The role of aid agencies within environmental cooperation in Congo Basin: facilitators or policy entrepreneurs?

Environmental cooperation in the Congo Basin region is facing a paradox as the region has an inflow of aid but deforestation and poverty continue to grow. We examine the role of aid agencies in this paradox, who we assume are policy entrepreneurs who influence and benefit from the process. To test these assumptions, we use policy entrepreneurship theory coupled with a comparative qualitative approach to conduct two case studies. The first case study is a climate change adaptation capacity building initiative with the German aid agency GIZ in the central role. We prove that GIZ led the project with high effectiveness, benefit from it but failed to align the initiative’s goals with the local needs. The second case study is the Congo Basin Forest Partnership (CBFP), which functions without an aid agency in the central role. We observe that although the CBFP’s actions strongly matched local needs, it lacked some effectiveness and could not yield relevant policy outcomes. Therefore, we suggest that suboptimal institutions meeting a minimum standard in both management and orientation toward local needs should be built. ARTICLE HISTORY Received 5 February 2021 Accepted 24 May 2021


Introduction: the 'third party' of environmental cooperation in the Congo Basin
Since the 1960s, the promotion of development aid as the cornerstone of economic growth and wellbeing in developing countries has seen the emergence of many aid agencies, known as the 'active hand' of donor countries (Dreher, Nunnenkamp, and Thiele 2011;Easterly 2008). According to Martens (2005), these aid agencies reduce the ex-ante transaction costs (i.e. the uncertainties related to funds or goods transfer), mediate between the diverging preferences of donors and recipients' countries, and package aid flows in contracts that reduce ex-post transaction costs. Within the development paradigm, the role of these aid agencies started to be questioned at some point, particularly in regard to their real contributions within this international development cooperation framework. Martens (2005) recognized that the outputs of aid agencies' facilitation work were not always undisputable. Gulrajani (2011) explored this issue further by questioning aid agencies' contributions to development aid effectiveness.
Despite the existing debate on the relevance of aid agencies within the development aid paradigm, when sustainable development and environmental assistance for the most vulnerable countries became a global priority after the Rio Earth summit in the 1990s (Sands 1992), the aid agencies that had been serving as development cooperation facilitators were suddenly regarded as a 'third party' in the international environmental cooperation paradigm. According to Macneil and Bray (2013), a third party is usually a trusted neutral actor of the cooperation system, and, based on its individual relationships with the different parties involved, facilitates the exchanges between the cooperating parties to help them achieve their different objectives.
As was the case in other global south regions, the former development cooperation aid agencies that were active in Congo Basin were relegated to 'third party' status in the international environmental cooperation paradigm. Their main objective is to significantly help developing countries by contributing to the sustainable management of the Congo Basin forest according to the needs of the countries in the Congo region. For the last 30 years, aid agencies have been helping the bilateral and multilateral processes of donors and recipient countries to achieve sustainable management of Congo Basin forests while supporting the national and regional development prospects.
However, a synopsis of the Congo Basin forest management presents a different picture (De Wasseige et al. 2012). There seems to be growing deforestation and poverty in Congo Basin forests and regions, despite flowing aid. We refer to this sober development as the 'the Congo Basin paradox 1 ' (Mvondo 2009;Nago and Krott 2020;Nago and Ongolo 2021;Trefon 2010). This observation has been extensively investigated. By reviewing the instruments and strategies used to promote cooperation between the donor and recipient countries, many analyses have been delivered to help tackle the paradox (Atyi et al. 2008;Tchatchou et al. 2015).
Nevertheless, we believe that previous literature has paid limited attention to aid agencies. In this study, we will consider as aid agencies, foreign aid agencies defined by Martens (2005) as agencies in charge of facilitating the cooperation between donor countries and recipient countries (mostly from the developing world). These foreign aid agencies differ from domestic aid. Indeed: … foreign aid agencies can be defined in contrast to domestic income redistribution agencies, such as government agencies dealing with medical and unemployment benefits. While domestic aid agencies redistribute income between donors and recipients living in the same political constituency, foreign aid agencies target recipients living outside the donor's constituency, usually in developing countries. Split constituencies have major implications for the decision-making process. In domestic aid, both donors and recipients have voting rights and can influence the political decision-making process. In contrast, in foreign aid the feedback loop between recipients and decision-makers is broken; only donors have political leverage over the decision-making process. This broken feedback loop explains the origins of the ownership problem that is so frequently mentioned in connection with foreign aid, but rarely if ever in connection with domestic income redistribution (Martens 2005, 644).
Aid agencies are presented in Congo Basin as neutral facilitators of the cooperation process and defenders of the cooperation parties' interests and thus as independent and rational actors of the environmental cooperation paradigm. Indeed, they have rarely been questioned on their role in the paradox of a flowing foreign aid and growing poverty in the Congo Basin. However, are they truly neutral? In this paper, we consider aid agencies as rational actors of environmental cooperation in the Congo Basin region to investigate their role in the system and whether they share responsibility for the Congo Basin paradox.
The paper proceeds as follows. Section II includes a presentation of the theoretical framework, Section III describes the methodological account, Section IV continues with the results, Section V provides a discussion of the results, and Section VI presents the conclusion of the solutions to the paradox.
Theoretical framework: policy entrepreneurship According to Goffee and Scase (1983), an entrepreneur is a risk taker and innovator who rejects the security of employment in a large organization to independently create wealth and accumulate capital. In this regard, theorizing about entrepreneurs has always been seen as an appanage of economists as they use their skills and innovation to boost the economy (Blanchflower and Oswald 1998). Kingdon (1984) linked the word entrepreneurship with policy by highlighting a group of very powerful actors, called 'policy entrepreneurs' within its Multiple Streams Framework (MSF). 2 These policy entrepreneurs are very important actors in policymaking and agenda-setting processes as they develop policy alternatives, associate them with problems, and present solutions to decision makers at the right time. The paper also presents these policy entrepreneurs as advocates who use innovative and non-traditional 3 ways to invest their resources to promote their investment with the ultimate aim to make profit.
Developing the theory, Mintrom and Norman (2009) defined policy entrepreneurship as a policy change theory that aims to frame the influence of public policy by individuals or institutions who take advantage of opportunities offered by their position to serve or increase their own interests. Strongly inspired by Kingdon (1984), they defined the elements that identify individuals or institutions as policy entrepreneurs.
First is social acuity display: A policy entrepreneur is an individual or institution who takes advantage of the 'window of opportunities' (Kingdon 1984) to promote policy change. In policymaking areas, once they have acknowledged the opportunity, policy entrepreneurs must display it to recognize it in order to achieve personal goals. They have two main strategies. The first is displaying the social acuity to its networks. The more the entrepreneur's networks acknowledge the acuity, the more likely the promotion of a policy change is successful (Balla 2001). The second strategy is understanding the ideas, motives, and concerns of others in their local policy context, and responding effectively.
The next element is problem definition. According to Mintrom and Norman (2009), how problems are defined determines which individuals pay attention to them. Problem definition therefore affects how people relate them to their own interests. In this regard, problem definition is very important for a policy entrepreneur. To meet their objectives, a policy entrepreneur usually presents a problem in a way that suggests that a crisis is at hand, that the current method of handling the crisis will fail, or that they are well-connected with those who can immediately solve the crisis (Henig 2008;Levin and Sanger 1994;Stone 1997).
Next is building teams: As in business policy, entrepreneurs know that their strength depends on their ability to work with others. They therefore develop several team-building strategies. The first is creating teams with outstanding individuals whose different knowledge and skills help find solutions to all their defined problems or pursue change (Meier 1995). The other strategy is developing and nourishing connections with different networks to help them either legitimize a problem or display their capacity to solve it.
The next element is leading by example. According to Mintrom and Norman (2009), the risk aversion prevalent amongst policymakers is one of the major challenges of policy change. Therefore, a policy entrepreneur must priorities reducing this perception of risk aversion amongst policymakers. Doing so allows the policy entrepreneur to convert ideas into action with the objective of demonstrating their genuine commitment to improved social outcomes, thereby increasing the credibility of policymakers (Quinn 2000).
The final element is interest fulfillment. Like their counterparts in business, policy entrepreneurs invest their resources to promote a position in return for anticipated future gains or appease a future concern in the form of material, purposive, or solidary benefits.
To properly answer the central question of this work on the responsibility of aid agencies within the environmental cooperation challenge of the Congo Basin, we investigate relevant policy implications. Relevant policy needs specific project objectives. By synthesizing these elements with the theory of policy entrepreneurship, we formulate the basic hypothesis that the independent variables: (i) match with local needs and (ii) produce a project output that effectively supports relevant policies. The match depends on the specific project objectives, display of social acuity, and problem creation. The effectiveness of a policy entrepreneur can be created by partnership building, leading by example, and interest fulfillment. Finally, the project outputs contribute to a relevant policy if a solution that serves the needs of local partners is achieved. Based on the above variables, we propose the following hypotheses: H 1: High effectiveness of a policy entrepreneur combined with a low match with local needs lead to irrelevant policy outcomes.
H 2 : Low effectiveness of a policy entrepreneur combined with a strong match with local needs lead to irrelevant policy outcomes.
H 3 : Relevant policy outcomes will be achieved only if there is at least a critical level of effectiveness of a policy entrepreneur and a critical level of match with local needs.
Relevant or irrelevant policies refer to regional policies (at Congo basin level) that match with the needs formally expressed in cooperation texts such as the COMIFAC 4 convergence plan 5 (local needs), and critical level refers to a level at which the effectiveness of a policy entrepreneur matches with local needs viably, but not necessarily perfectly.
Methodological account: the narrative of a comparative approach As described by Collier (1993, 1), 'comparison is a fundamental tool of analysis. It sharpens our powers of description, and plays a central role in concept-formation by bringing into focus suggestive similarities and contrasts among cases.' We distinguish three main comparison alternatives within political science: withincase analysis, quantitative techniques employing a relatively small number of cases, and systematic comparison of a small number of cases for causal analysis (Lijphart 1971). For this analysis, we use the third one.
This allows us to compare two case studies to test our hypotheses. Although we could have added other cases, we prefer to limit our work to these two because we wanted to focus on cases that we had sufficient access to during our fieldwork as to be able to answer our research questions in depth. Moreover, the two cases we chose represent the most-used cooperation facilitation models in the region and thus they allow us to comprehensively cover cooperation facilitation in Congo Basin. Environmental cooperation aid in Congo Basin, especially when it comes to the technical aspect, rarely follows the model of direct cooperation between the Congo Basin countries' administrations and the donor countries, and almost always pass either through renowned or accredited aid agencies or through facilitation bodies created for the purpose (Nago and Ongolo 2021).
The first case study will be an environmental cooperation initiative in the Congo Basin with the mediation of an aid agency, and the second case study will be an initiative with a facilitation body. The two cases will be examined for the elements designed by the theoretical framework in Section II. First, we examine whether aid agencies can be identified as policy entrepreneurs who shape and benefit from environmental cooperation in the Congo Basin. Second, we examine the consequences of their influence on environmental cooperation outputs.
Our first case study is a climate change project named Scenarios of Climate Change in the Congo Basin (SCB), which was implemented between 2011 and 2013 by the Climate Service Centre of Hamburg in Germany and the University of Wageningen in the Netherlands. It was funded by the German Ministry of Environment, Nature Conservation and Nuclear Safety for 1.5 million EUR. It aimed to strengthen the capacities of Congo Basin leaders in adaptation and mitigation (Hänsler, Saeed, and Jacob 2013). However, the project outputs ultimately had little utility for the Congo Basin countries as their expectancies and priorities, clearly expressed in their official cooperation manual (COMIFAC Convergence plan), were not considered during the elaboration stage of the project (Nago and Krott 2020). Moreover, power imbalances in favor of the northern participants of the project weakened the southern participants' abilities to modify the project objectives for a better match with their priorities. GIZ 6 (Gesellschaft International Zusammenarbeit), which is a German non-governmental aid agency with a long-time experience in international cooperation facilitation, facilitated the cooperation between the Netherlands, Germany, and the Congo Basin countries within this project.
The second case study is the Congo Basin Forest Partnership 7 (CBFP). It is a facilitation body created in 2006 with the objective to facilitate and enable a fruitful environmental cooperation in the Congo Basin that leads to a win-win partnership between donor and recipient countries for the sustainable management of the Congo Basin forests. Almost all the actor groups of Congo Basin forests management are represented within this facilitation body through different colleges (Congo Basin countries, regional governmental organizations, civil society, research institutions, and donor countries). The initiative's lead is successively assumed by donors. This study covers the CBFP facilitations between 2014 and 2018.
The data collection was qualitative and was conducted over 6 years from 2012 to 2018 at different arenas where the SCB project (Case Study 1) and the CBFP initiative (Case Study 2) activities were held. We collected data at the results presentation workshop of the SCB project in November 2012 in Douala, at the United Nations Climate Change Conference of parties in November 2017 in Bonn, and at the CBFP meeting of parties in November 2018 in Brussels (to see the different institutions' representatives that we interviewed, please consult our Appendix). Our empirical work included 20 interviews (10 experts from Case Study 1 and 2, respectively), two group discussions of mixed participants (one from Case Study 1 at the results presentation workshop of the SCB project, and another from Case Study 2 at the CBFP meeting of parties), the analysis of many published and confidential key documents, and participant observations at many formal and informal meetings. Since none of our interviewees granted permission to be recorded, we did not transcribe any data. Most interviews were conducted as either semi-structured dialogues or informal discussions. The interviews lasted about 1 h each, depending on the interviewee's availability and interest in participating. These interviews were mostly fragmented in at least two nonlinear work sequences (initial contact, core discussion, and complementary information voluntarily provided by the interviewee or requested by the researcher).
The data analysis included codification of qualitative data in Excel-Stat to avoid the omission of pertinent information collected in the field. The last step in our data analysis was the interpretation of the codified data based on the elements of policy entrepreneurship theory presented in Section II. The results of our analysis are presented below.

Results: entrepreneurship versus facilitation
The first package of results is from the scanning of Case Study 1 through the elements of the theoretical framework described in Section II. The objective of this scanning is to answer our central question on whether aid agencies in the Congo Basin are a 'third party' of the environmental cooperation process or policy entrepreneurs that influence the process and benefit from it. Table 1 shows the activities of the GIZ within the cooperation initiative in the Congo Basin (the SCB project). The activities are evaluated along the seven elements and have the following results.
Aid agency identification: The aid agency involved in the project scenarios of climate change in the Congo Basin was GIZ, a German cooperation agency that has been active for decades in the Congo Basin as a partner to Germany, and many other developed countries' institutions, in their dispatching of aid. GIZ was first involved in development cooperation 'facilitation.' As has been the case for many other aid agencies in the developing world, it added environmental expertise to its portfolio when the need arose (Sands 1992). In addition to simple observations in the field, we consulted formal unpublished documents of the project that clearly identified GIZ as the aid agency in charge of the cooperation between the northern and southern parties of the SCB project.
Project's objectives: The objectives were stated in the official documentation (Hänsler, Saeed, and Jacob 2013;Ludwig et al. 2012) and presented during the project's results presentation workshop in November 2012 in Douala. These project's objectives could mainly be summarized in terms of strengthening the adaptation capacity of the countries in the Congo Basin toward climate change through climate change modeling. These objectives have a potential for being a good match with some of the needs of the local partners.
Social acuity display: When it came to test whether GIZ displays social acuity as a policy entrepreneur, we first reviewed some of its reports on the socio-political context in Central Africa. We found that, on the official website 8 , GIZ often uses terms like, 'poor governance, corruption, unfavorable conditions for investment' to describe the governments of Congo Basin countries. Moreover, when we interviewed the northern researchers of the SCB project about why they approached GIZ, given that it is classed as a 'third party,' they said that they approached GIZ to know more about the Central African context as they had no experience with that region. The feedback they received from GIZ was that it is relatively difficult to directly cooperate with the Congo Basin administrations. Thus, it was obvious for them to let GIZ 'lead' their cooperation within the project (see Appendix, interview numbers 1-3). This social acuity focusses on corrupted local administration and hampers the connection with climate change adaptation capacity building, hence weakening the match with local needs.
Problem definition: Within the SCB project, the problem definition of the GIZ was easy to highlight as it was the logical implication from their social acuity display strategy. On the one hand, GIZ presents itself as being active in Central Africa since the 1960s, having collaborated with many German Ministries with positive results in the region; on the other hand, GIZ presents Central African countries as being corrupt or marred by poor governance. 9 The direct conclusion is that cooperating directly with Congo Basin countries on environmental initiatives is difficult and that it would be better to involve GIZ as they know the context. Thus, GIZ became the cooperation 'leader' within the SCB project. However, stressing the high risk of cooperating with Congo Basin countries weakens the potential of a good match with the needs of these countries.
Partnership building: Throughout the years, GIZ has built different partnerships within the Congo Basin, and can be mobilized at any time to implement different projects. These partnerships aim at both reducing the doubt of effectiveness in North-South cooperation, and guaranteeing GIZ a central position within this cooperation. In the specific case of the SCB project, GIZ has built partnerships with the Central African Network of Forestry and Environmental Training Institutions 10 and the Central African Forest Commission (COMIFAC), which demonstrated to the Climate Service Centre of Hamburg (the project leader) that they were capable of involving the most relevant actors of the Congo Basin in the SCB project. They could also guarantee their control over Congo Basin parties through good cooperation within the project. Leading by example: Once GIZ received the role of enabling the cooperation within the SCB project, they passed everything through the project's workshops, from participants' invitations to the results, to ensure the project would be a success. GIZ led by example, as their experts were on the ground to solve any issues. The experts also used their regional networks to involve the 'right' participants in the project as well as give the project commanders the audience they wanted. In this regard, some of our northern interviewees explicitly acknowledged the relevance of GIZ within the SCB project. For instance, 'GIZ made this project happen' (see Appendix, interview number 1) and 'GIZ made us feel very comfortable during our stay in Cameroon. We even left our hotels to stay at the private houses of some of their staff because, without them, we would have been lost when cooperating with Congo Basin countries. We would definitely not have achieved anything without their facilitation' (see Appendix 1, interview number 2).
Interest fulfillment: This can help identify an individual or institution as a policy entrepreneur (Mintrom and Norman 2009). According to Mintrom and Norman (2009), an entrepreneur invests their time, energy, networks, and resources on an initiative to fulfill some interests. These interests can be materialistic or not. In the case of GIZ, our informal consultation of the unpublished documents of the project revealed that GIZ had an estimated 15% of the budget for their 'facilitation work' marked as an overhead. This overhead is, in other words, a fee that the GIZ gets paid as part of the cooperation money to deliver its service. These fees are said to be used to pay the staff involved in the project, the offices, cars, and other costs related to their work. However, many interviewees from Congo Basin viewed these fees as too high, and they would have preferred a direct cooperation without the facilitation of the GIZ; the following quotations highlight this point: The so-called overhead fees are still cooperation money that could have been used to make concrete actions on the field. Expensive cars and offices are not contributing to climate change adaptation in the Congo Basin (Appendix 1, Interviewee 5) International experts hired by GIZ are incredibly expensive. They are paid five time more than the local experts to do a job that local experts could do better and with less costs. I believe that this is a waste of cooperation money, and I would have wished to cooperate directly with German institutions without the facilitation of GIZ. (Appendix 1, interview 6).
Besides this financial benefit, our personal observation as a project participant also led us to the conclusion that GIZ benefited from being renowned for delivering the SCB project. Although they built this reputation over years, the project contributed somewhat to solidifying their recognition as the key partner of environmental cooperation in the Congo Basin. The following quotation from one GIZ interviewee highlights the organization's target of good reputation and positioning through its actions in the Congo Basin: It is very important for us to have good relations with the scientists from Germany and Netherlands, as in case they are satisfied they will ask for our services again for the next round of the projects or at least for other projects. This is how we keep on running. (Appendix 1, Interviewee 5) Project's output: The project outputs are our dependent variable (ii). The basic claim of the cooperation project is that the output should support climate change adaptation capacities of Congo Basin countries. It is important to understand whether GIZ, the project cooperation 'leader' helps the parties reach their objectives. The first observation we made is that GIZ helps the northern parties to lead the project from the beginning to the end, and is, in itself, an achievement with regards to the latent conflict among the participants. Furthermore, we observed that GIZ could not help the donor to meet its objective of strengthening the climate change adaptation capacities of Congo Basin countries, as the southern project participants confessed to not having been able to acquire the knowledge transferred within the project (see Appendix 1, interview numbers 6-10). Among the reasons of this SCB project 'failure,' the southern parties interviewed mentioned the poor relevance of the transferred modeling results for the interests of the Congo Basin countries. In addition, the GIZ's role was seen by the southern partners as 'profiting from the cooperation money and preventing them from having a direct and sincere cooperation with the donor' (see Appendix 1, interview number 8).
Case Study 1 helped us to confirm our first hypothesis. GIZ is a highly effective entrepreneur, but, due to a weak match with the needs of Congo Basin countries, GIZ project outputs do not support relevant policy. In contrast to the strong activities of GIZ, the project failed to meet the aim of increasing the adaption capacity of Congo Basin countries; empirical evidence shows that a highly effective policy entrepreneurship does not guarantee the success of the project if the leader has weak coherence with local needs.
Case Study 2 follows the activities of the CBFP from 2014 to 2018. Table 2 shows the results for the different factors for a successful policy entrepreneur.
Aid agency: Generally, the CBFP functioned without the cooperation leadership of an aid agency, particularly from 2014 to 2018. Its technical and executive secretaries changed over the years according to the head of the facilitation. 11 Projects objective: The main objective of CBFP is 'to support the shared vision of the Central African Heads of State, notably by improving efficiency of measuresincluding technical and financial assistance, to promote biodiversity conservation and sustainable management of forest ecosystems, combat climate change, and reduce poverty in Central African countries'. 12 The improvement of the technical and financial assistance for biodiversity conservation and sustainable management of the Congo Basin forests should be achieved by coordination between the different actors' projects and programs. These objectives match well with the needs of the Congo Basin countries.
Social acuity display: The CBFP also displayed some social acuity to justify the relevance of a partnership. According to their website and our interviews (see Appendix 1, interview numbers 11-15), CBFP stresses the need for a better coordination for sustainable forest management, biodiversity conservation, and fighting against poverty in the Congo Basin. This matches the needs of Congo Basin countries.
Problem definition: As a direct consequence of the social acuity display, CBFP highlights the failed harmonization of international assistance as a key problem for sustainable management of the Congo Basin forests, and is positioning itself as the best solution for strengthening cooperation. In addition, the problem definition directly meets the needs of the Congo Basin countries as it covers the issue of deforestation, which causes a direct threat to Congo Basin countries.
Partnership building: The CBFP secretariat is not associated with any network to help guarantee it as a leader within environmental cooperation in the Congo Basin (see Appendix 1, interview number 20). The secretariat invites the actors involved in environmental cooperation in the Congo Basin to meet and work together to reach 'their common goal' of safeguarding the forests while fostering regional development. Due to the weak leadership and dependency from national and international actors, the meetings did not result in any relevant decisions being made or programs being designed. This result clearly shows poor effectiveness of the CBFP. Leading by example: The CBFP and its technical secretariat did not lead by example. They invited actors to collaborate and lead together for sustainable management but did not actively demonstrate their competence by example. The very limited number of permanent staff (approximately five) curbs the potential to make a strong difference in leading sustainable management of the Congo Basin forests. Instead, the CBFP provides a loose platform for different partners (scientists, donors, civil society, regional, multilateral, private sector, international NGOs) to pursue their individual agenda. In addition, the meetings and discussions occur without constraint, but are not productive. The CBFP 'depends on their college member action and good will' (see Appendix 1, interview number 19).
Interest fulfillment: CBFP does not get paid an overhead fee. It receives some facilitation acknowledgement in case it directly or indirectly helped institutions achieve their goals of sustainable management in the Congo Basin forests (see Appendix 1, interview number 17).
Project output: The CBFP helped connect the different actors of the environmental cooperation landscape in the Congo Basin. This platform helped them exchange and see how to collaborate for a sustainable management of the Congo Basin. However, the CBFP project did not lead to any significant change and improvement in the field.
Case Study 2 shows that CBFP does not have the capacity to act as an effective policy entrepreneur. Its contribution toward a positive output for the initiative objectives was negligible, despite the CBFP formulating social acuity and matching the problem with local needs. This empirical evidence is well in line with our second hypothesis: low effectiveness of a policy entrepreneur combined with a strong match with local needs lead to irrelevant policy outcomes.

Discussion: the environmental cooperation dilemma of the Congo Basin
The above results illustrate that environmental cooperation in the Congo Basin is in a dilemma.
Case Study 1 showed environmental cooperation with the leadership of aid agencies that function as entrepreneurs with rigor and effectiveness, but do not match local needs (H 1 ). The cooperation fails to positively contribute to the project's objectives. The following are the challenges of using an aid agency in international environmental cooperation.
Among the points raised about the central role of aid agencies, the first is that of the overhead received from the cooperation fund. Indeed, GIZ received an estimated 15% of the SCB project for their administrative fees and their experts' salaries. This is considered disturbing by Congo Basin representatives, who believe that the aid agency should be more transparent about the financial benefits derived from the cooperation and also be more modest given the context. This echoes Easterly and Pfutze (2008, 46), who observed that it is very difficult for a researcher or external expert to determine the exact overhead cost of aid agencies in a project 'because there does not seem to be completely standard definitions of concepts like number of aid agency employees and administrative costs. Also, some of these agencies have other purposes than granting aid, and the employees and costs of granting aid are not clearly separated out.' This lack of transparency created tensions within the SCB project, and had a negative impact on the project outputs. The Congo Basin representatives of the SCB project said that they felt as though they were 'played' within the project (see Appendix 1, interview numbers 6-10). This is a clear indicator of a low match with the needs of Congo Basin countries.
Another issue with the intermediation of an aid agency within our case study was the GIZ experts' familiarity with the project. Many researchers from the Congo Basin (see Appendix 1, interview numbers 6-10) mentioned that these GIZ experts did not have any knowledge of the climate modeling transferred within the project, and believed that this could hinder achieving the project objectives as they could not properly advise the northern and southern parties involved. Ika (2012) called this habit the one-size-fits-all trap: the aid agency has one functioning model, or one group of experts, for all types of projects. Other authors also framed this habit of having few experts for an unlimited amount of sectors as problematic for cooperation, and attempted to explain why this practice has continued since the 1960s (Easterly and Pfutze 2008;Easterly and Williamson 2011;Williamson 2010;Wright and Winters 2010). Easterly and Pfutze (2008, 38) argued that 'aid agencies split their assistance between too many donors, too many countries, and too many sectors for each donor, where 'too many' reflects the view that having multiple donors and multiple projects forfeits the gains of specialization and leads to higher-thannecessary overhead costs for both donors and recipients.' The lack of specific expertise in the highly difficult application of climate modeling into practice was neglected in the problem definition, and stressed the need for effective management instead of risky cooperation with corrupt administration.
The third point highlighted as being challenging for international environmental cooperation using an aid agency is the lack of accountability of the aid agency to the recipient countries. Within the SCB project, the southern party kept complaining about the 'bossy' attitude of GIZ. Be it through the group discussion, interviews, or participant observation, the frustration of the southern parties in not being consulted for the project elaboration or implementation was dominant. As GIZ played the cooperation facilitator role, this frustration could be attributed to its failure. We never observed GIZ taking any action to ease these tensions. Instead, the southern participants who complained considerably during an early workshop were not invited to the next workshop (observation made in Yaounde at a small meeting after the 2012 Douala workshop). This observation is well in line with Easterly and Pfutze (2008), who stated a complete unaccountability of aid agencies to their intended beneficiaries. The results again show a low match with the needs of the beneficiaries.
The above explains how a very low match with the local needs, despite high effectiveness, results in the failure to create a relevant policy.
Case Study 2 showed the failure of an initiative without the central role of an aid agency. It had a strong match with local needs, but, due to low effectiveness, generated a policy output that was not relevant.
The first challenge we could highlight is the lack of leadership. Aid agencies usually use their networks, experts, and expertise to ensure the donor countries can control the recipients and encourage them to achieve the cooperation objectives. Through leadership, they reduce the uncertainty and lack of trust that exist between the donor and recipient countries (Ika 2012;Martens 2005). The CBFP or its secretariat did no such thing. They simply enabled dialogue between donors and recipients and did not guarantee the success of the interactions.
Another challenge is the dependency on the CBFP members' personal activities agenda. The CBFP does not lead any project and depends on the individual working agenda of its members and hopes, through its facilitating harmonization, that its objectives of biodiversity, safeguarding against deforestation, and poverty reduction in the Congo Basin are reached. We agree with Rogerson (2005) that harmonization of aid objectives between the donor and recipients do not always guarantee concrete development benefits in the field.
In particular, the CBFP does not lead concrete actions in the field as a policy entrepreneur does. It has neither the staff nor the expertise to conduct any initiative on its own to reach its overall objective. This is a major challenge to its role of an effective entrepreneur.
Effectiveness was diminished by the non-binding nature of collaboration within CBFP. A major criterion of membership and collaboration within CBFP is that it is non-binding. 13 On the one hand, this is a strength as it guarantees the self-commitment of any partner to the goal of sustainable management of the Congo Basin forest; on the other hand, it echoes unaccountability. That is, the CBFP is not accountable for the success or failures of any collaboration developed amongst partners within its framework. Neither the donor nor the managers are accountable to the recipient countries. This lack of accountability of the CBFP means that there is no incentive to generate a relevant policy.

Conclusion: fostering local policy entrepreneurship
This analysis focused on the role of aid agencies within environmental cooperation in the Congo Basin. Based on the empirical results, we first highlight that aid agencies are not neutral third parties of environmental cooperation in Congo Basin as often pictured, but are rather highly effective policy entrepreneurs influencing the cooperation process and benefiting from it. Second, we could draw further conclusions at both theoretical and practical levels.
At the theoretical level, we expanded upon Mintrom's policy entrepreneurship theory (Mintrom and Norman 2009) to better capture weaknesses within environmental cooperation. By linking the policy entrepreneur's elements to policy matching and policy relevance, both the effectiveness of a policy and its relevance to local needs can be highlighted. In the specific case of Congo Basin countries, environmental cooperation is not intended for the display of effectiveness, but for the fulfillment of local needs. This enlarged framework is therefore more appropriate, and we hope that it can also be used in other regions to better pinpoint the weaknesses of environmental cooperation as well as brainstorm solutions.
At the practical level, we showed within our empirical section that an environmental cooperation in the Congo Basin that promotes high effectiveness and neglects to match local needs, as portrayed within Case Study 1, will always produce irrelevant outcomes for Congo Basin countries. In the same vein, we also conclude that an environmental cooperation initiative within the Congo Basin that matches local needs, but has the low effectiveness associated with the paucity of policy entrepreneurship, will not yield relevant policy outcomes, as leadership by example and rigor are needed to deliver a project from its design to realization in the field (Rogerson 2005).
The optimal solution for environmental cooperation in the Congo Basin seems to be a result of coupling at least a critical level of policy entrepreneur effectiveness to a critical level of match with local needs (H3) with the high competences the actor has already. We consider the aim of an actor who has both high entrepreneur effectiveness and strong match as unrealistic to achieve. Looking for such an ideal actor would result in failure and produce a symbolic aid policy. However, achieving the more modest solution is both possible for international aid agencies and local policy entrepreneurs, given that several conditions are fulfilled by each of these actors.
International or northern aid agencies need to be more accountable to recipient countries if they want to match their actions with local needs. In accordance with Easterly and Pfutze (2008), we postulate that the priorities of northern aid agencies to be financially accountable to Congo Basin countries is low on the agenda. We suggest, as a palliative, that Congo Basin countries should have a 'veto right' on aid agencies, in this case GIZ's annual budget received from different donors for the sustainable management of the Congo Basin forests. This would enable Congo Basin countries to prevent aid agencies from receiving further funds on their behalf in case there are mismatches between their actions and local needs. This non-objection right would constitute a good incentive for GIZ particularly and other northern aid agencies to meet a critical level of match with local needs within their different projects.
Local policy entrepreneurs who are currently nonexistent or poorly promoted within the environmental cooperation landscape in the Congo Basin could also be a very good option for guaranteeing the relevance of environmental cooperation outputs for Congo Basin countries in accordance with the Paris agenda on aid effectiveness. 14 Due to their regional proximity with recipient countries, local aid agencies strengthening their policy entrepreneur effectiveness skills could better guarantee the match of environmental cooperation objectives with the local needs of Congo Basin countries. Some conditions that surrender their actions could assure that a critical level of effectiveness will always be reached, such as self-budgeting to guarantee themselves a certain autonomy in their internal affairs management, as well as regular participation at capacity building courses on good management practices and tips.
As a final remark, we acknowledge that it is very challenging to identify the exact critical level of entrepreneurship and match with local needs to guarantee continuous policy relevance of environmental cooperation outcomes. Other experiments are warranted. The trial and error approach would certainly be of interest. However, the results of this work show that, so far, the strong management approach of GIZ will not contribute to fulfilling Congo Basin countries' needs as long as there is no match. The common northern policy of effectiveness will not contribute to the solution in this context either, as it is easier for the diverging interests of the North and South to meet a consensus on effectiveness of the policy entrepreneur than on the match with local needs.