each nation and the ECID providing complementary overall coordination” (Tibayrenc 1998). Rather than cooperation based on surveillance of disease, the ECID would be a more ambitious idea, as its inspiration the US Center for Disease Control (CDC), being consequential in terms of training, surveillance, and research, areas in which the Charter group had demonstrated no ambition. The proposal enriched the primeval soup but never met a strong support within the policy community. As such it remained an idea “on the shelf”: the only member of the policy community to put efforts and time in the idea was Michel Tibayrenc, but the idea never involved efforts from the community in designing a specific plan nor bared a cost. It is interesting to note that this policy idea is a cased for policy diffusion and as such cannot be considered to present increasing returns without the policy community making efforts to translate the US model to the European context.

The proposal for an EU agency triggered important debates among the members of the policy network and received thorough criticism from supporters and members of the Charter group. The journal for medical practitioners The Lancet featured an unsigned editorial titled “Not another European Institution” (Lancet 1998). The Editorial recalled the accomplishments of the Charter Group, in terms of trainings or surveillance. It was followed by a stream of back and forth open letters and articles dedicated to support one idea over the other one, in different public health journals (see Butler 1998; Dove 1998; Giesecke and Weinberg 1998; MacLehose, McKee, and Weinberg 2002; Newton, Grimaud, and Weinberg, 1999; Reichhardt 1998; Tibayrenc 1998; 1999).

This debate explains how these two policy ideas survived the primeval soup. The Charter Group seems to have won the battle: as a matter of technical feasibility, the Charter group was a “light structure” designed for the exchange of information not the funding and hosting of research facilities that is why a “bricks and mortar” solution was not deemed necessary. In terms of value acceptability, this was a clash between the creation of a European‐wide institution and the reluctance of national institutionalized experts to cede sovereignty over public health policy. More importantly, the debate revolved around how each idea had the potential to be used by policymakers. However, I will demonstrate later on that the idea of an agency had limited receptivity in the politics stream.

Beyond the criteria of survival, this analysis of the policy streams ought to assess the increasing returns of the two different ideas. The initiative of the Charter Group is the only one of the two to demonstrate a cost for the policy community and the policymakers and as such fulfill the criteria. This being demonstrated, the next part ought to investigate how the evolving ripeness of the politics stream.

The Ripeness of the Politics Stream

The European Commission had a long history with the Charter group. It supported them throughout the 1990s. In 1993, the soon to be fathers of the Charter group put a proposal to the European Commission for a grant to draw up an inventory of all the international surveillance and training collaborations that were currently taking place in the EU, the grant was accepted (Bartlett 1998). Then, a representative of the Commission (DG V: Employment & Social Affairs) participated in the meeting of the Charter group. Eventually, the results of the prioritization of communicable diseases were communicated to the EU Commission as “expert advice” (Newton, Grimaud, and Weinberg, 1999). The support for this idea culminated in 1999 with the proposal for a decision creating a network for the epidemiological surveillance and control of communicable diseases in the European Communities, mentioned earlier.

The institutional bargaining that occurred during the decision‐making on the proposal for a decision offers precise insights to map the institutions' preferences on the control and prevention of diseases. In the case of the EP, this happens in a context of growing hostility towards agencification where new agencies are seen as an “irresponsible” development of the European Executive that endangers the balance of power between elected bodies and technocratic institutions (Lord 2011, 912, quoting the Herman Report, European Parliament 1999). Nevertheless, the EP has demonstrated no suspicion toward the creation of an agency dedicated to the surveillance, control, and prevention of communicable diseases and even championed the creation of an institution from the beginning (Bowis 2004). What explains the role of the EP in championing the Agency approach rather than the Network approach? The agency idea presents technical characteristics that leave room for a greater oversight of the EP. While agencification is usually seen as a phenomenon that accentuates deparliamentarization (Lord 2011, 913), the case of the ECDC differs from policy sectors of exclusive competences. The EP's preferences were formulated in a policy vacuum and it is then not surprising that an institution traditionally eager to secure oversight makes the choice of the most institutionalized option. Lord comes to this conclusion about the EP when the other option is a “voluntary pattern of co‐operation” (2011, 915). Kelemen (2014) underlines that, where the EP has influence over agency design, it tends to weight on the creation of bureaucratic structures to enhance the transparency and accountability of agencies. The EP thus favors the Agency idea because it gives the Agency a stronger value acceptability than a network approach, based on oversight considerations. The amendments of the EP have been consistently suggesting creating a center rather than a network: “Having regard to the current shortcomings in the structures for the epidemiological surveillance of communicable diseases in the Member States and, therefore, the need to establish a permanent structure at Community level”; and suggests “collecting information relating to epidemiological surveillance and coordinating control measures in order then to forward them to a central body: the European Centre for the Surveillance of Communicable Diseases.” (Cabrol 1997).

This approach was systematically countered by the EU Commission and the Council: “The situation is so diverse that one cannot talk about ‘shortcomings in structures’. It would be wiser to refer to ‘increasing needs’. The common position has largely taken over the text of the Commission's amended proposal”; “The term ‘Eurocentres’ does not properly describe these structures, whose activities are mainly geared towards national surveillance. […] On the contrary, it is likely to precipitate objections from the Council” (Commission 1998). The Council of Ministers was indeed favoring the network approach, with three prominent advocates: Spain, Sweden, and the UK showing clear concerns over the financing arrangements for the system, and clear preferences to leave the operational costs of the network to be financed by member states themselves (Council of the European Union 1997a; 1997b; 1997c; 1997d). This position was interpreted within the policy community as the reluctance of the EU's member states to cede sovereignty over public health policy (Butler 1998).

In the late 1990s the politics stream was thus more receptive to the idea of a network rather than the idea of an Agency. The role of increasing returns is probing: the decision to set up a network bares no additional costs economically and very little cost politically. However, the value acceptability argument shows that, at this moment, the politics stream was not fully “ripe” to the idea of a European agency. Nevertheless, this was subject to change in 2002, with the European Commissioner for Health and Consumer Protection mentioning in a speech the ambition of setting up an agency, “we have committed ourselves to creating a European Centre for Disease Control by 2005. This will bring together the expertise in Member States and will act as a reference and co‐ordination point both in routine and in crisis situations.” (European Commission 2002).

A classic explanation of this change in preferences in the MSA is the renewal of the members of the Commission between. In 1999 the Santer/Marin Commission's mandate came to term and Pádraig Flynn was replaced by David Byrne as the European Commissioner for Health and Consumer Protection. The literature on the ECDC underlines the role of the severe acute respiratory syndrome (SARS) crisis in the Commission coming to terms with the idea of an Agency (Greer 2012; Greer and Löblová 2016). However, the dates do not exactly add‐up. The SARS crisis began in November 2002 with an outbreak in southern China, while Commissioner Byrne mentioned the ECDC, for the first time, in September 2002. In the member states corner as well, some receptivity is observable before SARS happened. As early as June 2001, the possibility of a “European Centre” was mentioned in the conclusions of the European Council at Gothenburg (European Commission 2003b), where concerns about bioterrorism were specifically underlined.

The evolution of the politics stream ripeness shows that on the eve of the SARS crisis, policymakers were ripe to both ideas. This is due to the Commission and the Council are becoming ripe to the idea of a new agency, while, in the meantime, increasing returns of the network were consolidated and its proponents still vocal in the Council. The next part underlines the changing nature of the problem stream.

Ambiguity in the Problem Stream

The evolution of the problem stream, and the way problems have been perceived, can be used to understand how eventually conditions were framed in a way that opened a problem window.

The problem of disease prevention and control in Europe in the 1990s differs from the general context of health policy in the EU. The development of a health policy in the EU has been described as an incremental development likened to a spillover dynamic: since most areas linked to health are progressively integrated, health will be eventually integrated (Greer 2006). If not relying on grand theories of integration, the assumptions have been that health policy was progressively Europeanized (Böhm and Landwehr 2013). However, the case of disease prevention is particular in the sense that the Maastricht Treaty (art. 129) gave, for the first time, legal competences to the European Commission to complement national policies, within the limits of disease prevention, health information, and education. The Maastricht Treaty paved the way for this opportunity for the fathers of the Charter group by asking the Commission to finance their initiative. This demonstrates that this policy idea emerged thanks to the realization of a problem that was recognized in the treaties, in line with the neo‐functionalist or the Europeanization assumptions presented earlier. The problem is thus perceived as a consequence of the spillover of European integration: since borders are now open and microbes know no borders, transnational cooperation is needed to tackle potential transnational health problems.

Through the 1990s, there was a debate among experts and institutions on the possibility to create an agency. The problem was framed by the policy community and the institutions as follows: what kind of integration is desirable in the field of prevention and control of communicable diseases? The terms of the problem were never properly defined by the Charter group or the proponents of the ECID beyond spillover considerations; however, the institutional archives provide a clear understanding of how the problem was framed depending on the solution championed by the institutions. While the EP describes the problem of disease prevention and control as a matter of “shortcomings in structure” (Cabrol 1997), the Commission and the Council are more restrained and, respectively, underline “growing” (Draft Minutes of the 2131st meeting of the Council (Health) 1998) and “increasing needs” (European Commission 1998).

Problem perception changed radically in the 2000s due to the persistence of health crises at the end of the 1990s, whereas food‐borne disease (“Mad Cow” disease) or bio‐terrorism (the post 9/11 anthrax contamination) shifted the issue from a functional problem to the recognition of threats. This trend culminated in the early 2000 with the events of the SARS crisis.

The following part is dedicated to understanding how the Commission framed this focusing event in a way that led to bricolage. To draw a clear picture of the opening of the policy window and bricolage I first present reflections on the reasons why changes in the politics and the problem stream before the SARS crisis never led to coupling the streams.

Policy Windows, Failed Entrepreneurship, and Contingent Framing of the Problem

The process tracing of the three streams has let us to draw a precise image of the context. Before investigating the framing of the SARS, some of these contextual elements must be clarified in order to understand why some conditions that could be interpreted as windows of opportunity did not lead to change and why some agents who could be interpreted as policy entrepreneurs did not succeed in coupling the streams.

The first case would be to address the Charter Group. The Charter Group was arguably demonstrating some elements of policy entrepreneurship. Nevertheless, it is not a good fit for the MSA: the Charter Group was an experiment that happened outside of the traditional decision‐making channels, its relationship with the decision makers was different from a policy entrepreneur seeking access to decision makers. The decision‐making process of setting up a network for the epidemiological surveillance in 1998 could hardly be considered a coupling of the streams: the change is minimal and the result is a formalization of the EU's sponsorship of the initiative. Therefore, there is here a form of entrepreneurship with minimalist goals and fostering change at a modest level rather than by starting a complete policy cycle.

The second case is the policy entrepreneurship behind the ECID that fits the traditional definition. It is a case of failed entrepreneurship, probably due to the limited outreach of the proposal presented by Michel Tibayrenc and his lack of access to EU institutions. There was a change in the politics stream that happened just before, but the relative weakness of the EP did not allow a proper window to be opened. Eventually, the ECID remained a policy idea that was stifled by the debate in the policy community and efforts to make it emerged never met the right opportunity.

The final case before the SARS crisis is the opening of policy window where a change in the problem streams due to Mad Cow disease and the Anthrax attacks led the Member States to become ripe to the idea of an agency. Moreover, the new Commissioner for health who had taken office in 1999 seemed more receptive to the creation of an agency. Despite a window of opportunity being arguably opened, no agent emerged as a policy entrepreneur to set the agenda. The Commissioner for Health, Byrne, only mentions that he and his team had “committed to themselves” to creating a European Centre for Disease Control by 2005. There were no clear preferences stated on the precise organizational elements of this agency and the policy idea was exposed vaguely. Moreover, he was giving a speech for an audience of young specialists of public health in a forum dedicated to health issues in the EU (European Commission 2002) and there was no other evidence that there had been a concrete effort made to put the issue on the agenda. Perhaps the Commission could have exercised some entrepreneurship if the salience of SARS had not had an overwhelming impact on the problem stream only 2 months after Commissioner Byrne's announcement. Empirical evidences only show that Commissioner Byrne was ripe to the idea, not that he engaged in policy entrepreneurship.

The SARS crisis has been a focusing event in the sense used by Birkland (1998a) that it was sudden, rare, and potentially harmful. Nevertheless, the impact of SARS on the European continent was limited. According to the WHO (2015) in the month preceding the Commission's proposal (July 2003), 33 cases had been reported in the European continent—excluding Russia (31 in the EU). At the end of year 2003, the WHO concluded that 25 cases were confirmed in the EU, 27 in total for the continent (excluding Russia), with one case resulting in the death of the patient. All cases were imported; there was no domestic spread of the epidemic, no local transmission. At global level, however, 8096 cases were confirmed.

The crisis seemed a matter of global scale rather than a continental one. But the SARS crisis was framed by the European Commission as the example of a problem that if inflicted to the European continent would be devastating without a European Agency. First, the Commission framed SARS as a threat, even if the threat was limited “Communicable disease outbreaks can pose a significant threat to the health and well‐being of the European Union's citizens, as shown during the recent spread of the SARS virus” (European Commission 2003a); but clarified that this was only one of the possible iterations of a systemic risk: “A major outbreak such an influenza pandemic could have catastrophic consequences” (European Commission 2003a). The risk of a similar outbreak was thus what the proposal suggested to tackle: the Commission framed the SARS to show that competencies were lacking at the EU level, defining “public health matters as an area where Community competencies should be consolidated” (European Commission 2003a; 2003b). The solution the European Commission was looking for was not a tool to solve the threat of the SARS outbreak; rather it was the result of considerations that the EU could be at risk. Therefore, the Commission defined the problem as the recognition of new needs for the EU and emerged as an agent taking responsibility for finding a solution.

I identify the first step of bricolage done by the Commission which is to open a window by framing conditions as a problem that can and must be fixed. The next step is to engage on a path of consequential search for fit. Taking cues from bricolage, I reason that the Commission pieced together a good fit for the definition of the problem: fill the capacity building gaps at the EU level in the field of disease prevention and control. Therefore, I now investigate the policy formulation aspect of bricolage.

Coupling by Bricolage: Policy Formulation

The last parts of this paper gave us the picture of the streams, up to the opening of the policy window. After the presentation of the mechanisms that led to bricolage rather than the emergence of a policy entrepreneur, this part focuses on bricolage as policy formulation and the consequences on the decision‐making process.

Bricolage, as an oblique way to formulate policy, will now guide our understanding of how different ideas were pieced together to create a solution that would go swiftly through the decision‐making phase. The European Commission had two different policy ideas at hand. On the one hand, the network had strong, consolidated increasing returns and, on the other hand, the policymakers had been increasingly riper to the creation of a new agency.

The process tracing showed that the Commission was wary of the Communicable Disease Network created in 1999 and used it as the basis for the proposal “The basic formula for cooperation amongst Member States and the Commission in the framework of Decision 2119/98/EC is not being questioned” (European Commission 2003a). This shows the Commission paying attention to spare the cost of exiting organizational choices set up previously. Nevertheless, it presents the agency idea as the core concept, while the agency idea is actually superimposed on the existing network.

The proposal did not retain key features of the suggested ECID such as financing and hosting research labs. The ECDC is not a European “CDC” based on the US model but a “hub” (Greer and Matzke 2012), a center that coordinates a network, composed of different authorities in charge of epidemiological surveillance in the EU. It retains all the existing features (including, for instance, the publication of Eurosurveillance) and is still based on the coordination and “synergies between the existing national centres for disease control” (European Commission 2003a). National information and expertise are still predominant in the functioning of the center, whereas for data exchange or for training purposes. The Commission pieced together a modest project “a large European Centre is not needed” (European Commission 2003a), which also prevents the most skeptical elements of the politics stream from opposing the proposal. The proposal also included the mention that the ECDC would be an agency without regulatory powers (European Commission 2003a), which echoed the concerns raised by two member states: the UK and Germany (Council of the European Union 2004).

This shows that bricolage was a process of creating a European Agency while incorporating many elements of the epidemiological network. By bricolage, the European Commission avoided the costs of exiting the initiative in which time and money had been invested while taking into account the changed ripeness of policymakers vis‐à‐vis the creation of agency.

Bricolage is thus seminal in understanding why a proposal is swiftly adopted, not the crisis itself. The Commission's proposal required only one reading facilitated by a conciliation meeting that was set up early on in the process, a practice that is relatively unusual in the inter‐institutional bargaining at the EU level. Rather than a beacon of the consensual culture of decision‐making in Brussels, the decision‐making process was the result of the European Commission formulating a policy solution that “fit”, which led to speeding‐up policy formulation and decision‐making.

Evidence that the European Commission did not act as a policy entrepreneur shows in the policy solution that was eventually formulated. Moreover, the proposal does not seem to serve the Commission's needs or interests. The literature on agencification would underline that the creation of an independent agency presents reputational elements for the Commission. While being a convincing explanation for agencification in general, it is limited in the case of the ECDC and not supported by evidence: the ECDC is a discreet agency. As demonstrated by the limit of its competences, the ECDC reflects the Member states' preferences more than the Commission's needs and interests in seeking reputational gains. The recombination of ideas is, however, in line with the assumption that the proposal was an “immediate action [is] motivated more by the need to avoid higher costs rather than the need to reap more benefits” as defined by Zahariadis in his definition of consequential coupling and search for fit (Zahariadis 2003, 72). The proposal of the Commission is thus the work of a bricoleur and not the pet solution of a policy entrepreneur who links her solution to the SARS crisis.

As a type of coupling in the MSA, bricolage shows that after framing the problem and with a process goal, the Commission acting as a bricoleur created a bespoke solution that was a new arrangement of elements. In creating a bespoke solution, the Commission made an arbitration between the different advantages of policy ideas and paid attention to the ripeness of policymakers, which led to a swift adoption and implementation of the Regulation.

Conclusion

The creation of the ECDC has been for long time interpreted in the light of the SARS crisis; however, our process tracing has shown that this crisis is a catalyst rather than the genesis of this agency. By showing that the ECDC is the result of a bricolage, this paper empirically proved that the traditional explanation found in the literature on the ECDC must be refined. Empirical elements such as the swiftness of the decision‐making process and the hybrid features of the ECDC find stronger explanatory leverage in the process of policy formulation by bricolage than in the reaction to the crisis.

The creation of the ECDC was the contingent result of the Commission framing a problem due to events that had a limited impact on the European continent and looking for the solution that would fit this ambiguous context. No agent championing a solution could couple the streams. No policy entrepreneur emerged as the “hero” with a solution to a “crisis.” This paper empirically confirmed an important property of bricolage: the bricoleur is an agent who frames the problem and sets her own goals: the consequential search for fit. This new take on agency in the MSA is a response to criticisms that underline that ambiguity of preferences clashes with the existence of a policy entrepreneur with a clear personal agenda (Zohlnhöfer and Rüb 2016). In bricolage, the agent does not have clear preferences on how to solve the problem. The micro‐foundations of the bricolage are that preferences regarding solutions are secondary; the issue of problem solving is at the core of this type of agency. The lack of preferences beyond looking forward to solving a problem explains why solutions are the result of an oblique mechanism. Bricolage produces a new arrangement of elements that are preconstrained because resources are scarce and cannot be expanded. In our case study, bricolage is performed by the European Commission who pieced together a policy solution that was based on previously consolidated organizational choices. Here the process tracing showed that the Commission's choice relates to the increasing returns of the network of epidemiological surveillance and the evolving ripeness of policymakers to the idea of an agency. Our bricoleur avoided the costs of exit from the network and combined elements of the agency idea to the solution because the member states were ripe to it.

Bringing bricolage into the MSA widens our understanding of agency in policy formulation. Unlike the policy entrepreneur who “softens‐up” the policymakers to their ideas, the bricoleur makes use of the policy environment to piece a solution. Bricolage is a process of arbitration and recombination that creates bespoke solutions to fit a problem. This paper empirically proved that a bricoleur evaluates policy ideas by the yardstick of two criteria: the ripeness of decision‐makers and the increasing returns of policy ideas.

Who can become a bricoleur? Is bricolage an individual property, or can a constellation of actors collectively engage in bricolage? The fact that the bricoleur is a process‐oriented actor chimes with the organizational properties of the Commission, which is an actor that follows process goals to outcome goals. In the EU system, other institutions may be problem‐minded rather solution‐minded depending on the issues at stake – this is an open question for further empirical research. I suspect that the role of bricoleur would suit a single Member of the European Parliament (MEP) or a national delegate within a Council formation. Collectively the MEPs could become a bricoleur when the EP identifies problems that the Commission is reluctant to deal with. While the policy entrepreneur is defined by their activities rather than by means of their position, the bricoleur is a concept that rather suits policymakers until empirically proven wrong.

A distinction should be made between bricolage and other concepts with which it shares some features. Cram (1993) introduced the purposeful opportunist using a very similar empirical case: a policy field in which the EU has reduced competences but in which and there is still a form of European integration. The purposeful opportunist is different from the bricoleur, while both are empirically identified as the Commission, and both can be seen as process goal driven rather than solution goal driven. Important differences remain. Cram's purposeful opportunist is an agent that accumulates knowledge until the moment it can seize an opportunity. The bricoleur is less strategic in cumulating reservoirs of knowledge that can be mobilized when the right moment comes. Bricolage can also be put in relation to models of decision. Prima facie it may resemble muddling through (Lindblom 1959), especially considering the importance of the concept ripeness of the politics stream. Bricolage and muddling through are, nevertheless, different. Muddling through is a collective process of partisan mutual adjustment. Bricolage takes partisan adjustments as one of the elements of the equation along with increasing returns of ideas. Moreover, these adjustments are not mutual, rather the bricoleur should be seen as one who has a precise idea of the policymakers' preferences and who is able to come up with a solution in the function of these preferences.

In conclusion, this article shows that bricolage can assist in the development of the MSA. It also provides a lens to re‐examine the relationship between framing problems and mobilizingideas, and provides a richer understanding of how exactly agency generates change. Finally, it allows us to scale down some claims made on crises as levers of change.

Acknowledgements

I would like to express my immense gratitude to Claudio Radaelli for his support during the process of writing this paper. I would like to thank the two reviewers with whom I engaged in a discussion with great pleasure and who provided excellent comments. Special thanks to attendees and discussants of the panel ‘The Multiple Streams Framework: Empirical Applications and Theoretical Innovations’ at the 2016 ECPR Conference in Prague for their comments, to Jonathan Kamkhaji for being my intellectual sparring partner and to Sophie Glaser for her inestimable support.