Expanding the politics of measurement in sustainable finance: Reconceptualizing environmental,social and governance information as infrastructure

Definition

While regulators, industry groups, and standard setters have tried to define sustainable finance, a universal definition remains elusive. This article leans towards a broad definition, encompassing historically prevalent practices like SRI and specific subfields like climate finance. I thus define sustainable finance as integrating ESG information into investment processes. This approach aligns with definitions from regulatory, industry bodies (EC, 2021; GSIA, 2020: 7), and academic contributions (Busch et al., 2016: 35; Curtis et al., 2021: 395)

The above stated definition of sustainable finance raises the follow-up question what “ESG information” refers to. As a working definition, one might think of ESG information as an interpretation of the ESG performance ¹ of a financial entity like a corporation or, by extension, a financial product like a corporate bond (cf. Berg et al., 2022: 5). Drawing from Parfitt’s discussion on “ethical capital” (2022:1), it can also be viewed as a method to incorporate sustainability and other non-financial factors into financial assets. Regardless of the approach, ESG information remains a crucial pre-condition for sustainable finance as claims about the relations between sustainability and financial concerns ultimately depend on the availability and reliability of such data (Konstantinos and Serafeim 2019: 50; Young-Ferris and Roberts 2023).

Beyond this working definition, in the context of this article ESG information is understood and problematized as an emerging and expanding infrastructure (see next section). As such, it comprises multiple classifications, data sources, indicators and methodologies regarding the attribution of ESG credentials to financial categories. Leaning on the work of STS scholars on these type of “objects” (e.g. Callon and Muniesa, 2005), in the following I use the terminology of “technical devices” as an umbrella category.

Sustainable finance and environmental, social and governance information from the macro-political perspective

Macro-political treatments on sustainable finance have come from several disciplines – including Finance, Management Studies, Law, and International Political Economy (IPE) – and broadly fall into two categories. One strand presents an evolutionary perspective, identifying various stages, key events, players, and regions (Busch et al., 2021; Dumas and Louche, 2016; Crifo et al., 2019; Dimmelmeier, 2021; Pollman, 2022; Robins and McDaniels, 2016; Schoenmaker, 2017; Sparkes, 2002). Another set of works sees sustainable finance’s rise as a manifestation of broader socio-economic trends, such as financialization or the growth of private environmental governance (Braun, 2022; Fichtner et al., 2023; Gabor, 2021; Paterson and Thistlethwaite, 2016; Pattberg, 2012; Smoleńska and van ’t Klooster, 2022).

Many evolutionary accounts offer a stage-based typology of the development of sustainable finance. While terms like “Sustainable Finance 1.0” to “Sustainable Finance 3.0” (Busch et al., 2021; Schoenmaker, 2017) have been used to describe this progression, it is widely recognized that the field emerged from the niche of religiously driven ethical investment. Often termed SRI, this ethical investment movement emerged in the 1960s and 1970s and was influenced by divestment campaigns against firms linked to the Vietnam War and South African apartheid (Busch et al., 2021: 3 ff, Sparkes, 2002: 49ff, Dumas and Louche, 2016: 439). In the next phase, institutional entrepreneurs like the UN Principles for Responsible Investment (UN PRI), UN Global Compact (GC), and others “mainstreamed” ethical concerns among major financial institutions by framing them as financial risks (Busch et al., 2021; 4; Dimmelmeier, 2021: 1611; Sparkes, 2002: 60ff; Pollman, 2022: 9ff). This shift also introduced the “ESG” acronym, first used in the UN GC’s influential 2004 report “Who Cares Wins” (Busch et al., 2021: 4; Pollman, 2022: 11).

While there is less consensus on the current and ongoing phase of sustainable finance, some trends can be identified. First, as sustainable finance expanded and solidified, standards and regulations emerged to counter ambiguity and misleading claims (Crifo et al., 2019: 134; Dimmelmeier, 2021: 1615ff). Second, in the aftermath of the 2015 Paris Agreement, there has been a notable shift in emphasis towards climate and carbon-related risks (Busch et al., 2021: 2; Robins and McDaniels, 2016: 13, Paterson and Thistlethwaite, 2016: 1198ff)

In summary, the contributions discussed portray an emerging and gradually maturing industry. Its evolution was shaped by institutional entrepreneurs, including investor coalitions and International Organizations from the UN system, and by historical events like the Vietnam War and the signing of the Paris Agreement. This literature underscores the progression from SRI to ESG, leading to contemporary debates on standardization and climate-related risk management While the literature valuably highlights actors, events, and discourses, it sometimes overlooks the significance of the technical devices that uphold sustainable finance in everyday practice and the actors that have developed them (but see Sparkes, 2002).

Contrasting the previously discussed works that primarily examine internal dynamics, another strand of literature delves into sustainable finance’s ties with broader macro-political and macro-financial trends. These often-critical perspectives focus on who governs and controls sustainable finance and the motivations of major financial institutions in participating in it. Regarding the question of control, critical accounts have tended to focus on the public versus private dichotomy. Gabor’s recent work on the financialization of development policies through the “Wall Street Consensus” emphasizes that investors strive to become “epistemic guardians of green taxonomies” (Gabor, 2021: 438), thereby “control [ing] the grammar of green finance” (ibid). In this account, a “private ESG regime” (Gabor, 2021: 439, my emphasis) is juxtaposed to “a public taxonomy” (ibid), the latter of which would benefit sustainable development outcomes but hurt the financial performance of investors. A recent analysis of European banking regulation also engaged with the debate over public versus private control. Accordingly, regulators’ dependence on climate risk assessments from financial institutions and other private data sources would prompt a “deferential transition”, intensifying greenwashing and triggering a race to the bottom (Smoleńska and van’t Klooster, 2022).

While the aforementioned contributions situate control over sustainable finance with investors and banks, other studies have focused on the power of private data intermediaries like the index providers (Fichtner et al., 2023) and private standard setters in sustainability accounting and reporting (Paterson and Thistlethwaite, 2016; Pattberg, 2012). These studies underscore the role of technically oriented actors such as the Global Reporting Initiative (GRI), the GHG Protocol and the CDP (formerly Carbon Disclosure Project), emphasizing how they make sustainable finance “governable”. In recounting the history of the relations between key players in these technical spaces, the aforementioned contributions have also problematized the public versus private dichotomy and argued that private governance leads to more nuanced outcomes than more general claims informed by abstract concepts like neoliberalism might suggest (Paterson and Thistlethwaite, 2016: 1215). This notwithstanding, by taking private governance as their main analytical framework, these contributions tend to focus on actors, that is influential organizations or individuals, rather than on the technical devices themselves.

Turning to the “why question”, that is asking what motivates financial institutions to engage in sustainable finance, two broader historical developments have been identified as potential drivers. For macro-financial analyses operating broadly within the financialization literature, the increasing importance of institutional investors and asset managers vis à vis banks after the financial crisis of the late 2000s is an important factor. In this perspective, institutional investors and asset managers push for sustainable finance to pressure states and other public institutions to generate and subsidize green investable assets – especially in the Global South – that can absorb investors’ capital (Dafermos et al., 2021; Gabor, 2021: 437, 443ff). A related point argues that asset managers employ the ambiguity of sustainable finance as a strategy to navigate geographically distinct political and regulatory demands concerning their engagement with social and environmental issues (Braun, 2022: 644).

Conversely, literature emphasizing private environmental governance associates the growing traction of sustainable finance with the ascent of non-state actors in global climate politics, particularly after the perceived inadequacies of interstate negotiations within the UNFCCC framework (Pattberg, 2012: 613). Here, sustainable finance is not so much seen as resulting from the strategic ambitions of specific actor groups, but rather as a manifestation of the increasing decentralization of global climate and environmental governance.

In summary, macro-political perspectives depict sustainable finance either as an outcome of financialization, with investors gaining increased control over climate and environmental policy, or as part of a rising regime of private environmental governance. While these views offer a valuable complement to largely endogenous evolutionary narratives, they risk oversimplifying the complexities that have accompanied sustainable finance’s history. This is particularly evident when examining the public versus private contrast, often used as a heuristic to understand sustainable finance’s (lack of) environmental and social outcomes. Moreover, even though some of the aforementioned contributions engage with and problematize technical devices like accounting standards or ESG ratings, they still remain relatively actor-centric. By emphasizing control over or participation in the creation and application of a technical device as power indicators, these studies may overlook the intrinsic politics and nuances of the devices themselves.

Sustainable finance and environmental, social and governance information from the micro-political perspective

Macro-political accounts thus treat technical devices mostly as a secondary concern. Micro-political accounts, by contrast, examine the creation and “life” of these devices in great detail. In discussing these micro-level accounts, I do not present a systematic and comprehensive literature review but instead aim for a representation of the different aspects (asset classes, financial instruments, standards) that are subsumed under the label sustainable finance (cf. Busch et al., 2021: 7 for a typology).

This search yielded contributions dealing with green bonds (Christophers et al., 2020; Langley et al., 2021; Triphathy, 2017), “carbon finance” including high and low carbon equities (Langley et al., 2021; Van Veelen et al., 2020), ESG rating agencies and their products (Beunza and Ferraro 2019; Déjean et al., 2004; Eccles et al., 2020; Giamporcaro et al., 2016), ESG integration into traditional (equity) investing (Parfitt, 2020; Young-Ferris and Roberts, 2023), impact investments (Barman, 2015), and corporate sustainability disclosures and the standards governing them (Barman, 2018; Parfitt, 2022; Talbot and Boiral, 2018).

While originating from diverse disciplines—evident from publication journals, such as organization studies, financial geography, anthropology, business studies, economic sociology, and accounting—the cited articles have overlapping methodological and theoretical foundations. Notably, 10 out of the 14 articles reference the works of the French sociologist Michel Callon, who has pioneered STS-inspired analyses on the constitution of markets through technical devices (e.g. Callon and Muniesa, 2005). ²

The STS emphasis on how technical devices mediate relations among different actors, ideas and values in complex, non-linear ways is also visible in the theoretical concepts and findings from these studies. Some articles employ the STS “translation” concept, demonstrating how tools like ESG ratings or green bond standards become spaces for diverse actors to negotiate the meaning and measurement of terms such as “value”, “risk”, or “sustainability” (Beunza and Ferraro, 2019; Tripathy, 2017; Van Veelen et al., 2020). Other contributions highlight processes of “valuation work”, where technical devices such as ratings and classification systems aim to assign a financial or non-financial value to a practice or entity (Barman, 2015) and “assetization” processes, where technical classifications and measurements of sustainability are mobilized to generate new categories of income yielding financial securities (Langley et al., 2021; Van Veelen et al., 2020).

These studies consistently highlight that translation, valuation and assetization processes are complex, contingent, and contested. Attempts by financial actors to bring ESG measurements in line with traditional financial metrics and valuation models are, therefore, often frustrated (Beunza and Ferraro, 2019; Young-Ferris and Roberts, 2023). Furthermore, different actor coalitions offer competing systems for the classification of sustainable financial assets (Langley et al., 2021) and multiple incommensurable systems for measuring financial and “impact”-value continue to exist side by side (Barman, 2015).

These insights provide a more detailed perspective compared to macro-level narratives linking sustainable finance to financialization and the rise of private governance. Contrary to broad assertions, micro-political studies emphasize that integrating financial tools with sustainability, using technical devices, is not merely a function of structural power. Instead, it involves intricate, sometimes unsuccessful negotiations between various actors. Additionally, even if a technical device is made to “work”, it is never stable or isolated from contestation.

Importantly, this nuance does not mean that technical devices cannot lead to clear distributional consequences and power shifts. Among the findings of the above presented studies are, for example, that technical devices connected to ESG integration articulate “risks” in a way that benefits risk trading financial intermediaries while creating disadvantages for at-risk populations (Parfitt, 2020: 583). Regarding ESG ratings, Young-Ferris and Roberts (2023) and Eccles et al. (2020) note that traditional financial accounting, focused on a narrow view of financial value, has crowded out alternative methodologies. Therefore, the nuanced perspective from micro-political studies is not that financialization is not occurring, but that its presence should be substantiated by closely examining the technical devices in question.

One of the most notable strengths of micro-political accounts is their meticulous attention to the intricacies of specific technical devices. However, this strength can also be viewed as a limitation. By honing in too narrowly, these accounts may overlook the interconnected nature of broader phenomena. For instance, Barman’s investigation of valuation devices targets exclusively impact investing, which it delineates from related practices like SRI, Responsible investment or sustainable investment (Barman, 2015: 15–16). Yet, as recently observed by Chiapello (2023: 15) and Busch et al. (2021) in practice the definitional boundaries between these categories are blurry as SRI or ESG investing also routinely use the impact label. Similarly, Parfitt (2020) and Young-Ferris and Roberts (2023) limit their investigation to ESG integration, which the GSIA (2020:7) defines as the integration of ESG factors into investment analysis. However, even though there are clear definitional distinctions between ESG integration and other sustainable finance strategies, like negative screens or thematic investments, most investors often blend them in practice. The extent of this blending is, for example, shown in recent market study of Germany’s sustainable finance landscape (FNG, 2023: 21)

The point of highlighting the connections between different sub-fields of sustainable finance is not to criticize the case selection of the mentioned articles. Instead, by highlighting these connections, I want to emphasize that STS-inspired research tends to focus on making more general claims by associating case study data with theoretical frameworks and structural contexts, whereas the aggregation of data across studies is uncommon (Vollmer et al., 2009).

Reconceptualizing environmental, social and governance information as emerging and expanding infrastructure

Summarizing, both macro- and micro-political viewpoints on sustainable finance offer unique insights but also have their limitations. Macro-political accounts adeptly trace the historical ties between diverse actors over time, illustrating how sustainable finance has evolved through various organizations and broader financial and political context changes. However, these accounts sometimes overlook the pivotal and intricate role of technical devices in making sustainable finance governable. As a result, they may either neglect the significance of these technical tools or simplify their importance, tying them to abstract notions like “control” or the public-private distinction.

Micro-political perspectives, by contrast, have engaged deeply and cogently with the complexities, contingencies and effects of technical devices, thereby nuancing reductionist explanations and showcasing potential intervention points for change. However, by zeroing in on specific cases and technical devices, they risk overlooking how broader trends in sustainable finance influence and interconnect over time and geography. Solely adopting a micro-political viewpoint might, contrary to observed industry practices and thus inaccurately, suggest that “sustainable finance” does not exist as a cohesive phenomenon but is just a collection of largely autonomous areas like ESG integration or impact investing.

One question that arises from the discussion of both perspectives is whether their strengths could be integrated to overcome their respective blind spots. In other words, would it be possible to narrate a history of sustainable finance as an evolving and connected system, but in such a way that the connections are not conceptualized between actors and discourses but, instead, between technical devices? In the reminder of this article, I will argue that such a re-narration is possible and, perhaps more importantly, offers complementary theoretical and political insights.

While some recent studies have already begun to map out the interconnectedness of various technical devices in sustainable finance in terms of their functional relationships (e.g. Condon 2023: 13ff; Fichtner et al., 2023: 5), I emphasize how these linkages developed historically. To denote this aspect, I conceptualize the aggregation of sustainable finance-related technical devices and their interactions as an emerging and expanding “ESG information infrastructure”.

The notion of infrastructure has been developed and used in STS to denote large, central but also oftentimes obscure socio-technical systems that enable the day-to-day operation and governance of specific domains (Star, 1999: 280ff). While historically the dynamics of physical objects like rail networks or power grids have been studied as infrastructures, the concept has also been applied to less “material” relations such as standards, classification frameworks or routinized ways regarding the production and use of information (ibid, see also Bernards and Campbell-Verduyn, 2019: 777).

Importantly, the infrastructure concept emphasizes the aggregation or “bundling together” of multiple technical devices into a larger system (Bernards and Campbell-Verduyn, 2019: 780). It not only explores the relationships between a device and its designers, users and otherwise affected parties, but also examines how new devices connect to pre-existing ones. In this context, STS scholar Susan Leigh Star introduced the “installed base” (Star, 1999: 822) concept, underscoring that infrastructure changes are layered and often inherit characteristics and limitations of previous systems.

In summary, for the purposes of this article two aspects of the infrastructure concept are particularly relevant. First, infrastructures are large, relational systems, in which different technical devices, and the parties related to them, interact. Second, infrastructures undergo gradual evolution, requiring new technical devices to either challenge or build upon the installed base. Moreover, conceptualizing infrastructures as slowly evolving systems means that the relations between technical devices emerge from historical processes rather than just from functional interactions.

Data and methods

Infrastructures are most effectively examined through a historical and qualitative lens (McCarthy, 2017: 12). I employ a genealogical approach in the subsequent sections, to re-narrate the evolution of sustainable finance through the infrastructure perspective. Genealogical methods, inspired by Foucault, unravel the contested and fluid histories behind seemingly solid concepts, offering a deeper understanding of their development (as discussed in the context of finance and accounting by Edgley, 2014: 256ff).

To reconstruct the history of sustainable finance through the lens of the ESG information infrastructure, I draw on secondary literature and textual information published by actors in the sustainable finance field. Additionally, I integrate original interview and participant observation materials gathered during a 2017–2019 study on the emergence and governance of sustainable finance. This research employed boundary specification methods from social network analysis (Knoke, 1993) to examine and group 145 organizations involved in sustainable finance from 1998 to 2018, based on their publication collaborations and citation patterns.

Using network analysis, I clustered different communities within sustainable finance to identify possible interviewees. This process resulted in 25 semi-structured elite interviews with representatives of different communities. I conducted interviews in person in Berlin, Brussels and Geneva, as well as over the phone. Interviewees came from various backgrounds, including International Organizations (IOs), central banks, regulatory agencies, asset management, sustainable finance trade associations, civil society and policymaking. The interview length varied between 20 min and one and a half hours. ³ Interviewees were asked a set of questions that was tailored to their position in sustainable finance. Nonetheless, a range of common questions querying inter alia their perception of the history of sustainable finance, the most important actors, geographical differences, the role of regulators and legislation and the importance of different ideas and frames were included in all interviews. During and after the interview process, statements were put into deductive as well as inductively emerging categories. This article specifically references categories tied to perceptions of the history of sustainable finance and ESG.

I also engaged in participant observation at 18 sustainable finance events from April 2017 to October 2019. These events, spanning the locations of Berlin, Brussels, Oxford, Paris, and Zurich, and online platforms, were hosted by EU institutions, IOs, universities, think tanks, industry groups, and civil society organizations. Event transcripts were categorized analogously to interview responses.

Period 1: Socially responsible investment

SRI started to grow in the 1970s and 1980s in the Anglo-American context. At the time, high-profile public campaigns targeted companies making profits from the Vietnam war, such as Dow Chemical, which produced the highly toxic herbicide “Agent Orange”, or from doing business with the South African Apartheid regime. These campaigns sensitized the public about the links between investments and ethics (Sparkes, 2002: 46ff). The preoccupation with avoiding investments in armaments led to the creation of pioneering SRI funds, which often came from a Christian religious background, such as the PAX fund, which two Methodist ministers launched in the US in 1971.

PAX and similar funds required information on the performance of companies concerning armaments production and other areas of other concern like gambling, alcohol and tobacco (Co-founder, Sustainability Consultancy, October 2018, Staffer, UNEP FI, July 2018). They repurposed research from NGO advocacy and consumer information campaigns to acquire this information. In the US, the Council on Economic Priorities, which had profiled US corporates’ exposure to the Vietnam war to inform consumer choices, became a reference point (Sparkes, 2002: 281). In addition, Kinder, Lydenberg, Domini and Co (KLD) started providing dedicated SRI information from 1988 onwards. In the UK, investors from the Methodist church and the Quakers built on NGO work that identified corporate exposures to arms trade to create their own ethical research service EIRIS in 1983 (Eccles and Stroehle, 2018: 10). In France, trade unions were closely involved in the process of SRI. This led to a greater emphasis on “social” issues like employment practices, while the “societal” implications of unethical products like tobacco were less of a priority (Co-founder, Sustainability Consultancy; European Commission-EC Official, March 2019). To cater to those needs, ARESE, which analyzed company practices and policies, was established in 1988. In Germany, corporate environmental performance was prioritized by Socially Responsible investors, which informed the focus of the 1993 created agency oekom research (Co-founder, Sustainability Consultancy).

To operationalize the values-investment alignment, Socially Responsible investors developed two approaches. The first consisted in punishing bad corporate performance through not investing (negative screening), while the second rewarded good performance through targeted investment (positive screening). Research agencies, meanwhile, developed technical devices to measure harmful outputs like armament production and unethical practices like bad labour relations. The measurement of harmful outputs was operationalized by setting thresholds (e.g. maximum 5% of turnover from arms production, Sparkes, 2002: 283), while practices and policies were evaluated through scoring systems (Déjean et al., 2004: 751ff).

Despite differences in scope and priorities of European and Anglo-American investors, the technical devices of this emerging SRI-focused infrastructure were relatively similar. Human analysts compiled information on companies from publicly available sources such as newspapers, government statistics, industry research and NGO reports to create scoring systems or exclusion lists. In addition, the newly founded research agencies consulted annual company reports, and Corporate Social Responsibility (CSR) reports. Finally, some agencies also sent dedicated questionnaires to companies and provided them with the opportunity to review ratings and fill in missing data (see Eccles and Stroehle, 2018: 5; Escrig-Olmedo et al., 2010; Hughes et al., 2021: 3; Sparkes, 2002: 280ff).

In the late 1990s, the comprehensive adoption of SRI remained a niche activity (Staffer, UNEP FI; Senior Researcher, UNEP Inquiry; see also McDaniels and Robins, 2016: 12). However, the attention of mainstream investors to CSR topics increased significantly due to a series of corporate scandals. The bankruptcy of the US energy utility Enron that followed a major accounting fraud, the deathly leakage of toxins at US multinational Union Carbide’s Bhopal chemical plant and the wreckage and subsequent oil spillage of the tanker Exxon Valdez respectively exposed the relevance of Governance (Enron), worker protection (Bhopal) and environmental (Exxon Valdez) criteria (Dumas and Louche, 2016; Sparkes, 2002; Pattberg, 2012). Other episodes that accentuated the relevance of the governance dimension were accounting frauds at the companies WorldCom, Tyco, Parmalat and Vivendi (Barron et al., 2006: 8; NGO representative, UK, December 2018). These scandals showed that information about companies’ environmental, social and governance performance was related to economic and financial losses.

Institutional entrepreneurs like UNEP FI saw this context as an opportunity to connect the heightened sensitivity of the financial community to corporate scandals with the infrastructure that had been built to operationalize SRI. On the discursive level, this expansion occurred through a change in language, where the label SRI, which had become associated with idiosyncratic values and negative screening, was replaced by the term “responsible investing” (Staffer, UNEP FI). Responsible investing differed from SRI and CSR by emphasizing financial returns through better risk management rather than ethics or good corporate citizenship (Barman, 2018: 292ff). As noted by one interviewee, in the late 1990s and early 2000s, the discussion was based on “the assumption that companies, which score better on social and environmental issues would also perform better financially” (EC Official; see also Staffer, UNEP FI; Senior Researcher, UNEP Inquiry; Barron et al., 2006).

However, while the discourse changed from SRI to Responsible Investing, the various technical devices for measuring ethical performance, that is the ESG information infrastructure, remained at first unaltered. With the entry of new actors and their associated demands and discourses, the infrastructure experienced, however, dynamics of change and contestation that the next section explores.

Period 2: Environmental, social and governance investing

The insertion of mainstream investors into the ESG information infrastructure was strongly linked to the concepts of “fiduciary duty” and “financial materiality”. Fiduciary duty is a legal concept that applies to institutional investors but not to smaller investors like church funds. A narrow interpretation of this concept argues that institutional investors must only pay attention to information that maximizes short-term shareholder value. However, in the 2000s, legal opinions across various countries concluded that the integration of ESG information was consistent with fiduciary duty because it could enhance long-term financial returns (Watchman et al., 2005: 13). Crucially, this interpretation assumed that ESG information contains so-called financially material information, whose non-consideration might lead to the misallocation of funds (cf. Barman, 2018: 292). To establish the case for financial materiality, proponents of ESG Investing presented case studies that showed a positive link between ESG information and financial performance (e.g. Barron et al., 2006). Subsequent review studies also turned to the data from established research agencies like EIRIS or KLD to explore the ESG-performance relationship (UNEP FI and Mercer, 2007).

The refocus on ESG’s contribution to financial value was, however, not total, as remnants of SRI’s preoccupation with values remained alive within mainstream financial institutions. One interviewee from the asset management industry noted that “a breach of values such as [financial] value creation from child labour has a dimension of financial risk, but the key driver in excluding such activities is that you simply do not want them.” (Asset Manager, Western Europe 1, February 2019, my emphasis). Another interviewee stated, “ESG is understood from a risk perspective, on the one hand, and from an impact perspective, on the other hand” (Asset Manager, Western Europe 2, January 2019). Thus, the movement from SRI to ESG was an expansion rather than a replacement. Equally, on the level of technical devices SRI research providers, now labelled ESG agencies, continued to offer products to faith-based investors (Researcher, ESG Agency, July 2018; Representative, ESG Agency, Policy Workshop, Brussels 2018) but also faced audiences that “do not have an intrinsic ethical motivation” (Researcher, ESG Agency).

The addition of mainstream financial institutions as demanders of ESG information changed the actor dynamics of the ESG infrastructure by shifting the definitional power from Socially Responsible investors to ESG agencies. Socially Responsible investors articulated their values and concerns and relied on ESG agencies for information. Institutional investors, by contrast, outsourced the identification of financially relevant issues to agencies, which, as mentioned by one interviewee, became the source of ESG definitions (Co-founder, Sustainability Consultancy).

This change in actor dynamics also affected the technical devices of the ESG information infrastructure. Rather than replacing the installed base including scoring methods and exclusion lists, ESG agencies answered the call for providing financially material information by expanding their coverage of issues, sectors and geographies. As these expansions necessitated additional resources, the agencies were subject to consolidation dynamics, which centralized the offered technical devices among a few leading organizations (cf. Eccles and Stroehle, 2018; Escrig-Olmedo et al., 2019). With the concept of ESG entering the dictionary of mainstream investors, agencies restructured and expanded their indicators under the three pillars of E, S and G. To do so, agencies identified subthemes and issues that went through various intervals of aggregation and weighting to produce an aggregate company ESG score.

Re-arranging and aggregating old and new indicators to arrive at supposedly financially material overall ESG scores obscured, however, some of the choices regarding indicator design and measurement. For instance, input measures like the (non-)presence of a policy for workplace safety were combined with output measures like the number of workplace accidents. Due to the proprietary nature of methodologies, outside observers were often unaware of whether aggregate ESG scores reflected input or output measures (Drempetic et al., 2020: 355). Moreover, output measures included absolute physical quantities like the amount of greenhouse gases released in a year and so-called economic intensity indicators that set physical quantities in relation to economic measures like sales, revenue or market capitalization. Importantly, intensity indicators and absolute indicators do not measure the same concept, with the former type being criticized for measuring economic efficiency rather than environmental performance and for being susceptible to economic volatilities (Raynaud et al., 2020: 120ff).

Instead of addressing these critiques, ESG agencies mostly focused on finding an answer to the demand for financially material information. In doing so they engaged with the appearance of aggregated metrics rather than with changing individual indicators. Accordingly, aggregated ESG scores were redesigned to mimick traditional financial metrics like credit ratings or risk scores by adding industry-based selection criteria and specific weights that reflected an agency’s judgement on financial materiality (cf. Escrig-Olmedo et al., 2010: 458ff; Hughes et al., 2021: 20; MSCI, 2022: 2ff).

Despite these transformations and additions, the relationship between mainstream investors and ESG information remained beset with conflicts and distrust. Surveys among Swiss pension funds (Hierzig, 2016: 17,32), US asset managers (Morgan Stanley and Bloomberg, 2019: 12), and stakeholder consultations (Eltogby et al., 2019: 14) found considerable confusion about the meaning, reliability and comparability of ESG information. The increasing availability of ESG information also enabled the academic community to document gross inconsistencies between the products of different agencies (e.g. Chatterji et al., 2016). Moreover, sustainability reporting from companies, which remained one of the core inputs for leading ESG agencies (cf. Hughes et al., 2021: 6), was found to be deficient and inconsistent (Busch et al., 2022; Dragomir, 2012).

When applying an infrastructural lens, these tensions do not appear as too surprising as they reflect the frictions that emerge from repurposing and amending the installed base of technical devices that catered to the needs of niche Socially Responsible investors to the demands of institutional investors and their preoccupation with financial materiality. The mentioned expressions of discontent with the state of the infrastructure already foreshadowed a potential erosion of the definitional power that the consolidated ESG agencies had acquired.

Period 3: Contemporary sustainable finance

Towards the second half of the 2010s, ESG Investing became a mainstream financial sector issue. The academic literature (Friede et al., 2015) and the discourse in the specialist press (Dumas and Louche, 2016: 450) at the time agreed that ESG Investing did not harm financial performance and, in many cases, contributed to enhanced risk-adjusted returns. In addition, with the adoption of the Paris Agreement in 2015, which made explicit reference to financial flows in article 2.1c, and the presentation of the SDGs in the same year, the integration of ESG issues and climate change in particular into financial regulation became a key concern for governments, regulators, central banks and IOs (Economist, Environmental Agency, Western Europe; NGO Staff, Climate Finance, North America, February 2019). ESG Investing was also increasingly subsumed under the broader label of Sustainable Finance, which shifted the debate to mostly climate change and emissions (Senior Researcher, UNEP Inquiry; Research Director, Think Tank).

Against the background of these developments, the market for ESG data doubled from US $ 300 million in 2016 to over US $ 600 million in 2019 (Foubert, 2020) and overall ESG spending stood at US $ 2.2 billion in 2020 (UBS, 2020). However, there were doubts whether the ESG information infrastructure, which had become dominated by a few leading agencies, could be trusted. The criticisms levelled against ESG information were that it lacked standardization as human analysts introduced subjectivity, that proprietary methodologies left it wanting regarding transparency, thereby introducing biases and that the reliance on corporate disclosures exposed it to greenwashing (Hughes et al., 2021: 4; see also Berg et al., 2022; Busch et al., 2022; Drempetic et al., 2020).

These issues were not merely of academic concern. One interviewee identified the comparability of ESG data as a key “catalytic factor” for the promotion of sustainable finance (Board Member, SRI Association, September 2017). Others complained about the “semantic confusion [of ESG]” (Asset Management Expert, North America) and cautioned that “data quality is not a given” (Economist, Research Agency, October 2018). Conference participants pointed out the “confusion of language” (Head of ESG, Global Financial Data Provider, EU-Stakeholder Dialogue, Brussels, June 2019), the “fluffiness [of ESG]” (NGO Representative, Academic-practitioner Conference, Zurich, January 2018) and the “challenge (…) to filter out the noise of greenwashing” (Head of Sustainable Investment, Global Bank, Academic-practitioner Conference, Zurich), while calling for a “common language” (Professor, Academic-practitioner Conference, Oxford, September 2019; Chair of Technical Expert Group (TEG), EU-Stakeholder Dialogue), “common standards and definitions” (Head of Strategy, Global Bank, OECD Forum, Paris, October 2019) and “ungameable metrics” (Professor, Academic-practitioner Conference, Oxford).

In light of the simultaneous growth of both use of and dissatisfaction with the ESG information infrastructure, private as well as public actors sought amend and challenge its technical devices. A key entry point for these actors were the corporate disclosures that informed many ESG products. Existing materiality-focussed company disclosure frameworks dating from around 2010 (cf. Barman, 2018) were thus complemented by a further “wave of standardization initiatives” from 2015 onwards (Board Member, SRI Association).

One prominent standardization initiative is the private sector-led Taskforce for climate-related financial disclosures (TCFD). The TCFD has been endorsed by 3000 organizations with a combined market capitalization of US $ 27.2 trillion (TCFD, 2022) and has been integrated into the mandatory disclosure requirements of the EU, Hong Kong, Switzerland and the UK (TCFD, 2021). On the one hand, the TCFD is reminiscent of earlier standards by adopting a high-level typology for corporate self-reporting and by focusing on financial materiality, thus reproducing important aspects of the ESG information infrastructure’s installed base. On the other hand, the TCFD also introduces novel aspects like the definition of financial materiality of climate-related risks and opportunities in relation to forward-looking information (TCFD, 2017: iii). Accordingly, companies and financial institutions are asked to report how the effects of different pathways of global warming (e.g. >3°, <2°) relate to their risks and opportunities. In practice, this is mainly operationalized through links with third-party scenarios from institutions like the International Energy Agency (IEA) (TCFD, 2017: 35ff; TCFD, 2021: 65).

A second influential initiative that partly reproduces and partly amends the ESG information infrastructure is the EU’s taxonomy for sustainable activities. The EU taxonomy regulation was adopted in 2020 and sought to address greenwashing by providing a common reference for the mandatory reporting of financial institutions and non-financial companies (EU, 2020: Recital 11; Article 1). EU officials referred to the taxonomy as a “grammar” (EC Official, OECD Forum) or “common language” (EC Vice-president Valdis Dombrovskis, EU-Stakeholder Dialogue). To establish the taxonomy as a trustworthy universal grammar of sustainable finance, the EU emphasized its ability to translate international commitments like the Paris agreement and the SDGs for investors (TEG, 2019a: 7) as well as its character as a “science-based” instrument (EC Vice-president Valdis Dombrovskis, EU-Stakeholder Dialogue).

The taxonomy sets out criteria that make an economic activity sustainable. Just as with the indicators developed by commercial ESG agencies this is operationalized by establishing quantitative output thresholds or input requirements for whether an economic activity counts as a substantial contribution to an environmental goal. Other thresholds or input requirements check that the activity does not constitute significant harm to other environmental goals. Despite its claims to standardization, at the level of individual indicators, the taxonomy perpetuates some of the problems of private ESG information such as the conflation of input and output indicators into one framework. By way of illustration, the taxonomy combines intensity thresholds like a maximum of 0.722 tons CO2 equivalent per ton of cement produced, absolute thresholds as in the requirement for trains to have zero tailpipe emissions and input-based information like the requirement for forestry projects to have a detailed forest management plan (EC, 2022).

While the taxonomy’s dichotomous classification of activities as sustainable or non-sustainable appears reminiscent of the values-based exclusion criteria of SRI, representatives from this group were marginal in its development (Board Member, SRI Association). This is not least observed in the deprioritization of the social part of ESG. While being a core priority of European SRI pioneers like French trade unions, in the taxonomy, social aspects were deemed to be “harder” and an issue for “another day” (Chair of Technical Expert Group, EU-stakeholder Dialogue). Instead, actors with a background in green bonds, like the European Investment Bank (EIB), informed the shape of the taxonomy (Co-founder, Sustainability Consultancy; see also list of members 2019b: 8–9).

In connecting ESG information to external, traditionally non-ESG-related, references like IEA scenarios or green bond evaluations, the TCFD and the Taxonomy challenged the definitional power of ESG agencies. Yet rather than seeing them as challenger projects that aim to assert private, investor-led (TCFD) or public (Taxonomy) control, an infrastructural perspective calls attention to how they translate between the installed base of ESG information and alternative technical devices.

These alternative technical devices include scenario methodologies that benchmark corporate decarbonization strategies to the IPCC’s climate pathways or the IEA’s energy models. These scenarios were developed amongst other by publicly funded think tanks (2°ii, 2022, Director Think Tank, Academic-practitioner Conference, Zurich) and UN and environmental NGO backed coalitions (SBTi, 2021). They have also been flanked by discursive shifts such as the equation of climate-related risks with carbon intensive “stranded assets”, likely to lose value in a below 2° policy environment. (Leaton, 2011, see also NGO Staff, Climate Finance, North America; Representative, ESG Agency, Policy Workshop).

Another influential technical device category are green bond classification frameworks that were first created by supranational development banks like the EIB and the World Bank (Tripathy, 2017: 241). Unlike traditional ESG ratings, which evaluate companies, green bond assessments focus on specific projects, such as the construction of a dam or a power plant (cf. EIB and Peoples’ Bank of China, 2017). The rising prominence of green bonds among investors, and the integration of their promoters like the EIB into policy-making mechanisms like the EU taxonomy, has progressively intertwined the calculative choices of green bond evaluations with other facets of the ESG information infrastructure.

Viewing these developments through an infrastructure lens highlights both shifts and consistencies. With regards to changes introduced through scenario analyses, the use of IPCC and IEA models arguably expands the timeframe for determining the risk of assets, thereby potentially challenging more short-termist notions of financial materiality. Moreover, some scenarios approaches (e.g. 2° ii, 2022) address criticisms over opaque and proprietary methodologies from ESG agencies by making their models open-source. At the same time, some scenarios reproduce existing practices such as the use emission intensity-based indicators that are popular with corporates and ESG agencies. Relying on these indicators can, however, can lead to the counterintuitive outcome, where companies’ alignment with a scenario are concurrent with absolute emissions surpassing the scenario’s GHG budget (cf. Raynaud et al., 2020: 53)

Integrating green bond evaluation frameworks into the ESG information infrastructure represents, likewise, both a departure from and a continuation of existing methodologies. On the one hand side, the taxonomy’s initial reporting cycle marked a significant shift, challenging companies and financial institutions with unfamiliar data requests (cf. Arnold et al., 2021: 45). On the other hand, there are also considerable continuities. As noted by Langley et al. (2021: 504), despite their different asset focus, green bond evaluation frameworks have also relied on common SRI metrics. Additionally, while the taxonomy might seem to undermine established ESG data intermediaries, many ESG agencies are now offering “Taxonomy Alignment Solutions” to assist companies with their reporting (IOSCO, 2021: 15).