2.3.1 Path Dependency

The simplest pattern of development in policymaking and policy design thinking is to continue what has been done in the past. This implies a continuing interest in furthering technocratic approaches to design based on earlier thinking about target compliance and actor motivations, even with the realization of the number and type of wicked and complex problems facing governments and policy designers. At the same time in third-wave studies the language in this path-dependent version of policy design thinking has changed: it is now more open to participatory methods than some of the earlier approaches even though many of the fundamental perspectives on design persist. In particular, this perspective on design continues to place the continuing effectiveness of earlier regimes and instrument, and to a lesser extent efficiency, in a central position in policy analysis.

The path-dependent approach to policy design has also been influenced by developments in the allied discipline of public administration. The spread of the New Public Management, and market thinking in general in the public sector, has emphasized effectiveness and efficiency in the implementation of public programs, and those values have dominated the design of programs as well as their implementation under such rubrics (Barzelay, Reference Barzelay2001). This approach to policy shares some of the technocratic elements associated with earlier design thinking, which assumed that management science could provide guidance for implementation, and to some extent formulation, of programs in a neutral nonpartisan way.Footnote ⁶

Interestingly, the path-dependent approach to design often returns design thinking to the realms of incrementalism and bounded rationality. Fundamentally, doing what one has always done does indeed provide some guarantee of a minimal level (status quo) of effectiveness. Unless a program has been an abject failure, its persistence will mean the continuation of some level of effective service for clients. However, the (significant) advances in efficiency or effectiveness that other forms of design might offer remain out of reach.

At the same time, saying that designing may be path dependent does not imply that changes cannot take place; rather, it means that change may be driven less by agency and purposeful action than by bureaucratic regimes and standard operating practices, among other factors. It may also mean, however, that policy change will be driven less by hubris than by political interactions among the multiple participants involved in the process of making and implementing policy, as Charles Lindblom (Reference Lindblom1959) suggested, in his early work on incrementalism and its penchant for marginal adjustment to the status quo.

2.3.2 Diffusion as Design

Much of the first- and second-generation policy design literature appears to have assumed that design occurred as a tabula rasa, in which a completely new policy was developed from scratch and put into place, often after an older one had been scrapped. But the simplest way to design a program is to copy another, existing policy. This practice, highlighted in the academic study of policy diffusion, is hardly a novel occurrence, but in the contemporary period there has been increased interest in learning from innovations in other governments and studying this learning more systematically.

This is sometimes done for technical reasons – to see “what works” and where with respect to specific tool deployments and designs – but there is also often a political aspect in that successful practice in some other jurisdiction can provide a good deal of protection for politicians who advocate policy reforms, helping them to avoid blame for any failure or to claim credit for success, as the case may be (Hood, Reference Hood2002; Mesequer, Reference Mesequer2006). Moreover, direct copying can also reduce the time and expense involved in designing and field-testing a new intervention.

Although diffusion is an inexpensive means of designing, or rather using policy designs, it is less clear that it is as capable of producing effective policies as its advocates might have us believe. The simpler and more technical the issue is, the more likely diffusion is to be effective, and the more likely that evidence from one setting can be carried over to another. However, for the major challenges facing most policymaking systems, diffusion may encounter a number of important social, cultural, and political barriers (Pritchett & Woolcock Reference Pritchett and Woolcock2004). What we have then is two extremes: while the more technocratic advocates of design simply assume that good ideas would triumph, others remain skeptical and suggest that mindless emulation may result.

2.3.3 Evidence-Based Policymaking

The diffusion of policy designs has also intensified interest in policy design studies of the practices and pitfalls of evidence-based policymaking or the application of knowledge about tools and designs to policy decisions (Pawson, Reference Pawson2006). While most policymakers think that they have been involved in evidence-based policymaking for their entire careers, this catchphrase about policymaking reflects an effort on the part of government policymakers and critics alike to inject into decision-making more knowledge and information and less ideology and partisanship (see Botterill & Hindmoor, Reference Botterill and Hindmoor2012). Several national governments have embarked on major efforts to enhance the use of available evidence in their policymaking, although some policy areas – health and medicine in particular – appear to have been more successful than others (see Sackett et al., Reference Sackett, Rosenberg, Gray, Haynes and Richardson1996).Footnote ⁷

The use of evidence-based policymaking has also emphasized effectiveness as a key criterion in evaluating policy design. The reason for copying programs from other jurisdictions, or perhaps more commonly, gathering evidence from a number of programs, is that the information collected provides evidence about programs that appear to “work” and those that don’t. “Working” in this sense is often defined rather narrowly in terms of reaching specific targets and goals, and may ignore many of the secondary and tertiary consequences of the programs in question, let alone the manner in which, and by whom, evidence is collected, stored, and disseminated, which may well bias results in particular directions. It is a minimalist version of effectiveness, albeit still a concern about effectiveness.

2.3.4 Actor-Centered Designing

A third emerging aspect of contemporary design thinking moves it from the passive voice toward the active voice. While identifying the exceptions is extremely easy, much of the earlier design literature had not been clear about who the “designer” was and the role of individuals and individual behavior in producing these designs. There may have been some notion that very clever policy analysis would be involved in this process, or in the case of democratic designing that it would be participatory, but this emphasis was rather vague in many cases. Similarly, even the work of Schneider and Ingram (Reference Schneider and Ingram1997) which focused on design behavior did not provide many insights into who, exactly, it was that carried stereotypical and prejudicial views of policy “targets” into policy design activities.

Mark Considine and his colleagues (Reference Considine, Alexander and Lewis2009, Reference Considine, Alexander and Lewis2014), however, have advanced the agent-centered approach to design and to policy formulation more generally (see also Jordan & Turnpenny, Reference Jordan and Turnpenny2016), which has been very useful in helping to close this gap in knowledge. This emphasis on the role of the designer and a conceptualization of design as a craft is to some extent the antithesis of the more technocratic approach previously described and harkens back to the work of David Weimer in the early days of policy design studies (Weimer, Reference Weimer1992). The technocratic approach assumes that an algorithm can be identified and applied to a range of cases. The craft approach, on the other hand, highlights the unique characteristics of each decision situation and how the experience and implicit knowledge of designers are utilized to produce the policy designs.Footnote ⁸

Reflecting what Hal Colebatch (Reference Colebatch2017) has termed “authoritative instrumentalism,” this approach to policy design places the emphasis in designing on the role of political leaders. In so doing, it assigns permanent civil servants, policy advisors, and those who usually are associated with policy formulation to a less important role in the design process (but see Page, Reference Page2012), and reduces the public and nongovernmental actors to an even more subordinate position. This raises a number of questions, not about the capacity of typically short-term and generalist ministers to be effective designers in complex policy areas requiring specialized knowledge and experience only gatherable over a long period of time (Hallerberg & Wehner, Reference Hallerberg and Wehner2013), but about how ministers use available policy advice and how such advice is organized into policy advisory systems in order to supplement their own knowledge and help them reach their decisions (Seymour-Ure, Reference Seymour-Ure and Plowden1987; Barker & Peters, Reference Barker and Peter1993).

Analytically it is very important to consider agency more carefully in the design process. Considine et al. (Reference Considine, Alexander and Lewis2009, Reference Considine, Alexander and Lewis2014) appear very sanguine about the policy capacity of ministers, but existing evidence challenges this prospect. As already noted, this opens up the consideration of issues of policy advice and the manner in which a range of actors involved in the process interact to generate designs (Halligan, Reference Halligan, Peters and Savoie1995; Craft & Howlett, Reference Craft and Howlett2012). As Considine et al. note, assessing the relative effectiveness of designs that are generated through different processes involving different actors forms a significant research agenda for the new policy design studies.

2.3.5 Coping with Wickedness Directly

Another aspect of policy design studies emerging in the past several years reflects analytic, and to some extent practical, attempts to cope with “wicked” or “messy” problems. As noted earlier, the revival of the notion of wicked problems as a focus for analysis has in earlier times dampened optimism about the capacity to design. At the same time, however, this revival has emphasized the necessity to think more creatively about how to intervene into these difficult situations (Verweij, Reference Verweij2011). Although the characteristics of wicked problems are in many instances antithetical to designing interventions (e.g., the absence of a stopping rule or even of clear definitions of the problem), the severity of these issues demands action (Newman & Head, Reference Newman and Head2017).

Many problems confronting contemporary policy designers are difficult, if not formally wicked (see Peters & Tarpley, Reference Peters and Tarpley2016), in the sense that problem causes and potential solutions may be known. However, they may also be “superwicked” (Levin et al., Reference Levin, Cashore, Bernstein and Auld2012) in the sense that not only are these unknowns present, but there may be time constraints and other limitations on gathering information and learning.

Conceptualizing the nature of the problems and understanding the difficulties they present for designing is thus even more important in these settings than it is for more “tame” policy problems. The wickedness of many contemporary policy problems will, in turn, also contribute to a movement away from more technocratic forms of intervention in favor of more process-oriented versions of design, as will be described in the following sections.

2.3.6 Emerging Design Thinking from Other Fields and Disciplines

While the earlier thinking about policy design was wedded to a rather mechanistic and technocratic form of thinking, and current work in the policy sciences remains heavily influenced by this intellectual debt and current, emerging ideas about design stemming from other design sciences may lead us down very different pathways of thinking about policy design process, and about their products. Indeed, at the most extreme the process may be the product, with building mechanisms for constant thinking and reform becoming institutionalized as a mechanism for making and implementing policies. Thus, in a significantly less technocratic perspective on the subject, the process of designing becomes the design in a process sometimes referred to as “co-design” (Lee, Reference Lee2008).

The preceding description may appear to be word play, but it is not. Rather, it is an attempt to point to the indeterminacy of some contemporary thinking about design, and the need to consider it more fully, as a remedy for hubris if nothing else, in all its aspects, from aim and goals to participants and actors. For many practitioners and scholars alike, the notion of arriving at a definite solution for a policy problem of any significance appears increasingly infeasible. Thus, where we are in policy design may reflect Ted Lowi’s notion (Reference Lowi1972) of the importance of affecting the policy environment rather than affecting (or effecting) the policy itself. That is, by attempting to shape, and to first understand, the conditions under which policies must be made may have a more enduring effect than any one intervention, and designing policies with robust processes rather than robust contents may be the order of the day (Capano & Woo, Reference Capano and Woo2017; Capano et al., Reference Capano, Howlett and Rameshforthcoming).

This less determinate perspective on design also reflects the recognition of the difficulties encountered in many policy areas with finding definitive solutions to increasingly complex problems. Following the more systemic logic, this version of designing suggests that any individual policy is itself impacted by, and impacts, other policies. Phrased differently, it is a component of a problematique (Warfield & Perino Jr., Reference Warfield and Perino1999) rather than necessarily a freestanding policy.Footnote ⁹ This embedded nature of policies in turn requires understanding better how policies are linked together and how they cohere, or fail to be integrated (Howlett et al., Reference Howlett, Vince and del Río2017). It further involves designing interventions that address the linkages and interactions among policies, even those policies not necessarily within the same domain. These interventions may need to go beyond conventional coordination (Peters, Reference Peters2017) or policy integration (Jordan & Lenschow, Reference Jordan and Lenschow2010) to address more fundamental linkages among policies and policy domains.

To date the research emphasis on policy design as a problematic and often conflict-laden process of policy formulation has focused on questions such as those exploring the tradeoffs and interactions between the various tools of governance used in policy “toolkits,” and the need to manage their inherent complementarities as well as contradictions and overlaps (Gunningham et al., Reference Gunningham, Grabosky and Sinclair1998; Howlett, Reference Howlett, Halpern, Lascoumes and Le Gales2014b; Howlett et al., Reference Howlett, Mukherjee and Woo2015). Studies have emphasized factors such as the different processes and patterns through which policy toolkits have developed over time by being layered on past design decisions (Thelen, Reference Thelen2004; Howlett & Rayner, Reference Howlett and Rayner2013) as often resulting in less “rational” mixes of policy tools than might have been originally desired or planned.

This focus on policy mixes, layering, and temporality differentiates the existing design literature from earlier approaches that examined aspects of policymaking and especially policy tool selection by concentrating on simple policy contexts and the selection of singular tools (Salamon, Reference Salamon, Salamon and Lund1989; Linder & Peters, Reference Linder and Peters1990; del Río & Howlett, Reference del Río and Howlett2013). However, this different focus, while adding an important set of new dimensions to the issues of “who designs what, when and how” (Howlett, Reference Howlett, Halpern, Lascoumes and Le Gales2014b) has also somewhat overshadowed the integration of “best practices” studies into formal policy design theory and thinking. That is, this approach to understanding and making public policy should integrate not only a solid understanding of policy processes but also a deep understanding of the characteristics of policy tools and how they operate both singly and in tandem with others conjointly.

3.1.1 The Role of Design Spaces: Defining Effective Design Environments

By the late 1990s, policy design scholars had begun to progress from the study of single instrument uses to the analysis of more complex policy mixes (Grabosky, Reference Grabosky1994; Gunningham et al., Reference Gunningham, Grabosky and Sinclair1998; Howlett, Reference Howlett2004). This era also saw a substantial shift in scholarly attention toward more “meta” level studies of policy governance, sparked by the emergence of globalization and its preference for market-based tools as well as the start of “governance” studies undertaken in Europe and elsewhere which emphasized the role of nonstate actors, especially networks, in policymaking (Howlett & Lejano, Reference Howlett and Lejano2012). This “globalization and governance turn” encouraged a polarization of discussions about effectiveness in design at a very abstract level such as that between instruments of the “market” and the “state,” or between dichotomous governance styles such as “hierarchies” and “markets” (Howlett, Reference Howlett2004; Koch, Reference Koch2013).

This often led to fairly simple design precepts about what constitutes effective policymaking being put forward by students of globalization and governance, such as an absolute preference for market-based tools or collaborative governance arrangements at all times (Hay & Smith, Reference Hay and Smith2010; Jarvis, Reference Jarvis2011; Tollefson et al., Reference Tollefson, Zito and Gale2012). These studies were undoubtedly biased but still useful in helping to link policy design thinking and studies to work on policy styles and other such “meta” governance activities. However, a wider vocabulary of tools and instruments is needed if effective programs and choices are to be elaborated and adopted (Hood, Reference Hood2007; Hood & Margetts, Reference Hood and Margetts2007).

Importantly, though, this orientation also raised questions about the environment of “spaces” in which formulation processes occur both at this “meta” and at the more specific level of policy tools to generate outcomes or designs, and whether or not certain kinds of spaces were more likely to produce better results than others.

The idea, of course, is that the nature of the overall policy design space can have a significant bearing on how effectively intended design activities take place and thus upon the likely effectiveness of policy designs which emerge from them. As early as 1991, Linder and Peters had suggested that policy design could be thought of as an area of study oriented toward the understanding of such spaces. That is, they noted that design was a systematic activity composed of a series of choices or design solutions which correspond to a set of possible locations in a design space (Linder & Peters, Reference Linder and Peters1991).

Since then, a major area with which contemporary policy researchers have been concerned is better understanding and demarking the nature of more effective design spaces – that is, those which allow deliberations and debates to occur which lead to superior designs rather than poorer or nonexistent ones (Howlett, Reference Howlett2011).

These spaces are delimited and characterized by on-the-ground political realities which shape overall public and elite preferences for certain kinds of tools and mixes over others, such as market-based portfolios over state-based ones. Policy design processes, for example, are embedded within, and their effectiveness is delimited by prevailing modes and styles of governance. As an example, Ansell and Gash (Reference Ansell and Gash2008, p. 543) underscored the significance of several “spatial” variables that are critical for the effectiveness of collaborative governance, including reconciling with “prior history of conflict or cooperation, the incentives for stakeholders to participate, power and resource imbalances, leadership and institutional design.”

Thinking about design effectiveness, in this sense, begins with an understanding of how the policy design space influences what kinds of policies are envisioned and how such envisioning takes place. Table 1 presents a schematic illustrating a contemporary understanding of such spaces, which are affected by the presence of significant policy legacies and unique political conditions that can affect whether or not policy changes follow a design-oriented pattern of analysis and deliberation or some other form, and whether design is likely to occur by whole measures (“packaging”) or in parts (“patching”) (Howlett & Rayner, Reference Howlett and Rayner2013; Howlett & Mukherjee, Reference Howlett and Mukherjee2014).

As Table 1 shows, in any specific design circumstance whether or not “design” takes place at all can be seen to depend on the political aims and intention of government to undertake systemic thinking on a subject (Anderson, Reference Anderson1975). But even having an intention to be formal and analytical in designing and evaluating policy alternatives is not enough in itself to promote a design-centered process, since this also depends on the government’s ability or capacity to undertake such an analysis and to alter the status quo. In many circumstances, even when an intention for design is present, capacity difficulties associated with a design situation result in the creation of alternatives which “patch” rather than “package” tools together in coherent and consistent ways.

Determining exactly what capacities are required in order to develop the design spaces needed to carry out complex design processes is thus a subject of much interest in the field today (Considine, Reference Considine2012). In order to address these issues, policymakers and formulators need to be cognizant about the internal mechanisms of their polity and constituent policy sectors which can boost or undermine their intention to systematically develop effective policies or simply, for example, adopt expedient or politically marketable ones (Braathen & Croci, Reference Braathen, Croci and Croci2005; Braathen, Reference Braathen2007; Grant, Reference Grant2010; Skodvin et al., Reference Skodvin, Gullberg and Aakre2010).

While work on intentionality remains rudimentary, however, recent work on policy capacity has outlined the fundamental nature of the skills and resources governments need to effectively formulate and implement policy (Bullock et al., Reference Bullock, Mountford and Stanley2001; Wu et al., Reference Wu, Ramesh, Howlett and Fritzen2010; Rotberg, Reference Rotberg2014; Howlett & Ramesh, Reference Howlett and Ramesh2015, Reference Howlett and Ramesh2016). These exist at three levels: individual, organizational, and systemic (Wu et al., Reference Wu, Ramesh and Howlett2015).

Individually, those striving for effective design need to possess technical expertise for substantive policy analysis and communication of knowledge, while necessary skills of those in management roles also include leadership and negotiation expertise. Individual political acumen for understanding the interests of various stakeholders and gauging political feasibility is also a fundamental capacity for successful governance. At the organizational level, information mobilization capacities to facilitate policy analysis, administrative resources for successful coordination between policymaking agencies, and political support all contribute toward overall policy capacity. At the system level, institutions and opportunities for knowledge creation and use need to exist alongside arrangements for accountability and securing political legitimacy.

In general, governments need high levels of capability and competence in all aspects of capacity in order to effectively design policies. While shortcomings in one or a few of the dimensions may be offset by strengths along other dimensions, no government can expect to be (fully) capable if it is lacking capacity across many dimensions (Tiernan & Wanna, Reference Tiernan and Wanna2006).

Shortfalls in specific kinds of capacity are especially critical in specific modes of governance and constitute their Achilles’ heel (Menahem & Stein, Reference Menahem and Stein2013; Howlett & Ramesh Reference Howlett and Ramesh2016) when it comes to effectively creating policy solutions for different policy problems. Hence, in recent years in many jurisdictions the default reform often adopted in practice by governments seeking to improve upon hierarchical governance has been to turn to a market or network mode of governance (Weimer & Vining, Reference Weimer and Vining2011). But in order to function effectively, markets require tough but sensible regulations that are diligently implemented. Having the technical knowledge to craft such regulations is thus a critical competence required for effective market-based governance. Analytical skills at the level of individual analysts and policy workers are key in this area, and the “policy analytical capacity” of government needs to be especially high in order to deal with complex quantitative economic and financial issues involved in regulating and steering the sector and preventing crises (Rayner et al., Reference Rayner, McNutt and Wellstead2013).

Similarly design within legal systems of governance requires a high level of managerial skills in order to avoid diminishing returns with compliance or growing noncompliance with government rules and regulations (May, Reference May2005). And system level capabilities are especially crucial in this mode of governance because governments will find it difficult to command and control in the absence of the target population’s trust. And while network governance may perform well when dealing with design for sensitive issues such as parental supervision or elderly care (Pestoff et al., Reference Pestoff, Brandsen and Verschuere2012), in other instances civil society and social capital may not be constructed or resourced sufficiently to be able to create beneficial network forms of governance (Tunzelmann, Reference Tunzelmann2010). Networks, for example, will fail when governments encounter at the organizational level capability problems, such as a lack of societal leadership, poor associational structures, and weak state steering capacities which make adoption of network governance modes problematic (Keast et al., Reference Keast, Mandell and Brown2006; Klijn & Koppenjan, Reference Klijn and Koppenjan2012).

3.1.2 Effective Instrument Mixes

A greater emphasis on tool mixes and on the processes that effectively create complex policy tool bundles has been a second feature of research into policy effectiveness over the last decade (Hood, Reference Hood2007; Howlett, Reference Howlett2011). These studies have increased awareness of the many dilemmas that can appear in the path of effective policy tool or “toolkit” designs and realities (Peters & Pierre, Reference Peters and Pierre1998; Doremus, Reference Doremus2003; Sterner, Reference Sterner2003; Klijn & Koppenjan, Reference Klijn and Koppenjan2012).

Such mixes are combinations of policy instruments that are expected to achieve particular policy objectives and which are generally seen as more efficient and effective than single instrument uses (Gunningham et al., Reference Gunningham, Grabosky and Sinclair1998; Rogge & Reichardt, Reference Rogge and Reichardt2016). Some instruments may work well with others by nature – as is the case with “self-regulation” set within regulatory compliance frameworks (Trebilcock et al., Reference Trebilcock, Tuohy and Wolfson1979; Grabosky, Reference Grabosky1994; Gibson, Reference Gibson1999) – while other combinations may not, such as, notably, independently developed subsidies and regulations. Some mixes may have evolved in a certain way which has undermined, or improved, their effectiveness, while others may have never been very effective in the first place and remain ineffective.

The new design orientation has engaged in a lengthy discussion as to how to effectively integrate policy mixes so that multiple instruments are arranged together in complex portfolios of policy goals and means (Gunningham et al., Reference Gunningham, Grabosky and Sinclair1998; Doremus, Reference Doremus2003; Briassoulis, Reference Briassoulis2005; Howlett, Reference Howlett2011; Yi & Feiock, Reference Yi and Feiock2012; Peters et al., Reference Peters, Eliadis, Hill, Howlett, Eliadis, Hill and Howlett2005; Jordan et al., Reference Jordan, Benson, Wurzel, Zito, Dryzek, Norgaard and Schlosberg2011, Reference Jordan, Benson, Wurzel, Zito and Richardson2012), often with a multilevel governance component (del Rio & Howlett, Reference del Río and Howlett2013). Effectively optimizing the choice of instruments in such mixes requires an additional level of knowledge of instrument-goal interactions and considerations of how mixes evolve over the long run, as well as an understanding of both long- and short-term processes of policy change. For example, a major concern of those working in the new orientation of policy design studies is whether combinations of different policy instruments, which have evolved independently and incrementally, can accomplish complex policy goals as effectively as more deliberately customized portfolios (Howlett, Reference Howlett, Halpern, Lascoumes and Le Gales2014b).

The components of such mixes include policy goals and policy means at various levels of generality (Cashore & Howlett, Reference Cashore and Howlett2007; Howlett, Reference Howlett2009; Kern & Howlett, Reference Kern and Howlett2009), and design and instrument selection in these contexts “are all about constrained efforts to match goals and expectations both within and across categories of policy elements” (Howlett, Reference Howlett2009, p. 74). The various elements of policies which are combined in the design process include those related to general goals and means, those linked to tools, and those linked to the settings or calibrations of those tools (see Table 2).

Achieving effectiveness with respect to deploying such policy portfolios relies upon ensuring mechanisms, calibrations, objectives, and settings display “coherence,” “consistency,” and “congruence” with each other (Howlett & Rayner, Reference Howlett and Rayner2013). Works on “smart regulation” by Neil Gunningham, Peter Grabosky, and Darren Sinclair (Reference Gunningham, Grabosky and Sinclair1998), for example, have led scholars to focus on how instruments within a policy mix or “portfolio” could effectively complement each other or conversely, lead to conflicts, resulting in guidelines for the formulation of more sophisticated policy designs in which complementarities were maximized and conflicts were avoided (Blonz et al., Reference Blonz, Vajjhala and Safirova2008; Barnett & Shore, Reference Barnett and Shore2009; Buckman & Diesendorf, Reference Buckman and Diesendorf2010; Roch et al., Reference Roch, Pitts and Navarro2010; del Rio et al., Reference del Río, Silvosa and Gómez2011).

Concerns regarding how to make the most of policy synergies while curtailing contradictions in the formulation of new policy packages are a major topic of investigation within the new design orientation (Hou & Brewer, Reference Hou and Brewer2010; Kiss et al., Reference Kiss, Manchón and Neij2013; Lecuyer & Quirion, Reference Lecuyer and Quirion2013). Evidence from the world concerning renewable energy and energy efficiency policy, due to climate change mitigation and energy security concerns, has revealed, for example, that policy packages combining voluntary compliance with command-and-control regulation can be inherently inconsistent, bringing out contradictory responses from targets of these policy combinations (Boonekamp, Reference Boonekamp2006; del Rio et al., Reference del Río, Silvosa and Gómez2011).

Scholars of the new orientation who are concerned with design effectiveness have also been interested in how “unintended” policy mixes, created and limited by historical legacies, can be hampered due to internal inconsistencies, whereas other policy instrument groupings can be more successful in creating an internally supportive combination (Grabosky, Reference Grabosky1994; Gunningham et al., Reference Gunningham, Grabosky and Sinclair1998; Howlett & Rayner, Reference Howlett and Rayner2007; del Rio, Reference del Río2010).

Studying complex design processes is thus concerned not only with gaining a better understanding of design “spaces” but also the temporal processes through which these spaces evolve. While some studies suggest that effective design could only occur in spaces where policy packages could be designed “en bloc” and “de novo,” as already described, more recent research recognizes that most design circumstances involve building upon the foundations created in another era, often by others, and working with suboptimal design spaces. New policy design scholars are thus interested in processes such as how policy formulators, like software designers, can issue “patches” in order to correct flaws in existing mixes or allow them to adapt to changing circumstances and become more effective (Howlett, Reference Howlett2013; Howlett & Rayner, Reference Howlett and Rayner2013; Rayner et al., Reference Rayner, McNutt and Wellstead2013).Footnote ¹⁰

In general, policy formulation is seen to take place while contained within present governance structures and the existing policy logic – restricting the number of alternatives that can be deemed feasible in such a context and decreasing the universe of policy alternatives to a smaller set of workable possibilities (Christensen et al., Reference Christensen, Laegreid and Wise2002; Meuleman, Reference Meuleman2009a, Reference Meuleman2009b). The “elbow room” or “degrees of freedom” designers have to maneuver in given policy design contexts is thus a subject of much interest. Many studies draw on the work of historical and sociological institutionalists such as Kathleen Thelen and colleagues (Reference Thelen, Mahoney, Rueschemeyer, Mahoney and Rueschemeyer2003, 2004), who noted how macro-institutional arrangements have normally been less the product of calculated planning and more the result of processes of incremental modifications or reformulations such as “layering” or “drift” (Beland, Reference Beland2007).

Contemporary research also asks questions regarding how some policy mixes may be comprised of redundant elements while others, despite containing repetitive elements, may in fact promote resiliency and adaptability (Braathen & Croci, Reference Braathen, Croci and Croci2005; Braathen, Reference Braathen2007; Swanson et al., Reference Swanson, Barg, Tyler, Venema, Tomar, Bhadwal, Nair, Roy and Drexhage2010; Walker et al., Reference Walker, Marchau and Swanson2010). In their discussion of policy element duplication, Yilin Hou and Gene Brewer (Reference Hou and Brewer2010) have noted that the real issue is not to simply eliminate all duplication on a priori grounds but rather to design policy toolkits containing tools that work together or complement each other while being effective, given stated goals and policy contexts (Swanson et al., Reference Swanson, Barg, Tyler, Venema, Tomar, Bhadwal, Nair, Roy and Drexhage2010; Hoffman, Reference Hoffman2011).

3.1.3 The Effectiveness of Specific Policy Tools

While the discussion of effective design environments and instrument mixes have developed in parallel over the last two decades, the literature consolidating lessons regarding the specifics of effective individual tool design has also been updated to become the third leg in this stool.

Classic works on policy instrument choice, such as those by Lester Salamon, Christopher Hood, Guy Peters, and others, laid the initial groundwork toward this goal in the early 1980s. At that time scholars and practitioners focused on more precisely categorizing policy instruments and better analyzing the reasons for their use (Salamon, Reference Salamon1981). Careful examination of instruments and instrument choices was expected to allow practitioners to more readily draw lessons from the experiences of others with the use of particular techniques in specific circumstances, leading to more effective designs (Linder & Peters, Reference Linder and Peters1984; Woodside, Reference Woodside1986; Bobrow & Dryzek, Reference Bobrow and Dryzek1987; Dryzek & Ripley, Reference Dryzek and Ripley1988).

The key information instrument studies needed to generate in order to help facilitate better or more effective policy designs was defined from these questions (Salamon, Reference Salamon1981; Timmermans et al., Reference Timmermans, Rothmayr, Serduelt and Varone1998; Hood, Reference Hood2007):

• What tools does a government have?

• How can these be classified?

• How have these been chosen in the past?

• Is there a pattern for this use?

• How can we explain these patterns?

• How can we improve on past patterns of use?

“Substantive” policy instruments are expected to alter some aspect of the production, distribution, and delivery of goods and services in society: broadly conceived to include both mundane goods and services such as school lunches to crude vices such as gambling or illicit drug use, to more common individual virtues such as charitable giving or volunteer work with the physically challenged, and include the attainment of collective goals like peace and security, sustainability, happiness, and well-being. They are those policy techniques or mechanisms designed to directly or indirectly affect the behavior of those involved in the production, consumption, and distribution of different kinds of goods and services in society.

This is a large field of action since it extends not only to goods and services provided or affected by markets, but also well beyond to state or public provision and regulation, as well as to those goods and services typically provided by the family, community, nonprofit, and voluntary means often with neither a firm market nor state basis (Salamon Reference Salamon, Salamon and Lund1989; Reference Salamon2002).

Substantive implementation instruments can affect many aspects of production, distribution, and consumption of goods and services regardless of their institutional basis. Production effects, for example, include determining or influencing:

1. 1 Who produces it – for example, via licensing, bureaucracy/procurement, or subsidies for new startups.

2. 2 The types of goods and services produced – for example, through bans or limits or encouragement.

3. 3 The quantity of goods or services provided – for example, via subsidies or quotas.

4. 4 The quality of goods or services produced – for example, via product standards, warranties.

5. 5 Methods of production – for example, via environmental standards or subsidies for modernization.

6. 6 Conditions of production – for example, via health and safety standards, employment standards acts, minimum wage laws, inspections.

7. 7 The organization of production – for example, via unionization rules, antitrust or anticombines legislation, securities legislation, or tax laws.

Consumption and distribution effects are also manifold. Some examples of these are:

1. 1 Prices of goods and services – such as regulated taxi fares or wartime rationing.

2. 2 Actual distribution of produced goods and services – affecting the location and types of schools or hospitals, forest tenures, or leases.

3. 3 Level of consumer demand for specific goods – for example, through information release, nutritional and dangerous goods labeling (cigarettes), export and import taxes and bans, and similar activities.

4. 4 Level of consumer demand in general – via interest rate, monetary and fiscal policy.

“Procedurally” oriented implementation tools, on the other hand, affect production, consumption, and distribution processes only indirectly, if at all. Instead they affect the behavior of actors involved in policy implementation. Policy actors are arrayed in various kinds of policy communities, and just as they can alter or affect the actions of citizens in the productive realm, so too can they affect and alter aspects of policymaking behavior. Procedural implementation tools are an important part of government activities aimed at altering policy interaction within policy subsystems but, as Erik-Hans Klijn et al. (Reference Klijn, Koppenjan and Termeer1995, p. 441) put it, they “structure … the game without determining its outcome.” That is, they affect the manner in which implementation unfolds but without predetermining the results of substantive implementation activities.

Some of the kinds of implementation-related activities that can be affected by the use of procedural tools (Klijn et al., Reference Klijn, Koppenjan and Termeer1995; Goldsmith & Eggers, Reference Goldsmith and Eggers2004; Klijn & Koppenjan, Reference Klijn and Koppenjan2006) include:

1. 1 Changing actor policy positions

2. 2 Setting down, defining, or refining actor positions

3. 3 Adding actors to policy networks

4. 4 Changing access rules for actors to governments and networks

5. 5 Influencing network formation

6. 6 Promoting network self-regulation

7. 7 Modifying system-level policy parameters (e.g., levels of market reliance)

8. 8 Changing evaluative criteria for assessing policy outcomes, success, and failure

9. 9 Influencing the pay-off structure for policy actors

10. 10 Influencing professional and other codes of conduct affecting policy actor behavior

11. 11 Regulating inter-actor policy conflict

12. 12 Changing policy actors’ interaction procedures

13. 13 Certifying or sanctioning certain types of policy-relevant behavior

14. 14 Changing supervisory relations between actors.

Procedural implementation tools and their effects are not as well studied or understood as are substantive instruments, although several procedural techniques, such as the use of specialized investigatory commissions and government reorganizations, are quite old and well used and have been the objects of study in fields such as public administration, public management, and organizational behavior (Schneider & Sidney Reference Schneider and Sidney2009; Lang, Reference Lang2016). Nevertheless, just like their substantive counterparts, they are a key part of policy designs and policy design activity, and deriving lessons and best practices about their deployment and use remains a key element of contemporary policy design studies (Hood, Reference Hood1983; Fung, Reference Fung2003; Redstrom, Reference Redström, Wijnand, Jsselsteijn, de Kort, Midden, Eggen and van den Hoven2006; Weaver, Reference Weaver2009a).

4.2.1 Three Clarifications on Effective Policy Design

The first clarification relates to the status of the “effective design” question itself. As previously discussed, the notion of “effective policy design” is and has been omnipresent in the design and public management literatures – even when not explicitly stated. That is, as the tale of design thinking presented at the outset has shown, the goal of design is effectiveness. This is why policymakers, implicitly or explicitly, engage in design in the first instance. As one of several analytical frameworks within policy sciences, the design orientation shares its foundational assumption that governments wish to have their goals effectively achieved in an efficient way.

As also previously discussed, however, it is not always clear to what “effectiveness” exactly refers. Is “effectiveness of policy design” a research question? Thus, does “effective policy design” become a research object that is constructed and analyzed (and eventually contested) by researchers? Or is “effective policy design” an objective for design scholars and a constitutive feature of the policy design approach’s ontology? That is, do policy design scholars and practitioners explicitly seek to define the conditions for effective policy design?

These two perspectives are interweaved in the policy design approach. As Michael Howlett and Ishani Mukherjee (Reference Howlett and Mukherjee2014, p. 62) put it, “The fervent wish of proponents of design orientation is generally to reduce instances of poor and non-design to as few as possible … This, is expected to result in policies more likely to resolve pressing problems, correct social ills and better serve the public good.”

It is thus necessary for the analysis of the international level to disentangle questions concerning design, effectiveness, and transnational public policy that have different ontological starting points. For instance, what does effectiveness mean in a policy world transformed by globalization? How can we define or identify effective design of transnational policies? Explicitly distinguishing these issues does not eschew the possibility that effectiveness may be the ultimate outcome for design scholars and practitioners interested in designing. Indeed, research findings can be policy relevant, but it is important to think about the existence of more than one approach in the search for effectiveness. Considerations of effectiveness can actually nourish a more policy-oriented debate on global policies, not by directly recommending how to design effective global public policies, but by providing insights on the meaning of effectiveness that might be of interest to policymakers.

A second clarification essential to pushing this perspective concerns the very understanding of the words “effective” and “effectiveness.” As already reviewed at the outset, an established literature in public policy and public management deals with these notions directly and indirectly. Evaluation of (effectiveness of) public policy is a significant topic of interest in public policy (Bovens et al., Reference Bovens, ’t Hart, Kuipers, Goodin, Moran and Rein2008), and has also become a very important field of study, with its own journals (American Journal of Evaluation; Evaluation: The International Journal of Theory, Research and Practice; Evaluation: Theory and Praxis; New Directions for Evaluation; etc.), scholarly networks, and methodological debates about how best to assess the effectiveness of policies. The perspective presented in this Element is different: the notion of effectiveness is unpacked to analyze what is at stake when actors are dealing with “effectiveness” when designing policy arrangements at the transnational level, and to show what it reveals about the transformations of policymaking in a globalizing world. In other words, this perspective is firmly anchored in a political science qualitative perspective, distinct from a policy evaluation viewpoint or a public management approach (see O’Toole, Reference O’Toole, Kim, Ashley and Lambright2014, on theory of context and how globalization context may affect the management-performance linkage).

In the common language, there are two definitions of “effective” and “effectiveness.” First, “effective” refers to successfully producing a desired or intended result; effectiveness is thus about success (outcome) and the ability to be successful (means). Evaluation approaches generally understand “effective” and “effectiveness” in this way, and therefore work at defining relevant templates and indicators to assess the production (and how much) of the intended results. But a second definition of “effective” is less about success and intended results than about the actual effects. “Effective” hence means operative, existing in fact, or producing an effect. Effectiveness in this second perspective would be about “operativeness,” or acknowledgment of the production of an effect.

In policy design studies, it can be observed that the understanding of effectiveness generally corresponds to the first one and that there is a tendency to neglect or elude the second perspective. In this way, focusing on the overall search for the effective design and specific “intended results,” the risk may be to miss all the other effects the policy instruments may trigger. The initial or observable effects may ultimately lead to the very effectiveness the designers seek, albeit through other combinations of effect or with the lapse of time (e.g., some instruments require a longer duration to take effect). Some effects may not be anticipated (this corresponds to the well-known category of unintended effects), but one can also imagine a specific type of policy design that leaves room for open effects. To understand the development of global policy arrangements, it is particularly important to adopt an effectiveness questioning which encompasses both meanings: effectiveness as success in producing intended results and effectiveness as “operativeness” and production of effects that are to be characterized.

A third clarification required in this discussion about design, effectiveness, and the transnational dimension deals with the “effectiveness of what” question. In policy design literature, and because of the approach’s optimism vis-à-vis the improvement of public policy, there tends to be an overlap between the ideas of an effective policy design on the one hand and an effective policy on the other.

Broadly, the core idea is that when a design is “smart,” governments then select the right implementation tools, consisting of substantive and procedural tools, for a rightly identified policy problem. Within this formulation, largely based on a rational choice perspective of actor behavior, any outcomes suggesting failure would simply mean that the wrong implementation tools were chosen or the “preconditions” for design were absent or unfavorable (Howlett & Mukherjee, Reference Howlett and Mukherjee2014). Once the right tools are picked when the favorable conditions are present, the assumption goes, the anticipated policy results should be observed. Hence in this approach effective policy design should lead to effective policy, and conceptually it is less important to work on the distinction between effective design and effective policy, which tends to belong to a single category.

However, there is no reason to assume that there is automaticity between effective design and effective policy. Policy design scholars’ argument on preconditions generally excludes the contemplation of different combination types of effectiveness between design and policy. In a world in which not all preconditions can always be foreseen, one can imagine a situation in which effective design leads to ineffective policy. While one may attribute this to bounded rationality, it can also be that some implementers simply refuse to execute the agreed upon and adopted policy for reasons that may not be political or rational.

A situation in which ineffective design ends up with an effective policy is equally conceivable. For example, in instances in which design is poor, policy actors may still achieve the policy goal through creative interpretation, or abandonment, of the policy means. While these outcomes may be less of a concern at the domestic level, where actor participation is more easily arranged and monitored, they may be more prevalent and important at the international level. So, for the study of global public policy it is important to consider that questions concerning “effectiveness of design” and “effectiveness of policy” may be more autonomous than usually assumed by the design literature. Table 3 summarizes these three clarifications, and the next section elaborates why policy effectiveness may take on a different meaning when examining transnational initiatives.

4.2.2 The Reasons Why Effectiveness Means Something Different at the Global Level

Although the more comprehensive understanding of effective design proposed here is not meant to be particular to global policy developments, there are some specifics to the design of global policies that ensure a revisited questioning on effectiveness is particularly useful.

The first thing that is different about the design of global policies, which then implies a different understanding of their effectiveness, has to do with the great variety of actors involved. Applying the policy design approach to studies of global public policy and transnational administration requires a renewed understanding of the types of actors who may be involved in policymaking (Chou & Ravinet, 2018). In the design literature, and even with the theoretical efforts of the new design orientation to develop a more open perspective on the design space (see the preceding section), the state is often seen as the key policy actor and is operationalized as the government (central or regional) responsible for decision-making across all stages of the policy cycle, sequential or otherwise. Taking this research design to the transnational level would imply that attention is focused primarily on analyzing the behavior of governments, which offers only a limited view. As the governance approach reminds us, interactive and collaborative forms of governance require the bringing together of public leaders with societal actors, as well as individuals from the business or scientific communities. Hence, while it is possible to keep the policy design core principle to concentrate on the notion of intentionality and view design as an act of purposive actors, studying global initiatives requires broadening the range of participant actors. In the case of global and transnational policymaking, the purposive actors involved in design are generally not predetermined, they may wear multiple hats, and they may follow a sectoral organizational logic that is extremely relevant and significant.

Indeed, as Diane Stone and Stella Ladi (Reference Stone and Ladi2015, p. 844) put it, “The constitutive actors of global processes and transnational administration may be different from the domestic level, or they may have demonstrated increased power beyond the state” – and thus their identification and relationship (if any) with the state should be elaborated through empirical research, which can only occur when the analytical perspective does not preclude this exploration. For instance, there are multiple design actors at the transnational level including, but are not limited to (1) formal or official actors (they include governments and international governmental organizations) and (2) nonstate actors and networks. International and regional organizations are involved in a variety of design activities, including, for instance, developing new programs for internal and external coordination (e.g., standards and benchmarks), for (client) countries, and for global or regional public-private partnerships and networks. Added to this grouping are the international consultancy firms (often contracted by governments and international/regional organizations to deliver design on procedural processes and evaluate implementation), think tanks, global commissions and taskforces, large international nongovernmental organizations (designers of indexes or rankings), business associations, and accreditation bodies. Although this list is not exhaustive, it does point to the many design actors at the transnational level who work with governments or on their behalf.

This has important implications for the understanding of effectiveness. In the design of global policies, not only might the number of actors involved be greater, but more fundamentally, the type of actors might be extremely diverse. This means that the possibility of convergence on the meaning of effectiveness may be much more challenging than could be feasibly conceived. Meanings are actually intersubjective, and the more diverse the actors involved, the more intersubjectivity is involved. Agreements upon ambivalent general objectives (sustainable development, for instance) or upon technical small objectives (that might be equally acceptable for different or even contradictory world visions), rather than on defined policy objective, are extremely likely in the design of global policies. In the case of ambivalence, what happens is that actors mask their real objectives and understanding; but when it comes to moments of evaluating effectiveness, ambivalence might not be manageable. In the case of small technical objectives, this might be problematic as well in terms of defining effectiveness: evaluating effectiveness is not only about looking at whether scattered small objectives were reached in isolation from one another, but also about how the combination of different indicators allows the evaluators to measure the overall policy outcome.

A second feature that is specific in designing global public policies deals with policy instruments. A design orientation to studying transnational developments should also adjust the potential range and types of policy instruments that actors may have at their disposal. The extant policy design literature depicts instrument selection (picking them up from a toolbox) as one of the core activities of the design process. This is a problematic formulation for transnational developments or initiatives in at least two ways. First, it assumes that instrument selection is a neutral act and reveals nothing about the relationship between the governing and the governed (Lascoumes & Le Gales, Reference Lascoumes and Le Gales2007). It is common knowledge in several strands of political science literature that governments use transnational policy arenas for domestic purposes, including tying their hands to supranational reforms they wished to introduce domestically but which are opposed by veto players (Guiraudon, Reference Guiraudon2000).

These supranational processes are not purely bargaining as they also involve deliberation of policy problems and matching potential solutions with outcomes. Any perception of instrument failure following these processes is thus less straightforward, and lessons to be learned are more challenging to extract. Design scholars acknowledge this possibility, but they often relegate such developments with references to complexity without adequately engaging in the unpacking of what complexity means for the design orientation. As noted already, the politics of making decisions about complex or wicked problems have yet to be fully explored.

Second, the existence of an instrument toolbox from which to “pick and choose” may be unavailable at the transnational level. This may be the result of some involved policy actors lacking the authority to invoke access to certain instruments, or it may be that transnational policy instrumentation is a process more about invention and innovation than about selecting among available tools. While more empirical observations are certainly needed to refine the analytical construct of the design orientation, existing findings already point to the general inaccuracy of categorizing these instances as preconditions for or against design.

Combined with the first specific feature of global public policies (diversity of actors involved in design), this has important implications for defining effectiveness. As noted earlier, actors might agree or disagree on the objectives, and as far as tools are concerned they might be available or unavailable. Four distinct scenarios can thus be envisioned (see Table 4). First, in a scenario in which the actors agree on the general objectives, the relevant tools available and actors entitled to use them, classical policy design assumptions may be applied; this scenario, however, is far less likely to emerge consistently at the transnational level for reasons mentioned earlier. Second, when actors disagree on the general objective in a scenario in which instruments are available, they would tend to agree on small technical instruments; this is a common scenario observed in transnational policy coordination. The third scenario sees actors disagreeing over general objectives in which there are no instruments readily available; this is another frequent scenario at the global level, where policy cooperation results in no concrete steps forward. Finally, the fourth scenario reveals actors agreeing on the general objective, but the tools do not exist (i.e., actors need to innovate) or the actors lack the authority to use existing tools.

The third specificity in the design of global public policies that has implications for the definition of effectiveness deals with the “line of command.” Any policy design incorporates an implementation template of not only which instruments will be used, but also of the different policy and administrative bodies that will be involved at different policy levels within this process. This is what can be designated as the notion of the “line of command.” What is specific in the design of global public policies is actually that the number of layers, and therefore intermediary policy bodies or delegated actors involved between the global objective and the final implementation on the policy ground, might be many. Indeed, the “line of command” between designing and implementing global policies is longer, more complex, and potentially filled with ambiguity in comparison to those associated with domestic policies where the “line of command” is comparatively shorter, clearer, and deeply institutionalized with taken-for-granted assumptions concerning actor behavior.

This specificity of global public policies from the vantage point of the “line of command” has implications for the definition of effectiveness. The universe of final implementation is likely to be very far from the one of the initial design (in space and time), and the initial design thinking may thus be very diffused or nonexistent among actors tasked with policy implementation. This results in a different understanding of the “effectiveness of what” question that has already been mentioned. Given the distance between policy design and final implementation, actors tasked with evaluation are far more likely to privilege an understanding of effectiveness to the design rather than what actually emerges (policy outcomes) on the ground. When final implementation is far and uncertain, appreciating design effectiveness is obviously a more conceivable, feasible, and satisfying approach than appreciating the final effects of implementing the policy. Table 5 summarizes the specific features of global policies from the perspectives of the types of actors, instruments, and line of command, as well as their implications for defining (and ultimately assessing) effectiveness at the transnational level.