Skip to main content

Conflicts in multi-level governance: an analysis of international climate policy implementation at the sub-national level


The complexity and multi-level nature of climate change requires governance systems able to manage and resolve conflicts across multiple scales. Drawing on the theories of collaborative governance and conflict resolution, this paper proposes a theoretical framework to understand three types of conflicts likely to occur in a multi-level governance (MLG) network: conflict in problem framing, conflict in benefit sharing, and conflict in capacity building. Using the implementation of the clean development mechanism in China as a case study, the paper finds the three types of conflicts emerged between supranational and subnational levels. The finding further indicates the role that the Chinese national government played in resolving conflicts through reinterpreting international norms, aligning presumably conflicting goals, eliminating the difference in benefits sharing, and narrowing capacity gaps across levels. The evidence suggests an improved MLG should be dedicated to integrating local level interests in centrally dominated policy processes. This study contributes new knowledge on MLG by investigating the hierarchical relations among participants and how authority and power differentials can shape their interactions across levels.


Modern societies are characterized by increasing complexity. Non-linear processes in economy, society, ecology, and others are hardly controllable, especially because of their interdependencies. The analytical framework of multi-level governance (MLG) has been applied to understand issues characterized by complex interactions and tight coupling—climate change governance, for example (Di Gregorio et al., 2019; Huang, 2019). Climate change governance has evolved into a complex polycentric structure that spans from global to national to sub-national levels, relying on both formal and informal networks and policy channels (Jordan et al., 2015). Such a complex governance structure reflects the ‘glocal’ nature of climate change: its distinct impacts are felt at and its solutions involve multiple levels of governance (Gupta et al., 2007). Although MLG has been applied to understand decentralized governance in promoting local initiatives, shifts in power and authority relations, and benefits achieved at multiple scales (Biermann et al., 2016; Underdal, 2010), many more theoretically significant questions are yet to be broadly discussed, especially in regard to the ‘politics of scale’ (Bulkeley, 2005). In the face of a multilevel challenge, governments at different levels have to develop robust and coherent strategies that build up to joint actions. A crucial scientific as well as practical question is how to take into account their diverging ambitions and capacities. Complex governance, network collaboration, and meta-governance theories propose strategies to manage a network of actors participating with various interests around a specific policy issue while most of them ignore the relations between actors with hierarchical relations (van Buuren & Gerris, 2008; Koppenjan & Klijn, 2004; Sørensen, 2006). How should the opinions from different hierarchical scales on problem framing and benefits sharing be reflected within the MLG framework? How can power and capacity differentials across jurisdictions influence their interactions (Young, 2006)? Who should take the initiative in resolving conflicts in MLG, and how?

This paper addresses these concerns by focusing on the implementation of the clean development mechanism (CDM), an international climate policy. CDM is the only mechanism embracing both developed and developing countries under the Kyoto Protocol, which allows an empirical study of climate-vulnerable developing counties with relatively fragmented or fragile institutional arrangements (Ishtiaque et al., 2021). China, as the biggest host country of the CDM, has rich experience in hierarchy interactions among participants for CDM project application and management. It can thus provide evidence on how an MLG network can successfully resolve conflicts across scales and sustain effective governance. To study the conflicts in MLG, this paper proposes a theoretical framework that combines conflict resolution theories (Page et al., 2018; Piatak et al., 2018), and the Collaborative Governance Regime (CGR) model raised by Emerson and colleagues (2012). The main research question of the present paper is twofold. First, how may conflicts occur in the CDM implementation at national and sub-national level? Second, who has played the role of conflict resolver in facilitating governance? The findings shed light on the deficiencies of the existing MLG networks for effective governance, and emphasize the role of national governments in orchestrating the MLG system.

Both quantitative and qualitative data were collected in this study. Quantitative data regarding CDM projects in mainland China was extracted from a global CDM data set. By identifying the location of each project through its Project Design Documents, this study built a dataset of 4429 CDM projects for the period from 2004 to 2015. Qualitative data was extracted from policies, documents, project papers, meeting minutes, and records from the United Nations and the Chinese governments.

Theoretical framework

Featuring traits of intergovernmental collaboration (IGR), MLG confronts various types of conflicts between governments (Huxham & Vangen, 2005). While IGR often looks at a few governmental agencies at the same level, MLG is prone to encounter more difficulties in making a deep, non-transactional collaboration as it involves governmental entities from two or more hierarchical levels (Ongaro et al., 2019). The network is constituted by a hierarchical structure extending beyond nation state countries and also extending to lower elements of the political and jurisdictional level. On each governance level, autonomous processes with their own dynamics are running, but these processes are also mutually connected or nested. In this complex and nested system, cross-level linkages have increased participants’ interactions and conflicts in building up joint actions (Duit & Galaz, 2008; Popering-Verkerk et al., 2015). Two questions then arise that need to be explored. First, what types of conflicts is MLG likely to encounter? Second, how do MLG actors address or cope with such conflicts?

This study borrows the concept of conflict from the work of Lan (1997), regarding conflicts as the opposing/differentiated ideas and actions of different entities that may ‘result in potentially destructive effects’ on collaboration outcomes. Conflict resolution literature classifies conflict into subjective conflict (e.g., cognitive conflict) and objective conflict (e.g., behavior conflicts) (Lan, 1997). Conflict resolution strategies are thus different due to the nature of conflicts. Drawing on the theories of conflict resolution and collaborative governance (Emerson et al., 2012; Piatak et al., 2018), the following section provides clues to understanding three types of conflicts in MLG and discusses the role of key actors in conflict-solving in MLG networks.

Conflicts in problem framing

MLG has been applied to remedy complex public problems with multiple actors. These actors normally have a different understanding of problem framing, from problem definition and root causes to acceptable goals and solutions. These conflicts may manifest as competing aims among actors (Huxham & Vangen, 2005) or as difficulties in gaining agreement on the definitions of and solutions to problems. MLG thus encounters conflicts regarding perceptions of problems, given that multiple institutional logics (Herranz, 2008) and accountability relationships (Thomson & Perry, 2006) often embody conflicting values. As Hoornbeek and Peters (2017) make clear, if a ‘wrong’ definition is adopted, it may mean that the ultimate ‘solution’ will be delayed or missed altogether. The CGR model thus calls for an iterative process for collaboration partners to build shared meaning and foster principled engagement, a key dimension of collaborative dynamics. Principled engagement consists of four basic process elements: discovery, definition, deliberation, and determination, in which initial success of discovery and definition is essential to the subsequent procedure.

Achieving principled engagement among participants is never easy for a controversial issue. Climate change is a ‘wicked’ problem exhibiting problematic characteristics including scientific uncertainty, public ambivalence, and significant lag in the effects of policy interventions (Jordan et al., 2015). It is highly cross-sectoral and cuts across jurisdiction borders, challenging the established institutional frameworks and having few obvious solutions. Because a self-reinforcing path dependency spiral determines the way people frame the problems, actors in MLG often struggle to frame the climate change issue within problems that suit their pre-existing political interests or policy competences (e.g., framing it as an energy, health, technology, or diplomatic policy). In that way, conflicts in problem framing may challenge accountability, because it determines the extent to which MLG is responsive to the claims of all authorizers and stakeholders (Page et al., 2018).

Uneven development and power imbalances across regions and scales also add to the challenges in building shared meaning of climate change governance. The breadth and complexity of the climate change issue span such a huge number of different sub-issues—ranging from poverty alleviation in less-developed places to long-term energy supply decisions in industrialized counties—that normal channels of inter- and intra-state policymaking struggle to cope. Policymakers and implementers may have distinct perceptions and definitions of climate issues due to their different socio-political environments. Ultimately the choice of how the climate change problem is handled within a jurisdiction is a reflection of the strength of the interests and power of the actors who define the problem. That has culminated in slow and sometimes even deadlocked negotiations at the international level (Biermann et al., 2016).

In addition to horizontal regional differences, conflicts in problem perception are generated across governance levels. Understanding climate issues requires consideration of the social construction of appropriate levels by institutions to further their own aims. An action that is successful for one level of government may not be classed as successful by another. For example, supranational/national actors (e.g., the European Commission or China) might be more concerned about the political and diplomatic attributes of international climate cooperation. On the other hand, for localities, climate governance is more related to environmental, economic, and social well-being. Such vertical differences in problem discovery and definition might debase future deliberation and determination, and thus hamper joint potential to achieve collaborative advantage (Huxham & Vangen, 2005).

Therefore, like-minded actors from different levels with their compatible belief systems are important elements of effective MLG. Conflicts in problem framing are normally subjective conflicts where explanations, communication, provision of scientific information, consensus-building, and rapport-building can effectively contribute to the resolution of the conflict. In CGR models, Emerson and Nabatchi (2015) highlight conflict resolution skills as requisite competencies for principled engagement as people have ‘differing substantive, relational, and identity goals to collaborate’ and solve problems (p. 58). These conflict resolution skills may include articulating common purposes and objectives, or clarifying and adjusting tasks and expectations for one another.

Conflicts in benefit sharing

The second dimension of collaborative dynamics in the CGR model is shared motivation, which focuses on how governance outcomes may benefit participants and the organization or constituency they represent (Emerson et al., 2012). The shared motivation component is a self-reinforcing and cyclical progression of four interactive elements: trust, mutual understanding, internal legitimacy, and commitment.

Trust and mutual understanding are essential to establish ‘boundary works’ that eliminate differences, address tensions, and facilitate communication and interaction across different levels (Piatak et al., 2018; Quick & Feldman, 2014). Actors in an MLG system have both common and diverse goals, while intrinsic tension between self-interest (e.g., economic development) and a collective interest (e.g., global carbon emission reduction) always exists (Thomson & Perry, 2006). Even with the hope of achieving a Pareto optimal improvement, winning and losing is an inescapable aspect of governing. Trust is a sine qua non for boundary works and is instrumental in reducing transaction costs of repeated interactions, improving stability in relations, and stimulating learning. Trust enables people to go beyond their own personal, institutional, and jurisdictional frames of reference and perspectives and toward understanding other peoples’ interests, needs, values, and constraints (Thomson & Perry, 2006). If the structure of cross-scale linkages reduces trust, then the robustness of an MLG network is in question.

The persistence and stability of collaboration also depend on the distribution of benefits from these cross-level linkages, which is demonstrated by the ability of the system to command legitimacy among participants. To promote a common vision and achieve accountability with all actors, an MLG system needs to reinforce confidence in internal legitimacy that enables participants to justify their continuing engagement to those they represent. An efficient MLG network thus has to choose which costs and benefits to take into account. But it would be difficult to decide how to allocate them and whether any losers should be compensated. Carbon trading for example, as a redistributive climate mitigation policy, indicates efforts to shift the allocation of wealth, income, property, or rights between countries (Caplan et al., 2003). If the intended losers in redistributive policies are the ‘haves’ (e.g., developed countries with the financial and technology resources to purchase carbon credits), they also have ample resources to resist such policies. Meanwhile, given the long-term benefits of environmental investments, it would be imprudent to assume sellers of carbon credits are the winners, because the performance of green financing or technology transfer often takes years to have effect. These challenges and conflicts in legitimizing participants’ benefits distribution may thus hamper their commitment in future collaboration.

Conflict in benefit sharing is normally a mixed conflict in which conflicting parties may gain by cooperating around some common interests while competing for resources that come only at the expense of the other (Brickman, 1974). Conflict resolution techniques, such as clarification of benefits and costs, negotiation, mediation, arbitration, bargaining, and alternative choice provision, can be used. In the CGR model, shared motivation can be achieved through facilitating the alignment of presumably conflicting goals within a hierarchy and keeping all actors, with relevant and significant different interests, accountable and satisfied with the governance outcomes (Romzek et al., 2012). Accountable benefits sharing is instrumental in enhancing trust between actors, as well as essential for improving stability in relations and stimulating the exchange of knowledge and information (Koppenjan & Klijn, 2004). Actors’ compatible and interdependent interests legitimize and motivate on-going collaborative actions and determine the performance of MLG.

Conflicts in capacity building

MLG intends to generate outcomes that no one level of governance could accomplish individually. However, to collaborate, MLG actors must generate enhanced or new capacities for joint action that did not previously exist. Capacity building enables MLG actors to accomplish their collective purpose. Capacity for joint action is the third and last dimension of collaborative dynamics. Borrowed from Emerson et al.’s (2012: p. 68) definition in the CGR model, capacity for joint action is composed of procedural and institutional arrangements, resources, leadership, and knowledge.

Capacity for joint action can be highly asymmetric with the powerful often obtaining more. Asymmetrical capacity exists between the international and subnational levels (provinces, cities), and conflicts associated with these asymmetries may eventually undermine governance structures. Sub-national governments are essential in MLG, as they are where international and national policies unfold and de facto implemented, and closer to where the consequences of climate change will be felt. However, due to the lack of motivation and capacity, subnational levels remain marginal in supranational or national policy decisions. As national governments are those that generally have an international commitment to climate policy, subnational governments may leave these policies to the upper levels or make more efforts to reach out to the national level for help. Overloaded with other local demands, subnational governments may put climate governance lower on their list of priorities, especially when lacking formal administrative institutions (Oliveira, 2009). The logic of economic rationality makes subnational governments more motivated to be the free-riders in global climate governance (as public goods). They have no incentive to implement certain climate policies if others do not, and they also lack the leadership to leverage and redistribute the shared knowledge and resources to effect common goals (Kousky & Schneider, 2003).

Examining cross-level dynamics could entail the challenge of integrating knowledge produced at differing scales that are characterized by quite disparate disciplinary approaches. The limited capacity of the subnational governments in MLG has concerned both scholars and practitioners (Ter-Minassian, 2020). Evidence shows that due to restricted cash flow, most local governments in the least-developed countries are not able to contribute effectively to climate change governance (Jachson & Coninck, 2019). Capacity building for subnational governments is determined by the availability of different forms of knowledge and resources, including but not limited to financial, political, and human resources (Ostrom, 2011; Williams et al., 2020). Embedded in local capacity building are associated values such as flexibility (possibly more achievable at local levels), accountability (public engagement is often harder to organize at higher spatial scales), and transparency (which is possibly easier to achieve locally).

Conflicts in capacity building pertain to ‘cooperation conflicts’ 1997, which refer to a situation in which the rewards and punishments that can accrue to each party are non-competitive (Lan, 1997: p. 29). Conflicts are more about how to achieve the goal. In a ‘shared power’ world there is a need for coordination between various levels of governments. An effective MLG system should facilitate the flow of resources across levels and jointly needed knowledge to be generated together. Knowledge and other resources need to be disseminated among a network of actors. The perceived and real fairness, legitimacy, and efficacy of MLG would depend on how well these capacity differences are managed.

Who could be the conflict resolver?

Conflicts as discussed above call for connective or integrative efforts to coordinate and synchronize joint actions from different levels (Chrislip, 2002; Crosby & Bryson, 2010). Who, then, could be the conflict resolver?

MLG features shared decision-making power across governance levels, but scholars differ when it comes to which level is the most influential. Some imply that the engagement and influence of all actors in MLG is equal, no level of activity being superior to another (Stephenson, 2013), while a greater number argue that MLG strengthens, or at least retains the power of nation governments (Hale & Roger, 2014; Yang et al., 2021). Gillard et al. (2017) and Tompkins et al. (2002) provided evidence confirming how national government actors are able to steer climate governance networks in their own favor by withholding access to information and resources. However, in the past decade, we have witnessed that in the absence of national leadership (e.g., the U.S. withdrawal from the Kyoto Protocol and the Paris Agreement), sub-national authorities have had high hopes in taking climate action as they are directly accountable to their constituents for their decisions, and are more nimble than state and national elected officials in taking decisive action. Linking the subnational and the international might sound promising, but in practice it is a highly complex process. Whether and how national governments, as key actors with dominating leadership, should maintain the roles of conflict resolver, facilitator, decision maker, sponsor, and expert to facilitate effective governance, deserved further discussion (Emerson & Nabachi, 2015).

Described above are three types of conflict that might emerge in MLG and the hypothetical role of a national government in resolving conflict. To examine the applicability of the analytical framework and a real-world example of conflict resolution, the following sections focus on the implementation of the CDM using China as an empirical case. The MLG system of CDM governance in China will be introduced first and then be followed by a discussion on the three types of conflict generated between the international and subnational levels. The findings highlight the role of the Chinese national government in resolving conflicts during the governing process.

The clean development mechanism (CDM) in China

The CDM is a climate governance mechanism written under the United Nations Framework Convention on Climate Change (UNFCCC, 2006). Defined in Article 12 of the Kyoto Protocol, the CDM allows a country with an emissions reduction or emissions-limitation commitment under the Kyoto Protocol (Annex B Party) to implement an emissions reduction project in developing countries. Such projects can earn saleable certified emissions reduction (CER) credits, each equivalent to one ton of CO2, which can be counted toward meeting Kyoto targets. The design of this mechanism enables developing countries to gain assistance in the form of capacity, technology transfers, and financial flow to promote local SD through increasing job opportunities, facilitating renewable energy use, and raising climate change consciousness, among other things. In exchange, CERs are issued to developed countries to help them meet their responsibilities in emissions reduction (Shrivastava & Bhaduri, 2019).

Although the implementation of the CDM has been criticized for skewed participation and unjust distribution of benefits (Torvanger et al., 2013), the implementation of the CDM in China has been successful. As the largest host country for CDM projects, China registered more than 5000 CDM projects between 2005 and 2014, spread across 30 provinces. An MLG network supported the implementation of the CDM in China; it consisted of the UNFCCC Secretariat, the CDM Executive Board, and Designated Operational Entity (DOE) at the international level; the Designated National Authority (DNA) and The National Leadership Group on Climate Change, and National Clean Development Mechanism Committee at the national level; and project developers, provincial governments, consulting companies, and research institutions at subnational levels. Table 1 summaries the key actors and their duties in the MLG network.

Table 1 Key actors in MLG networks

Despite taking on the largest share of projects around the globe, China has encountered multiple conflicts in problem framing, benefit sharing, and capacity building across jurisdictional levels during CDM implementation. These require a governance system able to manage and resolve conflicts across multiple scales and among diverse policy actors. Because the central government retains key steering power in cross-level interactions, this study will focus on the role of central government in conflict resolution.

The author collected both quantitative and qualitative data to understand different types of conflicts in the MLG networks and how a national government may act as a conflict resolver to facilitate governance. Quantitative data regarding CDM projects in mainland China were extracted from a global CDM dataset. By identifying the location of each project through its Project Design Documents, the author obtained a dataset of 4429 cases for the period from 2004 to 2015 (after excluding those with missing or incomplete information). Qualitative data were extracted from policies, documents, project papers, meeting minutes, and records from the UNFCCC. The content analysis of qualitative data provided preliminary data that were used for substantiation and enrichment in the interviews. Data for other economic, socio-political, and demographic contextual variables were obtained from the China Statistical Yearbook, the China Environmental Yearbook, and other sources.


Conflicts in problem framing

The environmentally sound transfer of technology to developing countries has been identified as a normative principle for achieving international climate justice and as essential in mitigating greenhouse gas emissions. As stressed in Article 4 of the UNFCCC (2006), the extent to which developing country parties will effectively implement their commitments depends on the financial resources and transfer of technology from developed country parties (p. 8). Although the CDM does not have an explicit technology transfer mandate, UNFCCC and other international communities set high expectations for its potential in promoting transnational technology transfer (Seres et al., 2009). In the CDM’s project design documents (PDDs), the project developers were asked to ‘include a description of how environmentally safe and sound technology and know-how to be used is transferred to the host Party(ies)’ (UNFCCC, 2006: p. 16).

However, the technical attributes of this new mechanism of international cooperation have been undermined by project developers at the subnational level. As a market-based mechanism, the CDM concerned local project developers with its high economic benefits; but often, the high economic benefits of a CDM project are inconsistent or even conflicting with its potential for technology transfer. A CDM project with technology transfer implies higher cost, a long-term return cycle, and high investment risk, all making it unattractive in the carbon market. Normally, foreign partners have low motivation to transfer technology because they want to maintain technological competitiveness. However, even if they were willing to apply their advanced technologies, local project developers might not welcome them due to their limited understanding of the technology transfer opportunities. Therefore, gas reduction projects (e.g., HFCs or N2O) are most attractive to the project buyers because they generate a large volume of CERs in a relatively short period with low abatement costs (Shin, 2010), even if their potential for technology transfer can be low.

The conflicts in problem framing between supranational level and localities called for the national government to be the broker to align the perceptions and expectations among diverse policy actors (Jing & Li, 2019). Facilitating cross-level communication and provision of scientific information would be helpful for these subjective conflicts. Notably, for an MLG network including an international level, cross-level interactions can often be one-way; a national government functions as an ‘interpreter’ to deliver the concern from its upper levels to its lower levels. Indeed, technology transfer based on CDM projects was taken as an essential part of China’s national interest and was highlighted in the nation’s legal framework for the CDM. Article 4 of the Measures for the operation and management of clean development mechanism projects (the Measures) states ‘…international cooperation on CDM projects should promote the transfer of environmentally sound technologies… should put focus on projects on energy conservation and energy efficiency, development of renewable energy, and methane utilization’ (NDRC, 2011). To clarify and adjust expectations of local project developers, the Chinese national government has adopted two strategies.

The first strategy was to reinterpret international norms in its own interest (Gillard et al., 2017). Indeed, the CDM failed to provide a clear definition of technology transfer, which leaves flexibility for the national/subnational government with regards to interpretation and implementation. For example, technology transfer based on renewable energy projects normally includes off-the-shelf foreign equipment purchase and basic knowledge on operation and maintenance; however, it barely includes data, designs, inventions, materials, software, or technical knowledge movement. It thus remains debatable whether hydro projects can achieve high-quality and successful technology transfer, as these are mature technologies that probably are already available in many developing countries, leading to low rates of technology transfer. However, to meet the national policy from 2004 on promoting renewable energies, the Chinese national government defined the ‘development of renewable energy’ through CDM as an approach to technology transfer in Article 4 of the Measures.

Besides, the National Development and Reform Commission (NDRC, Energy Research Institute) held several CDM training programs to disseminate scientific information on technology transfer. As stated in a summary report submitted by the Xinjiang Development and Reform Commission in 2009, biomass energy projects had rarely received attention from local project developers before the training program. These training programs helped local project developers to understand the opportunities and benefits of technology transfer in biomass power generation and heat utilization technologies, which had successfully facilitated the development of related CDM projects in Xinjiang in 2009 (and both validated in 2011). This successfully aligned local project developers’ understanding of the CDM’s technical attributes with its international norms.

The second strategy was applying policy intervention to adjust the market’s preference for projects with a low rate of technology transfer. Article 36 of the Measures states that the Chinese government takes 65% of the CER transfer benefit from HFC and PFC projects, 30% from N2O projects, and 2% for energy efficiency projects and the others (NDRC, 2011). This has substantially reduced the market size for greenhouse gas reduction projects with few technology transfer opportunities, and increased market enthusiasm for renewable energy and energy efficiencies projects. Table 2 presents the number of CDM projects hosted in China and worldwide. Most of the CDM projects hosted in China were renewable energy projects (especially hydropower and wind power) and energy efficiency own generation; all are prioritized by the national policy. Besides, compared to the other countries, the dominant share in renewable energy (40.57%), energy efficiency improvement (24.99%), and methane utilization (72.13%) implies the effectiveness of the national policy intervention and its ambitions on technology transfer.

Table 2 CDM projects on renewable energy, energy efficiency, and GHG reduction

Conflicts in benefit sharing

The CDM has two objectives: first, to assist developing countries in achieving sustainable development (SD), and second, to facilitate developed countries in meeting their emissions reduction targets. However, scholars have argued that the two objectives might be inherently incompatible (Shrivastava & Bhaduri, 2019). A recent study based on the 4429 CDM projects hosted in China has reinforced that the emission reduction effects of CDM projects had a significantly negative relationship with local SD (Xu & Zhang, 2021). The finding also supports the statements that renewable energy CDM projects (e.g., wind turbine generators) might cause new ecological challenges for local climate-vulnerable communities by constructing green infrastructures. In the western/central regions hosting more renewable energy CDM projects, CERs had a significantly stronger and negative relationship with local SD.

Imbalance of power and participation between the international and subnational levels provide explanations for their conflicts in benefits sharing. On the one hand, the dominance of nation states in international climate negotiations gave high priority to binding emissions reduction targets. Sophisticated and science-based measurement, reporting, and verification systems were designed to assure that financial assistance has been effectively converted into tradable credit issuance (UN, 2020). On the other hand, the lack of monetary benefits of SD benefits has undermined the local interest. Even though the SD benefits are required for each CDM application, this content was often ambiguous. The PDD of a biomass heat and power project states that ‘92% of respondents believed that the proposed project would improve their livelihoods (p. 34)’; but how SD goals are to be achieved still remains unclear. Within a market-based mechanism dominated by buyers, developing countries are more vulnerable to the rapidly changing results of international climate negotiations. Once the international demand for CERs declines (e.g., during the climate negotiation deadlock at the 2009 Copenhagen Conference), sellers of carbon credits may race to the bottom to attract investment by sacrificing their SD benefits. The conflicts in benefit sharing thus threaten the internal legitimacy of local participants and hamper their willingness for future engagement and commitment.

To facilitate an effective MLG with better benefits sharing across levels, the Chinese government redistributed the national revenue from registered CDM projects by establishing the China CDM Fund in 2007. Believing developed countries in general did not have much interest in the sustainable development of host countries, China is the only major host country to collect revenue from CDM projects, with the hope of establishing a domestic CDM Fund to support sustainable development (Shin, 2010). As a public institution that integrates government and market functions, the mandate of the Fund is to promote sustainable social and economic development. Here is an example in Inner Mongolia. Given the favorable geographical conditions, Inner Mongolia has developed the most wind power CDM projects in China from 2004 to 2012. However, the negative environmental and climatic impact of wind farms has concerned scholars and governments since the early 2010s (Leung & Yang, 2012). In 2017, the CDM Fund granted a loan of 50 million RMB to a vegetation coverage project in Inner Mongolia to compensate for the negative impact on local sustainable development by planting drought-tolerant plants.

The China CDM Fund also provides grants to support activities in climate-related capacity building and promotion of public awareness and invests in industrial activities contributing to local sustainable development. For example, the Fund sponsored a CCTV documentary in 2011, “Warm and Cold, We Share Together—The Journey to Climate Culture”, to raise public awareness on the climate change issue. In 2018, the Fund sponsored the publication of the Environmental and Social Safeguard (ESS) Handbook to provide knowledge on the environmental and social impacts of the on-going projects and guidance for risk management (CDM Fund, 2018). Through exercising its regulatory power, the national government thus facilitated better benefit sharing from CDM projects in the MLG networks.

Conflicts in capacity building

Although the CDM was designed as a market-based mechanism, the calculation method of CER and the market benefits generated are approved by an international public institution, the CDM Executive Board (EB). The legitimacy of CDM is thus rooted in local project developers’ knowledge and confidence in the institutional arrangement, regulations, rules, and methodologies established by the CDM EB. Asymmetrical capacity between the international and subnational levels has undermined the MLG structures.

The concept of the ‘additionality’ of the emission reduction lies at the heart of the CDM rules, emphasizing that the emissions reductions of a CDM project would not have happened under a ‘business-as-usual’ approach (Freestone & Streck, 2009). To develop a tool to determine consolidated additionality, the CDM EB established a specialized Methodology Panel to develop recommendations on guidelines for methodologies for baseline and monitoring the ‘additionality’. Till the end of December 2020, the methodology panel has updated its Methodology Booklet 12 times with over 200 methodologies (CDM, 2021). In addition, the Designated National Authority and Designated Operational Entity (DOE) are also encouraged to propose new methodologies or revise the approved ones. Up to the end of 2020, methodologies for large-scale CDM project activities were updated over 259 times, and methodologies for small-scale CDM project activities were updated 813 times (CDM, 2021). Although the evolved and abundant methodologies assist CDM project developers in accurately determining the amount of CERs generated, it nevertheless requires the host countries to have enough professionals with relevant knowledge.

In a busy period, it normally takes EB 1–2 years to review a CDM project. Because the methodologies keep changing, the methodology applied in a project’s application might be out-dated when it is eventually reviewed. Local project developers often failed to follow the updated methodologies due to the lack of timely and accessible information, which led to CER buyers’ queries, defaults, and termination of contracts. This has raised another issue: local project developers also have limited capacity to deal with international contractual disputes. Due to the complicated and rigorous procedures of the CDM project application, projects developers normally search for private CDM consultants or research institutes to launch a CDM application. However, most of these local private CDM consultants or members of research institutes are environmental science and engineering professionals with limited knowledge on international law and politics, transnational investment, and risk management. In 2012, more than half of China’s CDM projects faced cancelation or renegotiation of their contracts amid a sluggish global carbon market. An interview with the CDM consultant pointed out one typical case is that disputes in carbon trading normally adopt the Anglo-American law system for arbitrations, but few CDM consultants in China could provide professional support to project developers who are dealing with trading disputes (China Daily, 2012). Delayed notification on the updated methodology led to the CDM board declining and withdrawing over 120 CDM projects from China. All these have gradually made project developers lose confidence in the transparency and predictability of decision-making at the international level.

Due to the lack of direct international–subnational interaction on capacity building (e.g., the provision of training courses), the NDRC and the Ministry of science and technology (MOST) of the Chinese government performed important liaison functions across levels of government. As the co-chairs of the Board of the National Leading Group to Address Climate Change, NDRC and MOST have a great deal of influence over designing and implementing China’s CDM. From 2005 to 2011, the two agencies held over 20 capacity building seminars and courses for local governments and CDM consultants. Contents of these courses include introduction of the latest emission reduction technologies and methodologies for CER validation, knowledge on international laws, negotiation skills, technology transfer, and project risk analysis and management (CDM China, 2007).

Since the first CDM was registered in 2005, there have been 27 provincial Science and Technology Departments that successively established affiliated CDM technology services or consulting centers to provide technical support for projects developers. These consulting centers, identified as ‘public service units’, are not governmental bodies but are directed by local governments by way of sponsorship and assignments (Guttman et al., 2018). Although in the early stage of development, CDM was expected to be a new form of public policy network that provides a platform for a wide array of partnerships (Streck, 2004). Non-state actors indeed play a relative weak role in the CDM governance system is China.

The Chinese national government has also leveraged through state-owned enterprises (SOEs). In March 2009, the China Joint Environmental Certification Center (an SOE established with the approval of the then Ministry of Ecology and Environment) was certified as a CDM DOE by the UNFCCC. It became the first DOE in China accredited by the CDM EB with certification and verification qualifications. Given the large task of validating project proposals or verifying whether implemented projects have achieved planned greenhouse gas emission reductions, a DOE based in China could provide direct technical support to project developers and communicate their needs from many localities through DOE conferences. Up to the end of 2020, China had six DOEs registered by the CDM Executive Board (among the 29 DOEs in total). The national government thus acted as a mediator of capacity building in two ways: on the one hand, to narrow the gap across levels through devoting expertise and resources; on the other hand, to develop actors at an international level to deliver voices from the subnational level.

Table 3 summarizes the three types of conflicts between the supranational and subnational levels in CDM implementation and proposes the role the national government may play in the resolution of conflict. The finding indicates the role that national government plays in reinterpreting international norms, aligning presumably conflicting goals, eliminating the difference in benefits sharing, and narrowing capacity gaps across levels. The mediating role of national government at the second level of governance seems to show that geographic proximity facilitates interactions, while more critical determinants were indeed the power differentials and dominance in the mobilization of resources that make the national level the most influential level in the network. The case of CDM implementation indicates a high level of national ownership of the climate policy process in China. Like the network administrative organization in network governance, the national government acts as the disseminator of resources and capabilities to improve overall governance performance (Provan & Kenis, 2008).

Table 3 Three types of conflicts and the roles of the Chinese national government

MLG has led to some degree of ‘unraveling of the state’ or a deconcentrated global governance (Moloney, 2021). In the past decade, the international community has acknowledged the role of and contributions by local and subnational governments. Deliberate steps have been taken to build them into a global framework of policy-setting and accountability, for example, involving local governments in various UN processes and bodies (Otto-Zimmermann, 2011). However, evidence from China shows the national government retains decision-making authority and orchestrates multi-level climate change governance networks. In contrast with Yamin and Depledge’s (2004) argument that the influence of international actors on climate policy can be more relevant in the Global South due to increasing international aid, this study argues the national governments in developing countries retain key steering power in integrating and managing information and resources flows. It would be difficult for international actors to hollow out national power and exert direct impact at the subnational level (e.g., through international aid financing local climate change action), and vice versa.

Admittedly, power differentials may hamper the integration of policy decision-making processes across levels of governance and the motivation of other actors, especially subnational governments. They are on the front line experiencing and having to address impacts from climate change. However, in the implementation of CDM, the subnational governments were ‘not allowed to reject any CDM applications for any reason’ before the application was submitted to the NDRC (NDRC, 2011: p. 5). Formal administrative institutions of national government hamper MLG, and jurisdictional boundaries create barriers to cross-level interaction, reinforcing mismatches between the governance system and the cross-level climate change policy problem. The dilution of power discourages subnational governments to be the advocates of CDM activities. They have low motivation and no authority to participate in governance and decision-making, and thus habitually hand over power to the upper level. To facilitate an effective and accountable MLG, the ability and motivation of subnational governments should be strengthened by the governance structure in which they are imbedded. Networks of support with other levels of governments, NGOs, and private sectors should be created to improve the capacity of subnational government in implementing voluntary actions.


The analytical framework of MLG has been widely applied to understand issues characterized by complex interactions and tight coupling. With a nested and mutually connected structure, MLG provides a context to understand the hierarchy relations among participants and conflicts across levels, which have been rarely discussed in the existing literature on collaborative governance. Drawing on the theories of conflict resolution and collaborative governance, this paper proposed a theoretical framework consisting of the three major conflicts MLG is likely to encounter: conflict in problem framing, conflict in benefit sharing, and conflict in capacity building. Focusing on the implementation of the Clean Development Mechanism in China, this study discovers conflicts generated in all three formats in an MLG network, especially between the supranational and subnational levels. The finding further indicates the role national governments play in resolving conflicts by reinterpreting international norms, aligning presumably conflicting goals, eliminating the difference in benefits sharing and narrowing capacity gaps across levels. Admitting national governments’ authority in decision-making and orchestrating multi-level climate change governance networks, this study concludes by calling for networks of support to strengthen the capacity and motivation of subnational governments in implementing voluntary actions, especially joint forces of practitioners and policy analysts in addressing large-scale environmental challenges like climate change (Young, 2021). This study contributes new knowledge on MLG by investigating the hierarchical relations among participants and how authority and power differentials can shape their interactions across levels.

This study has several limitations. First, it is descriptive and exploratory in nature. It only examines the types of conflicts generated in MLG, as well as the players and strategies that might help to resolve conflicts. The impact of the conflicts on governance performance and the role of the conflict resolver in moderating the impact are not the key concern, which leaves room for future study to explore. Second, this paper addresses only one unique governance context—that of China. Since what works (and is appropriate) in one context may not work in another context (Jing et al., 2021), future research should verify the generalizability of our findings in global settings with various institutional arrangements and dynamics.


  • Biermann, F., Bai, X., Bondre, N., Broadgate, W., Chen, C. T. A., Dube, O. P., et al. (2016). Down to earth: contextualizing the anthropocene. Global Environmental Change-Human and Policy Dimensions, 39, 341–350.

    Google Scholar 

  • Brickman, P. (1974). Social conflict: readings in rule structures and conflict relationships. D. C. Heath.

    Google Scholar 

  • Bulkeley, H. (2005). Reconfiguring environmental governance: towards a politics of scales and networks. Political Geography, 24(8), 875–902.

    Google Scholar 

  • Caplan, A. J., Cornes, R. C., & Silva, E. C. D. (2003). An “ideal kyoto protocol”: emissions trading, redistributive transfers and global participation. Oxford Economic Papers, 55(2), 216–234.

    Google Scholar 

  • CDM. (2021). CDM Methodologies. Accessed 14 Nov 2021.

  • CDM China. (2007). Meeting files of CDM experience exchanging conference on Jan 22–23. Accessed 23 July 2021.

  • CDM Fund. (2018). Environmental and social risk management handbook. Accessed 23 July 2021.

  • China Daily. (2012). Half of China’s CDM projects face default in global carbon market downturn. Accessed 26 Oct 2021.

  • Chrislip, D. D. (2002). Collaborative Leadership Fieldbook. John Wiley & Sons Inc.

    Google Scholar 

  • Crosby, B. C., & Bryson, J. M. (2010). Special issue on public integrative leadership: Multiple turns of the kaleidoscope. Leadership Quarterly, 21(2), 205–208.

    Google Scholar 

  • de Oliveira, J. A. P. (2009). The implementation of climate change related policies at the subnational level: an analysis of three countries. Habitat International, 33(3), 253–259.

    Google Scholar 

  • Di Gregorio, M., Fatorelli, L., Paavola, J., Locatelli, B., Pramova, E., Nurrochmat, D. R., et al. (2019). Multi-level governance and power in climate change policy networks. Global Environmental Change-Human and Policy Dimensions, 54, 64–77.

    Google Scholar 

  • Duit, A., & Galaz, V. (2008). Governance and complexity—emerging issues for governance theory. Governance, 21(3), 311–335.

    Google Scholar 

  • Emerson, K., & Nabatchi, T. (2015). Collaborative governance regime. Public management and change series. Georgetown University Press.

    Google Scholar 

  • Emerson, K., Nabatchi, T., & Balogh, S. (2012). An integrative framework for collaborative governance. Journal of Public Administration Research and Theory, 22(1), 1–29.

    Google Scholar 

  • Freestone, D., & Streck, C. (2009). Legal aspects of carbon trading: Kyoto, Copenhagen, and beyond. Oxford University Press.

    Google Scholar 

  • Gillard, R., Gouldson, A., Paavola, J., & Alstine, J. V. (2017). Can national policy blockages accelerate the development of polycentric governance? Evidence from climate change policy in the United Kingdom. Global Environmental Change-Human and Policy Dimensions, 45, 174–182.

    Google Scholar 

  • Gupta, J., de Moel, H., & van der Leeuw, K. (2007). Climate change: a glocal problem requiring glocal action. Environmental Sciences, 4(3), 139–148.

    Google Scholar 

  • Guttman, D., Young, O., Jing, Y., Bramble, B., Bu, M., Chen, C., et al. (2018). Environmental governance in China: interactions between the state and “nonstate actors.” Journal of Environmental Management, 220, 126–135.

    Google Scholar 

  • Hale, T., & Roger, C. (2014). Orchestration and transnational climate governance. Review of International Organizations, 9(1), 59–82.

    Google Scholar 

  • Herranz, J. (2008). The multisectoral trilemma of network management. Journal of Public Administration Research and Theory, 18, 1–31.

    Google Scholar 

  • Hoornbeek, J. A., & Peters, B. G. (2017). Understanding policy problems: a refinement of past work. Policy and Society, 36(3), 365–384.

    Google Scholar 

  • Huang, Y. (2019). Multi-level governance: explaining the “climate-focused” behavior of Chinese exporting companies. Public Policy and Administration, 34(2), 165–188.

    Google Scholar 

  • Huxham, C., & Vangen, S. (2005). Managing to collaborate. Routledge.

    Google Scholar 

  • Ishtiaque, A., Eakin, H., Vij, S., Chhetri, N., Rahman, F., & Huq, S. (2021). Multilevel governance in climate change adaptation in Bangladesh: structure, processes, and power dynamics. Regional Environmental Change, 21(3), 75.

    Google Scholar 

  • Jachson, D. & Coninck, S. D. (2019). How supporting climate action on a local level can transform the world. Accessed 16 May 2021.

  • Jing, Y., & Li, D. (2019). Private roles in enhancing multi-level governance: China’s “Internet + ” national strategy. Public Policy and Administration, 34(2), 144–164.

    Google Scholar 

  • Jing, Y., Torenvlied, R., van Gerven, M.M.-L., & Cao, J. (2021). Nonprofit contracting and partnership in elderly care: a comparison between china and the Netherlands. Global Public Policy and Governance, 1(2), 1–23.

    Google Scholar 

  • Jordan, A. J., Huitema, D., Hilden, M., Asselt, H. V., Rayner, T., Schoenefeld, J. J., et al. (2015). Emergence of polycentric climate governance and its future prospects. Nature Climate Change, 5(11), 977–982.

    Google Scholar 

  • Koppenjan, J. F. M., & Klijn, E.-H. (2004). Managing uncertainties in networks a network approach to problem solving and decision making. Routledge.

    Google Scholar 

  • Kousky, C., & Schneider, S. H. (2003). Global climate policy: will cities lead the way? Climate Policy, 3, 359–372.

    Google Scholar 

  • Lan, Z. (1997). A conflict resolution approach to public administration. Public Administration Review, 57(1), 27–35.

    Google Scholar 

  • Leung, D. Y. C., & Yang, Y. (2012). Wind energy development and its environmental impact: a review. Renewable & Sustainable Energy Reviews, 16(1), 1031–1039.

    Google Scholar 

  • Moloney, K. (2021). Deconcentrated global governance, transnational administration, and the public administration discipline. Global Public Policy and Governance, 1(2), 175–201.

    Google Scholar 

  • National Development and Reform Commission (NDRC). (2011). Measures for the operation and management of clean development mechanism projects (in Chinese). Accessed 16 Nov 2020.

  • Ongaro, E., Gong, T., & Jing, Y. (2019). Toward multi-level governance in China? Coping with complex public affairs across jurisdictions and organizations. Public Policy and Administration, 34(2), 105–120.

    Google Scholar 

  • Ostrom, E. (2011). Background on the institutional analysis and development framework. Policy Studies Journal, 39(1), 7–27.

    Google Scholar 

  • Otto-Zimmermann, K. (2011). Global environmental governance: the role of local governments. Accessed 16 Oct 2021.

  • Page, S. B., Stone, M. M., Bryson, J. M., & Crosby, B. C. (2018). Coping with value conflicts in interorganizational collaborations. Perspectives on Public Management and Governance, 1(4), 239–255.

    Google Scholar 

  • Piatak, J., Romzek, B., LeRoux, K., & Johnston, J. (2018). Managing goal conflict in public service delivery networks: does accountability move up and down, or side to side? Public Performance & Management Review, 41(1), 152–176.

    Google Scholar 

  • Provan, K. G., & Kenis, P. (2008). Modes of network governance: structure, management, and effectiveness. Journal of Public Administration Research and Theory, 18(2), 229–252.

    Google Scholar 

  • Quick, K. S., & Feldman, M. S. (2014). Boundaries as junctures: collaborative boundary work for building efficient resilience. Journal of Public Administration Research and Theory, 24(3), 673–695.

    Google Scholar 

  • Romzek, B. S., LeRoux, K., & Blackmar, J. M. (2012). A preliminary theory of informal accountability among network organizational actors. Public Administration Review, 72(3), 442–453.

    Google Scholar 

  • Seres, S., Haites, E., & Murphy, K. (2009). Analysis of technology transfer in CDM projects: an update. Energy Policy, 37(11), 4919–4926.

    Google Scholar 

  • Shin, S. (2010). The domestic side of the clean development mechanism: the case of China. Environmental Politics, 19(2), 237–254.

    Google Scholar 

  • Shrivastava, M. K., & Bhaduri, S. (2019). Market-based mechanism and ‘climate justice’: reframing the debate for a way forward. International Environmental Agreements: Politics, Law and Economics, 19(4–5), 497–513.

    Google Scholar 

  • Sørensen, E. (2006). Metagovernance the changing role of politicians in processes of democratic governance. The American Review of Public Administration, 36(1), 98–114.

    Google Scholar 

  • Stephenson, P. (2013). Twenty years of multi-level governance: Where Does It Come From? What Is It? Where Is It Going? Journal of European Public Policy, 20(6), 817–837.

    Google Scholar 

  • Streck, C. (2004). New partnerships in global environmental policy: the clean development mechanism. Journal of Environment and Development, 13(3), 295–322.

    Google Scholar 

  • Ter-Minassian, T. (2020). Local public finance and capacity building in Asia: issues and challenges. Accessed 2 July 2021.

  • Thomson, A. M., & Perry, J. (2006). Collaboration processes: inside the black box. Public Administration Review, 66, 20–32.

    Google Scholar 

  • Tompkins, E., Adger, W. N., & Brown, K. (2002). Institutional networks for inclusive coastal management in Trinidad and Tobago. Environment and Planning A, 34(6), 1095–1111.

    Google Scholar 

  • Torvanger, A., Shrivastava, M. K., Pandey, N., & Tørnblad, S. H. (2013). A two-track CDM: improved incentives for sustainable development and offset production. Climate Policy, 13(4), 471–489.

    Google Scholar 

  • Underdal, A. (2010). Complexity and challenges of long-term environmental governance. Global Environmental Change-Human and Policy Dimensions, 20(3), 386–393.

    Google Scholar 

  • UNFCCC. (2006). Guidelines for completing the project design document (CDM-PDD), and the proposed new baseline and monitoring methodologies (CDM-NM), version 05, CDM-Executive Board, Bonn, Accessed 23 Apr 2021.

  • United Nations (UN). (2020). CDM methodology booklet (12th edition). Accessed 23 July 2021.

  • Van Buuren, A., & Gerrits, L. (2008). Decisions as dynamic equilibriums in erratic policy processes. Public Management Review, 10(3), 381–399.

    Google Scholar 

  • van Popering-Verkerk, J., Teisman, G., & van Buuren, A. (2015). Synchronising climate adaptation processes in a multilevel governance setting: exploring synchronisation of governance levels in the Dutch Delta. Policy and Politics, 43(4), 579–596.

    Google Scholar 

  • Williams, D. S., Rosendo, S., Sadasing, O., & Celliers, L. (2020). Identifying local governance capacity needs for enabling climate change adaptation in Mauritius. Climate Policy, 20(5), 548–562.

    Google Scholar 

  • Xu, J. & Zhang, Y. (2021). Has the international climate regime promoted climate justice? Evidence from clean development mechanism projects. Climate Policy.

  • Yamin, F., & Depledge, J. (2004). The international climate change regime: a guide to rules, institutions and procedures. Cambridge University Press.

    Google Scholar 

  • Yang, L., Du, Z., Cheng, C., & Shen, P. (2021). Building a compound and collaborative governance framework to improve international space sustainability. Global Public Policy and Governance, 1(2), 202–224.

    Google Scholar 

  • Young, O. R. (2006). Vertical Interplay among scale-dependent environmental and resource regimes. Ecology and Society, 11(1), 27.

    Google Scholar 

  • Young, O. R. (2021). Can practitioners and analysts join forces to address largescale environmental challenges? Global Public Policy and Governance, 1, 61–77.

    Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Jingyuan Xu.

Ethics declarations

Conflict of interest

The author states that there is no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Xu, J. Conflicts in multi-level governance: an analysis of international climate policy implementation at the sub-national level. GPPG 1, 401–420 (2021).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Multi-level governance
  • Conflict
  • Climate change governance
  • Policy implementation
  • China