Exploring the mechanics of city-to-city learning in urban strategic planning: Insights from southern Africa

Abstract The importance of city-to-city learning processes in urban development has been gaining traction over the last two decades, particularly in the global North. Little empirical work has been done however, to critically analyse exactly how knowledge is transferred and the conditions under which these happen in order to facilitate better policy transfer between southern cities and regions. This paper focuses on the experiences of urban planning practitioners and key stakeholders from cities in South Africa, Malawi and Namibia that participated in a mentorship program coordinated by the international United Cities and Local Governments (UCLG). Using this case study, it develops a staged learning model that unpacks in detail, the mechanics of how the learning occurs during the program. In order to inform good practice, key ingredients that contribute to knowledge transfer – in this instance around urban strategic planning – is distilled. The study employed a qualitative approach combining observation, focus group and in-depth interviews with 18 practitioners directly involved in the program. The study shows that effective knowledge transfer requires inter-alia an expertly-facilitated and highly structured program of mutual learning. Most importantly however, the paper argues that the building of trust between the practitioners involved in city-to-city learning is critical for transformative action in cities of the global South.


Introduction
The seriousness of the consequences of inadequate urban planning and effective governance in developing economies has been flagged by the World Economic Forum (WEF, 2015, p. 35) identifying it as a key global risk in their global risks report that is supposedly "threatening the sustainability of urban development". What is of particular interest in their analysis, however, is the observation that given the extreme pressures of urbanization, the governments of such rapidly growing cities make little time for "adjustment and learning" (WEF, 2015, p. 35). Whilst this may be the case, there is a strong counter-narrative emphasising the trend towards an active transfer of knowledge on urban planning and development between cities and regions (Calzada & Keith, 2017;McCann & Ward, 2011;Yigitcanlar & Bulu, 2016).
In an interesting methodological piece that traces policy movements and circulation, Wood (2016) acknowledges the advancements in growing scholarship that identifies the mobility of policy knowledge and models via policy actors and their policy organizations, and even across different sites of learning. What is most encouraging however, is the recent interest in the methods employed by southern cities as they respond to circulating policy ideas (Harrison, 2015;Wood, 2015). This paper develops on the conclusions made by both Harrison (2015) and Wood (2015) that knowledge transfer may lead cities towards more effective policy solutions. This theme of social learning within urban planning resulting in positive change is also explored by von Sch€ onfeld, Tan, Wiekens, and Janssen-Jansen (2019) who caution of the danger to frame the process as necessarily positive and with guaranteed desirable outcomes. Whilst their argument is accepted, given the exchange of knowledge around strategic planning in the case study that resulted in implementable urban plans, this paper further builds on existing scholarship on how social learning can contribute to desirable outcomes.
It does so by reporting on a southern African research project that documented the detailed mechanics of knowledge transfer between urban planning practitioners involved in a long-term mentorship program coordinated by the international United Cities and Local Governments (UCLG).
By answering the research question "how exactly does the knowledge transfer process occur between southern African urban planning practitioners occur?", the case study research explored in detail the knowledge generation and transmission process that resulted in the production of bottom-up strategic plans developed entirely in-house by the respective planning teams. It is contended that this work is a timely contribution on both an under-studied and highly relevant topic. Moreover, by using three southern African cities as reference points, the research also supports the current move towards a geographical realignment of urban studies that begins to profile the work in cities in the global South.
This qualitative study employed four methods: document analysis, observation, focus group discussions and 18 in-depth key informant interviews with participants involved in the mentorship program. By unpacking exactly how the knowledge exchange process unfolded in the southern African case study, this article aims to fill a knowledge gap through the development of a learning model that outlines the stages in the social learning process of urban planning practitioners. It begins by contextualising both the role of urban strategic planning in Africa and the UCLG case study before a quick survey of the theoretical landscape of city-to-city learning and policy transfer. After presenting an overview of the research methodology, the key findings from the research is shared by unpacking the five stages in the relationship-building process. The article concludes by reviewing the contributions of the study and suggesting key policy directions.

Urban strategic planning knowledge transfer in southern Africa: The UCLG case study
In a sobering commentary on the urban development crisis in Africa, Pieterse and Hyman (2014) isolate the underlying reasons why significant proportions of urban dwellers in Sub-Saharan Africa will continue to live in slums and remain trapped in cycles of urban poverty. In their analysis, most African cities and towns lack the "institutional, financial and political resources to deal with growing levels of urbanization" (Pieterse & Hyman, 2014, p. 42). More importantly, they argue that despite international evidence strongly suggesting the need for urban planning and institutional frameworks that are underpinned by well-resourced, capacitated, effective, relatively autonomous and democratic local government systems, most African governments still continue to under-invest in the establishment of such frameworks (Pieterse & Hyman, 2014).
This argument was made earlier and more emphatically at the highest policy level in the UN-Habitat 2010/2011 State of the World's Cities Report, 'Bridging the Urban Divide' which noted that much of the inequality in the 21st century actually stems from inadequate planning by local authorities and central governments (UN-Habitat, 2010). In attempting to offer a solution to the crisis, one of the catalysts for change identified during the in-depth policy analysis conducted by UN-Habitat is the development of a sustained and clear city vision that provides a robust framework for urban strategic planning. Investment of substantial time and energy in building a collective city vision with all stakeholders is mooted as the first step to building a more sustainable and inclusive city.
Arguing in a similar vein, the UCLG (2010a: 10) in their policy paper on urban strategic planning which is viewed as merely "providing a methodology which helps cities identify their strengths and weaknesses while defining the main strategies for local development", recognize the role of urban strategic planning in helping decision-makers select appropriate goals that help them achieve a collective vision. A clear distinction is made between urban strategic planning and urban planning. It is suggested that urban strategic planning usually results in a tangible planning product such as a City Development Strategy (CDS). More importantly, for the UCLG (2010a), whereas land use planning, urban planning, comprehensive or integrated development planning are often legally binding instruments or laws, the point is emphasized that strategies are flexible tools for longer term orientation that is dynamic and can easily change and be adapted over time.
It is important to note however, that the UN-Habitat (2010) World Cities Report adds that city visions must not only give a very clear sense of direction, but must also detail the exact steps that it will take to get there. It then goes on to cite as good practice the South African example of the eThekwini Municipality's "Integrated Development Plan (IDP) 2010 and Beyond", an urban strategic plan, which outlines specific goals, strategic focus areas, programs, projects, budget allocations and performance indicators to emphasize the importance of translating vision into reality and ensuring that planning is put into practice (UN-Habitat, 2010: 157). Similarly, what has inspired many other cities about the early eThekwini strategic planning approach is its participatory design of a "single holistic process that moves from strategic and visionary statements of intent, through to a process that combines planning and budgeting for development, through to implementation and evaluation" (Dlamini & Moodley, 2002, p. 7).
In contrasting Johannesburg's 2030 planning initiative with Durban's LTDF and IDP process, Robinson (2007: 9) also acknowledges the bottom-up approach of Durban, observing that Durban's strategic planning process was shaped by a strong participatory approach, part of which actually involved "adjusting budget splits -the proportion to be spent on different council priorities". Beyond its ability to translate vision into implementable action, some of the other unique characteristics of the strategic planning process included the "fostering of partnerships with key stakeholders in the community in order to develop strong ties of trust, engaging political leadership, working with schoolchildren, taking risks and experimenting and embracing new opportunities" (eThekwini Municipality, 2011: 18-28).
The UCLG recognized Durban's good practice in the field of participatory urban strategic planning as an opportunity to be emulated by other cities in southern Africa (UCLG, 2010b). As a result, a mentorship program was operationalized through City Future, which is a joint program of the UCLG and Cities Alliance, aimed at promoting CDSs based on the then Millennium Development Goals (MDGs). The UCLG (2010c) was very clear about the roles and responsibilities of each city in the mentorship process. It was expected that the eThekwini Municipality as mentor "provide technical assistance in the areas of strategic analysis, including a pre-evaluation of the Durban lessons learned and to plan the next activities taking into account the local context" (UCLG, 2010c, p. 2). It is this UCLG mentorship program that identified the eThekwini Municipality as a mentor to provide strategic planning support to the Mzuzu City Council in Malawi and the Otjiwarango City Council in Namibia that is used as a case study to provide insights on the knowledge transfer process. The reason that this case study has been selected is because it has been identified as a good practice by the UCLG who have made Durban one of the emblematic cases for effective knowledge transfer in Africa. The study therefore offers important opportunities for critical reflection and learning.
The UCLG had selected Mzuzu City Council for participation in the mentorship program due to its strategic location in the northern region of Malawi, which serves as the capital for the entire Northern region (UCLG, 2010b). Growing at 4.2% per annum, it is the third largest urban centre after Lilongwe and Blantyre (UN-Habitat, 2011: 8). With a population of nearly 150 000 people, Mzuzu serves a broader hinterland population of about 1.7 million and important to note is regarded as one of the fastest growing cities in Malawi (UN-Habitat, 2011: 11).
Turning briefly to Namibia which also has had a long and shared history of colonial rule and underdevelopment with its South African neighbour, the Namibian mentorship program with the eThekwini Municipality, unlike in the case of Malawi, was supported at a national level by the Namibian Association of Local Authority Officials (NALAO) and the Association of Local Authorities in Namibia (ALAN). For the purposes of this paper, the Otjiwarongo Municipality which was the lead municipality in the mentorship program is focused on. Otjiwarongo, which is considered one of the oldest towns in Namibia, is the capital of Namibia's Otjizondjupa region and considering its strategic location, serves as a node between the harbour towns, the Capital city and the northern regions (Otjiwarongo Municipality, 2009). As the case with Mzuzu City, Otjiwarongo is also very strategically located, forming the route axis for trade and cargo movements from the Walvis Bay Port to the rest of southern African countries. The first contact between the mentor and mentee cities was in July 2009 when the Namibian municipalities were exposed to the eThekwini Municipality's strategic planning process. It is worthy to note that the relationship between the eThekwini Municipality and the Namibian municipalities is still being nurtured by the planners more than ten years after the inception of the process.
3. Theorising city-to-city learning Toens and Landwehr (2009) observe that previously the policy literature on learning focused more on what was being learned by individuals and other political actors, rather than on how they learned from each other. So, whilst city-to-city cooperation has enjoyed much acclaim as an effective mechanism for municipal capacity building and policy transfer, very little is known about exactly how the knowledge and capacity transfer takes place and about the actual learning practices experienced by the practitioners involved (Bontenbal, 2013;Haupt, Chelleri, van Herk, & Zevenbergen, 2019). Campbell's (2012a) seminal work on city-to-city learning has begun to fill in this void, and his insights (which are used as key reference points in this paper), have revolutionized contemporary thinking on city learning and policy transfer with the introduction of the concept of learning typologies. Campbell (2012b) was commissioned by the UCLG in September 2012 to prepare a strategic paper on the international outlook for city learning. As the thought leader for the UCLG workshop on city-to-city learning in Barcelona, he built on the work in his book Beyond Smart Cities to include information from cities (including Durban) attending this strategic workshop. As indicated in Table 1 above, Durban is now the first African city to feature amongst other international examples of learning cities in Campbell's revised typology.
Through a categorization of types of learners, Campbell (2012b) makes reference to the purpose, motivation and intention of learning by cities. In this typology, some cities (as depicted at the bottom of the table) are represented by individuals or small groups that are involved only in casual and ad-hoc learning, whilst others in the middle are similarly represented but focused on special purposes or specific topics. There are those cities, however (as depicted at the top of the table), that involve larger groups who take part collectively in what Campbell (2012b: 11) describes as the "more intentional, committed and continuous pursuit of learning". It is interesting to note that Durban is represented in this category, given its commitment to learning and the institutional mechanisms including the establishment of its own Municipal Institute of Learning (MILE) that it has established to promote city-to-city learning (eThekwini Municipality, 2010).
One of the central arguments of Campbell (2012a) that is relevant for this paper, is that city learning depends on the relationships amongst those engaged in a major city project, plan or event. Employing a powerful and novel social network analytical tool that groups data collected from his many interview respondents, Campbell (2012a: 111) presents the networks of ties between the key actors in a single picture to reveal what he terms "clouds of trust" or constellations of network ties between persons. Using the case study of Turin, the research undertaken by Campbell (2012a) uncovered a tight system of trusted ties among the players in the planning processthe critical ingredient responsible for Turin's turn-around.
The proposition made by Campbell's (2012a: 11) on what makes city learning successful are the informal leadership networks, referred to as "clouds of trust" is a fundamentally important assertion that has been tested empirically in this southern African case study. A unique opportunity was provided to test the applicability of the notion of "clouds of trust" identified by Campbell (2012a) in the context of stakeholders within particular African cities to a community of stakeholders involved in city-to-city learning. In particular, during the interviews, focus group discussions and through the census survey, the issue of the importance of trust is explored, in order to validate Campbell's (2012a) assertions in this regard. This application has, to the best of the researcher's knowledge, not been done before in the African context.
From a learning theory perspective, Campbell's (2012a) assertion is in line with the earlier thinking espoused by knowledge management expert Nonaka (1994: 14) who argues that merely transferring information from one context does not make sense if it is "abstracted from embedded emotions and nuanced contexts that are associated with shared experiences". It is this personal, human side that is fundamental in making mutual learning possible. This point about human interaction in city-to-city learning has over the last decade been overshadowed due to advancements in communications and technology, and the focus on digital learning. In fact, Campbell (2012a: 41) laments the fact that leading academic thinkers have been swayed in part by the "all-powerful digital revolution" towards focusing only on the global connectedness of networks and interactions with digital media -at the expense of delving into a critically important understanding of what makes connectivity important. In particular, Campbell (2012a) stresses the point that human relationships and how these play a role in connectedness have not been given theoretical prominence.
In an attempt to unpack this phenomenon, Campbell (2012a) emphasizes two centrally important and related concepts that are critical for understanding knowledge transfer between cities. The first is that of an innovative milieu identified by Aydalot (1986) and understood as the creation of an atmosphere of complete trust that is helpful for both "collaboration and creativity in a particular locality" (Campbell, 2012a, p. 43). As introduced earlier, the second idea, developed by knowledge management thought leader Nonaka, is referred to as a ba (Nonaka, 1994;Nonaka, Toyama, & Konno, 2000). In an industrial management setting, this ba refers to the creation of a particular atmosphere, or openness, where managers and workers are without the "negative constraints of possible embarrassment, retribution or even fear" (Campbell, 2012a, p. 44). The argument in Campbell's (2012a) thesis is that it is this personal trust and sharing of values that is an essential ingredient in determining whether those involved in city learning will reach out to others in learning networks in any meaningful or significant way. It is interesting to note that in recent research conducted by Haupt et al. (2019) which involved in-depth interviews with forty-two stakeholders engaged in city-to-city learning processes across twenty-two countries, also revealed the critical role of personal relationships based on mutual trust in contacts facilitated by trans-municipal networks.
Drilling deeper into this key notion of trust and open dialogue as an important ingredient for city-to-city learning, another insight from the literature that sheds greater light on the mechanics of learning comes from the work of Johnson, Pinder and Wilson (2012). It is argued that with direct face-to-face interactions between practitioners from the collaborating cities, the opportunities not just for one-sided, but for mutual learning is maximized (Johnson et al., 2012). This happens because officials are from similar professional backgrounds and share a common framework, which is a critical component to sharing tacit knowledge. Table 1 Typology of city learning (Campbell, 2012b, p. 11 The learning literature also point to the interesting notion that beyond sharing this degree of similarity, it is difference that is considered to be of equal importance to practitioners involved in city-to-city learning (Johnson & Wilson, 2009;Van Ewijk, 2012). During interactions with practitioners from different social, economic, political and cultural contexts the "other" may provide an ideal opportunity for critical reflection of one's own approaches (Wulf, 2001cited in Devers-Kanoglu, 2009. This was an important observation for the purpose of the research as the planning and development contexts in South Africa differ radically from those in Malawi and Namibia. It is on the basis of this experience that people change their perspective after a process of internal reflection that transforms information into knowledge and ultimately learning (Higgins, 2011). These "situations which are complex and novel, calling for situational understanding and awareness", as argued by Higgins (2011: 583), should provide exactly those experiences to critically reflect upon. Again, a unique opportunity was presented to test these assertions in the UCLG case study, when planners from eThekwini Municipality share their stories of citizen-driven sustainable strategic planning that may be in contrast to those told by their Namibian and Malawian colleagues.

Methodology
This qualitative research study employed four deliberate methods. Commencing with document and textual content analysis was very useful, as vital information on processes and policies are contained in municipal and UCLG documentation. Key reports and narratives authored by the eThekwini Municipality's Strategic Planning Unit were analysed thematically. The reports, learning notes, terms of references, funding proposals, and other documentation of the UCLG's USPC were also analysed to fill in key knowledge gaps about the mentorship program.
In order to better understand the mentorship process, the planning methodology and the role played by the eThekwini Municipality planners and facilitators in the mentorship program, the second method involved conducting a focus group workshop. All five planners and facilitators in eThekwini Municipality who were directly and indirectly involved in the mentorship program participated in the focus group workshop, providing a "common communicative ground" (Hyde'n & Bülow, 2003, p. 311) around which reflective discussions could take place.
Given that the planners and facilitators had a common sense of belonging as they had all been involved in the program and were committed to the city learning process, the focus group helped them to feel safe to share information (Vaughn, Schumm, & Sinagub, 1996).
All other selected key role-players involved in the process were then interviewed as part of the third research method. The perspectives and insights of each of the participants were important in order to build a richer picture of city-to-city learning through the careful and deep probing into how the knowledge exchange process unfolded. As the mentorship program was a closed one, involving a fixed number of known protagonists, the issue of key informant interviewee selection was less complex. The criteria for inclusion related to direct involvement in the UCLG mentorship program in the role as planner, facilitator, funder, manager or Local Government Association support staff. A total of 18 interviews were conducted, drawn from international agencies, Namibian and Malawian planners/facilitators and support staff as well as support staff from within eThekwini Municipality, including two of the key staff (planner and facilitator) who were part of the focus group workshop.
Being a practitioner in the field of international city-to-city learning, and a known "insider" to the key role-players in the mentorship program, and one who was considered part of the broader mentorship team, the researcher had full access to all the sessions that were conducted during the various interventions in South Africa and abroad in all three municipalities. Whilst being a researcher and practitioner did pose some challenges, it is strongly contended that in being a practitioner in the field of city-to-city learning, and a known "insider" to the key roleplayers in the mentorship program, one of the major strengths that enriched the quality of the research was being able to experience firsthand the emotions, feelings, narratives and "back-office" stories of the planners involved in the mentorship program. Regarded as one of their own, the level of honesty in the reflections of the key actors ensured that the data that was more reliable and allowed the researcher a wider spectrum for reflection and introspection.

Findings: Unpacking the mechanics of city-to-city learning methodologies
Attempting to fill in the knowledge gap on exactly how knowledge and capacity transfer take place and to uncover the actual learning practices experienced by the practitioners themselves, the eThekwini Municipality's focus group participants were asked to explain exactly how the learning process unfolded. In explaining how the learning process was initiated, participants made reference to the first southern African UCLG mentorship program between Johannesburg and Lilongwe, Malawi. This mentorship program which commenced at the same time as that of the eThekwini Municipality proceeded much faster than eThekwini Municipality, given the appointment of a dedicated consultant on behalf of the Johannesburg Council to support the Lilongwe process. The eThekwini team contrasted this approach with theirs which differed fundamentally in that it involved a hands-on training and empowerment phase where the facilitators from the mentee cities were invited to Durban to understand their strategic planning process.
The Namibian and Malawian colleagues were provided an opportunity to begin to internalize the outcomes-based strategic planning methodology. This was done through a sharing of the eThekwini Municipality's transformation story which showed how the municipality was not being responsive to its own stakeholders, how this had led to poor quality of life results and which had prompted a new transformation plan (Dlamini & Moodley, 2002). During the training, in addition to field trips to key projects that were implemented as a result of the strategic planning prioritization process, Namibian and Malawian colleagues were also exposed to hands-on facilitation tools such as case studies. These tools were then used by the local planners during the process of identifying local challenges and creating vision statements when conducting their own workshops: Our process is essentially about empowering them (reference being made to the Namibian and Malawian colleagues) to really take full ownership and control of their own strategic planning process. Unlike the Lilongwe-Johannesburg model, our eThekwini team is seen to be much more in the background, playing a supportive role. Of course, this process involves a long timeframe, and commitment from all of us. Sometimes the commitment waxes and wanes over time, but generally it is there.

Participant 1, eThekwini Municipality
During the focus group interview, it emerged that the facilitation team did not have an explicit overall learning methodology. As a result, one of the key outputs from this research was the construction of a model that summarizes the five key stages that the municipalities appeared to follow during their mentorship journey. It is important to stress that this model, as depicted in Fig. 1 below, is merely an analytical construct that captures the process of knowledge transfer and learning exchange between the partners, as articulated by the focus group participants.

One: Courtship and acclimatization stage
This first stage of city courtship and acclimatization involved the mentor and mentee cities exploring whether there is a potential for collaboration. Officials from each city became familiar with each other and their respective municipal contexts. In this early stage of the relationship, there was a heavy reliance on an external broker, or to continue with the matrimonial analogy, a city 'match-maker'. In this instance. the program manager from the UCLG in Barcelona, a well networked, welltravelled and politically astute individual recognized the potential that Durban could play in southern Africa and played a key role as the 'gobetween' in this early courtship stage in what was to become a long-term city-to-city relationship. 1 From this analysis, the case study suggests that if city-to-city learning and policy transfer is to yield powerful results based on strong relationships of trust, then a foundation must be laid in the early stage of the engagement. This is the first important finding that emerged from the study.
The focus group participants admitted that this was a very timeconsuming stage. As the lead facilitator stated, it was important for city officials to "really find each other". It is interesting to note that three exploratory trips over a two-year period were made to Mzuzu by the lead facilitators from Durban even before the actual mentorship process commenced. The first trip was to present the Durban model and test the appetite for a relationship. The second trip was regarded as a 'reconnaissance' trip to understand the lay of the land, and plan a program of action with all the role-players including Cities Alliance who was the sponsors. The third trip was to develop the details of the exchange and plan the logistics of the exchange.
During the focus group discussion, it dawned on the participants that in Namibia the initial exposure period spanned five years before the Namibian team were ready to begin the process. This was a fact that was uncovered, only through reflection and made an impact on the participants. They recalled that during the five-year period a series of telephonic and e-mail interactions had ensued, including a visit from the Namibian team who were invited to watch how the Durban team supported another local municipality in KwaZulu-Natal.
Unlike in Malawi where visits were made to understand the context there, this was not the case in Namibia and according to the participants they felt that this was a less robust relationship than Mzuzu because of this shortcoming. Both the lead facilitator and lead planner agreed that more visits should have been made to Namibia to begin to understand the development context there before initiating the program. This is the second important finding that emerged from the study.

Two: Inspiration and reflection stage
With the cities declaring their learning interest in each other, the first stage of acclimatization was then followed by a critically important process of the mentee cities being inspired by the work and the actual results of their mentors' strategic planning efforts. In this stage the mentees saw what was actually happening in the mentor city and heard moving personal testimonies and stories shared by the Durban colleagues. The Durban story of city transformation from 2001 was narrated together with inside information, shared in an environment of complete trust. After two days of closed discussions, they then conducted a city tour to observe the transformation made in the city in the form of built environment projects that had been catalysed through the strategic planning process. They also took note of challenges that still needed to be addressed in the city. These honest, open and frank discussions, spurred on by the fact that what brought them together was a set of common challenges, as suggested by the respondents, helped to build what they regarded as a solid foundation of trust and mutual understanding between the officials of both cities.
It is acknowledged that it is difficult to test or measure this empirically as a researcher. However, it appeared that it was the power of the depth of the city-to-city interactions that helped shape the long-term learning and seemed to be the glue that held the program together. Officials engaged with each other over a prolonged period and became familiar with each other as they learned more about each other's 1 Note that beyond the Terms of Reference and project funding cycle, the relationship between the three cities is still being nurtured by Durban. challenges and unique contexts, keeping up the contact through ongoing e-mail dialogue and constant telephonic conference calls. A senior planner captured the essence of the uniqueness of the learning in her thoughtful consideration: It is really surprising how every interaction, every conversation can take you deeper and deeper as we hear more about the other city… but in the same way we learn about our own city too!

Participant 3, eThekwini Municipality
The issue of trust between planners from each of the cities appeared to be one of the most critical ingredients in allowing planners to speak their minds, challenge each other and reflect on how best to implement the strategic planning process. After much probing to get a better sense of how respondents understood and interpreted this notion of 'trust', key insights emerged: Trust is fundamental to make any relationship work. Long-term learning will only be successful if there is trust. Trust is built when both parties deliver on their commitments. Trust is easier to be built between peers, as opposed to a planner and a World Bank consultant. Trust is also built when all parties' behaviour is geared towards achieving the common goal of learning.

Member of Cities Alliance Secretariat
Trust means valuing what is being said and then adapting this to suit the local context. There are unspoken things, for example, when funders want information on the project that might be perceived as demanding, both sides gel and trust is developed.

Participant 3, eThekwini Municipality
Trust involves giving the mentee the freedom to express themselves in an open manner.

Participant 1, Otjiwarongo Municipality
The third key finding from the study is therefore the recognition of paramount importance of trust, not just for making the overall relationship work, but for ensuring that the experiences are filtered and applied to the respective unique local contexts as well as helping to ensure more open expression between the practitioners in the learning partnership. Of all the insights offered around the importance of trust, the most illuminating view was offered by one of the younger support facilitators from the Mzuzu City Council, who again used the powerful analogy of a mirror, to make the point that trust is critical for mindset and behavioural changes: For most of the city challenges to be appreciated, it must be explained in detail. Now, you can't give details of the challenges to someone you don't trust. You have to provide an environment where everyone feels that they are important, and you appreciate what they have done, and that is where it starts. It is like how you trust your mirror, you know, that is exactly what happens. If we are to change the mindset of people, they need to look at us from that point of view, where they trust us, and whatever we say, there is no partiality in what we are saying and it is the true picture of what is there… just as we do with the mirrorwhatever the mirror tells you, you believe it.
The powerful in-depth reflections volunteered by the interviewees represent a critical area of contribution that the case study begins to make in terms of the validation of theory in the southern African context. Three important observations will be focused on here. As alluded to earlier, in the first instance, Habermas' (1990: 135-136) notion that communicative reason is made easier when there is a "consensus created through shared 'life-worlds'shared contexts, cultures, values and propositions about social life", does appear to hold true, based on the views expressed by the practitioners. The fact that the practitioners were all planners and facilitators, and not a mix of different built environment professionals such as architects and engineers, and the reality that they all shared the same worldview and were particularly committed to developing powerful municipal vision statements and strategic plans, helped create a shared life-world. They spent most of their work-life grappling with similar challenges and they experienced similar frustrations. This facilitated open and honest communication with each other. The case study therefore does validate this aspect of Habermas' assertions.
Secondly, beyond this shared mindset, it is argued that what is fundamentally more important from this research is that the planners have noted that spending time with each other, relating personal stories, travelling together on the mini-bus during the field visits, having meals together and even sharing drinks after hours in relaxed settings whilst reflecting on how what was being learned can effectively be applied to their own situation, was an important part of the learning process in this second stage. The one eThekwini municipal respondent volunteering an extreme example of sharing an item of personal clothing with a colleague from abroad whose luggage was delayed, was found particularly entertaining by all the focus group participants. In a very concrete way, however, it actually reinforced the point about the extent to which practitioners can bond with each other and build up strong ties in a network of trust that is required for transformative action.
From the case study, it is clear therefore that a necessary condition for effective learning, beyond shared worldviews as identified in the literature by Habermas (1990), is this developing of trusted ties between the practitioners involved in the learning process. To reiterate this point given the gravity of the finding, this research suggests a much more nuanced understanding of learning in southern Africa. The emphasis placed by Habermas (1990) on the importance of shared worldviews is indeed accepted as a necessary; however, it is clearly not a sufficient condition for transformative action arising from the learning process. More than just sharing a planning worldview, this research shows that trust between the planners promotes critical dialogue and reflection which is fundamental for transformative action.
Thirdly, respondents also appeared to confirm the critical importance of the concept of Campbell's (2012a: 111) informal leadership networks referred to as "clouds of trust" that represent important linkages between key actors involved in a particular setting. Whilst this research did not physically map the networks, it is clear that there are key protagonists in each city that serve as the glue that hold the learning process together. These are the lead facilitators that are trusted by all the participants, and which in turn develop further networks. This was evident in Malawi, where the CEO of the Blantyre City Council was invited for the Mzuzu strategic planning workshop, as he had heard about the process from his peers in eThekwini and Mzuzu; peers that he had known and had trusted in these cities. He was excited by the process and was interested in replicating the similar strategic planning process with support from the eThekwini municipal facilitators in his own municipality. In a similar vein, the case study also appeared to re-confirm Nonaka's (1994: 14) argument that transferring information from one context to another makes very little sense without recognizing the "embedded emotion and nuanced contexts that are associated with shared experiences".
Fourthly, these findings also validate the assertions made by von Krogh and Geilinger (2014) who propose that powerful knowledge creation processes are actually embedded in the spaces of organizations. Having practitioners from the cities engaging in face-to-face interactions is indeed important, but more critical was ensuring that these were intimate spaces and safe spaces, away from outsiders, spaces in which practitioners could bond with each other and could speak their minds without fear.
The case study therefore suggests the carefully constructed program, the way the room was laid out by the MILE facilitators ensuring intimate group discussions, the deliberate organizing of bus trips and joint dinners, all contributed to what Nonaka (1994) and Nonaka et al. (2000) referred to as a "ba", or the creation of an atmosphere of openness without any fear of potential retribution, embarrassment or fear, facilitating transformative action. von Krogh and Geilinger (2014) made a recent call to pay more attention to these social and informal aspects of knowledge creation across organizational boundaries. This, it was argued, was necessary to begin to understand how knowledge flows informally in what is regarded as an "eco-system of organizations that are in geographically proximate spaces" (von Krogh & Geilinger, 2014, p. 156).
What the findings in this case study demonstrate are that during this stage in the mentorship process, practitioners from the respective organizations (eThekwini Municipality, the local government associations, Mzuzu City Council and Otjiwarongo Municipality) came together for a particular period of time into a learning eco-system with a single unified objective of improving municipal visioning processes. This eco-system, it seemed appeared to be a safe space for most that maximized the learning potential and helped challenge previously held mindsets about city planning and development. The learning impact achieved in this time could in no way compare to time spent through virtual learning platforms or other non-physical contact sessions. Yet again, Campbell's (2012a) thesis that human interactions and human relationships are critical to understanding city-to-city learning, comes to the fore.
In the interview with the senior UCLG respondent who was the most experienced city-to-city learning practitioner in the program, the point about officials being inspired by face-to-face practice as being an important turning point in itself was reiterated. She held the view that when people are exposed to the city realities in a particular safe setting: All the senses are utilized, and a much deeper identification happens. City visits are therefore fundamental… you must be exposed to the city and feel the reality of the city.

Participant 1, UCLG
Similarly, the UCLG program officer who had also been involved in other city-to-city learning between Brazilian and Mozambican municipalities stressed the importance of on-site visits and direct interactions between practitioners. She captured the spirit of the relationship most succinctly in observing: People must get a chance to see each other's faces. They must know each other; the process is fortified when the exchange happens. They spend quality time together; laughing, joking and sharing problems together, getting to know where each other comes from, in their own contexts. They must have empathy with each other, and be able to put themselves in each other's shoes.

Participant 2, UCLG
In summary then, it seems that Stage Two was the most critical stage in the learning process in that, according to the respondents, a solid basis for open and honest communication was laid, and trust between the planners was being built. Having been exposed to the Durban story, the practitioners from Namibia and Malawi now began to make the transition from inspiration to adoption.

Three: Adoption stage
Closely related to Stage Two, where the three-day intensive training in Durban takes place aided with manuals, role play, case studies and site visits; a process of reflection and challenging assumptions begins. A fundamental part of the training involves all role-players quietly but actively reflecting on what it would take to implement a transformation process in their respective mentee cities. The process is designed so that by this time, the planners from the mentee city become convinced of the need for change and are inspired and enthused to take action. The transition to the third stage of adoption is therefore smooth. Having been inspired and having gone through the process of internal reflection, they are ready to adopt the 'outcomes-based planning' methodology and are eager to roll out a similar strategic planning process in their own city, mindful of the need to make adaptations along the way to suit the specificity of their own unique local contexts.
What is theoretically relevant in this stage of the learning process is the observation of Nonaka et al. (2000) who recognize the process of organizational knowledge creation as involving the amplification of knowledge in social contexts and selectively connecting it to existing knowledge in the organization. The findings from the case study show that practitioners admitted to reflection about application to their own context, after engaging in these safe spaces. In such spaces, it appeared that practitioners engaged openly and honestly, and then made time to see how the new information which challenged their mindsets could be applied to their own contexts back home.
It is important to emphasize that the work done by academics in the field of organizational learning that make the case that having different contexts is just as important in city learning is relevant here (Johnson & Wilson, 2009;Van Ewijk, 2012). Their suggestion that interactions with practitioners from different contexts may provide an ideal opportunity for critical reflection of one's own approaches certainly holds true in the case study, as admitted by the respondents. The idea therefore that it is on the basis of experience that people change their perspective following a process of internal reflection that transforms information into knowledge and ultimately learning, as evinced by Higgins (2011), does appear to be confirmed by the study.
Another key theoretical insight from the case study in this regard is the demonstration that practitioners do not glibly import models from one context to another. The process of reflection and interrogation of the validity of the learning and the ability to apply it to practitioner's own unique local context comes through very clearly. This acknowledgement has been made in the literature by Robinson (2011) who accepts that city strategies are in fact significantly reconfigured when they arrive in particular places. This realization of the need to recognize the specificity of the local context and the unique geographies of space in implementing planning innovations from other cities is yet another key finding emerging from the research.

Four: Supported implementation stage
Equipped with the know-how and tools to run their workshops, facilitators from the mentee cities became activated to play an active leadership role, running the strategic planning sessions themselves in respective municipalities, with the full support of their local political leadership, and the technical support of the Durban team.
In reflecting on this learning process, the UCLG respondent stressed the point of having strong leadership to make sure that the planning process stays on track. For her: The learning of a city relies on the learning of a team that can fall or stand depending on the leader.

Participant 1, UCLG
This aspect of strong leadership it was suggested, particularly during the period of implementation was found to be important. Making the transition from planning to implementation, they agreed, was the point at which most processes failed.
The researcher accompanied the lead facilitator to one of the sessions held in Otjiwarongo to observe how this process unfolded in practice. What stood out was the patience from the eThekwini municipal planner, who resisted the temptation to intervene in the workshop to answer questions raised by the civil society stakeholders. Suggestions on how to move forward was not made publicly, but discussions were held over tea and lunch. The researcher recorded these as journal observations and during the debriefing session mentioned these to the eThekwini municipal planner, who confessed that resisting active participation required discipline as there was an urge to help shape the process. She reflected that this was all part of the personal learning and development process.

Five: After-care stage
In this final stage, the mentees have now taken ownership of the workshop with the Durban team merely providing support. Once a road map is developed, the targets identified are self-monitored by the local team. The Durban team did, however, play an oversight role, using opportunities at conferences and workshops to meet to discuss progress, over and beyond the telephone conferences and e-mail dialogues. As in any union, the post-honeymoon stage is perhaps the most challenging one, as the novelty of the learning relationship has worn out, and the hard work of implementation has begun. The need for open and continued communication between learning partners during this time is key to achieve the desired learning outcomes. It is interesting that the lead planners and facilitators are still in contact, with Durban playing a role in ensuring that the contact and relationship is still nurtured, years after the project funding had expired.

Summary and conclusions
The study put the learning process between planners in the eThekwini, Mzuzu and Otjiwarango Municipalities under the microscope, in order to better understand exactly how knowledge transfer takes place in this UCLG mentorship program. Using the insights gained during the indepth interviews and focus group discussions, the research was able to develop a learning model with five clearly delineated stages in the cityto-city learning process. In summarising some of the main findings, the paper concludes by providing pointers for policy considerations for those involved in city-to-city learning processes.
The research suggests that if city-to-city learning is to yield powerful results based on strong relationships of trust, then a foundation must be laid from the early stage of engagement, where planners from the respective cities begin to identify with each other. From a policy perspective, city-exchanges will require significant long-term investment in terms of time from stakeholders that will be expected to meaningfully engage in learning processes. Recognizing the reality of leadership changes and the fact that securing buy-in from a range of stakeholders takes longer than planned, it is important to ensure that realistic learning timeframes are set. Whilst the exact length of a typical city-to-city learning program cannot be prescribed as this would be context specific, there has to be a fine balancing act between ensuring that a long enough time to allow for meaningful interaction on the one hand and the need to begin to deliver on tangible products and outcomes without learning and participation fatigue on the other.
Perhaps more important to note is that committing to longer-term learning relationships has implications for applicants preparing funding proposals, as often most of the funding is reserved for travel of delegates either for training sessions or for the actual strategic planning workshops. Consideration of the need for reconnaissance visits and high-quality contact time throughout the process between the planners directly involved is not necessarily factored into budgets and business plans. This is therefore an important practical policy consideration for program managers, funders and learning practitioners themselves, in order to ensure that the returns of investment in learning is maximized.
The research also confirmed the power of the depth of the city-to-city interactions that helped shape the long-term learning, and appeared to serve as the glue that held the program together. Officials engaged with each other over a prolonged period and who became familiar with each other as they learned more about each other's challenges and unique contexts, maintained contact through ongoing e-mail dialogue and constant telephonic conference calls. The trusted ties referred to by learning theorists (Campbell, 2012;Nonaka et al., 2000) that allowed planners to speak their minds, challenge each other and reflect on how best to implement the strategic planning process cannot be underestimated in its role in transformative action. Two related policy considerations emerge from this finding. Firstly, attention must be paid to creating a learning methodology that values face-to-face interactions. This will include a range of interventions from ensuring that highly skilled facilitators manage the process, to the careful design of the training session content to maximize group participation, right down to the micro-design of the room layout. This suggestion may appear insignificant, as often study tours are hurriedly constructed, however, if thoughtful consideration of the program design is given, more meaningful engagements and better learning outcomes are more likely to be achieved.
Secondly, learning event organizers will do well to consciously create the space and opportunities for social engagement and interaction. This is particularly important for host cities to take note of, as often project budgets do not allow for these 'soft' expenses related to learning programs. The creative use of host city Mayoral receptions, for example, is a concrete mechanism to employ for augmenting limited project budgets to ensure more social interactions. At a more strategic level, it is important for decision-makers themselves to be aware of the importance of social interactions in building trust networks that are essential for social learning and transformative action.
Finally, the role of energised political and administrative leadership in policy transfer processes in cities and regions, particularly at the commencement of the implementation stage is interesting to note. Whilst the issue of visibility and credibility is important, fundamentally more important is that the political leadership is fully appraised of the process and feels confident to give strategic direction and actively help shape the learning agenda.
The development of this learning model is an important milestone as it is the first time that the documentation of and critical engagement with the mechanics of learning between cities has been undertaken in a UCLG program. With the renewed emphasis on learning and knowledge exchange between cities in a race to meet SDG targets, it is hoped that the model of learning is a good starting point for cities and regions that are considering embarking on a similar learning journey with a view to ultimately improve the quality of their urban strategic plans.