Learning through policy transfer? Reviewing a decade of scholarship for the field of transport

ABSTRACT

Attempts to pursue sustainable mobility face widespread challenges. One key way of approaching these challenges is through policy transfer and policy learning; indeed, the practice of learning from elsewhere is encouraged at various levels of government. This paper contends that a better understanding of what facilitates learning through policy transfer might support further change, yet such examinations remain underdeveloped in the field of transport. This paper synthesises key concepts and factors that drive this learning process, by reviewing 65 papers on transport policy published between 2011 and 2020. Our findings testify to the growing prevalence of policy transfer research and emerging critical perspectives on the transfer and translation of global ideas. We uncover critical factors of the learning process, including settings where learning takes place, inter-actor relations, and organisational and institutional patterns. While most papers reviewed here aimed to examine learning, few employ theories to measure the concept. Consequently, one of our main conclusions is that relatively little is known about how and to what extent learning, triggered by experiences from other contexts, is actually transformed into action. Suggestions include more systematically focusing on organisational and institutional dimensions and concerted trans-disciplinary efforts to close the gap between research and practice.

KEYWORDS:

• Policy transfer
• policy learning
• sustainable mobility
• transport
• learning
• critical drivers

Introduction

Attempts to pursue sustainable mobility agendas face a variety of challenges. One oft-promoted way of approaching these challenges is to “learn from abroad” (Dolowitz & Marsh, 1996). Indeed, there is a rich body of literature from the domain of political science that explores issues related to transferring, diffusing, or drawing lessons from policies enacted elsewhere (see Dolowitz & Marsh, 2000; Hall, 1993; Rose, 1993). Generally, these terms define processes in which “knowledge about policies, administrative arrangements, institutions and ideas in one political system (past or present) [are] used in the development of policies, administrative arrangements, institutions and ideas in another political system” (Dolowitz & Marsh, 2000, p. 344). The underlying principle is that since policy problems (such as congestion) are comparable across political and geographical boundaries, their solutions may also be capable of transcending a single, specific set of conditions (Dolowitz & Marsh, 2000).

Questions remain, however, about the role played by learning with regards to policy transfer, and the extent to which this learning influences policy outcomes, if at all (Bennett & Howlett, 1992; Benson & Jordan, 2011; Pojani, 2020). An important milestone in the literature on policy transfer as it relates to transport policy is Marsden and Stead’s (2011) review. In their conclusions, the authors find that although the transfer of policies and ideas does occur, and can in fact be influential in the adoption or implementation of transport policy, the role of learning remains overlooked, inadequately described and, therefore, poorly understood. If policy transfer is to be a valuable instrument for governments (Stead, 2016), the role played by learning must be further clarified (Marsden & Stead, 2011).

A decade on, the conclusions drawn by Marsden and Stead provide the starting point for this review: do they still stand, leaving the question of learning unaddressed, or has knowledge in the field evolved? The following two interconnected questions therefore guide this paper: How does the literature on transport policy transfer engage with the concept of learning? What key factors in the process of learning through policy transfer can we glean from the literature? Rather than gather further evidence to prove that policy transfer does indeed occur, we aim instead to (a) critically assess the degree to which transport literature engages with learning in the study of policy transfer, and (b) identify those factors that drive (or hamper) learning through policy transfer. By exploring drivers of, barriers to, and different concepts within learning, we hope to produce a fuller picture of learning processes, and in so doing, facilitate applications of such processes in future practice.

In order to unpack the current state of “learning from elsewhere” within transport policy, we review 65 empirical papers published between 2011 and 2020, using a codebook methodology adapted from a previous review (Gerlak, Heikkila, Smolinski, Huitema, & Armitage, 2018). The paper begins with a theoretical discussion before we lay out our methodological choices, and closes with a look at the implications of our analyses, as well as opportunities for future research.

Theories underpinning the practice of learning from elsewhere

The process of searching for, learning from, and spreading ideas and knowledge that originate outside a given body or institution are, of course, centuries old. Within the domain of political science alone, there is a long history of questioning both how learning occurs, and how it can improve the exercise of government or governance. While it is beyond the scope of this paper to fully detail this rich and vast body of knowledge, we outline here several strands which have brought focus to this review.¹

The topic of “learning from abroad” garnered significant academic interest in the last two decades of the twentieth century, when several terms emerged to describe the process. These included “policy transfer,” “policy diffusion,” “policy convergence,” and “lesson-drawing,” though there were others. One of the general assumptions undergirding this first wave of theoretical work was that because governments have limited capacity to solve complex problems, they must look to other governments that have attained a desirable outcome or implemented a particular policy instrument to meet a similar or related challenge (Rose, 1993).

A notable milestone in this era is the widely-cited policy transfer framework, as advanced by Dolowitz and Marsh (2000). The framework proposes a catalogue of concrete questions to define the process under review: why policies are transferred, who is involved, what is transferred, what sources are used, how effective the outcome is, and what barriers there may be. Though it offers a flexible heuristic (Dolowitz & Marsh, 2012), a group of scholars were critical of the framework’s rationalist underpinnings (see Evans & Davies, 1999; Stone, 2001), as well as its inability to examine or explain learning processes as they are involved in policymaking (McCann & Ward, 2012; Evans, 2009). This next wave of academic writing prioritised an emphasis on socio-political factors such as actor agency, power structures, and policy networks (see Meseguer, 2006; Radaelli, 1995; Stone, 2004).

Stemming from the “mobilities turn” in geography (see Sheller & Urry, 2006), this focus on policy mobilities pays particular attention to aspects of power, politics, and relationality in the study of policy (McCann & Ward, 2012; Peck & Theodore, 2010). This perspective positions learning as inextricably linked to “sites of encounter, persuasion, and motivation,” in which policy “elites” play a crucial role, both directing and invigorating ideas (Temenos & McCann, 2013, p. 346). One criticism of this approach, however, is its tendency “to downplay the importance of structures and institutions” (Dolowitz & Marsh, 2012, p. 343). Nevertheless, addressing the multifaceted process of learning in the spread and development of policy ideas calls for a variety of different theories and concepts, a point that is widely acknowledged.

The present review builds on this vast body of literature concerned with learning and its relation to policy, which continues to be of great interest within the domain of policy itself (Bennett & Howlett, 1992; Benson & Jordan, 2011). More specifically, we expand on Marsden and Stead’s (2011) work, to collate existing empirical evidence of how learning occurs not only through policy transfer, but also at “the sites of transfer” (Hudson & Kim, 2014, p. 496), a dynamic that remains underexplored (Moyson, Scholten, & Weible, 2017). On this topic, a preliminary contention is that empirical studies into the experience of learning for the actors themselves, as well as into the transfer of learning to wider circles, are largely undeveloped (Hudson & Kim, 2014; Marsden & Stead, 2011; Moyson et al., 2017). This review, therefore, sets out to address these gaps, by focusing on the learning process and exposing both its drivers and barriers.

Research methods

Research methods and inclusion criteria

While there is no consensus on the right way to conduct a literature review, our approach has been designed to follow the transparent, replicable steps recommended in the work of van Wee and Banister (2016). To begin, search words were devised and a set of explicit criteria developed. Our keywords included: “policy learning,” “policy tourism,” “policy transfer,” “policy diffusion,” and “policy mobilit*.” The output was then filtered for the keyword “transport*,” in order to exclude results that lay outside of the relevant field and cases. Then, to assess the nature of the progress made since Marsden and Stead’s (2011) review, the search was restricted once more to publications from between 2011 and 2020. Other inclusion criteria were simultaneously applied: language (English) and article type (peer-reviewed journal or book chapter). The search was conducted across three databases (Google Scholar, SCOPUS and Web of Science).² The results were saved in a spreadsheet, and the output from all databases was manually combined and duplicates removed.

To ensure the relevance of the remaining literature to the research questions, the articles and chapters identified thus far were submitted to one final round of screening, to confirm that they (a) demonstrate an analytical focus on the notion of policy transfer and related concepts; (b) include primary empirical data, obtained via qualitative and/or quantitative methods (i.e. not a commentary, literature review, or solely theoretical article); and (c) exhibit a primary focus on transport policy. This allowed us to exclude several papers which, for example, focused primarily on theoretical and conceptual developments, or examined a variety of policy areas (sometimes outside of the realm of transport altogether), or else compared and assessed the differences between multiple contexts (such as user behaviours or mode choice), as opposed to focusing specifically on the transfer and movement of policies.

Data analysis

To assess and synthesise the literature, a codebook methodology and coding instructions were adapted from Gerlak et al. (2018) and transformed into a tool fit for the research at hand (see Appendix A). The advantage of this method is that it allowed a team of researchers to code the papers, as well as systematically compare results across the sample, thereby increasing the validity of the findings. The codebook assessment covered 45 fields, including (a) research question(s), theoretical framework(s), definitions, and conclusions; (b) methodological details; and (c) the drivers of and barriers to learning. Items were assessed in a spreadsheet using yes/no questions, numerical coding, and qualitative fields (i.e. text citations, page numbers, and other assessments). Additional decision rules were applied to certain fields, so as to ensure consistency in the coding. For example, the fields for theoretical frameworks and drivers/barriers were transcribed from the papers into the codebook; when a given column was complete, the data were systematised into distinct codes through an inductive process. In the interests of internal validity, the table cross-checked once it had been completed by two researchers, and an additional researcher was randomly assigned 10 articles to code for further triangulation. For codebook questions yielding numerical values, the results were given statistical analysis (see Appendix B).

Results

The final round of screening produced 65 articles for analysis (see Appendix C for the list of included articles). In what follows, we begin by providing an overview of the literature sample (see Tables 3.1–3.2 and Appendix for details), and then proceed to address our research questions.

Research landscape overview

Scholarly attention

With notable summits in 2015 and 2020, it is abundantly clear that policy transfer and “learning from abroad” continues to be a subject of widespread scholarly interest. The 65 papers analysed here were published in 40 different peer-review journals (and two books) and 60% of the articles were featured in journals with a broader focus than transportation, covering the disciplines of policy, geography, urban planning, and urban studies. Particularly prevalent transport policies investigated in the articles include: cycling policy, integrating transportation and land use (that is, transit-oriented development, or TOD), adapting transport infrastructure, and bus rapid transit (BRT).³

Dominant theories & emerging critique

Table 3.1 provides a summary of the many theories represented in our corpus. Roughly half of the papers (32) make use of the policy transfer framework. Most use Dolowitz and Marsh’s (2000) framework as an operational tool to analyse, synthesise, or assess policy change, though rarely in its original form and rather as a starting point. The remainder of the sample is split between: (a) policy learning and derivatives (such as policy-oriented learning and lesson-drawing); (b) perspectives on policy mobilities; and (c) other social science theories, such as socio-technical transitions, cultural symbolism, discourse framing, education theory, and planned behaviour.

We observe a noticeable recent trend in the critiques of learning and transfer processes, which mostly derive from the literature on policy mobilities emanating from the Global South (see Montero, 2020; Si et al., 2020; Silva Ardila, 2020, references provided in online Appendix C). Several papers problematise the notion of “best practice,” considering it to be laden with non-neutral political power and concealing highly contextual factors, such as cultural norms, planning legislation, financial arrangements, and so on (Blake et al., 2020; Pojani & Stead, 2015; Wood, 2020, references provided in online Appendix C). A few papers express a sceptical attitude towards global think tanks involved in policy transfer, emphasising “the rising influence of [philanthropic]” organisations (Montero, 2020, p. 2277; Whitney et al., 2020, references provided in online Appendix C). The motivations of these organisations are queried, as they “intervene in the political realm by helping place particular topics and policy frames in the local and global agenda” (Montero, 2020, p. 2276, references provided in online Appendix C). Certainly, most of the papers in the sample pay heed to glaring imbalances in power and financial resources, as can be seen especially often in transfer processes between North and South (or developed countries and developing ones).

Methodological choices

Qualitative methods, and particularly single-issue case studies, dominate this area of research. Most of the papers in the sample (48) examined instances of policy transfer and policy learning ex-post, or retrospectively, and 11 articles conducted ex-ante, or hypothetical, research. In all cases, interviews are the preferred method of data collection, and sample sizes varied from one participant to “over one hundred” (Wood, 2014b, references provided in online Appendix C). The incorporation of multiple sources of data is also common, such as participant observation in addition to interviews; indeed, 33 papers draw from at least two types of data, and several use multi-sited ethnographic methods.

Geographically speaking, the sample is rather diverse, with roughly one third of papers focusing on cases within Europe, and a quarter representing transcontinental cases of “in-bound” transfer, knowledge assimilation, or the proliferation of policies either within or between countries or continents. While one-third of the papers (19) discuss the Global South, there is a noticeable recent increase in this area of research. Just over half of papers examine cases within one city or between cities or regions; 15 investigate cases relative to nation states. Finally, a number of papers assess the effect of policy transfer on a particular site (for example, an infrastructure project).

Most significantly, there was a wide range of methodological clarity, especially among the papers that use interview methods (46 papers). Five articles were coded as thoroughly transparent (Glaser et al., 2020; Hysing & Isaksson, 2015; Macmillen & Stead, 2014; Pojani et al., 2017; Pojani & Stead, 2015, references provided in online Appendix C), and included details such as recruitment of respondents, response rate, question schedule or protocol, interview location/duration, transcription, and coding procedures. Most papers, however, omit or offer a very limited description of methodological choices.

How does the literature engage with the concept of learning?

Learning as a research aim, variable, or outcome

All but five articles in the sample discuss learning, and the word often appears synonymous with the terms “transfer,” “innovation,” and “change.” Regardless of their primary theoretical framework, over a third of the articles explicitly reference learning within their research question or aim. Of these, many are interested in uncovering factors that may influence learning or barriers that might impede it, including the question of how to learn from failure. For example, Macmillen and Stead (2014, references provided in online Appendix C) offer a critique of “best practice” as “the de facto approach” to facilitating learning, an argument also found in Lawer (2019b, references provided in online Appendix C).

Learning is considered a variable, an outcome, or an implication in more than half of the papers (35), a trend that demonstrates quite how ambiguous the concept is. The following, for example, frames learning as a dependent variable: “Failures can be important to the learning process … In this article, we propose some conditions that can prevent governments from learning from policy failure” (Newman & Bird, 2017, p. 71, references provided in online Appendix C). Elsewhere, many articles found that learning was an implication of their research. May (2015, references provided in online Appendix C) suggests that city organisations would benefit from adopting “better learning cultures,” in order to be more receptive to outside ideas and policies, and therefore become better agents of policy transfer (p. 10). It was also frequently noted that learning is not a neutral activity; it can in fact be used in contradictory and problematic ways to, variously, accommodate political needs, legitimise power, and justify economic growth (see Haughton & Mcmanus, 2012; Lawer, 2019b, references provided in online Appendix C).

A variety of concepts of learning

A range of theories of learning are cited in roughly two thirds of the papers (43), however vast their differences in scope and specificity. Ten distinct types of learning were identified in the sample, with 16 papers referencing more than one type of learning (Table 3.3). Of those papers which do not explicitly identify learning as either a research goal, key variable, implication, or outcome, a sizeable number (13) do nevertheless associate their work with a specific type of learning. It should be noted, however, that few apply these concepts theoretically or empirically, such as through operationalisations. Indeed, only three papers provide a clear definition of learning. One such example is in Ma’s (2017, references provided in online Appendix C) examination of the influence of site visits on the transfer and implementation of bike-share programmes, where policy learning is defined as: “⁠either the direct adoption of concrete policies or indirect heuristics of ideas, and both approaches benefit policy design and implementation” (p. 584).

Policy learning was the single most cited concept related to learning, appearing in 34 papers. While no substantive pattern emerged for its use, origin, or associations, though it is rarely defined. In many papers, the term “policy learning” seems to be used synonymously with “transfer,” “transferability,” “adaptation,” and “knowledge exchange.” Given the relative dearth of conceptual analysis, many of the papers in the sample merely imply that policy learning equates to acquiring new information; the processes of applying or assimilating that information are rarely further examined.

Organisational learning and social learning also appear regularly. For example, Argyris and Schön’s (1978) theory of single and double-loop learning is called upon in an exploration of bio-fuel targets (Anderton & Palmer, 2015, references provided in online Appendix C) and, elsewhere, of knowledge transfer (Glaser et al., 2021, references provided in online Appendix C). Meanwhile, Marsden et al. (2012, references provided in online Appendix C) discuss the institutional constraints and organisational behaviours that are of major importance to any process of actor learning.

How is learning measured?

Although learning is a primary concept in half of the papers, only eight articles explicitly aim to measure learning and/or demonstrate what the outcomes of learning can be. Montero (2017b, references provided in online Appendix C), for example, examines the role that study tours play in policy transfer in order to “understand … how these learning practices result in policy change” (p. 333). One aspect that remains unclear, despite the various illuminations of processes and fora for learning within the sample, are explicit measurements of learning. Thomas and Bertolini (2015, references provided in online Appendix C), however, attempt such a measurement, by asking transport professionals working on TOD implementation to complete two surveys, one before and one after a workshop. Taking a more archival approach, Ben-Zadok (2018, references provided in online Appendix C) analyses government documents to chart the pace and extent of policy learning in smart-growth legislation. A final example is Glaser et al. (2021, references provided in online Appendix C), who borrow survey metrics from human resources in order to measure the knowledge transfer that occurs after study visits.

What factors influence learning?

A majority of the papers (52) indicate and discuss mediating factors. In some cases, these are ascribed to learning, and in others, to policy transfer or adoption (or the failure thereof), while many others either blend these concepts or make no distinction between them. Additionally, and depending on the context, certain factors are often identified as both facilitating a process and hampering it (a number of individual articles identify and draw out this dynamic). Therefore, the analysis presented here includes insights into drivers and barriers that have been identified as influential in any process of change, whether or not the article’s author(s) specifically cite learning or transfer.

Learning settings

The setting where learning takes place is by far the most frequently identified moderating factor, and is considered a driver in the learning process by 40 articles (in no case is the setting deemed a barrier). Learning settings are typically characterised by group activities where social interaction, whether informal or formal and often face-to-face, is encouraged if not required. Indeed, many papers report that these activities are not isolated episodes of learning but that they occur repeatedly, as part of a longer-term project, collaboration, or strategic exchange. Variables such as the specific activities, experiences, dialogue, and interactions that take place in learning settings (along with their relative effects), however, are rarely the subject of empirical exploration or analysis, despite being recognised as influential. Below, three of the most common learning settings are synthesised.

Networks, such as knowledge or policy networks, are the most commonly discussed setting and potential driver of learning. Most papers do not undertake an empirical investigation of how or why such networks form, or how they contribute to learning. Rather, networks are taken to “play a role in the spread of policy models” (O’Dolan & Rye, 2012, references provided in online Appendix C), or in other words, to facilitate the process of transfer (Bray et al., 2011, references provided in online Appendix C). Some papers, however, offer deeper insights into the scale and power of networks. On one hand, larger networks operating at the national level not only disseminate ideas (this is echoed by Lawer, 2019b; Ma, 2017; Parkes et al., 2013; Timms, 2011, references provided in online Appendix C), but also “shape and structure knowledge flows” (Sengers & Raven, 2015, p. 170, references provided in online Appendix C). On the other hand, smaller and more local networks consisting of “closer-knit policy actors” (Macmillen & Stead, 2014, p. 82, references provided in online Appendix C) can contribute to building legitimacy for certain policy models (Macmillen & Stead, 2014; Thomas & Bertolini, 2015, references provided in online Appendix C), especially when “key actors” within the network have attained a certain level of influence. In all cases, “trusted” and “personal” connections are foundational, regardless of scale (Marsden et al., 2010; Marsden et al., 2012; Montero, 2017b; Wood, 2014b, references provided in online Appendix C). Rarely examined or discussed, meanwhile, are the potentially perverse effects of networks, though some scholars point to issues of inclusion and exclusion (Sengers & Raven, 2015, references provided in online Appendix C), and the resistance of some networks to being mobilised has been identified as a structural barrier to policy implementation (Pojani & Stead, 2015, references provided in online Appendix C).

Organised group travel or study visits are seen to advance (group) processes of learning, transfer, or policy implementation. Wood (2014a, references provided in online Appendix C) describes study tours as a “common practice in which local actors travel elsewhere to see innovation and meet with those in the exporting locality who implemented it” (Wood, 2014a, pp. 2655–2656, references provided in online Appendix C). For some researchers, the geographical destination of the visits is particularly meaningful, and this is mostly due to historical successes with various policies. Examples of destinations associated with successful policy implementations include the Netherlands for airport master planning (Bok, 2015, references provided in online Appendix C) and cycling policies (Glaser et al., 2021; Pojani & Stead, 2015, references provided in online Appendix C), Bogota for BRT (Montero, 2020, references provided in online Appendix C), Seoul for highway infrastructure removal (Farmer & Perl, 2020, references provided in online Appendix C), and London for congestion charging (Attard & Enoch, 2011, references provided in online Appendix C). Several papers discuss the participation of certain “elite” local actors (Glaser et al., 2021; Ma, 2017; Montero, 2017b, references provided in online Appendix C), who are keen to “personally observe and experience” programmes (Ma, 2017, p. 584, references provided in online Appendix C). Others delineate the effects of study visits, where such can be said to have variously “improved relationships” (Wood, 2015a, p. 217, references provided in online Appendix C), “helped create consensus and influence public opinion” (Attard & Enoch, 2011, p. 550, references provided in online Appendix C), or facilitated dialogue and trust among actors (Glaser et al., 2021, references provided in online Appendix C).

Conferences and workshops also appear as drivers for learning. Conferences are cited as influential venues that facilitate the building of “trust, relationships, and collaboration between actors” (Montero, 2017b, p. 63, references provided in online Appendix C), and are also investigated as a setting in which to gain experiential knowledge (i.e. the case of cycling infrastructure in Glaser & te Brömmelstroet, 2020, references provided in online Appendix C). Workshops are sometimes used as a forum in which to carry out research into learning (Thomas et al., 2018, references provided in online Appendix C), or else take the form of a “serious game” for the simulation of TOD implementation (Duffhues et al., 2014, references provided in online Appendix C). However, as with study visits, these settings are rarely the subject of empirical study.

Inter-actor relations

Inter-actor relations were identified as both a driver and barrier to processes of learning through policy transfer, indicating a highly instrumental mechanism. Most articles referred to relations between ‘borrowing’ actors, those who were learning from elsewhere, with only one paper examining structural forms of collaboration between in-bound and out-bound parties (see Si et al., 2020, references provided in online Appendix C). Merely pointing to a relation is not always the end of a piece of research, however; certain qualities of relations are also highlighted. The degree of cooperation between local actors, along with previous collaboration experience (Hysing & Isaksson, 2015, references provided in online Appendix C), is considered a “mediating factor” (Rye et al., 2011, p. 541, references provided in online Appendix C), one that is capable, in turn, of influencing the degree of stakeholder participation (Kaszubowski et al., 2018; O'Dolan & Rye, 2012, references provided in online Appendix C). Here, the notion of trust was particularly prevalent (see Hysing & Isaksson, 2015; Montero, 2017a; Glaser et al., 2020, references provided in online Appendix C), though others included good communication (Thomas et al., 2018, p. 1211, references provided in online Appendix C), dialogue (i.e. Glavic et al., 2017; Kaszubowski et al., 2018, references provided in online Appendix C), and shared values (Rye et al., 2011; Walker, 2019, references provided in online Appendix C). Finally, cooperation between policy actors also surfaced in relation to other organisational and institutional characteristics, including leadership (described below).

Conversely, interaction deemed negative, of poor quality, or infrequent represents a mediating factor of some importance. For example, in a case-comparison study on congestion charging, Hysing and Isaksson (2015) describe how, despite a “crisis” of “internal resistance” faced by the governing coalition that their research focuses on, and a “continuous challenge to agree on the details” of the policy, the coalition was finally able to reach a compromise through additional measures unrelated to the original scheme (p. 57). As a result, the authors conclude from interview data that the coalition “came out stronger,” as internal communication and trust were improved (p. 57). In other cases, however, cooperation or agreement between actors proved impossible. In their “serious game” simulation of TOD in the Dutch context, Duffhues et al. (2014, references provided in online Appendix C) found that the game actually revealed barriers which “cannot be overcome easily” (p. 17), such as “a lack of incentives for actors to cooperate at the corridor or regional level” (p. 4). The authors attribute these constraints to individual perceptions, and lack of coordination (though this is not described in any depth).

Organisational factors

Organisational characteristics, including structure, capacity, and resources, emerged as a significant factor for learning, more often as a barrier than a driver. Canitez (2020, references provided in online Appendix C), for example, concludes that the fragmentation of public agencies in Turkey has been a contributing factor in the failure of attempts at policy transfer. In a similar vein, Marsden, Frick, and May (2013, references provided in online Appendix C) suggest that an organisation’s learning culture, especially when characterised by risk aversion, may impact that organisation’s ability to access information, exchange with “trusted peers,” and build networks. May (2015, references provided in online Appendix C) concurs, and in his research into the implementation of Sustainable Urban Mobility Plans, he calls on researchers to “tackle this issue” by “stimulating interactive learning” (p. 10).

Relatedly, and perhaps unsurprisingly, six other papers identify a lack of resources as a major barrier to learning. The capacity and resources of both humans and organisations are vulnerable to myriad constraints, including pressures on time and finances (Pojani & Stead, 2015; Werland, 2020, references provided in online Appendix C), a lack of structures or systems for learning (May, 2015; Glaser et al., 2021, references provided in online Appendix C), a shortage of readily available high-quality information (Blake et al., 2021; Duffhues et al., 2014; May, 2015, references provided in online Appendix C), unequal systems for rewarding learning (Newman & Bird, 2017, references provided in online Appendix C), and limited municipal staff expertise (Fink, 2019, references provided in online Appendix C). Several papers maintain that organisations without leadership capable of encouraging learning are likely to suffer from reduced learning in the future (Newman & Bird, 2017, references provided in online Appendix C), as well as diminished partnership building (Thomas & Deakin, 2017, references provided in online Appendix C), and poorer policy implementation (Kaszubowski et al., 2018, references provided in online Appendix C).

Institutional factors

Finally, institutional factors are found to influence the processes of learning through policy transfer. As a barrier, the cultural values of those involved in policy learning processes “impact the acceptability of new ideas” (Rye et al., 2011, p. 541, references provided in online Appendix C), and “should not be downplayed in policy transfer attempts” (Canitez, 2020, p. 3, references provided in online Appendix C). This finding is echoed by others (see Gray et al., 2017; Hysing & Isaksson, 2015; Ibsen & Olesen, 2018, references provided in online Appendix C). In an attempt to measure this, Ashmore (2018, references provided in online Appendix C) concludes that the strong symbolism, found in many countries, of cultures where “automobility” is dominant may influence the mindset and operations of policymakers, and therefore presents an obstacle to importing sustainable policies from elsewhere. Additionally, eight articles highlight the framing of and discourse around policy problems as a driver for learning. Anderton et al. (2015, references provided in online Appendix C), for example, describe the “ability to create or […] reframe” the narrative as a “critical source of political power,” and a “critical driver of policy learning” (p. 140).

Leadership and “mobilisers”

A factor with particularly hazy borders was leadership, identified in some places as a driver and less frequently as a barrier, but in all cases difficult to categorise firmly as either an institutional or organisational factor. Leadership usually refers to those individuals based in the importing party who hold positions of power, or have a particular ability to influence others. Examples of such figures in the sample include “an ‘innovation champion’” (Parkes et al., 2013, p. 17, references provided in online Appendix C), senior management support (Glaser et al., 2021; Marsden et al., 2012, references provided in online Appendix C), and “clear leadership from politicians with vision and credible pathways to achieve the vision” (Gray et al., 2017, p. 233, references provided in online Appendix C), which in all cases accelerated learning processes for those in the borrowing position. In their case study of two different partnerships, Thomas and Deakin (2017, references provided in online Appendix C) observe that “strong and engaged leadership” led to a successful outcome in one of the partnerships (p. 45), whereas in the other, ambiguous, disengaged, and unsupportive leadership presented a barrier.

Several papers also critically examine the role of “policy mobilisers”– individuals who might be citizen experts or international policy entrepreneurs, and who sometimes represent consultancies, think tanks, or philanthropic organisations others (Montero, 2017; Parkes et al., 2013, references provided in online Appendix C). These individuals have the “capacity to strengthen the narrative” around certain policies and act as “the instrument to communicate and promote the benefits” of them (Silva Ardila, 2020, p. 82, references provided in online Appendix C).

Limitations

Although every effort was made to design and conduct this literature review soundly, it is not without limitations. Aggregating and comparing results across three academic databases did have the benefit of minimising exclusions, however we acknowledge that the search terms, time limits, and inclusion criteria used may have overlooked relevant publications. As such, this design bounds the search, and thus necessarily the results. For example, the role of individual actors in the process of policy learning is poorly represented (see Olsson & Hysing, 2012), as is advocacy, along with the role played by activist and civil society groups in facilitating policy learning.

For the analysis of the articles, adapting and using the codebook methodology was also not without its challenges and potential limitations. For some codebook fields, a team of three researchers relied on deductive and narrative synthesis; however, the team was able to reach consensus on all fields through slight modifications of the coding instructions, and discussion. Nevertheless, many of these issues are inherent to any bounded literature review, and the strength and scope of the findings suggest that the process and search protocol were not too great a hindrance. Another important consideration here, one borne out by our analysis, is that a degree of theoretical saturation (Lewis-Beck, Bryman, & Futing Liao, 2003) appears to have been reached, since the main concepts were often repeated across the sample without the addition of any new ones. Despite these drawbacks, we leave this methodological framework open for future interrogation.

Discussion: the future of policy transfer research

Our analysis from 65 empirical articles reveals that learning from elsewhere continues to be common in contemporary transport planning research and practice. By no means, however, is it a fixed practice itself. Quite the opposite, in fact; learning from elsewhere encompasses a host of activities, actors, geographies, contexts, and other dimensions. Clearly, those involved in transport decision-making do look elsewhere for salient and potentially transferable examples, but only in rare cases do we find that same learning linked to specific action regarding policy outcomes (such as adoption or implementation). A major caveat of this finding is that few papers empirically examine the learning process, even though such an examination is often stated as a research objective.

It therefore appears that the conclusions of Marsden and Stead’s (2011) review do remain valid: we continue to know very little about how learning occurs in the context of policy transfer, and still less about how – and to what extent – learning is transformed into local action, triggered by experiences in other contexts. This is not to suggest that learning does or does not occur; rather, we suggest that extracting knowledge from policy transfer activities and applying it elsewhere may require more concerted efforts, both in the practice of these activities, and in researching them. In what follows, we identify potential avenues for future research to unpack these questions (a summary of dimensions and potential indicators can be found in the concluding section, Table 3.4).

A system of factors to examine learning

The literature studied here indicates that many scales of influence are relevant to the process of learning from elsewhere: the settings in which the learning takes place, the inter-personal or inter-actor dynamic, and organisational and institutional factors. This in turn invites us to explore how the practice of policy transfer might be leveraged to facilitate both the policymaking process and the application of knowledge, especially where those in the position of “learner” or “borrower” are concerned.

The sample suggests that the real value of learning from elsewhere might lie in the settings in which learning takes place (such as networks, conferences, study visits), and most notably the “site of transfer” (Hudson & Kim, 2014), the umbrella term for the variables of place, travel, time, and social interaction. These settings offer a promising venue for close studies of interactions and learning processes in situ. Indeed, policy scholars have long argued that experiencing a policy in action provides unique insights (May, 1992; Rose, 1993), but only limited research has been carried out into the many variables that characterise these events and experiences, as well as their effects. A number of papers in the sample suggest that network activities and study visits do not take the form of standalone events, but rather unfold continually, or in a successional manner. Therefore, we see a potential to embed these activities within longer-term processes of collective learning, with well-known and recent conceptual contributions regarding, for example, capacity building (Foster-Fishman, Berkowitz, Lounsbury, Jacobson, & Allen, 2001), mentorship (Gagliardi et al., 2009), participatory governance (Newig, Jager, Kochskämper, & Challies, 2019), institutional learning (Steele, 2011), and sustainability transitions (Van Poeck, Östman, & Block, 2020). Applying insights from these specific discussions within the literature would, then, help to frame learning as a non-linear process of discovery and debate.

The review also finds that examining an organisation or institution’s capacity for learning through policy transfer is an increasingly relevant focus. Our findings expand upon the work of others (Evans & Davies, 1999; Spaans & Louw, 2009; Stone, 2012; Wolman & Page, 2002), and concur that institutional factors ought to be addressed by policy transfer projects more systematically. Here, leadership, communications systems, and organisational culture emerge as factors that may either facilitate or hamper the process of learning through policy transfer. Many of the authors in the sample are broadly aware of these forces, however only a minority brave the challenge of exploring them empirically. Useful theoretical works on institutional learning include Pahl-Wostl’s (2009) framework for analysing adaptive capacity, along with Schmitt and Wiechmann’s (2019) perspectives on how governance theory can inform planning (see pages 27–28 for institutional learning in governance networks). Their ultimate recommendations are to expand the debate on cross-sectoral coordination, and to improve the understanding of institutional flexibility within governance systems (p. 29). An interesting question here might be whether local institutional barriers are built within or between organisations (that is, the relative impacts of organisational culture), or whether they derive from wider institutional settings, such as rules and norms in the transport sector.

Expanding our methodological portfolio to study learning from elsewhere

Many papers in this review take learning as a central concern, but neglect to scrutinise the term and its myriad applications through, for example, definitions, operationalisations, or existing seminal work in the social sciences. Instead, the papers in this review often present learning and transfer as an entangled pair, thereby obscuring distinctions between the two concepts. Without a clear idea of what we are measuring, “how do we know it when we see it?” (Gerring, 2001, p. 43). To better understand how learning from elsewhere is linked to policy outcomes, and under what conditions, it may be advisable to concentrate instead on a single type, or phase, of learning. Examples from the policy studies literature are valuable here (Dunlop, 2009, 2015; Heikkila & Gerlak, 2013; van Doren, Driessen, Runhaar, & Giezen, 2020), and in one instance, specific to transport policy (Glaser, te Brömmelstroet, & Bertolini, 2019).

Qualitative methods, and especially interviewing, dominate research into transport policy learning and policy transfer, which as a result stands in marked contrast to research into transport policy analysis, where quantitative methods dominate (Marsden & Reardon, 2017). As such, many papers in this review conducted small samples of cross-sectional, retrospective interviews. Seminal learning scholars have long questioned humans’ ability to retrospectively describe and give details about lessons learned, because experiences of learning accumulate as tacit knowledge which cannot be codified (Duguid, 2005; Polanyi, 1966). When knowledge or skills have reached “expert” levels, it becomes particularly difficult to identify the increasingly incremental gains in these same knowledge and skills sets (Dreyfus & Dreyfus, 1986). Therefore, experimental approaches, pre-test/post-test evaluation (see Rouwette, Vennix, & Van Mullekom, 2002), longitudinal process-tracking analysis, institutional mapping (Aligica, 2006), and in situ ethnographic studies, might offer a valuable approach to countering the challenges around self-reporting. Emphasis is needed, however, on enhancing the replicability of interview and document analysis (i.e. informant description and recruitment, protocol or question design, and coding).

Towards bridging a divide

This review observes a divide in how the subject of learning from elsewhere is studied. On one side of the divide, a faction frames and structures its research within the traditional “policy transfer” approach. The other group, meanwhile, trains their focus on process, relations, and context, making use of a wide variety of theories and engaging in a critique of their subject matter. While a majority of the former are published in transport journals, only a minority of the latter share the same terrain, demonstrating a divergence in particular approaches not only to policy transfer but also to learning. We see here a potential for these perspectives to merge in some way, and possibly augment or complement one another.

Our findings also testify to emerging critical perspectives on the learning and transfer of global ideas. One aspect of this is a critique of the reductionist “best practices” approach, as well as of the international philanthropic bodies and think tanks that promote best practices and intervene in issues of local or municipal transport. Many observe power imbalances and contextual differences in processes of learning and transfer from North to South (or developed to developing). While critical perspectives do make a valuable contribution, alternatives or paths forward tend not to be provided. One interpretation of such critical perspectives is to see them as being more concerned with pursuing institutional approaches to the learning and translation of policies. As described above, incorporating theories, methods and tools from areas such as mentorship, capacity building, governance, and institutional studies could be of use here.

One way to bridge this divide in the scholarship would be to encourage more structural collaborations and feedback loops between research and practice (Straatemeier, Bertolini, te Brömmelstroet, & Hoetjes, 2010), as well as between institutional partners at different levels, including philanthropy and think tanks, and the cities or regions that they serve. Such an approach might address the challenges discussed here, because it would bring together the complexities and autonomies of the various parties, especially on the borrowing side. Structural learning collaborations, then, might just be capable of empowering those on the receiving side, thereby grounding policy transfer in discovery and debate, as opposed to a discrete trigger for change, or a copy-and-paste operation.

Conclusions

Policy transfer is a somewhat serendipitous process, dependent on particular actors – those with influence, but also those with enthusiasm, openness to outside ideas and language skills – actually finding out about a new idea. (Rye et al., 2011, p. 541, references provided in online Appendix C)

With a decade of evidence taken from 65 studies, the present paper contributes to the growing body of knowledge in transport policy by (1) updating the empirical database within studies of transport policy transfer; and (2) examining how learning is conceptualised and measured, paying particular attention to drivers of and barriers to the process. The aim here is to glean common conceptualisations, drivers, and barriers as they relate to “learning from elsewhere,” in order to synthesise patterns and provide insights for future research.

The findings from this review highlight that learning from elsewhere is a widespread practice that continues to grow as a subject of global interest, and in which the Global South is increasingly represented. The battery of policy topics identified in this review reflect those needed in order to move towards sustainable mobility, such as cycling, transit, transportation and land use integration (TOD), and adapting (surface) transport infrastructure (Banister, 2008; Holden, Gilpin, & Banister, 2019).

Despite this development, our analysis finds that we still do not know exactly what learning from elsewhere can achieve, nor the extent of its potential to improve the policy process. In particular, very little is known about just how, and to what extent, learning is transformed into local action, triggered by experiences from other contexts. This does not suggest that learning does not occur, but rather that applying knowledge from exercises in policy transfer will not happen automatically and may require more concerted efforts. A key reason for this, as our analysis found, is that learning itself, as an operationalised concept, needs to be brought further into the centre of discussions. Indeed, a lack of conceptual clarity around learning has been noted in other recent reviews (see Gerlak et al., 2018; Van Poeck et al., 2020). Rye et al. (2011, references provided in online Appendix C) point to the particular challenge posed by the concept, as highlighted in the quote at the beginning of this section: how does one account for so many unique variables within a broad concept of learning? In what follows we summarise three avenues along which future research might better understand learning through policy transfer in the domain of transport. These suggestions are augmented by Table 3.4 which summarises dimensions and outlines indicators of learning factors identified in the review.

Firstly, analysis suggests that broadening the scope of theories of learning, while also demarcating clear boundaries between learning and transfer, is necessary. Learning processes could be better measured when examining types or phases of learning, as well as their conditions and settings. Here, methodological clarity in any definitions or uses of concepts would go some way to generating more transparent outcomes. Many seminal works exist that could function as examples and prove highly useful, from fields such as organisational learning and education and policy studies (see van den Bergh, van Leeuwen, Oosterhuis, Rietveld, & Verhoef, 2007; van Doren et al., 2020; von Schönfeld, Tan, Wiekens, Salet, & Janssen-Jansen, 2019).

Secondly, opportunities abound for exploring the organisational and institutional dimensions that may either facilitate or hamper the “on the ground” changes connected to exercises in policy transfer; this is a gap that others have also noted (Moyson et al., 2017; Wolman & Page, 2002). Indicators such as previous experiences with policy transfer and other rolling institutional dynamics might hold the key to unlocking potential underlying factors (see Table 3.4). Third, and finally, there have been calls to engage more actively “with real world policy” and “realities on the ground” (Marsden & Reardon, 2017, p. 245). To close the gap between research and practice, therefore, we propose that more assiduous trans-disciplinary engagement is required. Not only will this help to empirically unpack and measure learning processes but might augment, and perhaps even accelerate, breakthroughs in both the understanding and practice of sustainable transport policy change.

Acknowledgements

The authors would like to thank the anonymous reviewers for their helpful suggestions in previous versions of this publication. M. G. conceptualised the review, developed the methodology and protocol, curated and validated data, and supervised the investigation. O. B., C. E., and M. G. analysed included articles. M. G. with support of O. B. prepared the first draft of the article. L. B. and M. t. B supervised and reviewed subsequent drafts. M. G. prepared subsequent drafts and final version of the publication.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 For example, see seminal contributions from Hall (1993), Heclo (1974), and Sabatier (1988). Useful summaries include Dunlop and Radaelli (2013, pp. 599–601) and Grin and Loeber (2006).

2 The search was conducted on March 3, 2021.

3 Five of these articles are attributed to one author (Wood, 2014a, 2014b, 2015a, 2015b, 2015c, references provided in online Appendix C). For the entire list of references included in the review, see Appendix C.