An interesting starting point for the elaboration of a coherent theoretical framework concerning the impact of policy ideas is John W. Kingdon's agenda-setting theory.3 Although distinct from historical institutionalism, this theory can provide institutionalist scholars with insights about the specific role of policy ideas in policy-making. Drawing on Kingdon's distinction between agendas and alternatives, this section identifies actors and ideational processes present at different stages of the policy-making process. Defining these processes in a rigorous manner is crucial in order to avoid vague statements about ideas that have long discredited ideational theories (Berman 1998).
The concept of agenda refers to “the list of subjects or problems to which governmental officials, and people outside of government closely associated with those officials, are paying some serious attention at any given time” (Kingdon 1995: 3). Consequently, agenda-setting is the process that narrows the “set of conceivable subjects to the set that actually becomes the focus of attention” (ibid.). Starting from this definition, we need to take into account the well-known distinction between public and policy agendas. If public agenda generally refers to the interaction between public opinion and issue salience in the media, policy agenda concerns the problems policy-makers themselves perceive as significant at a specific moment in time (Soroka 2002: 7–8). Although public and policy agendas are related, we should acknowledge their relative autonomy. Kingdon essentially looks at the policy agenda, which explains why students of policy-making tend to draw on his work (for example, Hacker 1997).
In his work, agenda refers to a cluster of issues considered as the “pressing problems of the moment”, and alternatives represent the policy options available to solve these problems (Kingdon 1995: 4). Agendas and alternatives are the product of the interaction between three autonomous streams through which social and political actors mobilize in order to promote specific issues or policy options. A critical examination of these three elements—problem, policy and political streams—will lead to a broader discussion concerning the relations between ideas and institutions in policy-making. Furthermore, this discussion will show that policy alternatives are rooted in specific policy paradigms, and that political actors frame these alternatives in order to sell them to the public while constructing the need to reform.
Problem stream: agendas
The first element of Kingdon's model is the problem stream. It refers to the selection of issues that are considered significant social and economic problems. Potentially relevant issues are numerous, yet the state cannot address all of them at once. If many issues die away because attention fades, some problems are of a cyclical nature and tend to correct themselves over time (1995: 106). Because policy-makers can only focus on a few core issues simultaneously, the political construction and selection of the problems on the agenda constitute a key phase of the policy-making process. As a result, beliefs about what are the most pressing problems of the day must be taken into account in the study of social policy reform.
Bureaucrats, elected politicians and the public generally become aware of socially constructed economic and social problems through statistical indicators, spectacular “focusing events”, or feedback effects from previously enacted policies (1995: 90–164). If statistical indicators represent the various measurements of economic and social conditions (growth rates, unemployment rates, poverty rates), focusing events refer to dramatic episodes that attract media attention such as natural catastrophes or unexpected political developments. For example, in 1991, the surprising election to the Senate of Harris Wofford, a Democratic (Pennsylvania) candidate campaigning in favour of universal health insurance, was the focusing event that convinced many US federal officials to seriously address the issue of health care reform (Hacker 1997).
Finally, feedback effects from existing policies can take the form of complaints about the functioning of these programmes and their perceived socio-economic impact. From an institutionalist viewpoint, the problems on the policy agendas are frequently constructed through social learning processes discussed above. Furthermore, as Peter Hall and others have argued, principles and causal beliefs embedded in specific policy paradigms shape most if not all learning processes. Consequently, the autonomy of the three streams is probably more limited than Kingdon would argue: “a historical perspective shows that these streams are linked in important ways over time. Policies from an earlier period can affect each of these streams at a later time. The conception of what the problems are and how they should be defined very often depends on previous policies, which establish some groups as authoritative voices in a particular field and make other perspectives less credible” (Weir 1992: 18). While avoiding rigid historical determinism, scholars could benefit from studying the three streams from a long-term historical perspective. Although autonomous, the problem stream itself is at least partly structured by existing policy legacies.
A fine example of the impact of policy legacies on the problem stream is the debate over “social exclusion” in France during the 1980s and 1990s. At the time, endemic long-term unemployment and social exclusion appeared in part as a feedback effect of that country's social insurance system and labour market policies that had emerged decades earlier (Béland and Hansen 2000). Another example of the relationship between policy legacies and the problem stream is the debate concerning social assistance reform in the United States between the late 1960s and the enactment of the 1996 Personal Responsibility and Work Opportunity Reconciliation Act. Because US social assistance benefits had targeted single parents since the Progressive Era (Skocpol 1992), the debate focused on single mothers and family structure, issues that proved less central in Canada and in most Western European countries. As these two examples show, the construction of social problems and policy issues within the problem stream is frequently related to long-term institutional legacies.
Policy stream: alternatives and paradigms
The policy stream gathers together policy experts working for academic institutions, governmental agencies and interest groups. Involved in a “policy community” related to a specific policy area (for example, health care), these actors frame alternatives and legislative proposals that stew in what Kingdon calls a “policy primeval soup”, which contains many policy ideas floating around that “combine with one another in various ways” (Kingdon 1995: 117). Rejecting rationalistic visions of policy-making, which would start from the assumption that policy ideas always emerge as responses to well-known problems, he states that some experts and interest groups can promote a policy alternative in the absence of a clear problem to solve. Moreover, a policy solution designed as a response to a particular problem can actually get attached to another problem if needed.
Against the impression that policy ideas have little consistency, it must be stressed that most alternatives are grounded in a policy paradigm, which constitutes the structured intellectual background of policy decisions. These paradigms serve as “road maps” to experts and policy-makers by providing them with a relatively coherent set of assumptions about the functioning of economic, political and social institutions. For that reason, actors who share the same paradigm “will make similar choices over time, even as the environment changes” (Berman 1998: 32). Far from being purely cognitive, paradigms are inherently normative and programmatic: they help policy-makers decide how to reform existing programmes, or to create new ones. If “general paradigms” concerning gender roles or economic regulation impact on political decisions across policy areas, “sectoral paradigms” belong to a specific policy area (Merrien 2001). Sectoral paradigms are frequently embedded within a general paradigm.
Although they can integrate the state bureaucracy, these sectoral paradigms frequently emerge in the “parapolitical sphere”, located at the interstices of business, government and academia (Horne 2002). Because national institutions and policy settings shape the role of policy experts, the parapolitical sphere is organized differently from one country to another. In the United States, their actions are institutionalized through the establishment of “think tanks” and academic “research institutes”, which are generally autonomous from political parties and the state (McGann and Weaver 2000). Yet, in other countries, political institutions structure the production of expertise and policy paradigms. In France, for example, many policy experts work for the CNRS (Centre National de Recherche Scientifique), a large, state-financed research institute and a governmental cabinet ministériel also performs tasks generally associated with think tanks (Gaffney 1991). Moreover, state capacities and administrative structures frame the action of bureaucrats in charge of policy design (Skocpol 1992). Policy ideas and political institutions interact within and outside of state boundaries, and paradigms are institutionalized through policy implementation and formal social learning processes like commissions and expert panels. At a deeper level, a country's political institutions (for example, the level of territorial decentralization) can also shape the way policy-makers construct, perceive and select debated policy alternatives (Dobbin 2004). But policy networks transcend national boundaries, which allow specific alternatives and paradigms to spread at the international level. In order to understand the policy stream better, scholars should then take into account the international circulation of policy ideas. For example, in the field of social assistance and employment policy, there is strong evidence that policy ideas about workfare emanating from the US had an impact on British New Labour's activation programmes for the unemployed (Daguerre 2004). In recent years, scholars have also underlined the role of international organizations like the World Bank in the propagation of policy ideas (Merrien 2001). These remarks should not obscure the fact that national cultural representations and institutional legacies restrict the transferability of policy ideas from one country to another (Schmidt 2002b: 250; see also Katzenstein 1996).
Finally, alternatives and paradigms have a dialogical nature: each of them exists only in opposition to other policy ideas available in a particular policy environment at a precise moment in time. On the one hand, policy entrepreneurs supporting new alternatives not only depict them in a manner that appeals to the public, but also attempt to undermine public support for existing programmes. On the other hand, if new policy ideas become popular, those committed to established alternatives and paradigms either integrate them into their frameworks or justify their exclusion.
The content of the debate over “Social Security reform” in the USA during the second half of the 1990s illustrates the dialogical nature of alternative formation. On the one hand, US experts and politicians supporting “Social Security reform”—the shift from pay-as-you-go old-age insurance to forced savings—argued that the federal old-age insurance programme constituted a “bad deal” for current and future workers. According to them, population ageing will take enormous fiscal resources away from young and future adults. This represented an attempt to undermine the political support for the existing federal old-age insurance programme. On the other hand, excellent stock-market performances, the growth of private savings schemes, and the growing public profile of discourse over “Social Security reform” encouraged those who opposed this policy alternative to support the investment of social security surpluses in equity. Because it proved difficult to resist the financial logic in the context of booming stock-market performances, this form of investment appeared as a suitable way to benefit from financial returns without embracing a risky privatization of old-age insurance (Béland and Waddan 2000).
Political stream: frames and policy entrepreneurs
If policy experts constantly discuss suitable alternatives within their policy communities, their proposals need to be associated with a problem perceived as significant in order to reach the policy agenda. Yet this is possible only if these proposals receive the direct backing of a major political advocate. Consequently, what is occurring within the political stream is crucial in determining policy outcomes. Among the factors that set down this third stream are electoral results, pressures from interest groups, and the perceived state of public opinion. The election of a new government or the emergence of a powerful interest group can reshape the policy agenda and help push new policy ideas to the centre of the political debate. During these moments of political opportunity, policy entrepreneurs are instrumental in bringing together their own favoured solutions with a recognized social or economic problem. Policy entrepreneurs favour the convergence of the three streams necessary to the legislative triumph of their policy ideas (Kingdon 1995: 122).
The role of policy entrepreneurs underlines the relationship between timing, policy ideas, strategic interests, and political institutions in policy-making. First, as suggested above, in order to affect political decisions and become institutionally embedded, policy ideas need the support of powerful actors that have an interest in promoting them (Goldstein and Keohane 1993: 21). Without such support, policy ideas cannot find their way to the legislative arena: “Ideas are out there and they have an impact on politics only when seized upon by political actors and through this process find an entry point into politics” (Hansen and King 2001: 259). Although policy ideas rarely constitute a mere reflection of existing strategic interests, they have little policy influence if no powerful actor—individual or collective—emerges to promote them. Second, formal political institutions largely determine which actors are in a strong position to campaign for a policy alternative on the legislative agenda. From this perspective, it is important to take into account cross-national institutional differences in order to understand who the influential policy entrepreneurs are (Schmidt 2002a). Third, the moment at which a policy entrepreneur attempts to promote a policy alternative is crucial in determining its level of political influence. A policy alternative popular under particular economic and political conditions may become far less acceptable if these conditions change over time. For example, the advent of the Nazi regime in 1930s Germany reduced the legitimacy of eugenic policy alternatives that had been increasingly popular in Western Europe and North America since the late nineteenth century (Hansen and King 2001).
Policy entrepreneurs succeed in imposing certain policy ideas partly because they appeal to the public through the mobilization of political symbols ever-present in the shared ideological repertoires available in their society. The concept of repertoire refers to a relatively coherent set of cultural symbols and political representations mobilized during policy debates to frame the issues and shape public opinion (Marx Ferree 2003). During legislative debates as well as after the enactment of particular measures, policy-makers must justify their political and technical choices. In this context, policy actors need “symbols and concepts with which to frame solutions to policy problems in normatively acceptable terms through transposition and bricolage” (Campbell 1998: 394). In France, for example, references to the concepts of citizenship and solidarity central to the French Republican tradition justified the enactment of major social policy reforms during the 1980s and 1990s (Béland and Hansen 2000).
Drawing from existing ideological repertoires, elected politicians legitimize their programmes in order to reduce political risks or claim credit for their potentially positive social and economic consequences. Ideological frames “appear typically in the public pronouncements of policy-makers and their aides, such as sound bites, campaign speeches, press releases, and other very public statements designed to muster public support for policy proposals” (Campbell 1998: 394). The framing process is indeed a strategic and deliberate activity aimed at generating public support for specific policy ideas. Moreover, the capacity to communicate ideological frames to the targeted audience is essential to experts and political actors who seek to legitimize existing policy decisions (1998: 397). In a sense, the need for policy-makers to frame issues using culturally accepted repertoires shows that “public opinion matters” (Burstein 1998). Frequently, these frames take the form of “policy lessons” aimed at convincing the population to support a specific policy alternative. As noted above, frames are not policy ideas in the strict sense of the term: they constitute a discourse that helps political actors sell policy choices to the public.
Like policy alternatives and paradigms, frames are dialogical in nature: they anticipate what potential opponents could say to undermine the support for specific policy alternatives. From this perspective, frames have a preventive component, in the sense that those involved in policy debates frequently mobilize them in order to shield their policy proposals from criticism. Frames can take the form of “strategic misconceptions” that mask the actual functioning—or the negative consequences—of specific public policies. The postwar discourse of the US Social Security Administration—misleadingly—depicting the federal old-age insurance programme as a quasi-savings scheme is an example of “strategic misconception” related to the framing process (Derthick 1979). Policy-makers can also frame policy alternatives in a manner that hides their actual departure from a well-accepted paradigm. For example, Swedish politicians refer to social democracy in order to legitimize reform options that have little to do with the core assumptions of this global paradigm (Cox 2004).
More importantly, ideological framing contributes to “the social construction of the need to reform”. “In a political environment the advocates of reform need to employ strategies to overcome the scepticism of others and persuade them of the importance of reform. In other words, they must create a discourse that changes the collective understanding of the welfare state, because doing so ‘shapes the path’ necessary to enact reform” (Cox 2001: 475). When supporting significant changes, policy entrepreneurs have to justify the need to reform but, simultaneously, shake up the existing “policy monopoly” (Baumgartner and Jones 1993) that favours the reproduction of previously enacted measures through institutional inertia and ideological justification. For example, publications like Charles Murray's Losing Ground (1984) played a significant role in justifying the radical reform of the US federal social assistance system enacted in 1996. Conservatives like Murray described Democrats who opposed the shift from welfare to workfare (work-for-welfare) and time limits as blind ideologues unable to realize how unreasonable and perverse federal social assistance programmes were (Somers and Block, forthcoming). In the United Kingdom, Margaret Thatcher's discourse about the need to “end the dependency culture” proved effective during her first term as prime minister. Although institutional inertia did not allow her government to “dismantle the welfare state” (Pierson 1994), the consequences of the conservative reforms in labour relations and housing policy, for example, proved extremely significant. According to Vivian Schmidt, Thatcher's superior capacity to justify the need to reform explains why her neo-liberal campaign was more successful than the one launched more recently in New Zealand, a country where governing politicians failed to build a coherent and stable discourse to construct genuine neo-liberal imperatives that could convince the population to support their reform projects (Schmidt 2002a). Robert Cox (2001) develops a similar argument in order to explain why Germany, and not Denmark and the Netherlands, failed to implement massive social policy reforms during the 1990s. For him, governing German politicians, as opposed to their Danish and Dutch colleagues, failed to adequately construct the need to reform the welfare state. It is only recently that the German government enacted a substantial reform agenda.
The ability to frame a policy programme in a politically—and culturally—acceptable and desirable manner is a key factor that can help explain why some policy alternatives triumph over others and why elected officials decide to “do something” in the first place. Additional factors include support from key political constituencies, technical feasibility in the context of established policy frameworks, and the relative simplicity of the policy ideas themselves —very complex policy ideas are difficult to explain to the public and even to elected officials themselves. The fate of President Clinton's complicated Health Security proposal of 1993 provides ground to this claim. Because the plan seemed both ambiguous and difficult to explain to the public, Republicans found it easier to dismiss it and turn interest groups and the electorate against the Clinton administration (Skocpol 1996).
The political arena is a structured arena of conflict in which ideological frames form “weapons of mass persuasion” related to existing social and institutional forces. In such a constraining environment, political actors must master the institutional “rules of the game” while manipulating the symbols available in existing ideological repertoires. Their behaviour reflects ideational and institutional structures that create political obstacles as well as opportunities for reform. Although experts, public intellectuals, and political actors draw on existing cultural symbols, they can use them—or combine them—in original ways. As students of social movements have shown, framing is a dynamic and potentially innovative process (McAdam et al. 2001). In postwar Canada, for example, Québécois political leaders mobilized a nationalist rhetoric to oppose centralizing welfare state development before using the very same rhetoric to justify the enactment of progressive social programmes in the province of Québec (Béland and Lecours, forthcoming).
Paving the way for future research
In order to pave the way for future research concerning the relationship between policy ideas and social policy, I would like to suggest a few cautionary remarks about the role of policy ideas in welfare state development. First, policy ideas discussed in this paper are not necessarily confined to social policy debates; some alternatives and paradigms are influential across policy areas. The same remark applies to framing processes, which can impact on many policy areas simultaneously. Yet we could argue that, considering the large constituencies tied to modern social programmes, framing is especially crucial in welfare state politics. Because politicians willing to impose unpopular measures have to make constant efforts to avoid blame and justify the need to reform, frames have become even more central under the current “new politics of the welfare state”. To simultaneously impose pain and avoid blame, policy-makers must rely on frames that could convince the population to accept potentially unpopular reforms.
Second, the arguments put forward in this paper do not constitute an alternative to historical institutionalism, or even a fundamental challenge to Pierson's argument about path dependence (1996, 2000). Institutional processes tend to reproduce themselves over time, and drastic change is rare, at least in well-established policy areas where large constituencies have emerged. Even in these areas, path-departing change remains possible. In some cases, a policy sector may witness the emergence of path-departing logics that are not the direct product of the “external shocks”Pierson (2000) sees as the main source of large-scale institutional change. Scholars have shown that “third-order change” can occur if a paradigm shift takes place (Hall 1993), if the accumulation of incremental reforms slowly alters the institutional logic of existing policies (Palier and Bonoli 1999), and/or if power-holders are successful in constructing the need to reform that would legitimize path-departing reforms (Cox 2001). In this paper, I only argue that ideational forces can either favour significant policy change or reinforce existing institutional paths through the reproduction of a dominant paradigm and the production of frames justifying existing policy arrangements. More research is necessary to understand both institutional stability and institutional transformation, and a discussion about ideational processes should contribute to that task.
A more careful study of ideational processes can shed new light on crucial empirical puzzles that traditional historical institutionalism is unable to solve alone. For example, why are some issues becoming important for policy-makers while others are not even considered by them? The answer to this question lies in the study of the policy agenda and its change over time. How can scholars explain the content of specific social legislation? Although the study of institutional “veto points” and policy legacies can sometimes answer this question, the discussion above suggests that policy paradigms guide policy-makers in a way that is not reducible to institutional constraints. Shared economic and technical beliefs also influence the policy-making process. Finally, why do public views on particular policy issues change over time, sometimes in a shift way? Beyond broad changes in cultural values, economic structures and family relations, the way these issues are constructed in the political discourse may alter the manner in which people look at them. Overall, taking policy ideas seriously provides at least partial answers to the above questions.
Third, I recognize that ideational processes are not always decisive for explaining policy outcomes. For example, some policy processes dealing with highly technical matters may not involve much framing activity. Furthermore, political institutions influence the way frames and policy ideas (i.e. alternatives and paradigms) affect political debates. According to Schmidt, single-actor systems such as the United Kingdom and New Zealand, in which politicians take the main policy decisions, favour the domination of “communicative discourse” (frames) over “coordinative discourse” (paradigms). Inversely, multi-actor institutional systems like the Netherlands and Germany, in which politicians must seek agreements with “social partners”, favour the domination of “coordinative discourse” over “communicative discourse” because government officials have to convince labour and union officials to embrace their policy objectives before launching any reform (Schmidt 2002b). Although framing is probably more essential in multi-actor systems than Schmidt acknowledges (Cox 2001), her typology provides more ground to a central claim formulated in this paper: policy ideas and political institutions constitute analytically distinct levels of reality that intersect and impact on one another.
Finally, future empirical research on the role of ideas in social policy should seek to formulate a consistent methodological approach to evaluate the concrete influence of ideas on welfare state politics. For Berman (1998: 22), scholars have to address four questions in order to demonstrate that ideas, far from being purely epiphenomenal entities, can constitute an independent variable in the study of politics:
1. Are there real differences between the ideas held by different individuals or groups, and do they imply different policy choices on the part of those who hold them? 2. Is it possible to establish a plausible connection between these differences and the decisions made by political actors? 3. Did the relevant ideas predate the decisions being explained? 4. Is it possible to deduce the specific content of the ideas from knowledge of some other observable variable in the system at the time the decision was made?
Although these questions apply more to policy paradigms than to ideological frames, they can help in understanding the tasks necessary to demonstrate that policy ideas can shape particular legislative outcomes. Tracing the influence of policy ideas is possible; yet, because “smoking guns” are uncommon in ideational analysis, a cautious approach is needed in order to distinguish between types of policy ideas, and their concrete effects on policy-making and welfare state development. Because empirical evidence varies from one type of policy idea to another, data generation must take into account the distinction between alternatives and paradigms, and the one between these two types of policy ideas and ideological frames, i.e. discourses that help policy-makers sell policy alternatives to the public. For example, public opinion data may prove useful to evaluate the influence of framing activities on the electorate, yet they cannot contribute that much to our understanding of policy paradigms shared by experts and policy-makers. A careful and selective use of textual documents, public opinion surveys, and interviews with experts and policy-makers helps to provide evidence about the causal role of frames and policy ideas in legislative and policy processes. Considering significant ideational and institutional variations from one country to another, comparative analysis is especially useful to showing how and when ideas matter in politics (Cox 2001; Berman 1998; Hansen and King 2001; Schmidt 2002a, 2002b).