ESG and the Cost of Debt: Role of Media Coverage

Abstract

This study delves into the interplay between the Environmental, Social, and Governance (ESG) ratings and the debt costs incurred by Korean-listed companies, highlighting their pivotal significance in today’s corporate ecosystem. Our primary focus is to explore how the extent of media coverage moderates this relationship, thereby shedding light on the pivotal role that public scrutiny plays in shaping a company’s financial outcomes. Utilizing the Ordinary Least Squares (OLS) regression model, we rigorously control for industry and year effects, as well as firm-specific variations. Additionally, we conduct a series of supplementary analyses and robust tests to further strengthen the credibility of our findings. Our empirical analysis reveals that firms with poor ESG ratings, indicating corporate social irresponsibility, incur higher debt costs in the subsequent period. Notably, this adverse financial impact is significantly alleviated for companies that enjoy higher media coverage. This notable discovery underscores the potential of media scrutiny to reduce the financial burden imposed by inadequate ESG performance. Our results suggest that companies, especially those with limited media attention, should prioritize enhancing their ESG performance to mitigate potential financial implications. Overall, our research contributes to a more nuanced understanding of the intersection between corporate social responsibility, media coverage, and financial performance.

1. Introduction

With growing global interest in corporate social responsibility (CSR) issues, companies’ CSR performance is not only closely related to their operations and future growth but is also an important decision-making criterion for investment decisions. Initially, CSR primarily focused on philanthropy and corporate charity. However, it has evolved into a strategic business imperative that encompasses a broader range of activities aimed at addressing the Environmental, Social, and Governance (ESG) issues [1]. ESG, which quantifies the engagement of CSR with widely accepted standards, has become an important indicator for corporate value evaluation in the context of sustainable development. Media coverage of ESG failures has heightened public scrutiny and led to adverse economic consequences for many companies. A notable example is the Volkswagen emissions scandal, or “Dieselgate”, which broke in September 2015. This scandal involved the installation of defeat devices in approximately 11 million diesel vehicles worldwide, allowing them to pass emissions tests by temporarily reducing harmful emissions during testing conditions. However, under normal driving conditions, these vehicles emitted pollutants up to 40 times above legal limits. The media’s exposure of this discrepancy not only revealed a major environmental fraud but also led to severe repercussions for Volkswagen, including a substantial drop in stock price and billions of dollars in fines, legal fees, and vehicle buybacks.

Prior research has documented the positive effect of ESG in various ways: financial performance improvement [2,3,4,5,6], reputation enhancement [7], cost of capital reduction [8], and firm value increase [9]. Investors pay attention to companies’ ESG ratings not only because they choose to invest in ethical and sustainable projects but also due to the outperformance of ESG portfolios, especially in economic downturns [10,11]. Also, Bae, et al. [12] and Kim, et al. [13] found that higher ESG ratings lead to the reduction of the stock price crash risk in the US and Korean markets, respectively. This study explores how ESG performance influences the cost of debt, a practical measure for evaluating the ramifications of ESG activities in the sense of a direct indicator of firms’ capital burden, and in particular the moderating role of media coverage in that relationship.

Media can provide investors with information about a company’s financial condition and influence their decision-making [14,15]. Existing literature has documented the effects of media coverage on investment behavior and firm value. Media coverage serves as a proxy for public attention, which can reduce market friction [16], influencing investment behavior [17]. Furthermore, compared with information released by companies, market participants are more likely to react to information disclosed by the media [18]. Besides, media exposure is one of the main drivers of ESG transparency, which improves investment efficiency [19] and leads to higher firm value [20,21]. As argued by Servaes and Tamayo [22], the value of CSR is only associated with high-awareness firms, which is raised by media coverage. Taken together, given that media coverage plays a significant role in improving the information environment by releasing more credible information, we posit that media coverage induces creditors to place differential importance on the information provided by ESG ratings.

Building on the above-mentioned context, our research examines whether ESG ratings are associated with debt costs in Korean-listed companies. We first propose the question of whether poor ESG performance, which can be viewed as a form of corporate irresponsibility, increases the cost of debt. Drawing on resource-based theory and signaling theory, we aim to explain why investors may charge higher costs of debt for companies with poor ESG ratings. Furthermore, we center our study on how media exposure changes the effect of ESG activities by providing more information and raising public attention. Media coverage serves as a proxy for the information environment, disseminating information to the public, monitoring corporate conduct, and attracting investors’ interest. Our results indicate that high media coverage can amplify the adverse impact of inadequate ESG performance on debt costs.

This study targets the companies that are listed on the Korean Stock Exchange. To measure a company’s commitment to corporate social responsibility, we use ESG ratings generated by the Korea Corporate Governance Service (KCGS). We employ OLS regression analysis, controlling for industry and year-fixed effects, as well as clustering firm-specific standard errors. Our empirical results suggest that low ESG ratings lead to a higher cost of debt in the subsequent period. More importantly, we find that the impact of low ESG ratings on debt costs is weakened by media coverage, proxied by the number of media reports on firms. Further, to address the endogeneity problems, we conduct the Two-Stage Least Squares (2SLS) and dynamic Generalized Method of Moments (GMM) and find consistent results after controlling for endogeneity with these two methodologies. To enhance the depth of our research findings, a series of additional analyses are conducted, including the usage of other proxies of the cost of debt, investigating the moderating role of media over-attention, and grouping analyses of ESG’s impact on the cost of debt. In sum, the additional analyses firmly support our hypotheses, providing a more detailed explanation and bringing our findings more instructive to reality.

This study contributes to the current literature on ESG, cost of debt, and media coverage. Previous research has primarily focused on the direct association between ESG disclosure and the cost of debt, particularly in European nations and on a global scale [23,24]. However, limited attention has been paid to the impact of poor ESG ratings on the cost of debt within the context of the Korean market. Our study addresses this gap by examining how ESG factors influence the cost of debt for Korean companies. Additionally, this study is the first to present that the effect of ESG activities on debt costs is mitigated by the degree of media coverage, suggesting that media scrutiny can help alleviate the financial burden associated with unfavorable ESG ratings. Our results imply that firms should pay more attention to their ESG performance when they receive less media attention because companies with poor ESG ratings may face more significant repercussions. Our findings have several practical implications for various stakeholders. Companies can leverage this knowledge to attract lower-cost capital by enhancing their ESG practices and expanding their visibility in the media landscape. For investors, by incorporating ESG considerations into their investment strategies, investors can make more informed decisions, potentially leading to better portfolio performance. Lastly, for policymakers, our findings can serve as a reference point for shaping regulations and incentives in the area of ESG. Policymakers can design policies that encourage sustainable practices, ultimately contributing to societal progress.

The remainder of the paper proceeds as follows: Section 2 discusses the related literature about the impact of ESG on the cost of debt as well as the role of media coverage. This section also provides some theoretical background and posits a set of testable hypotheses. Section 3 describes the data and the variables, while the model used in this paper is also presented. The econometric analysis is implemented in Section 4, and the results are discussed. Section 5 provides a robustness test and additional analysis, and Section 6 concludes.

2. Literature Review Hypothesis Development

2.1. ESG Relevant Research

The association between CSR and corporate financial behavior has been investigated in previous literature. It is generally agreed that ESG activities contribute to an increase in firm value, improve financial and operational performance, as well as increase organizational value [25,26,27]. Furthermore, ESG enables firms to outperform in an economic downturn. The study by Amiraslani et al. [28] indicates that high-CSR firms enjoyed lower bond spreads during the financial crisis. CSR’s positive role in mitigating the risk of a stock price crash is examined by [12].

While most research focuses on the benefits of good ESG performance, some attention is paid to the negative economic consequences that result from low ESG scores. For instance, Ratajczak [29] conducted research on ESG and the cost of debt based on companies recognized by Corporate Knights as the most sustainable companies in the world, finding that a lower cost of debt is associated with environmental and social activities in the long run. Conversely, Kölbel, Busch, and Jancso [15] explore the relationship between corporate social irresponsibility and financial risk, positing that media coverage of social irresponsibility increases the potential for stakeholder sanctions.

2.2. ESG Ratings and COD

The resource-based view theory posits that a firm’s competitive advantage originates from its internal resources [30]. When considering the impact of ESG performance on debt costs, the resource-based view framework provides a compelling lens. When a firm’s ESG performance is inadequate, it reflects deficiencies or weaknesses within its internal resources related to environmental, social, and governance aspects.

First, from the perspective of business risk, ESG is one of the factors contributing to a lower cost of capital because it is associated with a lower level of risk [31]. There are two aspects to explain the association. Firstly, an organization with low ESG performance is likely to face higher risks of litigation and fines, whereas one with high levels of corporate social responsibility enjoys more stable relationships with the government and the financial community [32]. Secondly, according to the risk management theory, companies with good CSR performance can generate positive moral capital among various stakeholders in times of crisis, which provides them with insurance-like protections [33]. Li et al. [34] show that the bond default rate is negatively related to its social responsibilities and corporate governance. Lee et al. [35] demonstrate that a higher ESG score is associated with a lower VaR in stock return. In contrast, as a firm’s ESG performance is indicative of the quality and management of its internal resources, a firm with poor ESG performance is likely to be perceived as having inferior resources in terms of environmental sustainability, social engagement, and corporate governance. Such perceived deficiencies can increase the risk profile of the firm and raise concerns about the firm’s long-term sustainability and prospects, leading investors and lenders to demand higher returns to compensate for the increased risk.

Second, engagement in CSR has the effect of enhancing reputation. ESG performance plays a crucial role in shaping investors’ confidence and trust in a firm [36]. From the perspective of performance, the resource-based theory of reputation suggests that corporate reputation has a significantly positive impact on firm growth [37] and plays a mediating role between CSR and corporate financial performance [4]. CSR activities may act as a reservoir of goodwill, insulating them from the negative impacts of the crisis [38]. Iwu-Egwuonwu and Chibuike [39] review the empirical studies on corporate reputation and argue that cultivating a good reputation is a prerequisite for firms that intend to beat competitors, enhance their market outlook, increase their financial performance, and have a sustainable place in the global economy. A good CSR reputation can increase the value of a company’s predicted cash flows, reduce the variability of its cash flows, and increase its net income [7], thus reducing the cost of debt. Conversely, poor ESG performance can erode investors’ confidence, resulting in higher debt financing costs as investors.

Third, aside from avoiding risk and improving performance through ESG practice, companies that have engaged in ESG activities release more information to the public. The signaling theory that was originally proposed by Spence [40] suggests that in markets with asymmetric information, firms engage in signaling activities to communicate their true quality to stakeholders. As posited by signaling theory, firms seek to mitigate information asymmetry through credible signals [41]. Due to the fact that insiders, i.e., managers, have better access to information than stakeholders, outsiders may interpret any additional information as a signal to the market [7]. A company’s ESG practices provide a channel for communicating with outsiders, reducing agency costs, releasing positive signals to stakeholders, and demonstrating the company’s willingness and ability to care for society [42]. In contrast with good ESG practice, negative ESG media coverage is likely to trigger investor concern about increases in future contract costs, resulting in a downward adjustment in firm valuations [43]. In accordance with Dhaliwal, et al. [44], CSR reports are likely to provide investors with additional information that can be used to assess the long-term sustainability of businesses. ESG-related information benefits firms by attracting socially responsible customers, contributing to minimizing government regulatory risks, addressing activists and non-profit organizations, and reducing energy waste [43]. In this regard, it is reasonable to assume that the ESG-related information provided by the company or evaluated by a third party can exert a significant signaling effect and assist in mitigating the asymmetry of information. By affecting a firm’s ability to communicate its true value and increasing its perceived risk profile, poor ESG performance leads to a deterioration in market sentiment and an increase in the cost of debt. Taken together, regarding the role ESG practice plays in reducing idiosyncratic risk and enhancing performance as well as mitigating information asymmetry, we posit the following:

H1. 

A lower ESG score is associated with a higher cost of debt for the subsequent period.

2.3. How Media Coverage Moderate the Relationship between ESG and COD

The role of media reports shares common ground with ESG activities in disseminating information to the public [45], capturing investors’ interest [46,47], and monitoring corporate actions [48,49]. The three functions of media improve the information environment, which weakens the influence of ESG performance on the cost of debt.

In the first place, media is a channel for disseminating information. Media coverage can lower the cost of debt by helping to mitigate agency costs through the dissemination of information. As proposed by Bushee, et al. [50], the media press serves as an information intermediary that shapes firms’ information environments by packaging and disseminating information. Even if mass media does not supply genuine news, it can alleviate information friction and affect security pricing [46]. Media provides timely news to the public and markets react in minutes after insider news is disseminated by the public media [51]. In sum, media reports are pivotal in disseminating information to the wider public, thereby reducing investors’ information expenses, which intersects with the ESG’s role in information disclosure. Consequently, for firms with limited media attention, ESG disclosure is the source for investors to get information about the enterprises that are not able to secure funding at optimal rates, facilitating financially beneficial ventures.

Secondly, media coverage increases the visibility of the firms covered, and investors are buyers of information. Their investment decisions are based on the companies in the spotlight and those that are being zoomed in. Following the agenda-setting theory, mass media possess persuasive influence in directing public attention toward particular events and issues, as well as in shaping the significance individuals attribute to matters of public concern [52]. The media agenda is reflected in the ratings of the significance or prominence of issues among the audience. Media shapes the perception of investors and can predict the trading volume and price of stocks. Tang and Zhang [53] have observed that stocks located in cities experiencing elevated abnormal media attention exhibit superior performance compared to stocks situated in cities with lower abnormal media attention. Tetlock [47] finds that the attitude of media content will induce investors to trade on noise and therefore drive stock prices away from fundamentals. Interestingly, investors tend to ignore the content of information and make judgments based on the amount of attention they receive [54]. According to Solomon et al. [55], extra flows were only attracted when high past returns stocks were recently featured in the media. Fang and Peress [46] offer empirical evidence that stocks with lower media coverage need to offer higher returns to compensate their holders. Hence, the focus of media amplifies investor interest, considering ESG initiatives serve as a means to capture public and investor attention, particularly in specific domains. Media seems to supplant the role of ESG endeavors in terms of garnering attention.

Thirdly, the media functions as an external governance mechanism overseeing the behavior of enterprises. The positive and negative contents of media reports are both favorable from the perspective of reducing information costs. Managers are sensitive to the level and tone of media reports on their decisions; they “listen to the market” [14]. On one hand, media coverage changes the relationship between ESG rating and COD through the positive effect on public perception. Based on Khan and Sukhotu [56]’s findings, companies with positive media images and positive CSR activities are likely to attract positive attitudes from their customers. Also, companies pay attention to media coverage when planning to raise new funds, since they want to ensure that their actions are perceived favorably by the capital markets [49]. Meanwhile, negative media reports serve as a public watchdog to monitor corporate issues enabling the public to identify corporate financial misconduct [57]. Chen, Cheng, Li, and Zhao [48] demonstrate a negative correlation between media attention and both accrual-based and real earnings management, implying that the media functions as an external monitor, mitigating managers’ propensity for opportunistic manipulation of earnings. It functions as an external governance mechanism by providing objective information about a company [58]. Gao et al. [59] documented a strong positive interaction effect between media coverage and corporate governance, indicating that media improves management oversight and efficiency. Due to the threat of adverse media coverage, firms are also more likely to commit to sustainable practices [60]. In this context, media scrutiny can be a form of governance. Media do not only monitor firms with inadequate CSR practices: they are far more likely to report accidents when they occur at a company with a superior CSR record [61].

As we proposed earlier, one of the negative effects of low ESG ratings is the higher cost of debt. Media coverage reduces the relationship between the two variables by improving the information environment. Prior literature finds that firms that act more socially responsible receive more favorable or unfavorable media coverage [60,61]. Public media disclosure is the main driver of ESG transparency [19]. Thus, media as one of the channels for non-financial information disclosure with creditability may mitigate partially the influence of ESG information. Consequently, companies with a more robust media presence can utilize debt at a lower cost as the media oversees the company’s operations, and with more effective corporate governance, the credit risk is reduced [62]. Creditors perceive higher media coverage as a better information environment, thus reducing the influence of ESG ratings on the cost of debt. Thus, we propose that the media serves as a substitute for other sources of information, especially concerning ESG disclosure. It is expected that companies with greater information accessibility will exhibit a weaker correlation between ESG ratings and debt costs. In summary, the aforementioned points culminate in the following hypothesis:

H2. 

The impact of low ESG ratings on the cost of debt is less pronounced for high media coverage firms than for low media coverage firms.

3. Variables and Research Design

Previous literature focuses on the positive impact of good ESG practice or disclosure quality and access to finance [63,64]. Our focus in this study is on the investigation of the impact of poor ESG practice on the cost of debt. ESG rating data is obtained from the KCGS (Refers to the Korean Corporate Governance Service website at http://www.cgs.kr, accessed on 1 September 2022). Other financial data is collected from the FnGuide. Our final sample contains 672 firms with 4910 firm-year observations covered for 8 years. We collect the data of the number of news articles from BigKinds (Refers to the Bigkinds News Bigdata & Analysis website at https://www.bigkinds.or.kr/, accessed on 1 September 2022, which is an authoritative website that aggregates Korean news articles and is publicly accessible). Following the previous literature [65], we exclude financial sector firms due to the sector-specific regulations and the incomparability in terms of the cost of debt financing. Additionally, we screen out missing or incomplete data. The data screening process is illustrated as follows (

Table 1. Data screening process.

	Firm-Year Observations	Number of Firms
All observations with ESG performance available from 2012 to 2019 provided by KCGS	6849	1140
Exclude observations with missing cost of debt variable and financial firms	(1688)	(381)
Exclude observations with missing media coverage	(90)	(67)
Exclude observations with missing control variables	(160)	(20)
Final sample	4910	672

):

Dependent variables and control variables are winsorized at the 1st and 99th percentiles to reduce the impact of extreme observations. To control for serial correlation and heteroscedasticity Ghoul, et al. [66], our regression standard errors are clustered by firms. Also, our regression results are robust by controlling the industry and year effect.

Our first test examines whether poor ESG practice affects the cost of debt. The cost of debt is calculated as the ratio of a firm’s interest expense to its average liabilities. ESG is evaluated and graded by the authoritative rating agency of KCGS. We obtained the aggregated and individual ESG evaluation ratings reflecting the environment, social responsibility, and corporate governance dimensions. The 6 grades are A+, A, B+, B, C, and D. To proxy for corporate irresponsibility, we assign −1 point for A+, −0.8 point for A, −0.6 point for B+, −0.4 point for B, −0.2 point for C and 0 point for D, so the higher number here stands for poorer ESG practice. We then collect the number of articles related to each firm in Bigkinds.

We estimate the following equation and include industry fixed effects and year effects in order to test the impact of a poor ESG rating on the cost of debt for H1:

$$\begin{gathered} CO{D}_{i,t+1}={\beta }_0+{\beta }_{1}Poor\text{\_}ES{G}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t} \\ +{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}+\sum Year+\sum Industry+{\epsilon }_{i,t} \end{gathered}$$

(1)

where the dependent variable, $CO{D}_{i,t+1}$ proxies for the cost of debt of subsequent period, and our main variable of interest, $Poor\text{\_}ES{G}_{i,t}$ represents the aggregate scores of poor ESG as well as the three individual poor E, S, and G scores. As we focus on how ESG practices in the current year t influence the following period’s cost of debt in t + 1, Equation (1) reveals the time lag between a dependent variable and the independent variable. As we hypothesized in H1, ${\beta }_1$ is expected to be significant and positive, suggesting that poor ESG leads to a higher cost of debt in the subsequent period. We add control variables that may influence the cost of debt, such as $SIZE$, $LEV$, $ROA$, $MTB$, $FOR$, and $OWN$, based on the existing literature [65,67,68,69]. Appendix A describes the variables and methods of measurement.

To test the moderating role of media coverage in the relation between ESG and the cost of debt for H2, we developed Equation (2) as follows:

$$\begin{array}{ll}CO{D}_{i,t+1}={\beta }_0+& {\beta }_{1}Poor\text{\_}ES{G}_{i,t}+{\beta }_{2}MEDI{A}_{i,t}+{\beta }_{3}Poor\text{\_}ES{G}_{i,t}\times MEDI{A}_{i,t}+{\beta }_{4}SIZ{E}_{i,t}\\ & +{\beta }_{5}LE{V}_{i,t}+{\beta }_{6}RO{A}_{i,t}+{\beta }_{7}MT{B}_{i,t}+{\beta }_{8}FO{R}_{i,t}+{\beta }_{9}OW{N}_{i,t}+\sum Year\\ & +\sum Industry+{\epsilon }_{i,t}\end{array}$$

(2)

where $MEDI{A}_{i,t}$ is an indicator variable, which is set to 1 if the natural logarithm of the number of media articles is greater than the median, 0, otherwise. $MEDI{A}_{i,t}$ interacts with $Poor\text{\_}ES{G}_{i,t}$ to examine the moderating effect of media coverage on the relationship between poor ESG and the cost of debt. ${\beta }_3$ is expected to be negative to support our second hypothesis. Consistent with Equation (1), $Poor\text{\_}ES{G}_{i,t}$ represents the aggregate scores of poor ESG as well as the three individual poor E, S, and G scores.

4. Empirical Results

In

Table 2. Summary of descriptive statistics.

	N	Mean	Std. Dev.	Min	p25	Median	p75	Max
COD	4910	1.781	1.630	0	0.76	1.56	2.4	32.93
Poor_ESG	4910	−0.373	0.176	−1	−0.4	−0.4	−0.2	0
Poor_E	4910	−0.359	0.231	−1	−0.6	−0.4	−0.2	0
Poor_S	4910	−0.391	0.208	−1	−0.4	−0.4	−0.2	0
Poor_G	4910	−0.382	0.176	−1	−0.4	−0.4	−0.2	0
MEDIA	4910	−0.382	0.5	0	0	−0.4	1	1
ROA	4910	0.017	0.069	−0.299	0.001	0.024	0.052	0.181
SIZE	4910	20.245	1.589	17.29	19.133	20.036	21.132	24.696
LEV	4910	0.465	0.201	0.075	0.305	0.473	0.616	0.908
MTB	4910	1.234	1.217	0.176	0.533	0.841	1.411	7.442
FOR	4910	0.102	0.127	0	0.015	0.048	0.143	0.593
OWN	4910	0.438	0.163	0.089	0.32	0.443	0.546	0.818

This table represents the descriptive statistics for the variables used in this paper. The data include a sample of Korean-listed non-financial firms from 2012 to 2019. It contains 672 firms with 4910 firm-year observations. Variable definitions refer to Appendix A.

, we summarize the means and medians of the main and control variables used in this paper. COD has a mean and median of 1.781 and 1.56, respectively, which means that the cost of debt is positively skewed. For poor ESG practices, the mean value is about −0.373, while the minimum and maximum values are −1 and 0, respectively. It is worth noting that good ESG practices are indicated by the minimum value and poor performance is indicated by the maximum value. In our sample, therefore, ESG performance is skewed toward poor performance. The average level of Poor_E, Poor_S, and Poor_G are −0.359, −0.391, and −0.382, respectively, meaning that they are close to the B grade.

Table 3. Correlation coefficient matrix.

	COD	Poor_ESG	Poor_E	Poor_S	Poor_G	MEDIA	ROA	SIZE	LEV	MTB	FOR	OWN
COD	1
Poor_ESG	0.116 ***	1
Poor_E	0.046 ***	0.696 ***	1
Poor_S	0.096 ***	0.770 ***	0.514 ***	1
Poor_G	0.157 ***	0.722 ***	0.270 ***	0.469 ***	1
MEDIA	−0.111 ***	−0.286 ***	−0.197 ***	−0.314 ***	−0.239 ***	1
ROA	−0.450 ***	−0.130 ***	−0.055 ***	−0.127 ***	−0.159 ***	0.134 ***	1
SIZE	−0.106 ***	−0.609 ***	−0.470 ***	−0.631 ***	−0.454 ***	0.420 ***	0.176 ***	1
LEV	0.393 ***	−0.136 ***	−0.167 ***	−0.142 ***	−0.009	0.079 ***	−0.335 ***	0.251 ***	1
MTB	0.093 ***	−0.001	0.023	−0.039 ***	0.026 *	0.053 ***	−0.014	−0.128 ***	0.026 *	1
FOR	−0.219 ***	−0.412 ***	−0.304 ***	−0.406 ***	−0.353 ***	0.284 ***	0.264 ***	0.505 ***	−0.126 ***	0.126 ***	1
OWN	−0.198 ***	0.088 ***	0.098 ***	0.060 ***	0.054 ***	−0.088 ***	0.187 ***	0.003	−0.105 ***	−0.186 ***	−0.187 ***	1

This table presents the Pearson correlation coefficient for the variables used in our analyses. The data include a sample of 672 non-financial firms in Korea. *, *** indicate significance at 10% and 1% levels, respectively. The variable definition refers to Appendix A.

presents the Pearson correlations for the independent and dependent variables. COD shows a significant positive correlation with the four poor ESG indicators which firmly support our first hypothesis that poor ESG performance can lead to a higher cost of debt in the subsequent year. In regard to media coverage (MEDIA), there is a significant negative correlation between COD and MEDIA, and it exhibits a significant negative correlation between Poor_ESG and the three individual ESG variables. The results indicate that media coverage is negatively correlated with low ESG ratings, which is consistent with our prediction.

4.1. Does Poor ESG Rating Influence the Cost of Debt?

Before testing the moderation role of media coverage, we first conduct the test on whether poor ESG performance is associated with a higher cost of debt. In

Table 4. The impact of poor ESG ratings on the cost of debt.

$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t}+{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}$ $+\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	0.472 ***
	(2.712)
Poor_E		0.011
		(0.075)
Poor_S			0.273 **
			(1.974)
Poor_G				0.537 ***
				(3.535)
ROA	−6.325 ***	−6.349 ***	−6.331 ***	−6.276 ***
	(−10.312)	(−10.230)	(−10.216)	(−10.211)
SIZE	−0.021	−0.052 *	−0.031	−0.032
	(−0.757)	(−1.892)	(−1.061)	(−1.183)
LEV	2.466 ***	2.478 ***	2.467 ***	2.423 ***
	(11.878)	(11.892)	(11.840)	(11.684)
MTB	0.039	0.034	0.041	0.035
	(1.346)	(1.176)	(1.419)	(1.238)
FOR	−0.948 ***	−1.017 ***	−0.991 ***	−0.945 ***
	(−3.003)	(−3.161)	(−3.130)	(−3.062)
OWN	−0.882 ***	−0.854 ***	−0.855 ***	−0.875 ***
	(−4.068)	(−3.912)	(−3.941)	(−4.051)
Constant	1.885 ***	2.328 ***	1.983 ***	2.084 ***
	(3.372)	(4.192)	(3.365)	(3.908)
Observations	4910	4910	4910	4910
Adjusted R-squared	0.445	0.443	0.444	0.446
Clustering	Firm	Firm	Firm	Firm
YEAR	YES	YES	YES	YES
IND	YES	YES	YES	YES

This table presents the regression results of poor ESG ratings on COD and other control variables. The data include a sample of 672 non-financial firms in Korea. The reported statistics are obtained after clustering the standard errors by the firm, and controlling the industry and year-fixed effects to account for the potential time-series and cross-sectional dependency of each firm. *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively. Variable definitions refer to Appendix A.

, we report the results of an OLS regression examining the impact of poor ESG performance on the cost of debt after controlling other factors that may affect the cost of debt. At a 1% level of significance, Poor_ESG shows a coefficient of 0.472, which supports our first hypothesis that poor ESG leads to an increase in debt costs in the subsequent period. As for other control variables, the coefficient of LEV is significant and positive, suggesting that higher leverage leads to higher debt cost, and those of ROA, FOR, and OWN are negative and significant. According to the regression analyses for the three individual measurements of ESG, the regression results, Poor_S and Poor_G coefficients are all significant and positive at 0.273 and 0.537, respectively, suggesting that both poor social responsibility and corporate governance lead to higher cost of debt, which firmly support our first hypothesis. The Poor_E coefficient, however, is insignificant, which indicates that firms’ costs of debt are not significantly affected by environmental ratings in ESG performance. Above all, the results indicate that our first hypothesis is confirmed, not only in the aggregate poor ESG rating but also in the two individual components of poor governance and social responsibility.

4.2. The Moderating Role of Media Coverage on the Relationship between Poor ESG Performance and the Cost of Debt

Table 5. The moderating effect of media coverage.

$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}MEDI{A}_{i,t}+{\beta }_{3}MEDI{A}_{i,t}\times {\mathit{X}}_{i,t}+{\beta }_{4}SIZ{E}_{i,t}+{\beta }_{5}LE{V}_{i,t}+{\beta }_{6}RO{A}_{i,t}+$ ${\beta }_{7}MT{B}_{i,t}+{\beta }_{8}FO{R}_{i,t}+{\beta }_{9}OW{N}_{i,t}+\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	1.128 ***
	(3.486)
Poor_E		−0.175
		(1.208)
Poor_S			0.732 ***
			(−0.610)
Poor_G				−0.479 *
				(3.202)
MEDIA$\times$Poor_ESG	−0.696 *
	(−1.851)
MEDIA$\times$Poor_E		−0.175
		(−0.610)
MEDIA$\times$Poor_S			−0.479 *
			(−1.723)
MEDIA$\times$Poor_G				−0.921 ***
				(−2.884)
ROA	−6.623 ***	−6.713 ***	−6.673 ***	−6.546 ***
	(−9.735)	(−9.641)	(−9.572)	(−9.559)
SIZE	−0.015	−0.042	−0.022	−0.029
	(−0.492)	(−1.413)	(−0.721)	(−1.054)
LEV	2.355 ***	2.357 ***	2.355 ***	2.310 ***
	(11.326)	(11.211)	(11.239)	(11.056)
MTB	0.065 **	0.063 *	0.068 **	0.058 *
	(1.976)	(1.872)	(2.050)	(1.803)
FOR	−1.153 ***	−1.192 ***	−1.196 ***	−1.142 ***
	(−3.573)	(−3.658)	(−3.679)	(−3.605)
OWN	−1.073 ***	−1.066 ***	−1.047 ***	−1.058 ***
	(−4.607)	(−4.499)	(−4.483)	(−4.570)
Constant	2.000 ***	2.255 ***	2.003 ***	2.366 ***
	(3.485)	(3.957)	(3.473)	(4.388)
Observations	4910	4910	4910	4910
Adjusted R-squared	0.393	0.388	0.390	0.395
Clustering	Firm	Firm	Firm	Firm
YEAR	YES	YES	YES	YES
IND	YES	YES	YES	YES

This table presents the OLS regression results of aggregated poor ESG and the three individual components of poor ESG on COD, conditional on media coverage. MEDIA×Poor_ESG, MEDIA×Poor_E, MEDIA×Poor_S, and MEDIA×Poor_G are the variables of interest. The reported statistics are obtained after clustering the standard errors by the firm, and controlling the industry and year-fixed effects to account for the potential time-series and cross-sectional dependency of each firm. *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively. Variable definitions refer to Appendix A.

reports the moderating effect of media coverage on the relationship between poor ESG and the cost of debt. MEDIA is a dummy variable defined as 1 if media coverage exceeded the median and 0 otherwise. As expected in H2, the result in Table 5 shows that the main coefficient for our variable of interest (MEDIA×Poor_ESG) is −0.696, significantly negative at the 1% significance level. This indicates that media coverage may offset partially the adverse effects of poor ESG practices on debt costs. Furthermore, the individual effects of each of the three ESG rating components on the cost of debt are also reported. The coefficients of MEDIA×Poor_S and MEDIA×Poor_G are −0.479 and −0.921, respectively, which are significant at the conventional level. Conversely, the coefficient of MEDIA×Poor_E is not significant with respect to statistical significance, which is consistent with the result of the insignificant relation between environmental issues and cost of debt shown in Table 4.

The moderating role of media coverage is demonstrated in Figure 1. The red line denoting high media coverage exhibits a flatter slope compared to the blue line representing low media coverage. This indicates that firms subject to high media coverage manifest a diminished correlation between adverse ESG ratings and debt costs.

In sum, companies receiving high media coverage may be able to alleviate the adverse impacts of a poor ESG rating, which firmly supports our second hypothesis. In other words, our test results imply that companies with limited media exposure should place more focus on ESG performance enhancement, as the escalation in debt expenses resulting from unfavorable ESG ratings may not be alleviated.

5. Robustness Test and Additional Analysis

To strengthen our research hypothesis and expand our research fields, we perform several robustness tests and additional analyses.

5.1. Endogeneity Analyses

To resolve the concern of a potential endogeneity problem, mainly stemming from the correlated omitted variables that influence both poor ESG measurements and the cost of debt, we implement the 2SLS and GMM methods. In the relation between Poor_ESG and the cost of debt, we employ the Two-Stage Least Squares (2SLS) and the Generalized Method of Moments (GMM) methodologies in light of Larcker and Rusticus [70] and Ghoul, Guedhami, Kwok, and Mishra [66]. We use the ESG instrument variable with the industry average poor ESG score for the first stage of the 2SLS method. In the second stage of 2SLS, the predicted value of poor ESG score, P_ESG_HAT is generated and used in Equation (1). In terms of industrial classification, we use the Korean Standard Industrial Classification (KSIC), which is consistent with the International Standard Industry Classification. Additionally, the dynamic panel MM approach is also implemented. By using the lagged dependent variable as the instrument variable, we are able to control the impact of the current period’s cost of debt on the cost of debt in the following period.

Based on the aggregate poor ESG and the three individual ESG components,

Table 6. Regression analyses using 2SLS and GMM. (A) The effect of Poor ESG on the cost of debt using 2SLS. (B) The effect of Poor ESG on the cost of debt using GMM.

(A)
$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t}+{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}+$ $\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = P_ESG_HAT	X = P_E_HAT	X = P_S_HAT	X = P_G_HAT
P_ESG_HAT	1.299 ***
	(4.869)
P_E_HAT		0.509 ***
		(2.972)
P_S_HAT			0.693 ***
			(4.298)
P_G_HAT				0.778 ***
				(3.610)
ROA	−6.503 ***	−6.598 ***	−6.516 ***	−6.410 ***
	(−14.608)	(−14.666)	(−14.524)	(−14.330)
SIZE	0.026	−0.023	−0.003	−0.027 *
	(1.201)	(−1.276)	(−0.136)	(−1.700)
LEV	2.588 ***	2.585 ***	2.593 ***	2.568 ***
	(21.932)	(21.851)	(21.961)	(21.948)
MTB	0.075 ***	0.072 ***	0.076 ***	0.065 ***
	(3.756)	(3.565)	(3.784)	(3.310)
FOR	−0.870 ***	−0.962 ***	−0.933 ***	−0.895 ***
	(−5.066)	(−5.524)	(−5.423)	(−5.283)
OWN	−1.065 ***	−1.068 ***	−1.013 ***	−1.020 ***
	(−8.556)	(−8.524)	(−8.198)	(−8.292)
Constant	1.670 ***	2.368 ***	2.078 ***	2.596 ***
	(4.396)	(6.915)	(6.023)	(8.964)
Observations	4910	4910	4910	4910
Adjusted R-squared	0.439	0.440	0.442	0.446
YEAR	YES	YES	YES	YES
IND	YES	YES	YES	YES
(B)
$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t}+{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}+$ $\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	1.756 ***
	(3.657)
Poor_E		−0.240
		(−0.702)
Poor_S			0.875 *
			(1.754)
Poor_G				1.591 ***
				(3.420)
ROA	0.177	−0.689	−0.469	−0.956
	(0.071)	(−0.319)	(−0.201)	(−0.424)
SIZE	−0.086	−0.091	−0.016	0.011
	(−0.489)	(−0.518)	(−0.077)	(0.062)
LEV	4.952 ***	3.976 ***	4.315 ***	4.229 ***
	(6.416)	(4.754)	(5.471)	(5.444)
MTB	−0.165	−0.126	−0.207 *	−0.196 *
	(−1.489)	(−1.150)	(−1.862)	(−1.758)
FOR	−5.934 ***	−7.691 ***	−7.534 ***	−7.484 ***
	(−2.582)	(−3.296)	(−3.487)	(−3.349)
OWN	−6.989 ***	−7.640 ***	−7.210 ***	−6.550 ***
	(−3.765)	(−4.402)	(−3.863)	(−3.440)
Constant	5.775	6.036 *	4.652	4.121
	(1.591)	(1.712)	(1.104)	(1.076)
Observations	4910	4910	4910	4910
Number of STOCK	672	672	672	672
YEAR	YES	YES	YES	YES
IND	NO	NO	NO	NO

(A) presents the empirical results using the 2SLS for aggregate poor ESG and the three individual ESG components. In the column of P_ESG_HAT (IV), P_E_HAT (IV), P_S_HAT (IV), and P_G_HAT (IV), they instrument aggregate and the three individuals’ poor ESG components with the average ESG scores in the same industry. We regress COD on the predicted value of aggregate ESG. (B) presents the empirical results of the GMM methodologies, using the aggregate and the three individual ESG components of X = Poor_ESG, Poor_E, Poor_S, and Poor_G. *, *** indicate significance at 10% and 1% levels, respectively. Variable definitions refer to Appendix A.

presents the empirical results of the 2SLS and GMM models. As shown in Panel A, poor ESG performance is associated with a higher cost of debt given that P_ESG_HAT examines a coefficient of 1.299 at a significance level of 1%. The coefficients of P_E_HAT, P_S_HAT, and P_G_HAT are all positive and significant. Using the GMM method, Panel B shows that the coefficients of Poor_ESG, Poor_S, and Poor_G are significant and positive, while the coeffect of Poor_E does not show any significance, in accordance with the results in Table 4. Using 2SLS and GMM methods, we demonstrated a positive correlation between poor ESG scores and the cost of debt after controlling for endogeneity issues. The Hansen J-test results indicate that our instruments are valid.

5.2. ESG and Credit Rating

The positive relationship between credit rating and good ESG performance is examined by Kim and Lee [71] under the context of the Korean market. We use the bond rating of the following period as the substitute for the cost of debt in Equations (1) and (2). Credit rating is calculated as the lowest rating of the bonds issued by the same issuer. The higher the rating is, the better the credit quality of the issuer is. For example, AAA is assigned with 26 and D with 1.

Table 7. Alternative proxy for cost of debt: bond credit rating. (A) Test of H1 using bond credit rating as a dependent variable. (B) Test of H2 using bond credit rating as a dependent variable.

(A)
$CR{B}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t}+{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}+$ $\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	−2.269 ***
	(−6.359)
Poor_E		−0.717 **
		(−2.301)
Poor_S			−1.762 ***
			(−5.870)
Poor_G				−1.975 ***
				(−6.398)
Observations	1535	1535	1535	1535
Adjusted R-squared	0.733	0.726	0.732	0.733
Controls	YES	YES	YES	YES
IND	YES	YES	YES	YES
YEAR	YES	YES	YES	YES
(B)
$CR{B}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}MEDI{A}_{i,t}+{\beta }_{3}MEDI{A}_{i,t}\times {\mathit{X}}_{i,t}+{\beta }_{4}SIZ{E}_{i,t}+{\beta }_{5}LE{V}_{i,t}+{\beta }_{6}RO{A}_{i,t}+$ ${\beta }_{7}MT{B}_{i,t}+{\beta }_{8}FO{R}_{i,t}+{\beta }_{9}OW{N}_{i,t}+\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	−4.224 ***
	(−7.281)
Poor_E		−1.442 ***
		(−3.269)
Poor_S			−3.818 ***
			(−7.068)
Poor_G				−4.121 ***
				(−7.281)
MEDIA$\times$Poor_ESG	2.772 ***
	(4.285)
MEDIA$\times$Poor_E		1.165 **
		(2.361)
MEDIA$\times$Poor_S			2.708 ***
			(4.565)
MEDIA$\times$Poor_G				2.878 ***
				(4.539)
MEDIA	1.238 ***	0.579 **	1.358 ***	1.218 ***
	(4.145)	(2.432)	(4.527)	(4.281)
Observations	1535	1535	1535	1535
Adjusted R-squared	0.736	0.727	0.735	0.736
Controls	YES	YES	YES	YES
IND	YES	YES	YES	YES
YEAR	YES	YES	YES	YES

This table shows the regression results of poor ESG and bond credit rating, conditional on media coverage. ESG×MEDIA is the variable of interest. **, *** indicate significance at 5% and 1% levels, respectively. Variable definitions refer to Appendix A.

represents the results of poor ESG’s impact on bond credit rating. Poor_ESG is our variable of interest in Panel A, which displays a coefficient of −2.269, indicating that a low ESG score is associated with a lower bond’s credit rating. The results are consistent with Polbennikov, Desclée, Dynkin, and Maitra [5]. Panel B confirms our previous test results in Table 5 by documenting the moderating role of media coverage in the relationship between poor ESG rating and bond credit rating. The positive and significant interaction terms of Poor_ESG and MEDIA indicate that when bond issuers are covered by more media articles, they are less likely to obtain lower credit ratings due to their poor ESG performance.

5.3. Under-Attention and Over-Attention Control for Market Perception

The market-to-book ratio can provide insight into the market’s perception of the company’s growth prospects and risks. Firms with high market-to-book value may be perceived as being more financially successful, which could lead to greater media coverage. Therefore, controlling for the effect of MTB as a measure of growth prospects on media coverage could help to isolate the specific effect of media coverage on this relationship between ESG and the cost of debt. To examine the moderating effect of media coverage on the relationship between ESG performance and the cost of debt while controlling for the potential confounding effect of MTB, we conduct the regression and use Equation (3) as follows:

$$Medi{a}_t={\beta }_0+{\beta }_{1}MT{B}_t+{\epsilon }_t$$

(3)

and predict the residual by year. The positive residuals of the regression represent media over-attention, while negative residuals indicate that firms are receiving under-attention from the media. We convert the over-attention into a dummy variable by assigning it a value of 1.

The findings presented in

Table 8. The moderating effect of media over-attention.

$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}Over\text{\_}Atte{n}_{i,t}+{\beta }_{3}Over\text{\_}Atte{n}_{i,t}\times {\mathit{X}}_{i,t}+{\beta }_{4}SIZ{E}_{i,t}+{\beta }_{5}LE{V}_{i,t}+$ ${\beta }_{6}RO{A}_{i,t}+{\beta }_{7}MT{B}_{i,t}+{\beta }_{8}FO{R}_{i,t}+{\beta }_{9}OW{N}_{i,t}+\sum Year+\sum Industry+{\epsilon }_{i,t}$
VARIABLES	X = Poor_ESG	X = Poor_E	X = Poor_S	X = Poor_G
Poor_ESG	0.800 ***
	(4.807)
Poor_E		0.090
		(0.738)
Poor_S			0.478 ***
			(3.285)
Poor_G				0.904 ***
				(5.740)
Poor_ESG$\times$Over_Atten	−0.575 ***
	(−2.858)
Poor_E$\times$Over_Atten		−0.148
		(−1.013)
Poor_S$\times$Over_Atten			−0.343 **
			(−1.992)
Poor_G$\times$Over_Atten				−0.610 ***
				(−3.141)
Over_Atten	−0.247 ***	−0.102 *	−0.173 **	−0.268 ***
	(−3.201)	(−1.728)	(−2.472)	(−3.449)
ROA	−6.435 ***	−6.487 ***	−6.476 ***	−6.375 ***
	(−25.222)	(−25.325)	(−25.419)	(−25.023)
SIZE	−0.032 **	−0.056 ***	−0.037 **	−0.038 **
	(−1.975)	(−3.480)	(−2.227)	(−2.534)
LEV2	2.597 ***	2.588 ***	2.597 ***	2.577 ***
	(26.846)	(26.711)	(26.808)	(26.671)
MTB	0.068 ***	0.068 ***	0.070 ***	0.063 ***
	(4.729)	(4.725)	(4.831)	(4.377)
FOR	−0.959 ***	−0.986 ***	−0.976 ***	−0.926 ***
	(−5.768)	(−5.928)	(−5.867)	(−5.577)
OWN	−1.019 ***	−1.019 ***	−1.007 ***	−1.011 ***
	(−9.472)	(−9.420)	(−9.355)	(−9.424)
Constant	2.733 ***	2.993 ***	2.734 ***	2.906 ***
	(7.546)	(8.339)	(7.523)	(8.505)
Observations	4910	4910	4910	4910
Adjusted R-squared	0.446	0.443	0.444	0.447
IND	YES	YES	YES	YES
YEAR	YES	YES	YES	YES

This table reports the OLS regression results of aggregated poor ESG and the three individual components of poor ESG on COD, conditional on media over-attention. The variable Over-atten is set to 1 for firms with media over-attention, and 0 for firms with media under-attention. Poor_ESG×Over_Atten, Poor_E×Over_Atten, Poor_S×Over_Atten, and Poor_G×Over_Atten are the variables of interest. *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively. Variable definitions refer to Appendix A.

reveal a noteworthy interaction between Poor_ESG and Over_Atten, characterized by a negative coefficient with statistical significance. This suggests that over-media attention alleviate the adverse impact of low ESG ratings on the cost of debt. This implies a complex interplay between media visibility, ESG performance, and the financial burden for firms, wherein heightened media attention appears to offer a form of insulation against higher debt costs associated with poor ESG performance.

5.4. The Moderating Effect of Media Coverage: Grouping Analysis

We grouped our sample based on the median of media report article numbers. The report is shown in

Table 9. The impact of poor ESG rating on the cost of debt: grouped by media coverage.

$CO{D}_{i,t+1}={\beta }_0+{\beta }_1{\mathit{X}}_{i,t}+{\beta }_{2}SIZ{E}_{i,t}+{\beta }_{3}LE{V}_{i,t}+{\beta }_{4}RO{A}_{i,t}+{\beta }_{5}MT{B}_{i,t}+{\beta }_{6}FO{R}_{i,t}+{\beta }_{7}OW{N}_{i,t}+\sum Year+\sum Industry+{\epsilon }_{i,t}$
	High Media Coverage	Low Media Coverage	High Media Coverage	Low Media Coverage	High Media Coverage	Low Media Coverage	High Media Coverage	Low Media Coverage
VARIABLES	X = Poor_ESG		X = Poor_E		X = Poor_S		X = Poor_G
Poor_ESG	0.335 **	0.731 ***
	(2.499)	(3.703)
Poor_E			0.031	−0.017
			(0.279)	(−0.114)
Poor_S					0.267 **	0.528 ***
					(2.400)	(3.025)
Poor_G							0.338 ***	0.864 ***
							(2.980)	(4.764)
ROA	−4.575 ***	−6.658 ***	−4.575 ***	−6.710 ***	−4.551 ***	−6.691 ***	−4.533 ***	−6.556 ***
	(−12.915)	(−18.117)	(−12.871)	(−18.218)	(−12.846)	(−18.198)	(−12.798)	(−17.815)
SIZE	−0.042 **	−0.043	−0.064 ***	−0.080 ***	−0.043 **	−0.046	−0.050 ***	−0.056 **
	(−2.367)	(−1.443)	(−3.631)	(−2.725)	(−2.396)	(−1.537)	(−3.110)	(−1.990)
LEV	2.864 ***	2.532 ***	2.849 ***	2.561 ***	2.861 ***	2.552 ***	2.850 ***	2.516 ***
	(23.454)	(16.734)	(23.286)	(16.853)	(23.435)	(16.869)	(23.390)	(16.653)
MTB	−0.024	0.157 ***	−0.026	0.153 ***	−0.023	0.159 ***	−0.026	0.147 ***
	(−1.504)	(5.908)	(−1.609)	(5.737)	(−1.444)	(5.950)	(−1.642)	(5.548)
FOR	−0.839 ***	−0.850 ***	−0.877 ***	−0.883 ***	−0.847 ***	−0.875 ***	−0.837 ***	−0.789 ***
	(−4.506)	(−2.968)	(−4.714)	(−3.076)	(−4.550)	(−3.055)	(−4.501)	(−2.756)
OWN	−0.648 ***	−1.181 ***	−0.637 ***	−1.156 ***	−0.638 ***	−1.157 ***	−0.635 ***	−1.173 ***
	(−5.077)	(−6.686)	(−4.970)	(−6.470)	(−4.997)	(−6.549)	(−4.977)	(−6.660)
Constant	2.419 ***	3.001 ***	2.758 ***	3.525 ***	2.425 ***	3.008 ***	2.586 ***	3.354 ***
	(6.268)	(4.610)	(7.093)	(5.398)	(6.263)	(4.571)	(7.160)	(5.273)
Observations	2481	2429	2481	2429	2481	2429	2481	2429
Adjusted R-squared	0.487	0.470	0.485	0.467	0.486	0.469	0.487	0.472
IND	YES	YES	YES	YES	YES	YES	YES	YES
YEAR	YES	YES	YES	YES	YES	YES	YES	YES

This table shows the regression results of poor ESG ratings on COD and other control variables, grouped by high and low media coverage. The data include a sample of 672 non-financial firms in Korea. The reported statistics are obtained after controlling the industry and year-fixed effects to account for the potential time-series and cross-sectional dependency of each firm. **, *** indicate significance at 5% and 1% levels, respectively. Variable definitions refer to Appendix A.

. We construct two subsamples based on the median value of media coverage—high versus low media coverage—and run a regression on the different subsamples using the equation for H1. This test is designed to test and compare the significance of the two coefficients of ESG variables in high versus low media coverage subsamples. As predicted in H2, we find that the coefficients of ESG in high media coverage samples are greater than those in low media coverage ones. Based on these comparisons, we reconfirm that firms with high media coverage receive lower debt costs from their debt holders for insufficient ESG engagement.

6. Conclusions

This paper examines the relationship between corporate irresponsibility and the cost of debt, specifically focusing on the moderating role of media coverage. Our empirical results provide robust evidence that a poor ESG rating leads to an increase in the cost of debt, aligning with previous research [72]. However, the key contribution of our study lies in uncovering a novel finding: firms with higher media coverage are less likely to experience a significant rise in their debt costs despite possessing low ESG ratings.

This finding fills an important gap in the ESG literature, as previous research has primarily focused on the direct relationship between ESG performance and debt financing costs [73]. By introducing the role of media coverage, we shed new light on the complex interplay between corporate social responsibility, media scrutiny, and financial outcomes. Specifically, our results suggest that media coverage can serve as a protective mechanism, mitigating the negative financial consequences of poor ESG performance for firms. In essence, this paper offers a valuable contribution to the existing Environmental, Social, and Governance (ESG) scholarship by emphasizing the pivotal role of media coverage in tempering the impact of corporate irresponsibility on debt costs. A noteworthy discovery is that enterprises with a more extensive media presence are somewhat shielded from the financial implications of low ESG ratings, thus enriching our comprehension of the intricate interplay between corporate social responsibility and financial outcomes.

Furthermore, we implement a series of additional analyses to make our results more persuasive. Our results are robust when we use the 2SLS and GMM methods to control for endogeneity. We employ the corporate bond rating as an ex-ante proxy of debt financing costs and find that firms exhibiting poor ESG ratings are associated with lower bond credit ratings. Media coverage shows the moderating effect in the relation between poor ESG ratings and bond credit ratings, implying that the moderating effect of media coverage also operates in the relation between poor ESG ratings and bond credit ratings. Also, we extend the concept of media coverage using media over-attention to test the moderating effect and find that firms with media over-attention have lower costs of debt due to unfavorable ESG ratings. In addition, we construct the subsamples based on the median of number of media articles to examine how the impact of ESG ratings on the cost of debt is different for high and low levels of media coverage groups.

As a limitation of our study, it is noteworthy that among the individual components of ESG, the impact of environmental rating on the cost of debt is, in part, insignificant in our analyses. This finding suggests that environmental performance and debt costs may be more complex than initially anticipated and deserve further exploration. One potential reason for the insignificance of environmental ratings could be the challenge in measuring and evaluating environmental performance due to the diverse practices and their varying impacts. Future research could focus on unpacking the complexities of the relationship between environmental factors and the cost of debt, considering various contextual factors, and employing more nuanced measures of environmental performance. Moreover, as a prospective research endeavor, it would be immensely valuable to delve into the influence of media coverage on ESG impacts, not just in Korea but also across the United States and European nations. Such an investigation would significantly contribute to the broader generalization of this crucial topic.