Investigating the Influence of Non-state Actors on Amendments to the CITES Appendices

Abstract Non-state actors are playing an increasing role in global environmental governance. Elucidating the modalities and implications of this engagement is important to understanding international policy-making processes. CITES, the Convention on International Trade in Endangered Species of Wild Fauna and Flora, is the primary mechanism for regulating international wildlife trade. It functions by listing species in its Appendices with corresponding trade controls. Accurately listing species in the Appendices is therefore fundamental to the Convention’s effectiveness. We investigate the influence of non-state actors on amending the CITES Appendices using an established framework for assessing NGO influence in international environmental negotiations. We find that non-state actors have been successful in issue framing and agenda setting, and in influencing the position of other actors and final decisions. We also find evidence that NGOs have sought to abuse CITES in pursuit of “campaign” victories, including claiming unwarranted victories, thus undermining NGO legitimacy and accountability. We recommend that the CITES parties seek the most robust science to inform decision-making on proposed amendments to the appendices, which should be broadened to include socioeconomic and economic considerations in order that proposals are evaluated in their real-world context. We further recommend that NGOs should seek to fully understand decision-making in the Convention in order to maximise their legitimate contribution to CITES. Further research is needed to fully elucidate the influence of non-state actors in CITES.


Introduction
Non-state actors are playing an increasing role in global environmental governance. Elucidating the modalities and implications of this engagement is both of research interest and important to understanding contemporary international policymaking processes (Larsen & Brockington 2018;Challender et al. 2015a). CITES, the Convention on International Trade in Endangered Species of Wild Fauna and Flora, is the primary mechanism for regulating international wildlife trade (CITES 2018a;Wijnstekers 2018) and provides a unique arena for investigating non-state actor influence in international policy-making. This is because (i) by design, non-state actors have played a role in CITES proceedings since the Convention's inception in 1975; (ii) although CITES is concerned primarily with sustainability in international wildlife trade, the issues it addresses are inherently political, economic, and cultural, as well as emotive to some actors; and (iii) the conduct of non-state actors has recently been called into question by some CITES parties concerned that nongovernmental organisations (NGOs) are "overstepping the mark," resulting in calls for a code of responsibility for NGOs (CITES 2017).
CITES functions by listing species in one of three Appendices, with corresponding trade controls based on scientific criteria (CITES 2018b;CITES Res. Conf. 9.24, Rev. CoP17). Although the contribution that non-state actors make to CITES meetings has been explicitly recognised (see Res. Conf. 13.8 [Rev. CoP16]; Duffy 2013), it has also been argued that some NGOs seek to manipulate the science-based approach of CITES in pursuit of their own often narrow agendas, primarily by attempting to ensure that species become or remain listed in specific appendices (Duffy 2013;Webb 2013). This can be manifested through building anti-trade alliances with receptive parties and/or by "soft-steering," whereby NGOs appeal to science to enhance the legitimacy of their own policy positions (Riise 2004;Bryant 2009). It is also clear that a number of NGOs measure success in terms of increased regulation or curtailment of trade and then demonstrate their influence in CITES to their membership and donors (e.g., Humane Society International 2014, 2019a,b, IFAW 2016see Jepson 2003). However, this approach can undermine the scientific credibility of CITES when it results in the unwarranted inclusion, transfer between, or retention of species in the Appendices (Jepson 2003;Webb 2013) and may overburden the appendices and dilute conservation funding (Martin 2000;Mofson 2000). It can also result in one-size-fits-all policies defined at the global level, but which disregard national management regimes, local livelihood considerations, and the cost of implementing appendix listings (Cooney & Abensperg-Traun 2013;Challender & MacMillan 2014).
Despite scientific criteria for amending the Appendices, listing decisions are frequently characterised by heated and polemic debate (Hutton 2000;Gehring & Ruffing 2008) and in reality are made for political, economic, philosophical, and even emotional reasons, as well as scientific reasons (Vincent et al. 2013;Webb 2013). This has led to criticisms over inconsistencies in the application of the listing criteria. For example, it is argued that charismatic megafauna are overrepresented in the Appendices (Webb 2000;. Debate occurs among parties as well as observers who attend and contribute to Conference of the Parties (CoP) meetings but cannot vote. Observers include non-Party countries, United Nations agencies, and other technically qualified organisations, including intergovernmental organisations (IGOs) and NGOs, which can vary in philosophy, raison d'être, size, and agenda and include preservationist, animal welfare organisations, and pro-hunting groups (see CITES 2013).
As of 2018, 2.8 percent of CITES-listed species (1003 species) were included in Appendix I, which bans commercial international trade; 96.6 percent of listed species (34,596 species) were included in Appendix II, in which trade is closely regulated; and <1 percent of listed species (202 species) were included in Appendix III (CITES 2018c). Amendments to the Appendices, including the listing, transfer between Appendices, and deletion of species, are based on an evaluation of proposals against biological and trade criteria (the CITES listing criteria) and may be proposed only by parties (CITES Res. Conf. 9.24,Rev. CoP17). Amendments are typically considered at CoP meetings, which are held approximately every three years, and are adopted by consensus or otherwise subject to at least a twothirds majority vote of parties present and voting (CITES 2018d).
In this article, we investigate the influence of NGOs and IGOs (together the main observer groups in CITES), or collectively non-state actors, on decisions to amend the CITES Appendices, an identified research need (Challender et al. 2015a). Our aim was to investigate if and how these organisations influence decision-making, to consider any associated implications of this influence for CITES, and, where appropriate, to make recommendations supportive of science-based decision-making. We found that non-state actors can frame specific issues and help shape the position of CITES parties on final agreements. In particular, we found that NGOs are able to successfully raise awareness of species trade issues and set the agenda in terms of proposed amendments to the appendices. We also found evidence of instances where NGOs have sought to abuse CITES in pursuit of campaign victories and claim unwarranted victories for the listing of species, which threatens the Convention's scientific credibility and calls into question NGO legitimacy and accountability. On the basis of our analyses we make recommendations supportive of science-based decisionmaking in CITES.

Methods
In order to identify and elucidate instances where non-state actors exerted influence on decisions to amend the CITES Appendices, we used Betsill and Corell's (2008) analytical framework developed specifically for assessing NGO influence in international environmental negotiations. This framework provides a means of assessing NGO (or here non-state actor) influence along two dimensions: (1) participation of non-state actors, and (2) subsequent effects on the behaviour of other actors (e.g., states) by gathering and analysing data along both dimensions. We focus primarily on subsequent effects on the behaviour of other actors, recognizing that the number of observers attending CITES CoPs has increased over time (Figure 1), and that provisions have been adopted to ensure that observers have greater access to decision-making forums in CITES (and thereby greater opportunities to exert influence) and can actively participate in proceedings (see CITES Res. Conf. 13.8,Rev. CoP16 and Decisions 11.14,11.70,11.71,11.73,and 11.128).
We focused on influence both in the negotiation process, including issue framing (how a problem was conceptualised prior to and during discussions), agenda setting (identifying a problem and calling on states to act), shaping the positions of state actors, and the outcome of the negotiation, including any final agreements regarding procedural issues (e.g., decisions on future participation from non-state actors) and substantive issues (e.g., whether decisions reflect non-state actor positions on particular issues) (Betsill & Corell 2008). Focusing on the negotiation process as well as outcomes is important in recognizing that influence extends beyond official voting forums to other informal settings (e.g., social events, corridor talk). We defined influence as "when one actor intentionally communicates to another so as to alter the latter's behaviour from what would have occurred otherwise," to ensure our analyses were robust (Betsill & Corell 2008). We also expanded Betsill and Corell's framework to include an additional influence indicator to capture dynamics of non-state actor influence not covered in their original framework.
We used two main methods to determine whether and how non-state actors influenced listing decisions in CITES-process tracing and counterfactual analysis (Betsill & Corell 2008). Process tracing entails assessing causality by constructing causal chains linking non-state actor activities with other actors' responses. Counter-factual analysis entails considering what might have happened if one examined variable were removed from the chain of events (Betsill & Corell 2008). In this case, for example, determining what would have happened if a non-state actor had not sought to exert influence in a particular way or at a certain time. We collected data on non-state actor activities and intentional communications with state actors within CITES, along with the processes and outcomes of specific decisions. We sourced data by conducting semistructured interviews with 38 CITES experts (see Table 1) and by drawing on primary and secondary data sources (e.g., NGO lobbying materials, policy briefings, CITES documents, and published literature).
Experts were identified from key texts on CITES (e.g., Hutton & Dickson 2000;Reeve 2002;Oldfield 2003) and through snowball sampling (Newing 2010;Bernard 2006). Respondents included NGO and IGO staff actively working on CITES issues, current and former party representatives, the CITES Secretariat, and other experts; respondents typically had many years of experience in CITES. Interviews were conducted face-to-face or by Skype/telephone and lasted 70.99 ± 25.20 minutes (mean ± SD, n ¼ 36). On three occasions two or more experts were interviewed together. In all but three interviews, a verbatim record of the interview was made and transcribed for analysis; when interviews were not recorded, detailed notes were made on interview content.
Interviews were conducted inductively, but an interview guide was used to introduce topics including means of non-state actor influence; examples of influence; the position of key actors; forms of communication between state and non-state actors; and outcomes following communication with non-state actors. All interview transcripts were imported into QSR International's NVivo 9 software and coded to sort the data by influence indicator, species, and issue (Newing 2010;Bernard 2006). Once data were coded, we applied process tracing and counter-factual analysis to construct detailed narratives of instances of non-state actor influence, and any points of uncertainty were clarified and/or additional details were obtained by following up with respondents by email. We also triangulated between interview responses and other data sources wherever possible to ensure authenticity in reported communications, cases of influence, and outcomes (Betsill & Corell 2008;Bernard 2006). We selected quotes from respondents where they best supported the narrative constructed and present them in our results.

Results
We were able to construct detailed narratives for six species case studies and other issues demonstrating that non-state actors have an influence on the negotiating process and on outcomes. This includes issue framing, agenda setting, and shaping the position of other actors, and on the final agreement on procedural issues and on substantive issues. We also identified negative implications from the engagement of non-state actors. The number of detailed case studies is low, especially in relation to interview effort, which we attribute to our respondent selection. While respondents provided interesting insights on influence within CITES, many were unable or reluctant to provide in-depth details on specific instances of non-state actor influence. Nonetheless, our analyses demonstrate that non-state actors have a high level of influence in CITES listing decisions based on Betsill and Corell's (2008) analytical framework-that is, they have some success in shaping the negotiating process, and effects of their participation can be linked to outcomes. There is no issue more controversial in CITES than measures adopted to control trade in the African elephant, specifically its ivory. Originally listed in Appendix III by Ghana, the African elephant was transferred to Appendix II in 1976 (CoP1), and controversially to Appendix I in 1989 (CoP7). As a result of poaching for the ivory trade, Africa's elephant population fell from 1.5 million to just over 600,000 in the 1980s, and the poaching crisis of the 1970s-1980s was used in the build-up to and during CoP7 in 1989 to frame the issue as an unprecedented crisis, and one that could be addressed only by a global ban on ivory trade (Anon. 1989;EIA 1989;Bonner 1993). The framing of this issue as a crisis was achieved by a number of non-state actors, promoting the same goal as the solution to saving the African elephant, including shaping public opinion. The proposal to transfer the African elephant to Appendix I at CoP7 was submitted by seven countries (Austria, Gambia, Hungary, Kenya, Somalia, United Republic of Tanzania, and the United States) and led by Tanzania, but the issue was framed by a coalition of non-state actors. As asserted by one interviewee: "NGOs drove the listing of the elephants at CoP7, there is no question" [R-22].
A number of initiatives contributed to framing the issue (Table 1). The African Wildlife Foundation (AWF), a US-based NGO, launched the year of the elephant in 1988 and started to raise awareness about the decline of the African elephant through a direct mailing campaign in the US (Bonner 1993). The EIA (Environmental Investigation Agency) also solicited the support of American NGOs in backing Tanzania to propose the transfer of the species to Appendix I. The EIA also produced emotive film footage of deceased elephants and the ivory trade taking place, and they utilised media connections, both in television and print, to raise awareness of the issue and to build and shape public support, particularly in receptive countries, including the US and United Kingdom, which included high-profile television news coverage (Thornton & Currey 1991). The EIA further raised awareness and the profile of the issue by publicly questioning the CITES Secretariat's stance on, and involvement in, the ivory trade (EIA 1989). The same organisations also produced an extensive report, A System of Extinction: The African Elephant Disaster, which was distributed to all participants at CoP7 and which argued that a complete ivory trade ban was the only solution with which to save the African elephant from extinction (EIA 1989;Thornton & Currey 1991). These activities ran parallel to a coalition of NGOs comprising the African Elefund, Zoocheck (now the Born Free Foundation), the World Society for the Protection of Animals (WSPA -now World Animal Protection), and Care for the Wild International, which launched an "Elefriends" campaign. This resulted in a petition signed by more than one million people pledging not to buy, sell, or wear ivory, which was presented at CoP7, and Elefriends called for an Appendix I listing as critical to saving the African elephant (Thornton & Currey 1991).
The Ivory Trade Review Group (ITRG; mainly representatives from IUCN, TRAFFIC, WTMU [the Wildlife Trade Monitoring Unit of the World Conservation Monitoring Centre, Cambridge UK], and the CITES Secretariat), experts on the ivory trade and an authoritative source of information on decision-making, also recommended that a ban was the most effective way of controlling the ivory trade, and thereby saving the African elephant. This was despite all but two members of the group acknowledging that an Appendix I listing would be unlikely to result in the control of trade in ivory (see Anon. 1989). These actions, together with the nature of CITES at the time, characterised by a lack of agreement among African range states within CITES, and relationships between range states and NGOs, characterised by the latter's superiority and direction, meant that it was, as one respondent stated: "[F]airly easy to get a number of Parties very concerned about elephants" [R-25].
As a result of these activities, each of which reflected the position of non-state actors, the CITES parties transferred the African elephant from Appendix II to Appendix I at CoP7, with a vote of 76 in favour of the ban, 11 against it, and four abstentions. This occurred despite dissenting voices from southern African states, where elephant populations were well managed and seemingly increasing, but which were effectively silenced by the largely northern NGOs acting as key knowledge brokers (Duffy 2013). Without these activities, which were, and arguably remain unprecedented ('t Sas-Rolfes 2000), it is likely the proposal to transfer the African elephant from Appendix II to Appendix I would not have been adopted.

Agenda setting
We found multiple examples demonstrating the influence of non-state actors in agenda setting, including proposals concerning the basking shark (Cetorhinus maxmius), Central and South American tree frogs (Agalychnis spp.), bluefin tuna (Thunnus thynnus), and African scorpions (Pandinus spp.). Agenda setting is seemingly the most significant role that non-state actors play in the process to amend the CITES appendices (see Table 2).

Basking shark (Cetorhinus maximus)
The basking shark, the second-largest fish in the world, was listed in Appendix III by the UK in 2000 and in Appendix II at CoP12 (2002). Traded for its fins, oil, cartilage and meat and threatened by targeted and incidental bycatch from fishing (CITES 2002), awareness of the need for greater regulation of trade in its derivatives came from non-state actorsspecifically, a UK NGO, the Shark Trust, which played a key role in getting the species on the CITES agenda. Prior to CoP11 (2000), and in recognition of the potential threat international trade posed the basking shark, the Shark Trust submitted a draft proposal to include the species in Appendix II, to the UK CITES Scientific Authority, JNCC (Joint Nature Conservation Committee). Following evaluation of the proposal against the listing criteria (Res. Conf. 9.24, Rev. CoP17) and consultation with DEFRA (Department for Environment, Food, and Rural Affairs), the UK Management Authority, and with ministerial approval, the UK decided to pursue the listing of this species in Appendix II and worked with the Shark Trust on the compilation of data and preparation of the proposal (see Table 2). The UK also invested significant time and resources in ensuring the proposal was robust-for example, by commissioning DNA studies on the species and conducting satellite tagging of sharks around the UK to determine whether they comprised one or multiple populations and determining migration patterns. Although the proposal at CoP11 (2000) was rejected, an improved proposal was submitted to CoP12 (2002), which was adopted. Without the Shark Trust's bringing the issue to the attention of the UK CITES authorities, which were otherwise unaware, these events would not have happened; the Shark Trust therefore played a key role in getting this species on the agenda.  (2010), and, having become aware of the trade and the potential need for greater regulation, a coalition of NGOs played a crucial role in getting them onto the CITES agenda. Despite Agalychnis spp. not being CITES listed, Humane Society International (HSI) became aware of large volumes of trade in these species from US importation data. As an example, trade into the US involved more than 200,000 tree frogs between 1999 and 2008, from a range of exporters, including Guatemala and Nicaragua (CITES 2010a). On the basis that this trade was potentially unsustainable (see CITES 2010a) and in seeking further regulation, a group of NGOs, including HSI, Defenders for Wildlife, the Species Survival Network (SSN), and Pro-Wildlife, started working with local frog biologists and CITES authorities in Central America to draft an Appendix II proposal. Specifically, this NGO coalition held meetings with the CITES authorities in each country from which the frogs were originating. They presented trade data to the authorities in writing, over the phone, and through personal visits. They also gathered further data to support the needs of local biologists for the proposal, answered any questions the CITES authorities and biologists had, and facilitated communication between the parties involved. The result of these activities was a proposal to list Agalychnis spp. in Appendix II, which was submitted by Honduras and Mexico; the submission of this proposal demonstrates that these NGOs actively participated in setting the agenda for CoP15.
These NGOs also generated support for the proposal at CoP15 by sending letters directly to parties urging their support. They also prepared factsheets for lobbying at CoP15 in all three official CITES languages and included a recommendation to support the proposal in SSN's Digest, which is its brochure of recommendations on listing proposals at CoPs. By engaging in all these activities, this NGO coalition was able to build support for the proposal, which got the support of all range states, as well as build regional camaraderie around its submission. In sum, they created and raised awareness of the trade and the potential need for greater trade regulation, which would not have occurred otherwise. By working with the receptive range states in this way, they were able to participate in setting the agenda.

Bluefin tuna (Thunnus thynnus)
The bluefin tuna is another controversial issue in CITES. As a commercially exploited aquatic species, it has been subject to debate about the most appropriate management regime to ensure harvest and trade remain sustainable. The species is currently managed under the International Commission for the Conservation of Atlantic Tunas (ICCAT); the role that CITES could play in its conservation has polarised debate in the Convention (Fleming et al. 2013). It was originally subject to two CITES proposals in 1992 (CoP8), in which Sweden proposed listing the western Atlantic population in Appendix I and the eastern Atlantic population in Appendix II, but neither was adopted. However, in response to the ongoing deterioration of bluefin tuna stocks (e.g., Taylor et al. 2011;CITES 2010b), a proposal to list the species in Appendix I was submitted by Monaco for consideration at CoP15 (2010), and non-state actors played a significant role in ensuring it was on the agenda.
The first draft of the proposal was written by a World Wide Fund for Nature (WWF) employee and was then subject to input from other WWF staff members, as well as personnel from TRAFFIC, and lastly sent to Monaco, which submitted the final version to the CITES Secretariat for consideration by the Parties at CoP15 (2010). In other words, the proposal was written on the basis that Monaco would submit it, and which had been sounded out prior to the proposal being written, as a potential proponent country. This was possible because of close personal connections between WWF staff members and senior officials in Monaco, who agreed to submit the proposal if WWF drafted it. As stated by one respondent: [the proposal] was written by a WWF staff member … the first draft was done by him, and then with input from other WWF and TRAFFIC people, and then taken to Monaco; … Monaco had been sounded out beforehand as being a potentially winning country. [R-16] By utilising personal connections in this case, WWF were able to set the agenda by engineering the submission of the Appendix I proposal, which would not have occurred otherwise. The bluefin tuna is a further example of an NGO working with a receptive Party to set the agenda.

African scorpions (Pandinus spp.)
Three African scorpions-the emperor scorpion (Pandinus imperator), the giant Senegalese scorpion (P. gambiensis), and P. dictator-which are popular in the live pet trade, provide another example of how non-state actors, or more specifically, individuals, can set the agenda in CITES. Each of these species was listed in Appendix II at CoP9 (1994). However, their being on the agenda emerged from conversations between a recognised CITES expert (interview respondent R-24), who is a regular observer and participant at CoPs, usually on behalf of NGOs, and scorpion experts, who suggested that regulation of the trade in these species was necessary for their conservation.
Although the exact distribution of these species is unknown, they are found in West and Central Africa, including Ghana, and because of the personal relationship between the head of the Ghanaian Management Authority and this expert, the Appendix II proposal was submitted to CoP9 in 1994. Dialogue between the two (by telephone) resulted in the proposal being written by the expert and submitted by Ghana, and it was adopted without opposition. As stated by respondent 24: I made a point of actually telephoning the head of the management authority in Ghana at the time, whom I knew, and asked if he would be interested. He was very interested, so I drafted the proposal and sent it to him, and it actually went through unopposed. [R-24] Without this particular expert's bringing the issue to the attention of the Ghanaian CITES authorities, the proposal probably would not have been submitted. This is another example of how a non-state actor, or in this case an individual, in collaboration with a receptive Party, set the agenda within CITES.
The above case studies show that non-state actors, primarily NGOs, but also individuals, influence decision-making in CITES by agenda setting, which is seemingly the most prominent role that they play in amending the appendices. This is consistent with the experiences of many respondents who corroborated this point: it was only when I came to [interviewee organisation] that I really saw the extent of NGO involvement in preparing proposals, it's arguably the most significant part that they play in the whole process. [R-16] 3.3. Shaping the position of other actors, and final agreement (substantive issues)

Polar bear (Ursus maritimus)
The polar bear is a controversial topic in CITES, but non-state actors have played a role in shaping the position of state actors on proposals to transfer the species from Appendix II to Appendix I at CoP15 (2010) and CoP16 (2013). The species is threatened primarily by climate change (Wiig et al. 2018), but there also have been claims that it is increasingly threatened by international trade and therefore in need of strict regulation in the form of an international trade ban (e.g., NRDC 2013). Both proposals to transfer this species to Appendix I were submitted by the US. However, despite some support for the proposals at CoP15 and CoP16 from observers, not all non-state actors were supportive of them. The analyses of proposals submitted to CoPs, produced jointly by IUCN and TRAFFIC for CoP15 and CoP16, stated that the species did not appear to meet the biological listing criteria for inclusion in Appendix I; similarly, the TRAFFIC recommendations produced for CoP15 and CoP16 also recommended that the proposal should be rejected. These analyses directly influenced the decision of state actors at CoP15, simultaneously contributing to the final agreement made by the parties. As stated by one respondent: like in the Polar bear intervention when six countries said they did not support the Appendix I proposal because the IUCN/TRAFFIC Analyses, and the TRAFFIC recommendations said that it did not merit, the scientific information did not merit its inclusion. [R-3] The position of six state actors at CoP15 was directly influenced by the IUCN/TRAFFIC Analyses and the TRAFFIC Recommendations. Parties cited information from these organisations directly in their deliberations, and the proposal at CoP15 was ultimately rejected by the parties, the position of state actors and the final agreement having been influenced by IGO and NGO analyses.

Procedural issues
Influence in procedural terms is demonstrated by the adoption of CITES Res. Conf. 13.8 ("Participation of Observers at Meetings of the Conference of the Parties") at CoP13 (2004), though this resolution has since been repealed (see infra, Part 4., Discussion). This resolution recognised the valuable contributions that observers make to CoPs (non-state actors are observers in CITES), recommended that observers have space in meeting rooms and time to make interventions during proceedings, and that knowledgeable observers be invited to participate in CoP working groups. Acknowledgement of the role of observers in this resolution means that they directly secured a future role in CoPs, demonstrating their influence in procedural terms. As stated by one respondent: The parties … , they … passed a resolution recognising the value of NGO's and what they provide, including technical information and all sorts of other things that NGO's provide to the CITES discussions. [R-35] Specifically, IUCN and TRAFFIC are seen as credible, scientific, and objective sources of advice in this regard, according to multiple respondents representing government agencies: we value most the IUCN Analyses and the TRAFFIC Analyses erm, you know, because they're talking to experts in the field, they're not for or against, they don't come with a fixed position before they even look at the proposal. [R-12] and I think IUCN are an important, in terms of information … the specialist groups I think are quite important, and you know, the IUCN Analyses, I think people turn to a lot as a more, impartial, objective assessment of proposals against the criteria.
[ R-11] 3.5. Claiming unwarranted influence and seeking to undermine credibility We also identified negative implications from the engagement of non-state actors, in the form of NGOs, claiming unwarranted influence on listing decisions and apparently seeking to undermine the scientific credibility of CITES for organisational benefit. When decisions are taken in line with non-state actor policy positions (and whether influence was exerted by non-state actors during the decision-making process or not), they are widely considered victories among NGOs. Conversely, decisions that are contrary to non-state actor policy positions are considered losses, as asserted by a respondent from an animal welfare NGO: "because … we have a scorecard, and the ones that went the way that we advised, are considered victories, and the ones that didn't are considered losses" [R-9].
Moreover, there are instances when NGOs, specifically, have claimed victories on listing decisions that were unwarranted, because they exerted little or no influence on the outcome. This undermines NGO legitimacy and accountability, especially if these organisations count these victories amongst their organisational achievements as some NGOs do (e.g., Born Free Foundation 2016;Humane Society International 2014;Pacelle, 2010). For instance, an NGO claimed a victory for setting the agenda and for the successful listing of a species in the Appendices despite it playing no causal role in either. As one respondent, formerly of a CITES Party, stated: [A country] co-sponsored several [B species] proposals. I know for a fact how it happened. I know for a fact who in the government of [A country], and I have heard from some NGOs said, 'We did it. I sent that email to the management [authority].' I don't correct them because it sounds like ego if I say, 'I know how it really works.' [R-22] In this instance, the NGO claimed a victory that was unwarranted because it played no causal role in setting the agenda, which arguably demonstrates a lack of understanding of the decision-making processes of CITES Parties. This is attributable, at least in part, to the fact that many NGO employees have never worked for governments and therefore do not understand government, i.e., Party, decision-making mechanisms. This was asserted by a former CITES Party representative: in all fairness many NGOs have never worked for government. Now I hired two people to work for me in [A city], they'd both once worked for government. I worked for government and a lot of people I work with once worked for government. That means we understand how governments make decisions. [R-22] We also found evidence of instances where NGOs specifically have sought "campaign" victories, i.e., they have campaigned for a change in the appendix in which a species is listed and have done so for the sake of claiming a victory and/or because securing a specific listing would enable their organisation to secure future funding for conservation activities (see Table 2). As stated by one respondent: There was an NGO who worked on, from a protectionist NGO said, 'See, if you worked on [species A] like we did this time … you could have won. Stop working on things that are so hard to win.' [R-22] In this instance, the successful listing of a species was seemingly pursued for a campaign victory, beyond the potential conservation benefit that could be conferred to the species through an amendment to the CITES appendices. In another example, an NGO pursued the listing of a particular species in Appendix I primarily because it would assist the organisation in fundraising (see Table 2).

Discussion
Our aim was to investigate if and how non-state actors influence decisions to amend the CITES appendices, to consider any associated implications of this influence, and, where appropriate, to make recommendations supportive of science-based decision-making in CITES. Although the number of case studies we were able to construct was limited, which, combined with our non-random selection of case studies restricts how generally our results can be interpreted, we have demonstrated multiple ways in which non-state actors, especially NGOs but also IGOs, influence decisions to amend the CITES Appendices. Using Betsill and Corell's (2008) analytical framework, our analyses demonstrate that non-state actors have a high level of influence on amendments to the CITES Appendices, in that they have some success in shaping the negotiating process and effects of their participation can be linked to outcomes.
Despite criticisms of the role and focus of non-state actors in amending the CITES appendices (e.g., Duffy 2013; Webb 2013), our results demonstrate both positive and negative aspects of their engagement, with the most significant role that they play, at least based on this evidence and with reference to NGOs in particular, being agenda setting. That is, NGOs raise awareness of the potential need for greater trade controls for species in trade and work with receptive Parties through a number of means (e.g., holding workshops to coordinate with parties, utilising individual relationships) to develop listing proposals for submission. This is demonstrated by case studies on the basking shark, Central and South American tree frogs, bluefin tuna, and African scorpions. By doing so, NGOs actively contribute to global environmental governance by putting these issues on the agenda at CoP meetings and presenting the parties with the opportunity to deliberate on the merits of the proposals. Without NGO engagement in the examples discussed, it is unlikely that the proposals would have been submitted otherwise and changes to trade controls adopted. While we characterise agenda setting as positive influence, we do so only insofar as it extends to NGOs supporting the international conservation community by raising awareness of species trade issues. We purposefully make no evaluation of the impact of amendments to the appendices on species' status, recognising the complexity of causally attributing the implementation of CITES to species' conservation status (see Challender et al. 2015a).
Other ways in which non-state actors exert influence include through issue framing, shaping the position of other actors and the final agreement, and procedural issues. The African elephant case study demonstrates how nonstate actors, led by NGO coalitions, framed the African elephant poaching and ivory trade crisis of the 1970s-1980s as one of unprecedented concern for the species. This was despite most experts in the CITES ITRG recognising that an international trade ban was unlikely to solve the poaching crisis (Anon. 1989). Critical here was the unprecedented and concerted NGO activity in the build-up to CoP7 in 1989, including courting press and media attention and shaping public opinion, largely in the most developed nations, to the point that decisions contrary to those being supported by NGOs became unrealistic options for many parties. This example demonstrates how non-state actors played a role in global environmental governance, specifically by limiting the acceptable decisions open to the CITES parties, in light of the crisis facing the African elephant, or in other words, framing this issue as having a singular policy response-an international trade ban.
Our results indicate that in terms of listing decisions, non-state actors are not inherently predisposed to increasing trade regulations. This is demonstrated by evidence that IUCN and TRAFFIC shaped the position of state actors on the proposal concerning the polar bear at CoP15, as well as the final agreement. The Analyses prepared by IUCN and TRAFFIC were specifically referred to by Parties at CoP15 as being the basis of their decisions not to support the transfer of the species to Appendix I, on the basis that international trade does not present a threat to the species. This is an example of an IGO and NGO playing a direct role in global environmental governance by helping to shape both the position of state actors and simultaneously the final agreement.
Since its inception, CITES has encouraged the contribution of technical and scientific knowledge from non-state actors, and the adoption of Res. Conf. 13.8 (Rev. CoP16) in 2004 demonstrates the influence of non-state actors at CoP meetings in procedural terms, i.e., it provided observers with an ongoing mandate to provide input as key knowledge brokers (Bryant 2009;Duffy 2013). This reflects influence in terms of the negotiating outcome as well as the negotiating process (see Table 2; Betsill & Corell 2008). Resolution Conf. 13.8 (Rev. CoP16) explicitly recognised the value of nonstate actors as sources of technical expertise and contained provisions to ensure that they have adequate space at CoP meetings, that they are permitted a voice in discussions, and that they are permitted to take part in working groups at CoP meetings. Although this resolution was repealed at CoP17 in 2016, it was done so in recognition of how mainstream observers have become in the Convention's proceedings, to the point that the resolution was, in the opinion of the CITES Secretariat, considered redundant (CITES 2016). At CoP17, relevant provisions from this resolution were otherwise included in revised Rules of Procedure for CoPs and other CITES meetings, thus ensuring that observers continue to have an ongoing mandate in CITES (CITES 2016). Despite this recognition, a number of parties felt that NGOs in particular were subjecting parties, as sovereign states, "to undue pressure and other unacceptable activities" and proposed a code of responsibility for NGOs at the 69th meeting of the CITES Standing Committee in November 2017 (see CITES 2017). This code of responsibility included guidance for NGOs on making interventions at meetings, the submission of documents, and behaviour of delegates, including, inter alia, publicity and advertising surrounding meetings. However, it attracted little support from parties and was not adopted.
Despite the valuable contribution that non-state actors make to CITES, our results demonstrate negative aspects of their engagement in the process to amend the appendices, in particular NGOs. This is specifically through claiming unwarranted influence on listing decisions and seeking to undermine the scientific credibility of CITES by ensuring certain outcomes of listing decisions for organisational benefit. Although quantifying how commonplace these dynamics are is not possible from our results, we have presented evidence of NGOs actively pursuing campaign victories regarding listing decisions-that is, they have sought to ensure specific outcomes are achieved so that they can claim "victories" at CoP meetings and convey their success to members and donors (see Table 2; e.g., Humane Society International 2014; IFAW 2016; Pacelle 2010). This corroborates Jepson (2003) on outcomes from decisions to amend the CITES Appendices constituting organisational achievement among NGOs. Our results also reveal that there are instances when NGOs have claimed such victories without having played a causal role in decision-making, i.e., without having influenced any other actors, specifically state actors, on decisions. Thus they claim unwarranted victories for the successful listing or transfer of species between appendices. This seemingly demonstrates a lack of understanding of decision-making processes among CITES parties, a point made explicitly by one former CITES party representative. While these tactics may be beneficial for NGOs in terms of conveying successes to members and donors and associated revenue streams, especially around the time of CoP meetings, it raises a number of questions about NGOs in terms of their legitimacy and accountability (Jepson 2005;Christensen 2003). For example, are such organisations misleading their members and donors about their organisational achievements? Are they overstating their achievements at CITES meetings and the role they play in decision-making and thereby global environmental governance? If so, are they doing so in order to bolster revenues, and how are they accountable to members and donors?
Our results also indicate that there have been instances when NGOs have pursued tighter trade controls for specific species because the adoption of such controls would support future fundraising efforts. Again, this raises major questions about NGO legitimacy and accountability: What systems of accountability are in place for such organisations? Are they principally concerned with the potential benefits that listing species in CITES could confer or the financial sustainability of their organisation? These dynamics threaten to undermine the scientific credibility of CITES if they result in listing decisions that are not based on the listing criteria and/or the unwarranted listing of species (i.e., the inclusion of species in specific appendices despite their not meeting the criteria). Again, it is difficult to determine how widespread this dynamic is from our results. However, such cases could have broad implications. If decisions were made to amend the Appendices that were framed by non-state actors, or species were included on the agenda largely due to NGOs, and/or listing decisions were made at CoPs that were influenced by NGOs but which resulted in adverse consequences for local peoples' rights and livelihoods, NGOs would have achieved victories; but these victories would have placed implementation costs on range states and their local people, many of whom are likely to have been marginalised and constitute the rural poor, thus raising serious ethical concerns (Cooney & Abensperg-Traun 2013). The same applies for the impact on national management regimes more broadly and other implementation costs associated with appendix listings. There is also no guarantee that an amendment to a species' listing would confer a benefit to the species in question either. It is well known that the listing of species in CITES is no guarantee of ensuring sustainability in international trade or improving the status of species (Challender et al. 2015b;'t Sas-Rolfes 2000).
More broadly, the costs associated with appendix listings-for example, the loss of livelihood options-highlights the narrow focus of the CITES listing criteria. Although these criteria offer detailed guidance to parties submitting listing proposals on biological parameters (e.g., intrinsic and extrinsic vulnerability factors, extent of population declines), they fail to adequately address the reality of wildlife trade as a socioeconomic activity (see Res. Conf. 9.24 [Rev. CoP17]). This is limited to the chapeau text of Res. Conf. 9.24 (Rev. CoP17), which acknowledges the need for decision-making to take into account socioeconomic factors and Section 8.1 (Management measures) which considers these factors tangentially, and in reality decision-making on amendments to the appendices remains focused on trade controls (Abensperg-Traun 2009). The explicit consideration of the socioeconomic impacts of amending the CITES appendices, and the economic rationale for doing so, would allow debate on proposed amendments in their real-world context (Challender et al. 2015b) and could avoid situations in which setting policy at the global level has adverse consequences on the livelihood strategies of local communities and indigenous peoples.
This research has direct relevance to conservation policy. Non-state actors participate in global environmental governance through their varied roles in amending the CITES Appendices. The contribution that non-state actors make to CITES is important and valued, and their contribution to CoPs and other meetings has been recognised through the adoption of Res. Conf. 13.8 and subsequently through the Rules of Procedure for CoPs and other committee meetings. However, by design, CITES is a scientific Convention with clear criteria for listing species in its Appendices, but the Convention, and its scientific credibility, have the potential to be undermined where NGOs, in particular, pursue campaign victories, claim unwarranted influence over decision-making, and/or seek to manipulate decision-making away from the criteria for organisational benefit. In order to ensure that future listing decisions are based on the best available science, we recommend that the CITES Parties seek the most robust and authoritative scientific analyses to inform decision-making on amendments to the Appendices and exercise caution when interpreting policy advice from NGOs that may have narrow, specific agendas and inherent self-interest tied to the outcome of listings decisions. We also recommend that NGOs engaging with CITES processes should seek to ensure that they possess an in-depth understanding of decision-making among CITES parties, in order to avoid claiming unwarranted victories surrounding listing decisions. This, in combination with greater transparency about the role that NGOs play at CITES meetings and their impact on specific decisions, or not, would ensure that donors, members, and other stakeholders are not misled and would otherwise maximise the legitimate contribution that NGOs make to the Convention. This is the first application of Betsill and Corell's (2008) framework for assessing NGO influence (or here, more broadly, non-state actor influence) in CITES. The framework has provided a robust means of analysing non-state actor influence on amending the CITES Appendices, though the framework required adaptation to account for other dynamics (i.e., negative aspects) in terms of NGO engagement in this process. Our results and analyses do have limitations, not least the number of case studies we were able to construct, and further research is needed to more widely and deeply understand the influence of non-state actors in CITES decision-making in terms of amending the Appendices and otherwise. If future research focuses on amendments to the Appendices specifically, it could complement our characterisation of where, along the decision-making process, influence occurs and through what means. We recommend the prior identification of specific case studies and respondents for such research in order to help ensure that an adequate evidence base exists on which to draw. Thought should also be given to the sensitivity of data collection, meaning that respondents may be reluctant to provide detailed information. We encourage researchers to use Betsill and Corell's framework as presented here and apply it to other case studies in order to further elucidate non-state actor influence in CITES.
Funding D.W.S. Challender was supported by a joint Economic and Social Research Council (ESRC) and Natural Environment Research Council (NERC) studentship (number ES/I028420/1).