TRANSFER AND LEARNING one Coin Two elements

Over the past 20 years, three bodies of literature (and their numerous outgrowths) have developed relatively independently of each other: studies of transfer, studies of learning, and studies of knowledge updating. This article will seek to illustrate how each has developed in relation to policymaking and then to link them through a discussion of how policy transfer can be better used to explain policymaking if viewed through the lens of knowledge updating as it occurs during the policy cycle.


TRANSFER AND LEARNING
one Coin Two elements David P. Dolowitz*

INTRoDuCToN
Over the past 20 years, three bodies of literature (and their numerous outgrowths) have developed relatively independently of each other: studies of transfer, studies of learning, and studies of knowledge updating.At their most basic, transfer studies examine the process by which one political system integrates policies that have been previously used in another political system.In a similar way, at their most basic learning studies tend to examine how and under what conditions information is acquired in the individual and institutional level and what types of psychological and outlook changes might occur as a result to this information.
[*] University of Liverpool, Liverpool, UK. d.dolowitz@liverpool.ac.uk http://dx .doi.org / 10.25091/ S0101-3300201700010002 Finally, at its base the knowledge utilization literature is interested in how knowledge is used within the policymaking process and what changes this utilisation might bring about.
This article suggests that the concepts of policy transfer, learning, and knowledge updating should be brought together.The goal will be to view how the three theoretically thought the prism of a typical policymaking model while illustrate some of the linkages through empirical examples of policies working their way through the actual policy process on its way to implementation.This should start to bring the different literatures together while advancing each one of them independently.Most importantly by linking the three literatures it should start the process of allowing the transfer literature to move forward from simple descriptive frameworks toward more advanced modelling that could have better predictive power.

Diffusion
During the 1960s American and European scholars began discussing how innovations appeared to spread amongst political jurisdictions in close proximity to each other.Out of these came Everett Rogers' Diffusion of Innovations.This text explored the way innovations and technologies spread from one jurisdiction to other neighbouring systems through direct information sharing and networks of communication.Following this publication diffusion studies emerged as a focus of policy analysis. 1These early diffusion studies improved our understanding of how communication channels between neighbouring jurisdictions were involved in the spread of information that led to similarities in policies.Diffusion studies also brought attention to the fact that diffusion tended to follow an s-shaped adoption curve across a region.The initial period of the s-curve sees fairly slow diffusion from one system to another.However, there is a "take-off" point where a number of jurisdictions join a bandwagon and adopt the innovation (reform) in short succession.This period lasts for a relatively short period, after which the number of systems adopting the reform trails off fairly quickly.When linked into the learning process a number of authors have found that the s-curve has a final period where the adoption process appears to go into reverse as nations who avoided jumping on the initial bandwagon, actively decided not to adopt the innovation as they see the results of the reform in a range of systems and others start to dismantle the reform as a result of unexpected and undesirable outcomes. 2hile diffusion studies have added much to our knowledge of transfer, 3 particularly those associated with the European Union and its internal and external interactions, 4 the concept and study of policy [5] Berry; Berry, 1990; Berry; Baybeck, 2005.[6] Gilardi, 2003; Phillips; Ochs,  2004.[7] Also known as spurious diffusion and is related to what is more commonly known as Galton's Problem.
[8] Braun; Gilardi, 2006; Jahn,  2006.[9] Bennett, 1991; Bennett;  Howlett, 1992; Rose, 1991; 1993; Wolman, 1992.diffusion has been criticized on a range of fronts.First, diffusion studies have tended to equate the spread of innovation with borrowers being involved in a deliberate and active effort to learning from the innovative system.In this formulation, an innovation can be seen as relating to a new idea, policy innovation or way of doing something, or a new procedure in how a policy is implemented or developed.However, as the concept of herding around suboptimal innovations illustrates, innovations do spread in the absence of reliable learning about the consequences of introducing a foreign idea into one' s own sociopolitical system.Second, many diffusion studies have been criticised for ignoring how the inherent characteristics of an innovation and/or the borrowing system interact to influence its subsequent spread and adoption pattern.Thus innovations that are seen as simple technical changes are more likely to spread than an innovation involving substantial political, fiscal, and procedural capital.Due to the differences in the ability of innovations to spread or for accurate information to spread with the innovation much of the diffusion literature fails to account for the importance of the adaptation processes involved in the spread and adoption of an innovation.
Third, diffusion studies have been criticised for underplaying or ignoring micromechanisms involved in the spread of innovations: particularly when this spread occurs across national boundaries. 5As such, diffusion studies underplay or ignore the role played by personalities, ideologies, and politics in the movement of innovations.A final criticism worth mentioning relates to spurious diffusion.Because diffusion studies neglect micromechanism there is a tendency to offer few definitive ways to prove that observed similarities are the result of the spread of an idea or policy versus similarities that emerge result of simple coincidence. 6,7 n other words most diffusion studies offer no way around Galton' s Problem. 8

Lesson drawing
Building on diffusion studies a new wave of scholarship emerges in the 1990s that became known as lesson drawing. 9Like diffusion, lesson drawing is interested in how ideas and policies spread.Unlike diffusion, lesson drawing is interested in the microprocesses involved in the movement of ideas and policies across geographic boundaries.More importantly, lesson drawing places rationality at the core of understanding who becomes involved in the movement of ideas and policies, how they are involved in the transfer process, and how and why lesson drawing occurs.The lesson drawing logic is that policymakers are able to draw lessons from localities where policies are successfully working in a real world situation.This allows them to reduce the risks and costs associated with the development of an entirely new policy in their home system when a similar problem arises. 10Thus, while diffusion studies are concerned with tracing the spread pattern of an innovation, lesson drawing examines the logics and mechanisms driving policymakers to look to exogenous systems for solutions to existing problems (whether real or perceived).
While the focus on rationality adds insights into why policymakers are interested in and engage with the movement of ideas and policies, lesson drawing' s genuine innovation was its focus on the microprocesses of the movement of ideas and policies from one system to another.In addition, by focusing on the role of rational learning in the transmission process lesson drawing allowed for a more systematic link to the policymaking literature.Finally, by focusing on the micro, lesson drawing was able to divide learning along a continuum running from copying a foreign model in total to using information from a range of different policy models as an inspiration for the development of a "new" or hybrid model. 11s with diffusion, a number of shortcomings have been raised in relation to lesson drawing.Probably the most frequent relates to an overreliance on the rational actor model as its primary explanatory variable.While some form of rationality might underpin the actions of some agents, it is equally clear that other agents and decisions rely on considerably less rationality than suggested by the lesson drawing literature.For instance, it has been well documented that there is a tendency in the international community to irrationally herd around suboptimal norms in the areas of economic policy and utility regulation. 12Similarly, studies have demonstrated that a range of transfers have occurred not out of the rationally but rather as a result of the fears and ideologies of policymakers. 13Indeed cultural factors have been shown to override rational impulses when agents engage in the transfer of sustainable development models; often being shaped more by tacit, culturally embedded beliefs, than any rational analysis of what is occurring elsewhere. 14 second criticism of the literature is its tendency to argue that lessons travel as a result of a voluntary processes, where borrowers actively look for solutions to their problems.While it is likely that a range of bilateral transfers between advanced industrial nations are more-or-less voluntary, it is much less likely to be true of interactions occurring between advanced and developing nations, between international lending organisation and developing nations, international governing organisations (igos) and developing nations, or between advanced nations and organisations and underdeveloped nations. 15gos and non-governing organisations (ngos) have also been shown to include more-or-less coercive mechanisms involving conditionality or obligations in exchange for financial support.This is clearly il-lustrated in the activities of the European Union and its interactions with a range of African nations, border states, and those wanting to join the Union.Beazer and Woo 16 capture this process while studying the International Monetary Fund: The International Monetary Fund (imf) often seeks to influence countries' domestic public policy […].One increasingly exercised took at the imf's disposal is conditionality, or explicitly linking financial support to borrowing governments' commitment to policy reform […].The imf and other international institutions use conditionality to encourage governments in crisis to adopt difficult […] reforms that domestic leaders might otherwise avoid […].imf programs have enormous economic and social consequences for participating countries. 17ese interactions are occurring between agents of unequal power and operate under unequal power configurations.Not only are these unlikely be conducive to voluntary interactions but they are less likely to involve hard learning (or learning as intended) because, "changes in the thinking of political leaders […] are not usually the sort of things that can be 'engineered' by actors who are external to the country in question". 18As such, one should expect a number of unexpected outcomes as a result of policymakers failing to develop the knowledge necessary to understand why something worked well in one system when they are coerced to transfer a policy into their own system.Similarly, in these situations it is likely that once the policy enters the system that indigenous actors will attempt to co-opt the policy in ways that minimises their need to change or how the policy impacts the existing order.

Policy transfer
Building on the diffusion and lesson drawing literatures a group of scholars began discussing what has come to be known as policy transfer. 19At its core policy transfer is a process where knowledge of how to make things work in one political system is used in the development of similar solutions in another political system. 20To help frame the analysis a series of questions were developed to assess the meso-and microprocesses involved in the movement of policies, ideas, techniques, and information from one political system to another, from one time to another, and/or from the international to the national or local. 21By focusing on what, whom, and into what "conditions" a policy was moved the transfer literature was better able to examine the question of what was learned (or not) during the process of movement, development and implementation than either the diffusion or lesson drawing literatures.It was also better able to address issues associated with why transfer often appeared to end in the failure of the transfer to achieve the desired goals of policymakers. 22hile the policy transfer literature has much to commend, it too has shortcomings.First, much of the literature focuses on voluntary transfer process.This has allowed the role and processes involved in less voluntary forms of transfer to remain fairly understudied, particularly as they interact within political systems, semi-independent subunits of political systems (e.g.British Council), bilateral arrangements and pilot projects, and the role of international financial institutions (ifis), igos and domestic and international ngos in the spread of global paradigms and "best practice" models.Transfer studies have also tended to neglect the human questions; the political, ideological, and unobservable (tacit beliefs) involved in transfer.As a result, much of the existing literature looks overly mechanistic: If you have problem x, look to system y, borrow z, and then will follow.In reality this seldom occurs, at a minimum lessons combine prior to entering the political process.Similarly, at any stage of the policymaking or implementation processes it is possible that the "lesson" will be transformed and/or translated into something new or more palatable to the importing/receiving system.By focusing on the voluntary and logical side of transfer, the literature has tended to overlook the possibility that lessons might emerge and be transformed by ideological predisposition, who policymakers are willing (able to) talk to, what the transferring systems shows (or the borrowing system sees).Or as Meseguer and Gilardi argue, "even if a particular policy is showing good results elsewhere, it may not spread if it is found to be ideologically alien, electorally risky and/or unlikely to be passed". 23his is itself linked to the often overlooked role of motivation.If transfer studies integrated motivations underpinning the activities of lenders and receivers it might begin to better explain where and why actors turn to one location for information but not others.Motivation can also start to draw in concepts learning and how deeply agents become involved in the learning process when they engage in transfer, policymaking and policy implementation.For instance, a politician might engage in transfer to find a model they can use to symbolically appear to be doing something.From a different position, this same politician might engage in transfer to create a policy capable of meeting an indigenous need.The same policymaker working at implementing a policy may be motivated to look to other systems to discover how the technical aspects of the policy operate in the originating system in order to discover the minutiae of how it is embedded into the wider policy arena. 24hile much of the literature suggests that policies are taken whole scale, in reality, what is borrowed will generally undergo substantial transformations.Rather than x being borrowed from y and subse-quently being implemented exactly as x appeared in its originating system, policymakers involved in transfer and decision making "are like a composer writing a symphony for a number of instruments; the quality of the symphony will depend upon the melody written for each instrument and also upon the combination of the many melodic lines". 25By neglecting the complexities of the process associated with the creation and implementation of a policy and the nature of learning involved in the movement and transformation of ideas, information and policies, the transfer literature is missing much of the microdetail of the end product.

Translations and mobilities
While transfer is an improvement on what came before, its shortcomings has led a new batch of scholars to relabel and refocus the study on the translations that occur to initial ideas and policies as they work their way from one system to another and then through to the implementation process. 26,27 hile there are differences between the mobilities (grew out of human geography) and translation (grew out of public policy) literatures, at their core both are concerned with how hybridization, adaption and mutation occur during the movement development and implementation of a policy; thus how ideas and policies are translated while in movement. 28As such, both are concerned with how "place, space and scale, coupled with an anthropological/ sociological attention to social relations, networks and 'small p' politics, both within and beyond institutions of governance, promise to deepen and strengthen" understanding of the role "foreign" has in the policy process. 29Both literatures also attempt to broaden the study of policy transfer so that the transfers involve "a global-relational social product -one produced by its circulation […] among cities, as much as its development in cities […]". 30For these literatures policy is not something waiting to be "taken" or "sent" but, rather something that is developed in networks that live in the ether surrounding and engulfing the decision making process.The very act of moving policies leads to mutation as they combine with other ideas, policies and experiences.As such, translation can be seen as "a series of interesting, and sometimes even surprising, disturbances can occur in the spaces between the 'creation', the 'transmission' and the 'interpretation' or 'reception' of policy meanings". 31The overall goal is to understand how the settings in which policies move transform the policy and influence the impact of what is received and how the policy continues its transformation as it moves toward implementation in the new system. 32hile the idea of translation is appropriate for some transfers, it is less obvious for others, particularly where one sees a photocopy emerge out of the process.In a similar vein, one of the most difficult [34] Weyland, 2004, p. 5.
[35] Westney, 1987; Weyland, 2004.issues with adopting a translation perspective to the study of transfer is attempting to fit the analysis into a social constructivist analysis.To date, most translation studies attach a few sentences or paragraphs on social constructivist analysis, but this takes a back seat to a standard who, what, when, where and why found in other transfer studies.
CAN A LeARNING LINK LeAD To A BeTTeR PoLICY TRANSFeR?
The current literature looks mechanistic: if you have problem x, you can look to system y (or w, x and y), and then borrow z and will automatically follow.This is normally not the situation.The link between learning and policy is considerably more.At a minimum the nomenclature associated with transfer implies different degrees to which knowledge is acquired and applied exist and that this is true no matter what the situation.For instance, lesson drawing implies that the actors involved in the transfer process engage in a bounded learning process.And, that complete analysis takes place before they borrow a policy and then again after the transfer takes place.Diffusion on the other hand often appears to happen in the absence of any formal learning.So, "while learning is central to 'the human condition', it is not easy to define".Particularly, as "people learning need not process information correctly, draw valid inferences, nor improve diagnosis and policy recommendations". 33It is equally clear that "humans never look at 'the facts' with complete neutrality and objectivity, but always interpret them in light of general cognitive schemes and the specific theories they embrace". 34Thus, while learning is occurring, what is learned does not necessarily lead to a better understanding of the existing situation, the situation existing in the system(s) originating the models, or necessarily a better understanding of the overall impact that the global policy environment might have on the transfer and transfer process.
Just as transfer is influenced by sociological and contextual factors that policymakers are embedded, so are their abilities to learn about foreign models.At a minimum ideological blinders and cognitive shortcuts shape what is seen and heard when looking at a foreign political system.As such, learning is itself an adaptive process that will cause many transfer situations to end in unexpected outcomes. 35uch of the existing learning literature focuses on observable change.However, when viewed through the transfer process, learning can (and does) occur in the absence of immediate or directly observable change.For instance, when a learner is not in a position to introduce a lesson when it is acquired it may not ever be utilised by the core policymakers.However, if the lesson is able to stay in a stream until a window of opportunity opens then there is a chance that it can shape [36] See Kingdon, 1995.
[37] In these circumstances it is less likely that one will be able to trace the origin and process by which foreign knowledge impacted a policy in any directly observable way.
[43] Heclo, 1974.a policy or political output. 36In a similar way individuals operating in larger institutions often find lessons they feed to their superiors must first go through institutional processing before it can find expression. 37In these circumstances, information originating elsewhere may be expressed in ways that make it unrecognizable as the lesson that was initially transferred.This is particularly likely if the information is modified to better match the preexisting internal psychological filters an agent or institution uses to understand the world, or when the information is recombined with other ideas that have been "stored" for later use. 38his links to another issue: the quality of information.Most studies present learning as if it was easy, but as already shown, what is learned is actually translated in a number of ways.Moreover, quality often relates to the level of engagement in the learning process, and this can vary along a continuum running from full engagement to no engagement.Based on the level of engagement, it is possible to see learning as falling along a continuum running from perfect learning to marginal learning.Stone 39 refers to these two extremes as soft learning and hard learning.Others have discussed the process as single-looped learning and double-looped learning. 40No matter how referred to, each end leads a different form of transfer as "learning is uneven and imperfect […][it] can be of different 'orders': shallow, tactical or instrumental […] as opposed to deeper social or policy learning". 41In light of this, the poorer the quality of information the less likely an accurate image of original system will emerge and the less likely an accurate knowledge map will form.
There are likely to be different forms of learning.For lesson drawing (and much of the transfer literature) policymakers are seen as engaging in rational learning.Policymaker' s turn to foreign ideas and policies in order to solve a problem by transferring (in an unbiased, educated fashion) information necessary to develop a similar solution in their own system.Even scholars who relax the model to "bounded rationality" assume that the process is undertaken voluntarily and that information is sought to address a specific problem.Unfortunately, while there is a lot written on learning and transfer from the rational position as Mesegue and Abel state, "learning is hardly tested; and […] the few attempts that do exist […] suggest that there is little support for the hypothesis that learning proceeds in a rational fashion". 42

Learning and transfer
Given the above, it would appear that the transfer literature could benefit from a more explicit understanding of learning.One of the first to see this was Heclo. 43 In Modern Social Politics in Britain and Sweden, Heclo linked learning into the international movement of policies [44] Heclo, 1974, pp.305-306.[45] Sabatier, 1988.
[48] Hall, 1988.between Britain and Sweden.For Heclo learning amounted to a social process of "collective wondering what to do", so that: "Policymaking is a form of collective puzzlement on society' s behalf; it entails both deciding and knowing […].Much political interaction has constituted a process of social learning expressed through policy". 44Applied to transfer what policymakers learn will relate to the outcome of their collective puzzlement relating to information involving the activities of a foreign political system and how they translate this during the transfer and policymaking processes.As this implies, learning is not just random puzzlement on the part of policymakers, but is shaped by the individuals involved in the movement of information; the relationship between the individuals and institutions they are embedded in; and it is shaped by what already exists in the transferees home system.These paths create flows that help explain what information moves from one system to another and how it is subsequently shaped and used in its new sociopolitical setting.When applied to the transfer process it becomes unlikely that learning will ever be fully observable or involve formulaic copying in which the original model is precisely replicated in the new location.
Probably the most ambitious attempts to link learning into the policymaking paradigm was developed by Paul Sabatier 45 and subsequently Sabatier and Jenkins-Smith 46 and became known as the advocacy coalition view policymaking.This sees policy as something that develops over at least a 10-year periods of time and that in this learning occurs in coalitions and the subcoalitions associated with it.Learning is possible because coalitional actors engage in a large range of interactions across multiple levels of governance.This allows actors operating inside and outside traditional governing institutions to share information and develop a policy or policy area over time and across space. 47he core insight for the transfer process is that learning is constrained.Or as Sabatier and Jenkins-Smith argue, a coalition ability to learn is constrained by a set of "deep core" beliefs.These beliefs act as filters on transferred information so that it is reinterpreted according to the coalition partners shared views of reality.As a result, learning, even when applied to "policy cores" and "secondary matters" (progressively more open to change) is inherently political.Because of this, one of the difficulties in studying transfer is that it is likely that any observable change will be evolutionary, making the location and timing of transfer difficult to capture and measure.
Another conception of learning was developed by Hall, 48 who conceives of learning as a social process where learning is a "deliberate attempt to adjust the goals or technique of policy in the light of the consequences of past policy and new information so as to better attain [49] Hall, 1988, p. 6, emphasis added.
[53] There are other conceptions of learning that have been used by public policy scholars including the rational learning models used by lesson drawing, and the government learning models forwarded by Etheredge (1982).
the ultimate objects of governance". 49Learning itself can be expressed through degrees of change.Hall discusses first order change as routine or incremental change, such as when the procedures governing the application processes associated with the receipt of welfare are slightly altered.This can be involved in some of the more basic transfers, it is unlikely to involve complex learning but rather a much more "monkey-see-monkey-do" processes.Or as Dobbin, Simmons and Garrett argue, first order change is likely to be "based on fads, revered exemplars, or abstract theories, rather than sold evidence". 50econd order change involves more "strategic thinking" and tends to be associated with changes in policy instruments and plans.When engage in transfer this is likely to be the realm where it will be observed, as it will probably involve looking for and thinking about policies and instruments to see if others do it better.Third order change involves altering ones deep core beliefs (paradigm shift).To undergoing third order change a policymaker entire worldview is fundamentally altered, which includes their underlying assumptions, beliefs and goals.Not only is third order change outside the normal course of policy development and change, it is unlikely to be a core element of individual instances of transfers. 51If third order change is to be tied to the transfer process it is most likely to occur over long periods of time and involve cascading transfers that end in a paradigm shift (such as the spread of the neoliberal economic and social policies since the mid-1980s across a range of nations and international lending organisations).
When linking Hall' s model to transfer it is important to remember two points.First, the core of the model is about change in the behaviours and beliefs of "elite" policymakers: not transfer and policymaking per se.Second, according to Hall policy deliberation "takes place within a realm of discourse […] [where] much of it is taken for granted and unnameable to scrutiny as a whole". 52Because deliberations are confined to discourse, when linked to transfer, learning will most likely be constrained and most closely linked to the translation processes which surround information as it moves and works its way through the policymaking and implementation processes. 53 WHeRe Do We Go FRom HeRe?
If the transfer literature is to move forward it needs more than a basic link to the learning literature, rather it needs to begin integrating the knowledge utilization and policymaking literatures.While a very diverse set of literatures, at the core of most knowledge utilization studies is a discussion and understanding of how knowledge develops, works its way into the political/policymaking processes, and is used by policymaking actors/institutions in the development, implementation, and evaluation processes. 54The first thing to note about knowledge utilisation and transfer is that new knowledge about a foreign political system is likely to enter a policy system slowly and initially likely to lead to only small changes in a policymaker' s conceptual frame (possibly leading to second and third order change in the long-run).Second, when knowledge about a foreign political system enters a system directly and in identifiable ways, the information is likely to be transformed in some way so that reshaped knowledge is used in order to make sure the information fits the political-culturalideological needs of the agent using the knowledge (this may or may not be the same agent who engaged in the initial transfer).Moreover, based on the transfer, learning and knowledge utilisation literatures, when an agent uses knowledge gathered from a foreign systems, it is likely that they will not only transform it to fit their "world view", but they are also likely to use it in fairly selective ways: "as a political weapon legitimizing an already advocated political position". 55his is problematic for policymaking.When information is used as a weapon instead of as its initial intent and as it operates in its original setting (or as the lenders intended), the bits of knowledge used (out of the totality of what was transferred) is likely to lose vital information and in selective ways, which is likely to lead to problems with any resultant policy. 56Even when agents attempt to be true to transferred knowledge, it is unlikely their lessons will be complete enough to truly understand all the factors that led to success (or otherwise) in the originating system (or systems).This is particularly true when those using transferred knowledge are not party to all of the material that was transferred.Part of this is a result of information itself; nothing can be fully known, no matter how scientific.Rather knowledge and information are packaged and moved and unpackaged as policymakers attempt to understand what is being transmitted in light of their own cultural/structural needs.
The difficulty in examining learning from foreign systems and seeing if and how appropriately policymakers apply this knowledge should not be seen as a reason for jettisoning the idea of learning and transfer.Rather by focusing the lens of knowledge utilization into the transfer process and directing both at the policymaking, implementation and evaluation processes (i.e.evidence based policymaking), it is likely that the study of policy transfer will become even more useful for scholars and policymakers alike.By using the learning and knowledge utilization frameworks the transfer process can better be analysed longitudinally (as occurring over time) and in a series of stages, which culminate (or not) in a recognisable policy action.Once this is incorporated in to transfer studies, the view of transfer as being little more [57] Hertin et al., 2009; Radaelli,  2009.than a process of actor in a looking to b and taking c to solve problem x must be restructured.
While the focus of much of the knowledge updating literature has been on long-term change it must be acknowledged that when agents engaged in the transfer of information knowledge updating is not confined to long-term change.Knowledge utilisation can lead to immediate change in how policies are understood, used, and in the way issues are comprehended.As such, knowledge can make an immediate impact, particularly when it is used as instruments in the battles of ideas surrounding the development of a new policy or in the alteration of an existing policy. 57ll told, while transfer can lead to change when it involves knowledge updating and use, this is not necessarily as straightforward as taking and implementing in total a photocopy of what is observed.In the first instance, knowledge that arises out of transferred information may sit within a policymakers conceptual knowledge map and be impossible to trace back to its origin by the time they utilise it.There will also be instances where an agent' s knowledge base is the result of the merging of a range of different transferred ideas into a single model that they forward.Similarly, knowledge may be held unaltered until a window of opportunity opens allowing them to forward it into an active policy system; its origins getting lost while it works its way along the policy stream or after being coupled with other ideas and models.Here it is probably worth considering if there is anything different about this form of knowledge utilisation and formation from a similar situation involving indigenous knowledge or at what point do we stop calling something transfer and begin referring to it as knowledge utilisation?
Most lessons, whether indigenous or exogenous tend to get transformed as they go through the policymaking and implementation processes.As such, transferred data is little more than a piece of many different pieces of data that combine into a final solution during the policy cycle.It is this combination of data, including that which comes from the evaluation processes associated with pilots or small scale programs used to provide data for "evidence based policymaking" that help policymakers update their knowledge as transferred data encounters other policymakers and other types of information exogenous and indigenous to the system.At any moment in the policy cycle an individual policy will be in a state of flux and undergoing modification.As such, any imported idea or model is constantly confronted by and mixing with other ideas and models.
Attaching a knowledge updating view to this development pattern suggests that to understand transfer scholars must look beyond where policymaker a sees a policy in system b and then uses it in an unal- [58] Dobbin; Simmons; Garrett,  2007; Hoppe, 2009; Meseguer,  2005; Radaelli, 2009.tered form to a similar policy in system a.Rather, it is far more likely that due to the complexity of the policy process policymakers in system a are likely to see a range of policies in systems b, c and d and that these (or part of these) policies are combined with e, f, g to create a "new" policy z.And, that as z works its way through the policymaking process it will be further altered and modified as new ideas (some transferred some indigenous) mix with it.When the former does occur, it is far more likely to be linked to systems that allow for a great deal of unitary government control, involve a form of technocratic learning, is used to initiate very small alterations, or is used in some symbolic fashion.
When knowledge updating occurs as a result of transfer it is likely to occur over a series of stages rather than a single instance of all-encompassing updating.At a minimum as a policy works its way through the system and develops the types of knowledge (and/or lessons) needed to progress it will change (as will the actors and institutions involved in the policies development).For instance, when ideas first transfer, the knowledge updating process may involve little more than an enlightenment function helping policymakers focus on a situation in their home system that is similar to a problem occurring elsewhere.As a result of awareness those who transferred the data might use it symbolically or as a way to frame a situation as a problem that was not initially viewed that way.In a similar way the information may be used in combination with other ideas to forward political goals or develop a strategic weapon in a political battle over the importance and definition of the newly discovered problem.Moving forward, the information used to finalise a policy solution is far more likely to be a combination of a number of different knowledge bases that have built up over the policy process.This is true for, as a policy develops, new actors and institutions come to the table, bringing in different collections of knowledge, interests, motivations and goals. 58All told, if something approaching second and third orders, change is to emerge from the transfer process it will likely result from the culmination of a range of learning experiences and processes involving the utilisation of both hard and soft lessons.
When and where an agent becomes involved in the transfer and policymaking (or implementing) process is also important for understanding what is transferred and how this information will eventually be utilised.For instance, when an agent becomes involved in the processes can say a lot about what motivation they have for using a lesson and the type of knowledge they offer (and the strategies they will have to employ in the use of this knowledge).In fact, outside instances where a policy is the result of a photocopy, a great deal of information is likely to be gathered, even if it not subsequently used.Unfortunately, due to the nature the policy process and the institutions involved, a great deal of this information will be "lost" or "held" until it is no longer relevant.For instance, many examples of transfer and knowledge underuse are due to a misfit between an existing indigenous situation and the situation in the originating system.This is true despite the fact that transferred knowledge can often be used to address more than one situation.Situations can also be similar but not recognised as such.Knowledge of a foreign system can be lost when an agent or institution with the information is not placed (or moves) in the policymaking process so that the knowledge cannot be accessed.An agent of transfer' s institutional role can also impact what knowledge they have access to and what they can do with this.By way of illustration, information held by a low level bureaucrat who is not in a position to forward it to those involved in policy development or implementation higher up in the institution might find it disappearing.The same information can find its way into the policy process and utilised in the development or implementation of a policy if it is taken-up and forwarded by and policy entrepreneur (or policy champion).In a similar way foreign ideas being held in limbo may emerge if a situation changes or a core actor who had been blocking the use of the information moves (or the one with the information moves to a more favourable position or institution).
Transferred knowledge not only interacts with the policymaking and implementation processes as a result of the power and position of actors within the policymaking system, it is also shaped by the institutional settings it finds itself embedded in (and the level of governance it is operating at).For instance, many actors find that institutional constraints interfere with their desire and ability to introduce data on foreign policies into the policy process.This can involve factors running from institutional tacit knowledge constraints (such as when administrators argue that German environmental solutions can't work here because our system is too different), to budget constraints prohibiting actors from looking to distant systems for ideas, to having a technician in charge of urban planning, who "wanted to transfer the green roof model but was blocked by his political superior". 59Looking at the same issue from the opposite end of the governing structure an actor sitting in a similar institutional setting at the international level will often have to rely on faith that the receiving system will understand the correct lessons, have the technocratic ability/skills needed to act upon the information, and that the political will to carry out any required alterations to the existing policy mix exists.Or as Sharma illustrates: "Decentralization is sometimes designed merely to receive loans from international agencies.The design of decentralization in such cases cannot be ex- Sharma, 2009, p.   [61] Bennett, 1991, p. 32.pected to aim to bring about systemic reforms". 60If any of these are missing, it may require too many new institutions and practices to be developed for a transfer to occur successfully.
All told, transferred information can be used by a range of actors, operating on different levels, holding different degrees of power, and having different institutional and structural capacities.This creates a situation where the same information may lead to a range of different outcomes as a result of different levels and types of knowledge emerging and being used.This is particularly true when information emerges from outside a political system or existing problem situation.Thus whether knowledge is used for instrumental, conceptual, or even organizational purposes will depend on a number of factors often overlooked by the existing literature.

CoNCLuSIoN
The literatures surrounding the transfer of information from one political system to another in its many guises has added to our understanding of the policy process; including those driving the globalisation of policy and governance.However, in this there is a considerable absence of what is learned and how this interacts with the transfer process (from the perspective of transmitters and receivers) and the policymaking and implementation processes (including how transfer fits into the evaluative and evidence based policymaking processes).By linking transfer to learning, knowledge updating, and knowledge utilization processes it will be considerably easier move beyond statements that policy x or idea y was transferred from system a to system b.While this might be a visible output it doesn't necessarily equate with what was learned.Rather while policy x may appear as an output of the policy process, the learning might have involved a study of the strategies and techniques used to pass a policy in the home system rather than the policy that emerged.Or it might have involved learning about the best ways to implement policy x once passed rather than x itself.A better understanding of learning and knowledge updating will also help overcome issues associated with spurious diffusion/ Galton' s Problem that few transfer studies (particularly large-n diffusion studies) address.
By considering what is learned (by whom), it should also be easier to explain why it often appears that "bad" or inappropriate models spread quickly while "good" or more suitable models are never transferred.Stated slightly differently: "Confusion tends to arise […] from the inference that evidence of the diffusion of a program is equivalent to diffusion of knowledge about that program". 61Rather to understand the linkages between learning, transfer and utilisation, it is nec- Bennett, 1991, p. 32, italics in original.
[63] Unalan, 2009, p. See also Campbell, 2002.essary to "distinguish clearly between knowledge of a foreign program, utilization of that knowledge, and the adoption of the same program". 62econd, the transfer literature would be greatly enhanced if it began to examine the factors underpinning the causes motivating pushers and borrowers to engage in transfer, learning and knowledge utilisation.As Unalan notes: "When studying policy transfer, one of the key factors that we need to understand is what drives actors to engage in the process, as these reasons can influence the whole process, including the outcomes, and the application of knowledge, beyond the initial selection of candidate 'lessons'". 63hile a considerable amount of literature has been written on the role of entrepreneurs and policy advocates, not nearly enough has been done to link their transfer activities to the overall policymaking process and how this interacts with knowledge utilisation.More importantly very little has been done to examine how those opposing the forced (or voluntary) importation of a lesson use the policy process to advance their own counter agendas or transform an undesirable use of transferred knowledge into something more suited to the indigenous environment.Or, how those opposed to a transferred policy can alter its effects through the implementation process.Finally, the role of national and international policy cycles in the spread of knowledge and thus, when and why international ideas become common knowledge and subsequently linked into national policymaking cycles need considerably more attention by those interested in transfer.
Third, a better understanding of the role the political plays in the transfer, knowledge updating and implementation processes needs to emerge.To date, the transfer literature has done little to understand or integrate how the politics surrounding the transfer of a policy or idea shape the overall use of data.Fewer transfer studies have attempted to analyse how the general political climate surrounding a transfer situation influences the transfer and knowledge utilisation processes.By integrating some of these ideas into the transfer process this it should be possible to begin understanding the role games play in the policy cycle.For instance, clearly there are situations where international agents actively work at bypassing national level actors and processes and direct their efforts at transferring data directly to the local.Similarly there are local agents who use lessons from (and actors floating around at) the international level in an attempt to block national actions and policies that might lead to negative (in a broad sense) local impacts.More to the point, the investigation of the political in the transfer process necessitates an integration of the ways actors promoting (or resisting) transfer operate in the confines imposed by the institutional, cultural, ideological and sociopolitical systems of the receiving political system.Without considering these types of influ-ences on the possible, political and personal how actors perceive a situation, even what they see as a valid model or idea, what might at first appear to be an illogical decision or policy may turn out to be perfectly logical in light of the political factors surrounding the transfer and policymaking situation.
Transfer studies must also start taking into account how information and knowledge is gained during the transfer processes and how this is subsequently altered and reformed as it works its way through the policy and implementation cycles.This will help shift the focus from the current "have-policy-will-travel model" to one that examines how a range competing transferred models develop and work their way into and through the policy process at the same time, competing and being combined into new models.As Manning argues, "we might surmise that policies will only fit depending on their place in policy sequences, dependent paths and slow processes […] a transferred policy might have quite different consequences depending on how, when and where it is adopted into a particular setting". 64inally, with some exceptions found in the translation literature, the role of social constructivism and discourse has not been properly integrated into the transfer literature; let alone integrated with how the discourse surrounding transferred information interacts with knowledge perceptions and updating.Noticeably absent is work examining the way in which some "ideas consist of taken-for-granted assumptions about values, attitudes, identities, and other 'collectively shared' expectations […] these […] lie in the background […]  but constrain action by limiting the range of alternatives that elites are likely to perceive as acceptable and legitimate". 65As a result "what gets transferred may well differ from what has been learned", what is learned may not be what was originally transmitted, and what is used may have little to do with either. 66This is because the "same communication will be interpreted and received differently by different individuals and organisations, the differences reflecting their different contexts, sensitivities and perspective". 67This is doubly true when one considers that much of the information relating to foreign models appears in snippets that tend to be filtered by agents and institutional settings in which the lesson becomes operationalized.All told, what one agent learns or thinks they have learned might be quite different from the agent they are sitting next to and from those who provided the lesson.Because of this, the focus on agents and their roles and understandings and use of information should take a more central role in future transfer studies.
Overall, those engaged in the transfer process face a range of obstacles that are likely to alter the way knowledge emerges from the information that was initially moved and then how this is subsequently used in the policy process.While some such as when actor is placed in a position to feed information directly into practice (such as a politician uses rhetoric they heard being used effectively in a foreign system) most instances of transfer involve many twists and turns.In this, much of the initial information that enters a system will be taken away, filtered and mixed before it is finally used.The filtering and mixing process will continue as the policy is moved through the political system, encountering and entering different organizations, and being adopted and adapted by different actors.Once a foreign idea or policy has entered the policy process it can be examined within a single organization or a range of organizations, simultaneously or sequentially.As a result, while a great deal has emerged from the existing literature with a degree of learning and adaptation the transfer concept will be able to provide many more years of useful analytic and conceptual study.

David
Dolowitz is director of Post Graduate Research at School of Histories, Languages and Cultures/Department of Politics/ University of Liverpool.