An Empirical Study on the Relationship between Corporate Social Responsibility and Default Risk: Evidence in Korea

Abstract

Focusing on the Korean stock market, this paper expands on previous research on the relationship between corporate social responsibility (CSR) and default risk. By using a comprehensive sample of 6977 firm-year observations during the 2011–2021 period, we employ the year fixed effects and industrial sector dummies classified by FnGuide Industry Classification Standard to control for shifting economic conditions over time and effects within industries. The Z-score is employed as the primary proxy for default risk, and the finding of the study confirms a negative association between CSR practices and default risk. Moreover, testing the three components of CSR, we also find that each component has a negative impact on the default risk. To ensure the robustness of our findings, we conduct a robustness check using two additional proxies of default risk: the K-score, a specific measure of default risk for the Korean market, and the distance to default (DTD), a market-based model. Our results remain consistent and robust even when utilizing alternative proxies, further confirming the negative relationship between CSR and default risk. This has significant implications for businesses and regulators who aim to decrease the risk of default through implementing CSR initiatives.

1. Introduction

The primary purpose of this study is to investigate the relationship between corporate social responsibility and the default risk of the firm. Attention is given to addressing this linkage because commitment to corporate social responsibility (CSR) activities should not only increase stakeholder value but also strengthen a firm’s ability to reduce risks [1]. While there are numerous studies that have investigated the association between CSR and risks faced by firms, there has been only scant literature available on examining the nexus between CSR and default risk [2].

Investigating the link between CSR and a company’s financial risk using various risk measures, a set of prior studies has discovered a negative relationship. These studies have shown that CSR can decrease the risk of financial distress, increase a firm’s value, reduce idiosyncratic and systematic risk, protect against legal risks, and reduce the risk of stock price crashes by promoting transparency and avoiding negative news. Moreover, higher CSR practices are associated with lower default risks and improved creditworthiness, providing access to more funding [3,4,5,6,7,8]. These findings demonstrate the significance of CSR in managing a company’s financial risk.

Nevertheless, only a limited number of empirical studies have exhibited a positive association between CSR and a firm’s risk. Preceding studies have found that higher social responsibility activities expose companies in the U.S. market to higher risks [9,10]. Assessing a firm’s idiosyncratic risk, it has been found that, while CSR activities can increase idiosyncratic risk, they also reduce risks for stakeholders [11]. Despite the mixed results, previous empirical studies have affirmed that CSR has a significant impact on a company’s risk, making it worthwhile to explore further the relationship between CSR and a firm’s likelihood of default [12].

In this study, our focus is to reveal the link between CSR and default risk in the South Korean stock market, where most previous studies have focused on the effect of CSR on firm values [13,14,15,16,17]. The unique governance system of Chaebol [15,18], the large family-controlled conglomerates that dominate the South Korean economy, makes this market particularly interesting to study. South Korea’s distinct economic, cultural, and political environment provides an opportunity to examine this relationship.

Previous research has yielded mixed results regarding the connection between CSR and a company’s default risk. This study aims to make substantial contributions to the existing literature by addressing the relationship between corporate social responsibility (CSR) and default risk in the South Korean stock market. Firstly, to the extent of our understanding, it addresses a gap in the literature by providing a comprehensive analysis of default risk using three distinct methods of measurement, including the Altman Z-score, K-score, and Distance to Default (DTD). The Altman Z-score, as the main proxy for our analysis, is an accounting-based measure, while the K-score and DTD are used in robustness checks. Secondly, the study also evaluates the impact of each component of CSR (environmental, social, and governance) on default risk, providing valuable insights into the significance of each component. Finally, the study also presents robustness tests to confirm the consistency of the results across multiple proxies for default risk. In the context of the Korean market, employing the K-score and DTD as other measures of default risk in our study provides an informative comparison with the Altman Z-score. The K-score model, as another accounting-based measure, was also established by Altman et al. in 1995 to specifically predict the failure of Korean firms [19]. The DTD, as a market-based model, has long been considered as an essential indicator of a company’s default risk, and is examined by adopting Merton’s structural model, established in 1974, with the same suppositions on debt maturity and also size as in the Kealhofer, McQuown, and Vasicek (KMV) implementation [20]. As a result, this study seeks to contribute a more in-depth understanding of the relationship between CSR and default risk in the unique context of the Korean market.

In brief, the finding of this study verifies that CSR performance, as measured by a company’s ESG score, is negatively linked with default risk. It aligns with the findings of prior literature [2,21,22,23,24,25]. It suggests that a firm with a high ESG profile has low financial distress or low default risk levels. In other words, firms are able to lessen their default risk by implementing CSR activities. Additionally, after examining the three components of ESG (environmental, social, and governance), our findings present that the social and governance aspects have a consistently negative and significant effect on default risk across all three measures used. Only one measure showed a significant negative link between the environmental component and default risk. Overall, the study provides strong evidence that CSR activities can effectively mitigate default risk in the Korean market.

2. Literature Review and Hypothesis Development

The link between CSR and a firm’s default risk has not been extensively studied despite the extensive investigations on the relationship between CSR and financial performance [2,10]. The connection between the two can be explained by two opposing theories. From a shareholder theory perspective, which is based on the idea of agency problems, the concept of corporate social responsibility (CSR) is viewed from a singular economic perspective, with the central aim being the maximization of shareholder wealth. This perspective raises concerns regarding the potential for agency conflict, where the interests of shareholders may not align with those of corporate management [26,27]. Additionally, CSR is considered as a deviation from the main objective of a company, which is to make profit for itself and its shareholders, rather than to engage in socially responsible activities or charitable causes [28]. On the other hand, the stakeholder theory argues that a firm should consider the needs of various parties that have an impact on the firm’s performance [11,28,29,30,31,32]. This theory, proposed by Freeman, has become a widely utilized base for further research, although it has been criticized for only focusing on the relationship between the companies and the stakeholders [33]. Also, to promote a good relationship between firms and stakeholders, CSR is acknowledged as an effective device for boosting relationships with stakeholders and enhancing a company’s reputation and overall performance in society and the market [34]. Furthermore, the implementation of socially responsible and environmentally sustainable practices by businesses can lead to cost savings through a reduction in production-related environmental risks, to the improvement of relationships with stakeholders, and ultimately to enhanced competitiveness and long-term financial success [30].

Both perspectives acknowledge the notion that addressing the needs and expectations of stakeholders is a critical aspect of a company’s business strategy. This aligns with the implementation of CSR initiatives within a corporate context as a means of effectively catering to the interests of stakeholders. In 2021, the Commission of the European Communities described CSR as a concept whereby firms incorporate their social and environmental involvements in their operations and their interactions with stakeholders voluntarily [35]. From the discussion, we agree with the statement that satisfying stakeholder wealth is a crucial key to a business. Thus, the implementation of CSR in business is beneficial to stakeholders.

In researching the correlation between CSR and firm risk, multiple studies present varying findings. Some studies have demonstrated a negative relationship between CSR and firm risk. A theory posits that, due to a product differentiation strategy, companies with strong CSR practices face lower price-sensitive demand, leading to reduced systematic risk [4]. Another study utilizing a unique method to assess ESG scores found that CSR initiatives are linked to better bond ratings, which can be influenced by investor clientele effects, local competition, and/or social interactions [23]. One study showed that activism focused on environmental CSR initiatives may lower a firm’s downside risk [36]. The adoption of CSR practices can also decrease distress and default risk, improving a company’s financial stability and creating a more attractive corporate environment and resilient economy during crises [2].

In contrast to the studies that confirm an inverse association between CSR and firm risk, there are a limited number of studies that show evidence of a positive association. In the context of US market, previous studies have found that excessive investment in CSR can lead to increased risk, a phenomenon that is commonly perceived as having a detrimental effect on shareholder value [9]. The prevalence of agency conflicts, which are often rooted in managerial actions driven by self-interest at the expense of other stakeholders, serves as a contributing factor to this phenomenon. Additionally, the presence of strengths in diversity and employee relations also exposes the firm to elevated risk. As a general principle, a commitment to employees results in a fixed cost, thereby transferring business risk to the firm’s shareholders [10]. Moreover, a study carried out previously supports the idea that CSR practices increase the idiosyncratic risk of a firm, but that, at the same time, they decrease stakeholder risk. This happens because CSR activities limit a firm’s ability to respond to negative productivity shocks, which then decreases stakeholder well-being. It leads to higher CSR stocks having less predictable returns [11].

In the Korean market, research has been conducted on the relationship between CSR and the association between distress and earnings management. The findings indicate that CSR acts as a moderating factor, diminishing the positive influence of distress on earnings management, with the environmental aspect emerging as the subcategory of CSR with the most significant impact [37]. Another study found that CSR is negatively related to credit risk, with social factors having a substantial role in reducing credit risk [38]. In general, if CSR activities reflect management quality, those who engage in them are less likely to experience distress risk [39].

Based on the risk management standpoint, a firm dealing with default risk can gain positive moral capital for various stakeholders that could be the safeguard for the firm’s management [32,40]. Moreover, CSR activities as an emerging trend can be viewed as an essential part of a firm’s risk management [41,42]. Capital constraints can be minimized by enhancing CSR action by including higher stakeholder involvement as well as by enhancing clarity, and reporting regarding CSR practices should be established [22].

We focus on examining the relationship between CSR and default risk specifically in the Korean stock market. In line with preceding studies, it is expected that companies that engage in positive CSR practices will have a reduced likelihood of default risk. Firms that prioritize quality CSR are less prone to default risk due to their effective governance and management. We, in addition, also elaborate on the partial influence of the ESG components on a company’s default risk. The aim is to determine which of the environmental, social, and governance components has the most significant impact on default risk. Our study also employs two alternative approaches—using the K-score and DTD measurement—which are accounting- and market-based models of default risk, respectively. Taken as a whole, it is expected that the results will affirm the conclusion that CSR practices can mitigate the risk of default of a firm.

Therefore, the hypothesis of this study is developed as follows:

Hypothesis 1.

The CSR performance of a firm, which is measured by its ESG score, is negatively associated with the default risk of the firm.

3. Data and Methodology

3.1. Data

The main purpose of the study is to test whether CSR performance and default risk have a negative relationship with each other. Since we are investigating the Korean stock market, the CSR data measured by the ESG score was obtained from the KCGS (Korea Institute of Corporate Governance and Sustainability), whereas the financial data to calculate the default risks and all the control variables were acquired from two data providers, the FnGuide database and CRI (Credit Research Initiative) database.

The sample for the study included all Korean firms except financial firms between the years 2011 and 2021. These firms were excluded from the sample since this sector has specific capital structure requirements [43]. Additionally, a high level of leverage is considered as a normal circumstance for financial firms, while for non-financial firms, higher leverage is more likely to reflect the firm’s level of distress [44].

Since CSR initiatives do not instantly impact a firm’s financial risk, we decided to factor in a one-year lag time in our analysis. This approach is based on the understanding that CSR typically does not result in immediate impacts [31,45,46], and that stakeholders need time to incorporate new information into their decisions. By doing so, we assumed a non-direct impact of a firm’s CSR activities on default risk, meaning that the CSR performance at the end of time t predicts the default risk of the firm at the end of time t + 1.

3.1.1. Corporate Social Responsibility

The definition of CSR and its scope are very diverse and lack uniformity. To measure CSR, some studies use the standardized ESG scores provided by international data collecting agencies, such as Bloomberg, Thomson Reuters Eikon, and the MSCI KLD index, to minimize measurement bias. The KLD index is the most widely used data source and has been commonly utilized in recent studies [15,47]. However, these databases have limitations as they do not offer detailed ESG information about Korean companies [17]. To examine the Korean market, most previous studies have used the KEJI index provided by Korea Economic Justice Research Institute [13,16,18,31] and KCGS (Korea Institute of Corporate Governance and Sustainability) [15,17,18,37,48,49].

We collected the ESG data from KCGS considering that this database has a key advantage over other databases in Korea. Since it examines all the KOSPI-listed companies and selected KOSDAQ companies, it provides wider coverage of Korean firms’ ESG data t the other index, KEJI, which has been mainly used in previous studies. Unlike KEJI, which only provides data on 200 annual firms, usually large-size firms [15,17], the KCGS database covers information on approximately 900 firms [14]. Furthermore, the latter’s ESG scores are used in several SRI indexes in the Korea Stock Exchange (KRX) [37]. Based on KCGS, the ESG scores of Korean firms, as well as each firm’s environmental, social, and corporate governance practices, have been available since 2011. The score of a firm’s environmental performance is evaluated based on categories such as environmental management, performance, and stakeholder relations. The social performance score is reviewed based on categories including suppliers and competitors, consumers, and the local community. The corporate governance score is evaluated based on categories such as shareholder rights protection, the board of directors, auditing body, and disclosure.

There are underlying reasons why prior studies have suggested adopting ESG scores as a proxy for CSR performance. The KCGS database only disclosed the letter grade of a company’s CSR performance, without providing detailed scores, prior to 2010. Since 2011, both the letter grade and numeric score data have been made available. However, the limitation of the ESG letter-grade-based system is that researchers cannot accurately determine a firm’s exact CSR score within a specific letter grade. [17]. For example, grade “A” of a firm refers to a specific range of ESG total scores, starting from 4.50, 4.55, and 4.60 to 5.45. On the other hand, using the ESG scoring system as a proxy for CSR performance could explicitly reflect the value of CSR performance. KCGS data offers access to numeric scores on which the letter-grade ratings are based [14]. Therefore, based on prior studies that have employed ESG scores as their proxy for CSR performance [17,37,50], this study also adopted the ESG score as a proxy for the CSR variable. The overall ESG score used in this study specifically refers to the final score provided by KCGS.

Table 1. Sample distribution of overall ESG score and each ESG component by year.

		ESG Score	Environmental	Social	Governance
Year	N	Mean	Mean	Mean	Mean
2011	306	2.18	0.22	0.26	0.36
2012	680	1.97	0.35	0.29	0.35
2013	745	2.81	0.42	0.34	0.34
2014	718	2.86	0.42	0.35	0.35
2015	711	2.72	0.30	0.26	0.22
2016	738	2.80	0.30	0.30	0.23
2017	755	2.58	0.27	0.22	0.27
2018	751	2.50	0.27	0.24	0.26
2019	792	2.61	0.23	0.25	0.28
2020	796	2.82	0.25	0.29	0.30

shows the statistics overall ESG scores and each ESG component score by year. The final sample of this study is an unbalanced panel of data on 6977 firm-year observations, reflecting 958 unique Korean firms (excluding financial firms) between 2011 and 2020. Of note is that the ESG score was obtained from KCGS data. The overall score, which is used in this study, refers to the final score provided by KCGS. The detailed formula and weights used by KCGS to calculate overall ESG performance are not publicly disclosed [14]. As a result, our study provides environmental, social, and governance scores, which are partially scaled down to 1:100 for calculation purposes [51]. A higher score for ESG indicates better ESG activities by the companies. On average, the highest and lowest ESG scores found in 2014 and 2012 were 2.86 and 1.97, respectively.

The environmental score, which reflects how firms take actions to protect and minimize damage to the environment, had the same highest score in 2013 as in 2014, which was 0.42. However, since 2015, a downtrend steadily happened until 2020. The social score represents how firms treat their employees and the communities they serve. Between 2011 and 2016, the average social score gradually fluctuated. Although it significantly dropped in 2017, a steadily increasing trend was shown in the subsequent years. The governance score reflects how firms’ managements direct and supervise their organizational authority. The average score slightly fluctuated until 2014 and suddenly bottomed out in 2015. However, since 2016, the trend gradually increased.

3.1.2. Default Risk

Based on a wide body of prior studies, there are two commonly used approaches to measure a firm’s default risk. The first approach is an accounting-based model [19,52,53,54,55,56], and the second approach is a non-accounting-based model, namely a market-based model [57]. Many preceding works have predicted financial distress risk by applying accounting and market-based measurements [53,58,59]. To calculate default risk, our study adopted two accounting-based measures (Z- and K-score) and a market-based measure (DTD). The Altman Z-score was the primary proxy for the analysis, while the K-score and Distance to Default (DTD) served as robustness tests.

As an accounting-based measurement, the Altman Z-score has been used as the main proxy for default risk by previous studies. However, there are differing perspectives regarding the use of the Z-score as a measure of distress or default risk. Although the Z-score by Altman surpasses the hazard and other models to predict a firm’s bankruptcy using an international dataset [60,61], the accounting statements are of limited sufficiency since they are formulated on a going-concern basis [62]. Other works have criticized the Z-score as a not-really-fit model, stating that the coefficients in the Z-score calculation weaken the predictive ability to the point where it does not perform better than its most predictive predictor ratio [63,64]. Despite differing views, a significant amount of literature still features the use of the Z-score model as the primary essential proxy in the research [53,58,59].

To obtain more meaningful results and increase the robustness of the analysis, two other proxies for measuring the default risk of the firm, the K-score and the Distance to Default (DTD), were used in this study. We obtained financial data from two data providers based on the default risk measurements. Data for the first two proxies (the Z-score and K-score) were derived from the FnGuide database as the reliable and comprehensive data provider for the Korean stock market. Meanwhile, data for the DTD measurement were collected from the CRI database by the National University of Singapore’s Risk Management Institute (RMI). This database has been used in prior studies that have focused on default risk [25].

• Z-score by Altman (1968) [52]

The first model to address the research problem is the Altman Z-score, 1968, as an accounting-based model. The equation of the Z-score is formulated as follows [52]:

Z = 0.012X1 + 0.014X2 + 0.033X3 + 0.006X4 + 0.999X5,

(1)

where Z is the Z-score, X1 is the working capital/total assets, X2 is the retained earnings/total assets, X3 is the earnings before interest and taxes/total assets, X4 is the market value of equity/market value of total liabilities, and X5 is the sales/total assets. Each X1 to X5 reflects the liquidity, profitability, leverage, solvency, and activity of the firm. According to this model, a firm is predicted to fall into “the safe zone or strong financial condition” when the score is higher than 2.99, whereas it is predicted to fall into “the distress zone or high probability of bankruptcy” when the score is lower than 1.81. To explain the negative relationship between CSR and default risk, we multiply the Z-score by −1 (negative one) as a measure of financial distress risk. It is important to confirm that a greater negative value for the Z-score variable suggests a lower level of default risk.

2.

K-score by Altman et al. (1995) [19]

We adopt another bankruptcy model, which was established particularly for a sample set of Korean companies by Altman et al. in 1995, as the alternative test of this study. According to their study, there are two model classifications established for the Korean market. The first model, the K1-score, is used only for public and private firms, whilst the second model, the K2-score, is used for publicly traded firms [19]. These accounting-based models perform with excellent accuracies in the first two years before the default. Since the sample of this study focuses on publicly traded Korean firms, we adopted the K2-score model and we simplified the terms of this model as the K-score. The formula is written as follows:

K-score = −18.696 + 1.501X1 + 2.706X2 + 19.760X3 + 1.146X4,

(2)

where K-score represents K2-score as a default risk’s proxy, X1 is log (total assets), X2 is log (sales/total assets), X3 is retained earnings/total assets, and X4 is the market value of equity/total liabilities. X1 to X4 represents the size, asset turnover, profitability, and leverage of a firm. Under this model, when the score is higher than 0.75, the firm is determined as “sound firm or strong financial health”, whereas when the score is lower than −2.3, the firm is determined as “insolvent firm or high probability of bankruptcy”. Using the same approach as the previous model, we multiplied the K-score by −1 (negative one) to evaluate the default risk. A greater negative value for the K-score variable suggests a lower level of default.

3.

Distance to Default (DTD)

Lastly, we utilized distance-to-default (DTD) as a market-based model that has long been observed as an essential indicator of a firm’s credit quality in relation to the default risk. It is determined by using Merton’s structural model [20], assuming that the firm value follows the Wiener process. The equity value, as a function of the total firm value, can be described by the Black–Scholes–Merton formula [65]. The default exists only if the value of the firm is less than the face value of debt, and it will happen only during the maturity period. The formula of DTD is defined as follows:

$${\mathrm{DTD}}_{\mathrm{t}}=\frac{\log \left(\frac{V_t}{L}\right)+\left(\mu –\frac{{\sigma }^2}{2}\right)\left(T–t\right)}{\sigma \sqrt{T-t}}$$

(3)

where $V_t$ is the asset value reflecting a geometric Brownian motion with drift ($\mu$) and volatility ($\sigma$), and L is the default point. There have been many variants of this model to estimate. We download the data estimated by NUS-CRI. NUS-CRI modifies the model by doing certain treatments on its own DTD computation to address the several drawbacks. It is not easy to estimate the $\mu$ with reasonable precision unless the very long time span of data is calculated [66]. The expression of the DTD forming by NUS-CRI is:

$${\mathrm{DTD}}_{\mathrm{t}}=\frac{\log \left(\frac{V_t}{L}\right) }{\sigma }$$

(4)

where the default point is $L=\mathrm{current}\mathrm{liabilities}+0.5\times \mathrm{long}-\mathrm{term}\mathrm{liabilities}$, and the maturity of liabilities is set to one year. For more details, refer to the CRI technical report on the CRI website.

Theoretically, firms with higher DTDs will have lower default risks. Using the same approach as the two previous models, we multiplied the score by −1 to be consistent with the other measures used; a greater negative DTD score suggests a lower default risk level.

Table 2. Average of default risks by year.

		Z-Score	K-Score	DTD
Year	N	Mean	Mean	Mean
2012	306	−2.54	−2.09	−2.86
2013	680	−2.8	−1.65	−3.7
2014	745	−3.36	−2.7	−4.49
2015	716	−4.28	−5.08	−4.14
2016	708	−3.59	−3.63	−4
2017	736	−3.91	−4.27	−4.96
2018	754	−3.35	−2.84	−4.35
2019	750	−3.12	−2.6	−3.9
2020	789	−3.76	−3.58	−3.41
2021	793	−3.62	−3.48	−3.69

Note: The negative average values of each score are determined to explain the negative association between CSR and default risk by multiplying each proxy’s score of default risk by −1 (negative one).

shows the average value of each proxy of default risk, which is measured by the Z-score, K-score, and DTD, by year. As we explained before, considering a non-direct effect of a firm’s CSR activities on default risk, we assume that the default risk at the end of time t + 1 is predicted by the CSR performance at the end of time t. Thus, Table 2 shows the proxies of default risk statistics from 2012 to 2021 including the mean, standard deviation, minimum, and maximum values of the distribution across companies. The sample comprises 6977 South Korean firm-year observations for the sample period. As stated earlier, the negative values for the Z-score, K-score, and DTD are determined to explain the negative relationship between CSR and default risk by multiplying the proxies by −1 (negative one). The high negative score refers to the low risk of default.

Moreover, in consonance with prior studies [2,67], a set of control variables that influence a firm’s default risk were included and explained in this study.

• The market-to-book value (MTB) is the comparison ratio between the market value of equity and the book value of equity. Based on prior studies, it was used as a proxy for a firm’s growth opportunities. High-growth opportunity firms have promising prospects and are more attractive to investors [2]. Therefore, we anticipate an inverse association between MTB and a firm’s default risk. In this study, by multiplying the proxy of default risk by −1, we expect the MTB to have a negative association to the minus Z-score.
• Volatility (VOL) reflects the amount of uncertainty related to the extent of changes in a firm’s value. High volatility means that a firm’s value can potentially change dramatically in either direction over a period. It is measured by computing the standard deviation of monthly stock returns over the year. Thus, we assume a positive relationship between volatility and default risk [68]. In this study, by multiplying the default risk by −1, we expect a positive association between VOL and the minus Z-score.
• The stock return (RET) shows the ability of a firm to maximize the wealth of its shareholders. It describes the average of the firm’s monthly return over the year. Providing a higher stock return means that a firm is less likely to have a distress risk. Therefore, we expect RET to have a negative relationship with default risk, which means that by multiplying the default risk by −1, we expect a negative relationship between the RET and the minus Z-score.
• Financial slack (SLACK) is the comparison between cash and cash equivalents and the firm’s total assets. High financial slack refers to less dependency on external financing, which leads to less debt. Accordingly, we expect that SLACK has an inverse relationship with default risk. By multiplying the default risk by −1, we expect a negative link between SLACK and the minus Z-score.
• The size of the firm (SIZE) is derived from the natural logarithm of the firm’s total assets. It reflects how big or small is the scale or volume of a firm’s operation. Larger firms are likely to have higher debt ratios, which lead to a higher opportunity for distress [2,67]. Thus, we anticipate that SIZE has a positive relationship with default risk, which means that by multiplying the default risk by −1, we expect a positive link between SIZE and the minus Z-score.
• Asset tangibility (TANG) refers to physical measurable assets that are utilized in a firm’s operations. It is measured by calculating the percentage of total fixed assets of a firm that contribute to its total assets. Theoretically, having more tangible assets will enhance the ability of the firm to collateralize its debt and could produce higher debt financing [69]. In this study, a positive relationship between asset tangibility and a firm’s default risk is expected to exist. In other words, by multiplying the default risk by −1, we expect a positive association between TANG and the minus Z-score.

Table 3. Descriptive statistics.

	N	Mean	Std.	Min	Max
Z-score	6977	−3.49	7.37	−459.16	46.45
K-score	6977	−3.27	16.77	−860.14	458.21
DTD	6977	−4.01	2.32	−22.95	1.21
ESG	6977	2.62	0.95	0.25	6.00
Environmental	6977	0.31	0.22	0.00	0.93
Social	6977	0.28	0.17	0.00	0.93
Governance	6977	0.29	0.10	0.00	0.77
MTB	6977	1.57	6.62	−55.51	513.67
VOL	6977	0.12	0.08	0.01	1.39
RET	6977	0.01	0.04	−0.13	0.48
SLACK	6977	0.08	0.08	0.00	0.85
SIZE	6977	20.19	1.56	15.78	26.66
TANG	6977	0.33	0.19	0.00	0.96

presents the descriptive statistics of the dependent and independent variables, as well as the control variables. The sample consists of 6977 South Korean firm-year observations of the sample period.

Table 4. Correlations between regression variables.

	VIF	ESG	Environmental	Social	Governance	MTB	VOL	RET	SLACK	SIZE	TANG
ESG	3.21	1
Environmental	2.08	0.66 ***	1
Social	3.14	0.77 ***	0.64 ***	1
Governance	1.71	0.62 ***	0.38 ***	0.56 ***	1
MTB	1.03	−0.04 ***	−0.06 ***	−0.03 **	−0.03 ***	1
VOL	1.60	−0.16 ***	−0.13 ***	−0.14 ***	−0.16 ***	0.09 ***	1
RET	1.56	−0.06 ***	−0.06 ***	−0.06 ***	−0.06 ***	0.09 ***	0.59 ***	1
SLACK	1.17	−0.06 ***	−0.16 ***	−0.10 ***	−0.06 ***	0.11 ***	0.08 ***	0.07 ***	1
SIZE	1.82	0.58 ***	0.47 ***	0.64 ***	0.43 ***	−0.08 ***	−0.19 ***	−0.11 ***	−0.14 ***	1
TANG	1.23	0.09 ***	0.26 ***	0.08 ***	−0.00	−0.06 ***	−0.07 ***	−0.05 ***	−0.35 ***	0.12 ***	1

Note: The table shows correlation coefficients between ESG, each ESG component, and control variables. ***, and ** indicate significance at 1% and 5%, respectively. The Variance Inflation Factor (VIF) is a measurement to check multicollinearity problems in regression analysis.

reports the examination of the multicollinearity problem between explanatory variables using a Pearson correlation analysis by following the prior study’s test [2]. Each ESG component shows a positive and high correlation with the aggregate ESG score, reflecting that these components contribute to firms’ ESG. Based on the table, no correlation value exceeds 59% (RET and VOL) between control variables. Furthermore, a variance inflation factor (VIF) computation is also conducted to test the absence of a multicollinearity problem. The larger the VIF value, the more “troublesome” or collinear the variable independent. As a rule of thumb, if the VIF value exceeds 10, that variable is indicated as highly collinear [70,71]. In other words, a VIF above 10 indicates that a multicollinearity problem might exist, and even some studies suggest a VIF value above 5 indicates that a multicollinearity problem exists [72]. As reported in the table, the highest VIF value was 3.21, which belongs to the ESG score, indicating that multicollinearity problems did not exist in this study.

3.2. Methodology

The objective of this study was to investigate the impact of implementing CSR activities on a firm’s default risk in the Korean stock market. The first regression was aimed at exploring the association between the overall ESG score and the first proxy of default risk, the Z-score. Our regression controlled for several firm attributes, namely market-to-book (MTB), firm volatility (VOL), stock returns (RET), financial slack (SLACK), firm size (SIZE), and asset tangibility (TANG), that are presumed to affect default risk. The model used for the analysis was a fixed effect or least-squares dummy variable (LSDV) regression model. This regression method was conducted to control the influence of unidentified industry and year characteristics and to address the omitted variable concern [2,31,47,73,74]. Industrial sector dummies classified by FICS Korea (FnGuide Industry Classification Standard Korea) and year-fixed effects were implemented in this model to control for changing economic conditions over time and effects within industries. In accordance with prior research, it has been established that there exists a temporal lag in the relationship between CSR initiatives and default risk. This lag is due to the lack of direct influence of an organization’s CSR performance on the default risk. As a result, the level of risk at time t + 1 was estimated based on the CSR performance at the end of time t. The expression of the regression model is as follows:

$$\begin{gathered} D{R}_{i,t+1}=\alpha +{\beta }_{1}ES{G}_{i,t}+{\beta }_{2}MT{B}_{i,t}+{\beta }_{3}VO{L}_{i,t}+{\beta }_{4}RE{T}_{i,t}+{\beta }_{5}SLAC{K}_{i,t}+{\beta }_{6}SIZ{E}_{i,t}+ \\ {\beta }_{7}TAN{G}_{i,t}+{\beta }_{y}Year\_D_{i,t}+{\beta }_{ind}Industry\_D_{i,t}+{\epsilon }_{i,t+1}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \end{gathered}$$

(5)

where i is for a firm, t is for a year, DR is for default risks (we employed three different proxies and we used t + 1 for each default risk to emphasize the indirect effect of CSR at time t), and ESG refers to the CSR performance of the firm. The set of control variables includes MTB (market-to-book), VOL (volatility), RET (stock return), SLACK (financial slack), SIZE, and TANG (asset tangibility). Year_D is a set of year dummies and Industry_D is for the industrial sector dummy. In this model, the ESG score, as well as the control variables, at the end of time t will be regressed with the t + 1 of response variable default risk.

To account for the negative relationship between CSR and default risk, the default risk proxies were multiplied by −1 (minus one) to serve as an indicator of financial distress risk. It should be noted that a greater negative value of the Z-score, K-score, and DTD variable indicates a decreased level of default risk. Additionally, we also standardized our data by using Z-score standardization because it has advantages, including the following: it allows one to check which variable has the greatest impact on the dependent variable, and it minimizes the impact of outliers and ensures that the results of statistical analyses are not biased towards extreme values. Furthermore, an examination of the individual influence of the CSR components on default risk is also reported in this study. The purpose is to give a more in-depth overview of which ESG components affect the decrease in default risk when a firm performs CSR activities. Thus, the three components, E, S, and G, were used as explanatory variables to estimate the financial distress risk as the dependent variable. The findings of this analysis disclose the contribution of each ESG component to reducing the financial distress risk. In other words, we expect this study to reveal which component is most advantageous for a firm to decrease the level of default risk.

We repeated the regression analysis with control of past default risk measures to address a causality problem. If a concurrent or past default risk affects the ESG score, the regression coefficient of ESG in Equation (5) could merely be an effect of the autocorrelation of the default risk. To discuss this more specifically, assume that ESG is determined by:

$$ES{G}_t=a+bD{R}_t+cD{R}_{t-1}+u_t$$

(6)

Then, Equation (5) becomes:

$$\begin{gathered} D{R}_{t+1}=\alpha +\beta ES{G}_t+{\epsilon }_{t+1} \\ =\alpha +\beta \left(a+bD{R}_t+cD{R}_{t-1}+u_t\right)+{\epsilon }_{t+1} \\ =\delta +\eta D{R}_t+\psi D{R}_{t-1}+{ϵ}_{t+1} \end{gathered}$$

(7)

The estimation of the coefficients for ESG in

Table 5. The relationship between the CSR and the minus Z-score.

	Dependent Variable: Minus Z-Score
	(1) ESG	(2) Environmental	(3) Social	(4) Governance
Constant	−0.000	−0.000	−0.000	−0.000
	(0.011)	(0.011)	(0.011)	(0.011)
ESG	−0.052 ***
	(0.015)
Environmental		−0.021
		(0.015)
Social			−0.050 ***
			(0.016)
Governance				−0.066 ***
				(0.016)
MTB	0.008	0.008	0.009	0.009
	(0.012)	(0.012)	(0.012)	(0.012)
VOL	0.091 ***	0.095 ***	0.094 ***	0.087 ***
	(0.015)	(0.015)	(0.015)	(0.015)
RET	−0.096 ***	−0.097 ***	−0.097 ***	−0.094 ***
	(0.015)	(0.015)	(0.015)	(0.015)
SLACK	−0.073 ***	−0.074 ***	−0.074 ***	−0.072 ***
	(0.013)	(0.013)	(0.013)	(0.013)
SIZE	0.110 ***	0.090 ***	0.114 ***	0.108 ***
	(0.015)	(0.014)	(0.016)	(0.014)
TANG	0.120 ***	0.121 ***	0.118 ***	0.115 ***
	(0.013)	(0.013)	(0.013)	(0.013)
Year dummies	Yes	Yes	Yes	Yes
Industry dummies	Yes	Yes	Yes	Yes
Observations	6977	6977	6977	6977
R2	0.089	0.088	0.089	0.090
Adjusted R2	0.086	0.085	0.086	0.087
Res. Std. Error	0.956	0.957	0.956	0.956
F Statistic	28.403 ***	27.966 ***	28.318 ***	28.693 ***

The table shows the regression results between the default risk proxied by the minus Z-score and the overall ESG score, as well as each ESG component. The values in the parentheses are standard errors. *** denotes significance at the 1% level.

,

Table 6. The relationship between the CSR and minus K-score.

	Dependent Variable: Minus K-Score
	(1) ESG	(2) Environmental	(3) Social	(4) Governance
Constant	−0.000	−0.000	−0.000	−0.000
	(0.011)	(0.011)	(0.011)	(0.011)
ESG	−0.054 ***
	(0.015)
Environmental		−0.007
		(0.014)
Social			−0.039 **
			(0.015)
Governance				−0.086 ***
				(0.015)
MTB	0.278 ***	0.278 ***	0.279 ***	0.279 ***
	(0.011)	(0.011)	(0.011)	(0.011)
VOLAT	0.236 ***	0.240 ***	0.240 ***	0.231 ***
	(0.015)	(0.015)	(0.015)	(0.015)
RETURN	−0.199 ***	−0.201 ***	−0.200 ***	−0.197 ***
	(0.014)	(0.014)	(0.014)	(0.014)
SLACK	−0.078 ***	−0.078 ***	−0.078 ***	−0.076 ***
	(0.012)	(0.012)	(0.012)	(0.012)
SIZE	−0.015	−0.042 ***	−0.019	−0.010
	(0.014)	(0.014)	(0.016)	(0.013)
TANG	0.103 ***	0.102 ***	0.101 ***	0.097 ***
	(0.012)	(0.012)	(0.012)	(0.012)
Year dummies	Yes	Yes	Yes	Yes
Industry dummies	Yes	Yes	Yes	Yes
Observations	6977	6977	6977	6977
R2	0.165	0.163	0.164	0.167
Adjusted R2	0.162	0.160	0.161	0.164
Res. Std. Error	0.916	0.916	0.916	0.914
F Statistic	57.145 ***	56.466 ***	56.773 ***	58.071 ***

The table shows the regression results between the minus K-score and the overall ESG as well as each ESG performance. The values in the parentheses are standard errors. ** and *** denote significance at 5% and 1% levels, respectively.

and

Table 7. The relationship between the CSR and the minus Distance to Default (DTD).

	Dependent Variable: Minus DTD
	(1) ESG	(2) Environmental	(3) Social	(4) Governance
Constant	−0.000	−0.000	−0.000	−0.000
	(0.010)	(0.010)	(0.010)	(0.010)
ESG	−0.140 ***
	(0.013)
Environmental		−0.060 ***
		(0.013)
Social			−0.114 ***
			(0.014)
Governance				−0.179 ***
				(0.014)
MTB	−0.025 **	−0.025 **	−0.023 **	−0.024 **
	(0.010)	(0.010)	(0.010)	(0.010)
VOLAT	0.516 ***	0.527 ***	0.526 ***	0.507 ***
	(0.013)	(0.013)	(0.013)	(0.013)
RETURN	−0.233 ***	−0.238 ***	−0.237 ***	−0.230 ***
	(0.013)	(0.013)	(0.013)	(0.013)
SLACK	−0.116 ***	−0.118 ***	−0.117 ***	−0.113 ***
	(0.011)	(0.011)	(0.011)	(0.011)
SIZE	0.158 ***	0.107 ***	0.153 ***	0.153 ***
	(0.013)	(0.012)	(0.014)	(0.012)
TANG	0.141 ***	0.146 ***	0.137 ***	0.127 ***
	(0.011)	(0.011)	(0.011)	(0.011)
Year dummies	Yes	Yes	Yes	Yes
Industry dummies	Yes	Yes	Yes	Yes
Observations	6977	6977	6977	6977
R2	0.308	0.299	0.304	0.314
Adjusted R2	0.306	0.297	0.301	0.312
Res. Std. Error	0.833	0.839	0.836	0.830
F Statistic	129.083 ***	123.692 ***	126.290 ***	132.685 ***

The table shows the regression results between minus DTD and overall ESG as well as each ESG performance. The values in the parentheses are standard errors. ** and *** denote significance at 5% and 1% levels, respectively.

could have been subject to bias or merely a reflection of the autocorrelation of the default risk. To address this issue, we controlled for the lagged values of the default risk to assess the persistence of the ESG coefficients. The regression results are presented in

Table 8. Regression results after controlling for the lagged values of default risks.

	Z-Score		K-Score		DTD
Constant	−6.395	−5.490	−2.813	−3.427	−2.594	−3.247
	(2.322)	(1.869)	(3.739)	(4.492)	(0.307)	(0.323)
Lag1	0.594 ***	0.425 **	0.656 ***	0.473 **	0.789 ***	0.739 ***
	(0.212)	(0.195)	(0.186)	(0.187)	(0.012)	(0.022)
Lag2		0.226		0.237		0.052 ***
		(0.143)		(0.144)		(0.018)
ESG	−0.177 **	−0.189 **	−0.394 **	−0.433 ***	−0.084 ***	−0.073 ***
	(0.078)	(0.084)	(0.167)	(0.167)	(0.019)	(0.020)
MTB	0.075 ***	0.080 ***	0.731 ***	0.744 ***	−0.001	−0.0002
	(0.015)	(0.012)	(0.063)	(0.062)	(0.003)	(0.003)
VOL	3.804 ***	3.759 ***	20.352 **	17.052 **	3.489 ***	3.458 ***
	(1.406)	(1.404)	(8.140)	(7.207)	(0.285)	(0.319)
RET	−3.955	−7.968 *	−31.533 **	−35.638 ***	−1.344 ***	−1.328 **
	(4.140)	(4.669)	(13.382)	(13.533)	(0.455)	(0.517)
SLACK	−3.018	−2.127	−8.147	−6.955	−0.867 ***	−0.726 ***
	(3.177)	(3.026)	(5.933)	(5.618)	(0.238)	(0.249)
SIZE	0.232 ***	0.222 ***	0.018	0.097	0.075 ***	0.070 ***
	(0.078)	(0.064)	(0.151)	(0.171)	(0.011)	(0.012)
TANG	1.646 **	1.305 ***	2.663 **	2.196 **	0.386 ***	0.366 ***
	(0.665)	(0.485)	(1.234)	(0.914)	(0.087)	(0.093)
Year dummies	Yes	Yes	Yes	Yes	Yes	Yes
Ind. dummies	Yes	Yes	Yes	Yes	Yes	Yes
Observations	6977	6019	6977	6019	6977	6019
R2	0.423	0.419	0.516	0.535	0.748	0.748
Adjusted R2	0.421	0.417	0.514	0.533	0.747	0.747
F Statistic	63.51 ***	60.81 ***	91.47 ***	99.16 ***	367.3 ***	327.4 ***

The table shows the regression results between the "minus default risk proxies” and the overall ESG score. We control for one and two lag values of default risks. The values in the parentheses are standard errors. *, ** and *** denote significance at 10%, 5% and 1% levels, respectively.

.

4. Results

4.1. ESG and The Z-Score

The objective of this study was to examine the relationship between the CSR performance, measured by the ESG score, and the default risk of the firm. Using the LSDV regression model, this study controlled for firm characteristics (market-to-book (MTB), volatility (VOL), returns (RET), slack (SLACK), size (SIZE), and tangibility (TANG)) as well as for dummies for year and industry. The year and industry dummies were incorporated to account for both cross-sectional and time-series dependence in the analysis.

Table 5 shows the regression result of the default risk (proxied by Z-score) toward ESG performance as well as control variables. Column 1 discloses the results of the LSDV regression analysis model (1). The result reveals the negative relationship between ESG and the default risk of the firm (proxied by the minus Z-Score) at the 1% significance level. This finding concludes that firms engaging in CSR activities are able to mitigate the likelihood of default risk. Regarding the magnitude of the default risk’s reduction, this study found that, when everything else was equal, the increase in one standard deviation through the ESG score generated a 0.049 reduction in the distress risk (derived from −0.052 (ESG coefficient) × 0.95 (ESG standard deviation) = −0.049), representing a 1.4% (−0.049/−3.49 = 0.014) reduction over the average default risk. The result supports the first hypothesis stating that CSR performance, measured by the ESG score, is negatively linked with the default risk of the firm. Additionally, the outcome of this study is in accordance with prior research that has indicated that firms engaged in high environmental, social, and governance (ESG) activities exhibit reduced default risk and are perceived as more creditworthy, resulting in improved access to financing [2,21,22,23,24,25].

Analyzing the three ESG components, Columns 2, 3, and 4 exhibit the regression findings of each ESG component toward the Z-score to obtain a more comprehensive description of which component has the truly negative association with default risk. The two components, social and governance, showed negative and significant coefficients for default risk whilst the environmental component had no significant effect on a firm’s default risk. These findings are in line with several preceding studies that have stated that the employee and human rights attributes of a firm’s CSR negatively affect the unsystematic risk of the firm [75], and that social constraint has the impact of lowering the capital constraints [22]. These results suggest that the CSR social and governance components could reduce the default risk of the firms in the Korean market.

The findings related to control variables are reported in Columns 1 to 4. Nearly all control variables showed statistically significant results for distress risk. As expected, the results confirm our theoretical background, with volatility, firm size, and asset tangibility having a positive effect, and stock return and financial slack having a negative effect, on the firm’s default risk. All coefficients were statistically significant at the 1% level [2]. However, the market-to-book ratio was the only control variable that showed no significant result. In general, these results align with prior studies, demonstrating that a firm’s CSR activity has a reducing impact on the firm’s default risk [2,4].

4.2. ESG and The Alternative Proxies of Default Risk

We conducted two other tests to address the association between CSR and default risk, utilizing the K-score and DTD (distance to default) as the default risk’s proxies for accounting-based measurement and market-based measurement. The findings based on each measurement are reported in a different table. Table 6 presents the result of the K-score measurement, whereas Table 7 presents the findings of applying DTD measurement.

In Table 6, Column 1 reports the result of the regression between the overall ESG score and K-score. The result shows that the ESG performance had a minus and statistically significant impact on the minus K-score, as a proxy for the default risk, showing that high ESG firms express low default risk. This result confirms our prior finding using the minus Z-score as the proxy for the default risk, showing that firms with high CSR activities are capable of reducing their default or potential bankruptcy risk. Economically, using the K-score, this study found that, when everything else was equal, the increase in one standard deviation through the ESG score generated a −0.051 reduction in the distress risk (derived from −0.054 (coefficient of ESG) × 0.95 (standard deviation of ESG) = −0.051), representing a 1.6% (−0.051/−3.27 = 0.016) reduction in the average default risk.

The analysis of the three ESG components in Columns 2, 3, and 4 reveals the results of the effect of each component on the default risk. Out of the three components, only the environmental component did not have a significant impact, while the social component had a significant negative effect on the default risk. Furthermore, the governance component had a more significant effect on the default risk compared to the social component, and the former was statistically significant at the 1% level. This result is consistent with our previous findings, which confirm that the social and governance components have a stronger impact on reducing the default risk of Korean firms. In line with a prior study in Korea, socially responsible practices and good governance are crucial in enhancing a firm’s value [15].

The results in Columns 1 to 4 indicate that almost all control variables have a significant impact on default risk. Volatility, return, financial slack, and asset tangibility are all positively associated with default risk, while market-to-book and size have a significant effect but with opposing results.

To support our findings further, we present another test in Table 7 using market-based measurement. We employed distance to default (DTD) as a measure of default risk and examined its relationship with Corporate Social Responsibility (CSR) activities. Column 1 reports the result of the regression between the overall ESG score and the minus DTD variable. The result shows that ESG performance has a negative and statistically significant impact on a firm’s default risk, showing that firms with high ESG activities exhibit low default risk. The result in Table 7, which uses the minus Z-score and minus K-score as proxies for default risk, supports our prior findings and shows that firms with high levels of CSR activities reduce their default risk. It also confirms the findings of a previous study that used the probability of the default as a proxy for the default risk [25].

We studied the impact of the three ESG components on reducing default risk (DTD). Columns 2, 3, and 4 show that all three components (environmental, social, and governance) have a negative and significant effect on default risk. Compared to prior measurements using the minus Z- and K-scores, we found evidence that environmental, social, and governance activities significantly contribute to reducing default risk, with statistical significance at the 1% level. Therefore, the three ESG activities performed by Korean firms were proven to reduce the default risk of the firms. We argue that, since all three components induce significant impacts on default risk, an interesting result is produced, which is that, when everything else is equal, the increase in one standard deviation through the ESG score generates a −0.133 reduction in the distress risk (derived from −0.14 (coefficient of ESG) × 0.95 (standard deviation of ESG) = −0.133), representing a 3.3% (−0.133/−4.01 = 0.033) reduction in the sample’s average default risk. Compared to the two previous proxies, when using DTD as a default risk proxy, ESG was proven to have a greater capability to lessen the default risk of a firm.

The findings in Columns 1 to 4 show that all control variables have a significant impact on default risk. Market-to-book, return, and financial slack each have a negative and significant relationship with default risk, while volatility, size, and asset tangibility have a positive impact on it. These control variables were significant at the 1% level and support our understanding of their relationship with default risk. Overall, our results align with the idea that CSR reduces firm risk and are consistent with the findings of some prior studies. Companies with CSR-focused strategies are believed to have improved access to financing and are viewed as more creditworthy, lowering the risk of financial difficulties and default risk [2,4,21,22,23,24,25].

To avoid any causality issues, we conducted a regression analysis while taking into consideration past and concurrent default risk measures. This is because, if there is a relationship between past or concurrent default risk values and ESG scores, then the regression coefficient of ESG in Equation (5) may simply be a result of the correlation with default risk. In Table 8, we exhibit the regression results after controlling for the lagged values between ESG and the proxies for default risk.

The results reveal several interesting insights. Firstly, we observed a significant serial correlation in default risk, with the Z-score and K-score being affected by their one-year past value, while the Distance to Default (DTD) was impacted by up to two lagged values. This finding implies that firms with high default risks are more susceptible to future defaults. Secondly, in spite of the presence of autocorrelation, the negative impact of ESG on future default risk remains statistically and economically significant. Interestingly, the economic magnitude of the ESG coefficient varied based on the measure of default risk employed, with an increase observed when using the Z- and K-scores, and with a reduction observed when employing DTD. This discrepancy may have arisen from the fact that the former are measured through accounting values, while the latter is based on market values, leading to a potential misalignment between the actual timing of the variables and our assumptions. The results demonstrate that the DTD regression was relatively less biased towards the absence of lagged values.

5. Conclusions

Previous studies have assumed that, through CSR engagement, firms can mitigate their financial risk [2,3,4,8]. Other studies have also found that a firm that intensively implements CSR activities can gain positive moral capital for various groups, which could be the safeguard for the firm’s management [32,40]. This study was conducted to complement the preceding literature by investigating whether CSR performance is significantly beneficial in mitigating the default risk of a firm, especially in the Korean market.

Addressing this problem, firstly, we tested our hypothesis by examining the relationship between ESG as a proxy for CSR performance and the Z-score as a proxy for default risk. We also conducted the alternative tests by using the K-score and DTD as accounting-based and market-based measurements of the default risk. In addition, by testing the individual effects of the ESG components, we examined which component had an impact on default risk. We assumed a non-direct effect of the firm’s CSR activities, meaning that the impact of CSR performance on the prediction of default risk was shifted by a year. The findings show that CSR performances, indicated by overall ESG scores, successfully mitigate the default risk of a firm. Considering the three components partially, we concluded that social and corporate governance activities have impacts on reducing the financial distress risk for the Z-score and K-score measurements. Meanwhile, by using DTD measurement, we confirmed the compelling finding that all components of ESG, namely environmental, social, and governance, have significant and negative influences on default risk. These findings strengthen our main finding that firms engaging in CSR practices are able to mitigate their default risk. Furthermore, the finding also reveals that firms with a higher risk of default are more likely to face future defaults. Despite the issue of correlation, the effect of ESG on future default risk is still significant both statistically and economically.

Based on the Korean market context, there are some practical implications related to our findings. First, policymakers should continue promoting socially responsible behavior in companies as this leads to lower default risks and a stronger economy. In line with this, South Korea will be applying mandatory ESG disclosure at the beginning of 2025 to the listed firms with more than KRW 2 trillion in assets and to all KOSPI firms at the beginning of 2030 [76]. For managers or the top management of companies, this finding reveals that, beyond socially responsible activities, there is a huge benefit for firms to obtain since CSR enables the reduction of their default risk. As such, managers should consider allocating resources to increase investments in such activities. Moreover, a robust engagement with CSR enhances the relationship between a firm and its stakeholders, providing the firm with access to alternative forms of capital. Finally, for investors, our findings provide valuable insights that can assist in evaluating a firm’s financial default risk based on its level of CSR engagement. They can use this information as part of their investment decision-making processes. Additionally, investors can exploit the knowledge that CSR activities can lower default risk to assess a firm’s long-term viability and stability, as well as to determine if the company’s social responsibility aligns with their own values and investment goals.

This study provides useful insights, but it has some limitations. One limitation is the choice of dependent variables, which could be improved by a comparison to other proxies for a firm’s default risk. Additionally, exploring different evaluations of CSR would deepen the understanding of the relationship between CSR and firm risk. Further research that considers different market types (developed vs. emerging), legal systems, and industries would provide a more comprehensive understanding of the relationship between a firm’s CSR and its default risk. Furthermore, it is important to consider the possibility that the relationship between default risk and CSR could be impacted by reverse causality. Examining this possibility will enhance the accuracy of future research results and strengthen the conceptual framework and approach. This reflection will help to ensure the validity and reliability of the findings and the robustness of overall research designs in future.