Understanding Agenda-Setting and Outcomes in Global Tuna Governance

• In Event: Collaborative and Polycentric Governance

Abstract

Over 5 million tonnes of tuna are caught globally every year by fisheries in over 80 countries. This catch is valued at nearly USD 40 billion (McKinney et al., 2020). Since tuna are highly migratory and the catch must be sustainable throughout their entire trans-oceanic distributions, the world’s tuna fisheries are governed through five regional intergovernmental organizations called Regional Fisheries Management Organizations (RFMOs), each of which has jurisdiction over a sperate ocean basin. Each year, fishery management measures are negotiated by governments with national interests in tuna fisheries at general meetings. In addition to government officials, these meetings also include non-state actors (NSA) including representatives from fishing companies (and other related seafood sectors) and non-governmental organizations. Although NSAs do not participate directly in the negotiations, NGO observers are allowed to submit position letters at the annual meetings. Typically, these letters focus on areas of improvement for the fishery as well as aspects related to transparency at sea. Recent work has investigated how NSAs participate and influence RMFO decision-making processes (Schiller et al. 2021, Schiller et al. 2023), which is in keeping with ongoing work aimed at better understanding regime complexity in global governance including the role of NSAs in navigating multiple forums (Henning and Pratt, 2023; Eilstrup-Sangiovanni and Westerwinter, 2022; Abbott and Faude, 2021).

One of the key challenges associated with obtaining information on the socio-political dimensions of such NSA engagement is the diverse and copious qualitative data sources that could be investigated (e.g., meeting summary documents, transcripts, participant lists that typically include hundreds of individuals, annual position letters and advocacy statements, etc.) and the manual effort required to gather and analyze this information in a systematic way. Although such content analysis-based work can lead to meaningful results and observations, the acquisition of data is especially time consuming and includes the potential for human error. To this end, we leverage a natural language processing model that enables faster and more reliable data acquisition. Here, we test this approach relative to an existing analysis on NSA advocacy at tuna RFMO meetings. Drawing from Schiller et al. (2021) and context-specific knowledge of global fisheries governance, we use a seeded LDA model to identify the topics discussed at RFMO meetings in position letters and advocacy statements as well as outcome documents. The analysis allows us to identify the topics, backed by which NSAs, that made it to the RFMO agenda and those topics that did not, and it will inform a subsequent assessment of whether and through what channels NSA have influenced RFMO governance.

Authors

• Takumi Shibaike, Syracuse University

• Graeme Auld, Carleton University

• Laurenne Schiller