Out of Sync: The Making and Remaking of Data and Regulations on Greenhouse Gases at the International Maritime Organization

Abstract

This article explores the entanglements of the making of data and climate regulations at a specialized UN agency, the International Maritime Organization (IMO). Based on observant participation, document analysis, and interviews, I examine the politics of making and remaking a new data infrastructure—the Data Collection System (DCS)—and how the limitations and hopes invested in this infrastructure shape the global governance of greenhouse gas emissions in international shipping. I analyze how the DCS shapes the design and pacing of new regulations based on what data are available and when they are available, and how at the same time the politics and regulatory process of the IMO shapes the making and remaking of the DCS and how the data can be used. I unpack the politics that unfold when the production of data and regulations are coupled but are out of sync and shed light on the present and future roles of data in international negotiations. As the DCS and regulatory development at the IMO become entangled, I argue this creates a form of governance-through-data which slows down the regulatory process but nonetheless supports consensus-building among parties at a time marked by deep political tensions.

Introduction

This article is about the politics of making and remaking a new data infrastructure at a UN specialized agency, the International Maritime Organization (IMO), and how the limitations and hopes invested in this infrastructure shape the global governance of greenhouse gas emissions in international shipping. In 2016, the IMO adopted the Data collection system for fuel oil consumption of ships, requiring ships globally to record and report data to a centralized database. ¹ After adopting new regulations on ship energy efficiency in 2011, the IMO struggled to decide on any new climate regulations due to the disagreement between developing and developed countries over who should be responsible for global emissions reductions. Member states could, however, agree to start collecting data from ships, as part of a new three-step approach to regulatory decision-making: first, data collection, second, data analysis, and third, “decision-making on what further measures, if any, are required.” ² This new IMO Data Collection System (DCS) is now emerging as a central pillar in the decarbonization of international shipping, a sector responsible for 2–3 percent of global greenhouse gas (GHG) emissions (Faber et al. 2020).

Previous work in Science and Technology Studies (STS) has shed light on how data infrastructures are not neutral artifacts and how they can be performative, highlighting the importance of studying how data are produced in order to explore how politics are embedded in emerging infrastructures (Bowker 2000; Bowker and Star 1999; Edwards 2010; Flyverbom and Murray 2018; Gray, Gerlitz and Bounegru 2018). Jackson's (2015) work exploring the usefulness of the analytic of “repair” also sheds light on the need to attend to the politics that unfolds over the lifespan of infrastructures. We need to think temporally when studying data infrastructures, paying attention to the repair work that keep infrastructures running. While literature on data infrastructures is burgeoning in STS (Hoeyer, Bauer and Pickersgill 2019; Kitchin 2022; Vertesi et al. 2019; Ziewitz 2016), there has been less attention to the development of data infrastructures within international political institutions and their relations with regulatory processes (but see some exceptions within EU studies, for example Trauttmansdorff and Felt (2023)). As data and data infrastructures are becoming increasingly important in “evidence-based” and “data-driven” policymaking at international organizations, studying how such infrastructures are made and used over time is critical for understanding contemporary global governance. The small but rich subfield of studying parliamentary politics in action in STS has highlighted the fruitfulness of applying an STS lens to the study of traditional political institutions, not just politics in new sites (Asdal and Hobæk 2020; Barry 2020; Dányi 2020). This includes studying the ordinary procedures (Asdal and Hobæk 2016) and “temporal logic” (Dányi 2020) of political institutions, or how “implicit or explicit procedural rules govern the organization of parliamentary time and space” (Barry 2020, 337). Many STS studies of regulatory procedures and processes have, however, focused on institutions at the national or regional levels, for example national parliaments and EU institutions, with little attention so far to UN agencies and other international organizations (for an exception see Riles's (2001, 2006) earlier work on UN conferences). This article aims to widen the discussion and improves our understanding of international organizations and their decision-making processes.

This article looks at the politics of data at the IMO, examining the IMO DCS as a new and evolving data infrastructure and its entanglements with the production of international regulations on shipping greenhouse gas emissions. Aalbu and Longva (2022) have analyzed how a three-step approach to decision-making at the IMO and the focus on collecting ever more data before making regulatory decisions has had a function of both progressing and delaying the regulatory process on greenhouse gases at different points in time. This paper expands on this topic and drills into the performative and temporal dimensions of the politics of data at the IMO, providing new insights into the relations between data infrastructures and global governance, and the present and future roles of data in international decision-making processes.

Building on ethnographies of infrastructures (Amelang and Bauer 2019; Anand 2017; Star 2002; 1999; Star and Ruhleder 1996), I follow the DCS in the making, in action, and in its remaking. I draw on observations from IMO meetings over several years, an analysis of IMO negotiation documents and meeting reports, as well as interviews with participants from IMO member states, observer organizations and the IMO Secretariat. ³ I use the term observant participation (McNeill and St Clair 2011; Seim 2024) to capture my role as an advisor in the Norwegian delegation to the IMO committee and working group meetings, participating as a researcher. Observing the negotiations through a member state delegation has given me unique access to the regulatory process, as well as invaluable insights into how the documents I analyze are produced and used in the negotiations. This positionality constrains both what I can observe, and what I can analyze and comment on. I have used document analysis as well as interviews and informal conversations with representatives from different member states and observer organizations to gather different views, as well as details from processes that I have not been able to observe. While I have access to all negotiation documents, I only cite publicly available documents and I also refrain from naming specific delegates and delegations unless I cite a publicly available document.

I first describe my analytical framing, and how insights from the literature on infrastructures, regulatory processes, and the politics of numbers and statistics are a useful point of departure for analyzing the dynamics at the IMO. Next, I describe the IMO as a data collector and consensus-based organization. I then dive into the politics of making the DCS from 2014 to 2016, and the many political compromises that led to a new data infrastructure. The article then examines the use of DCS data in the making of a new climate regulation in the period 2019 to 2022, and how the DCS emerges as an undisputed basis for regulatory development. I analyze how the DCS shapes the design and pacing of new climate regulation based on what data are available in the DCS and when they are available; and how IMO politics and regulatory processes simultaneously shape the making and remaking of the DCS, including how the data can be used. I unpack the politics that unfold when the production of data and regulations are coupled but are out of sync. Here, out of sync refers to both misalignment between data infrastructure design and regulatory use, and misalignment in regulatory timelines. As the DCS and regulatory development at the IMO become entangled over time, I argue this creates a form of governance-through-data which slows down the regulatory process but nonetheless supports consensus-building among parties at a time marked by deep political tensions.

Analyzing the Present and Future Roles of Data in International Negotiations

Literature in STS has shown how the making of data infrastructures requires myriad decisions, often as a result of intense negotiations between experts, with material consequences for someone or something (Bowker and Star 1999). In discussing biodiversity databases, Bowker (2000) sheds light on the consequences of design choices in the making of databases, arguing it is critical not to lose sight of the political, social, and scientific contexts in which databases are developed. Other studies have similarly emphasized that specific attention is needed to the data that are not produced, or “absent,” and “to the negotiations regarding what counts as data, for whom, when, where, why—and how this changes” over time (Leonelli, Rappert and Davies 2017, 195). Jackson (2014, 2015) has also highlighted how repairing an infrastructure is as fraught with politics as its making, turning our attention to the often taken-for-granted work that takes place to keep infrastructures alive.

This article builds on and contributes to this literature by examining the politics of the making of a data infrastructure in an explicitly political setting—international negotiations—and its entanglements with the making of regulations over time. The IMO is a particularly interesting case because the data infrastructure is designed through member state negotiations where there is a need to reach consensus, rather than by technical experts in a bureaucratic setting. The political context, to use Bowker's words, is particularly relevant here, as is also the regulatory context in which the data infrastructure is produced. I examine how the politics of climate change, and the mode of operation and temporal logic (Dányi 2020) of the IMO—core elements of the regulatory context—shape how, what and when data are made and used.

To examine the central role of data in the governance of emissions at the IMO, and its functions in the negotiations, I also build on literature on the politics of numbers and statistics, highlighting the role of numbers in providing trust and in depoliticizing public decision-making. Studies have shown how “governance through numbers” has been an intrinsic feature of modern democratic societies, and how statistics have functioned as an infrastructure for national governance (Desrosières 2002; Porter 2020; Rose 1991). Statistical tools and infrastructures have transformed modes of governance, as illustrated by Didier (2020) in his study of the introduction of representative sampling. These tools can also have direct political effects, for example by creating new categorizations (Desrosières 2002). Particularly relevant for my case is Porter's (2020) work on numbers as technologies of trust, where trust in numbers is produced through “mechanical objectivity,” an understanding of objectivity as adhering to strict rules where there is no room for human judgment. Numbers, Porter argues, are used to legitimize and defend public decisions where trust is lacking: “Quantification is a way of making decisions without seeming to decide” (Porter 2020, 8). Rottenburg and Merry (2015) elaborate on this perspective in a global governance setting, discussing how numbers and quantifications more broadly can be an effective means of translating the political into the “technical,” and can be used to depoliticize situations where there are doubts about other evidentiary practices. More recent literature on the politics of data links the current popularity of “evidence-based” and “data-driven” policymaking among both national and international policymakers to mechanical objectivity (Rieder and Simon 2016). Recent literature has also discussed how data can have a different political effect. Hoeyer (2019) has argued that data can take on a promissory function, promising to provide better insights for policymakers in the future, when more and “better” data are available. The collection of data in the present can legitimize the postponement of action to the future as policymakers argue for the need to wait for more data before making decisions.

I build on these insights to explore the politics of data at the IMO and the present and future roles of data in the negotiations. I describe how the making of the IMO DCS emerges not because of a lack of trust in existing evidentiary practices or from a need to defend decision-making to actors outside the IMO, but as a solution when member states cannot agree on how or when to regulate greenhouse gas emissions. I explore how the production of data is decoupled from the politicized negotiations on greenhouse gases to move the negotiations forward, and the tensions that arise when data and regulatory development are recoupled but are out of sync. I analyze how the misalignment between the DCS and regulatory use opens for a new politics, where the lack of the “right” data at the right time is used to argue for delaying action, as Hoeyer (2019) describes, but where hopes that future data will solve disagreements also push political decisions into the future. At the same time, the prevalent trust in quantified data supports making a delicate consensus in the present. I analyze how the DCS emerges as a trusted basis for decision-making among member states with diverging interests, and how delegations work to amend the DCS to get data production and the regulatory process back in sync.

The IMO as a Data Collector and Consensus-Based Organization

The IMO, headquartered on the embankment of the river Thames in London, is a United Nations specialized agency with the mission “to promote safe, secure, environmentally sound, efficient and sustainable shipping through cooperation.” ⁴ The Organization develops international conventions, regulations, and guidelines on these issues, and recommends implementation to its 174 member states. As member states are responsible for both implementing and enforcing international regulations, the IMO operates as a consensus-based organization to ensure a widely accepted and effective regulatory regime.

Climate change entered the IMO's scope of work in 1992, yet regulatory progress has been slow. The Organization has been the main governor of GHG emissions from international shipping since the 1997 Kyoto Protocol to the UN Framework Convention on Climate Change (UNFCCC) mandated developed countries to work through the IMO to limit or reduce shipping emissions. ⁵ In the same year, a Conference of Parties to the International Convention for the Prevention of Pollution from Ships (MARPOL Convention) adopted a new Annex VI on air pollution, and invited the IMO to consider strategies for emission reduction. ⁶ The IMO's Marine Environment Protection Committee (MEPC)—hereafter referred to as the Committee—has been discussing GHG emissions ever since. After more than a decade, the Committee adopted the first international regulations on GHG emissions from ships in 2011—the Energy Efficiency Design Index and the Ship Energy Efficiency Management Plan—using the language of energy efficiency rather than CO₂ or greenhouse gases to depoliticize the issue and move the focus to technical details during the negotiations.

The pace and form of the IMO's regulatory process is governed by detailed “rules of procedure” and specific “method of work” guidelines, giving the IMO a particular temporal logic. ⁷ The meetings of the Committee and its various sub-groups are attended by IMO member states, associated members of IMO, representatives from other United Nations bodies and specialized agencies and accredited observer organizations, such as intergovernmental and non-governmental organizations representing industry associations, environmental organizations, and professional associations. Member states and observers convene once or twice per year for week-long Committee meetings, with various sub-groups meeting between the Committee meetings. Thus far, new regulations on greenhouse gases have been adopted as amendments to the existing legal framework, MARPOL Annex VI. While amending the MARPOL convention is quicker than developing a new legal framework due to its tacit acceptance procedure, it is still a cumbersome process with set steps and time intervals. Amending MARPOL starts with a specific text proposal from member states or observers, which is then usually discussed by a working group before the Committee decides on further actions. Once a proposal is approved by the Committee after multiple reworkings, it is circulated among member states for adoption at the following Committee meeting six months or so later. If adopted, there is then a sixteen-month period before the amendments can enter into force. Moving through these set intervals takes years. The IMO's efforts to reach consensus often leads to prolonged negotiations until the Committee reaches a point where there are no more objections by member states.

The Politics of Making a New Data Collection System

In this section, I trace the development of the DCS from a simple proposal by the United States that the IMO should start collecting data from ships, through the negotiations to a stand-alone system. In 2012, the United States submitted a proposal to the Committee arguing that the IMO should develop operational energy efficiency standards for ships as a new regulation through a phased approach, with data collection as the first phase to support assessing which energy efficiency metrics were appropriate for different ship types. ⁸ At the time, the only established infrastructure that provided information about GHG emissions from shipping was so-called IMO GHG studies, ad hoc studies requested by member states before taking major decisions in the Committee. When the United States submitted this proposal, the IMO was in the trenches of difficult discussions on the next generation of emission reduction regulations. While member states recognized that the new 2011 energy efficiency regulations were not enough to adequately reduce emissions, there was little political will to do more. Member states from developed and developing countries disagreed fiercely on who should be responsible for emission reduction, with the IMO's principle of non-discriminatory regulation of all ships in international trade irrespective of flag conflicting with the “common but differentiated responsibilities” principle of the UNFCCC and the Kyoto Protocol, which limited binding emission reductions to developed countries.

While member states could not agree on what new emission regulations to develop or when to do so, they could agree to start collecting data from ships. The EU had already started developing its own monitoring, reporting and verification (MRV) system for ships, and had communicated that it would move ahead with regional emission regulations if the IMO did not progress with international regulations (European Commission 2013). After a series of Committee meetings, the initial US proposal was transformed into a “three-step approach” to decision-making (Aalbu and Longva 2022): “data collection, data analysis, followed by decision-making on what further measures, if any, are required.” ⁹ The work on developing new regulations to reduce GHG emissions was paused while the Committee put all its efforts into the first phase of the three-step approach, developing what was to become the IMO Data collection system for fuel oil consumption of ships.

No Agreement on What the Data Should be Used For

But what should the data be used for, and what type of data should be collected from ships? Despite disagreements on these core questions at the 2014 meeting, the Committee tasked a working group with considering “the development of a data collection system for fuel consumption of ships,” and set up a correspondence group with the same mandate, to work over email and report back to the next Committee meeting. ¹⁰ Over the next few meetings the Committee continued working on this initial foundation, slowly making a DCS for ship's fuel consumption, continuing to use politically neutral language in a similar vein as energy efficiency—rather than CO₂ or greenhouse gases—to separate the discussions on data collection from discussions on emission reduction regulations.

The question of what the data should be used for remained open throughout the following meetings, with several stated purposes floating in the discussions. Was the purpose of collecting data to inform the development of metrics for ships’ energy efficiency, as originally proposed; or to provide an overview of emissions to generally inform the IMO's regulatory work on GHG? The Committee had noted in 2014 that “the purpose of establishing a data collection system is to support possible, if agreed, further measures to enhance the energy efficiency of international shipping.” ¹¹ However, responses in the correspondence group that continued to work on the emerging DCS made it clear that delegations still had different perspectives. At the core of the conflict was whether the IMO should develop a new operational energy efficiency standard for ships to complement the Energy Efficiency Design Index from 2011. Several member states and especially industry associations were arguing strongly against this, for example as highlighted by one industry association:

the operational efficiency of a ship cannot be regulated in any meaningful manner, nor can it be assessed uniformly across ship type and size. ¹²

Other delegations argued for the importance of collecting data for developing appropriate efficiency metrics, highlighting the role of data in future decision-making at the IMO. ¹³ Member states and observers continued to express diverging purposes throughout the negotiations, as stated by a delegate in an interview:

I think at that point nobody really envisioned what the DCS would ever be used for, because nobody had ever collected this data….So people went in to discuss and negotiate something which they had no idea what it was going to be used for, and it was all a bit experimental….And I think you have to remember that most of the people at the IMO have no background in large-scale data analysis or data collection. And so, there is no conception of what amazing things you could actually do with properly collected and structured data. So it's just, yeah, we just collect some data, right, and just try to figure out, without even knowing what it is going to be used for, let's collect some data. It's just the worst way to start this. (Interview 9, Representative from observer organization)

This illustrates that early negotiations on the DCS were shaped not only by different interests on how to regulate ship emissions, but also by lack of experience with data collection in the IMO.

Negotiating What Data to Include

These disagreements on the purpose of data collection trickled down into the negotiations on the design of the emerging system. What would count as relevant data for the DCS (Leonelli, Rappert and Davies 2017), and who should have access? A particularly contentious issue was whether the system should include data on ships’ so-called transport work—a product of the cargo carried by a ship and the distance traveled. For example, an industry association argued in a submission to the Committee that including data on transport work “is a very complicated matter in view of the different characteristics of ship types, trade routes, cargoes and various other factors,” and warned that deliberations on this issue “might risk slowing down progress” on the development of the DCS. ¹⁴

Others argued that collecting data on transport work was essential for assessing ships’ energy efficiency. Canada, Denmark, Germany, Japan, Netherlands, Norway, the United Kingdom, and the United States highlighted for example the importance of collecting different kinds of data on transport work, including cargo data, to assess which metrics were appropriate for assessing the energy efficiency of different ship types—an issue that I will return to later. ¹⁵ The other sticking point was who should have access to the collected data. Some member states and industry associations argued for example that the data were commercially sensitive and feared that the data would be used to benchmark individual ships, while others argued for openness, highlighting the role of transparency in facilitating more efficient shipping. ¹⁶ At the following meeting in May 2015, delegations struggled to come to agreement on these issues and agreed to set up an intersessional meeting to hash out the remaining details. ¹⁷ After lying dormant for three years, the Committee's agenda point on greenhouse gases came to life again at the May meeting. The Marshall Islands, as part of the run-up to the Paris Agreement on climate change, proposed that the IMO should set a GHG reduction target for international shipping in line with the upcoming global climate goals. ¹⁸ Yet there was little support and member states agreed instead to prioritize finalizing the DCS to enable the Committee “to take an informed decision about next steps to be taken.” ¹⁹

At the intersessional meeting in September 2015, delegations managed to close the remaining questions on the DCS. Many delegations were still fiercely opposed to including cargo data and providing access to the data, and moving forward with the DCS was not possible without aligning with these interests. Delegations agreed to use ships’ cargo carrying capacity as a proxy for actual transported cargo in the calculation of transport work, and also agreed to enshrine the principle of data anonymization in the DCS. EU states cautioned that the DCS diverged from the new EU MRV system, which would include data on ship's transported cargo and be open to the public. A year later, the IMO Data collection system for fuel oil consumption of ships was finally adopted in October 2016 as an amendment to Annex VI of the MARPOL convention, firmly embedding the DCS in the IMO legal framework and regulatory process. From 2019, all ships of 5,000 gross tonnage and above engaged in international trade—representing around 85 percent of GHG emissions from international shipping—are required to submit data on annual consumption and CO₂ conversion factor per fuel type, distance traveled, hours underway, as well as data on ship particulars. While member states can access non-anonymized data for ships in their own fleet, they can only see anonymized data for other ships. The data are also rounded to make it more difficult to identify ships. ²⁰

The Data Collection System as a Political Compromise

The adopted DCS, a product of a series of consensus decisions in international negotiations over a four-year period, had traveled far from the initial proposal to collect data from ships. While the initial proposal from the United States to collect data was coupled with developing a new regulation for ships, the final DCS was decoupled from specific regulatory development and stripped down to a stand-alone system with a few data points and limited access to anonymized data. This negotiated DCS was explicitly recognized as a compromise by delegations, but still a step in the right direction for those member states that wanted a more comprehensive regulatory framework on GHG. Developing the DCS was not a response to a lack of trust in existing evidentiary practices at the IMO (Rottenburg and Merry 2015), but became a solution when nothing else could be done on greenhouse gases. Member states created a negotiation space that separated “data” from the politicized greenhouse gas discussions, allowing the Committee to move forward and keep the IMO's work on greenhouse gases alive (Aalbu and Longva 2022). Yet the political positions of delegations on how and when to regulate GHG emissions seared through the negotiations on the DCS and shaped the final system, highlighting the critical importance of studying both the political and regulatory contexts in which data infrastructures are produced. In the next section I describe the tensions that arise when the DCS and regulatory production are recoupled but are out of sync.

Out of Sync: the Data Collection System and the Making of Emission Regulations

While the IMO finalized the development of the DCS, the Paris Agreement was adopted in December 2015 and gave new momentum to the IMO's work on shipping emissions. In 2018 the IMO adopted its Initial strategy on reduction of GHG emissions from ships (Initial GHG Strategy), including the ambition to reduce the carbon intensity of shipping by at least 40 percent by 2030, and to halve total GHG emissions from the sector by 2050. ²¹ The first step in the strategy was to develop “short-term measures” to support the strategy's carbon intensity ambition. After intense negotiations, most of which took place remotely during the COVID-19 pandemic, in July 2021 the Committee adopted the first new regulations on ship GHG emissions in a decade, as an amendment to MARPOL Annex VI. This included an Energy Efficiency Existing Ship Index, similar to the Energy Efficiency Design Index for new ships from 2011, but for ships in operation. They also developed an operational carbon intensity indicator rating scheme for defined ship types (hereafter referred to as the rating scheme), which requires ships to calculate their annual carbon intensity indicator and assigns a rating to ships on a scale from A to E.

This new rating scheme picked up several elements that had been previously discussed by the Committee in relation to operational energy efficiency standards, but that had been paused during the development of the DCS. Making a carbon intensity standard for ships without the data that had originally been envisaged for this purpose (as outlined in the initial proposal by the United States) proved to be difficult. In what follows I describe how the DCS shaped the design and pacing of the rating scheme, slowing down the regulatory process as delegations argued for waiting for the right type and amount of data to be collected, while at the same time supporting the making of a delicate consensus in the present.

The Absent Data Reappear on the Agenda

When work commenced on the carbon intensity indicator rating scheme, it was the first time the DCS data were used in regulatory development. As stated by a representative of the IMO Secretariat,

With the CII we saw the first requests for use of DCS data, which was used to set baselines, and reduction factors etc. This was quite demanding, and really, I think people started to realize at that moment that it's a key element of the IMO GHG regulatory framework. (Interview 8, Representative of the IMO Secretariat)

The two most contentious issues during the making of the DCS—cargo data and data access—came to the fore during the development of the rating scheme. The regulatory text on the rating scheme in MARPOL Annex VI sets out the overall requirements for ships, but many of the details were specified in a set of supporting guidelines developed by a correspondence group led by technical experts after the regulation was approved. Guidelines included which carbon intensity metric to use and the carbon intensity reduction targets for ships. The first set of guidelines specified the metric itself—how should the carbon intensity of ships be calculated? The design of the DCS gave the group limited options.

The correspondence group considered several metrics, including the established Energy Efficiency Operational Indicator (EEOI), measured as CO₂ emitted per tonne of cargo transported, and the Annual Efficiency Ratio (AER), which is based on a similar formula but uses the ship's carrying capacity as a proxy for the actual cargo carried by ships. Delegates argued that it was important for the carbon intensity metric to be based on data submitted by ships to the DCS, referred to by delegates as “real” and “robust” data, as opposed to data produced by emission models as in the previous GHG studies. As stated in the report of the correspondence group:

To ensure the consistent implementation of the short-term measure, the Group recognized that only one single metric should be taken as the mandatory CII for each and every ship type and the mandatory CII should be calculated on the basis of the Data Collection System (DCS). In its most simple form, the attained annual operational CII of individual ships is calculated as the ratio of the total mass of CO₂ emitted to the total transport work undertaken in a given calendar year. However, since the data on cargo/passenger on board a ship is not covered by the IMO DCS, the deadweight tonnage (DWT) or gross tonnage (GT) has to be taken as a proxy. ²²

The correspondence group agreed that only the AER metric could be used to measure carbon intensity, as all the required variables are reported by ships to the DCS. The group also agreed to set 2019 as the baseline year for the rating scheme because it was the first year of DCS data, instead of using uncertain estimates for 2008, the baseline year for the ambitions in the Initial GHG Strategy. That the DCS should form the basis for the metric and the calculation of baselines was not questioned but taken for granted, even though the limited design of the DCS was not made for this tight coupling of data production and regulatory production. In the end, the use of the DCS data ended up being very similar to what had originally been proposed by the United States back in 2012.

When delegations turned to discuss the carbon intensity reduction factors—how much each ship needs to reduce each year—the metric choice came up for debate, once the consequences of selecting the AER metric became clear. While the IMO's Initial GHG Strategy set an ambition on carbon intensity, measured as CO₂ emissions per transport work, it did not define transport work and whether carbon intensity should be measured based on actual cargo transported (tonne-mile) as in the EEOI metric, or be based on a ship's carrying capacity (capacity-mile) as in the AER metric. The crux of the issue was that ships had historically made better progress on carbon intensity when measured on the EEOI metric—around 30 percent reduction in carbon intensity from 2008 to 2019—compared to the AER, with around 22 percent reduction in the same period (Faber et al. 2020). As an individual ship's progress on the Carbon Intensity Indicator in the rating scheme would be measured on the selected carbon intensity metric and would determine what grade the ship gets from A to E, this issue sparked intense negotiation. Several member states and observers, many of which had argued against including cargo data in the DCS, were now critical of the choice of carbon intensity metric, arguing that the EEOI metric better captured ships’ carbon intensity. Other member states who had been supportive of including cargo data in the DCS agreed that the EEOI was a better metric but pointed out that it would be impossible to use the EEOI in the rating scheme as the DCS did not include cargo data.

During the final negotiations on the rating scheme, member states accepted the simplified AER metric, but disagreed fiercely over which emission reduction factors should be required, and how and when the rating scheme should be enforced. How much should ships reduce annually and should ships with poor carbon intensity ratings lose their seaworthiness certificates? Having only one year of DCS data to work with (from 2019), made several delegations challenge the basis on which the regulation was developed—arguing both for delaying the implementation of the regulation and weakening its enforcement regime while waiting for more data. Many member states argued that the DCS had not collected sufficient data yet to analyze how the rating scheme would affect ships and whether it was an appropriate way to regulate ship operations, arguing strongly against penalizing underperforming ships. Others argued that the reduction factors were not ambitious enough and that the rating scheme would remain toothless without proper enforcement powers. To solve this disagreement and to move forward, member states added a review clause to the regulatory text; the rating scheme would enter into force in 2023 with “soft enforcement,” before a review of the regulation by 1 January 2026, to evaluate whether there is a need for more stringent reduction targets and enforcement—based on analysis of DCS data from the first three years after the regulation begins operation. This review clause was critical for the adoption of the CII rating scheme, pushing matters that member states could not agree to when more data from the DCS are available.

This misalignment between the design of the DCS and its expanding regulatory role sheds light on the consequences of the many political compromises in the making of this data infrastructure. While member states had agreed to use data from the DCS to inform regulatory development at the IMO, the absence of cargo data and the lack of access to the data challenge the tight coupling of data and regulatory development. The political disagreements during the making of the DCS have not only limited how emissions are governed, but also when. The misalignment between the DCS and its regulatory use opens up a new politics during the negotiations. As Hoeyer (2019) has argued, the promise of future data can legitimize the postponement of action in the present as decision-makers wait for more data. Yet in this case we also see something different. The lack of the “right” data at the right time is used to legitimize delaying the rating scheme, similar to what is suggested by Hoeyer, but hopes that future data will solve political disagreements on the rating scheme also postponed a decision on potentially more stringent enforcement of the rating scheme from 2023 to 2026. The review clause was acceptable to both opponents and proponents of more stringent regulatory action. Proponents were in favor of the clause because, perhaps unintuitively, they hoped that more data would support more stringent enforcement measures being agreed to in the future than what would have been possible in the present. At the same time, the prevalent trust in the data already produced by the DCS supported reaching conclusions during the final negotiation meeting and the adoption of the rating scheme. The DCS was needed for both closing the debates on CII reduction factors and the technical details on CII metrics and reference lines.

While a delicate consensus was reached, few delegations were content with the regulatory outcome. Many member states and industry associations were unhappy with a carbon intensity metric that did not represent their ships’ operations. Others argued that the regulation would have little effect on reducing the carbon intensity of shipping and were discontented with the regulation's delayed implementation and weak enforcement. However, the IMO did manage to adopt the first new regulations on ship GHG emissions in a decade. This was a critical outcome for both sides in the negotiations to demonstrate that the IMO is capable of governing greenhouse gas emissions from international shipping.

Attempts to Get in Sync: the Politics of Amending the DCS and the Rating Scheme in Parallel

While the Carbon Intensity Indicator rating scheme relies on DCS data both for its design and pacing, the full dataset was not available to member states and observers during the development of the rating scheme. As noted by the Royal Institution of Naval Architects in a submission to the Committee,

while the design of the Fuel Oil Consumption database predates the CII discussion (and its current use was perhaps not envisaged), it is clear that in the current development of CIIs, the anonymized and restricted nature of the database has hampered transparent analysis and discussions between stakeholders on a very complex topic. ²³

The limitations of the DCS do not, however, seem to affect how delegates perceive the quality and legitimacy of the data, or the hopes invested in the data. Nor has it raised questions on whether the DCS should indeed provide the evidence base for future new regulations. During meeting discussions and interviews, the DCS is presented by delegations as an undisputed basis for regulatory development at the IMO and a critical part of the Organization's “evidence-based decision-making.” Even though the DCS is recognized by delegations as a political compromise, the data produced by the DCS are presented as neutral and depoliticized. This prevalent trust in the quantified DCS data is produced through something akin to mechanical objectivity (Porter 2020), namely the procedures for collecting and verifying data from ships and the management of the database, which are considered robust enough for how the data are used at the IMO.

A representative of the IMO Secretariat specifically highlighted the importance of trusted data for consensus making, arguing that data are

a prerequisite for developing robust regulations, especially to get the consensus based on facts from all the parties involved in the negotiations. So that there is no discussion on, you know, the data quality, data sources, etc. (Interview 8, Representative of the IMO Secretariat)

This illustrates the value of trusted data in supporting consensus in international negotiations even in light of—or perhaps specifically in situations of—political disagreements. The role of data as a trusted decision-basis in supporting consensus is important for the drive to continuously amend the DCS. But as Jackson (2015) reminds us, repairing an infrastructure is as fraught with politics as its making, and as we see here, this is particularly the case in the context of amending an infrastructure through international negotiations. The experience of using DCS data in the development of the rating scheme prompted some member states to propose that the IMO should amend the DCS to match regulatory needs—to bring the DCS and regulatory production in sync. In a document submitted to the Committee meeting in November 2021, after the adoption of the rating scheme, a large group of member states including EU countries, Norway, and the United States argued that the “DCS should remain the main source of information on ship's energy efficiency and carbon intensity performance,” and that the IMO should launch a work stream to amend the DCS to make sure the system “remains fit for this purpose.” ²⁴ Yet what is “fit for purpose” is a political question and depends on who you ask at the IMO and when. With this proposal, the member states reopened discussions that were closed when the DCS was adopted in 2016, arguing for including cargo data and improving data access. Due to time constraints, this proposal was pushed to a working group meeting in March 2022, where discussions made it clear that delegations still had different positions on the DCS.

Some member states and industry associations had, however, during this same period become open to collecting more data in the DCS and improving access if this supported the use of new carbon intensity metrics in the rating scheme to better reflect ships’ operational efficiency. What count as relevant data in the DCS have therefore changed over time for many delegations, based on experiences with how the data have been used. At the July 2023 Committee meeting, there was finally political agreement to do a more comprehensive revision of the DCS, and several delegations highlighted that the system was “at the cornerstone of GHG measures and a crucial element for the Committee's decision-making.” ²⁵ A working group during the meeting revised the regulatory text, and delegations finally agreed to include data on transport work, allowing the potential use of new indicators in the rating scheme. However, the issue of access to data remained unresolved.

The temporal logic of the IMO regulatory process makes it impossible to complete the DCS revision and collect data in time for the review of the Carbon Intensity Indicator rating scheme by 1 January 2026. The attempt to bring the DCS and the rating scheme in sync by collecting the data needed for an improved rating scheme is challenged by another misalignment in regulatory timelines. What is peculiar in this case is how the data infrastructure itself is adopted as a regulation and adding just simple amendments set in motion all the steps in the IMO regulatory process. Revising both the DCS and the rating scheme requires amendments to the MARPOL convention, following the IMO regulatory process with its set steps of approval, adoption and entry into force: The DCS amendments approved in July 2023 were adopted in March 2024, and then there will be a 16-month period before the amendments enter into force. This means 2026 will be the first year when data on transport work (and other new parameters) from ships will be collected, with data not available for analysis before mid-2027. Yet the rating scheme review process needs to be finalized by the Committee by mid-2025. When the same working group that undertook the revisions to the DCS also developed a plan to review the rating scheme, the misaligned timelines became an explicit concern. The review of the rating scheme will involve a “data-gathering stage” until early 2024, however this does not match up with the timeline of the DCS. Several delegations—who had earlier fought adamantly against including transport work—were now concerned that the new data on transport work would not be available in time for the review.

It remains to be seen how the politics of data will play out in the review of the rating scheme without the “right” data being available at the right time to support closing debates. This constant amendment process driven by political disagreements creates a continuously evolving DCS, which in turn shapes and reshapes the political process. However, there will likely always be a misalignment between the DCS and regulatory needs at a particular time—both in terms of content and timing—and something that policymakers will have to grapple with in practice. What is fit for purpose changes continuously over time as political interests change, and the regulatory process continues whether data is available or not.

Conclusion

This article explored the entanglements of the production of data and regulations on greenhouse gas emissions at a specialized UN agency, showing the importance of data infrastructures for global governance. While few had imagined the potential of the emerging DCS when IMO member states first agreed to start collecting data from ships, this new data infrastructure has now taken center stage in the IMO's regulatory framework on greenhouse gas emissions in international shipping.

This case adds to the existing literature in STS on data infrastructures and governance by showing how both the political and regulatory contexts in which data infrastructures are produced and used matter for global governance. Both the politics of climate change at the IMO (the political context), and the IMO's mode of operation and temporal logic (the regulatory context) shape the production and use of data—and how and when emissions from ships are governed. This case illustrates how deeply intertwined data infrastructures and regulatory processes are. A constant regulatory amendment process creates a DCS in motion, which in turn shapes and reshapes how climate change is governed.

The analysis of the dynamics at the IMO also provides specific insights into the present and future roles of data in international decision-making processes. I have described how the DCS and regulatory development are coupled based on a prevalent trust in the data, and the tensions that arise when they are out of sync—a misalignment between DCS design and regulatory use, and a misalignment in timelines. I argue the misalignment between data infrastructure design and regulatory use opens up for a new politics, where the lack of the right data at the right time is not only used to argue for delaying action, similar to what is suggested by Hoeyer's (2019) “promissory data,” but where hopes that future data will solve political disagreements also strategically pushes decisions to the future. This case highlights how being out of sync also has a productive dimension that has so far not been discussed in the literature. In situations of political disagreement, waiting for more data can also be used to push difficult decisions forward with the hope that this will lead to better regulatory outcomes in the long run. At the same time, prevalent trust in the DCS and the data it already has produced supports the making of a delicate consensus in the present. The DCS has emerged as the gold standard for data on ship emissions, creating a trusted decision-basis for IMO member states that supports consensus-making at a time in the negotiations marked by deep political tensions. Having a trusted knowledge base to support consensus-making is particularly critical at an international regulator like IMO, as the actual implementation and enforcement of the international regulations it develops are the responsibility of individual states.

This case suggests that being out of sync may indeed be an intrinsic feature in the relations between data infrastructures and regulatory processes, not just in the IMO but in other international organizations that operate based on formal negotiations. Attempts to bring the DCS and regulatory use back in sync meets another misalignment—this time in regulatory timelines. While a deeper revision of the DCS was approved in July 2023, the timeline of the IMO regulatory process makes it impossible for new data points to be available in time for the review of the rating scheme. What count as relevant data changes continuously over time as political interests and regulatory needs change.