Policy Research Working Paper 11226 Green versus Conventional Corporate Debt From Issuances to Emissions Juan J. Cortina Claudio Raddatz Sergio L. Schmukler Tomas Williams Development Economics Development Research Group October 2025 Policy Research Working Paper 11226 Abstract This paper investigates how firms use green versus conven- green bond and loan issuances are systematically followed tional debt and the associated firm- and aggregate-level by sustained reductions in carbon intensity (emissions over environmental consequences. Employing a dataset of income) of up to 50 percent. These reductions correspond 127,711 global bond and syndicated loan issuances by to as much as 15 percent of global annual emissions. Green non-financial firms across 85 countries during 2012–23, the bonds contribute to reducing emissions by providing financ- paper documents a sharp rise in green debt issuances relative ing to large, high-emitting firms, whose improvements in to conventional issuances since 2018. This increase is par- carbon intensity have significant aggregate consequences. ticularly pronounced among large firms with high carbon Syndicated loans do so by channeling a larger volume of dioxide emissions. Local projections difference-in-differ- financing to a wider set of firms. ences estimates show that, compared to conventional debt, This paper is a product of the Development Research Group, Development Economics. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at jcortinalorente@worldbank.org; clraddatz@fen.uchile.cl; sschmukler@worldbank.org; tomaswilliams@email.gwu.edu. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Green versus Conventional Corporate Debt: From Issuances to Emissions∗ Juan J. Cortina Claudio Raddatz World Bank Central Bank of Chile, School of Economics and Business, Universidad de Chile Sergio L. Schmukler Tomas Williams World Bank George Washington University Authorized for distribution by Deon Filmer, Director, Development Research Group, Development Economics, World Bank Group Keywords: carbon emissions, corporate bonds, firm growth, green debt, green transition, sustainability, syndicated loans JEL Codes: F33, G00, G01, G15, G21, G23, G31 * We are grateful to Shawn Cole, Caroline Flammer, Galina Hale, Guojun He, and participants at the 2024 Annual Meetings of the Chilean Economic Association, Bank of Korea, HKIMR-ECB-BOFIT conference, IMF MCM Policy Forum, 7th Summer School of Sustainable Finance of the European Commission, and UCSC Center for Analytical Finance workshop for useful comments. Brian Castro, Paolo Sabella, Matias Soria, and Patricio Yunis provided excellent research assistance. The World Bank’s Chile Research and Development Center and Knowledge for Change Program (KCP) and Chile’s ANID Fondecyt Project 1220115 provided financial support. The findings, interpretations, and conclusions are those of the authors and do not necessarily represent those of the World Bank, its executive directors, or the governments they represent. Email addresses: jcortinalorente@worldbank.org; clraddatz@fen.uchile.cl; sschmukler@worldbank.org; tomaswilliams@gwu.edu. 1 Introduction Concerns about the economic and environmental consequences of climate change have heightened interest in the role of the financial system in supporting industrial decarbonization (Stern, 2006; IPCC, 2018; IEA, 2021; Draghi, 2024). In response, financial actors have integrated climate considerations into capital allocation decisions at an accelerating pace (GFMA and BCG, 2020). Between 2013 and 2023, assets under management in sustainability- focused funds grew tenfold, reaching over $3.4 trillion, or 7.2 percent of total global portfolios. By 2023, institutions representing more than half of global banking assets had endorsed the UN Principles for Responsible Banking, committing to align their strategies with sustainability goals (United Nations, 2023; Morgan Stanley, 2025). These developments reflect a broader shift toward environmental accountability in corporate finance (OECD, 2021a; CPI, 2023). Green debt instruments—comprising bonds and syndicated loans—have emerged in this context as leading tools for mobilizing capital toward environmental objectives (ICMA, 2021; OECD, 2021b). These instruments incorporate climate considerations by earmarking proceeds for green uses or by linking repayment terms to environmental performance. While designed to promote decarbonization, the actual relation between green bonds and syndicated loans and changes in firm behavior and carbon emissions—both at the micro and macro levels—remains insufficiently understood. a-vis conventional debt instruments In this paper, we analyze how firms use green vis-` to raise capital, and how these financing choices relate to subsequent changes in firm- and aggregate-level performance. We systematically compare corporate debt issuance along two key dimensions: (i) green versus conventional debt and (ii) bonds versus syndicated loans. We examine how debt market participation relates to issuer characteristics and how firms evolve after issuance in terms of scale (measured by assets and operating income or sales) and carbon intensity (defined as carbon emissions per unit of income). To assess the aggregate significance of green debt, we combine firm-level responses with observed issuer profiles and baseline emissions, offering new evidence on the relation between green debt borrowing and global emissions. To conduct the analysis, we construct a novel global dataset that links green and 1 conventional corporate bond and syndicated loan issuances to firm-level balance sheet and carbon emissions data. The dataset covers 120,711 conventional debt and 6,412 green debt transactions issued in domestic and international markets by 50,832 non-financial firms across 85 countries between 2012 and 2023.1 We define green debt broadly to include both green-labeled and sustainability-linked instruments, providing a comprehensive view of the market aimed at directly aligning financing with environmental objectives. We document a sharp expansion of green debt issuance after 2018, coinciding with a broader slowdown in conventional corporate borrowing. Between 2017–2018 and 2022–2023, annual green issuance increased nearly ninefold, reaching 12 percent of total corporate debt issuance by the end of the period. While the United States and China continue to dominate the conventional debt issuance activity, they play a much smaller role in the green segment. In contrast, Europe has emerged as the global leader, accounting for nearly half of all green issuances between 2012 and 2023. Over time, green syndicated loans have surpassed green bonds in total credit volume and reached a broader range of countries, sectors, and firms. Green debt is issued primarily by large incumbent debt issuers with high carbon footprints, most of which rely on conventional borrowing before incorporating green instruments into their financing choices. Among these “hybrid” issuers using both conventional and green debt, the share of green debt rose from just 1 percent of their total debt issuance in 2013–2014 to roughly 35 percent by 2022–2023. The median hybrid issuer is about five times larger—both in total assets and carbon emissions—than firms that rely exclusively on conventional debt. The relation between firm size and green debt issuance has strengthened over time, suggesting that high entry costs into a new, untested debt market do not drive this result. Compared to green bonds, green loans are used by smaller companies and first-time debt issuers. To assess the environmental performance, we estimate firm-level trajectories of scale and carbon intensity, the two components that jointly determine total emissions. We implement a local projections difference-in-differences (LP-DiD) approach (Dube et al., 2025) to trace dynamic firm responses to debt issuance and compare outcomes across green and conventional instruments. The results indicate that scale trajectories are broadly similar following green 1 We focus on the non-financial corporate sector due to its central role in global emissions (CDP, 2023). 2 and conventional debt issuances. In contrast, carbon intensity shows a marked divergence: it remains flat or rises after conventional issuance, but declines steadily in the aftermath of green debt issuance, reaching a cumulative reduction of approximately 50 percent by year four. These improvements are also observed within firms that issue both green and conventional debt. While green bonds and loans exhibit comparable reductions in carbon intensity, bond issuance is followed by larger post-issuance expansions in firm scale. We extend the analysis to the aggregate level by combining our LP-DiD estimates with the observed characteristics of issuers and their baseline emissions. This aggregation suggests that green debt issued between 2018 and 2023 could be associated with projected cumulative carbon dioxide (CO2 ) reductions of 4.5 billion to 5.7 billion metric tonnes by 2025. These figures represent roughly 12 to 15 percent of one year’s global energy-related CO2 emissions. A disproportionate share of this aggregate abatement is driven by large firms. For example, the top quartile of firms in our sample accounts for more than 85 percent of the total estimated reductions. Overall, the aggregate relevance of green debt reflects the profile of its issuers and how their emissions trajectories evolve afterward. Relative to conventional borrowing, green issuance remains concentrated among large firms with substantial carbon footprints. At the same time, issuers of green debt display consistent reductions in carbon intensity. The participation of large global emitters in green debt markets and their post-issuance performance are especially relevant for aggregate outcomes. Within green debt, bonds and loans shape aggregate carbon emissions outcomes through different channels. Bonds are issued predominantly by large firms with high baseline emissions, so reductions in their carbon intensity translate into sizeable aggregate declines, even when post-issuance expansion tempers these gains. Loans, by contrast, extend to a broader and more heterogeneous group of borrowers. Moreover, their aggregate issuance volume exceeds that of bonds. At the firm level, loan issuances are followed by carbon intensity reductions of a similar order to those of bonds, but with a smaller offsetting expansionary effect. Overall, the larger scale and emissions profile of bond issuers explain why bonds account for most aggregate reductions, while loans complement this by channeling green finance more widely 3 across the corporate sector. This paper adds to a growing literature on the drivers and consequences of the rapid expansion of the green debt market. A significant portion of this literature examines whether rising investor demand for green debt instruments translates into financial benefits for issuing firms—either through a lower cost of debt (the so-called “greenium”) (Hachenberg and Schiereck, 2018; Zerbib, 2019; Bachelet et al., 2019; Baker et al., 2022; Caramichael and Rapp, 2024), or via a lower cost of equity through signaling effects that attract environmentally minded shareholders (Tang and Zhang, 2020; Flammer, 2021). Another strand investigates whether, given the convenience of issuing green debt, firms use it to improve their environmental outcomes (Fatica and Panzica, 2021; Flammer, 2021; Dursun-de Neef et al., 2023), or instead to finance business-as-usual activities (Ehlers et al., 2020; Tuhkanen and Vulturius, 2022; Lam and Wurgler, 2024). A smaller number of studies explore the characteristics of firms issuing green debt in specific markets (Cicchiello et al., 2022; Dutordoir et al., 2024). Most of this research focuses on green bonds, with only a few papers analyzing the fast-growing green syndicated loan market (Aleszczyk et al., 2022; Du et al., 2023; Dursun-de Neef et al., 2023). We contribute to this literature by systematically analyzing which firms issue green debt and how they use it relative to conventional debt across green debt products. Unlike prior work, our focus is on how the development of the green debt market has altered firms’ financing patterns and what consequences this has for their economic activity and environmental footprint. Our joint analysis of issuance decisions, firm expansion, and carbon intensity offers two additional advantages. First, it enables a more detailed characterization of the relation between firms’ financing choices and their environmental outcomes. Second, it allows us to project the macroeconomic implications of the green debt issuance by combining firm-level dynamics with baseline emissions and issuance volumes. Importantly, by incorporating the syndicated loan segment—which exceeds the green bond market in size—we capture firms’ use of different green financing instruments and better gauge the significance of the overall green debt market. Understanding how green debt markets have reshaped firm financing is both important 4 and not obvious. On the one hand, green finance can enhance access to capital and enable the expansion of innovative, environmentally focused firms, which might otherwise not have easy access to conventional markets. On the other hand, green debt issuance entails stricter requirements—including third-party certification and environmental reporting—which can raise issuance costs and restrict eligibility, especially among smaller firms (ICMA, 2020a; Abraham et al., 2021a). Furthermore, the swift rise of the green syndicated loan market relative to green bonds can influence these dynamics. Banks are traditionally more adept at evaluating and monitoring opaque borrowers, potentially broadening green finance access. Yet, several studies suggest that syndicated loans have become increasingly “market-like,” being distributed to non-bank investors and traded on secondary markets (Ivashina and Scharfstein, 2010; Aldasoro et al., 2022; Albuquerque et al., 2025), which could limit their distinction from bonds. The environmental impact of green debt issuance is similarly nuanced. While green instruments are designed to finance environmentally friendly activities, the empirical evidence on post-issuance improvements in firms’ carbon footprint remains mixed. More fundamentally, how much carbon firms emit depends not only on changes in carbon intensity but also on the scale of their operations. Here again, the market structure can influence outcomes. Syndicated loans involve more direct monitoring than bonds issued in arm’s-length transactions to dispersed investors (Diamond, 1984; Rajan, 1992; Holmstrom and Tirole, 1997; Boot and Thakor, 2000; Acharya et al., 2011). This monitoring could constrain opportunistic uses of proceeds and enhance the credibility of environmental commitments. In contrast, bonds are subject to looser ex-ante scrutiny and less ongoing oversight, which could allow for greater discretion in capital deployment (Bruno and Shin, 2017; Abraham et al., 2021b; Acharya and Plantin, 2025; OECD, 2025). Ultimately, the aggregate environmental relevance of green debt depends not only on firm-level emission reductions but also on the number and scale of participating firms. Since large emitters account for a disproportionate share of global emissions, their climate actions carry outsized aggregate consequences (CDP, 2023; Acharya et al., 2025). Our paper contributes to this literature by assembling comprehensive new global data to 5 examine how green debt adoption relates to firm-level performance and aggregate emissions. We show that, unlike conventional debt, both green bonds and loans are systematically associated with reductions in carbon intensity. A key novelty of our approach is to analyze the two instruments jointly rather than in isolation. This perspective reveals that the use of green debt largely mirrors that of conventional debt markets: bonds concentrate among large emitters, and loans reach a broader and more diverse set of firms. Firm-level improvements in carbon intensity are comparable across the two instruments and appear tied to their green design. At the aggregate level, however, differences in market reach generate distinct outcomes for bonds and loans. By analyzing them within a unified framework, our study sheds light on the channels through which green debt markets contribute to emissions abatement. The remainder of this paper is structured as follows. Section 2 provides an overview of the green debt market and details the main dataset used in the analysis. Section 3 examines the distribution of green and conventional debt over time and across different regions. Section 4 shows the types of firms issuing green and conventional debt. Section 5 investigates the performance patterns at both the firm and aggregate levels associated with conventional and green debt issuances. Section 6 offers concluding remarks. 2 Structure of Corporate Debt Markets and Data Collected 2.1 Conventional and Green Debt Markets Corporate bonds and syndicated loans are the primary instruments through which firms raise debt at scale. Both markets have expanded significantly over the past two decades, increasing their importance relative to traditional bank lending and playing a central role in the global growth of corporate leverage (Abraham et al., 2021a; Aramonte et al., 2023; Acharya et al., 2024). Bonds and syndicated loans increasingly operate as part of a common market for corporate debt. Syndicated loans—while originated by banks—have evolved to share some characteristics of public bonds in key respects: they are often syndicated to non-bank institutional investors, structured off banks’ balance sheets, tranched by risk, and actively traded in secondary markets (Gatev and Strahan, 2009; Ivashina and Scharfstein, 2010; 6 Aldasoro et al., 2022; Albuquerque et al., 2025). At the same time, syndicated loans retain features that reflect their banking origins. Even when packaged into collateralized loan obligations (CLOs), banks typically remain responsible for origination and loan servicing. In contrast, bonds are fully intermediated through market-based underwriting and servicing. Since the early 2010s, both instruments have evolved to incorporate environmental objectives, giving rise to the emergence of green bonds and syndicated loans. We define green debt instruments as those that either allocate proceeds for environmentally beneficial projects or connect financial terms to achieving sustainability targets. This includes traditional green-labeled instruments tied to the use of proceeds and newer sustainability-linked bonds and loans tied to environmental performance. While there is no universal regulatory definition of green debt, our approach reflects the two main ways individual debt transactions align with climate goals. Green-labeled instruments publicly commit capital to green projects, whereas sustainability-linked instruments incorporate pricing incentives based on sustainability outcomes. We classify all other debt instruments as conventional. In bond markets, green-labeled bonds retain the same financial structure as conventional bonds but include a public commitment by the issuer to allocate proceeds to environmentally beneficial projects—such as renewable energy, energy efficiency, or clean transportation. These commitments are typically guided by voluntary market standards rather than enforced through binding contractual clauses. As a result, compliance is driven by reputational concerns, investor oversight, and, in some cases, third-party verification. Issuers can self-label or align green bonds with standards such as the ICMA Green Bond Principles (ICMA, 2021). A subset of these bonds is aligned with the Climate Bonds Initiative (CBI) taxonomy, which builds on ICMA standards by specifying more detailed eligibility criteria for green projects. Some issuers seek formal CBI certification, while others simply declare alignment without certification. In contrast, sustainability-linked bonds tie financial terms—typically coupon step-ups—to the issuer’s progress toward predefined environmental or ESG performance targets (ICMA, 2020b). A similar green debt structure has emerged within syndicated loan markets. Green- labeled loans follow a use-of-proceeds model in which borrowers commit to allocating funds 7 to eligible environmental investments. As with green bonds, these commitments are generally governed by voluntary guidelines, such as the Green Loan Principles published by the Loan Market Association (APLMA, LMA, and LSTA, 2019). Sustainability-linked loans, by contrast, incorporate environmental incentives directly into the loan contract, with pricing terms (e.g., interest rates) adjusting based on the borrower’s performance against specified environmental targets. Introduced in 2017, these instruments have grown rapidly, and thus, represent a substantial share of green lending activity (Du et al., 2023). 2.2 Data To study the importance of the green debt market, we construct a comprehensive dataset that links transaction-level data on green and conventional corporate debt issuances to firm-level financial and carbon emissions from 2012 to 2023. The transaction-level data come the Securities Data Company (SDC) Platinum database from LSEG (the London Stock Exchange Group), formerly from Refinitiv and Thomson Reuters. This database offers detailed coverage of global corporate bond and syndicated loan transactions, including issuance volume, nature of the debt and its terms, and borrower characteristics. LSEG also classifies syndicated loans into conventional, green-labeled, and sustainability-linked categories. Loan transactions include term loans, fixed-amount instruments with scheduled repayments, and revolving credit facilities, allowing borrowers to draw down funds up to a predefined limit as needed. While these structures differ in flexibility and repayment, both are syndicated through similar bank-led arrangements, serve comparable financing functions, and share key contractual features. In our analysis, we group sustainability-linked term loans and revolvers together, as they represent conceptually similar green pricing-based instruments. Moreover, more than half of revolver transactions occur with term loans attached in the same package, and both aim to incentivize environmental performance through margin adjustments tied to sustainability targets.2 For green bonds, we supplement the transaction-level dataset with Govsearch, also from LSEG, which provides systematic green bond identification based on issuer disclosures 2 The main results in this paper hold when sustainability-linked revolver transactions are excluded from the analyses. 8 and external certification frameworks. This source enables the classification of green bonds into five categories: CBI-aligned, CBI-certified, self-labeled green bonds, sustainability bonds, and sustainability-linked bonds.3 The firm-year balance sheet and income statement data are obtained from Worldscope Fundamentals, also from LSEG, which provides standardized global coverage of publicly listed firms. Total debt is defined as the stock of all outstanding liabilities, encompassing both green and conventional instruments. Measures of firm size or operational scale are total assets and operating income.4 As a complementary size proxy, we compute the average transaction size across all debt issuances by each firm during the sample period, using the global transactions dataset. This variable extends coverage to both listed and unlisted issuers and is highly correlated with balance-sheet size measures (Appendix Figure 1). Therefore, it provides a useful benchmark for assessing whether the characterization patterns observed for listed firms are consistent across the broader set of debt issuers. All monetary variables from Worldscope (e.g., assets, income, debt), as well as the transaction-level issuance volumes, are deflated to constant 2011 U.S. dollars and reported in millions.5 Firm-year carbon emissions data are primarily sourced from LSEG, with supplementary data from MSCI used only when a firm has missing LSEG emissions data for all years in the sample. The dataset includes Scope 1 (direct emissions) and Scope 2 (indirect emissions from purchased energy), reported in millions of tonnes. Emissions figures include CO2 and CO2 equivalent greenhouse gases.6 For simplicity, we refer to all these as “carbon emissions” throughout the paper. We use this dataset to construct total carbon emissions and carbon intensity, defined as total Scope 1 and 2 emissions (in metric tonnes) divided by firm operating income (in millions of U.S. dollars). 3 Sustainability bonds are a type of green-labeled bond, typically aimed at financing a mix of environmental and social projects. They differ from sustainability-linked bonds, which are pricing-based instruments rather than green-labeled instruments. 4 For comparison, we also use physical capital (property, plant, and equipment) in the summary statistics. 5 To deflate all monetary values to constant 2011 U.S. dollars, we use the U.S. Consumer Price Index (CPI). Domestic currency values are first converted to U.S. dollars using contemporaneous exchange rates. This choice is unlikely to bias the empirical results on green versus conventional debt. Both types of debt issuers within a country are similarly exposed to domestic conditions, and all regressions include country-year fixed effects to absorb macroeconomic shocks. 6 CO2 equivalent includes methane (CH4 ), nitrous oxide (N2 O), hydrofluorocarbons (HFCs), perfluorinated compounds (PFCs), sulfur hexafluoride (SF6 ), and nitrogen trifluoride (NF3 ). 9 To construct a matched panel, we aggregate transaction-level data to the firm-year level, summing total issuance volumes for each firm separately by instrument type. For firm-years with no recorded issuance, we explicitly assign zeros to ensure a balanced panel coverage. We then merge the debt data with the firm-level financials and carbon emissions using a combination of identifiers, including permanent firm IDs, ISINs, SEDOLs, and CUSIPs. The final dataset includes 120,711 annual conventional debt transactions and 6,412 green debt transactions across 85 countries during 2012–2023 (Table 1).7 Green debt transactions tend to be larger than their conventional counterparts. The median issuance size for green debt is $200 million, compared to $149 million for conventional debt. The median maturity is similar across categories, approximately 5.2 years for green debt and 5.0 years for conventional debt. Still, within each category, bonds exhibit longer maturities than syndicated loans (by about 1 year on average). Across all green debt transactions, 58 percent of issuances are sustainability-linked, while 42 percent are green-labeled instruments. However, this composition varies significantly by instrument type. Among green bonds, the majority (88 percent) are green-labeled instruments, whereas 70 percent of green syndicated loans are sustainability-linked.8 There is also significant heterogeneity within each category of green bond and loan instruments (Appendix Table 2). Among green bonds, self-labeled green issuances are systematically smaller and shorter in maturity than the other types of green issuances. In particular, their median size is roughly half that of CBI-aligned or CBI-certified bonds, and their maturities are 1 to 2 years shorter. Similar differences hold when comparing self-labeled green bonds with sustainability-linked bonds. In the loan market, green-labeled loans are also substantially smaller than sustainability-linked transactions, though they tend to have somewhat longer maturities (about 2 years). In terms of firms, the dataset includes 50,832 unique non-financial firms across 85 countries (Table 2). Of these, 46,987 firms issued only conventional debt, while 3,845 firms participated in the green debt market. Among green debt issuers, 1,334 firms issued green 7 Appendix Table 1 provides the full list of countries and the total amounts each country raised through green and conventional debt. 8 Approximately 91 percent of green debt issuances are classified as investment-grade, compared to 70 percent for conventional debt. 10 bonds, 2,773 issued green loans, and only 157 issued both types of instruments (Table 3).9 This limited overlap suggests that firms tend to specialize in one form of green debt—either bonds or loans—rather than combining them. Approximately 40 percent of green debt issuers report balance-sheet and carbon emissions. This includes the world’s largest firms, which are also the most significant emitters. Because reporting requirements are generally required for listed firms or for those subject to enhanced ESG disclosure standards, this dataset provides the most comprehensive available basis for analyzing the relation between corporate debt and carbon emissions. 3 Evolution of Green and Conventional Debt To examine the evolution of green debt relative to conventional corporate debt over the 2012–2023 period, we begin by showing changes in overall issuance volumes and instrument composition. We then analyze the geographic and sectoral distribution of green versus conventional debt. Green debt issuance has expanded rapidly since 2018, contrasting with a decline in conventional borrowing. Between 2017–2018 and 2022–2023, annual green issuance rose nearly ninefold, from $64 billion to about $574 billion (Figure 1, Panel A). Over the same period, annual conventional debt issuance declined from $5.6 trillion to $4.1 trillion. As a result, the green share of total corporate debt issuance rose from 2 percent in 2017 to approximately 12 percent by 2023 (Figure 1, Panel B). Growth in green issuance occurred in both bond and loan markets, although with distinct trajectories. Green bond issuance peaked at approximately $200 billion in 2021 but declined to $145 billion during 2022-23 (Figure 2, Panel A). This expansion was primarily driven by the issuance of CBI-aligned instruments.10 Green syndicated loans grew even more rapidly, fueled by the adoption of sustainability-linked contracts. By 2023, green loans had become the dominant vehicle for corporate green finance, with total financing reaching $428 billion—nearly triple the volume of green bonds. Even when excluding revolving credit 9 For comparison, conventional debt markets exhibit a similar segmentation pattern: of the 46,987 conventional issuers, 42,094 issued only bonds or syndicated loans, while 3,893 firms issued both types of instruments. 10 Appendix Figure 2 shows the green debt evolution across different types of green debt instruments. 11 facilities, green loan volumes totaled $197 billion in 2023, exceeding green bond volumes. Green loans also exhibit a higher market share relative to their conventional counterparts. By 2023, they accounted for nearly 14 percent of total syndicated loan issuance, compared to roughly 10 percent for green bonds within the bond market (Figure 2, Panel B). The global distribution of green debt between 2012 and 2023 differs markedly from that of conventional debt (Figure 3, Panel A). Europe has emerged as the global leader in green debt markets, accounting for about 51 percent of total green volumes—well above its 19 percent share of conventional debt. This expansion reflects broad-based growth across European countries, most of which improved their global rankings in green debt volumes relative to conventional benchmarks (Appendix Table 1). Other regions have also modestly expanded their presence: East Asia (excluding China) increased its share from 3 to 6 percent, while Latin America (and the Caribbean) rose slightly from 2 percent to 2.4 percent. By contrast, the United States—long the dominant player in corporate bond and syndicated loan markets—lags in green finance. It accounts for 47 percent of global conventional debt but just 20 percent of green issuance. China shows a similar pattern, representing 13 percent of conventional issuance but about 5 percent of green debt. Penetration rates, measured as the share of green in total corporate debt, follow a similar pattern (Figure 3, Panel B). In 2022–2023, over 30 percent of European corporate debt was classified as green, around 18 percent in East Asia and Latin America, and only 6.5 percent and 3 percent in the United States and China, respectively.11 One possible driver behind the broader adoption of green debt in Europe is regulation (Demski et al., 2025). Europe’s leadership in green debt coincides with the rollout of a comprehensive regulatory framework for sustainable finance. The EU Green Taxonomy, finalized in 2020, is the world’s only legally binding classification system for environmentally 11 Regional groups are defined as follows. “Europe” includes all European Union member states plus Norway and the United Kingdom. “Other Advanced Economies” comprise Australia, Canada, Iceland, Japan, New Zealand, and Switzerland. The United States and China are reported separately. “East Asia” covers Hong Kong SAR, China, Indonesia, Malaysia, the Philippines, Singapore, Taiwan, China, Thailand, and Viet Nam. “Latin America (and the Caribbean)” includes Argentina, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, Mexico, Panama, Peru, and Uruguay. “Other Emerging Economies” comprise Albania, the Arab Republic of Egypt, Bahrain, Bangladesh, Bulgaria, Cambodia, the Democratic Republic of Congo, Cyprus, Georgia, Ghana, India, Israel, Kazakhstan, Kenya, Lao PDR, Lesotho, Mauritius, Mongolia, Morocco, Nigeria, Oman, Qatar, Russian Federation, Saudi Arabia, Senegal, Serbia, South Africa, T¨ urkiye, Ukraine, United Arab Emirates, Uzbekistan, and Zambia. 12 sustainable activities (Merler, 2025). It requires companies to disclose the share of operations aligned with sustainability criteria, offering a consistent reference for green investment. The Sustainable Finance Disclosure Regulation (SFDR), proposed in 2018 and implemented in 2019, mandates ESG disclosures from financial market participants (European Commission, 2024).12 In contrast, the United States lacks a unified federal regulatory framework for ESG investing. Its system remains fragmented, with variations across states and frequent political shifts that generate uncertainty for both issuers and investors (DataFisher, 2024). This divergence in the institutional context might explain the slower adoption of green instruments in the United States, despite its longstanding dominance in conventional debt markets. At the sectoral level, the distribution of green debt also differs from that of conventional debt. Utilities account for 43 percent of total green issuance compared to 25 percent of conventional issuance (Figure 3, Panel A). Renewable energy captures 0.5 percent of conventional issuance and 4 percent of green issuance. Manufacturing has a similar share in both markets, at around 30 percent of total issuance. By contrast, services, trade, construction, and fossil fuels have smaller shares in green relative to conventional markets. For instance, fossil fuels account for 10 percent of conventional issuance and 4 percent of green issuance. By 2022–2023, about 50 percent of all debt issuance in the renewable energy sector and 20 percent in utilities was classified as green, the highest sectoral penetration rates observed (Figure 3, Panel B). These regional and sectoral differences in green debt adoption are consistent across bonds and syndicated loans (Appendix Figure 3). However, green loans reach a wider set of countries and sectors than bonds. Among the 85 countries in our sample, 31 use green loans but not green bonds, while only 5 use green bonds but not green loans (Appendix Table 1). Importantly, the countries that use only green loans are the same ones that use only conventional loans. A similar pattern is present across industries: green loans cover a broader range of two-digit SIC sectors than green bonds, mirroring the differences in conventional 12 These efforts are further reinforced by the Corporate Sustainability Reporting Directive (CSRD) and the broader European Green Deal, which together strengthen investor confidence and align regulatory incentives for green issuance. The CSRD establishes harmonized sustainability reporting standards across firms and financial institutions. The European Green Deal, launched in 2019, outlines Europe’s sustainability and climate strategy. 13 loans and bonds. Thus, green loans might play an important role in broadening the reach of sustainable finance across countries and industries. 4 Firms Issuing Green and Conventional Debt To analyze how green debt is allocated across firms, we compare issuer characteristics with patterns observed in conventional corporate debt markets. We focus on three dimensions: prior use of debt markets, operational scale, and environmental profiles, with distinctions between green bonds and syndicated loans. Green debt markets are dominated by hybrid issuers—firms that issue both green and conventional instruments. Between 2012 and 2023, these firms accounted for 65 percent of all green debt issuers (Table 2), capturing nearly 85 percent of total green debt volume (Figure 4, Panel A). Among them, 91 percent had issued conventional debt before their first green transaction, indicating that green market participation has expanded mainly through firms with pre-existing access to traditional debt markets.13 These firms are also the largest in the sample, with a median asset size of $4.9 billion, more than 5 times that of both pure conventional issuers and pure green issuers (Table 2). They carry higher levels of debt, income, and physical capital, and are also substantially more polluting. Their median emissions reach 0.42 million tonnes, compared to 0.11 million tonnes for conventional issuers.14 This reflects the close link between firm size and emissions, with the top 10 percent of firms in our sample accounting for about 87 percent of reported emissions between 2012 and 2023 (Appendix Figure 1).15 Over time, hybrid issuers have increased their use of green instruments. By 2022–2023, green debt represented 35 percent of their total issuance (Figure 4, Panel B). This dominance of large firms is further reflected in the concentration of green debt markets. The top quartile of firms by asset size accounted for 87 percent of total green issuance over the sample period, compared to 77 percent in conventional markets (Figure 5, 13 Appendix Table 3 lists the top 20 green debt issuers, showing that the largest borrowers in green markets are well-known corporations, which are also prominent conventional debt issuers. Appendix Table 4 provides the corresponding ranking for lenders, indicating that the same global banks that dominate conventional syndicated loans also lead the bulk of green syndicated lending as arrangers. 14 The size differential between green and conventional issuers holds across regions and sectors (Appendix Figure 4). 15 Hybrid issuers are also more carbon intensive, with a ratio of 92.7, compared to 50.8 among conventional debt issuers. 14 Panel A). The gap is even wider when measured by the number of transactions. The top quartile of green issuers generated 61 percent of all green debt transactions, compared to 39 percent in the conventional market (Figure 5, Panel B). Green debt issuance is thus more concentrated at the top than in conventional markets, both in terms of issuance volume and activity. The patterns are broadly consistent when green debt is split into bonds and loans, but important differences arise. Concentration is particularly pronounced in the green bond segment: the top quartile of green bond issuers accounts for 92 percent of total issuance volume and 73 percent of transactions, compared to 84 and 53 percent, respectively, for green loans (Figure 6). Green bond issuers are also more than twice as large as green loan issuers in terms of median assets (Table 3). In contrast, the green loan market features a more balanced issuer composition, with greater participation from smaller firms and pure green issuers. Furthermore, for hybrid loan issuers, green debt constituted over 40 percent of total issuance by 2022–2023, whereas for hybrid bond issuers, it captured just 22 percent (Figure 4, Panel B). Overall, while green bonds remain concentrated in the largest firms, syndicated loans play a distinct role in broadening participation. To formally assess the link between firm size and participation in green debt markets, we estimate a linear probability model that links firm characteristics to the likelihood of issuing green rather than conventional debt, controlling for country- and industry-specific trends. We estimate the following equation: GDi,t = γ + βXi,t + θc,t + θs,t + ϵi,c,s,t , (1) where t ∈ {2012-2022} and GDi,t is a binary indicator equal to one if a firm issues green debt in a given year and zero if a firm issues conventional debt. In addition, to directly compare green bonds and loans, we redefine the dependent indicator GDi,t to equal one if a firm issues a green loan in a given year, and zero if it issues a green bond. The vector Xi,t includes the logarithm of the average transaction size (2012-23), lagged asset values, lagged income values, lagged carbon emissions, and lagged carbon intensity. The model controls for regional and sectoral heterogeneity using country-time and industry-time fixed effects θc,t and θs,t , respectively. We estimate Equation 1 using the full sample of firm-year observations for debt 15 issuers over the 2012–2023 period.16 The results confirm that firm size is a key predictor of green issuance (Table 4). For example, a doubling of total assets is associated with a 2 percentage point higher probability of issuing green rather than conventional debt, equivalent to a 34 percent increase relative to the baseline probability of 5.8 percent. Similar patterns hold across other firm-level characteristics: firms with higher income, larger fixed assets, and greater carbon emissions are significantly more likely to issue green debt. For example, a doubling of carbon emissions is linked to a 27 percent increase in the probability of green issuance. These relations are robust across model specifications and remain statistically significant after controlling for time-varying country and industry effects. The analysis also indicates that more carbon-intensive firms are more likely to issue green debt than conventional debt. However, once industry-year and country-year fixed effects are included, the magnitude and significance of this difference decrease. Size differences between green and conventional issuers persist across bond and loan markets (Table 5). Within each segment, firm size is positively associated with the probability of issuing green relative to conventional debt. However, across green instruments, larger firms and higher emissions are more associated with bond issuance than with loan issuance, consistent with the greater preponderance of bonds among large firms observed in the descriptive evidence. One plausible explanation for the prevalence of large firms among green debt issuers, relative to those relying on conventional instruments, is the presence of substantial fixed costs associated with green issuance. Regulatory compliance, sustainability reporting, third-party verification, and instrument structuring are generally more demanding for green debt than for conventional borrowing. For example, bonds aligned with or certified by the CBI must adhere to a detailed taxonomy of eligible activities and undergo formal pre- and post-issuance verification in the case of certification. These requirements impose added administrative and 16 We use a linear probability model (LPM) for ease of interpretation, as the coefficients can be directly read as changes in probability. Results are robust to estimating the same specification using logit and probit models, with nearly identical significance and directional patterns. Since our primary interest lies in comparing relative associations across firm characteristics and not in modeling probabilities at the extremes, the LPM provides a transparent and tractable baseline. 16 audit burdens. Similarly, sustainability-linked instruments require firms to define and track environmental performance targets, supported by ongoing data collection, external assurance, and public reporting. These tasks involve substantial up-front investments, which are more easily absorbed by large firms with an established ESG infrastructure and internal compliance capacity (ERM, 2022; OECD, 2021b). Consistent with this interpretation, firm size varies systematically across green debt instruments (Figure 7). Issuers of green bonds (including CBI-certified, CBI-aligned, sustainability-linked, or self-labeled green bonds) are the largest participants in corporate debt markets. They are typically larger than conventional bond issuers, all types of green loan issuers, and conventional loan issuers. By contrast, green loan issuers, especially those relying on self-labeled instruments, are significantly smaller, though still larger than conventional loan issuers. Within the loan segment, sustainability-linked loan issuers are larger than those issuing green-labeled loans, consistent with the added compliance burden associated with setting and tracking performance targets. A similar hierarchy is evident in the volume of funds raised per transaction. One exception is self-labeled green bonds, which are relatively low-volume transactions issued by otherwise large firms. The evolution of green debt issuances sheds light on how large incumbent market participants use and can benefit from the rise of a new type of instrument. On the one hand, conditions in the early stages of the green debt markets, when there was a narrower understanding of the characteristics of these instruments, less standardization, and fewer participants in their origination and certification, could have resulted in higher entry costs that benefited large incumbent firms. On the other hand, the lack of standardization and bureaucratization of the certification process at earlier stages could have resulted in lower entry costs. The results show that the relation between green debt issuance and firm size gets tighter over time (Table 6).17 The results show that, before 2018, firm size is only weakly associated with green issuance. However, the relation becomes significantly stronger in the 2018–2023 period, consistent with the growing dominance of larger firms as green markets mature. 17 To construct the table, we re-estimate Equation 1 for distinct subperiods. 17 This supports the view that the standardization of green activities through taxonomies, improvements in the certification process, and increased interest of institutional investors in this type of instrument might have benefited access to green debt markets by larger firms, relative to earlier stages of the market. 5 Performance Because green instruments are explicitly designed to support environmentally beneficial activities, a first-order question is whether their issuance is associated with observable differences in firm behavior relative to conventional debt. We therefore examine how firms evolve following the issuance of green and conventional instruments. We also distinguish between bonds and loans, given that differences between green and conventional borrowing can partly reflect the market through which financing takes place. The empirical strategy proceeds in two steps. First, we estimate firm-level changes in total debt, assets, operating income, and carbon intensity following different types of debt issuance. Changes in total debt indicate whether green borrowing is used to expand financing or to alter its composition. These choices, in turn, shape firm scale as reflected in assets and income. Together with carbon intensity, scale determines the trajectory of firms’ total emissions, since reductions in emissions per unit of output could be offset—or even outweighed—by post-issuance expansion in firm activity. Second, we aggregate firm-level responses using observed green debt issuance patterns to approximate the contribution of green debt to global emissions trends. This aggregation reflects not only firm-level performance but also the scale of participating firms and the patterns of issuance across markets. 5.1 Firm-level Outcomes To examine how firm performance evolves following debt issuance, we implement the LP-DiD approach. This method estimates dynamic treatment effects while addressing biases that arise in staggered estimation settings, particularly the use of already-treated units as controls. To avoid this problem, we impose a “clean control” condition: we exclude from the estimation any firm-year observation where the firm issues the same type of debt in the four years before 18 the treatment (debt issuance) year. This ensures that control observations have not been recently treated, and thus serve as a valid control group. The treated observations correspond to different types of debt issuance, as described below. We estimate the following specification for each outcome variable: LP-DiD h h yi,t+h − yi,t−1 = βh ∆Di,t + θc,t + θs,t + ϵi,c,s,t , (2) where yi,t+h denotes the log of the firm i’s outcome (total debt, assets, operating income, and carbon intensity) h years after issuance. The treatment indicator ∆Di,t is equal to one if the firm issues the relevant type of debt in year t, and zero otherwise. Importantly, the estimation is restricted to a sample in which no firm has issued the same type of debt in the four years before year t.18 This restriction applies to both treated and control units and ensures that control observations are not contaminated by residual effects from prior treatments. h θ h ) control for time-varying country and Country-year and sector-year fixed effects (θc,t s,t LP-DiD captures the cumulative post-treatment industry-specific shocks. The coefficient βh change relative to the baseline year t − 1. This framework enables structured comparisons of firm trajectories following different financing choices. We assess whether green debt issuance is associated with distinct post- issuance dynamics relative to two benchmarks: (i) firms without new debt issuance in those years, and (ii) firms with conventional debt issued in those years. The regression results show that, relative to non-issuance periods, green debt issuance is associated with significant firm-level expansion in total balance sheet debt and size (Figure 8, Panel A). Total debt rises by about 15 percentage points (p.p.) in the year of issuance and the two years that follow, accompanied by increases of roughly 10 p.p. in assets and operating income. These dynamics are evident across bond and loan markets, though the magnitude differs. Green bond issuance is followed by stronger expansion of balance sheet variables, with total debt rising by 20 p.p. and operating income by 13 p.p., whereas green loans are associated with more moderate increases of around 10 p.p. in debt and 5 p.p. in income. 18 We adopt a four-year cleaning window to align with the event-study horizon, which considers outcomes up to four years before and after issuance. Results are robust to using a five-year window (the median maturity of corporate debt) and a three-year window. 19 Beyond firm growth, green debt is also consistently followed by reductions in carbon intensity (emissions per unit of operating income). Relative to non-issuance periods, carbon intensity falls by 25 p.p. by year three and 45 p.p. by year four (Figure 8, Panel A). These improvements are robust across both bonds and loans: firms issuing either instrument experience statistically significant and sustained declines in carbon intensity (Panels B and C). By contrast, conventional debt issuance is associated with firm growth but not with improvements in carbon intensity (Figure 9). Following the issuance of conventional bonds or loans, firms increase total debt by approximately 25 p.p., and assets and operating income by up to 13 p.p. However, carbon intensity remains broadly unchanged—or increases slightly—over the same period, indicating that conventional borrowing is not systematically linked to post-issuance decarbonization. Direct comparisons between green debt (the treated group) and conventional debt (the control group) underscore these differences (Figure 10). Because both types of borrowing are associated with statistically similar increases in debt and income, the LP-DiD estimations show no clear pattern after the green debt issuance.19 However, environmental outcomes diverge significantly. Green debt is systematically associated with reductions in carbon intensity, with the gap relative to conventional debt reaching about 30 p.p. by year three and 50 p.p. by year four. These differences are even more pronounced than those observed when comparing green debt to non-issuance periods, highlighting the distinct environmental alignment of green borrowing relative to conventional financing.20 Lastly, we explicitly compare responses to green bond and green loan issuances to assess whether the two instruments generate systematically different outcomes (Figure 11). Bond issuance is associated with significantly larger increases in outstanding debt and firm size than loan issuance, indicating a stronger association between bonds and post-issuance expansion. By contrast, the reductions in carbon intensity following green borrowing do not 19 Green loans are linked to somewhat smaller firm expansion than their conventional counterparts. 20 Appendix Figures 5 and 6 decompose carbon intensity trends into Scope 1 and Scope 2 emissions. The results show that the post-issuance decline in overall intensity is primarily driven by significant reductions in Scope 1 emissions, while Scope 2 exhibits a noisier pattern with weaker evidence of sustained declines. Taken together, this suggests that firms are adjusting their production processes to reduce direct emissions more efficiently, while indirect emissions linked to purchased energy witness less systematic changes. 20 differ significantly between the two instruments. Overall, these results suggest the distinction between bonds and loans is most relevant for firm growth trajectories, while environmental outcomes are statistically similar across the two types of green debt. These results are robust to several sensitivity checks (Appendix Figure 7). First, restricting the sample to green debt issuers enables comparisons within those firms and yields estimates similar to the main results. Second, we construct a propensity score-matched (PSM) sample of green and conventional debt issuers based on average firm-level income and carbon intensity across the sample period.21 The post-issuance trajectories observed in the green issuer and PSM samples closely mirror those in the full data, and in some cases yield slightly larger estimates, with carbon intensity falling by 40–60 p.p. by year four relative to the two counterfactual comparisons (no debt and conventional debt). Third, results are also consistent across different types of debt contracts: both use-of-proceeds instruments and sustainability-linked loans are followed by sustained declines in carbon intensity. Together, these patterns indicate that green debt is systematically associated with environmental improvements across samples and instruments. 5.2 Aggregate Outcomes To assess the macro-level implications of green debt, we aggregate firm-level responses using observed issuance patterns alongside the dynamic effects estimated in the previous section. This exercise combines three elements of the analysis: the volume and timing of green debt issuances, the distribution of green debt issuances across firms, and the post-issuance performance. Together, these components allow us to estimate the total change in corporate carbon emissions associated with green debt. We project each firm’s emissions path by applying our LP-DiD estimates of post- issuance changes in carbon intensity and operating income to the firm’s initial scale and emissions profile. We then aggregate these projections to calculate the associated changes in total emissions at the global level. 21 Each green debt issuer is matched to the nearest-neighbor conventional issuer using average operating income and carbon intensity over the entire sample window. The matched sample includes 1,149 green issuers, down from the full sample of 3,845. 21 Formally, firm-level emissions can be expressed as the product of carbon intensity and firm size: Carbon Emissions Carbon Emissions i,t = Firm Size i,t . (3) Firm Size i,t Letting ci,t denote carbon intensity and si,t denote firm size (measured by operating income), total emissions across firms at time t are: Ct = ci,t si,t . (4) i The change in aggregate emissions over the post-issuance horizon h can then be decomposed as: ∆Ct+h = [(∆ci,t+h si,t ) + (∆si,t+h ci,t ) + (∆ci,t+h ∆si,t+h )] , (5) i where ∆ci,t+h and ∆si,t+h represent changes in carbon intensity and firm size, respectively. The decomposition separates aggregate emissions changes into three parts: (i) reductions due to lower carbon intensity (efficiency gains), (ii) increases due to firm growth (scale effects), and (iii) their interaction. This structure provides a transparent and additive mapping from firm-level debt dynamics to economy-wide climate outcomes. Aggregation Procedure. Building on the decomposition in Equation (5), we quantify the macro-level emissions impact of green debt by applying the LP-DiD estimated changes in carbon intensity (βc,h ) and firm size (βs,h ) to the baseline characteristics of green debt issuers in each year t, allowing us to simulate aggregate emissions outcomes.22 In implementing Equation (5), we apply the pooled LP-DiD estimate for carbon intensity, as firm-level trajectories do not differ significantly between green bonds and green loans. For firm size adjustments, we use instrument-specific coefficients, which reflect systematic differences in post-issuance changes in scale between bond and loan markets. We consider the same two counterfactuals used to estimate firm-level performance: (i) green debt issuance relative to no issuance, and (ii) green debt versus conventional debt issuance. The projected change in aggregate emissions h years after green debt issuance is 22 To ensure complete firm-level information for the aggregation exercise, we impute missing baseline characteristics using the median value of each variable by industry and instrument type (bond or loan). This approach preserves heterogeneity across firm types while maintaining the representativeness of the issuer sample. 22 given by: t ∆ Ch = [(βc,h ci,t−1 ) si,t + (βs,h si,t−1 ) ci,t + (βc,h ci,t−1 ) (βs,h si,t−1 )] Ii,t , (6) i where si,t and ci,t denote firm-level carbon intensity and size, and Ii,t is an indicator equal to one if firm i issued green debt in year t. The first two terms capture the marginal effects of changes in intensity and scale; the final term captures their interaction. This formulation reflects the cumulative emissions change attributable to green debt issued in year t, evaluated h periods later. To compute the total impact over time, we sum across issuance cohorts. Let t0 be the baseline year and T the final year, assuming treatment effects persist for H years. Then: T −H H −1 t T −h CT − Ct0 = ∆CH + ∆ Ch , t=t0 h=1 where the first term sums fully realized H -year effects, and the second term captures partial effects for more recent issuances. Substituting Equation (6) into this expression yields: T −H CT − Ct0 = (βc,H ci,t−1 ) si,t + (βs,H si,t−1 ) ci,t t=t0 i + (βc,H ci,t−1 ) (βs,H si,t−1 ) Ii,t H −1 + (βc,h ci,T −h−1 ) si,T −h + (βs,h si,T −h−1 ) ci,T −h h=1 i + (βc,h ci,T −h−1 ) (βs,h si,T −h−1 ) Ii,T −h . (7) More generally, for any given calendar year t, the cumulative emissions change since baseline year t0 is: t−t0 Ct − Ct0 = (βc,h ci,t−h−1 ) si,t−h + (βs,h si,t−h−1 ) ci,t−h h=1 i + (βc,h ci,t−h−1 ) (βs,h si,t−h−1 ) Ii,t−h . (8) This structure enables forward-looking simulations of aggregate emissions under different green debt issuance scenarios and translates firm-level behavioral effects into cumulative macroeconomic outcomes. Aggregate Results. Our analysis indicates that green debt issuance during 2012-23 is associated with substantial reductions in aggregate carbon emissions relative to the no-issuance and conventional-issuance scenarios. 23 The results show that cumulative emissions reductions associated with green debt issuance amount to approximately 4.5 billion metric tonnes in the no-issuance scenario, and 5.7 billion metric tonnes in the conventional-issuance scenario by 2025 (Figure 12). For comparison, global energy-related CO2 emissions reached a record 37.4 billion metric tonnes in 2023, according to the International Energy Agency (IEA, 2024). Thus, our estimates correspond to roughly 12 to 15 percent of one year’s global emissions, underscoring the macro relevance of firm-level financing choices in these markets. The largest issuers drive a disproportionate share of the aggregate effect. In our sample, the top quartile of firms accounts for approximately 85 percent of total estimated abatement, reflecting their high baseline emissions and dominant role in issuance volumes. While approximate, these estimates provide a transparent and scalable benchmark for assessing the global climate impact of green debt markets. Both green bonds and green loans contribute substantially to the abatement of aggregate emissions. Green bonds are associated with cumulative reductions of 2.8 billion metric tonnes under the no-issuance scenario and 3.6 billion metric tonnes under the conventional-issuance scenario, while green loans contribute 1.6 billion and 2.1 billion metric tonnes, respectively. These differences reflect both firm-level dynamics and issuer composition. Larger expansions in firm size follow green bond issuances, partially offsetting the gains from reduced carbon intensity relative to a no-debt benchmark (Figure 13, Panel A). In contrast, green loans generate emissions reductions more directly through sustained improvements in carbon intensity, coupled with more modest firm expansion (Figure 13, Panel B). Because a common carbon intensity trajectory is applied across instruments, the larger aggregate contribution of green bonds primarily reflects the composition of issuing firms.23 Green bonds are more frequently issued by large firms with high baseline emissions, whose post-issuance adjustments exert outsized influence on aggregate outcomes. As a result, the distribution of issuer size across instruments plays a central role in shaping the macro relevance of green debt. 23 Results are broadly consistent when applying instrument-specific carbon intensity trajectories rather than the pooled estimate. Although bond-specific coefficients are moderately larger, issuer composition remains the dominant driver of the greater aggregate abatement linked to green bonds. Using a common trajectory thus yields a conservative estimate of bond-financed impact. 24 6 Conclusion This paper analyzed the expansion of green corporate bond and syndicated loan markets around the world, focusing on how firms use green financing relative to conventional debt. We examined which types of firms issue different debt instruments and how they perform after they raise green funding, focusing on firm size and carbon intensity. We constructed an aggregation framework to complement the firm-level analysis to quantify the macro-level climate impact associated with green debt issuance activity. We distinguished green debt overall versus conventional debt, as well as bonds versus syndicated bank loans. We documented a sharp expansion of green debt since 2018, contrasting with a broader deceleration in conventional corporate borrowing. The growth of green debt has been increasingly concentrated in syndicated loan markets and in Europe. Syndicated loans not only capture the largest share of green debt financing, but are also used by a broader and more diverse set of borrowers. Still, green debt issuance overall remains dominated by large incumbent firms with elevated baseline emissions, which are central to the climate transition due to their scale and outsized contribution to global emissions. In contrast to conventional debt, which is not systematically related to environmental improvements, green debt is associated with sustained post-issuance reductions in carbon intensity by firms. These differences are robust across multiple empirical specifications, including within-firm comparisons and matched control samples. Aggregating firm-level trajectories, we estimated that green debt issued between 2018 and 2023 is associated with cumulative CO2 reductions of 4.5 billion to 5.7 billion metric tonnes by 2025—equivalent to approximately 12 to 15 percent of yearly global energy-related emissions. The largest issuers drive a disproportionate share of this aggregate abatement: the top quartile of firms accounts for the vast majority of the estimated reductions in our sample. These results underscore the macro-relevance of firm-level financing decisions and the potential role of debt markets in supporting large-scale decarbonization. Whereas both green bonds and loans are associated with comparable reductions in carbon intensity, they contribute to emissions abatement through complementary channels. Bonds are prominently used by large firms with high baseline emissions. Thus, they have a 25 larger effect on CO2 reductions when their carbon intensity falls after issuance. However, the expansion of firms following bond issuance partly offsets those gains. Loans, by contrast, reach a broader and more diverse set of firms, sectors, and countries, including smaller borrowers and those in less developed capital markets. Firms that issue green loans also expand after issuance, but to a lesser extent than bond issuers, resulting in a smaller offsetting effect. Furthermore, the total volume of loan issuance exceeds that of green bonds. At the aggregate level, bond financing accounts for larger emissions reductions than loans. This outcome reflects the large size and high emissions of bond issuers, as well as the fact that they capture a large share of bond issuance activity. The similar proportional within-firm reduction in emissions after bond and loan issuances implies a larger contraction in the volume of CO2 for the larger bond issuers. Taken together, these findings suggest that the environmental effectiveness of green debt is linked to the incentives it generates for firms to change their polluting practices. Moreover, the characteristics of bond and loan markets shape which firms around the world use them, which is essential for estimating the contribution of green debt to global emissions abatement. A natural caveat with the results is the issue of endogeneity. Issuance could reflect pre-existing plans to undertake greener projects that would occur despite the green debt financing. Thus, our analysis does not allow us to claim that green debt per se causes firms to decarbonize. Rather, it shows that reductions in carbon intensity are systematically associated with firms raising funds through green debt, not conventional debt. Our decomposition between bonds and loans and the aggregation of firm-level changes also show how the type of financial instrument and the type of firms that use green debt could shape the relation between firms’ financing choices and climate outcomes. Lastly, our results raise several questions for future research and policy design. What role can taxonomies, verification standards, and regulatory frameworks play in enhancing the credibility and scaling of green debt? Why has Europe emerged as the global leader in green debt adoption? To the extent that sustainable finance continues to grow, understanding how contractual structures, firm behavior, and institutional environments interact can be essential to realizing its full potential in reducing carbon emissions. 26 References Abraham, F., J. J. Cortina, and S. L. Schmukler. 2021a. Growth of global corporate debt: Main facts and policy challenges. Oxford Research Encyclopedia of Economics and Finance . Abraham, F., J. J. Cortina, and S. L. Schmukler. 2021b. The rise of domestic capital markets for corporate financing: Lessons from East Asia. Journal of Banking & Finance 122, 105987. Acharya, V. V., Y. Amihud, and L. Litov. 2011. Creditor rights and corporate risk-taking. Journal of Financial Economics 102 (1), 150–166. Acharya, V. V., I. Engle, Robert F, and O. Wang. 2025. Strategic commitments to decarbonize: The role of large firms, common ownership, and governments. NBER Working Paper 33335. Acharya, V. V. and G. Plantin. 2025. Monetary easing, leveraged payouts, and lack of investment. Management Science , 10.1287/mnsc.2022.01440. Acharya, V. V., G. Plantin, P. Reggiani, and I. Yao. 2024. Monetary easing, lack of investment and financial instability. Journal of Financial Intermediation 59, 101100. Albuquerque, B., E. Cerutti, N. Chen, and M. Firat. 2025. From banks to nonbanks: Macroprudential and monetary policy effects on corporate lending. IMF Working Paper 25/96. Aldasoro, I., S. Doerr, and H. Zhou. 2022. Non-bank lenders in the syndicated loan market. BIS Quarterly Review, March, 15–29. Aleszczyk, A. A., M. Loumioti, and G. Serafeim. 2022. The issuance and design of sustainability-linked loans. Working paper, SSRN 4287295. APLMA, LMA, and LSTA. 2019. Sustainability linked loan principles. Aramonte, S., A. Schrimpf, and H. S. Shin. 2023. Chapter 7: Non-bank financial intermediaries and financial stability. Research Handbook of Financial Markets , 147–170. Bachelet, M. J., L. Becchetti, and S. Manfredonia. 2019. The green bonds premium puzzle: The role of issuer characteristics and third-party verification. Sustainability 11 (4). Baker, M., D. Bergstresser, G. Serafeim, and J. Wurgler. 2022. The pricing and ownership of US green bonds. Annual Review of Financial Economics 14, 415–437. Boot, A. W. A. and A. V. Thakor. 2000. Can relationship banking survive competition? Journal of Finance 55 (2), 679–713. Bruno, V. and H. S. Shin. 2017. Global dollar credit and carry trades: A firm-level analysis. Review of Financial Studies 30 (3), 703–749. Caramichael, J. and A. Rapp. 2024. The green corporate bond issuance premium. Journal of Banking and Finance 162, 107126. CDP. 2023. The carbon majors database: 2023 update. Cicchiello, A., M. Corugno, S. Monferr` a, and S. Perdichizzi. 2022. Which are the factors influencing green bonds issuance? Evidence from the european bonds market. Finance Research Letters 50, 103190. CPI. 2023. Global landscape of climate finance 2023. 2025-02-07. DataFisher. 2024. ESG regulations: Comparing the EU and US approaches. Demski, J., Y. Dong, P. McGuire, and B. Mojon. 2025. Growth of the green bond market and greenhouse gas emissions. BIS Quarterly Review, March, 53–71. Diamond, D. W.. 1984, 07. Financial intermediation and delegated monitoring. Review of Economic Studies 51 (3), 393–414. Draghi, M.. 2024. The future of European competitiveness. Report, European Commission. 27 Du, K., J. Harford, and D. D. Shin. 2023. Who benefits from sustainability-linked loans? European Corporate Governance Institute—Finance Working Paper, No 917/2023. Dube, A., D. Girardi, O. ` Jord` a, and A. M. Taylor. 2025. A local projections approach to difference-in-differences event studies. NBER Working Paper 31184. Dursun-de Neef, , S. Ongena, and G. Tsonkova. 2023. Green versus sustainable loans: The impact on firms’ ESG performance. Swiss Finance Institute Research Paper, No. 22–42. Dutordoir, M., J. Quariguasi, and S. Li. 2024. When green is no longer a win - new evidence on the shareholder wealth effects of green bond offerings. Working paper, SSRN 4287295. Ehlers, T., B. Mojon, and F. Packer. 2020. Green bonds and carbon emissions: Exploring the case for a rating system at the firm level. BIS Quarterly Review . September, 31–47. ERM. 2022. Costs and benefits of climate-related disclosure activities by corporate issuers and institutional investors. Sustainability Institute. European Commission. 2024. Sustainability-related disclosure in the financial services sector. Fatica, S. and R. Panzica. 2021. Green bonds as a tool against climate change? Business Strategy and the Environment 30 (5), 2688–2701. Flammer, C.. 2021. Green bonds as a tool against climate change? Journal of Financial Economics 142 (2), 499–516. Gatev, E. and P. E. Strahan. 2009. Liquidity risk and syndicate structure. Journal of Financial Economics 93 (3), 490–504. GFMA and BCG. 2020. Climate finance markets and the real economy. Technical report, Global Financial Markets Association and Boston Consulting Group. Hachenberg, B. and D. Schiereck. 2018. Are green bonds priced differently from conventional bonds? Journal of Asset Management 19 (6), 371–383. Holmstrom, B. and J. Tirole. 1997, 08. Financial intermediation, loanable funds, and the real sector. Quarterly Journal of Economics 112 (3), 663–691. ICMA. 2020a. Bond market size and statistics: Global developments in primary and secondary markets. International Capital Market Association. ICMA. 2020b. Sustainability-linked bond principles: Voluntary process guidelines for issuing sustainability-linked bonds. International Capital Market Association. ICMA. 2021. Green bond principles: Voluntary process guidelines for issuing green bonds. International Capital Market Association. IEA. 2021. Net zero by 2050: A roadmap for the global energy sector. IEA. 2024. CO2 emissions in 2023 – Analysis. IPCC. 2018. Summary for policymakers. In Global Warming of 1.5°C: An IPCC Special Report. Cambridge University Press, 3–24. Ivashina, V. and D. Scharfstein. 2010. Loan syndication and credit cycles. American Economic Review 100 (2), 57–61. Lam, P. and J. Wurgler. 2024. Green bonds: New label, same projects. NBER Working Papers 32960. Merler, S.. 2025. How to improve the European Union’s sustainable finance framework. Policy brief, Bruegel. Morgan Stanley. 2025. Sustainable fund returns lag peers’ in second half of 2024. Morgan Stanley Institute for Sustainable Investing. OECD. 2021a. Public financial institutions and the low-carbon transition: Five case studies on low-carbon infrastructure and project investment. OECD Environment Working Papers, No. 72. OECD. 2021b. Scaling up green, social, sustainability and sustainability-linked bond issuances in developing countries. OECD Development Perspectives, No. 11. 28 OECD. 2025. Global debt report 2025: Financing growth in a challenging debt market environment. Rajan, R. G.. 1992. Insiders and outsiders: The choice between informed and arm’s-length debt. Journal of Finance 47 (4), 1367–1400. Stern, N.. 2006. The Economics of Climate Change: The Stern Review. Cambridge University Press. Tang, D. Y. and Y. Zhang. 2020. Do shareholders benefit from green bonds? Journal of Corporate Finance 61, 101427. Tuhkanen, H. and G. Vulturius. 2022. Are green bonds funding the transition? Investigating the link between companies’ climate targets and green debt financing. Journal of Sustainable Finance & Investment 12 (4), 1194–1216. United Nations. 2023. Principles for responsible banking: Key statistics. Zerbib, O. D.. 2019. The effect of pro-environmental preferences on bond prices: Evidence from green bonds. Journal of Banking and Finance 98, 39–60. 29 Figure 1. Global Corporate Debt Issuance over Time A. Debt Amounts Issued in Green and Conventional Markets 800 7,000 700 6,000 600 5,000 USD Billion 500 USD Billion 4,000 400 3,000 300 2,000 200 100 1,000 0 0 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 Green Debt Conventional Debt (RHS) B. Green Debt over Total Debt 14% 12% 10% 8% Percent 6% 4% 2% 0% 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 Green Debt over Total Debt This figure shows trends in green and conventional debt issuances from 2012 to 2023. Panel A reports total issuance volumes in billions of 2011 U.S. dollars (USD). Panel B shows the share of green debt issued per year as a percentage of total debt issuance (green plus conventional). “RHS” denotes the right-hand side axis. 30 Figure 2. Global Corporate Bond and Syndicated Loan Issuance over Time A. Debt Amounts Issued in Green and Conventional Markets Corporate Bonds Syndicated Loans 600 4,000 600 4,000 3,500 3,500 500 500 3,000 3,000 400 400 2,500 2,500 USD Billion USD Billion USD Billion USD Billion 300 2,000 300 2,000 1,500 1,500 200 200 1,000 1,000 100 100 500 500 0 0 0 0 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 12 13 14 15 16 17 18 19 20 21 22 23 Green Loans Excluding Revolving Green Bonds Conventional Bonds (RHS) Conventional Loans (RHS) 31 B. Green Debt over Total Debt Share of Green Bonds Share of Green Loans 14% 14% 12% 12% 10% 10% 8% 8% Percent Percent 6% 6% 4% 4% 2% 2% 0% 0% 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 12 13 14 15 16 17 18 19 20 21 22 23 Green Bonds over Total Bonds Green Loans over Total Loans This figure shows trends in green and conventional debt issuances from 2012 to 2023, separately for bonds (left-hand side) and syndicated loans (right-hand side). Panel A reports total issuance volumes in billions of 2011 U.S. dollars (USD). Panel B shows the share of green bonds and syndicated loans issued per year as a percentage of total bond and loan issuances (green plus conventional). “RHS” denotes the right-hand side axis. Figure 3. Regional and Sectoral Patterns of Green Debt Issued A. Distribution of Green and Conventional Debt across Regions and Sectors Regions Sectors 55% 50% 50% 45% Share of Total Green Debt Issued Share of Total Green Debt Issued 45% 40% 40% Share of Total Conventional Debt Issued Share of Total Conventional Debt Issued 35% 35% 30% Percent Percent 30% 25% 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% Latin Other China East Other United Europe Aggriculture Renewable Cons- Fossil Trade Services Manu- Utilities America Emerging Asia Advanced States and Energy truction Energy facturing Mining 32 B. Green Debt over Total Debt Regions Sectors 65% 65% Europe Renewable Energy 60% 60% Latin America Utilities 55% 55% East Asia Manufacturing 50% 50% Other Emerging Trade 45% 45% Other Advanced Agriculture and Mining 40% 40% United States Services Percent 35% Percent China 35% 30% Construction 30% 25% Fossil Energy 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 12 13 14 15 16 17 18 19 20 21 22 23 This figure shows the allocation of green and conventional debt issuances across regions (left-hand side) and industries (right-hand side). Panel A reports the share of each region and industry in global green and conventional debt issuance volumes raised from 2012 to 2023. Panel B shows, for each region and industry, the share of green debt issued per year as a percentage of total debt issuance (green plus conventional). Figure 4. Hybrid and Pure Green Debt Issuers A. Distribution of Green Debt across Types of Issuers 100% Share of Total Green Bonds Issued 90% Share of Total Green Loans Issued 80% Share of Total Green Debt Issued 70% 60% Percent 50% 40% 30% 20% 10% 0% Pure Green Hybrid B. Green Debt over Total Debt for Hybrid Issuers 50% Green Debt over Total Debt 45% Green Loans over Total Loans 40% Green Bonds over Total Bonds 35% 30% Percent 25% 20% 15% 10% 5% 0% 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 This figure shows green debt issuance patterns for hybrid and pure green issuers. Hybrid issuers are defined as firms that issue both green and conventional debt, while pure green issuers exclusively issue green debt. Panel A reports the share of each type of issuer in global green debt issuance volumes raised from 2012 to 2023. Panel B reports, for hybrid issuers, the share of green debt per year as a percentage of total debt issuance (green plus conventional). 33 Figure 5. Debt across the Issuer Size Distribution A. Amount Raised 100% 0.8 90% 0.7 80% 0.6 70% Carbon Emissions 60% 0.5 Percent 50% 0.4 40% 0.3 30% 0.2 20% 10% 0.1 0% 0 1st 2nd 3rd 4th Firm Size (Assets Quartile) Share of Total Conventional Debt Share of Total Green Debt Carbon Emissions (RHS) B. Number of Issuances 100% 0.8 90% 0.7 80% 0.6 70% 0.5 Carbon Emissions 60% Percent 50% 0.4 40% 0.3 30% 0.2 20% 10% 0.1 0% 0 1st 2nd 3rd 4th Firm Size (Assets Quartile) Share of Total Conventional Debt Share of Total Green Debt Carbon Emissions (RHS) This figure shows the distribution of green and conventional debt issued from 2012 to 2023 across the firm size distribution of issuers. Firms are grouped into quartiles based on their average total assets over the period. Panel A presents the distribution in terms of total debt volume raised, while Panel B shows the distribution based on the number of debt transactions. Each panel also reports the median carbon emissions for firms in each quartile on the right-hand side (RHS). Carbon emissions are expressed in millions of metric tonnes. 34 Figure 6. Bonds and Syndicated Loans across the Issuer Size Distribution A. Amount Raised Corporate Bonds Syndicated Loans 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% Percent Percent 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 1st 2nd 3rd 4th 1st 2nd 3rd 4th Firm Size (Assets Quartile) Firm Size (Assets Quartile) Share of Total Conventional Bonds Share of Total Green Bonds Share of Total Conventional Loans Share of Total Green Loans 35 B. Number of Issuances Corporate Bonds Syndicated Loans 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% Percent Percent 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 1st 2nd 3rd 4th 1st 2nd 3rd 4th Firm Size (Assets Quartile) Firm Size (Assets Quartile) Share of Total Conventional Bonds Share of Total Green Bonds Share of Total Conventional Loans Share of Total Green Loans This figure shows the distribution of green and conventional debt issued from 2012 to 2023 across the firm size distribution of bond and loan issuers. Firms are grouped into quartiles based on their average total assets over the period. Panel A presents the distribution in terms of total debt volume raised, while Panel B shows the distribution based on the number of debt transactions. Each panel distinguishes between bonds (left-hand side) and syndicated loans (right-hand side). Figure 7. Median Issuer Size across Debt Instruments 7,000 350 Issuer Size 6,000 300 Issuance Size (RHS) Issuance Size (USD Million) Issuer Size (USD Million) 5,000 250 4,000 200 3,000 150 2,000 100 1,000 50 - 0 Conventional Green-Labeled Conventional Sust.-Linked CBI-Aligned CBI-Certified Self-Labeled Sust.-Linked Loans Loans Bonds Loans Bonds Bonds Green Bonds Bonds This figure illustrates the median size of issuers—measured by total assets—and the median size of issuances across different types of green debt instruments. Issuance sizes are shown on the right-hand side (RHS). 36 Figure 8. LP-DiD Outcomes: Green Debt versus No Debt A. Green Debt vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 30 10 0 20 20 20 -10 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 10 -20 0 0 0 -30 -40 -10 -10 -10 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Green Bonds vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 30 10 20 20 20 0 -10 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 10 -20 37 0 0 0 -30 -40 -10 -10 -10 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) C. Green Loans vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 10 30 0 20 20 20 -10 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 10 -20 0 0 -30 0 -40 -10 -10 -10 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green debt issuances to periods without any debt issuance. Panel B compares outcomes around green bond issuances to periods without any debt issuance. Panel C compares outcomes around green loan issuances to periods without any debt issuance. The dependent variables are the log of balance sheet debt, assets, operating income, and carbon intensity, measured as differences from the year before issuance (t = –1). Figure 9. LP-DiD Outcomes: Conventional Debt versus No Debt A. Conventional Debt vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 20 30 10 20 20 20 0 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 -10 10 -20 0 0 0 -30 -40 -10 -10 -10 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Conventional Bonds vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 30 20 25 10 20 20 20 15 0 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 10 -10 38 5 -20 0 0 0 -30 -5 -40 -10 -10 -10 -15 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) C. Conventional Loans vs. No Debt Outstanding Debt Assets Income Carbon Intensity 30 30 30 20 10 20 20 20 0 Percentage Points Percentage Points Percentage Points Percentage Points 10 10 -10 10 -20 0 0 0 -30 -40 -10 -10 -10 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around conventional debt issuances to periods without any debt issuance. Panel B compares outcomes around conventional bond issuances to periods without any debt issuance. Panel C compares outcomes around conventional loan issuances to periods without any debt issuance. The dependent variables are the log of balance sheet debt, assets, operating income, and carbon intensity, measured as differences from the year before issuance (t = –1). Figure 10. LP-DiD Outcomes: Green Debt versus Conventional Debt A. Green Debt vs. Conventional Debt Outstanding Debt Assets Income Carbon Intensity 20 20 20 20 15 15 15 10 10 10 10 0 Percentage Points Percentage Points Percentage Points Percentage Points 5 5 5 -10 0 0 0 -20 -5 -5 -5 -30 -10 -10 -10 -40 -15 -15 -15 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Green Bonds vs. Conventional Bonds Outstanding Debt Assets Income Carbon Intensity 20 20 20 20 15 15 15 10 10 10 10 0 Percentage Points Percentage Points Percentage Points Percentage Points 5 5 5 -10 39 0 0 0 -20 -5 -5 -5 -30 -10 -10 -10 -40 -15 -15 -15 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) C. Green Loans vs. Conventional Loans Outstanding Debt Assets Income Carbon Intensity 20 20 20 20 15 15 15 10 10 10 10 0 Percentage Points Percentage Points Percentage Points Percentage Points 5 5 5 -10 0 0 0 -20 -5 -5 -5 -30 -10 -10 -10 -40 -15 -15 -15 -50 -20 -20 -20 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green debt issuances to periods around conventional debt issuances. Panel B compares outcomes around green bond issuances to periods around conventional bond issuances. Panel C compares outcomes around green loan issuances to periods around conventional loan issuances. The dependent variables are the log of balance sheet debt, assets, operating income, and carbon intensity, measured as differences from the year before issuance (t = –1). Figure 11. LP-DiD Outcomes: Bonds versus Loans A. Green Bonds vs. Green Loans Outstanding Debt Assets Income Carbon Intensity 40 40 20 20 15 15 30 30 10 10 20 20 Percentage Points Percentage Points Percentage Points Percentage Points 5 5 10 10 0 0 -5 0 0 -5 -10 -10 -10 -10 -15 -20 -20 -15 -20 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Conventional Bonds vs. Conventional Loans 40 Outstanding Debt Assets Income Carbon Intensity 40 10 10 20 15 30 5 5 10 20 Percentage Points Percentage Points Percentage Points Percentage Points 5 10 0 0 0 -5 0 -5 -5 -10 -10 -15 -20 -10 -10 -20 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green bond issuances to periods around green loan issuances. Panel B compares outcomes around conventional bond issuances to periods around conventional loan issuances. The dependent variables are the log of balance sheet debt, assets, operating income, and carbon intensity, measured as differences from the year before issuance (t = –1). Figure 12. Green Debt and Global Carbon Emissions Green Debt vs. No Debt Green Debt vs. Conventional Debt 0.5 0.5 0.0 0.0 -0.5 -0.5 -1.0 -1.0 -1.5 -1.5 Tonnes, Billions Tonnes, Billions -2.0 -2.0 -2.5 -2.5 -3.0 -3.0 -3.5 -3.5 -4.0 -4.0 -4.5 -4.5 41 -5.0 -5.0 -5.5 -5.5 -6.0 -6.0 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 0 0 1 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 8 9 0 Green Bonds vs. No Debt Green Bonds vs. Conventional Bonds Green Loans vs. No Debt Green Loans vs. Conventional Loans Total Total This figure presents the aggregate estimates of changes in total carbon emissions associated with green debt issuance. It compares two counterfactual scenarios: green debt versus no debt issuance (left-hand side), and green debt versus conventional debt issuance (right-hand side). Figure 13. Aggregate Carbon Emissions’ Decomposition A. Green Bonds Green Bonds vs. No Debt Green Bonds vs. Conventional Bonds 1.5 1.5 1.0 1.0 0.5 0.5 0.0 0.0 -0.5 -0.5 Tonnes, Billions Tonnes, Billions -1.0 -1.0 -1.5 -1.5 -2.0 -2.0 -2.5 -2.5 -3.0 -3.0 -3.5 -3.5 -4.0 -4.0 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Intensity Scale Cross-correlation Total Intensity Scale Cross-correlation Total 42 B. Green Loans Green Loans vs. No Debt Green Loans vs. Conventional Loans 1.0 1.0 0.5 0.5 0.0 0.0 Tonnes, Billions Tonnes, Billions -0.5 -0.5 -1.0 -1.0 -1.5 -1.5 -2.0 -2.0 -2.5 -2.5 -3.0 -3.0 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Intensity Scale Cross-correlation Total Intensity Scale Cross-correlation Total This figure presents aggregate estimates of changes in total carbon emissions associated with green debt issuances, decomposed into intensity, scale, and cross-correlation effects. It compares two counterfactual scenarios: green debt versus no debt issuances (left-hand side), and green debt versus conventional debt issuances (right-hand side). Panel A shows the aggregate decomposition for corporate bonds. Panel B shows the aggregate decomposition for syndicated loans. Table 1. Summary Statistics of Green and Conventional Debt Issuances Share of Share of Total Median Median Type of Debt Debt Raised Green- Sustainability- No. of Size Maturity Instrument (USD, Bn) Labeled Linked Transactions (USD, Mn) (Years) Instruments Instruments Conventional Debt 120,711 61,600 149 5.0 . . 43 Bonds 36,799 21,900 229 6.0 . . Loans 83,912 39,700 133 5.0 . . Green Debt 6,412 2,496 200 5.2 42% 58% Bonds 2,255 745 192 6.0 88% 12% Loans 4,157 1,751 205 5.0 30% 70% This table shows summary statistics for conventional and green debt transactions from 2012 to 2023. Dollar amounts are expressed in 2011 U.S. dollars (USD). The units used are: Mn = million and Bn = billion. Table 2. Summary Statistics of Green and Conventional Debt Issuers Debt Issuance Activity Firm-Level Characteristics Number of Debt Issued Balance Sheet Carbon Type of Debt Carbon Issuers (USD, Tn) (USD, Bn) Emissions Issuer Intensity Share Fixed (Tonnes, Total Total Conventional Green Debt Assets Income - Listed Assets Mn) Pure Conventional 46,987 21% 48.1 48.1 . 0.25 0.93 0.25 0.70 0.11 50.8 44 Green 3,845 37% 16.0 13.5 2.5 1.31 4.19 1.32 2.81 0.38 89.6 Hybrid Green 2,503 52% 15.6 13.5 2.1 1.53 4.91 1.52 3.24 0.42 92.7 Pure Green 1,342 8% 0.4 0.0 0.4 0.15 0.55 0.21 0.35 0.08 65.6 This table shows summary statistics for conventional and green debt issuers for 2012–2023. It displays the total number of issuers, the share of publicly listed firms, and the total amount of conventional and green debt issued by type of firm. For firm-level balance sheet and carbon emission characteristics, the table presents the median values. Hybrid issuers are defined as firms that issue both green and conventional debt, whereas pure green issuers exclusively issue green debt. Debt issuance and balance sheet values are expressed in 2011 U.S. dollars (USD). Carbon emissions are expressed in metric tonnes. The units used are: Mn = million, Bn = billion, Tn = trillion. Table 3. Summary Statistics of Green and Conventional Debt Issuers in Different Markets A. Bonds Bond Issuance Activity Firm-Level Characteristics Number of Bond Issued Balance Sheet Carbon Carbon Type of Bond Issuers (USD, Tn) (USD, Bn) Emissions Intensity Issuer Share Fixed (Tonnes, Total Total Conventional Green Debt Assets Income Listed Assets Mn) Pure Conventional 12,878 38% 16.4 16.4 . 0.68 2.12 0.62 1.40 0.25 67.4 Green 1,334 49% 6.3 5.5 0.7 2.42 7.36 2.88 4.85 0.79 130.2 Hybrid Green 1,107 56% 6.2 5.5 0.7 2.72 8.27 3.14 5.36 0.86 133.8 Pure Green 227 14% 0.1 0.0 0.1 0.17 0.54 0.13 0.32 0.09 76.1 B. Loans 45 Loan Issuance Activity Firm-Level Characteristics Number of Loan Issued Balance Sheet Carbon Carbon Type of Loan Issuers (USD, Tn) (USD, Bn) Emissions Intensity Issuer Share Fixed (Tonnes, Total Total Conventional Green Debt Assets Income Listed Assets Mn) Pure Conventional 38,808 20% 34.9 0.0 . 0.25 0.96 0.26 0.78 0.10 47.7 Green 2,773 35% 6.54 1.8 1.8 1.01 3.50 1.06 2.55 0.29 65.7 Hybrid Green 1,652 54% 6.24 4.8 1.4 1.18 3.87 1.16 2.94 0.31 66.9 Pure Green 1,121 7% 0.31 0.0 0.3 0.13 0.56 0.22 0.43 0.03 45.3 This table shows summary statistics for conventional and green debt issuers for 2012–2023. It displays the total number of issuers, the share of publicly listed firms, and the total volume of conventional and green debt issued, disaggregated by type of firm. For firm-level balance sheet and carbon emission characteristics, the table presents the median values. Hybrid issuers are defined as firms that issue both green and conventional debt, whereas pure green issuers exclusively issue green debt. Debt issuance and balance sheet values are expressed in 2011 U.S. dollars (USD). Carbon emissions are expressed in metric tonnes. The units used are: Mn = million, Bn = billion, Tn = trillion. Table 4. Firm Characteristics around Green versus Conventional Debt Issuances Dependent Variable: Dummy = 1 if a firm issues green debt in a given year and 0 if it issues conventional debt Base Value: 0.037 Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Fixed Effects: Beta N Beta N Beta N Beta N Beta N No 0.012∗∗∗ 95,064 0.020∗∗∗ 28,492 0.017∗∗∗ 28,531 0.012∗∗∗ 16,304 0.010∗∗∗ 16,788 46 Industry-Time 0.017∗∗∗ 94,962 0.020∗∗∗ 28,383 0.017∗∗∗ 28,425 0.016∗∗∗ 16,222 0.014∗∗∗ 16,712 ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ Industry- and Country-Time 0.017 94,958 0.018 28,367 0.017 28,408 0.012 16,180 0.005∗ 16,671 This table reports linear probability regression estimates of the likelihood of issuing green debt, conditional on debt issuances. The dependent variable is a binary indicator equal to one if a firm issues green debt in a given year and zero if it issues conventional debt. Independent variables are lagged log values of firm-level characteristics. Standard errors are clustered at the country-year level. Asterisks (*, **, ***) denote statistical significance at the 10%, 5%, and 1% levels, respectively. Table 5. Firm Characteristics around Green versus Conventional Bond and Loan Issuances A. Green Bonds vs. Conventional Bonds Dependent Variable: Dummy = 1 if a firm issues green bonds and 0 if it issues conventional bonds Base Value: 0.072 Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Fixed Effects: Beta N Beta N Beta N Beta N Beta N No 0.011∗∗∗ 29,241 0.019∗∗∗ 11,896 0.012∗∗∗ 12,090 0.011∗∗∗ 7,848 0.012∗∗∗ 8,189 Industry-Time 0.011∗∗∗ 29,124 0.017∗∗∗ 11,826 0.008∗∗∗ 12,020 0.014∗∗∗ 7,769 0.014∗∗∗ 8,107 ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ Industry- and Country-Time 0.012 29,084 0.014 11,781 0.009 11,978 0.007 7,685 0.001 8,023 B. Green Loans vs. Conventional Loans Dependent Variable: Dummy = 1 if a firm issues green loans and 0 if it issues conventional loans Base Value: 0.051 Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Fixed Effects: Beta N Beta N Beta N Beta N Beta N 0.010∗∗∗ 69,590 0.016∗∗∗ 19,820 0.015∗∗∗ 19,781 0.008∗∗∗ 11,328 0.008∗ 47 No 11,546 Industry-Time 0.014∗∗∗ 69,484 0.016∗∗∗ 19,710 0.016∗∗∗ 19,671 0.011∗∗∗ 11,222 0.008∗∗∗ 11,450 ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ Industry- and Country-Time 0.015 69,480 0.015 19,666 0.016 19,629 0.011 11,173 0.005 11,403 C. Green Bonds vs. Green Loans Dependent Variable: Dummy = 1 if a firm issues green bonds and 0 if it issues green loans Base Value: 0.371 Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Fixed Effects: Beta N Beta N Beta N Beta N Beta N No 0.056∗∗∗ 4,439 0.106∗∗∗ 1,670 0.045∗∗∗ 1,691 0.052∗∗∗ 1,678 0.038∗∗∗ 1,675 Industry-Time 0.060∗∗∗ 4,321 0.086∗∗∗ 1,599 0.053∗∗∗ 1,616 0.051∗∗∗ 1,611 0.045∗∗∗ 1,599 ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ Industry- and Country-Time 0.067 4,239 0.083 1,510 0.065 1,527 0.038 1,535 0.008 1,533 This table reports linear probability regression estimates of the likelihood of issuing green debt, conditional on debt issuances. In Panel A, the dependent variable equals one if a firm issues green bonds and zero if it issues conventional bonds in a given year. In panel B, the dependent variable equals one if a firm issues green loans in a given year and zero if it issues a conventional loan. In Panel C, the dependent variable equals one if a firm issues green bonds in a given year and zero if it issues green loans. Independent variables are lagged log values of firm-level characteristics. Standard errors are clustered at the country-year level. Asterisks (*, **, ***) denote statistical significance at the 10%, 5%, and 1% levels, respectively. Table 6. Linear Probability Regressions over Time A. Green Debt vs. Conventional Debt Dependent Variable: Dummy = 1 if a firm issues green debt in a given year and 0 if it issues conventional debt Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Period Base Beta N Beta N Beta N Beta N Beta N 2012-18 0.007 0.003*** 27,054 0.003*** 6,365 0.002** 6,320 0.00 3,152 -0.001 3,713 2018-20 0.045 0.020*** 17,813 0.023*** 4,267 0.020*** 4,272 0.019*** 3,449 0.004 3,453 2021-23 0.141 0.052*** 19,462 0.072*** 4,117 0.067*** 4,094 0.034*** 3,713 0.013* 3,707 B. Green Bonds vs. Conventional Bonds Dependent Variable: Dummy = 1 if a firm issues green bonds in a given year and 0 if it issues conventional bonds Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Period Base Beta N Beta N Beta N Beta N Beta N 2012-18 0.013 0.004*** 7,516 0.004* 2,970 0.002 2,953 0.00 1,593 0.001 2,032 2018-20 0.064 0.015*** 4,939 0.018*** 1,817 0.007 1,849 0.010** 1,904 -0.005 1,911 2021-23 0.170 0.049*** 5,027 0.060*** 1,607 0.042*** 1,595 0.019*** 1,969 0.000 1,953 C. Green Loans vs. Conventional Loans Dependent Variable: Dummy = 1 if a firm issues green loans in a given year and 0 if it issues conventional loans 48 Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Period Base Beta N Beta N Beta N Beta N Beta N 2017-18 0.004 . . . . . . . . . . 2019-20 0.035 0.014*** 13,740 0.012*** 3,108 0.015*** 3,124 0.014*** 2,416 0.007 2,402 2021-23 0.129 0.043*** 15,194 0.058*** 3,071 0.058*** 3,074 0.027*** 2,572 0.016** 2,556 D. Green Bonds vs. Green Loans Dependent Variable: Dummy = 1 if a firm issues green bonds in a given year and 0 if it issues green loans Independent Variable: Issuance Size Total Assets Income Carbon Emissions Carbon Intensity Period Base Beta N Beta N Beta N Beta N Beta N 2017-18 . . . . . . . . . . . 2019-20 0.431 0.085** 588 0.029 150 -0.014 147 0.002 158 -0.054 159 2021-23 0.341 0.064*** 2,327 0.112*** 692 0.080** 684 0.041*** 770 -0.001 755 This table reports linear probability regression estimates of the likelihood of issuing green debt over time, conditional on debt issuance. Regressions are estimated separately by period and include country-industry fixed effects. In Panel A, the dependent variable equals one if a firm issues green bonds in a given year and zero if it issues conventional bonds. In Panel B, the dependent variable equals one if a firm issues green loans and zero if it issues conventional loans. In Panel C, the dependent variable equals one if a firm issues green bonds and zero if it issues green loans. Independent variables are lagged log values of firm-level characteristics. Standard errors are clustered at the country-industry level. Asterisks (*, **, ***) denote statistical significance at the 10%, 5%, and 1%, respectively. Appendix Figure 1. Firm Size and Carbon Emissions Firm Size and Issuance Size Firm Size and Carbon Emissions 20 10 Log of Avarage Issuance Size Log of CO2 Emissions 15 5 10 0 49 5 -5 0 -5 0 5 10 15 -5 0 5 10 15 Log of Total Assets Log of Total Assets Conventional Issuers Green Issuers Conventional Issuers Green Issuers This figure illustrates the relation between firm size and carbon emissions over 2012–2023. The left panel shows the correlation between firm size (log of total assets) and debt issuance size (log of total funds raised per issuance). The right panel plots the relation between firm size and carbon emissions (in logs). Each point represents a firm-level average over 2012–2023. Appendix Figure 2. Green Debt Instruments over Time A. Green Bond Instruments 250 CBI-Aligned Bonds Sustainability-Linked Bonds 200 Self-Labeled Green Bonds No. of Transactions Sustaibability Bonds 150 CBI-Certified Bonds 100 50 0 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 B. Green Loan Instruments 600 Sustainability-Linked Loans (Revolver) 500 Sustainability-Linked Loans (Term) Green-Labeled Loans No. of Transactions 400 300 200 100 0 20 20 20 20 20 20 20 20 20 20 20 20 12 13 14 15 16 17 18 19 20 21 22 23 This figure depicts the time trends in green debt instruments. Panel A shows the annual number of different types of green bond transactions, while Panel B displays the corresponding trends for green syndicated loan transactions. 50 Appendix Figure 3. Bond and Loan Distribution across Regions and Industries A. Corporate Bonds Regions Sectors 60% 60% Share of Total Green Debt Issued Share of Total Green Debt Issued 50% 50% Share of Total Conventional Debt Issued Share of Total Conventional Debt Issued 40% 40% Percent Percent 30% 30% 20% 20% 10% 10% 0% 0% Latin Other China East Other United Europe Aggriculture Renewable Cons- Fossil Trade Services Manu- Utilities America Emerging Asia Advanced States and Energy truction Energy facturing Mining B. Syndicated Loans 51 Regions Sectors 60% 60% Share of Total Green Debt Issued Share of Total Green Debt Issued 50% 50% Share of Total Conventional Debt Issued Share of Total Conventional Debt Issued 40% 40% Percent Percent 30% 30% 20% 20% 10% 10% 0% 0% Latin Other China East Other United Europe Aggriculture Renewable Cons- Fossil Trade Services Manu- Utilities America Emerging Asia Advanced States and Energy truction Energy facturing Mining This figure shows the allocation of green and conventional bonds and loans across regions and industries. Panel A reports the share of each region (left-hand side) and industry (right-hand side) in global green and conventional bond issuance volumes raised from 2012 to 2023. Panel B shows the percentage share of each region and industry in global green and conventional syndicated loan issuance volumes raised over the same period. Appendix Figure 4. Issuer Size across Regions and Industries A. Debt Issuer Size per Region 12,000 Conventional Debt Issuer 10,000 Green Debt Issuer 8,000 USD Million 6,000 4,000 2,000 0 Europe United Other China East Other Latin States Advanced Asia Emerging America B. Debt Issuer Size per Industry 9,000 Conventional Debt Issuer 8,000 Green Debt Issuer 7,000 6,000 USD Million 5,000 4,000 3,000 2,000 1,000 0 Utilities Aggriculture truction Energy facturing Services Renewable Trade Fossil Cons- Manu- Energy Mining and This figure shows the median size of debt issuers in conventional and green markets, disaggregated by region (Panel A) and industry (Panel B). Issuer size is measured in terms of total assets, in millions of 2011 U.S. dollars (USD). 52 Appendix Figure 5. Carbon Scopes 1 and 2: Green Debt versus No Debt A. Green Debt vs. No Debt Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Green Bonds vs. No Debt Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) C. Green Loans vs. No Debt Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green debt issuances to periods without any debt issuance. Panel B compares outcomes around green bond issuances to periods without any debt issuance. Panel C compares outcomes around green loan issuances to periods without any debt issuance. The dependent variables are carbon intensity Scope 1 and Scope 2, measured as differences from the year before issuance (t = −1). 53 Appendix Figure 6. Carbon Scopes 1 and 2: Green Debt versus Conventional Debt A. Green Debt vs. Conventional Debt Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Green Bonds vs. Conventional Bonds Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) C. Green Loans vs. Conventional Loans Carbon Intensity Scope 1 Carbon Intensity Scope 2 30 30 20 20 10 10 0 0 Percentage Points Percentage Points -10 -10 -20 -20 -30 -30 -40 -40 -50 -50 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green debt issuances to periods around conventional debt issuances. Panel B compares outcomes around green bond issuances to periods around conventional bond issuances. Panel C compares outcomes around green loan issuances to periods around conventional loan issuances. The dependent variables are carbon intensity Scope 1 and Scope 2, measured as differences from the year before issuance (t = −1). 54 Appendix Figure 7. LP-DiD Green Debt Estimates: Carbon Intensity across Different Samples A. Green Debt vs. No Debt Green Debt Issuers PSM Sample Green-Labeled Sustainability-Linked 20 20 20 20 10 10 0 0 0 0 -20 Percentage Points Percentage Points Percentage Points Percentage Points -20 -10 -10 -40 -20 -20 -40 -30 -30 -60 -40 -40 -60 -80 -50 -50 -100 -80 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) B. Green Debt vs. Conventional Debt 55 Green Debt Issuers PSM Sample Green-Labeled Sustainability-Linked 20 20 20 20 0 10 10 0 0 0 -20 -20 Percentage Points Percentage Points Percentage Points Percentage Points -10 -10 -40 -40 -20 -20 -60 -30 -30 -60 -80 -40 -40 -80 -100 -50 -50 -100 -120 -60 -60 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 -5 -4 -3 -2 -1 0 1 2 3 4 Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) Difference relative to t = -1 Confidence Interval (90% CI) This figure presents the estimated beta coefficients and 90 percent confidence intervals from LP-DiD estimations of Equation 2. Panel A compares firm-level outcomes around green debt issuances to periods without any debt issuance, while Panel B compares outcomes around green versus conventional debt issuance. The dependent variable is the log of carbon intensity, measured as the difference from the year before issuance (t = –1). The different charts show results across subsamples: all green debt issuers, a propensity-score-matched (PSM) sample, green-labeled instruments, and sustainability-linked instruments. The PSM sample matches green and conventional debt issuers based on average income and carbon intensity levels. Appendix Table 1. Green and Conventional Debt by Country Green Debt Conventional Debt Share Green Debt Conventional Debt Share Country Total Share Share Total Share Share of Country Total Share Share Total Share Share of Volume of of Volume of of Green Volume of of Volume of of Green Raised Bonds Loans Raised Bonds Loans Debt Raised Bonds Loans Raised Bonds Loans Debt United States 508,471 25% 75% 28,809,927 28% 72% 2% Philippines 1,895 61% 39% 97,015 40% 60% 2% France 245,798 31% 69% 2,001,954 38% 62% 11% Czechia 1,870 74% 26% 64,312 25% 75% 3% United Kingdom 179,972 16% 84% 2,561,999 28% 72% 7% Peru 1,813 69% 31% 47,484 36% 64% 4% Germany 172,695 19% 81% 1,830,765 20% 80% 9% Iceland 1,708 0% 100% 5,022 15% 85% 25% Spain 144,242 17% 83% 682,671 18% 82% 17% Ghana 1,465 0% 100% 30,871 0% 100% 5% Netherlands 139,939 48% 52% 1,206,649 52% 48% 10% Viet Nam 1,417 0% 100% 41,748 11% 89% 3% China 130,664 94% 6% 7,810,397 86% 14% 2% Mauritius 1,278 0% 100% 11,413 24% 76% 10% Italy 127,297 30% 70% 717,399 31% 69% 15% Lithuania 1,056 64% 36% 3,937 20% 80% 21% Japan 97,765 43% 57% 3,023,276 27% 73% 3% Egypt, Arab Rep. 1,044 0% 100% 39,270 0% 100% 3% Canada 83,259 19% 81% 2,990,543 21% 79% 3% Colombia 1,038 35% 65% 71,440 40% 60% 1% Taiwan, China 54,705 11% 89% 360,069 27% 73% 13% Lao PDR 847 33% 67% 4,460 37% 63% 16% Sweden 46,966 26% 74% 385,314 29% 71% 11% Kenya 743 0% 100% 4,115 0% 100% 15% Switzerland 43,000 7% 93% 964,296 14% 86% 4% Qatar 689 0% 100% 29,395 35% 65% 2% Singapore 42,879 5% 95% 404,967 17% 83% 10% Ukraine 685 100% 0% 17,759 19% 81% 4% Australia 42,760 18% 82% 927,015 22% 78% 4% Georgia 562 74% 26% 2,658 52% 48% 17% Denmark 39,470 37% 63% 158,115 21% 79% 20% Congo, Dem. Rep. 461 0% 100% 1,043 0% 100% 31% Norway 33,614 32% 68% 317,513 26% 74% 10% Panama 446 51% 49% 17,453 33% 67% 2% Hong Kong SAR, China 30,295 33% 67% 431,916 28% 72% 7% Slovak Republic 409 0% 100% 16,083 20% 80% 2% 56 Finland 29,384 31% 69% 168,274 23% 77% 15% Uruguay 399 67% 33% 6,187 27% 73% 6% India 28,629 30% 70% 575,406 25% 75% 5% Oman 375 0% 100% 46,175 2% 98% 1% Mexico 27,203 43% 57% 416,851 51% 49% 6% Latvia 374 100% 0% 669 29% 71% 36% Belgium 25,330 13% 87% 318,560 27% 73% 7% Nigeria 371 0% 100% 46,874 2% 98% 1% Brazil 18,925 57% 43% 427,082 61% 39% 4% Romania 328 0% 100% 13,187 9% 91% 2% Ireland 18,668 28% 72% 360,134 21% 79% 5% Morocco 309 0% 100% 7,900 52% 48% 4% Saudi Arabia 17,934 11% 89% 268,296 12% 88% 6% Malta 268 0% 100% 11,191 31% 69% 2% Portugal 16,106 37% 63% 47,126 29% 71% 25% Dominican Republic 247 100% 0% 2,294 15% 85% 10% Luxembourg 15,367 44% 56% 425,237 55% 45% 3% Costa Rica 247 100% 0% 3,543 24% 76% 7% United Arab Emirates 15,101 12% 88% 344,526 9% 91% 4% Uzbekistan 203 0% 100% 9,953 8% 92% 2% Poland 14,628 10% 90% 83,190 9% 91% 15% Senegal 190 0% 100% 2,280 7% 93% 8% Chile 11,899 57% 43% 129,479 46% 54% 8% Bulgaria 81 0% 100% 6,003 9% 91% 1% Thailand 9,400 47% 53% 224,962 69% 31% 4% Croatia 76 0% 100% 15,518 14% 86% 0% Austria 8,893 23% 77% 107,997 34% 66% 8% Cyprus 70 0% 100% 7,229 27% 73% 1% New Zealand 7,435 36% 64% 117,991 18% 82% 6% Zambia 67 0% 100% 3,206 0% 100% 2% South Africa 6,807 0% 100% 133,088 7% 93% 5% Kazakhstan 52 0% 100% 36,157 34% 66% 0.1% Russian Federation 6,419 21% 79% 346,911 41% 59% 2% Serbia 51 0% 100% 4,593 13% 87% 1% Indonesia 5,715 23% 77% 245,540 29% 71% 2% Ecuador 40 0% 100% 2,282 0% 100% 2% Greece 5,012 38% 62% 47,516 13% 87% 10% Lesotho 35 0% 100% 168 0% 100% 17% Bahrain 4,514 0% 100% 16,647 11% 89% 21% Albania 34 0% 100% 81 0% 100% 29% Malaysia 3,831 44% 56% 155,764 62% 38% 2% Mongolia 22 0% 100% 12,523 8% 92% 0.2% T¨urkiye 3,774 44% 56% 123,782 12% 88% 3% Cambodia 15 100% 0% 2,988 15% 85% 0.5% Argentina 2,309 72% 28% 49,503 52% 48% 4% Bangladesh 0.3 0% 100% 12,700 2% 98% 0.0% Estonia 2,002 0% 100% 3,715 35% 65% 35% Israel 1,974 0% 100% 73,579 11% 89% 3% Hungary 1,921 29% 71% 36,843 9% 91% 5% This table reports total volumes of green and conventional debt issued between 2012 and 2023. All amounts are expressed in millions of 2011 U.S. dollars (USD). Appendix Table 2. Types of Green Debt Instruments A. Corporate Bonds Median Total Total Median Type of Bond Transaction Green Sust. No. of Yearly Debt Raised Maturity Instrument Size Labeled Linked Transactions (USD, Bn) (Years) (USD, Mn) CBI-Aligned 912 363 212 6.0 ✓ CBI-Certified 84 27 218 7.0 ✓ Sustainability 176 67 220 7.6 ✓ Self-Labeled Green 442 107 131 5.0 ✓ 57 Sustainibility-Linked 280 113 292 6.0 ✓ B. Syndicted Loans Median Total Total Median Type of Loan Transaction Green Sust. No. of Yearly Debt Raised Maturity Instrument Size Labeled Linked Transactions (USD, Bn) (Years) (USD, Mn) Green-Labeled 1,253 333 115 7.0 ✓ Sustainibility-Linked (Term) 1,240 397 137 5.0 ✓ Sustainibility-Linked (Revolver) 1,664 1,022 244 5.0 ✓ This table shows summary statistics for conventional and green debt transactions from 2012 to 2022. Dollar values are expressed in 2011 U.S. dollars (USD). The units used are: Mn = million, Bn = billion. Appendix Table 3. Top Green Debt Issuers Green Debt Issued Conventional Debt Issued Share of Top Company Name Country Industry (USD, Mn) (USD, Mn) Green Debt 1 Ford Motor Co. 42,277 116,718 27% United States Manufacturing 2 Engie SA 27,455 52,475 34% France Utilities 3 Alphabet Inc. 23,774 12,834 65% United States Services 4 RWE AG 21,636 24,148 47% Germany Utilities 5 Enel SpA 20,127 66,115 23% Italy Utilities 6 Airbus SE 18,255 30,434 37% France Manufacturing 7 TenneT Holding BV 16,550 5,169 76% Netherlands Utilities 8 ´ e de France SA EDF – Electricit´ 16,108 72,461 18% France Utilities 9 ENI SpA 15,749 26,028 38% Italy Fossil Energy 58 10 China Three Gorges Corp. 15,670 27,970 36% China Utilities 11 Iberdrola SA 14,595 21,339 41% Spain Utilities 12 Orsted A/S 13,979 5,687 71% Denmark Utilities 13 Pfizer Inc. 13,100 108,806 11% United States Manufacturing 14 Intel Corp. 12,665 59,543 18% United States Manufacturing 15 Siemens Energy AG 12,276 – 100% Germany Renewable Energy 16 Sanofi SA 12,102 58,752 17% France Manufacturing 17 EDP – Energias de Portugal SA 11,692 16,442 42% Portugal Utilities 18 E.ON SE 11,677 19,704 37% Germany Utilities 19 Crown Castle International Corp. 11,485 69,591 14% United States Utilities 20 Terna Rete Elettrica Nazionale SpA 10,562 9,514 53% Italy Utilities Rankings based on total volumes of green and conventional debt issued during 2012–2023. Amounts shown in millions of 2011 U.S. dollars (USD). Appendix Table 4. Top Lead Arrangers in Green and Conventional Syndicated Lending Top Green Lenders Top Conventional Lenders Green Debt Lent Conventional Debt Lent Top Bank Name Top Bank Name (USD, Mn) (USD, Mn) 1 BNP Paribas 500,472 1 JP Morgan 6,384,464 2 JP Morgan 142,922 2 Bank of America 4,445,685 3 Bank of America 78,283 3 BNP Paribas 3,298,894 4 ABN AMRO Bank 77,275 4 Citigroup 2,749,404 5 Australia & New Zealand Banking Group Ltd 47,577 5 Wells Fargo 1,788,545 6 Banco Santander SA 47,201 6 MUFG 1,187,601 7 Mizuho Bank Ltd 45,401 7 Barclays 1,042,085 8 MUFG 43,987 8 Credit Suisse AG 1,022,068 9 Citigroup 43,440 9 Deutsche Bank 794,150 59 10 Banco Bilbao Vizcaya Argentaria SA 43,107 10 Australia & New Zealand Banking Group Ltd 761,676 11 Credit Agricole 31,697 11 Mizuho Bank Ltd 754,301 12 Wells Fargo 28,850 12 Goldman Sachs & Co 681,954 13 Barclays 27,984 13 ABN AMRO Bank 626,064 14 CaixaBank SA 23,686 14 Sumitomo Mitsui 602,001 15 Canadian Imperial Bank of Commerce 21,465 15 Scotiabank 522,691 16 Sumitomo Mitsui 20,837 16 Morgan Stanley Group Inc 497,354 17 Agricultural Bank of China 19,460 17 Banco Bilbao Vizcaya Argentaria SA 482,547 18 HSBC 16,830 18 Bank of China Ltd 448,004 19 Bank of China Ltd 15,857 19 The Royal Bank of Canada 415,037 20 Deutsche Bank 15,195 20 Credit Agricole 322,596 Rankings are based on total syndicated loan volumes during 2012–2023 in which each bank participated as a lead arranger. Amounts are expressed in millions of 2011 U.S. dollars (USD). Reported figures reflect the full size of the syndicated loans, not the individual lending commitments of each bank.