Climate Finance Coordination From the Global to the Local: Norm Localization and the Politics of Climate Finance Coordination in Zambia

Abstract The transfer of climate finance to developing countries is a key feature in global climate change agreements. This article examines how conflicting preferences between climate finance donors and domestic actors affect the diffusion of emerging international norms on climate finance coordination at national and subnational levels in Zambia. Informed by literature on norm localization we trace the diverging preferences and interactions among donors and domestic actors over a 12-year period as they shape climate finance coordination in Zambia with support from the global Climate Investment Funds. We find that this has been a highly conflictual and dynamic process with contestation centered on norm application rather than norm validity. While the World Bank have had a strong initial influence on the enactment of coordination norms, domestic actors have over time undermined and reconfigured coordination arrangements to align better with their own preferences. Our findings show (i) how emerging international norms on climate finance coordination may be localized and reshaped as they are enacted in developing countries; (ii) that preferences may differ significantly between and among donors and domestic actors in this regard, and (iii) that past relationships from development cooperation may be carried forward but also challenged in this process.


Introduction
Climate finance is a key feature in the implementation of the Paris Agreement on climate change. Article 9 of the agreement commits developed countries to mobilize USD 100 billion per year to support climate change adaptation and mitigation in developing countries. While this goal remains to be fulfilled, finance is now flowing from a variety of climate funds and donors to developing countries, with reported public climate finance totaling USD 62.2 billion in 2018 (OECD, 2020).
The institutional architecture of climate finance is still developing, including the norms that guide how it should be governed and managed. While there has been much focus on the global politics of climate finance and associated North/South dynamics, less is known about the ways in which emerging international norms around the governance of climate finance are diffused and contested within developing countries, and in particular how climate finance donors and domestic bureaucracies interact in this respect (Clar, 2019;Pickering, Betzold, & Skovgaard, 2017).
Significantly, a number of the key actors involved in this process are not strangers to each other, but long-standing actors in the field of development cooperation. The Multilateral Development Banks (MDBs) and bilateral aid agencies thus play a major role in channelling public climate finance to developing countries (Buchner et al., 2019;OECD, 2020). While the accounting of this finance is contentious (Weikmans & Roberts, 2019), the emergence of climate change as a major emphasis in the strategies and portfolios of aid donors is clear (MDBs, 2020;UNDP, 2021;WB, 2021). Understanding these actors' role in shaping the norms and mechanisms for climate finance governance in developing countries is therefore important.
Our focus here is on one particular aspect of this issue, namely the emerging norms around the coordination of climate finance among and between climate finance donors and recipient governments. 1 The issue of coordination is not new on the development agenda, but with the advent of global climate change agreements and associated public climate financing, new efforts are made to address the coordination of climate finance specifically. The rationale is to enhance coordination between and among climate finance donors and recipient governments, and thereby move beyond the so far very fragmented climate finance landscape (Pickering et al., 2017;Zelli, 2011). The lack of coordination has caused concerns over inefficient use of already inadequate climate funds, misalignment between donor interventions and government policies, and contradictions between climate change interventions and other development efforts (Amerasinghe, Thwaites, Larsen, & Ballesteros, 2017;Nakhooda & Jha, 2014).
So far, the limited academic literature on climate finance coordination has focused mainly on coordination at the international level or across multiple countries (Pickering et al., 2017). Very few studies have examined the ways in which the emerging norm of climate finance coordination is diffused and evolves in domestic settings, how donors and domestic actors interact in this process, and what the outcomes are (Clar, 2019).
In this article we offer insights on this issue by applying a norm diffusion lens to the evolution of climate finance coordination in Zambia. Specifically, our aim is to examine how conflicting preferences between and among donors and government actors affect the ways in which emerging international norms of climate finance coordination are transformed into practical institutional arrangements at national and subnational levels.
Our empirical analysis focuses on the evolution of World Bank administered support to the establishment of climate finance coordination arrangements in Zambia. Funded through the global Climate Investment Funds, Zambia's Pilot Program for Climate Resilience (PPCR) forms part of a global suite of national adaptation programs with an emphasis on coordinating and mainstreaming support to climate change adaptation. The 12-year implementation period in Zambia provides a rich case for examining how coordination has evolved over time and at different organizational levels.
Drawing on norm localization literature (Acharya, 2004), we understand norm contestation not only as struggles over the validity of norms, but also their application (Zimmermann, Deitelhoff, & Lesch, 2017). On this basis we examine the conflicting preferences (shaped by strategic interests and organizational norms) of donors, central government agencies and local government authorities in Zambia vis-a-vis climate finance coordination, trace their interactions as they pursue these preferences, and discuss the nature of climate finance coordination that results from this process.
We find that the emergence of climate finance coordination in Zambia has been a highly conflictual and dynamic process with contestation centered on norm application, that is, the organizational 'how' and 'who' questions of coordination. While the World Bank has initially been successful in pursuing its preferences for climate finance coordination in Zambia, central and local government actors have over time gradually reconfigured donor-dominated coordination arrangements to match their preferences better or pursued competing coordination mechanisms. The resulting climate finance coordination arrangements are of a hybrid nature that remain contested and continue to evolve.
Our findings contribute new insights to the developing literature on the governance and politics of climate finance and its practical application in the context of development. Firstly, our findings show how emerging international norms on climate finance coordination may be reshaped and localized as they are turned into practical arrangements in domestic settings. Secondly, our findings highlight the competing preferences and actions of donors and domestic actors as they enact climate finance coordination in developing countries, and how these preferences may differ significantly not only between donors and government actors, but also internally among them. Thirdly, our findings add insights on the interactions between and among multilateral development donors and domestic actors in the new arena of climate finance, including how existing norms and relationships from development cooperation may be carried forward but also challenged as climate finance norms are turned into practical arrangements.

Conceptual framework
In the following we first outline climate finance coordination as an emerging international norm. We then discuss the literature on norm localization and associated work which helps us understand the dynamics of how such norms are diffused and contested in domestic settings. Lastly, we describe our understanding of actor preferences and associated power dynamics as actors engage in the diffusion of climate finance norms.

Climate finance coordination as an emerging norm
Coordination can be defined as a process aimed at aligning interdependent tasks (Okhuysen & Bechky, 2009). In everyday social and organizational life humans continuously coordinate as they relate to each other. This includes ad hoc micro-actions and more organized, regularized practices (Isabelle et al., 2012). Our emphasis here is on the latter, and especially on the intentional efforts of actors to establish and enact particular coordination arrangements. We focus mainly on formal arrangements, but with attention to how their evolution is influenced by informal practices.
Coordination of development aid is a long-standing concern and normative principle in development aid (Bourguignon & Platteau, 2015). The 2005 Paris Declaration on Aid Effectiveness emphasized coordination as a means to make aid more effective, while the 2008 Accra Agenda for Action called for recipient countries to play a stronger role in aid coordination. However, studies of coordination in conventional aid have pointed to competing interests among donors (Saltnes, 2019) and perceived value clashes with national governments (Olivi e & P erez, 2016), leading some to conclude that the 2005 Aid Effectiveness norms have simply failed (Brown, 2020).
Despite and separately from these experiences, emerging norms for the specific coordination of climate finance are developing as part of the overall efforts to establish global principles for the governance of climate finance. The UNFCCC Standing Committee on Finance has been charged with the responsibility to improve the coherence and coordination of climate finance (SCF, 2021). The major multilateral climate funds have publicly committed to take active steps to improve coordination (AF et al., 2021) and the strategies of climate finance donors often commit to coordination (GCF, 2020;WB, 2021). A number of policy studies have furthermore called for better coordination (Amerasinghe et al., 2017;Nakhooda & Jha, 2014;Pickering, Jotzo, & Wood, 2015). As an ideal, coordination is thus increasingly articulated in international Climate Finance Coordination From the Global to the Local 1901 debates on climate finance as 'a standard of appropriate behaviour for actors with a given identity', to quote a common definition of norms that we adopt here (Finnemore & Sikkink, 1998, p. 8).
In practice, efforts to develop the coordination of climate finance to developing countries seek to (i) address coordination between different multilateral funds, agencies, and donors at the international level; (ii) mainstream climate goals in development aid and align development and climate finance; and (iii) establish climate finance coordination arrangements in the recipient countries. The rationale for the latterwhich is our concern hereis founded in the growing international climate finance directed at climate change adaptation and mitigation in developing countries. This has led to efforts to establish national and subnational arrangements through which climate funded activities can be coordinated. The rationale is to ensure that climate funded programs are coordinated with each other, with other non-climate activities in related sectors, and with national and subnational policies and plans. The ideal aim of coordination is thereby typically to avoid duplication, promote integration and synergies, and ensure policy coherence.
Importantly, in developing countries efforts to develop climate finance coordination not only involve domestic actors, but often also multilateral and bilateral development donors. As contributors and managers of climate finance these actors not only participate in domestic coordination but may also promote and fund the establishment of national coordination mechanisms in the first place, thereby potentially taking on a significant role in the diffusion of climate finance coordination norms.
So far, academic studies of climate finance coordination have mainly focused on the global level, on why there is an absence of coordination, and how it may ideally be addressed. This work has discussed the highly fragmented nature of the overall global climate governance architecture (Zelli, 2011); the disconnects between different climate funds and donors (Pickering et al., 2017); the contrasting framings of what counts as climate finance (Weikmans & Roberts, 2019); and different perceptions of fairness in its distribution (Pittel & R€ ubbelke, 2013). Related work has shown the contrasting norm positions between those who approach climate finance as an extension of development aid, and those who frame it as a separate domain (Moore, 2012). A few studies have explored how this latter issue plays out inside donor agencies (L. Peterson & Skovgaard, 2019;Rai & Tanner, 2016).
Studies of ongoing efforts to actually create climate finance coordination are however very limitedespecially in developing countries, where the few existing studies tend to apply an efficiency perspective (van Rooij, 2014;Yoseph-Paulus & Hindmarsh, 2018). There has been little attention to our main emphasis here, that is how conflicting preferences among donors and government actors affect the ways in which the emerging norms of climate finance coordination are transformed into practical institutional arrangements at national and subnational levels (Clar, 2019). We suggest that the literature on norm localization can help illuminate this.

Climate finance coordination and norm localization
Within norm diffusion studies, World Society literature has examined how international systems may shape the diffusion of ideas and norms at the national level, thereby internalizing international norms in domestic settings and 'socializing' countries into global systems (Meyer, Boli, Thomas, & Ramirez, 1997). Related work has focused on international norm entrepreneurs, showing how international actorsincluding donor organizations and multilateral financial institutions such as the World Bankcan be significant proponents and bearers of international norms to developing countries (Kukkonen et al., 2018;Park, 2007).
However, the diffusion and evolution of international norms is rarely direct or smooth. Constructivist research on norm diffusion has in recent years sought to highlight the agency of domestic actors. Acharya's work on norm localization examined how local 'norm-takers' re-work international norms to fit local contexts (Acharya, 2004(Acharya, , 2011. This involves ' … the active construction through discourse, framing, grafting, and cultural selection of foreign ideas by local actors, which results in the former developing significant congruence with local beliefs and practices' (Acharya, 2004, p. 245). Alternatively, domestic actors may resist international norms through the development of counter-measures such as alternative practices and rules (Acharya, 2011).
The literature has explored different dimensions of this agency. Work on norm translation has emphasized how local actors re-interpret and adapt international norms to become meaningful in local cultural and political contexts (Berger, 2017). Related work on norm appropriation shows how domestic actors may harness and rework international norms to serve their own alternative interests. This may involve exploiting norms to leverage or consolidate authority vis-a-vis other domestic actors (Lorentzen, 2017), or re-interpreting norms as a means to counter external domination while still engaging broader normative orders (Großklaus, 2018;van H€ ullen, 2017).
Importantly, norm contestation does not necessarily imply outright rejection. Zimmermann et al usefully distinguish between contestation of the validity of norms or of the application of norms (Zimmermann, 2016;Zimmermann et al., 2017). The former questions whether the norm is fundamentally valid. The latter accepts the norm's basic premise, but questions whether its proposed meaning and/or shape is appropriate for a given context, and what actions are required. Domestic actors may therefore accept a norm's validity but contest the nature and form of its application. Shared acceptance of a norm's validity may thus only be the starting point for contestation over its practical enactment (Krook & True, 2012;Wiener, 2009).

Preferences and power in the shaping of climate finance coordination
As actors seek to shape the localization of climate finance coordination, they draw on differing preferences. These preferences are shaped partly by the actors' strategic interests, and partly by the prior organizational norms within which actors are embedded (Acharya, 2004;Lundsgaarde, Fejerskov, & Skovgaard, 2018).
A range of strategic interests are at stake for actors in the development of climate finance coordination. Coordination arrangements inevitably establish certain premises and framings that distribute the benefits and costs for those involved, thereby potentially favoring some actors over others. For some, a given coordination arrangement may provide leverage and increased command of resources, while for others it may lead to surrender of autonomous decision-making power and intrusion into organizational domains.
The preferences of actors in shaping climate finance coordination are however not only influenced by strategic self-interest, but also by the existing organizational norms of which they form part. This includes the internal norms of the individual organizations such as established practices and normative perceptions, typically formed by each organization's history and identity. Moreover, prior norms may exist for how different organizations relate to each other. As a case in point, the relationship between climate finance donors and African governments is emerging in a context where decades of development assistance have patterned the ways in which these actors interact (Scoville-Simonds, 2016).
When examining the interplay of actors' preferences, two things are key. Firstly, a purely voluntaristic perspective must be avoided. Actors are differently positioned to pursue their preferences, and power is therefore critical. This is not least so in public climate finance to developing countries, where multilateral and bilateral donors are strongly positioned visa-vis recipient countries (Ojha et al., 2016;Rai & Tanner, 2016). As providers of the funds, they not only automatically qualify as key players in the coordination process, but also have significant leverage in determining the nature of coordination arrangements.

Climate Finance Coordination From the Global to the Local 1903
Secondly, the power of actors in a localization process is rarely absolute and always contextual. While climate finance donors may be strongly positioned vis-a-vis African representatives in the boardrooms of the multilateral climate funds, the situation is different when these actors engage in domestic settings. In aid studies, a small but important body of literature has shown how recipient governments may employ a variety of strategies for 'managing' donors in order to shift interventions towards national political interests (Birkholz, Scherf, & Schroeder, 2017;Whitfield, 2008). Moreover, the way coordination arrangements evolve does not necessarily reflect the preferences of any one actor. Organizational conflicts may lead to outcomes that neither party really desired, or to 'ambiguous compromises' where actors have different interpretations of the arrangements in question (Mahoney & Thelen, 2009).

Methodology
The article draws on two main processes of data collection: From 2012 to 2016 one of the authors conducted research on the evolution of institutional frameworks for climate change finance in Zambia, including the PPCR program, under the Climate Change and Rural Institutions (CCRI) research project. From 2019 to 2020 both authors conducted field research focusing specifically on climate finance coordination within the Share or Spare (SOS) research project.
The article is based on (i) a total of 51 qualitative interviews with climate finance donors and domestic actors relevant to climate finance coordination in Zambia; (ii) study of PPCR project documents and progress reports and relevant government documents; and (iii) secondary literature on the political and institutional landscape of climate finance in Zambia.
We conducted 10 interviews in Washington, DC with individuals in the World Bank and other organizations connected to or knowledgeable about the PPCR and Climate Investment Funds, including CIF staff members, civil society representatives, donor agency representatives, and researchers. We conducted a further 41 interviews at national and subnational levels in Zambia with (i) national staff in the key government ministries involved in coordination (ministries of finance and planning, natural resources and environment, disaster management, and agriculture); (ii) staff members of subnational authorities involved in climate finance coordination (Rural District Councils and deconcentrated sector agencies); (iii) staff members in Zambian offices of the involved donor agencies (World Bank, UNDP, African Development Bank and key bilateral donors); (iv) members of local and national civil society organizations; (v) independent observers and researchers in Zambia. Interviews centered on permanent staff as these deal with long-term interactions with other actors but included questions on the activities of short-term consultants. The interviews at subnational level focused on three of the 25 PPCR target districts. 2 The key themes in the interviews were (a) how actors perceived climate finance and its coordination, and what their preferences were for its organization; (b) actors' coordination practices in relation to climate finance and generally; and (c) actors' specific roles and actions during the evolution of the PPCR program. The interviews used qualitative probing techniques and emphasized rich narratives of specific situations (Brinkmann & Kvale, 2018).
Data analysis had three main components: (1) Contextual analysis of the institutional landscape of climate finance in Zambia, using available literature and interviews; (2) Mapping the preferences of key donors and domestic actors vis-a-vis climate finance coordination. This was based on actors' written and verbal proposals for coordination; preferences expressed during our interviews; and their actions and practices in support of coordination arrangements; (3) A process study of the evolution of PPCR-supported climate finance coordination 2009-2020. Drawing on our interviews and document studies we charted the actions of donors and domestic actors over time, that is whether and how they supported, adapted or opposed climate finance coordination arrangements, and how these arrangements changed (Langley, 1999). The analysis further benefitted from contextual knowledge on the broader institutional dynamics of climate change in Zambia gained through the CCRI project (Funder, Mweemba, & Nyambe, 2018).

The PPCR and climate finance in Zambia
In the following we briefly describe the overall nature and context of PPCR support to climate finance coordination in Zambia, before moving on to examine the differing preferences involved.

The global Pilot Program for Climate Resilience
Initiated in 2008, the global Pilot Program for Climate Resilience (PPCR) is a multi-country climate change adaptation program under the global Climate Investment Funds. So far, approximately $1.2 billion in grants and loans have been allocated to PPCR activities in 28 countries in Asia, Africa and Latin America. The funds are managed by the World Bank and sourced from 14 donor countries, including the US, UK, and Japan. The global PPCR has two overall aims, namely (i) strengthening climate responsive development planning at national and subnational levels, and (ii) increasing the resilience of communities through support to specific climate change adaptation activities. Our emphasis here is primarily on the former, wherein support to cross-sectoral coordination arrangements at national and subnational levels is a key element.
Although a climate finance program, the global PPCR is built directly on the scaffolding of the development aid architecture. As the main trustee and manager of Climate Investment Funds, the World Bank is centrally positioned in the program's planning and implementation at both global and domestic levels. This has significant implications. Firstly, the World Bank's long experience in development aid informs its approach to the management of climate finance, includingas we shall seeits preferences for how and with whom finance should be coordinated. Secondly, the PPCR sits outside the UN framework for climate finance, replicating a long-standing political fault line between World Bank and UN structures in both development aid and climate finance. This has resulted in direct competition between the World Bank managed Climate Investment Fundsincluding the PPCRand the UN governed Green Climate Fund (Moore, 2012). The World Bank's efforts to promote the PPCR can thereby be seen as an attempt by a long-established development actor to expand its mandate and legitimacy into the climate change financing arena (Seballos & Kreft, 2011).

The Zambian climate finance context
In Zambia, the PPCR program was initiated in 2009, with a proposed end date of 2022 and a budget so far of USD 51 million. The program aims to strengthen Zambia's institutional framework for climate resilience at national and subnational levelsincluding coordinationand to support specific climate change adaptation projects among vulnerable communities in 25 districts.
The PPCR program in Zambia was not however initiated on a blank canvas, but in a context where decades of international funding has supported the evolution of the country's policy frameworks. Zambia was particularly dependent on foreign aid during the 1990s and 2000s, reaching as much as 56 per cent of GNI in the mid-1990s (Abrahamsen, 2000). Although overall aid to Zambia has been reduced in recent decades, the economy remains fragileas emphasized by a recent return to high foreign debt levels. Food insecurity and poverty rates are among the highest in Africa, and the population is highly vulnerable to recurring drought and flood hazards. 3 Climate Finance Coordination From the Global to the Local 1905 Public climate funding to Zambia rose from USD 68 million in 2015 to USD 113 million USD in 2018. While much of the funding is provided from bilateral sources to climate action in specific sectors, the main thrust in support to overall climate policies and frameworks have come from multilateral donors, including the World Bank. Long established with a physical presence in Zambia, the World Bank has been a major factor in Zambia's development funding for several decades and has at timesalongside the IMFhad substantial influence on the country's economic policies. Another key donor in the evolution of Zambia's climate change frameworks has been the UNDP, who since the Rio Convention in 1992 have provided financial support and technical assistance to the development of national agencies and policies on environmental management, disaster management and, since the early 2000s, climate change adaptation.

Formal objectives of the PPCR's support to climate finance coordination
Prior to the PPCR, there existed no formally designated mechanism for climate finance coordination in Zambia. A UNDP-supported 'Climate Change Facilitation Unit' within the Ministry of Lands & Natural Resources 4 acted as a go-to place for donors seeking to obtain an overview of government activities on climate change, but most project planning was done on a bilateral basis between individual donors and government departments, and amongst donors themselves.
However, with the arrival of the PPCR project in 2009, a more ambitious approach was proposed. A key component in the program was support to enhancing government coordination of incoming climate finance. The program's main planning document states that ' … the interest of external and internal partners in climate change activities is high and growing. In the absence of a coordinated national programme, it threatens to overwhelm Zambia's fragile capacity' (WB, 2011, p. 29). A key success indicator for the PPCR in Zambia was stated as: 'Harmonized mechanisms to coordinate and manage climate change funds developed and agreed' (WB, 2011, p. XII). Support to coordination was thus from the outset a key theme in Zambian PPCR activities. However, behind these overall aims were a variety of different preferences regarding how climate finance coordination should in fact be enacted, which soon became a contested issue.

Actor preferences in PPCR-supported coordination
In the following we discuss the differing preferences of donors and government actors that have pervaded and shaped PPCR supported climate finance coordination in Zambia. While there has been no controversy over the need for coordination, preferences have diverged sharply in terms of how coordination should be undertaken, that is, through what mechanisms, and who should lead coordination. These differences first emerged during an initial 'pilot' phase of the PPCR in 2009 but have continued during the course of the program.

Coordination preferences of the World Bank and other donors
As the main managers of the PPCR funds, World Bank staff in Zambia have significant influence on the development and implementation of the program and its support to climate finance coordination. As the PPCR got underway, they actively argued for their preferred approach to coordination.
In terms of how coordination should place, World Bank staff argued in meetings with government staff that if climate finance coordination was to succeed it would have to operate without the constraints of conventional bureaucratic frameworks and take the form of a cross-sectoral mechanism through which donors, government agencies and NGOs could freely negotiate and allocate climate funding. During our interviews World Bank staff expressed similar sentiments. For example, one leading staff member explained that ministries were 'silos that have to be broken down' in order to create 'a whole new playing field'. Another World Bank interviewee likened such an arrangement to ' … basically a marketplace … [where donors] … can turn up with funds and work out with the ministries how best to use them'. World Bank staff further emphasized the importance of efficiency and transparency in the coordination process. They were sceptical of embedding coordination too firmly in government mechanisms. Instead, they argued for a more independently facilitated coordination mechanism with minimal transaction costs, which would avoid politicisation and sectoral turf wars.
With respect to who should be involved in climate finance coordination, the World Bank saw donors as a natural and integral part of the coordination process alongside government agencies. In terms of domestic actors, the Ministry of Finance and Planning was considered particularly important by the World Bank, as was the finance and planning departments of Local Governments. The rationale for this was that climate finance coordination was essentially a question of development planning, and that these agencies were key in that respect. For the World Bank, a primary aim of climate finance coordination was to mainstream climate action into national development budgets and sector plans, thereby 'integrating climate resilience into development strategies and programmes' (WB, 2011, p. 2), as stated in one key PPCR project document. In our interviews, World Bank staff explained how other agencies, such as environmental and agricultural ministries, were important for technical input but that only finance and planning departments could bring about real change because 'development work is climate work' as one interviewee put it.
These preferences were further reflected in specific proposals by the World Bank for how climate finance coordination could be supported and arranged in Zambia: For national level coordination, the World Bank proposed establishment of a semi-autonomous secretariat which would act as a clearing house for climate finance, charged with facilitating coordination and interaction between donors, government agencies and NGOs. The secretariat would function under the auspices of the Ministry of Finance and Planning but operate and be located outside established government mechanisms for sectoral coordination. At the subnational level, the World Bank proposed establishment of multi-disciplinary 'Project Teams' consisting of Local Government planners and technical staff drawn from local branches of the ministries of agriculture, forestry, community development. The teams would be anchored in the planning and finance departments of Local Governments, but again would operate outside existing institutional mechanisms.
Apart from reflecting normative perceptions, these proposals were also strategically enabling for World Bank interests. An open coordination process facilitated by a semi-autonomous secretariatas opposed to simply embedding coordination in existing government functionsensured the World Bank and other donors a certain room for maneuver. It also allowed them to better monitor funding and transparency, thereby addressing long-standing donor concerns about corruption and delays in aid (Faust, 2011) and more recently in climate finance (Williams & Dupuy, 2019). Anchoring coordination in the Ministry of Finance & Planning furthermore allowed the World Bank to draw on its long-established connection to this ministry. This provided leverage and opportunities to sustain collaboration. As one World Bank staff member said: 'We know those guys and they know us'.
The World Bank's proposals were supported by a variety of other multilateral and bilateral climate donors in Zambia. One notable donor was however skeptical. For the UNDP, a firmly government-led coordination process was critical, and the notion of a semi-autonomous climate secretariat was therefore considered problematic. Moreover, having already supported a Climate Facility Unit in the Ministry of Lands & Natural Resources for a number of years, the Climate Finance Coordination From the Global to the Local 1907 UNDP was skeptical of anchoring coordination elsewhere. This hinged on more than mere protection of prior investments. The Ministry of Lands and Natural Resources was the formal anchoring point for the UNFCCC process in Zambia, and thus for the UNDP a natural lead on climate finance coordination.

Coordination preferences of central and local government agencies
The World Bank's coordination preferences were met with reluctance by domestic actors. At the national level, there was general skepticism towards the proposal that coordination should be undertaken outside of established government mechanisms. Staff in national ministries and agencies saw coordination as an integrated part of the regulatory work that formed part of government planning and decision-making procedures. They pointed to existing government mechanisms for general coordination in which sector agencies meet to align plans and argued that climate finance coordination could be integrated within these. During our interviews, a representative of the Ministry of Finance and Planning stated that once climate finance had been provided to a country, its further allocation was 'the country's own matter as a sovereign state and government'. The Ministry of Lands & Natural Resources echoed this, arguing that climate finance was a feature of UNFCCC agreements and its associated commitments to developing countries, and should therefore not be subject to the influence of development donors.
Differences between ministries were however also apparent. Although the Ministry of Finance & Planning was unhappy with the notion of a semi-autonomous coordination secretariat, it wasnot surprisinglysupportive of the idea that the secretariat should operate under the Ministry's own auspices. Like the World Bank, this Ministry saw climate change financing as a natural extension of decades of development aid and felt it should logically be integrated in national development planning. Other ministries and government agencies were however opposed to this. The Ministry of Lands & Natural Resources was particularly critical. During high-level meetings, representatives of this ministry argued that climate change was primarily a technical matter and should be coordinated by organizations knowledgeable of the field. As one interviewee stated: 'Things can also become too mainstreamed'. Accordingly, the Ministry of Lands & Natural Resources suggested that its own existing Climate Facilitation Unit should be lead on climate finance coordination.
At the subnational level, the involved authorities also had concerns. At district level, Zambia has a dual institutional structure consisting of democratically elected local governments and deconcentrated branches of the central government's sector ministries. While local governments appreciated the World Bank's suggestion to anchor coordination in their quarters, local sector agencies preferred the District Commissioner as the lead on coordination, who represented central government. Of particular concern to both sides, however, was the risk that the Bank's proposed arrangements would fuel underlying tensions between local governments and local sector agencies and upset their existing means of coordinating. These included a formal District Development Coordination Committee, and informal networks through which local practitioners collaborated. We return to this below, but a quote from a local agricultural officer in one district expresses the sentiment: 'We have our ways of working here [ … ] We solve problems between us. We don't want donors to change that'.

Divergence of preferences
Overall, the differing preferences between the actors involved were thus not about norm validity, that is, that climate finance coordination was a good idea, but rather norm application, that is how coordination should be done in a Zambian context. In this respect, the preferences varied across two major dimensions. Firstly, the World Bank articulation of coordination as a 'market-like' arena with limited bureaucratic control stood in contrast to that of government agencies, for whom coordination should be seen as an integral part of the standard processes and planning procedures of the state. Linked to this were World Bank interests in ensuring room for maneuver and efficient aid delivery, vis-a-vis government concerns to protect the autonomy and integrity of formal and informal governance mechanisms.
Secondly, a distinction was evident between actors who articulated climate finance coordination as an extension of past institutional practices in development aid (the World Bank and Ministry of Finance and Planning), and those who argued that it should be developed as a new institutional domain specific to climate governance (most notably the Ministry of Lands and Natural Resources and the UNDP). Intertwined with this were the individual strategic interests of sector agencies and donors to sustain or expand their role as legitimate actors vis-a-vis climate finance.

The evolution of PPCR-supported coordination
A number of different coordination preferences were thus evident from the outset of the PPCRprogram in Zambia. In the following we examine how the different actors have sought to assert their preferences during the period 2009-2020.

Donor domination and government pragmatism (2009-2013)
In pursuit of their preferred coordination arrangements, World Bank staff relied heavily on strategic engagement with powerful actors through their organization's political and economic leverage in Zambia. This included organizing meetings with high-ranking Ministerial staff and arguing their case during negotiations with the government during subsequent phases of the PPCR, thus linking it to possible future funds.
The Ministry of Finance & Planning, on their part, pursued a strategy of pragmatic opportunism. While the Ministry would have preferred the World Bank's proposed coordination secretariat to be fully embedded within its own organization, its leadership and staff were well versed in donor concerns over transparency and efficiency, and eventually accepted the World Bank proposition as a trade-off that could be re-opened later. Such pragmatism in dealing with donors was not new, but rather a well-established tradition within the Zambian state (Fraser, 2008).
As a result, the Ministry of Finance & Planning accepted the World Bank's proposal for climate finance coordination. In 2013, the two parties entered into an agreement in which coordination was to take place through a semi-autonomous secretariat dubbed the 'Interim Climate Change Secretariat', financially supported by the PPCR and physically located outside the Ministry of Finance Planning, with separate budgets. The Ministry of Finance and Planning was the main institutional anchoring point for the secretariat. Staff from other ministries were to be seconded to the secretariat, tasked with facilitating coordination.
At the subnational level, the involved local government authorities also accepted the World Bank's proposed coordination arrangement. Financially constrained yet charged with considerable developmental mandates, a substantial support program such as the PPCR was a major opportunity that could not be dismissed. During our interviews, local government staff recounted how their meager budgets usually left them with little option but to accept donor rationales, and then 'deal with the problems later' as one staff member said. The PPCR project was thus anchored in local governments, and the cross-sectoral 'Project Teams' proposed by the World Bank were established.

Climate Finance Coordination From the Global to the Local 1909
It appeared, then, as if the World Bank had succeeded in shaping Zambia's developing climate finance coordination mechanism according to its preferences. However, subsequent events proved otherwise.

Undermining donor-driven coordination (2013-2015)
Despite initiation of the new coordination mechanism in 2013, it ceased to gain widespread legitimacy. The mechanism was not popular with several ministries and agencies at the central level, who resisted the semi-autonomous secretariat and the role of the Ministry of Planning and Finance. Most notably, the Ministry of Lands & Natural Resources continued to operate its own existing climate finance coordination unit, thereby indirectly asserting its claim as the rightful lead on the matter. In this they were supported by the UNDP, who continued to fund the unit. Other ministries and agenciesincluding the Ministry of Agriculture and the Disaster Management and Mitigation Unitformally agreed to align with the semi-autonomous secretariat, yet at the same time continued developing direct financing agreements with donors independently of the new coordination secretariat. They furthermore protracted the secondment of senior staff to the World Bank-backed secretariat, thereby restraining the secretariat's capacity, and limiting its channels for everyday coordination and liaison with ministries. Staff members of some departments furthermore excluded the secretariat from involvement in everyday informal coordination, by not informing them of ad hoc meetings or arising matters.
At subnational level, the PPCR-supported coordination mechanism was also challenged. In 2014, the cross-sectoral PPCR 'Project Teams' began implementing community-level adaptation projects in the target districts within farming, water management and alternative income generation. In order to comply with PPCR funding conditions, such activities could only be carried out in areas identified as particularly climate vulnerable. This led to objections from local government politicians outside the targeted areas, and complaints from District Development Committees that their authority to coordinate district-wide planning was being sidelined. Central level ministries furthermore held back on allocating salary and time for their local staff to join the cross-sectoral PPCR teams, seeing little incentive to allocate scarce resources to a coordination process anchored in local governments rather than in their own local branches.
By not engaging with the new mechanism, withholding staff, pursuing alternative mechanisms, and raising questions of authority, domestic actors thus gradually undermined the legitimacy of the new climate finance coordination mechanism and the World Bank preferences it was built upon. This significantly weakened the new coordination mechanism's ability to function effectively. Although some donors did channel funds through it, most did not. What was intended as an all-encompassing coordination arrangement came to operate in piecemeal fashion alongside internal government coordination practices and the competing UNDP-supported coordination mechanism.

Embedding coordination in government arrangements (2015-2016)
In 2015, a general re-organization of the ministerial landscape triggered a further development. The Ministry of Finance & Planning was split into two separate ministries, and the former ministry's role in climate finance coordination was transferred to the new Ministry of Planning. The new ministry continued the alliance with the World Bank but argued that it was time to integrate the semi-autonomous coordination secretariat fully within the new ministry. Facing the possibility of an increasingly marginalized national climate coordination mechanism, the World Bank agreed. The semi-autonomous climate secretariat was disbanded, and the coordinating functions moved to the new Ministry of Planning.
This shift to a more government-controlled coordination mechanism was consolidated and expanded upon in Zambia's first national climate policy, finalized in 2016, which set out the roles of different actors in climate finance coordination. In principle, all new climate finance proposals in Zambia now had to pass through assessments in appropriate Technical Committees, andif acceptedrecommended for approval at higher levels. A key feature of this arrangement was that it involved other government agencies in coordination to a greater extent, thereby reducing their grounds for resistance. While the Ministry of Planning was central to the process, other agencies including the Ministry of Lands & Natural Resources were allocated roles as chairs and secretariats for the coordinating committees. Final decision-making power was located in the Council of Ministers, thus allowing all agencies influence at this level. The previously semi-autonomous climate secretariat initiated by the PPCR was thereby replaced with a mechanism much more firmly embedded in government mechanisms.
At the subnational level a similar process of localization took place, although in more informal ways. As the cross-sectoral local 'Project Teams' became increasingly contested and restrained in their work, the participating staff began to gradually re-configure the PPCR initiated coordination mechanism into less contentious and more locally adapted forms. This included side-stepping inter-organizational conflicts by drawing on and expanding the existing informal personal and professional relationships that characterize the everyday work of local bureaucracies in many African countries (De Herdt & Olivier de Sardan, 2015). Through these relationships, local staff developed casual, low-cost but regularized coordination practices such as informal phone briefings to align PPCR coordination with meetings in the District Development Committee. In some cases, the Project Teams set up informal working groups to overcome tensions and coordinate between local agencies. In everyday practice, the Project Teams also expanded the definitions and boundaries of 'climate vulnerable areas', thus expanding their work to non-PPCR communities and aligning better with district plans.
Subnational staff in the PPCR Project Teams thereby drew on the discretionary room for maneuver available to street level bureaucrats worldwide (Lipsky, 2010), and on the 'practical norms' through which local bureaucracies in African settings creatively manage their complex role at the frontline of the state (De Herdt & Olivier de Sardan, 2015). At both central and local level, the original coordination mechanism was thus gradually reworked and innovated by domestic actors to better align with their preferences for a coordination process which aligned better with existing formal and informal practices and enhanced domestic sovereignty in climate finance coordination.

Continued contestation (2016-2020)
At both the national and local levels, the PPCR supported climate finance coordination arrangements thereby evolved from being initially strongly dominated by World Bank preferences towards being morebut not completelyadapted to domestic preferences and practices. While these arrangements remain active at the time of writing, efforts to circumnavigate and further develop them continue. At the national level, the Ministry of Lands & Natural Resources have persisted in developing their own national coordination unit in parallel. Meanwhile, other Ministries and agencies continue to develop climate projects with donors on a separate bilateral basis, arguing that the PPCR supported mechanism is intended mainly for multilateral climate funds.
While accepting the current situation as a necessary compromise, World Bank staff harbor hopes for a return to the more semi-autonomous coordination secretariat than they originally promoted. During our interviews, they pointed out that when the current policy expires, there will be an opportunity to propose a new arrangement. Meanwhile donors conduct informal meetings amongst themselves to coordinate climate finance, despite requests by the government to do so via government frameworks. At the subnational level, the informal practices of local government planners and technical sector staff have managed to subdue the initial contestations Climate Finance Coordination From the Global to the Local 1911 over PPCR coordination arrangements and have gradually established stronger links between PPCR supported activities and existing district planning processes.

The reshaping of climate finance coordination through norm localization
Our findings show how donors and government agencies have pursued diverging preferences for climate finance coordination in Zambia. Whereas the World Bank has emphasized facilitation of a free-flowing interaction between donors and domestic agencies outside the conventional bureaucracy, government agencies have sought to ensure sovereignty of the state in the coordination process, and continuation and evolution of existing formal and informal coordination practices.
The initially very dominant role of the World Bank in shaping the PPCR-supported climate finance coordination in Zambia corresponds to studies of PPCR implementation in Mozambique and Nepal (Rai & Tanner, 2016). Broader studies in other settings have likewise illustrated how donor agendas may shape the framing of national climate change policies (Ojha et al., 2016;Taylor, 2014). Our findings add to these insights by showing how donor preferences can also be carried forward into domestic climate finance coordination arrangements, which form part of how climate finance is governed at national and subnational levels.
Yet our findings also show that, over time, the PPCR-supported climate finance coordination in Zambia has evolved from being initially dominated by World Bank preferences towards increasingly corresponding to domestic preferences. This adds an important dimension to our understanding of the way climate finance coordination arrangements are emerging, namely the active work of ministries and local governments in norm localization. In Zambia this has taken two overall forms, that is (i) gradually re-configuring the initial PPCR supported arrangements, and (ii) undermining and competing with the PPCR supported arrangements.
The efforts of domestic actors to challenge the World Bank's finance coordination preferences correspond to studies of norm localization in other fields. This includes engaging in subversive rather than overt forms of contestation (Lorentzen, 2017), and either re-working the norms of international actors or establishing competing framings (Acharya, 2004). Our findings further resonate with work in development studies on how African bureaucracies seek to discretely reshape aid interventions as they are enacted in practice (Birkholz et al., 2017;Whitfield, 2008), suggesting that such strategies may be carried forward into the shaping of climate finance coordination.
In Zambia these actions are, we suggest, more than attempts to simply translate norms to the local context and ensure congruence. Common to all domestic actors in the process described here is an effort to counter World Bank preferences in defense of domestic bureaucratic norms and interests. While often pragmatically done, it is nevertheless persistent. The work of domestic actors in Zambia can thus be seen as attempts to engage a normative order whilst minimizing external actors' control over it, similar to Grossklaus' findings of gender norm localization in Nigeria (Großklaus, 2018(Großklaus, , p. 1255; see also van H€ ullen, 2017).

The divergence of preferences beyond the donor-government divide
Significantly, our findings show how the localization of climate finance norms in Zambia is shaped not only by conflicting preferences between donors and domestic actors, but also by diverging preferences within these two groups of actors. In this respect, a distinction can be seen between actors who see climate finance coordination as an extension of the development architecture (that is, the World Bank and ministries responsible for finance and planning), and those who see it as a new institutional domain under the UNFCCC (that is, the Ministry of Lands and Natural Resources and UNDP).
These differing positions echo broader normative debates on whether global climate finance is a form of aid or a separate obligatory compensation (Moore, 2012). However, they also reflect strategic interests. While our emphasis here has been on the World Bank, the competing efforts of the UNDP to support parallel coordination arrangements illustrate how broader global struggles between World Bank and UN-affiliated climate funds may play out in recipient countries, and how donors may compete to diffuse the same international norm of climate finance coordination through alternative variants.
Likewise, for domestic actors in Zambia climate finance coordination represents an opportunity to upset or consolidate existing power relations in the domestic institutional landscapenot only with respect to climate financing but also in terms of how authority and mandates are distributed more broadly in the governance of sectors and resources. This diversity of domestic preferences echoes recent findings from studies of norm diffusion in other areas of natural resources governance (Gustafsson, Merino, & Scurrah, 2020;Macdonald & Nem Singh, 2020).

Localization of climate finance norms as a co-produced process
Our findings suggest that the diffusion of climate finance coordination norms may have other outcomes than either full adoption or full rejection by domestic actors. The situation so far in Zambia shows a situation in which domestic actors accept the fundamental validity of the norm that is, that climate finance coordination is a good ideabut have sought to reshape its form according to their own preferences. This corresponds with the points by Zimmermann et al. that domestic actors may not necessarily contest the fundamental validity of norms, but rather their particular application; and that outcomes may constitute a middle ground between rejection and adoption (Zimmermann, 2016;Zimmermann et al., 2017).
In extension of this it would be wrong to conclude that donors have been entirely sidelined in the evolution of climate finance coordination in Zambia. Firstly, it is clear that the current coordination arrangements still contain elements of World Bank preferences. These arrangements have therefore effectivelybut not intentionallybeen co-produced by donors and domestic actors, similar to what has been shown in some aid interventions (J. Peterson, 2012). Secondly, the current arrangements are not stable, but in flux and remain contested by donors and domestic actors. The localization of climate finance coordination in Zambia is continuously evolving, shaped by struggles between different preferences and practices of donors and government actors.
In this respect, history does matter. In the development of climate finance coordination in Zambia, past relationships and modes of interaction between donors and government actors are carried forward from the development field to that of climate change. Donors continue to influence the process, and governments continue to localize arrangements as far as possible. Nevertheless, a deterministic perspective must be avoided. The advent of climate financing also introduces new institutional dynamics and normative framings, as evident in the efforts by the Ministry of Land and Natural Resources to contest that climate finance should be coordinated by donors and ministries of finance and development.

Conclusion
Our study shows how emerging international norms for the coordination of climate finance may be subject to a dynamic process of norm localization as donors and domestic actors pursue conflicting preferences in the enactment of coordination in developing countries. In Zambia, efforts by the World Bank to steer climate finance coordination in a particular direction have been gradually reconfigured by national and subnational agencies. While the validity of the norm has been accepted, its practical application has not and remains contested.
Climate Finance Coordination From the Global to the Local 1913 Our findings further highlight how donors and domestic actors may have widely differing preferences in the diffusion of norms on climate finance coordination, and how preferences may also differ within these groups. This divergence in preferences is influenced partly by differing organisational norms and associated perceptionssuch as the role and efficacy of the state in climate finance coordination, and whether climate finance should be part of the development architecture or separate from itand partly by differing strategic interests such as opportunities for challenging or consolidating power relations between and among donors and domestic actors.
Lastly, our findings show how both donors and domestic actors may draw on and reproduce patterns of interaction inherited from development aid as they go about enacting climate finance coordination norms. Existing organizational norms and strategic interests from development cooperation may thus be carried forward into climate finance, although we also find examples of efforts to challenge these past relationships by contesting that climate finance is a development issue.
Our study thus highlights the dangers of assuming that international norms around the governance of climate finance are diffused and adopted in a straightforward manner in developing countries. We suggest that analysis of such processes can benefit from examining how these norms are contested and reshaped as they are enacted in domestic settings, and how they evolve over time. In so doing, attention to conflicting preferences between donors and domestic actors is important, but our findings also highlight the importance of moving beyond a simple 'donors vs recipients' divide to examine the diverging preferences and strategic alignments amongst different donors and domestic actors.
From a normative perspective, we sympathize with calls for better climate finance coordination, but suggest that climate donors and governments should abandon any notion that climate finance coordination is merely an administrative process, and instead acknowledge it for the political space that it is, and as a component of climate governance. Doing so will help focus attention on ensuring that climate finance coordination processes are broadly inclusive of government and non-government actors, and that they evolve around explicit interest articulation and negotiation as per other political processes. Notes 1. We define coordination as a process aimed at aligning interdependent tasks. See Section 2. 2. The districts have been anonymized due to the informal activities discussed in the article. 3. For a discussion of the broader politics of climate change adaptation in Zambia see (Funder et al., 2018). 4. This ministry was re-named during the period studied. For clarity we use the current name throughout.