Policy Design and Non-Design: Towards a Spectrum of Policy Formulation Types

Public policies are the result of efforts made by governments to alter aspects of behaviour -- both that of their own agents and of society at large -- in order to carry out some end or purpose. They are comprised of complex arrangements of policy goals and policy means matched through some decision-making process. These policy-making efforts can be more or less systematic in attempting to match ends and means in a logical fashion or can result from much less systematic or rational processes. "Policy design" implies a knowledge-based process in which the choice of means or mechanisms through which policy goals are given effect follows a logical process of inference from known or learned relationships between means and outcomes. This includes both 'good design' in which means are selected in accordance with experience and knowledge and 'bad' or poor design in which principles and relationships are incorrectly or only partially articulated or understood. In other circumstances, however, policy decisions are more highly contingent and driven by situational logics, bargaining or opportunism than result from careful deliberation and assessment. To distinguish these from poor design, these results can be thought of as "non-designs". This paper considers the question of both design and non-design modes and formulates a spectrum of policy formulation types which helps clarify the nature of each type and the likelihood of each type of policy process unfolding.

result from tool interactions, as well as the different means and patternssuch as layering -through which policy mixes evolve over time (Thelen 2004). This 'toolbox' and temporal emphasis distinguishes this new approach from earlier efforts towards understanding policy formulation which tended to focus on single tool choices in simple (level, time and space delimited) policy contexts (del Rio and Howlett 2013).
This orientation, however, raises the issue of the difference between design and non-design processes and the frequency or likelihood of occurrence of each. That is, not all policy-making is logic or knowledge driven and it is debatable how closely policymakers approximate the instrumental logic and reasoning which characterizes a design situation . Many formulation situations, for example, involve information and knowledge limits ("poor design") or involve multiple actors whose relationships may be more adversarial or competitive than is typically associated with a 'design' process and outcome ("non-design") (Schon 1988;Gero 1990). This paper addresses the differences between good and poor design and 'nondesign' policy-making processes and the likelihood of each occurring. By engaging in a discussion of intentionality in policy designswhether towards public interest or more politically driven opportunismand of the capacity of governments to undertake design efforts, the paper develops a typology of the several formulation processes that exist between pure design and more contingent non-design ones.

What is Policy Design?
Policy design involves the deliberate and conscious attempt to define policy goals and connect them in an instrumental fashion to instruments or tools expected to realize those objectives (Majone 1975;May 2003;Gilabert and Lawford-Smith 2012). Policy design, in this sense, is a specific form of policy formulation based on the gathering of knowledge of the effects of policy tool use on policy targets and its application to the development and implementation of policies aimed at the attainment of specific desired policy outcomes and ambitions (Weaver 2009 andBobrow and Dryzek 1987;Bobrow 2006;Montpetit 2003).
Within the policy sciences, 'design' has been linked both to studies of policy instruments and implementation (May 2003) and of the impact of policy ideas and advice on policy formulation (Linder and Peters 1990). In this sense, policy designs can be seen to contain both a substantive componenta set of alternative arrangements thought potentially capable of resolving or addressing some aspect of a policy problem, one or more of which is ultimately put into practiceas well as a procedural componenta set of activities related to securing some level of agreement among those charged with formulating, deciding upon, and administering that alternative on its relative merits vis-àvis other alternatives (Howlett 2011). Design thus overlaps and straddles both policy formulation and policy implementation and involves actors, ideas and interests present at both these stages of the policy process (Howlett, Ramesh and Perl 2009). However it also posits a very specific form of interaction among these elements, driven by knowledge and evidence of alternatives' merits and demerits in achieving policy goals.
Conceptually, a policy design process begins with the analysis of the abilities of different kinds of policy tools to affect policy outputs and outcomes and the kinds of resources required to allow them to operate as intended Hood 1986).
This instrumental knowledge is contextual in the sense that it requires a special understandings of how the use of specific kinds of instruments affects target group behaviour and compliance with government aims. It thus includes knowledge and consideration of many constraints on tool use originating in the limits of existing knowledge, prevailing governance structures, and other arrangements which may preclude certain options and promote others (Howlett 2009 and. It also requires government analytical and evidentiary capacity as well as the intention to exercise it. Such a means-ends understanding permeates the policy design orientation to policy-making but, of course, is only one possible orientation or set of practices which can be followed in policy formulation and result in policy-outputs (Tribe 1972;Colebatch 1998). In the design case policy formulators are expected to base their analyses on logic, knowledge and experience rather than, for example, purely political calculations or bargaining or other forms of satisficing behaviour which characterize other forms of alternative generation (Sidney 2007;Bendor et al 2009). These other forms can be termed 'non-design' ones.
Policy design studies, of course, acknowledge that not all policy work is rational in this instrumental sense and often deals with the vagaries of policy formulation processes by separating out two dimensions of the design experience: on the one hand the exploration of the procedural aspects of designthe specific types of policy formulation activities which lead to design rather than some other form of policy generation -and on the other the substantive componentsthat is, the substance or content of the design itself in terms of the instruments and instrument setting of which it is composed. This is the policy-relevant articulation of the well known distinction in design studies generally between 'design-as-verb' ('policy formulation') and 'design-as-noun' (policy tools and instruments). 1 The idea is that even when policy processes are more contingent, the design of a policy, conceptually at least, can be divorced from the processes involved in its actual creation. Thus regardless of the nature of actual alternative formulation in a specific context, it is still possible to imagine a more instrumental world and hence consider design alternatives "in-themselves" as ideal-type artifacts which can be developed and studied in preparation for decision-making circumstances which might be propitious to their adoption either in 'pure' form or with some minor adjustments or amendments. Of course this is the bread-and-butter of policy analytical work undertaken by think tanks, policy institutes and policy schools which generally criticize existing arrangements and propose more 'rational' alternatives; that is, ones felt more likely in the abstract or in practice to achieve their goals.
Again, however, this does not preclude, but rather is built upon the recognition and acceptance of the fact that some policy decisions and formulation processes are highly contingent ones in which 'design' considerations may be more or less absent and where the logical or empirical relations between policy components are ignored (Kingdon 1984;Cohen, March and Olsen 1979;Dryzek 1983;Eijlander 2005;Franchino and Hoyland 2009;Sager and Rielle 2013). This includes a variety of contexts in which formulators, for example, may engage in trade-offs or log-rolling between different values or resource uses or, more extremely, engage in venal or corrupt behaviour in which personal gain from a decision may trump other evaluative criteria. These are 'nondesign' situations and the extent to which such considerations as political gain or blame avoidance calculations outweigh instrumental factors in policy formulation is an empirical question, however, and can be studied systematically (Hood 2010).
As Junginger (2013) recently argued, however, at the present time we continue to know too little about many important aspects of design and non-design work, especially about the nature of the kinds of policy formulation activities which bring about either result. As she put it, we know very little about "the actual activities of designing that bring policies into beingof how people involved in the creation of policies go about identifying design problems and design criteria, about the methods they employ in their design process " (p. 3).
That is, while many commentators, pundits and jaded or more cynical members of the public may assume that all policy-making is 'political' and hence irrational in a design sense, policy scholars have noted many instances in which processes of policy formulation are governed less by considerations of interest accommodation and bargaining than by concerns about criteria such as the practical efficiency and effectiveness of policy alternatives which involve policy formulators thinking more systematically and analytically about their options (Bobrow and Dryzek 1987;Bobrow 2006). This highlights the continued need for better understanding the mechanics of policy formulation involved in translating ideal-type models into context-sensitive solutions to public problems (Linder and Peters 1988;Wintges 2007) and to distinguish more carefully between design and non-design processes.

What is Policy Non-Design?
The academic enquiry of policy designthat is, self-consciously dealing with both policy processes and substance under an instrumental rubric -emerged and flourished throughout the 1970s and 1980s (see for example, Salamon 1981Salamon , 1989. Studies of policy design with this general orientation towards policy formulation began at the very origins of the policy sciences when many pivotal early works contained within them the idea of improving policy outcomes through the more systematic application of knowledge to policy formulation activities (Lasswell and Lerner 1951;Wildavsky 1979;May 2003). 2 In his foundational work on the policy sciences, for example, Harold Lasswell argued for the separation of the processes of policy formulation from decision-making and implementation, highlighting the centrality and significance of policy instruments and instrument choices made in the formulation process for policy outcomes and arguing for the need to bring interdisciplinary knowledge to bear on the development of the appropriate means to resolve public problems and issues (Lasswell, 1954).
For the "old" policy design studies which emerged from this foundational work, the historical and the institutional context of policy-making was seen to bear significant weight in policy formulation, and this was often argued to be determinant of both the content and activities of designs and designing (Clemens and Cook 1999). In this view, as the policy context and conditions changed and evolved, so too did the set of policy means or alternatives which were deemed acceptable or feasible by an evolving set of policy actors involved in policy-making, themselves informed by shifting ideas and calculations of the appropriateness of a particular design and its consequences (Majone 1975 and1976;March and Olsen 2004;Goldmann 2005;Howlett 2011).
This highly contextual orientation in early policy studies (Torgerson 1985 and) led some policy scholars in the 1970s to argue that policy decisions were by nature the result of processes so highly contingent and fraught with uncertainty that decisionmaking would invariably involve a high degree of 'irrationality'; that is, be informed more by the opportunistic behaviour of decision-makers within fluid policy-making contexts than by careful deliberation and 'design' thinking about the logical or functional merits and demerits of specific alternative arrangements of policy goals and means (Lindblom 1959;Cohen et al. 1979;Dryzek 1983;Kingdon 1984). 3 This led some to express serious doubts that policy could truly be 'designed' in the way that proponents of a policy design orientation advocated (Dryzek and Ripley 1988;deLeon 1988).
Many other scholars, however, questioned the extent of this emphasis on contextuality and contingency (Dror 1964) and in a series of path-breaking articles in the 1980s and early 1990s authors such as Linder and Peters (1984;1990b;1990c and sought to re-orient design studies by arguing that the process of policy designing as a type of formulation activity was conceptually distinct from a policy design, in the same way that an analytical distinction can be made between the development of an abstract concept or plan in architecture and the manifestation of that conception through engineering and construction practices followed on the ground (Schon 1988(Schon , 1992. Incorporating this distinction between design-as-formulation-process and design-aspolicy-content, design studies in the 1980s shifted from the study of 'designing' to the study of 'designs' themselves, with a specific focus on better understanding how individual implementation-related policy tools and instruments such as taxes and subsidies or regulation and public ownership operated in theory and practice (Sterner 2003;Woodside 1986;Mayntz 1979). This marked the beginning of modern studies of policy tools and this tools orientation sparked interest in a range of related subjects, such as the study of target group behaviour, implementation failures and their role in policy success, and the linkages connecting the two; with policy scholars turning their attention to the description and classification of alternative implementation instruments and the factors which conditioned their effective use and deployment (Mayntz 1979;O'Toole 2000;Goggin et al 1990;Schneider and Ingram (1990;1994). These works provided a deeper understanding of the social and behavioural factors underpinning the use of specific kinds of policy designs in practice. 4 Students of public policy making were joined in this effort by scholars of economics and law who studied the evaluation of policy outputs in terms of their impacts on outcomes as well as the role of law and legislation in effecting policy tool choices and designs (Stokey and Zeckhauser 1978;Bobrow and Dryzek 1987;Keyes 1996). And studies in management and administration at the time also sought to explore the linkages between politics, administration and implementation in the effort to better understand policy tool choices and patterns of use (Trebilcock and Hartle 1982). Researchers also looked at how policy instrument choices tended to shift over time (Lowi 1966(Lowi , 1972(Lowi and 1985, examples of which during this period included the rise of privatization and deregulation (Howlett and Ramesh 1993) and the first wave of governance thinking advocating the use of network management or non-governmental tools (Peters and Pierre 1998). 5 By the early 1980s, this tools literature had merged with the policy design orientation and emerged as a body of policy design literature in its own right. Students of policy design consequently embarked upon theory building, developing more and better typologies of policy instruments that sought to aid the conceptualization of these instruments and their similarities and differences, and attempting to provide a greater understanding of the motivations and reasons underlying their use (Salamon 1981;Tupper and Doern 1981;Hood 1986;Bressers and Honigh 1986;Bressers and Klok 1988;Trebilcock and Hartle 1982). Other scholarly work during this period continued to further elucidate the nature and use of specific policy tools, especially tools such as "command-and-control" regulations and financial inducements such as tax incentives but also many others (Landry, Varone and Goggin 1998;Tupper and Doern 1981;Hood, 1986;Vedung et al 1997;. In general it was believed that a greater understanding of implementation instruments and the reasons underlying instrument choice would benefit policy design both as a practice and a theoretical body of knowledge, contributing to more positive policy outcomes (Woodside 1986;Linder and Peters 1984;Mayntz 1979). Studies on pollution prevention and professional regulation conducted at the time, for example, benefited from advances in the systematic study of policy instruments which influenced the design and creation of new alternative instruments in these and other fields (Hippes 1988;Trebilcock and Prichard 1983).
Most of this work focused on tool design-as-a-noun, however, and ignored the issues involved in policy-design-as-formulation-process. Understanding the difference between "non-design" and design thus remains very much a part of the outstanding research agenda in contemporary policy design studies.

Modeling Non-Design: Revisiting the Pre-Conditions of Policy Design
The modern policy studies movement began with the recognition that public policymaking results from the interactions of policy-makers in the exercise of power, legitimate or otherwise (Lasswell 1958;Arts and van Tatenhove 2004;Stone 1988). Although some of these policy-making efforts could be seen to be arbitrary or capricious, most were viewed as representing the concerted efforts of governments to act instrumentally; that is, to achieve a particular policy goal or end through the use of a set of relatively well known set of policy means developed over many years of state-building and experience (Lasswell and Lerner 1951). It was acknowledged that these goals can be wide-ranging and often posed no small amount of difficulty and complexity in both their definition and diagnosis, with the implication that the formulation of solutions that were likely to succeed in addressing them necessitated the systematic consideration of the impact and feasibility of the use of specific kinds of policy means or instruments (Parsons 1995 and.
This work thus depicted policy design as a specific kind of policy-making in which knowledge of the policy impacts of specific policy tools was combined with the practical capacity of governments to identify and implement the most suitable technical means in the effort to achieve a specific policy aim. This activity was expected to occur ex ante and independently of other considerations such as political or personal gain which might also affect formulation processes. Significantly, this 'design' activity was recognized as requiring a situation where there was support for knowledge-based policy analysis and design work on the part of policy-makers and also one where there is a low policy "lock-in" on existing tool arrangements which could preclude adoption of superior alternatives.
Such favorable design circumstances had to be coupled with the presence of a high level of technical capacity and expertise on the part of policy analysts if knowledge was to be mobilized effectively so that policy instruments were effectively and efficiently matched to policy goals and targets (Howlett 2009 andDunlop 2009;Radaelli and Dunlop 2013;Howlett and Rayner 2014).
When all such conditions are present, purposive design activity resulting in good alternative generation and assessment was thought to be possible, much as is expected in the current era with recent efforts at improving knowledge mobilization in policy-making in the form of an emphasis upon 'evidence-based policy-making' (Bhatta 2002;Locke 2009). When they are not, however, either poor designs could ensue from incomplete knowledge and information even with the best government intent, or less technical and more overtly political forms of non-design policy-making are more likely to ensue

Developing a Spectrum of Design and Non-Design Activities: The Significance of Layering and Temporality
In itself this suggests that a spectrum of design and non-design formulation processes exists between "good" processes which are informed and "poor" ones which are not. However in order to be more precise about the nature of these processes, it is necessary to examine the nature of some of the other constraints on government intentions which can negatively affect both design and non-design processes Here scholars in the new design orientation have been concerned with factors such as how policy processes can be affected by the previous existence of a policy design.
That is, not all design processes, in fact very few, begin de novo. Most must deal with already created policies and are limited by these historical legacies, which can be hampered due to internal inconsistencies. Although other policy instrument groupings can be more successful in creating an internally supportive combination (Howlett and Rayner 2007, Grabosky 1994, Gunningham, Grabosky and Sinclair 1998, del Rio 2010 it may be very difficult to accomplish or propose wholesale change and designs instead will often tend to focus on reform rather than replacement of an existing arrangement. In this case legacies from earlier rounds of decision-making affect the introduction of new elements which conflict with pre-existing policy components. Policy development strongly marked in this way is typically one where new elements were added to the policy mix without the removal of older ones and existing elements are stretched to try to fit new goals and changing circumstances. This creates a mix that contained various incompatibilities, tending to frustrate the achievement of policy goals. A key concept in this regard is that of 'layering" (van der Heijden 2011) or the result of (re)design in altering only some aspects of a pre-existing arrangement and it is a distinction between different types of layering which allows us to further distinguish different kinds of design and non-design processes from each other.
Layering, of course is a concept developed in the neo-institutional sociological literature by some of its leading figures, namely Beland (2007), Thelen (2004), Hacker (2004; Beland and Hacker (2004); and Stead and Meijers (2004) to explain the pattern through which social and political institutions have evolved over long-periods of time. As applied to policy-making, 'layering' connotes a process in which new elements are simply added to an existing regime often without abandoning previous ones so that polices accrete in a palimpsest-like fashion (Carter 2012).
This adds a second, temporal, dimension to design and non-design formulation contexts that most early policy design studies neglected. That is, most design studies have focused on what in fact is the exceptional case of 'replacement' or 'exhaustion' in which an existing policy is scrapped and a new one adopted in its entirety. However this is not a common event. Much more common is some process of layering in which some aspects of a policy are layered on top of pre-existing ones. This context 'lock in' can impact the formulation process by restricting a government's ability to evaluate alternatives and plan or design in a purely instrumental manner (Howlett 2009;Oliphant and Howlett 2010;Williams 2012). This is an issue in part for temporal reasons as policy arrangements are often the result of transformation pathways that can easily lead to internal contradictions emerging between tools and goals within policy mixes (Hacker 2004). Mixes of policy elements may emerge over long stretches of time as a result of successive policy decisions which are not necessarily congruent. As a result, even when the initial logic of each decision matching policy tool and target may have been clear, through multiple layering processes they can gradually transform into degenerated mixes over time (van der Heijden 2011, Bode 2006, Orren and Skowronek 1998, Rayner et al. 2001, Torenvlied and Akkerman 2004, Hacker 2004. Optimizing the choice of instruments when a pre-existing mix exists thus requires an additional level of knowledge of instrument-goal interactions and considerations of both long and short-term processes of policy change. That is, in addition to questions relating to the logic of integration of policy tools, the evolution and history of existing policy mixes are also of concern. While the old orientation tended to suggest that design would always occur in spaces where policy packages could be designed 'en bloc', the new orientation recognizes that most design circumstances involve building on the foundations created in another era, working within already sub-optimal design spaces (Howlett and Rayner 2013).
In such situations of significant policy legacies, designers often attempt to "patch" or restructure existing policy elements rather than propose completely new alternative arrangements even if the situation may require the latter for the sake of coherence and consistency in the reformed policy mix" Thelen et al. 2003;Thelen 2004;Eliadis et al. 2005). Hence even where intentionality to design may be high it may only be partial in the sense that patching and not replacement is on the table.
Hence a key first distinction among design formulation processes concerns whether they involve 'packaging' a new policy mix or 'patching' an old one. Layering is often thought to be inherently sub-optimal but 'patching' in itself is not 'non-design', as very often the new layer is designed in an effort to overcome anomalies or problems with earlier mixes (Howlett and Rayner 2013). Policy design scholars are thus very interested in processes such as how policy formulators, like software designers, can issue such 'patches' to correct flaws in existing mixes or allow them to adapt to changing circumstances (Rayner 2013;Rayner 2013). And they are also interested in related subjects such as how policy experiments can help reveal the possibilities of re-design (Hoffman 2011) or how building temporal properties into tool mixes -"adaptive policy-making" (Swanson et al 2010) -can make designs more flexible or resistant to shifting conditions (Walker et al. 2010, Haasnoot et al. 2013). Patching can be either a form of 'smart' layering if done well, or not so smart if done poorly.
Another second phenomenon which can occur as layering unfolds is 'stretching' (Feindt and Flynn 2009). This is where, operating over periods of decades or more, elements of a mix are simply extended to cover areas they were not intended to at the outset. "Stretching" is especially problematic as small changes in the mixture of policy elements over a decade or more can create a situation where the elements can fail to be mutually supportive, incorporating contradictory goals or instruments whose combination create perverse incentives that frustrate initial policy goals. When these problems are identified, they set the stage for further rounds of tinkering that may make them worse (Feindt and Flynn 2009). This second process is associated with a particular form of 'tense layering' (Kay 2007) which occurs when repeated bouts of layering lead to both incoherence amongst the goals and inconsistency with respect to the instruments and settings used in a policy area. As Kay (2007) and Feindt and Flynn (2009) noted, destructive layering can be initiated by a process of stretching the regulatory framework to accommodate new and potentially incoherent policy goals. In such cases tense layering introduces progressively more severe inconsistencies and incongruences and tensions between layers.
Stretching is more problematic as a design process than patching since the addition of new goals or objectives increases the risk of incoherence, as does the introduction of policy instruments that suppose new kinds of implementation preferences, for example, when a market orientation is introduced into an instrument set that has been based on a regulatory approach (Howlett and Rayner 2007). Inconsistencies also arise where the means work at cross-purposes, "providing simultaneous incentives and disincentives towards the attainment of stated goals" (Kern and Howlett, 2009: 6). And incongruence occurs when an otherwise consistent mix of instruments fail to support the goals.
Layering thus has two sides to it. On the one hand negative stretching or destructive layering exacerbates tensions between regime elements and leads to wholesale change.
However layering can also have a positive side and help ameliorate or reduce tensions through patching. Moderate layering can be successfully accommodated through a process of learning and patching, leading to a policy mix that exhibits a high degree of coherence, consistency and congruence.
Both these processes fall between the design and non-design ends of a spectrum of design processes which moves from highly intentional and instrumental replacement efforts to those which are more partial and less intentional such as 'smart' patching and ultimately to those which involve poor design such as 'stretching' and poor or 'dumb' layering (see figure 2 below).

|-------extent of attempted or possible alteration of status quo by design type ------------|
All of these design efforts can be done well or poorly but reflect some wholesale or partial effort to match policy goals and means in a sophisticated way linked to improving outcomes. Non-design types also vary in the same way but more by process of decision-making than by their sphere of activity. Non-design mechanisms, as highlighted above, include activities such as alternative generation by bargaining or log-rolling, through corruption or co-optation efforts or through other means which are not instrumental in the same sense as are design efforts. Again such efforts can also be done poorly or well (for example, maximizing the return from a bargain or the returns from corruption) depending on the context and situation but do not involve the same appraisal activities and competences or intentions on the part of governments (see Figure 3).

Figure 3: A Spectrum of Policy Non-Design Types
These processes vary in terms of the extent to which the policy goal is linked to individual and political interests rather than public ones. Most have been studied extensively in the political science literature but less systematically in the policy sciences (Saward 1992;Goodin 1980;Frye et al 2012;Gans-Morse et al 2014).

Conclusion: Distinguishing Design from Non-Design-Based Formulation
Transforming policy ambitions into practice is a complex process. The efforts of policy makers often have failed due to poor designs which have inadequately incorporated this complexity in policy formulation (Howlett 2012;Cohn 2004). These experiences have led to a greater awareness of the various obstacles that can present themselves to policy design and have gradually fueled a desire for better understandings of the unique characteristics of policy formulation processes and the spaces in which design efforts are embedded.
The new design orientation calls for a broadening of thinking about design beyond policy tool choices, examining combinations of substantive and procedural instruments and their interactions in complex policy mixes. It also has focused on more detailed study of the actual formulation processes involved in tool and design choices as these occur and have evolved over time (Linder and Peters 1990;Schneider and Ingram 1997;Considine 2012).
As the discussion here has shown, both design and non-design formulation processes are not unique but vary along several important dimensions. For design situationsthat is those characterized by a government desire to systematically match ends and means in the attainment of policy goals, the processes vary according to the nature of the resources available for design purposes and the constraints imposed by policy legacies, with the latter generating non-replacement spaces in which processes such as patching and stretching unfold. In the non-design world where the intention to design is lacking, constraints on outcomes also exist as do different processes which vary in their distance from the design ideal of public service and improving the public good (Holmberg and Rothstein 2012;Rotberg 2014).
Students of policy design must be aware of these differences and the situations governments are in or want to be in while developing policy options and making recommendations and providing advice to governments. More systematic study of these formulation contexts and processes can help move this area of policy design studies forward.
Endnotes 1 This is similar to the general orientation towards design found in other fields such as architecture, urban planning or industrial design. See Hillier, Musgrove and O'Sullivan 1972;Hillier and Leaman 1974;Gero 1990. 2 Policy design studies have been undertaken since at least the 1950s (Tinbergen 1952;Dahl and Lindblom 1953;Kirschen et al 1964). Most of the early studies focused on policy tools and had a strong focus on policy implementation issues and processes; paying much less attention to policy development or formulation issues which are the hallmark of current studies with a design orientation (Hood 1986;Hood and Margetts 2007). 3 Of course this is a view some continue to hold. See for example Eijlander 2005;Franchino and Hoyland 2009. 4 Subsequent contributions would further advance the study of the behavioural aspects of the design process and raised the issue of the difference between design and non-design to the fore Schneider and Ingram 1997;Mondou and Montpetit 2010;Timmermans et al 1998;Hood 2007). At this time, for example, Bardach (1980) and Salamon (1981) went so far as to argue that the definition of policy in terms of "issues" or "problems" originally made by scholars at the outset of the policy studies movement (Mintrom 2007) was misguided and that policy should instead have been defined from the start in terms of the "instruments" used in policy-making. They advocated shifting the focus of policy studies squarely towards the study of the design and operation of such tools, later defined to include both traditional 'substantive' tools such as regulation and public ownership and more 'procedural' ones such as the use of advisory commissions and public participation exercises (Howlett 2000). 5 Of course, not all work on policy instruments has restricted itself to implementation issues. Work on the exploration of "instrumentation" for example, has considered larger issues about feedback processes from instrument choices to the politics of policy formation, as has some work on instruments and network governance (see Legales 2007 andde Bruijn andten Heuvelhof 1997). However these can still be distinguished from the new design studies, given the latter's almost exclusive emphasis on formulation and its resulting concern for understanding the inherent nuances involved in developing mechanisms for meeting policy goals, couched within contextual realities, which the former studies still lack.