The value effect of sustainability: evidence from Latin American ESG bond market

ABSTRACT

We use the event study methodology to examine the effect of 115 environmental, social, and governance (ESG) bond issuances on the stock price of Latin American listed firms over the period 2016–2022. Consistent with the signaling theory, firms sending a signal of commitment to sustainability obtain a positive and significant average Cumulative Abnormal Return (CAR) of 1.97% during an event window of 18 days. The CAR is higher for first-time issuers, reaching 2.28%. Firms with smaller and more diverse boards are associated with higher CARs. However, ownership concentration reduces the CARs due to the potential misbehavior and expropriation risk from controlling shareholders. These results suggest that firm-level corporate governance mechanisms are critical to the ESG bonds signaling effect. Overall, our findings indicate that investors in Latin American capital markets positively value the reduction of information asymmetries regarding firms’ sustainability efforts, especially in a firm that mitigates potential agency conflicts.

KEYWORDS:

• ESG bonds
• signaling theory
• information asymmetry
• stock prices
• Latin American capital market

1. Introduction

Since the World Bank issued the first green bond in 2008, ESG bonds have grown as instruments to fund sustainable projects required to meet the United Nations Sustainable Development Goals (SDGs). However, at the firm level, it is unclear what are the motivations behind these financing sources. For example, Flammer (2021) did not find evidence supporting the cost of capital argument, that is, the issuance of green bonds to get cheaper financing. Moreover, there is no evidence of a ‘green premium’ for green bonds, where investors buy them at a higher price (Larcker and Watts 2020; Vulturius, Maltais, and Forsbacka 2022). These findings, together with the additional issuing and external verification costs and limitation of the use of proceeds, suggest that this form of financing is costly. Flammer (2021) shows that the cost of issuing green bonds in the US is explained by the benefits firms get by signaling to the market their environmental commitment (signaling argument). Thus, several firms and governments in Latin America have been increasingly issuing these ESG debt instruments (Núñez, Velloso, and Da Silva 2022).

Latin American institutional context makes it challenging to use these instruments. This region is characterized by a high level of opaqueness, deep information asymmetries, and small capital markets (Chong and López-de-silanes 2007; Djankov et al. 2008; González et al. 2021b, 2021a; OECD 2018; Villarraga, Giraldo, and Agudelo 2012). On the one hand, the cost of capital of Latin American firms substantially increases due to the limitation of the use of proceeds and the extra auditing and verification costs, which represent an additional restriction in underdeveloped capital markets with high information asymmetries; but on the other hand, issuing firms in the region could also benefit the most by their required information disclosure and signaling commitment toward sustainability (Klapper and Love 2004).

The region is worth studying for several reasons. First, Klapper and Love (2004) argue that good corporate practices are a strong signal, especially in weak legal systems with high levels of information asymmetries, such as Latin America (Chong and López-de-silanes 2007). Second, emerging markets have shown stronger market reactions to corporate news (Guzmán et al. 2020; Morck, Yeung, and Yu 2013), which makes this context especially attractive to test the information content of ESG bond issues. Third, less opaque firms are able to find higher levels of external financing and tend to have better valuations in stock markets (Durnev and Kim 2005). Finally, the region requires special efforts toward sustainability due to its geographic, socioeconomic, environmental, economic, and demographic characteristics (World Meteorological Organization 2022).

In this study, we examine the effect of ESG bond issuances on the stock prices of listed firms in Argentina, Brazil, Chile, Colombia, Mexico, and Peru, the largest capital markets of Latin America. We use the event study methodology to analyze the impact of 115 ESG bond issuances on the stock price of 67 issuers. Consistent with the signaling theory, we find that Latin American capital markets have positive and significant reactions to the issuance with Cumulative Abnormal Returns (CAR) of 1.97%; the CARs are higher for first-time issuers showing an average of 2.28%. When using the Cumulative Average Abnormal Return (CAAR), we document a positive and significant reaction of 1.11% for the full sample and 1.13% for first time issuers. We also explore the determinants of these stock price reactions. In one respect, we found that firms with smaller and more diverse boards are associated with positive CARs. Contrarily, more concentrated ownership structures and the potential misbehavior and expropriation risk of controlling shareholders reduce the positive value effect of sustainable bond issuance. These results indicate that in the Latin American context, with high levels of information asymmetries and agency tensions, firm-level governance matters for firms to take advantage of the premium market participants grant to sustainable bond issuers. Therefore, our results suggest that investors value the signal of the commitment of the issuer firms and that the signal is more potent in firms with better firm-level corporate governance and lower ownership concentration.

This study contributes to the existing literature on several aspects. First, while previous studies have focused on the issuance of green bonds (Flammer 2021; Glavas 2019; Lebelle, Lajili Jarjir, and Sassi 2020), we include a set of social, sustainable, and sustainability-linked bonds that strengthens the analysis in the broader scope of ESG bonds. Second, in contrast to previous studies, we focus exclusively on Latin American capital markets known for large information asymmetries and represent an understudied region in this recent literature. There is a gap in the empirical research regarding the Latin American context, whereas, for example, Glavas (2019), Lebelle, Lajili Jarjir, and Sassi (2020), and Flammer (2021) include no more than two Latin American countries. Third, we document higher abnormal returns for Latin American ESG bond issuers than those in other event studies focused on developed economies. This is consistent with the higher sensitivity to corporate news in emerging markets (Morck, Yeung, and Yu 2013). Finally, we show that this signaling effect is more robust for firms where their corporate governance mechanisms and ownership structure reduce agency conflicts.

The remainder of the paper is organized as follows: In the next section, we review the literature and set up the context of our research questions. Next, we explain the evolution of the ESG bond market in Latin America. We then explain the methodology, describing the data and the statistical approach. We follow by presenting our results and discussing our findings. Finally, we conclude.

2. Literature review

Financial markets display asymmetric information in any transaction where one of the two parties implicated has more information than the other. These asymmetries can give rise to adverse selection on markets (Akerlof 1970) and increased agency conflicts between managers and outside investors (Healy and Palepu 2001; Jensen and Meckling 1976). Moreover, large information asymmetries reduce the capital flows to these markets due to higher monitoring costs (Leuz, Lins, and Warnock 2009). Spence (1973) demonstrated that well-informed firms could improve their market outcome under certain conditions by signaling their private information to poorly informed agents. For this reason, firms with good performance try to lessen information asymmetries by sending credible ‘signals’ to the market. According to the signaling theory, a signal is convincing if it is difficult to imitate by firms with inferior performance (Riley 1979; Spence 1973).

Consistent with signaling theory, Patel, Balic, and Bwakira (2002) argue that higher transparency and more explicit disclosure diminish the information asymmetry between a firm’s management and equity and bondholders, lessening agency problems (Ross 1977). Furthermore, there is evidence that disclosure practices can help firms grow by relaxing external financing constraints, thereby allowing capital to flow to positive net present value projects (Khurana, Pereira, and Martin 2006).

ESG bonds, financial instruments designed to promote sustainable goals, have become increasingly popular in recent years in part because of the finance sector’s role in allocating capital efficiently (Maltais and Nykvist 2021). However, the literature still seeks to understand the rationale for issuing these instruments and their implications because, except for the ‘sustainability’ component, there is limited evidence regarding ESG bond convenience in comparison to other bonds with similar characteristics. As proposed by Flammer (2021), one of these rationales can be explained through signaling theory for the following reasons: First, with the issuance of ESG bonds, firms dedicate larger amounts of money to sustainable projects. Second, third parties certify the use of ESG proceeds (Banga 2018; Zerbib 2019). Moreover, fulfilling the ESG bond principles (ICMA, 2021) involves considerable managerial efforts and resources, representing a substantial cost to the issuer and brand reputation alone may not be sufficient to help firms successfully issue these instruments as they may need superior corporate social responsibility performance in the form of high ESG scores (Avramov et al. 2022).

The external verification cost and the restriction in using proceeds make ESG bonds costly for issuing firms, creating a good opportunity to send credible signals regarding firms’ sustainability commitment to the market. Lyon and Maxwell (2011) state that a company’s commitment to sustainability is frequently made difficult to evaluate by investors and most projects require years to improve firms’ sustainability performance (Benlemlih, Jaballah, and Kermiche 2023). Moreover, the emergence of the ESG bond markets raises important issues in terms of governance. Who determines what an ESG bond is? Can all bonds be ESG? Who controls greenwashing and what can be done to prevent it? (Câmara and Morais 2022). This situation requires a credible signal to differentiate between firms committed to sustainability and those that are not. Therefore, by issuing ESG bonds, firms can signal their sustainable actions.

Previous research has shown that shareholders might react positively to firms’ commitment to sustainability and that ESG bonds can potentially play a major role in increasing economy’s sustainability without financially penalizing the issuers. For instance, Ge and Liu (2015) studied how a firm's corporate social responsibility (CSR) performance is associated with the cost of its new bond issues in the US market. They found that firms with better CSR performance were able to issue bonds at lower cost. Bauer and Hann (2012) analyzed a large cross-industrial sample of US public corporations and found that environmental concerns are associated with a higher cost of debt financing and lower credit ratings. Moreover, proactive environmental practices were associated with a lower cost of debt. Gianfrate and Peri (2019) found that green bonds were more convenient for the issuer than conventional bonds because, on average, they offered the investors returns 0.18% lower.

Consistent with the signaling argument, many event studies show positive abnormal returns in response to companies’ implementing green projects, thus improving their valuation. In North America, Flammer (2015) examined the effect of shareholder proposals related to corporate social responsibility (CSR) on market valuation and found that the stock market reacted positively to the adoption of close-call shareholder proposals advocating the pursuit of eco-friendly policies. In Europe, Baulkaran (2019) reported a CAR of 1.48% for firms issuing green bonds. The studies of Zhou and Cui (2019), Wang et al. (2020), and Verma and Bansal (2023) have also shown that Asian firms issuing green bonds had significant positive effects around 1% on their stock price after the issuance.

Some studies have focused on more than one continent. Tang and Zhang (2020) studied the response of stock prices to green bond announcements for firms in 28 countries and found a CAR of 1.39% for first-time issuers. This finding is consistent with Glavas (2019), who also reported positive stock responses of 0.46% for firms in 27 countries. Flammer (2021) revealed that investors responded positively to the issuance announcement with CARs of 0.49%, a stronger response for first-time issuers and bonds certified by third parties. The issuers also improved their environmental performance post-issuance and experienced increased ownership by long-term and green investors.

Previous literature had been concentrated on green bonds but muted on information content on a broader spectrum of ESG bonds. Furthermore, the Latin American region has been mostly absent from this discussion, with the exception of Glavas (2019), Lebelle, Lajili Jarjir, and Sassi (2020), and Flammer (2021), which include one or two Latin American countries with at most ten (10) observations.

In the Latin American context, this signaling effect could be valuable for market participants despite the associated issuing costs. For example, restricting the use of proceeds is especially costly in undeveloped capital markets, possibly affecting ESG bonds’ issuance size (Barua and Chiesa 2019). The third party’s verification and required reporting generate additional costs compared to ordinary bonds. However, these extra costs could be compensated given the credible signal firms committed to sustainability can send to the market (Klapper and Love 2004).

As explained above, conditions regarding the high level of information asymmetries could produce a stronger signal to the Latin American capital markets (Klapper and Love 2004), a more effective reaction to corporate news (Guzmán et al. 2020), and the valuation effect may differ from those in developed markets (Durnev and Kim 2005). Additionally, Núñez, Velloso, and Da Silva (2022) highlight that the corporate bond market in Latin America has experienced a significant boost. Hence, substantially increased information disclosure regarding ESG issues creates an opportunity to analyze the impact of these dynamics on firm value. In this context, we post the following hypothesis:

H1: There is a positive stock market reaction for firms issuing ESG bonds in Latin America.

As mentioned before, good governance practices are essential in opaque markets with high levels of information asymmetries, such as Latin America (Chong and López-de-Silanes 2008). Therefore, firms’ level governance mechanisms could be important determinants of the market reaction due to ESG bond issuances. Market participants receive information regarding corporate governance practices at the firm level and their commitment to sustainable practices, and they update their beliefs regarding the issuer’s long strategy, potentially impacting its market valuation.

Several authors have explored the effect of corporate governance characteristics on sustainability and firm value. For example, board size may improve decision-making processes, including sustainability decisions, by pushing forward these issues in the board’s agenda (Albitar et al. 2020). Some studies have shown that when participating in ESG practices, a larger board represents stakeholders more efficiently (Jizi et al. 2014). This finding is consistent with the position that a larger board might have more competent directors who can handle various issues, such as sustainability (Almaqtari et al. 2022). Moreover, de Villiers, Naiker, and van Staden (2011) study reveals a clear relationship between higher environmental performance and larger boards. Nevertheless, there is strong evidence that communication and coordination problems could affect the decision-making process in large groups (Jensen 1993). In finance, this negative size-effect was confirmed by Yermack (1996) and Bhagat and Black (2001) where they show that firms with bigger boards may destroy value. Their sample was composed of big US firms rising the question whether this side effect is negative also for smaller firms or firms in other institutional context (Gilson and Roe 1993). Eisenberg, Sundgren, and Wells (1998) show a strong negative relation between board size and firm value for a sample of smaller firms.

Other studies have reported a positive and significant relationship between board diversity and sustainability performance. Specifically, the participation of women on the board might be related to new perspectives on board discussions (Husted and Sousa-Filho 2019), higher ethical behavior in the use of information, and attention to investors (Adams, Licht, and Sagiv 2011; Matsa and Miller 2011), improved supervision of managers (Adams and Ferreira 2009), better relations with stakeholders (Hussain, Rigoni, and Orij 2018), and a sounder background in human sciences (García et al. 2023). In addition, previous literature has also demonstrated a positive impact of board gender diversity on firm performance (Post & Byron, 2015), suggesting the presence of women on boards as an efficient governance mechanism.

There is also evidence that companies with a more independent boards will likely to be related with more sustainability initiatives (Amran, Periasamy, and Zulkafli 2014; de Villiers, Naiker, and van Staden 2011; Giannarakis, Andronikidis, and Sariannidis 2019; Hu and Loh 2018; Liao, Luo, and Tang 2015). According to Husted & De Sousa-Filho (2019), independent directors understand the relevance of community, environmental, and other stakeholder interests and are likelier to promote better sustainable practices. Furthermore, as Ibrahim, Howard, and Angelidis (2003) note, independent directors are more involved in benevolent and philanthropic matters associated with corporate social responsibility than executive directors.

Other board characteristics, such as CEO duality, might harm firm propensity to take sustainability practices in favor of different stakeholders due to the agency conflict that arises when CEOs also lead the board that supervises him or her (Naciti 2019). In this agency context, firm's decisions are only sometimes geared to maximize value to all shareholders. However, it has been shown that CEO duality could also positively impact firm sustainability performance due to reputational and compensation incentives (Velte 2020).

Ownership concentration as other corporate governance mechanism is especially relevant in Latin America given the high concentrated ownership structure in the typical firm in this region. Previous literature stresses the agency tensions that arise from ownership concentration. The large shareholder may use its power to obtain what Grossman and Hart (1980) label ‘private benefits of control’ at the expense of the minority shareholders (Villalonga et al. 2019). Controlling shareholders also influence sustainability decisions, strategies, and performance, and they may focus more on sustainability practices to enhance the firm’s reputation in the long term (Duong et al. 2022; Lin and Nguyen 2022). However, the presence of large shareholders might also generate agency problems as they can use their power to reduce environmental and social sustainability practices that decrease the firm’s value (Chen et al. 2021).

In general, as supported by the previous evidence, corporate governance characteristics such as board size, independence, diversity, CEO duality, and ownership concentration could be determinants of how the market reacts to the value of Latin American firms when they issue ESG bonds. Consistent with this argument, we post the following hypothesis:

H2: Firms’ level corporate governance mechanisms are significant determinants of market reaction for firms issuing ESG bonds in Latin America.

2.1. Latin American ESG bonds market

The issuance of ESG bonds from Latin America in global capital markets began in 2014 when the Peruvian firm Energía Eólica issued a US $204 million green bond. As Figure 1 shows, the participation of Latin America in the global ESG bond market has increased since then, with the exception of 2018. In 2021 approximately one-third of the amounts raised by ESG bonds in the world came from the Latin American region. Regarding issuances, the Latin American region went from 0.2% in 2014 to almost 16% in 2021; this trend is expected to continue in the coming years (Bocquet et al. 2021; Núñez, Velloso, and Da Silva 2022). However, not all bond issuers trade their stocks in the capital markets, making it difficult to link the information impact of these issuances on the stock price.

Figure 1. Latin American corporate ESG bond issuances in international markets. Source: Núñez, Velloso, and Da Silva (2022).

ESG bond issuers have accepted the International Capital Markets Association (ICMA)’s principles as a standard for issuing ESG bonds. ICMA circulates these technical standards yearly (ICMA 2021). The Green Bond Principles (GBP), the Social Bond Principles (SBP), the Sustainability Bond Guidelines (SBG), and the Sustainability-Linked Bond Principles (SLBP) have become increasingly relevant in international capital markets, with numerous countries and corporations around the world using them as a reference to produce their own frameworks for the issuance of sustainable bonds (Núñez, Velloso, and Da Silva 2022). Furthermore, the credibility of management disclosures is enhanced by regulators, standard setters, auditors, and other capital market intermediaries (Healy and Palepu 2001). A description of ESG bonds is presented in Table 1.

Table 1. Types of ESG bonds.

Type of instrument	Description
Green bond	The proceeds are used toward projects and activities promoting a wide range of environmental targets.
	This category includes instruments such as climate, blue, and transition bonds.
Social bond	The proceeds are used to finance projects addressing or mitigating specific social issues to achieve a positive social impact.
	This category includes instruments such as COVID-response bonds and affordable-housing bonds.
Sustainability bond	Combines social and environmental targets.
Sustainability-linked bond	The coupon of the bond is tied to the achievement of sustainable targets.

Source: MSCI (2021).

In Latin America, a group of countries has issued green bond guidelines and sustainable bond frameworks to control the issuance of green, social, sustainability, and sustainability-linked bonds, taking the ICMA’s Guidelines and Principles as a reference (Argandoña, Rambaud, and Pascual 2022). Some have also taken as a guide the Helsinki Principles, which are the shared principles of the Coalition of Finance Ministers for Climate Action. The Helsinki Principles are aspirational and give a common purpose to member countries committed to taking collective and domestic action on climate change and achieving the Paris Agreement’s objectives (Núñez, Velloso, and Da Silva 2022). The adoption of standards in the international markets strengthened the signal to the market for the LAC issuers, given the cost associated with the compliance of these principles.

3. Methodology

3.1. Data

We use Bloomberg, EMIS, Refinitiv, sustainability reports, and stock exchange web pages to create a database of corporate ESG bonds issued in Argentina, Brazil, Chile, Colombia, Mexico, and Peru, the largest capital markets of Latin America. According to the World Bank¹, these six countries account for 84% of the total gross domestic product in purchasing power parity in the 32 Latin American and Caribbean countries. Table 2 summarizes the main characteristics of our sample. Panel A shows the construction of the sample. To identify ESG bonds, the ESG indicator in each database was marked as ‘YES’; this criterion generated a combined set of 429 issuances. Then, governments and non-listed companies were removed from the database, leaving a total of 145 corporate ESG bonds. This filter generated three outcomes: (i) firms that issued ESG bonds before being listed on a stock exchange, (ii) firms that were delisted and then issued the ESG bond, and (iii) firms that were listed on a stock exchange in the event window. Since outcomes (i) and (ii) could not be processed because there was insufficient data to compute Cumulative Abnormal Returns (CAR), only outcome (iii) was considered, leaving a total of 115 ESG bonds. Finally, first-time issuers were identified.

Table 2. Sample description.

Panel A. ESG bond issuances in selected countries
Total issuances	429
Issuances of listed firms	145
Sample issuances	115
First-time issuers	54

Panel B. ESG bond listed issuers per country
Country of issue	Number of issuers	Percentage
Argentina	5	7.46%
Brazil	35	52.24%
Chile	10	14.93%
Colombia	7	10.45%
Mexico	9	13.43%
Peru	1	1.49%
Total	67	100.00%

Panel C. Corporate ESG bonds by country and year
Country	2016	2017	2018	2019	2020	2021	2022	Subtotal
Argentina	0	0	0	0	1	1	3	5
Brazil	1	1	1	9	20	24	4	60
Chile	0	0	1	9	0	6	3	19
Colombia	1	1	4	1	4	6	0	17
Mexico	0	2	0	0	3	4	4	13
Peru	0	0	0	0	0	0	1	1
Subtotal	2	4	6	19	28	41	15	115

Panel D. Corporate ESG bond issuances in Latin America by type of bond.
Type of ESG bond	2016	2017	2018	2019	2020	2021	2022	Subtotal
1. Green	2	4	6	17	24	35	15	103
2. Social	0	0	0	1	2	3	0	6
3. Sustainability	0	0	0	1	2	0	0	3
4. Sustainability-linked	0	0	0	0	0	3	0	3
Subtotal	2	4	6	19	28	41	15	115

Table 2 Panel B summarizes the number of ESG bonds for listed issuers per country. Sixty-seven companies were identified, and Brazil leads the list with 35, representing 52.24% of our sample. Ten Chilean companies correspond to 14.93%, nine (9) Mexican issuers are 13.43%, seven (7) Colombian firms denote 10.45%, five (5) Argentinian corporations embody 7.46%, and one (1) Peruvian business exemplifies 1.49%. The evolution of ESG bonds in our sample is presented in Panel C of Table 2. Our sample covers bond issuances for listed firms from 2016 to 2022. Only two companies issued ESG bonds in 2016 and 15 in 2022. The highest number of issuers, 41, was reported in 2021.

The types of ESG bonds are described in Panel D, Table 2. 103 of 115 (89.56%) of the issuances are represented by green bonds whose proceeds were used for projects related to eco-efficiency, renewable energy, sustainable transportation, circular economy, energy transmission, and forest protection. Social bonds represent 5.22% of the sample (6 of 115 issuances), including water sanitization and affordable basic infrastructure projects. 2.61% of the sample (3 out of 115) is characterized by sustainability bonds linked to financial inclusion and renewable energy projects. Finally, 2.61% of the ESG bonds (3 out of 115) include sustainability-linked bonds whose proceeds were used for sustainable agriculture projects.

3.2. Cumulative abnormal return and event study

Finance theory suggests that capital markets reflect all available information about firms in the firms’ stock prices. Given this basic premise, it is possible to examine how a specific event modifies a company's prospects by measuring the event’s effect on the company's stock returns (Mackinlay 1997).

Return event studies measure the economic effect of an event in what is called an abnormal return. Abnormal returns are determined by subtracting the returns that would have been obtained if the researched event had not happened (normal returns) from the real returns of the stocks adjusted by dividend. The real returns are observed empirically. However, the expected returns are estimated using the window between t₁ = −120 and t₂ = −10 before the event [−120, – 10] using a normal distribution.

The market model is built based on the real returns of the reference market and the correlation of each stock with its reference market for the firm i in the day t. The abnormal return (AR_i,t) on each day in the event window symbolizes the difference between the real stock return (R_i,t) on that specific day and the expected return, which is projected based on two inputs; the typical relationship between the firm's stock and its reference index (parameters α and β), and the actual reference market's return (R_m,t). (1) $A{R}_{i,t}=R_{i,t}-({\propto }_i+{\beta }_i{R}_{m,t})$(1) The event study window of our study begins nine days before each bond issuance and ends nine days after the event [−9, 9]. To measure the impact of each issuance over the event window, the abnormal returns are added, and a cumulative abnormal return (CAR) is calculated. The CAR is widely used in corporate finance and governance literature to measure the impact of specific events on firms’ value over a short window of time (Strong, 1992). (2) $\mathrm{CAR}(-9,9)=\sum _{t=-9}^{9}A{R}_{i,t}$(2) To test across all events and calculate the CAR for all the companies as a group, a linear regression is performed for each day of the event window, beginning nine days before and ending nine days after the event [−9, 9]. The CAR of firm i in the event day t is presented in equation (3)(3) $\mathrm{CA}{R}_{i,t}=\propto +e_i$(3) . To differentiate the CAR between first-time and repeated issuers, a second set of regressions is developed using a dummy variable where 1 represents a first-time issuer and 0 otherwise. (3) $\mathrm{CA}{R}_{i,t}=\propto +e_i$(3) Where α represents the estimated coefficient for the cumulative abnormal return of firm i in day t, the p-value on the constant of each regression provides the significance of the CARs across all firms using robust standard errors.

3.3. Cumulative abnormal return determinants

Once the CARs are obtained, a linear regression analysis with robust standard errors is used to study CAR determinants. Equation (4)(4) $\mathrm{CA}{R}_i={\beta }_0+{\beta }_1{X}_i+e_i$(4) exhibits the linear regression form. (4) $\mathrm{CA}{R}_i={\beta }_0+{\beta }_1{X}_i+e_i$(4) Where: β₀ = Intercept. β_i = Coefficient that represents the effect of predictor variable i. X_i = independent variable i. e_i = error term.

All the financial, governance, and performance variables were extracted from EMIS, Refinitiv, and the annual reports of each firm. The description of firm characteristics is presented in Table 3. Following Harris and Raviv (2008) and Villalonga et al. (2019), we count the number of directors on the board and subtract the mean board size of our sample (nine directors) to obtain the Board size deviation from the mean and its square value (Board size deviation from the mean ²). Board Independence is estimated as the ratio of independent directors on the board as a fraction of board size (Rosenstein and Wyatt 1990). Board diversity is measured as the ratio of women on the board as a fraction of board size (González et al. 2020). Concerning CEO duality, as proposed by Boyd (1995), is a dummy variable equal to 1 if the CEO is also the president of the board and 0 otherwise.

Table 3. Variable description.

Variable	Description
CAR	Cumulative Abnormal Return
CAAR	Cumulative Average Abnormal Return
Board size	Number of directors on the Board
Board size deviation from the mean	The difference between the Board size and the average board size of the sample
Board Independence	Number of independent directors on the Board / Board size
Board Diversity	Number of female directors on the Board / Board size
CEO Duality	1 if CEO is President of the Board; 0 Otherwise
Rating	Bond rating. 10 if the bond is rated as AAA
Size	Natural logarithm of firm assets
Growth	Average percentage of assets growth / 100
Leverage	Debt / Assets
Largest Shareholder’s Equity Stake	Percentage of shares owned by largest shareholder / 100
ESG	ESG Score by Refinitiv
Private	1 if firm is controlled by private owners; 0 Otherwise
Financial	1 if firm is part of financial sector; 0 Otherwise
Energy	1 if firm is part of energy sector; 0 Otherwise
Industry	1 if firm is part of an industry; 0 Otherwise
Services	1 if firm is part of services; 0 Otherwise

Financial and firm characteristics are represented as follows. Bond rating (Rating) is measured as an ordinal variable for which a value of ten (10) represents a rating of AAA, AA + rated bonds as nine (9), and goes until four (4) equals CAA. Firm size (Size) is measured as the natural logarithm of total assets in thousands of dollars (Shalit and Sankar 1977). Asset growth (Growth) is the annual percentage change of firms’ assets (Titman, Wei, and Xie 2013). The measurement of firms’ debt (Leverage), as proposed by Lang, Ofek, and Stulz (1996), is represented as the ratio of short-term and long-term debt to the book value of total assets. Following Demsetz and Lehn (1985), ownership concentration is given by the percentage of common equity owned by the largest shareholder or the Largest Shareholder’s Equity Stake. We control by industry and use the ESG Score by Refinitiv (ESG) as another control variable.

3.4. Summary statistics

Table 4 indicates our variables’ mean, standard deviation, minimum, and maximum values. The average CAR obtained from the event study is 1.97%. Regarding the firm’s characteristics, the average size of the board of directors is 9, and the deviation from the mean goes from 4 (boards composed of 5 directors) to 11 (boards composed of 20 directors). The mean percentages of independent directors and women on the board are 43.3% and 18.6%, respectively. Only 4.3% of the firms are also Chairman of the Board (CEO duality). On average, the firms of our study have US $27,808.31 million assets (equal to the exponential value of 15.89), and those assets grew 4.9% with respect to the previous year of the ESG bond issuance. The mean leverage ratio is 66.8%, and the largest shareholder has, on average, 41.3% of the company. Finally, the mean ESG score for our sample was 60.05, ranging from a minimum of 16.81 to a maximum of 89.82.

Table 4. Descriptive statistics.

Variable	Obs	Mean	Std. Dev.	Min	Max
Dependent variable
Cumulative abnormal return (CAR)	115	0.0197	0.1019	−0.1911	0.5871
Cumulative Average Abnormal Return (CAAR)	115	0.0111	0.0554	−0.0945	0.2814
Independent variables
Board size	115	9.0435	2.9747	5	20
Board size deviation from the mean	115	0.0435	2.9747	−4	11
Board Independence	115	0.4337	0.2159	0	1
Board Diversity	115	0.1862	0.1107	0	0.4285
CEO Duality	115	0.0434	0.2048	0	1
Rating	115	8.0086	1.7696	4	10
Size	115	15.8899	1.5232	12.5718	19.6915
Growth	115	0.0493	0.1762	−0.2378	0.8664
Leverage	115	0.6686	0.1892	0.1278	0.9437
Largest Shareholder’s Equity Stake	115	0.4126	0.1758	0.0523	0.8770
ESG	115	60.0455	14.7175	16.8100	89.8200
Private	115	0.9130	0.2830	0	1
Financial	115	0.2435	0.4310	0	1
Energy	115	0.2261	0.4201	0	1
Industry	115	0.3391	0.4755	0	1
Services	115	0.1913	0.3950	0	1

4. Results

Table 5 summarizes the CARs for the total sample of 115 ESG bonds. The first column presents the days before and after the issuance. The second column shows the cumulative abnormal return coefficient for each day of the event window, and columns three, four, and five highlight the standard error, t-value, and statistical significance, respectively. Consistent with the signaling argument and hypothesis 1, a positive and significant CAR is found for different event windows. For example, we have a statistically significant CAR at a 5% level for the [−9, 9] event window. We also find statistically significant CARs at a 5% significance level for other event windows such as [−9, 2], [−9, 4], and [−9, 8].

Table 5. Cumulative Abnormal Return (CAR) regressions for [−9, 9] window.

Days before and after issuance	CAR	Std. Err.	t	P > | t |
−9	0.0004	0.0014	0.29	0.7720
−8	0.0028	0.0023	1.24	0.2190
−7	0.0043	0.0035	1.23	0.2210
−6	0.0048	0.0038	1.26	0.2090
−5	0.0057	0.0040	1.43	0.1540
−4	0.0055	0.0039	1.40	0.1630
−3	0.0079	0.0038	2.10	0.0380
−2	0.0109	0.0056	1.95	0.0540
−1	0.0137	0.0068	2.02	0.0460
0	0.0170	0.0103	1.65	0.1020
1	0.0160	0.0089	1.80	0.0750
2	0.0155	0.0073	2.13	0.0350
3	0.0141	0.0072	1.95	0.0540
4	0.0157	0.0073	2.13	0.0350
5	0.0142	0.0078	1.83	0.0700
6	0.0120	0.0079	1.52	0.1310
7	0.0121	0.0082	1.49	0.1400
8	0.0176	0.0089	1.99	0.0490
9	0.0197	0.0095	2.08	0.0400

On the final day of the window, the CAR reaches 1.97% with a significance level of 5%. As discussed in our H1, these results suggest that the markets recognize the signal in which issuers show their commitment to more sustainable practices, with investors reacting positively, thus increasing the stock price. Moreover, consistent with Morck, Yeung, and Yu (2013) and Guzmán et al. (2020), Latin American emerging markets seem to show stronger market reactions to corporate news as our CAR is higher than previous studies in developed markets (Baulkaran 2019; Tang and Zhang 2020; Wang et al. 2020).

First-time issuers represent 49.95% of our sample, and their CARs are presented in Table 6. Column two presents the CAR coefficient for each day of the event window. All coefficients after the event day are positive, and the CARs for days eight and nine are significant at the 5% level. First-time issuances generated positive and significant CARs of 2,27% during the event window compared to 1,97% of the entire sample. This result aligns with the studies of Tang and Zhang (2020) and Flammer (2021), who also found higher CARs for first-time issuers. As expected, these results suggest that the market gives a higher premium to new issuers that signal to the market their commitment to sustainability by issuing ESG bonds for the first time.

Table 6. Cumulative Abnormal Return (CAR) regressions for [−9, 9] window. First-time issuers.

Days before and after issuance	CAR	Std. Err.	t	P > | t |
−9	−0.0018	0.0019	−0.93	0.3560
−8	0.0014	0.0035	0.42	0.6760
−7	0.0061	0.0058	1.04	0.3020
−6	0.0079	0.0064	1.23	0.2240
−5	0.0121	0.0063	1.93	0.0590
−4	0.0105	0.0057	1.82	0.0750
−3	0.0108	0.0054	1.97	0.0540
−2	0.0107	0.0060	1.79	0.0800
−1	0.0133	0.0060	2.21	0.0310
0	0.0135	0.0064	2.10	0.0400
1	0.0151	0.0066	2.26	0.0280
2	0.0156	0.0069	2.26	0.0280
3	0.0125	0.0072	1.74	0.0880
4	0.0174	0.0081	2.16	0.0360
5	0.0153	0.0084	1.82	0.0750
6	0.0162	0.0086	1.87	0.0670
7	0.0162	0.0088	1.85	0.0700
8	0.0199	0.0088	2.24	0.0290
9	0.0227	0.0097	2.34	0.0230

We then analyze the determinants of the CARs and report the results in Table 7 for six (6) different model specifications. Regarding governance characteristics, we find a negative, significant, and non-monotonic association between CAR and the size of the board. Specifically, when boards are smaller than the mean in the sample (9 directors), the impact on the CAR is positive and significant but non-monotonic. However, when the board increase in size, the effect becomes negative. These results imply that firms with large boards cannot fully capitalize on the benefits in terms of CAR from ESG bond issuance, suggesting information and coordination problems.

Table 7. Impact of firm characteristics on Cumulative Abnormal Return (CAR).

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	CAAR	CAR	CAR	CAR	CAR	CAR
Board size deviation from the mean	−0.0050	−0.0060*	−0.0070*	−0.0070*	−0.0140***	−0.0150***
	(0.0040)	(0.0040)	(0.0040)	(0.0040)	(0.0040)	(0.0050)
Board size deviation from the mean^2	0.0010**	0.0020**	0.0020**	0.0020***	0.0030***	0.003***
	(0.0010)	(0.0010)	(0.0010)	(0.0010)	(0.0010)	(0.0010)
Board independence	−0.0160	−0.0140	−0.0110	−0.0100	−0.0220	−0.0340
	(0.0410)	(0.0410)	(0.0410)	(0.0390)	(0.0380)	(0.0480)
Board diversity	0.3620***	0.3490***	0.3480***	0.3840***	0.3370***	0.3310***
	(0.0840)	(0.0830)	(0.0850)	(0.0810)	(0.0790)	(0.0810)
CEO duality	−0.0050	−0.0110	−0.0080	−0.0050	0.0300	0.0450
	(0.0440)	(0.0440)	(0.0440)	(0.0420)	(0.0410)	(0.0440)
Rating		−0.0080	−0.0080	−0.0060	−0.0070	−0.0060
		(0.0050)	(0.0050)	(0.0050)	(0.0050)	(0.0060)
Size			−0.0050	−0.0140**	−0.0100*	−0.0120
			(0.0060)	(0.0060)	(0.0060)	(0.0090)
Growth			−0.0380	−0.0730	−0.0750	−0.0850
			(0.0540)	(0.0520)	(0.0500)	(0.0540)
Leverage				0.1770***	0.1580***	0.2000***
				(0.0500)	(0.0480)	(0.0600)
Largest Shareholder’s Equity Stake					−0.1900***	−0.2270***
					(0.0570)	(0.0600)
ESG score						−0.0010
						(0.0010)
Constant	−0.0520*	0.0150	0.0940	0.0870	0.1410	0.4410**
	(0.0280)	(0.0490)	(0.1040)	(0.0980)	(0.0950)	(0.1790)
Observations	115	115	115	115	115	115
R-squared	0.1990	0.2180	0.2280	0.3110	0.3790	0.4800
Industrial sector dummies	No	No	No	No	No	Yes
Controlling shareholder identity dummies	No	No	No	No	No	Yes
Number of firms	67	67	67	67	67	67

Standard errors in parentheses.

*** p < 0.01, ** p < 0.05, * p < 0.1.

We also observe a positive association with board diversity measured as the percentage of women on the board (Board diversity); a boost of 10% on this variable increases the CAR by 3.31% (column 6 in Table 7). These findings suggest that board characteristics such as size and diversity can impact sustainability-related decisions and strategies by bringing a wider variety of issues to the board’s agenda, including social and environmental matters, and the ability of the board to deal with them appropriately. Regarding Board independence and CEO duality, we did not find significant associations with CAR.

Ownership concentration has a negative and significant association with the dependent variable as an increment of 10% in the percentage owned by the largest shareholder decreases CAR by 2.27% (column 6 in Table 7). Like Chen et al. (2021), this finding is consistent with agency problems between large and minority shareholders, as controlling shareholders can use their power to reduce environmental and social sustainability practices by adopting a rent-seeking behavior or managerial myopia favoring short-term results. Overall, consistent with our H2, firm-level governance provisions significantly affect the market reaction for firms issuing ESG bonds in Latin America. Unlike ownership concentration, the rating of the ESG bond does not have a statistically significant impact on CAR.

Concerning financial characteristics, we document a positive and significant impact of leverage on the CAR, where a 10% increase in debt boosts CAR by 2.0% (column 6 in Table 7). This finding is consistent with Flammer (2021)’s signaling argument, where firms needing financing instead of going to traditional bonds and other funding sources decide to issue ESG bonds, showing their commitment to sustainable practices. Firm’s size shows negative coefficients but are only significant in models (4) and (5).

As illustrated in Table 7, we include several control variables to test the robustness of the model. The linear regression in column 6 uses controls with dummy variables by industry. We are also controlled by the firm’s controlling shareholder (private or state-owned) and the ESG performance rated by Refinitiv. Our results remain similar in the different regression models.

4.1. Robustness checks

To enhance the robustness of the study, we used the announcement date (AD) of the issuance as a proxy to measure the impact of ESG bond issuances on firm’s stock price. As Tang and Zhang (2020) argue, when firms announce the issuance of ESG bonds, media exposure increases dramatically compared with conventional corporate bond issuance. This happens because ESG bond issuers publish formal press releases, thereby sending good signals to the market. However, since Latin American markets are characterized by high levels of opaqueness, deep information asymmetries (González et al. 2021b, 2021a; OECD 2018), we were able to collect the announcement date for only 75 issuances to calculate the CAR. As suggested by Nguyen and Wolf (2023), we also computed the Cumulative Average Abnormal Return (CAAR) in equation (5)(5) $\mathrm{CAAR}(-9,9)={\displaystyle \frac{1}{N}\sum _{t=-9}^9\mathrm{CA}{R}_{i,t}}$(5) . (5) $\mathrm{CAAR}(-9,9)={\displaystyle \frac{1}{N}\sum _{t=-9}^9\mathrm{CA}{R}_{i,t}}$(5) As presented in Table 8, the CAR is higher for the announcement date (2.6%) than it is for the issuance date (1.97%), which means the signal is stronger for the announcement date as it represents new information to the market. When analyzing the CAAR, we find that it is statistically significant, reaching 1.11% for the full sample and 1.13% for first time issuers, confirming H1.

Table 8. CAR for announcement dates and CAAR for issuance dates.

Variable	Coefficient	Std. Err.	t	P>|t|
CAR (AD)	0.0265	0.0129	2.05	0.0440
CAAR (issuance date)	0.0111	0.0052	2.15	0.0330
CAAR (issuance date – first time issuers)	0.0113	0.0056	2.01	0.0500

The impacts of firm characteristics CAR for announcement dates (AD) and CAAR for issuance dates are shown in Table 9. Board size deviation from mean has negative and significant coefficients of – 0.016 and – 0.02 for CAR (AD) and CAAR respectively. Board diversity preserves positive and significant coefficients for CAR (AD) (0.336) and issuance date CAAR (0.160). As with previous models, this confirms H2. Leverage keeps positive and significant coefficients of 0.199 and 0.082 for CAR (AD) and CAAR. Ownership concentration preserves negative coefficients of – 0.226 and – 0.088 respectively.

Table 9. Impact of firm characteristics on CAR for announcement dates and CAAR for issuance dates.

Variable	(1)	(2)
	CAR (AD)	CAAR
Board size deviation from the mean	−0.0160**	−0.0040**
	(0.0070)	(0.0030)
Board size deviation from the mean^2	0.0030***	0.0010**
	(0.0010)	(0.0000)
Board independence	−0.0430	−0.0080
	(0.0610)	(0.0230)
Board diversity	0.3360***	0.1600***
	(0.1050)	(0.0490)
CEO duality	0.0270	0.0120
	(0.0570)	(0.0260)
Rating	−0.0100	−0.0050
	(0.0080)	(0.0030)
Size	−0.0210*	−0.0010
	(0.0120)	(0.0060)
Growth	−0.0410	−0.0050
	(0.0690)	(0.0320)
Leverage	0.1990***	0.0820***
	(0.0730)	(0.0300)
Largest Shareholder’s Equity Stake	−0.2260***	−0.0880**
	(0.0730)	(0.0340)
ESG score	−0.0000	−0.0010*
	(0.0010)	(0.0000)
Constant	0.4760**	0.0620
	(0.1790)	(0.0900)
Observations	75	115
R-squared	0.5900	0.3160
Industrial sector dummies	Yes	Yes
Controlling shareholder identity dummies	Yes	Yes
Number of firms	67	67

Standard errors in parentheses.

*** p < 0.01, ** p < 0.05, * p < 0.1.

4.2. Endogeneity

To deal with endogeneity issues that might come from omitted variables, simultaneity, and measurement error, we follow a Propensity Score Matching (PSM) approach and estimate the impact of board diversity on Abnormal Return (AR), Cumulative Abnormal Return (CAR), and Cumulative Average Abnormal Return (CAAR). This technique constructs an artificial control group by matching firms with board diversity higher than 20% with non-treated firms of similar characteristics. As Table 10 shows, firms with more than 20% Board Diversity have 1.49% more AR, 12.11% more CAR, and 5.64% more CAAR than similar matches firms with lower board diversity.

Table 10. Propensity Score Matching (PSM) to test the validity of the results in Table 7, regarding the effect of Board Diversity on AR, CAR, and CAAR.

Variable	Treated (Women on the Board)	Controls	Difference	S.E.	T stat
AR	0.0149	−0.0002	0.0151	0.0062	2.42**
CAR	0.1211	0.0009	0.1202	0.0237	5.07***
CAAR	0.0564	0.0027	0.0537	0.0133	4.02***

Even though we cannot argue that board diversity is a genuinely exogenous event and therefore cannot dismiss potential endogeneity in our outcomes, we believe that the robustness of our main results, including a complete set of control variables and the PMS method, lessen the endogeneity issues, which are a typical problem in corporate governance literature.

5. Discussion and conclusions

In this study, we assembled a comprehensive dataset encompassing 115 corporate ESG bond issuances from listed firms in six Latin American countries and studied the signaling effect through the stock market´s reactions. Consistent with the signaling argument, we empirically document that stock markets had positive and significant reactions to the issuance of ESG bonds measured by a Cumulative Abnormal Return (CAR) in a [−9,9] window with a positive price reaction of 1.97%. For the sample of first-time issuers, these reactions were higher, with a significant 2.27% CAR. These findings remain robust after using alternative proxies such as Cumulative Average Abnormal Return (CAAR).

We studied ESG bonds as a broader approach to sustainability. Our results suggest that not only green bonds, which have been the focus of previous studies, have a significant signaling effect. No previous studies have focused only on Latin American markets, and we show stronger price reactions to corporate sustainability practices in this region. This finding suggests that when a firm tags its ESG bond, it increases its visibility to the media, possibly attracting new investors, leading to more demand for its shares, and increasing its stock price (Tang and Zhang 2020). This result adds to the literature which argues that in emerging markets, corporate practices serve to send strong signals to mitigate information asymmetries and enhance transparency regarding firms’ ESG performance.

We then analyze the determinants of the CARs. The board of directors is a relevant corporate governance characteristic as directors push sustainability issues into the firm’s agenda. We document a negative, significant, and non-monotonic association between CAR and board size, which implies that smaller boards have a positive impact on CAR. These results imply that firms with large boards cannot fully capitalize on the benefits in terms of CAR from ESG bond issuance, suggesting information and coordination problems that affect decision-making processes (Avramov et al. 2022). We also find that board gender diversity positively impacts the value that the market recognizes to ESG bond issuers, suggesting that board composition impacts sustainability-related decisions and strategies as female directors might bring these new perspectives to board discussions, thereby enhancing governance mechanisms (Husted and Sousa-Filho 2019). Finally, ownership concentration has a negative and significant association with CAR, consistent with the idea that controlling shareholders might generate agency problems as they pursue short-term returns on investment and can use their power to reduce environmental and social sustainability practices.

All these results suggest that ESG matters for market participants. Firms that engage in ESG practices are recognized by investors, creating an incentive for other listed firms to join these initiatives. We also take from our results the importance of governance structures to lead these efforts towards sustainability.

6. Limitations and future research

The sample considered in this study includes 115 ESG bond issuances from 67 issuers. It is a small sample because most ESG bonds in Latin America are issued by governments and non-listed firms, for which we cannot measure a stock price reaction. Moreover, issuers come from the six largest Latin American countries, but firms from the remaining markets may have also issued ESG bonds. For this reason, a generalization of these results must be conducted carefully.

Another important limitation is related to the announcement date of the issuance. As Tang and Zhang argue (2020), the signal of the ESG bond issuance is stronger for this date as it represents new information to the market. However, Latin American markets have high levels of opaqueness and information asymmetries, for which it was not possible to compile all the announcement dates. As more information is available and more firms issue these instruments, future research might explore the impact of ESG bond announcements on a firm’s stock price, cost of capital, and performance.

The focus of our study was positioned toward listed firms, and our results suggest that good corporate governance practices also represent good signals to the market. Since listed firms have higher requirements regarding corporate governance than non-listed firms, future research may explore how corporate governance influences the decision to issue ESG bonds in privately held firms.

Ethical Approval

This article does not contain any studies with human participants or animals performed by any of the authors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Universidad de los Andes.

Notes

1 Report for Selected Countries and Subjects. World Bank Database, October 2022.