Bus Strat Env - 2021 - Atif - Environmental Social and Governance Disclosure and Default Risk

Received: 5 May 2020 Revised: 19 May 2021 Accepted: 2 June 2021
DOI: 10.1002/bse.2850
RESEARCH ARTICLE
Environmental, social and governance disclosure and

default risk
Muhammad Atif1 | Searat Ali2
1
Department of Accounting and Corporate
Governance, Macquarie University, Sydney, Abstract
New South Wales, Australia We investigate whether environmental, social and governance (ESG) disclosure is
2
School of Business, University of
related to default risk. Using a sample of US nonfinancial institutions from 2006 to
Wollongong, Wollongong, New South Wales,
Australia 2017, we find that ESG disclosure is positively related to Merton's distance to default
and is negatively related to the credit default swap spread, which suggests that firms
Correspondence
Searat Ali, School of Business, Faculty of with a higher ESG disclosure have lower default risk. Our analysis further indicates
Business and Law, University of Wollongong,
that the inverse effect of ESG disclosure on default risk is through increased profit-
Northfields Ave, Wollongong, NSW 2522,
Australia. ability and reduced performance variability and cost of debt. We also document that
Email: [email protected]
the negative impact of ESG disclosure on default risk is existent only for mature and
older firms. These results are important for all stakeholders of firms, including share-
holders and bondholders to consider firm's ESG disclosure in conjunction with life
cycle stage before making their investment decisions.
KEYWORDS
corporate life cycle, cost of debt, default risk, ESG disclosure, performance variability
1 | I N T RO DU CT I O N Corporate default, generally defined as a firm that is failing to make

contractual payments to debt holders (Valta, 2016), can lead to devas-
This study has examined whether disclosures of environmental, social tating results for stakeholders (e.g., investors, creditors, customers,
and governance (ESG) affect default risk in firms in the US context. employees and regulators).2 The likelihood of default depends on
Nonfinancial performance evaluated by ESG disclosure continues to whether the firm's future cash flows are sufficient to cover both its
attract the attention of practitioners, investors, customers, employees debt service costs (interest payments) and the principal amount. In
and suppliers, as well as scholars. All major stakeholders source infor- general, the likelihood of default increases when a firm experiences a
mation from ESG disclosure: about product safety (Deegan & downward shift of its average future cash flows or when the future
Rankin, 1996), employee rights and welfare at workplace cash flows become more volatile. Given the significance of default risk
(Waddock, 2008) and firms' relationship with society.1 Growing for investors and other stakeholders, it is imperative to examine
empirical evidence regarding the influence of ESG disclosure on firm whether ESG disclosure can help firms to mitigate default risk.
performance (e.g., Endrikat et al., 2014; Horvathov
a, 2010) suggests a We use the resource-based theory (RBT) to conjecture the associ-
positive impact on cash flows and a reduction in cost of capital ation between ESG disclosure and default risk. RBT (Barney, 1991)
(Plumlee et al., 2015). Extant evidence also suggests that ESG disclo- asserts that firms are capable of managing internal dynamic factors
sure has an impact on firms' financial risk management (e.g., Plumlee that affect their success. Firms with higher dynamic capabilities may
et al., 2015; Limkriangkrai et al., 2017). However, the role of ESG dis- combine ESG strength with other corporate strategies to successfully
closure in influencing default risk has received limited attention in the capitalize ESG benefits in their financial risk management (Teece
literature (e.g., Boubaker et al., 2020). Default risk is one of et al., 1997). For instance, ESG disclosure is associated with brand
the important yardsticks of firm financial health (Rego et al., 2009). value and with customer satisfaction that leads to higher sales and
Abbreviations: ESG, environmental, social and governance; GFC, global financial crisis; OLS, ordinary least square; PSM, propensity score matching; RBT, resource based theory.
Bus Strat Env. 2021;30:3937–3959. wileyonlinelibrary.com/journal/bse © 2021 ERP Environment and John Wiley & Sons Ltd. 3937
10990836, 2021, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/bse.2850 by Indian Institute Of Management Ranch, Wiley Online Library on [15/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3938 ATIF AND ALI
thus to profitability (Brown & Dacin, 1997; Luo & cycle following the Dickinson (2011) model. We find that the negative
Bhattacharya, 2006). Such customer loyalty helps to ensure a stream impact of ESG disclosure on default risk exists mainly in mature and
of consistent cash flows, reducing a firm's financial distress. Moreover, older firms.
in case of negative events, firms continue to enjoy stable future cash Our study makes two important contributions to the literature.
flows due to moral capital among customers. Further, disclosure of First, we contribute to the stream of research that explores the deter-
ESG information contributes to a reduction in information asymmetry minants of firms' default risk. This recent strand of literature has identi-
(Cho et al., 2013), leading towards investors trust and loyalty. The fied important determinants of default risk: for example, stock market
symmetric flow of information increases the availability of debt capital liquidity (Brogaard et al., 2017; Nadarajah et al., 2020), incentive struc-
and reduces the cost of funds (Cheng et al., 2014; El Ghoul ture (Bennett et al., 2015), innovation performance (Hsu et al., 2015),
et al., 2011; Waddock & Graves, 1997). We therefore argue that ESG debt maturity choice (Goyal & Wang, 2013), stock market return and
disclosure reduces firms' default risk by increasing their level of profit- volatility (Giesecke et al., 2011) and corporate governance characteris-
ability and by decreasing their performance variability and cost of tics including CEO compensation, board independence, gender
debt capital.3 diversity and ownership structure (see, e.g., Acrey et al., 2011; Ali,
We extend this baseline idea to examine if the ESG–default link- 2017; Ali et al., 2018, 2021; Chiang et al., 2015; Chiu & Wagner, 2012;
age is moderated by firm life cycle stages. Existing literature suggests Kabir et al., 2020; Miglani et al., 2015; Nadaraja et al., 2020; Schultz
distinct capabilities and resources, cash flow volatilities and policy et al., 2017; Switzer & Wang, 2013; Vallascas & Hagendorff, 2013). We
decisions in firms at different life cycle stages (e.g., Faff et al., 2016; extend this literature by providing robust evidence regarding the role of
Miller & Friesen, 1984; Porter, 2008). For instance, firms in the early aggregate ESG disclosure as a critical determinant of default risk for US
stages of their life cycle are financially constrained, lack liquid nonfinancial firms. Prior studies investigate the effect of ESG on vari-
resources, may experience uncertainties about future cash flows and ous firm-level outcomes including firm performance, cost of capital and
may have difficulty in raising additional capital, all of which raises the financial distress (e.g., Boubaker et al., 2020; Endrikat et al., 2014;
likelihood of default (Hasan et al., 2015). Younger firms, due to their Horvathov
a, 2010; Plumlee et al., 2015). However, our study goes
implicit concerns about achieving financial objectives, are less likely to beyond cause-and-effect relationship between ESG and default risk
focus on ESG disclosure than on showing legitimacy with key stake- and differs from existing studies in at least three ways. First, our data
holders (Ramaswamy et al., 2007). Firms in their mature phase have are sourced from Bloomberg, which uses the Global Responsible Initia-
greater competitive advantage, sufficient resources and more cer- tive (GRI) and the United Nations (UN) guidelines in collating ESG dis-
tainty of future cash flows (Hasan & Habib, 2017), leading to lower closure information. Second, our default risk measures are based on
chances of default. In addition, mature firms are more likely to market data, rather than on accounting information that lacks currency.
embrace ESG disclosure due to their reputational concerns and Third, we examine various mechanisms through which ESG disclosure
greater interaction with key stakeholders, including regulatory author- affects default risk, and document that higher profitability, lower per-
ities (Hasan & Habib, 2017). Given this evidence, we expect to find formance variability and lower cost of debt are the channels through
that the inverse relation between ESG disclosure and default risk which ESG disclosure reduces default risk.
should be relatively stronger (weaker or insignificant) for firms at Second, we contribute to the firm life cycle literature. Previous lit-
mature stage (younger and other stages) of the life cycle. erature on the firm life cycle has described its role in influencing divi-
Using a panel of nonfinancial firms from 2006 to 2017, we find dend policy (DeAngelo et al., 2006; Grullon et al., 2002), equity
that ESG disclosure is negatively related to default risk, as proxied by offerings (DeAngelo et al., 2010), merger and acquisitions (Owen &
Merton's distance to default and credit default swap spread in base- Yawson, 2010), cash flow patterns (Dickinson, 2011), firm financial
line estimations. This finding is robust to alternative variable specifica- policies (Faff et al., 2016) and financial resources and corporate social
tions and to alternative estimation methods, including lagged responsibility (Hasan & Habib, 2017). We show that the corporate life
independent variables, instrumental variable approach and propensity cycle moderates the effect of ESG disclosure on default risk. Specifi-
score matching (PSM). We further explore mechanisms through which cally, we find that firms in their mature stage have a significantly
ESG disclosure reduces default risk. Using the profitability mechanism, higher level of ESG disclosure and thus have lower default risk.
we find that the inverse effect of ESG disclosure on default risk is Overall, our study complements the current discussion regarding
through increased profitability (measured through return on assets the importance of ESG disclosure for firms and for investors and
and return on equity) and reduced performance variability (measured stakeholders. This study offers new insights for policy makers into the
through quantile-based risk metrics) (see, e.g., Konstantinidi & business case for ESG disclosure (i.e., doing well by doing good). More
Pope, 2016). In addition, we examine whether ESG disclosure reduces specifically, the role of ESG disclosure in reducing default risk
default risk via cost of debt (measured as a ratio of interest expense strengthens its importance for firms and regulators. Our findings also
divided by average interest-bearing debt). We find specific evidence provide insights to shareholders and debtholders for considering a
that firms with a lower cost of debt have a lower default risk and that firm's disclosure about sustainability initiative before making invest-
firms with a higher ESG disclosure have a lower cost of debt. Finally, ment decisions. Given our findings, policy makers aiming at sustain-
we investigate whether the relationship between ESG disclosure and ability should encourage firms to disclose their sustainability
default risk relationship is magnified in different stages of a firm's life initiatives.
ATIF AND ALI 3939
The rest of the paper is organized in seven sections. Section 2 (Brown et al., 2006). ESG disclosure also represents a firm's strategic
reviews the literature then develops the hypotheses. The sample investment towards its integrity (McWilliams et al., 2006). These soci-
selection and summary statistics are discussed in Section 3. Sections 4 etal initiatives improve the welfare of society (Davis &
and 5 present the main empirical results. Section 6 shows additional Blomstrom, 1975, p. 6). All those important activities help in creating
analysis. Section 7 concludes the study. a firm's integrity and its better linking with stakeholders and society at
large. Therefore, ESG disclosure enhances a firm's ability to create
reputation and to enjoy the benefits of such internal strength in its
2 | LITERATURE REVIEW AND financial risk management. RBT embraces this argument: that a firm's
HYPOTHESES internal strength, along with its corporate strategies, helps to capital-
ize the benefits for firms in their financial risk management.
Default is an abrasive event in the life of a corporation. It causes dis- However, investment on societal initiatives is considered to be an
ruptions in the productivity of firms through supply chain; it deters expense for shareholders. For instance, prior studies suggest that the
employee and customer retention; it increases legal and administra- main objective of a firm is to maximize its shareholders' wealth while
tive costs (Brogaard et al., 2017). The risk of default increases if firm focusing on profit making (e.g., Berle, 1930; Friedman, 1970), rather
cash flows are shaky, and/or there is no access to market. This type of than spending their wealth on stakeholder orientation. To that end,
peril becomes a yardstick to evaluate a firm's financial health (Rego some traditional studies offer support for shareholder-only orienta-
et al., 2009). The literature concurs that debtholders carry special tion, while regarding societal and environmental consideration to be
interest in firms while shareholders are residual owners (Anderson & undesirable. However, earlier literature provides strong support for
Mansi, 2009). This implies that debtholder will be mainly impacted in stakeholder orientation, contemplating that firms exist not only
case of a firm's likelihood of financial distress. Given the importance for profit making but also to offer social services (Dodd, 1932). More
of default risk in a firm's financial journey, it has received wider atten- recent studies have shown a positive impact of ESG disclosure on firm
tion in the literature.4 outcomes, including on performance (Endrikat et al., 2014), on finan-
Default risk is considered to be an important factor in firm evalua- cial returns and risk-taking (Limkriangkrai et al., 2017; Sassen
tion. Investors and debtholders take into account such risk in their et al., 2016) and on cost of debt (Plumlee et al., 2015). This study,
investment screening process (Campbell et al., 2008; Dichev, 1998). which focuses on the impact of ESG disclosure on default risk, is con-
As capital market is the largest source of external funding and to sistent with this area of the literature.
ensure availability of sufficient funds in future at relevant cost, man- ESG disclosure may help firms in reducing default risk in several
agers endeavour to keep default risk lower (Anderson & Mansi, 2009). ways. First, it is directly linked with brand value and with the customer
In turn, lower default risk may foster market confidence, resulting in a satisfaction that leads to higher sales turnover and results in higher
lower cost of capital ensuring smooth operations. profitability (Brown & Dacin, 1997; Luo & Bhattacharya, 2006). Prior
Prior research has found an array of factors associated with literature concurs that propensity to default has a direct association
default risk, including a firm's operational and production efficiency with incoming cash flows because a reliable stream of liquidity pro-
(Becchetti & Sierra, 2003; Rendleman, 1978), cash flow uncertainty vides support to operations and prevents firms from financial distress
(Chava & Purnanandam, 2010), information asymmetry (Bhojraj & (D'Aveni & Ilinitch, 1992). ESG disclosure ensures consistency of cash
Sengupta, 2003), brand loyalty (Rego et al., 2009), customer satisfac- flow and is therefore likely to affect the default risk.
tion (Anderson & Mansi, 2009) and social factors (Allen et al., 2004). Second, firms may experience default in cases of volatile cash
All those factors are associated directly or indirectly with future cash flows because infrequent income flows create cash shortfall in critical
flows and firm performance. For instance, customer satisfaction and operational requirements (Chava & Purnanandam, 2010). ESG disclo-
brand loyalty demonstrate the stability of future incoming cash flows. sure helps in stabilizing financial performance through corporate
Such a manifestation is considered among major determinants of image, reputation and better relationship with governments and finan-
default risk (Acharya et al., 2006; Asvanunt et al., 2007; cial markets (Carter, 2005; Cornell & Shapiro, 1987; McGuire
Merton, 1974). Moreover, business-related factors (e.g., business et al., 1988). For example, in the case of a negative event, stake-
cycle and economic) and social factors both play an important role in holders are more likely to punish the firm; however, a firm with ESG
default risk (Allen et al., 2004). In summary, default risk reflects market known initiatives is less likely to have a negative effect on cash flows
confidence and a firm's internal strength to minimize the risk associ- due to its image and its moral capital among stakeholders
ated with a number of factors, including those that are social, eco- (Godfrey, 2005). ESG disclosure creates loyalty and trust, as well as
nomic and business-related (e.g., life cycle). brand equity among customers; thus, income stream and profitability
ESG signals firm-specific information regarding environmental, are less volatile, even during the time of a crisis (Godfrey, 2005;
social and governance initiatives towards society (Brown & Godfrey et al., 2009). In addition, performance variability is less likely
Dacin, 1997). Those initiatives provide information about the firm rev- to happen when there are stable cash flows. Therefore, ESG disclo-
enue prospectus (Lev et al., 2010), including employee welfare and sure acts as an ‘insurance’ for firms, protecting them from default by
labour relations (Edmans, 2011; Waddock & Graves, 1997) and envi- ensuring stable cash flows. Prior literature has also supported the
ronmental and social compliance initiatives, based on firm operations notion that socially responsible investors (SRI) are less sensitive than
3940 ATIF AND ALI
conventional fund providers to negative events (Bollen, 2007; the form of numbers ranging from 0.1 to 100. Our additional analysis
Renneboog et al., 2011), suggesting a continuous supply of funds to employed the individual scores of environmental, social and gover-
firms with better ESG disclosure. nance separately. These individual scores represent firm performance
Third, ESG disclosure reduces agency cost and information asym- on each factor individually. For instances, environmental scores are
metry in firms (Cormier et al., 2011). The availability of symmetric based on firm performance relating to climate change policies, hazard-
information across the market mitigates several risks: regulatory, con- ous wastes, nuclear energy, and sustainability indicators. Social scores
troversy, managerial and reputational. Investors evaluate firms based are based on firm performance relating to consumer protections,
on the availability of nonfinancial information to determine default diversity, human rights, animal, welfare, child labour and employee
probability. The provision of such firm-specific information to the health and safety indicators. Governance scores are based on firm
market builds investors' trust and loyalty, which increases the avail- performance of management structure, executive compensation and
ability of funds at relatively lower cost and reduces the cost of capital conflict of interest indicators.
(Cheng et al., 2014; Cormier et al., 2010; El Ghoul et al., 2011;
Waddock & Graves, 1997).5
In a nutshell, ESG disclosure represents important initiatives that 3.3 | Measuring default risk
create a valuable image, that increase cash flow, that decrease perfor-
mance volatility and that garner support for the easy availability of Our first measure of default risk, Merton's (1974) distance to default
low-cost funding, leading towards a reduction in default risk. There- (DD), remains the most widely used market-based credit risk metric
fore, we hypothesize as follows: and performs better than accounting-based approaches for the esti-
mation of default risk (Bharath & Shumway, 2008; Das et al., 2009;
H1. ESG disclosure is negatively associated with a Duan et al., 2012; Duan et al., 2018; Hillegeist et al., 2004; Jessen &
firm's default risk. Lando, 2015; Miao et al., 2018).8 Merton's (1974) structural models
used option pricing methods to compute a probability of default from
H2. ESG disclosure is negatively associated with a the level and volatility of the market value of assets, with a strike price
firm's default risk through increased profitability and equal to the face value of the firm's liabilities (Bharath &
reduced performance variability and cost of debt. Shumway, 2008). The larger the positive distance between firm value
and firm liabilities, the lower would be the probability of default. In
other words, the higher the DD, the lower the default risk, meaning
3 | DATA AND METHOD more firm stability.
Our second proxy of default risk is the credit default swap (CDS)
3.1 | Sample and data spread. CDS are credit derivatives that allow the transfer of the firm's
default risk between two agents for a predetermined time period. In a
Our sample, for the period from 2006 to 2017, is sourced from typical CDS contract, the protection seller offers the protection buyer
Bloomberg and Singapore National Universal liquidity data for US insurance against the default of an underlying bond issued by a certain
firms listed in the New York and Chicago exchanges.6 First, we collect company (the reference entity). In the event of default by the reference
the ESG disclosure scores and firm characteristics from Bloomberg, entity, the seller commits to buy the bond for a price equal to its face
disregarding the difference in industries, and receive 9455 firm-year value from the protection buyer. In exchange for the insurance, the
observations.7 Since our research is not applicable to financial service buyer pays a quarterly premium, called the CDS spread, which is quoted
firms, we receive 7412 firm-year observations. The selection of firms as an annualized percentage of the notional value insured. Therefore,
is also based on the availability of aggregate and individual ESG disclo- by definition, the CDS spread is the pricing of the default risk (Das
sure scores, as well as on firm characteristics. Second, we collect the et al., 2009). The higher the default risk of the reference entity, the
default risk variables from the Singapore National University data for higher is the CDS spread. We extract the CDS spread from the National
all the US firms and merge them with the ESG data. We eliminate the University of Singapore's Credit Research Initiative (CRI) database. In
missing firm-year observations across all the variables. Finally, we that database, CDS spread, or ‘Actuarial Spread’, is constructed using
receive 5206 firm-year observations as our selected sample. the traditional CDS design but without an upfront fee, on the assump-
Appendix A presents the detail of our sample selection criteria. tion that market participants are risk neutral. Thus, actuarial spread has
the same features as the standard CDS spread. This adjustment allows
the calculation of CDS spread for a large number of firms.
3.2 | Measuring ESG
In order to measure a firm's ESG disclosure following prior literature 3.4 | Measuring control variables
(e.g., Cheng et al., 2014; Sassen et al., 2016), we use ESG disclosure
scores as an umbrella measure provided by Bloomberg to examine the Following prior studies (e.g., Ali et al., 2018; Chiang et al., 2015;
effect on the firm's default risk. ESG disclosure scores are provided in Schultz et al., 2017), we control for several variables which might
ATIF AND ALI 3941
potentially influence default risk. We first control for firm size (SIZE), disclosure is measured as the aggregate environmental, social and
measured as the natural log of total assets in millions of US$. In gen- governance disclosure scores of firm i in year t. Controls represents
eral, larger firms are more stable and consequently have less default control variables as defined in Section 3.4. Industry effects
risk; thus, we expect a negative relationship between firm size and represents the two-digit industry codes based on global industry clas-
default risk. Second, we control for leverage, the ratio of total debt to sification standards (GICS), and year effects shows the sample years
total assets (TLTA). Leverage reflects the capital structure of the firm: (2006–2017). See Appendix B for the description of variables.
that is, how a firm finances its assets, as well as the ability of the firm
to meet its financial obligations. Leverage increases the financial risk
of firms and is thus expected to have a positive relationship with 3.6 | Descriptive statistics
default risk (Chiang et al., 2015; Schultz et al., 2017). Third, we control
for firm profitability using return on assets (ROA), which is equal to Table 1 presents the summary statistics of various measures of ESG dis-
the ratio of net income to total assets. ROA, based on accounting closure, default risk and firm-specific variables. The ESG disclosure vari-
information, indicates the ability of a firm to generate sufficient ables show that, on average, the overall ESG score is 19.94, the
returns for the smooth functioning of the firm. Higher ROAs would environmental score is 22.49, the social score is 19.46 and the gover-
indicate lower default risk. Fourth, we control for liquidity (LIQ), which nance score is 52.29. The default risk variable shows that the average
indicates the ability of the firm to pay its short-term obligations. DD is 5.88 in the full sample. The mean DD of firms with high ESG dis-
Default risk should be less in firms that have a higher liquidity ratio closure (i.e., 6.60) is considerably higher than the mean DD of firms with
than in those with a lower liquidity ratio. We capture liquidity through low ESG disclosure (i.e., 5.24), suggesting that the firms with more ESG
the current ratio (current assets divided by current liabilities). disclosure have lower default risk. The average size (SIZE) of the firm-
Our fifth variable, Tobin's Q (TOBINQ), considers the market infor- specific variables is 11,899 million USD. On average, the sample firms
mation and is arguably a better measurement of both performance have 2% profit (ROA), have 2.58 USD in current assets to repay 1 USD
and growth opportunity. Tobin's Q, measured by summing fair market of current liabilities (LIQ), have 52% debt in their capital structure (TLTA)
value and total liabilities divided by total assets, is expected to have a and have 0.97 sales as a proportion of total assets. Asset tangibility
negative impact on default risk since better performance leads to (TANG) and growth opportunities (TOBINQ) average 0.27 and 2.07,
higher stability (Chiang et al., 2015). The sixth control variable, the respectively. Those firms with high ESG scores are larger, more profit-
sales to asset ratio (SALES), is measured as the ratio of sales to total able and levered and have more tangibility. However, these firms are
assets. SALES represents management efficiency and is expected to less liquid, with lower sales to total assets and lower growth.
have a negative influence on the default risk (Miglani et al., 2015).
Seventh, we control for tangibility (TANG), measured as the ratio of
net property plant and equipment divided by total assets. Tangibility 3.7 | Correlation analysis
is found to be negatively related to firm value (Fukui &
Ushijima, 2007; Nakano & Nguyen, 2013). Hence, we expect tangibil- Table 2 presents the pairwise correlation among the variables used in
ity variables to be positively related to default risk (Chiang empirical analysis. The correlation between CDS and DD is .49,
et al., 2015). which is not very high, suggesting that these are the alternative prox-
To reduce the influence of outliers in our estimates, we have win- ies of default risk. In the relationship between ESG and default risk,
sorized all other control variables at 1%. We keep the key indepen- the ESG scores have a significant negative (positive) correlation with
dent variable (ESG) at its original values as we observed no problem of the CDS (DD), as anticipated. These results suggest that for firms
outlier in the descriptive statistics. We also keep the dependent vari- with a greater ESG score, the default risk is lower. However, the corre-
ables (DD and CDS) at their original values, as extreme values on DD lation analysis does not control other factors that affect default risk,
and CDS indicate worst performance (bankruptcy). so these results should be viewed with caution. For correlation
between the control variables and default risk, we find that large firms
with more profitability (ROA), firm value (TOBINQ), sales and liquidity
3.5 | Estimation model (LIQ) have a lower default risk. However, leverage (TLTA) and tangibil-
ity (TANG) increase default risk. The correlation analysis also indicates
We estimate the following model to examine the effect of ESG disclo- that collinearity is generally moderate between the explanatory vari-
sure on firm default risk: ables. As a rule of thumb, a correlation of .70 or higher in absolute
value may indicate a multicollinearity issue. The highest correlation
Default riskit ¼ α þ βX
1 ðESG_disclosureÞit þ β 2 ðControls
X Þit coefficient (.63) is between ESG and SIZE variables. Thus,
þ δ3 ðindustry effectsÞi þ δ4 ðyear effectsÞt þ εit , multicollinearity may not be an issue in our analysis.
ð1Þ Figure 1 presents the industry-wide differences in the average
DD between firms with high and low ESG scores. It is evident from
where dependent variable default risk is measured by Moreton's dis- the figure that the firms with high ESG disclosure have a lower default
tance to default (DD) and credit default swap spread (CDS). ESG risk than the firms with low ESG disclosure for all industries. However,
3942
TABLE 1 Descriptive statistics
Whole sample Low ESG High ESG
Variables Obs. Mean SD Min p25 p50 p75 Max Mean SD Mean SD
ESG 5206 19.94 12.86 1.24 11.16 14.05 22.31 76.03
E factor 2048 22.49 18.28 1.38 6.98 16.67 35.66 82.17
S factor 3452 19.46 15.49 3.13 8.77 14.04 28.07 79.69
G factor 5240 52.29 6.38 3.57 48.21 51.79 55.36 85.71
DTD 5206 5.88 3.22 0.96 3.52 5.38 7.75 24.99 5.24 2.94 6.60 3.37
CDS 5206 7.96 25.12 0.00 0.35 1.73 6.91 388.62 8.52 19.23 7.49 30.89
CDS (Ln) 5206 0.30 2.26 11.87 1.04 0.56 1.93 5.96 0.78 2.03 0.23 2.40
SIZE 5206 7.50 2.11 1.07 6.09 7.63 9.00 13.59 6.32 1.72 8.86 1.64
SIZE ($) 5206 11,899 41,286 3 443 2066 8127 797,769 1971 4800 23,208 58,136
ROA (%) 5206 0.02 0.18 1.18 0.02 0.05 0.09 0.40 0.01 0.23 0.05 0.09
TLTA (%) 5206 0.52 0.23 0.07 0.37 0.52 0.66 1.31 0.48 0.25 0.58 0.19
LIQ 5206 2.58 2.22 0.37 1.30 1.94 2.98 14.95 3.11 2.57 1.97 1.53
TANG (%) 5206 0.27 0.24 0.00 0.09 0.19 0.39 0.90 0.23 0.23 0.32 0.25
SALES 5206 0.97 0.65 0.00 0.50 0.83 1.28 3.70 0.99 0.67 0.93 0.63
TOBINQ 5206 2.07 1.56 0.67 1.22 1.59 2.27 13.75 2.23 1.87 1.89 1.08
Note: This table presents summary statistics for the sample of 5206 firm-year observations over the period 2006–2017. Definitions of all variables are described in Appendix B.
ATIF AND ALI
ATIF AND ALI 3943
TABLE 2 Correlation analysis

DD CDS ESG SIZE ROA TLTA LIQ TANG SALES TOBINQ
DD 1.00
CDS .49 1.00
ESG .26 .30 1.00
SIZE .19 .23 .63 1.00
ROA .33 .24 .15 .32 1.00
TLTA .24 .09 .21 .39 .09 1.00
LIQ .06 .06 .24 .42 .18 .49 1.00
TANG .06 .04 .14 .22 .07 .16 .32 1.00
SALES .02 .05 .12 .13 .27 .01 .14 .11 1.00
TOBINQ .25 .10 .05 .27 .28 .05 .22 .25 .03 1.00
Note: This table shows the correlation analysis among the independent and dependent variables. See
Appendix B for variable definitions.
FIGURE 1 Industry-wise default risk for firms with high and low ESG disclosure scores [Colour figure can be viewed at wileyonlinelibrary.com]
some industries benefit more from high ESG than others do. For are reported in Columns 1–3 and 4–6, respectively. Columns 1 and
instance, default risk reduction with high ESG disclosure is the highest 4 include univariate regression between ESG and default risk; regres-
in Health and Telecommunication Services (47% each) and is the low- sions in Columns 2 and 5 include ESG and all control variables;
est in Energy and in Consumer Discretionary (17% each). Overall, regressions in Columns 3 and 6 include ESG, all the control variables
Figure 1 demonstrates that industries with high ESG disclosure have a and the industry- and year-fixed effects. All of these alternative model
lower default risk. Given this evidence, it is important to include this specifications suggest that ESG is positively (negatively) related to DD
industry effect in the empirical analysis. (CDS) at the 1% level, which might indicate a lower default risk for
firms that have a greater level of ESG disclosure. Specifically, a
1-point increase in ESG increases DD by 0.025 points (Column 3 of
4 | E S G D I S C L O S U R E A N D D E FA U LT R I S K Table 3). These results also have economic significance. For instance,
(H1) one standard deviation increase in ESG raises DD by 5.40%.9 These
findings are consistent with our prediction that ESG disclosure influ-
4.1 | Baseline results ences default risk: specifically, we find a significantly positive (nega-
tive) association with DD (CDS). The firm-specific variables are also
Table 3 presents the OLS results of Equation 1 between ESG and important in determining default risk. For instance, larger, more profit-
default risk (as measured by DD or CDS). The results for DD and CDS able and high growth firms all have a lower default risk, while the
3944 ATIF AND ALI
TABLE 3 ESG and default risk—baseline results
DD DD DD CDS CDS CDS

(1) (2) (3) (4) (5) (6)
ESG 0.064*** (7.73) 0.041*** (5.13) 0.025*** (3.47) 0.052*** (6.24) 0.038*** (4.23) 0.035*** (4.19)
SIZE 0.316*** (7.11) 0.409*** (9.22) 0.128** (1.98) 0.115* (1.79)
ROA 5.670*** (13.80) 5.262*** (12.10) 2.678*** (6.04) 2.733*** (5.88)
TLTA 4.363*** (13.54) 4.834*** (14.60) 1.825*** (5.01) 2.144*** (6.09)

LIQ 0.004 (0.10) 0.014 (0.36) 0.089** (2.38) 0.094** (2.49)
TANG 0.159 (0.54) 0.082 (0.25) 0.700** (2.00) 0.959** (2.53)
SALES 0.039 (0.38) 0.049 (0.47) 0.256** (2.36) 0.333*** (2.61)
TOBINQ 0.796*** (13.32) 0.706*** (12.15) 0.272*** (5.01) 0.262*** (4.82)
Constant 4.595*** (26.79) 3.307*** (8.30) 3.572*** (6.45) 1.348*** (8.36) 1.020** (2.02) 1.663*** (3.22)
Year effects No No Yes No No Yes
Industry No No Yes No No Yes

effects
SE robust Yes Yes Yes Yes Yes Yes
N 5234 5206 5206 5234 5206 5206
Adj. R2 .065 .355 .513 .087 .188 .249
Note: This table presents the regression results between ESG and default risk using pooled OLS. Dependent variable is distance to default, that is, DD (or credit
default swap spread, i.e., CDS), in Columns 1–3 (Columns 4–6). Figures in parenthesis are the t statistics. See Appendix B for variable definitions.
*Statistical significance at 10% level.
**Statistical significance at 5% level.
***Statistical significance at 1% level.
default risk is higher in firms that have more debt in their capital struc- control for reverse causality. Regressions based on contemporaneous
ture. Overall, our findings are consistent with prior studies, which variables are susceptible to endogeneity bias due to reverse causality,
have stated that ESG disclosure decreases firm risk (Sassen whereas regressions based on lagged values of independent variables
et al., 2016), influences cash flows and cost of capital (Plumlee help to control for reverse causality and thus tend to be less suscepti-
et al., 2015) and therefore influences default risk. ble to endogeneity effects (Harjoto & Jo, 2015; Luo &
Bhattacharya, 2009). Second, we adopt an instrumental variable
approach, two-stage least squares (2SLS), to address the possible
4.2 | Endogeneity bias endogeneity of the ESG disclosure score. Third, we execute PSM,
which can help solve the reverse causality and self-selection bias.
The relationship between firm default risk and the ESG may possibly
be affected by endogeneity. Two of these possible sources of endo-
geneity (i.e., omitted variable bias and reverse causality) are likely to 4.2.1 | Lagged variables
bias our main results regarding how ESG influences default risk. First,
omitted variables (e.g., time varying and time unvarying; observable To mitigate these reverse causality issues, we estimate alternative
and unobservable) may simultaneously affect both ESG and default specifications of Equation 1. Specifically, we test the influence of the
risk. It is difficult, if not impossible, to capture all the determinants of ESG score in the previous year on the default risk (DD) in the current
default risk in the empirical models, which leads to the omitted vari- year. We report the results in Columns 1–2 (Lag 1) and Columns 3–4
able bias. Second, the direction of causation between ESG disclosure (Lag 2) of Table 4. As can be seen from these results, ESG is positively
and default risk is not clear. For instance, Hermalin and (negatively) related to DD (CDS), suggesting that the prior-year ESG
Weisbach (2001) suggest that a firm's governance structure is endog- inversely affects the current year's default risk. These results suggest
enously determined. Bouslah et al. (2013) find a bidirectional causality that the direction of causation runs from ESG disclosure to default risk
between corporate social performance and firm risk. Given this, it is but not vice versa.
possible that ESG disclosure and default risk are determined simulta-
neously: that is, not only does ESG disclosure influence firm risk—
firms also adjust ESG disclosure based on the current risk exposure. In 4.2.2 | Instrumental variable approach
such circumstances, current ESG disclosure is likely to be influenced
by a past realization of default risk (i.e., reverse causality). Hence, as The instrumental variable approach requires an instrument that is cor-
part of our identification strategy, we address these endogeneity con- related with the endogenous variable (i.e., ESG) but that does not have
cerns in three ways. First, we employ lagged independent variables to a direct influence on the dependent variable (i.e., default risk), except
ATIF AND ALI 3945
TABLE 4 ESG and default risk—endogeneity bias (lagged independent variables)
Lag 1 Lag 2
DD CDS DD CDS
(1) (2) (3) (4)
ESG 0.022*** (2.97) 0.035*** (4.15) 0.019** (2.34) 0.035*** (4.01)
SIZE 0.419*** (9.18) 0.118* (1.80) 0.452*** (9.18) 0.135** (2.03)
ROA 5.224*** (12.76) 2.562*** (5.51) 5.183*** (11.64) 2.599*** (5.44)
TLTA 4.540*** (12.77) 2.311*** (6.46) 4.331*** (11.05) 2.474*** (6.50)
LIQ 0.022 (0.57) 0.108*** (2.80) 0.038 (0.91) 0.102*** (2.60)
TANG 0.315 (0.86) 0.973** (2.49) 0.516 (1.27) 0.986** (2.52)
SALES 0.092 (0.79) 0.326** (2.50) 0.105 (0.81) 0.304** (2.32)
TOBINQ 0.610*** (11.02) 0.262*** (4.72) 0.562*** (9.01) 0.272*** (4.78)
Constant 4.173*** (8.40) 1.704*** (3.25) 1.242** (2.41) 1.575*** (2.95)
Year effects Yes Yes Yes Yes
Industry effects Yes Yes Yes Yes
SE robust Yes Yes Yes Yes
N 4638 4635 4034 4032
Adj. R2 .471 .250 .400 .257
Note: This table presents the regression results between ESG and default risk using lagged independent variables. Columns 1–2 are based on 1-year lag,
and Columns 3–4 are based on 2-year lag. Figures in parenthesis are the t statistics. See Appendix B for variable definitions.
through the endogenous variable (Kennedy, 2003, p. 159). Specifi- than the critical value (at the 10% level) given by Stock and
cally, a valid instrument must satisfy two conditions. First, the rele- Yogo (2005), this again rejects the null hypothesis that the instrument
vance condition requires that, after controlling for the set of variables is not weak.
in our model, the instrument correlates with ESG. Second, the exclu- Columns 2 and 3 of Table 5 present the second-stage regression
sion restriction requires that, conditional to the full set of control vari- results in which we regress the ESG (fitted) on default risk while con-
ables, the instrument influences default risk only through its trolling other firm-specific variables. Consistent with our earlier find-
correlations with ESG. ings, the coefficient on ESG is statistically significant and negative
Our instrument in this study, the industry median ESG, is com- (positive) for CDS (DD), indicating that a higher ESG disclosure leads
puted as the average score of ESG of all the firms in a particular year to lower default risk. Therefore, our results are robust to the use of
except firm i's score of ESG in that year.10 The intuition behind using the 2SLS approach and seem not to be driven by an endogeneity bias.
this instrument is that a firm's ESG might be highly related to the
industry peers due to their similar business mix and investment oppor-
tunities, but such an industry average is unlikely to directly affect a 4.2.3 | PSM estimate
firm's default risk. Moreover, firm-level default risk may affect firm-
level ESG, whereas it is less likely to affect industry-level ESG. Given To further eliminate the endogeneity bias, we execute two-step PSM,
these arguments, the industry median ESG should be a valid instru- in which firm-years with higher ESG are matched with firm-years with
ment, since it is related to firm-level ESG but unrelated to firm-level lower ESG, but with no significant differences in terms of all other var-
default risk. iables (see, e.g., Atif et al., 2021; Chen et al., 2017). Specifically, in the
Column 1 in Table 5 reports the first-stage regression results, first step, we create a dummy variable (high-ESG) that takes the value
where the dependent variable is the ESG. The explanatory variables ‘1’ if ESG score is greater than the sample median and ‘0’ otherwise.
include the instrumental variable and the same control variables as in Firms with ESG scores greater than the sample median are considered
Equation 1. The coefficient on ESG (IND) is statistically significant at in the treatment group whereas firms with ESG scores lower than
the 1% level, suggesting that the instrument is valid because of its rel- sample median are included in the control group. We run the logit
evance and its statistical power to explain ESG. We also check the regression for high-ESG with all the explanatory variables specified in
validity of the instrument. First, the reported F statistics of the first- Equation 1. The predicted estimates from the logit regression are used
stage regression is very high, which indicates that our instrument is as the propensity scores for each firm-year observation. The logit
not weak. Moreover, as the Cragg–Donald's Wald F statistic is greater regression results are reported in Column 1, Panel A, Table 6. The
3946 ATIF AND ALI
T A B L E 5 ESG and default risk—

First stage Second stage
endogeneity bias (Instrumental variable
ESG DD CDS approach)
(1) (2) (3)
ESG (fitted) 0.338*** (3.63) 0.296*** (3.33)
ESG (IND) 1.013*** (3.76)
SIZE 4.317*** (17.06) 0.950** (2.27) 1.008** (2.55)
ROA 4.016*** (2.81) 6.427*** (9.43) 3.665*** (5.71)
TLTA 5.507*** (3.19) 3.013*** (3.78) 0.602 (0.86)
LIQ 0.171 (1.35) 0.037 (0.65) 0.043 (0.89)
TANG 0.528 (0.32) 0.486 (0.81) 0.986* (1.72)
SALES 0.268 (0.50) 0.123 (0.60) 0.212 (1.26)
TOBINQ 1.054*** (5.32) 0.411*** (3.58) 0.017 (0.17)
Constant 9.358*** (5.70) 3.718** (2.35)
Year effects Yes Yes Yes
Industry effects No No No
F statistics 27.7
Cragg–Donald 14.03
N 5206 5206 5206
Note: This table presents the regression results between ESG and default risk using 2SLS. Column 1
presents the result of first-stage regression, and Columns 2–3 present second-stage regression. Figures
in parenthesis are the t statistics. See Appendix B for variable definitions.
regression results suggest that firms with higher ESG are larger in size In the second diagnostic test, we examine the difference for each
and have lower leverage, more tangibility, more sales and a higher firm observable characteristic between the treated firms and the control
value (TOBINQ). The pseudo-R2 for the regression is high, with a value firms, reporting the results in Panel B, Table 6. Again, we find no sig-
of 34.05%. nificant difference in observable characteristics between the two
In the second step, we use the propensity scores to form groups. These results suggest that PSM removes other observable dif-
one-to-one matched pairs, ensuring that the firms with ESG higher ferences, increasing the probability that any difference in the default
than the median (treatment group) are sufficiently similar to the firms risk between the treated and control groups can be attributed to dif-
with ESG lower than the median (control group). Specifically, each firm ferences in ESG rather than to differences in other variables.
with higher ESG was matched to a firm with lower ESG using the clos- The PSM estimates and the multivariate results using the matched
est propensity score. We employ matching with replacement, requir- sample are reported in Panels C and D, Table 6, respectively. As evident
ing that the difference in the propensity scores of the treatment and in Panel C, Table 6, we find significant differences in both default risk
the matched firms does not exceed 0.1% in absolute value. With the measures between the treatment and control group.11 In particular, we
application of these criteria, we are able to effectively match 1922 find that the CDS spread is lower and distance to default is higher in the
firm-year observations; hence, our treatment and control groups are firms with higher ESG than in the otherwise indistinguishable firms with
nearly identical along all explanatory variables, with one relatively lower ESG. Likewise, the multivariate results reported in
exception: ESG. Panel D, Table 6 show that firms with higher ESG have lower default
We run two diagnostic tests to verify that the treatment and con- risk. These results from the propensity match method suggest that the
trol groups are indistinguishable in terms of observable characteristics. variation in default risk is attributable to the systematic differences in
In the first diagnostic test, we re-estimate the logit regression for the ESG. Hence, we infer that greater ESG leads to greater stability in firms.
post-match sample. The results are reported in Column 2, Panel A, Natural experiment is one of the state-of-the-art solutions to the
Table 6. Most of the regression coefficients are statistically insignificant endogeneity problem, as suggested by several empirical studies (see,
and are smaller than those in the column pre-match, suggesting that e.g., Black et al., 2015; Chen et al., 2015; Gippel et al., 2015). How-
both groups are almost similar in terms of observable characteristics. ever, the implementation of such a methodology requires a purely
2
The pseudo-R drops substantially from 34.05% for the pre-match sam- exogenous natural event, such as mandatory ESG disclosure reforms.
ple to 0.1% for the post-match sample. This implies that the PSM In the context of our study, we are unable to use natural experimenta-
removes all observable differences other than the difference in ESG. tion because of the absence of any mandatory ESG disclosure reforms
ATIF AND ALI 3947
TABLE 6 ESG and default risk—endogeneity bias (propensity score matching)
Panel A: Pre-match regression and post-match diagnostic test
Post-match
Pre-match
High-ESG
High-ESG
(1) (2)
SIZE 0.992*** (31.73) 0.014 (0.22)
ROA 0.539 (1.41) 0.333 (0.60)
TLTA 0.928*** (4.57) 0.123 (0.28)
LIQ 0.013 (0.54) 0.033 (0.70)
TANG 0.882*** (5.47) 0.156 (0.40)
SALES 0.263*** (4.47) 0.030 (0.20)
TOBINQ 0.102*** (3.39) 0.020 (0.34)
Constant 7.945*** (26.21) 0.353 (0.58)
Year effects No No
Industry effects No No
Observations 5212 1922
2
Pseudo-R .3405 .001
Panel B: Mean differences
Treated Control Difference t test

SIZE 7.593 7.594 0.002 0.03
ROA 0.046 0.042 0.004 0.84
TLTA 0.539 0.542 0.003 0.27
LIQ 2.422 2.494 0.072 0.81
TANG 0.280 0.284 0.004 0.38
SALES 1.060 1.059 0.000 0.01
TOBINQ 1.924 1.947 0.023 0.40
Panel C: PSM estimator
Variables Treated Controls Difference t statistics

DD 6.27 5.56 0.71 4.99***
CDS 0.17 0.56 0.39 4.28***
Panel D: Multivariate regression
DD CDS
ESG 0.022*** (3.64) 0.022*** (4.50)
SIZE 0.420*** (9.32) 0.093** (2.51)
ROA 6.841*** (14.42) 3.610*** (9.24)
TLTA 5.123*** (18.00) 1.955*** (8.34)
LIQ 0.050 (1.59) 0.065** (2.51)
TANG 0.200 (0.69) 0.737*** (3.07)
SALES 0.094 (1.19) 0.320*** (4.88)
TOBINQ 0.809*** (18.20) 0.239*** (6.54)
Constant 3.733*** (5.44) 0.008 (0.01)
Year effects Yes Yes
Industry effects Yes Yes
N 1922 1922
Adj. R2 .518 .203
F 80.266 19.846
3948 ATIF AND ALI
Note: This table reports the regression results between ESG disclosure and default risk using propensity score matching estimation. Panel A reports the
parameter estimates from the logit model used to estimate the propensity scores. The dependent variable, High-ESG, is an indicator variable set to 1 if ESG
score is greater than median and 0 otherwise. Panel A reports the pre-match propensity score regression (in Column 1) and post-match diagnostic
regression (in Column 2). Panel B reports the univariate comparisons with corresponding t statistics of firm characteristics between firms with high and low
ESG. Panel C reports the average treatment effect on default risk for the matched sample. Panel D reports multivariate results relating to ESG and default
risk for matched sample. The dependent variable is DD in Column 1 and CDS in Column 2. Figures in parenthesis are the t statistics. See Appendix B for
variable definitions.
in the United States during our sample period. Therefore, the disclo- equals 1 if ROA < 0 and 0 otherwise. ACC is the difference between
sure of ESG is still at the discretion of the firm due to the absence of net income and operating cash flow, scaled by total assets. PAYER, a
any credibly exogenous reform. Overall, we acknowledge that it is not dummy variable, equals 1 if PAYOUT > 0 and 0 otherwise. PAYOUT is
possible to completely eliminate endogeneity. However, our results, the ratio of dividend paid divided by net income.
based on lagged independent variables, an instrumental variable and First, we capture conditional dispersion in the future earnings dis-
the PSM technique, provide a certain degree of comfort that even tribution using the predicted interquartile range (IQR) = Qi75 Qi25, a
when considering endogeneity due to omitted variable and reverse commonly used measure of earning dispersion. The higher the IQR,
causality, the main findings regarding the relationship of ESG disclo- the higher is the uncertainty in future earnings. Second, we estimate
sure to default risk do not change. conditional skewness in the future earnings distribution as SKEW = [(
Qi75 Qi50) (Qi50 Qi25)]/IQRi, which captures the balance between
upside risk relative to downside risk in future earnings within the two
5 | ROLE OF PROFITABILITY, middle quartiles. Third, we estimate conditional kurtosis in the future
PERFORMANCE VARIABILITY AND COST OF earnings distribution as KURT = [(Qi87.5 Qi62.5) + (Qi37.5 Qi12.5)]/
DEBT (H2) IQRi, which measures the density of the distribution close to Q25 and
Q75, relative to the density close to Q50. Finally, we decompose the
In Section 2, we argue that profitability, performance variability and numerator of KURT to capture asymmetry in the relative densities of
cost of debt/capital are the important mechanisms through which the distribution surrounding Q25 and Q75 as UPi = [(Qi87.5 Qi62.5)]/
ESG disclosure influences default risk. In this section, we empirically IQRi and DOWNi = [(Qi37.5 Qi12.5)]/IQRi. The higher the SKEW and
test these assertions. KURT, the higher is the uncertainty in the future earnings realization.
First, we investigate the relationship between ESG and default Table 8 presents the results between ESG and the five perfor-
risk via profitability. We use two common measures of profitability: mance variability measures, IQR (Column 1), SKEW (Column 2), KURT
(1) return on assets (ROA), measured as net income divided by total (Column 3), UP (Column 4) and DOWN (Column 5). We find that the
assets, and (2) return on equity (ROE), measured as net income divided coefficients of ESG are negative and significant for all performance
by common shareholder equity. Table 7 presents the interaction variability proxies expect SKEW. Overall, these results provide empiri-
effect of ESG and profitability (using ROA and ROE) on default risk in cal support to our argument that ESG disclosure enhances perfor-
Columns 1–2 and 3–4 for DD and CDS, respectively. As expected, we mance or earning stability and thus lowers the level of default risk.
find that the coefficients on the interaction terms (ESG * ROA and Finally, we test the linkage between ESG and default risk via cost
ESG * ROE) are statistically significant, positive for DD and negative of debt. On the one hand, the firms experience default if they are
for CDS. These results suggest that firms with better ESG disclosure unable to meet contractual debt payment (i.e., principal and interest
have higher profitability and thus have a lower level of default risk. or cost of debt). On the other hand, firms with more investment in
Second, we examine the association between ESG and perfor- ESG activities are considered safer by debtholders, and thus, they
mance variability. We measure performance variability using the qua- require less risk premium, which reduces the cost of debt (Eliwa
ntile regression forecasts of the interquartile range of the distribution et al., 2019). We tested if this is true for our sample. We obtain the
of profitability: that is, ROA (see, Konstantinidi & Pope, 2016). This cost of debt data for our sample firms from the Bloomberg database.
approach, which relies only on cross-sectional fundamental character- First, we test the effect of cost of debt on default risk, reporting
istics, requires time series data for computation. We forecast perfor- our results in Table 9, with Column 1 for DD and Column 2 for CDS.
mance in different quantiles using the following equation: We find a significant negative (positive) coefficient of cost of debt
with DD (CDS), suggesting that firms have higher default risk if the

QUANT q ðROAit Xit Þ ¼ βq0t þ βq1t ROAit1 þ βq2t LOSSit1 þ βq3t ðLOSSit1 ROAit1 ð2Þ cost of debt is higher. Second, we test the impact of ESG on cost of
þβq4t ACCit1 þ βq5t PAYERit1 þ βq6t PAYOUTit1 : debt and reported the results in Column 3, Table 9. We find a signifi-
cant negative coefficient on ESG with cost of debt, which implies that
Our model resembles that by Correia et al. (2018), with the firms with higher ESG disclosure have a lower cost of debt. Taken
exception that we forecast return on assets (ROA), rather than return together, these results indicate that ESG reduces default risk by
on net operating assets equity (RNOA). LOSS, a dummy variable, reducing the cost of debt.
ATIF AND ALI 3949
TABLE 7 ESG and default risk—role of profitability
DD DD CDS CDS
(1) (2) (3) (4)
ESG 0.010 (1.56) 0.017*** (4.90) 0.021** (2.57) 0.030*** (3.51)
ROA 2.359*** (2.61) 0.259 (0.28)
ESG * ROA 0.210*** 0.217***
ROE 0.536*** (5.73) 0.153 (1.28)
ESG * ROE 0.007* (1.67) 0.012*** (2.59)
SIZE 0.416*** (9.48) 0.593*** (24.94) 0.122* (1.91) 0.213*** (3.44)
TLTA 4.780*** (14.55) 6.260*** (35.75) 2.089*** (6.15) 2.890*** (8.54)
LIQ 0.014 (0.34) 0.042** (2.25) 0.094** (2.49) 0.109*** (2.77)
TANG 0.025 (0.08) 0.043 (0.22) 0.900** (2.41) 0.934** (2.42)
SALES 0.050 (0.48) 0.369*** (6.44) 0.332*** (2.62) 0.166 (1.27)
TOBINQ 0.674*** (11.63) 0.619*** (25.97) 0.229*** (4.41) 0.213*** (3.95)
Year effects Yes Yes Yes Yes
Industry effects Yes Yes Yes Yes
Constant 3.563*** (6.43) 1.100 (0.65) 1.674*** (3.28) 2.942*** (5.99)
N 5206 5206 5206 5206
Adj. R2 .519 .468 .261 .227
Note: This table reports the regression results between ESG, profitability and default risk using pooled OLS. Figures in parenthesis are the t statistics. See
TABLE 8 ESG and default risk—role of performance variability
IQR SKEW KURT UP DOWN

(1) (2) (3) (4) (5)
ESG 0.0000*** (3.02) 0.0002 (1.11) 0.0018*** (6.20) 0.0006** (2.39) 0.0013*** (8.18)
SIZE 0.0015*** (12.40) 0.0051*** (3.71) 0.0015 (0.64) 0.0017 (0.86) 0.0005 (0.39)
ROA 0.2595*** (38.91) 1.9362*** (26.40) 3.9140*** (30.19) 3.3458*** (30.29) 0.5524*** (11.32)
TLTA 0.0063*** (5.51) 0.0265* (1.81) 0.1599*** (7.31) 0.0780*** (4.07) 0.0782*** (7.62)
LIQ 0.0005*** (4.55) 0.0030** (2.39) 0.0000 (0.01) 0.0049*** (2.77) 0.0048*** (3.83)
TANG 0.0058*** (6.12) 0.0851*** (7.15) 0.1441*** (7.73) 0.1953*** (12.01) 0.0557*** (5.88)
SALES 0.0011*** (4.16) 0.0077** (2.41) 0.0229*** (4.50) 0.0072* (1.65) 0.0155*** (5.84)
TOBINQ 0.0005*** (2.72) 0.0025 (1.12) 0.0026 (0.64) 0.0150*** (4.45) 0.0178*** (7.99)
Constant 0.0325*** (17.31) 0.0791*** (4.06) 1.9704*** (60.81) 0.7763*** (28.07) 1.1878*** (68.57)
Year effects Yes Yes Yes Yes Yes
Industry Yes Yes Yes Yes Yes
effects
N 4187 4187 4187 4187 4187
Adj. R2 .697 .565 .624 .628 .226
Note: This table reports the regression results between ESG performance variability using pooled OLS. Figures in parenthesis are the t statistics. See
3950 ATIF AND ALI
T A B L E 9 ESG and default risk—role

DD CDS COD
(1) (2) (3)
of cost of debt (COD)
COD 0.249*** (7.10) 0.170*** (4.79)

ESG 0.008*** (3.57)
SIZE 0.551*** (16.64) 0.273*** (6.36) 0.115*** (5.83)
ROA 3.663*** (10.88) 2.178*** (5.89) 0.614*** (3.24)
TLTA 4.190*** (13.16) 1.896*** (6.28) 2.083*** (13.31)
LIQ 0.020 (0.67) 0.042 (1.30) 0.036** (2.26)
TANG 0.402 (1.25) 0.520 (1.47) 0.301** (2.03)
SALES 0.071 (0.80) 0.225** (2.05) 0.173*** (3.74)
TOBINQ 0.594*** (10.84) 0.293*** (6.61) 0.116*** (5.94)
Constant 6.031*** (2.85) 1.396* (1.67) 3.726*** (16.35)
Year effects Yes Yes Yes
Industry effects Yes Yes Yes
N 7255 7262 5196
Adj. R2 .516 .234 .438
Note: This table reports the pooled OLS regression results between cost of debt and default risk in
Columns 1–2 and between ESG and cost of debt in Column 3. Figures in parenthesis are the t statistics.
See Appendix B for variable definitions.
6 | ADDITIONAL ANALYSES variables are the same as in Equation 1. Column 1 includes the regres-
sion results using the Dickinson (2011) life cycle proxy. Compared to
6.1 | Role of corporate life cycle the shake-out stage, the introduction, growth and decline stages are
significantly negatively associated with DD, while the mature stage of
In this section, we report our exploration of the effect of ESG disclo- the life cycle is positively associated with DD (p < .01). Columns 2 and
sure on default risk at various stages of the firm life cycle, as argued in 3 include the regression results using the RETA and AGE life cycle
the introduction. We use three alternative methods to classify firms proxies, respectively. These results suggest that both RETA and AGE
into different life cycle stages. First, we adopt the cash flow-based are positively and significantly associated with DD (p < .01). Overall,
method developed by Dickinson (2011), who argues that cash flow these results suggest that mature and older firms have lower default
captures differences in firm profitability, growth and risk. Thus, one risk whereas introductory and younger firms have higher default risk.
may use operating (CFO), investing (CFI) and financing (CFF) cash Second, and more importantly, we investigate the interactive
flows to classify firms into five stages: introduction, growth, mature, effect of ESG and firm life cycle on default risk. These results are
shake-out and decline. The firm is considered in the introduction stage reported in Panel B, Table 10. All the control variables are the same as
if CFO < 0, CFI < 0 and CFF > 0; growth, if CFO > 0, CFI < 0 and in Equation 1. Columns 4–6 include the regression results using
CFF > 0; mature, if CFO > 0, CFI < 0 and CFF < 0; and decline, if Dickinson (2011) life cycle stages, RETA and AGE life cycle proxies,
CFO < 0, CFI > 0 and CFF ≤ 0 or CFF ≥ 0, with the remaining firm- respectively. The results suggest that the coefficient on ESG and Dick-
years classified under the shake-out stage.12 inson (2011) life cycle stages with DD is significant and positive only
Second, we follow DeAngelo et al. (2006) and use retained earn- for the mature stage. Moreover, both the coefficients on the interac-
ings to total assets (RETA) as a proxy for the life cycle. RETA measures tion between ESG * RETA and ESG * AGE with DD are positive and sta-
the extent to which a firm is reliant on internal or external financing. tistically significant. These results imply that firms with higher ESG
Firms with high RETA are mature or older, with declining investment, disclosure reduce default risk for only those firms that are mature
whereas firms with low RETA are in the young and growth stages. and older.
Finally, as life cycle stages naturally relate to firm age, we use firm age Finally, we use a split sample strategy to investigate the effect of
as a simple and natural proxy of life cycle. We measure firm age as a ESG on default risk in different life cycle stages. For a Dickinson (2011)
difference between the current fiscal year and the year of life cycle proxy, we run separate regressions for firms in introduction,
incorporation. growth, mature and decline stages. For RETA and AGE life cycle prox-
First, we investigate the direct effect of firm life cycle on default ies, we split the sample into old firms, with RETA and AGE higher than
risk. The results are reported in Panel A, Table 10. All the control the sample median, and as young, with RETA and AGE lower than the
TABLE 10 ESG and default risk—role of corporate life cycle
ATIF AND ALI
Panel A: Life cycle and default risk Panel B: ESG, life cycle and default risk
DD DD DD DD DD DD
(1) (2) (3) (4) (5) (6)
INTRO 0.949*** (5.42) 1.606*** (4.11)
GROWTH 0.335*** (2.58) 0.439* (1.95)
MATURE 0.660*** (5.74) 0.113 (0.54)
DECLINE 0.539*** (2.96) 0.513 (1.09)
RETA 0.705*** (9.69) 0.083 (0.30)
AGE 0.012*** (4.90) 0.003 (0.61)
ESG 0.004 (0.50) 0.006 (0.66) 0.002 (0.17)
ESG * INTRO 0.004 (0.17)
ESG * GROWTH 0.004 (0.39)
ESG * MATURE 0.026*** (2.88)
ESG * DECLINE 0.018 (0.58)
ESG * RETA 0.065*** (3.19)
ESG * AGE 0.000** (1.96)
SIZE 0.480*** (14.52) 0.432*** (13.23) 0.510*** (14.95) 0.374*** (15.72) 0.299*** (6.50) 0.420*** (9.19)
ROA 3.036*** (9.49) 2.083*** (7.05) 3.694*** (10.59) 4.226*** (18.12) 2.876*** (6.55) 5.221*** (11.70)
TLTA 4.530*** (15.12) 4.112*** (14.16) 4.815*** (14.96) 4.526*** (26.39) 3.877*** (11.66) 4.990*** (14.55)
LIQ 0.035 (1.21) 0.020 (0.69) 0.009 (0.30) 0.020 (1.11) 0.018 (0.46) 0.019 (0.50)
TANG 0.192 (0.62) 0.062 (0.20) 0.315 (0.96) 0.190 (1.06) 0.199 (0.64) 0.023 (0.07)
SALES 0.157* (1.79) 0.144* (1.65) 0.075 (0.83) 0.080 (1.42) 0.049 (0.48) 0.035 (0.34)
TOBINQ 0.629*** (11.63) 0.704*** (15.26) 0.624*** (10.80) 0.718*** (31.74) 0.742*** (12.37) 0.708*** (12.33)
Constant 5.934** (2.50) 5.845*** (2.62) 1.960*** (5.49) 4.635*** (2.91) 4.808*** (8.43) 3.756*** (6.63)
Year effects Yes Yes Yes Yes Yes Yes
Industry effects Yes Yes Yes Yes Yes Yes
N 7277 7182 7134 5206 5184 5098
Adj. R2 .528 .533 .517 .538 .553 .524
Note: This table reports the pooled OLS regression results between corporate life cycle and default risk in Columns 1–3 (Panel A) and between ESG, life cycle and default risk in Columns 4–6 (Panel B). Figures in
parenthesis are the t statistics. See Appendix B for variable definitions.
3951
3952 ATIF AND ALI
sample median. The unreported results indicate that ESG reduces to check if the relationship between ESG disclosure and default risk is
default risk only during Dickinson's mature stage. ESG disclosure sig- consistent across different periods. Our results reveal that the impact
nificantly reduces default risk for old age firms only and significantly of ESG on default risk exists for both pre-GFC and post-GFC periods.
reduces default risk only for high RETA firms (i.e., mature firms). Finally, we test the interaction effect of ESG disclosure and GFC on
default risk and find that the interaction term is positive and statisti-
cally significant at the 1% level. These results indicate that firms with
6.2 | Components of ESG and default risk high ESG disclosure had lower default risk during the GFC period.
Table 11 reports the results between the components of ESG disclo-

sure (environmental, social and governance) and default risk. These 6.4 | Alternative measure of ESG
analyses are important to rule out the possibility that our results are
driven by one of the components. Columns 1–3 of Table 11 show that Our analyses so far have been based on the ESG disclosure data
ESG disclosures are positively and significantly associated with DD: (instead of ESG performance) obtained from the Bloomberg database.
that is, a lower default risk. Overall, these results suggest that all the Cho et al. (2015) argue that managers may engage in the deceitful
components of ESG are important determinant of default risk for US behaviour of creating organizational facades for signalling purposes.
firms. This practice may result in ESG performance differing from ESG dis-
closure (Tashman et al., 2019). With this in mind, we use an alterna-
tive measure of ESG from the Thomson Reuters Eikon (Asset4)
6.3 | Excluding GFC database which is designed to transparently and objectively measure
a company's relative ESG performance, commitment and effective-
For this section, we exclude the extreme observations during the ness, in terms of environmental aspects (i.e., resource use, emissions
global finance crisis (GFC) and then investigate the impact of ESG on and innovation), social aspects (workforce, human rights, community
default risk. We consider observations during 2008 and 2009 as GFC. and product responsibility) and governance aspects (management,
Our untabulated results suggest that even when we exclude GFC stakeholders and corporate social responsibility strategy). The results
years, our main inference does not change: that is, ESG disclosure are reported in Table 12. Consistent with our baseline results reported
reduces default risk. We also conduct pre-GFC and post-GFC analyses in Table 3, we find that overall ESG performance and its factors
TABLE 11 ESG components and default risk
DD DD DD CDS CDS CDS

(1) (2) (3) (4) (5) (6)
G factor 0.042*** (3.82) 0.053*** (3.99)
E factor 0.016*** (2.79) 0.020*** (2.61)
S factor 0.015*** (2.62) 0.024*** (3.29)
SIZE 0.437*** (10.71) 0.271*** (3.62) 0.376*** (6.41) 0.169*** (3.05) 0.128 (1.22) 0.094 (1.13)
ROA 5.187*** (11.88) 5.391*** (4.85) 7.112*** (9.77) 2.620*** (5.63) 3.163*** (3.44) 4.378*** (6.33)
TLTA 4.875*** (14.82) 5.044*** (9.49) 4.698*** (10.36) 2.216*** (6.27) 2.310*** (3.01) 2.083*** (4.07)
LIQ 0.016 (0.41) 0.061 (0.76) 0.067 (1.21) 0.097** (2.56) 0.003 (0.04) 0.134** (2.42)
TANG 0.021 (0.06) 0.343 (0.84) 0.211 (0.54) 0.871** (2.29) 1.233** (2.20) 1.376*** (2.82)
SALES 0.050 (0.48) 0.537*** (3.18) 0.150 (1.14) 0.334** (2.57) 0.610** (2.54) 0.494*** (2.68)
TOBINQ 0.711*** (12.19) 1.400*** (9.62) 0.900*** (9.40) 0.273*** (5.00) 0.815*** (5.45) 0.347*** (3.90)
Constant 1.635*** (2.63) 4.409*** (4.31) 4.171*** (4.94) 4.160*** (5.90) 1.159 (0.97) 0.468 (0.60)
Year Yes Yes Yes Yes Yes Yes

effects
Industry Yes Yes Yes Yes Yes Yes
effects
N 5206 2034 3429 5206 2036 3428
Adj. R2 .512 .608 .536 .240 .362 .250
Note: This table reports the pooled OLS regression results between individual components of ESG and default risk. Figures in parenthesis are the t statistics. See
ATIF AND ALI
TABLE 12 Alternative measure of ESG—Asset4
DD DD DD DD CDS CDS CDS CDS

(1) (2) (3) (4) (5) (6) (7) (8)
ESG 0.007* (1.82) 0.013*** (3.77)
G factor 0.003* (1.85) 0.006** (1.97)
E factor 0.004*** (2.78) 0.007*** (2.63)
S factor 0.005*** (3.04) 0.011*** (3.26)
SIZE 0.406*** (6.88) 0.443*** (17.80) 0.417*** (15.70) 0.419*** (15.72) 0.233*** (3.59) 0.296*** (4.41) 0.253*** (3.88) 0.244*** (3.77)
ROA 8.208*** (8.52) 8.175*** (16.28) 8.238*** (16.40) 8.339*** (16.43) 5.700*** (6.85) 5.627*** (6.76) 5.754*** (6.85) 5.856*** (6.93)
TLTA 5.799*** (10.91) 5.824*** (28.54) 5.804*** (28.41) 5.799*** (28.35) 3.140*** (6.04) 3.180*** (6.08) 3.148*** (5.99) 3.121*** (5.97)
LIQ 0.052 (0.96) 0.055** (2.05) 0.053** (2.00) 0.052** (1.96) 0.070 (1.30) 0.072 (1.33) 0.070 (1.30) 0.070 (1.30)
TANG 0.376 (0.85) 0.348** (2.07) 0.359** (2.13) 0.354** (2.10) 1.103** (2.39) 1.064** (2.28) 1.075** (2.32) 1.085** (2.35)
SALES 0.274* (1.82) 0.271*** (4.61) 0.266*** (4.52) 0.270*** (4.58) 0.477*** (2.77) 0.474*** (2.70) 0.461*** (2.66) 0.471*** (2.76)
TOBINQ 0.789*** (8.74) 0.802*** (20.08) 0.794*** (19.87) 0.786*** (19.66) 0.385*** (4.15) 0.407*** (4.31) 0.392*** (4.18) 0.376*** (4.06)
Constant 3.409*** (4.51) 3.144*** (9.48) 3.440*** (10.20) 3.377*** (10.13) 2.114*** (2.74) 2.605*** (3.30) 2.052*** (2.66) 2.101*** (2.73)
Year effects Yes Yes Yes Yes Yes Yes Yes Yes
Industry effects Yes Yes Yes Yes Yes Yes Yes Yes
SE robust Yes Yes Yes Yes Yes Yes Yes Yes
N 4714 4714 4714 4712 4725 4725 4725 4723
Adj. R2 .553 .552 .552 .553 .338 .331 .333 .337
Note: This table presents the regression results between ESG and default risk using pooled OLS. ESG and its factors are from Thomson Reuters Eikon (Asset4). Dependent variable is distance to default, that is, DD (or credit
default swap spread, i.e., CDS), in Columns 1–3 (Columns 4–6). Figures in parenthesis are the t statistics. See Appendix B for variable definitions.
3953
3954 ATIF AND ALI
3
(i.e., E, S and G) are significantly and positively (negatively) associated Despite the increasing importance of ESG, critics (e.g., Friedman, 1970)
with DD (CDS), suggesting that the firms with higher ESG performance have supported the view that businesses exist to make profit, and thus,
no social responsibility should be assigned to businesses that incur
have a lower level of default risk.
extra expenses.
4
Prior literature examines default risk from an array of factors: stock
market liquidity (Brogaard et al., 2017), incentive structure (Bennett
7 | C O N CL U S I O N et al., 2015), innovation performance (Hsu et al., 2015), debt maturity
choice (Goyal & Wang, 2013), stock market return and volatility
(Giesecke et al., 2011), as well as corporate governance characteristics
ESG performance (also termed nonfinancial or extra-financial informa-
such as CEO compensation, board independence and ownership struc-
tion) has emerged as an important consideration for evaluating invest- ture (see, e.g., Acrey et al., 2011; Ali et al., 2018; Chiang et al., 2015;
ment decisions. On the other hand, default risk has become an Chiu & Wagner, 2012; Miglani et al., 2015; Schultz et al., 2017;
important yardstick of firms' financial health due to its severe conse- Switzer & Wang, 2013; Vallascas & Hagendorff, 2013).
5
quences for firms. However, previous literature has not revealed how A better ESG disclosure also indicates better management quality, sig-
ESG disclosure influences default risk and through which mechanisms. nalling superior management of firm financial risk.
6
We explore the effect of ESG disclosure on default risk for the non- The ESG disclosure data from Bloomberg is mainly available from 2006
onward which is the main reason we start our sample period
financial US firms over the period from 2006 to 2017. Our results sug-
from 2006.
gest that ESG disclosure is positively related to default risk as proxied 7
Firms' ESG performance and relevant nonfinancial disclosure have
by the distance to default (DD) measure and is negatively related to gained global focus. With this ever-increasing attention, three main
credit default swap spread (CDS). We also explore various mechanisms agencies (Bloomberg, MSCI and Thomson Reuters) provide ESG ranking
through which ESG disclosure reduces default risk. We find that the across different industries and geographic locations. These agencies
implement different scoring frameworks, along with uniform interna-
inverse effect ESG disclosure on default risk is through increased prof-
tional ESG guidelines such as from the GRI and the UN. Bloomberg pro-
itability and reduced performance variability. We also find that ESG vides real-time individual and aggregate ESG disclosure scores ranging
reduces default risk via cost of debt. We specifically find that firms from 0.1 to 100, mainly from 2006, which marks the start of our sample
with a lower cost of debt have a lower default risk and firms with high period. These scores are mainly derived from ESG information publi-
shed by the firms (e.g., annual reports and websites of firms), from
ESG have a lower cost of debt. Finally, we document that the negative
information provided by public sources (e.g., super funds) and
impact of ESG on default risk exists only for mature and older firms.
from questionnaires sent to firms in order to get information regarding
Our findings are consistent when using alternative variable specifica- ESG concerns.
tions and alternative estimation methods, including lagged indepen- 8
Market-based default risk measures overcome the criticism of
dent variables, an instrumental variable approach and PSM. accounting-based models: the forward-looking nature of market data
Our study contributes to the stream of research that explores the reflects the expectations of a firm's future cash flows and hence should
be more appropriate for prediction purposes (Beaver et al., 2005).
determinants of firms' default risk. We extend this literature by pro-
9
We multiply the standard deviation of ESG, that is, 12.88, with the coef-
viding robust evidence on the role of aggregate ESG disclosure as crit-
ficient on ESG, that is, 0.025, in Column 3 of Table 3 to get 0.322.
ical determinants of default risk for nonfinancial US firms. Moreover, Therefore, one standard deviation increase in the ESG increases the
our study extends the literature on the consequences of ESG by distance to default (DD) by 0.322 points. As the mean distance to
examining various mechanisms through which ESG disclosure reduces default (DD) is 5.98, an increase by 0.322 denotes a change by 5.40%
of the average default risk.
default risk. Finally, we contribute to the firm life cycle literature by
10
Several governance studies have used this instrument (see, e.g., Ali
providing empirical evidence that the corporate life cycle moderates
et al., 2018; Jiraporn et al., 2011; Liu et al., 2014; Liu et al., 2015;
the effect of ESG disclosure on default risk. Our findings are impor-
Yang & Zhao, 2014).
tant for the potential of SRI to screen investment and to facilitate 11
The difference in the means between the treatment and control groups
fund availability for firms with a better ESG outlook. Our findings also is the PSM of the average treatment effect on the treated (ATT).
provide insights for debtholders to consider a firm's sustainability ini- 12
The life cycles of firms are categorized into five stages: introduction,
tiative before making decisions. Given our findings, policy makers growth, mature, shake-out and decline. Five dummy variables are cre-
aiming at sustainability should encourage firms to disclose their sus- ated for each of the five stages. However, to avoid the problem of
dummy variable trap multicollinearity in the regression model, one
tainability initiatives.
of the stages is dropped. As the shake-out stage of the life cycle is
ambiguous in theory (Dickinson, 2011), we drop this stage in the
ENDNOTES regression model.
1
For instance, investors taking nonfinancial information into consider-
ation before making investment decisions have become essential in the
light of UN Principles for Responsible Investment (Capelle-Blancard & DATA AVAILABILITY STAT EMEN T
Monjon, 2014). The default risk data is sourced from publicly available database
2
Default (a) adversely affects productivity through supply chain interrup- whereas ESG data is sourced from privately subscribed database.
tions and employee attrition, (b) incurs legal and administrative costs,
(c) destroys some or all of the value of investment the shareholders
have made and (d) harms customer retention and employee well-being OR CID
(e.g., Brogaard et al., 2017). Searat Ali https://orcid.org/0000-0002-7250-8116
ATIF AND ALI 3955
RE FE R ENC E S Bouslah, K., Kryzanowski, L., & M'Zali, B. (2013). The impact of the dimen-
Acharya, V., Huang, J. Z., Subrahmanyam, M., & Sundaram, R. K. (2006). sions of social performance on firm risk. Journal of Banking & Finance,
When does strategic debt-service matter? Economic Theory, 29(2), 37(4), 1258–1273. https://doi.org/10.1016/j.jbankfin.2012.12.004
363–378. https://doi.org/10.1007/s00199-005-0035-9 Brogaard, J., Li, D., & Xia, Y. (2017). Stock liquidity and default risk. Journal
Acrey, J. C., McCumber, W. R., & Nguyen, T. H. T. (2011). CEO incentives of Financial Economics, 124(3), 486–502. https://doi.org/10.1016/j.
and bank risk. Journal of Economics and Business, 63(5), 456–471. jfineco.2017.03.003
https://doi.org/10.1016/j.jeconbus.2010.09.002 Brown, T. J., & Dacin, P. A. (1997). The company and the product: Corpo-
Ali, S. (2017). Women in the boardroom and their impact on default risk: A rate associations and consumer product responses. Journal of Market-
pitch. Accounting Research Journal, 30(2), 137–146. https://doi.org/10. ing, 61(1), 68–84. https://doi.org/10.1177/002224299706100106
1108/ARJ-07-2016-0092 Brown, W. O., Helland, E., & Smith, J. K. (2006). Corporate philanthropic
Ali, S., Hussain, N., & Iqbal, J. (2021). Corporate governance and the insol- practices. Journal of Corporate Finance, 12(5), 855–877. https://doi.
vency risk of financial institutions. The North American Journal of Eco- org/10.1016/j.jcorpfin.2006.02.001
nomics and Finance, 55, 101311. https://doi.org/10.1016/j.najef.2020. Campbell, J. Y., Hilscher, J., & Szilagyi, J. (2008). In search of distress risk.
101311 The Journal of Finance, 63(6), 2899–2939. https://doi.org/10.1111/j.
Ali, S., Liu, B., & Su, J. J. (2018). Does corporate governance quality affect 1540-6261.2008.01416.x
default risk? The role of growth opportunities and stock liquidity. Inter- Capelle-Blancard, G., & Monjon, S. (2014). The performance of socially
national Review of Economics and Finance, 58, 422–448. https://doi. responsible funds: Does the screening process matter? European
org/10.1016/j.iref.2018.05.003 Financial Management, 20(3), 494–520. https://doi.org/10.1111/j.
Allen, L., DeLong, G., & Saunders, A. (2004). Issues in the credit risk model- 1468-036X.2012.00643.x
ing of retail markets. Journal of Banking & Finance, 28(4), 727–752. Carter, C. R. (2005). Purchasing social responsibility and firm performance:
https://doi.org/10.1016/S0378-4266(03)00197-3 the key mediating roles of organizational learning and supplier perfor-
Anderson, E. W., & Mansi, S. A. (2009). Does customer satisfaction matter mance. International Journal of Physical Distribution and Logistics Man-
to investors? Findings from the bond market. Journal of Marketing agement, 35(3), 177–194. https://doi.org/10.1108/096000305105
Research, 46(5), 703–714. https://doi.org/10.1509/jmkr.46.5.703 94567
Asvanunt, A., Broadie, M., & Sundaresan, S. (2007). Managing corporate Chava, S., & Purnanandam, A. (2010). CEOs versus CFOs: Incentives and
liquidity: Welfare and pricing implications. In Columbia University corporate policies. Journal of Financial Economics, 97(2), 263–278.
Working Paper. https://doi.org/10.1016/j.jfineco.2010.03.018
Atif, M., Hossain, M., Alam, M. S., & Goergen, M. (2021). Does board gen- Chen, J., Leung, W. S., & Goergen, M. (2017). The impact of board gender
der diversity affect renewable energy consumption? Journal of Corpo- composition on dividend payouts. Journal of Corporate Finance, 43,
rate Finance, 66, 101665. https://doi.org/10.1016/j.jcorpfin.2020. 86–105. https://doi.org/10.1016/j.jcorpfin.2017.01.001
101665 Chen, T., Harford, J., & Lin, C. (2015). Do analysts matter for governance?
Barney, J. (1991). Firm resources and sustained competitive advantage. Evidence from natural experiments. Journal of Financial Economics,
Journal of Management, 17(1), 99–120. https://doi.org/10.1177/ 115(2), 383–410. https://doi.org/10.1016/j.jfineco.2014.10.002
014920639101700108 Cheng, B., Ioannou, I., & Serafeim, G. (2014). Corporate social responsibil-
Beaver, W. H., McNichols, M. F., & Rhie, J. W. (2005). Have financial state- ity and access to finance. Strategic Management Journal, 35(1), 1–23.
ments become less informative? Evidence from the ability of financial https://doi.org/10.1002/smj.2131
ratios to predict bankruptcy. Review of Accounting Studies, 10(1), Chiang, S. M., Chung, H., & Huang, C. M. (2015). A note on board charac-
93–122. https://doi.org/10.1007/s11142-004-6341-9 teristics, ownership structure and default risk in Taiwan. Accounting
Becchetti, L., & Sierra, J. (2003). Bankruptcy risk and productive efficiency and Finance, 55(1), 57–74. https://doi.org/10.1111/acfi.12045
in manufacturing firms. Journal of Banking & Finance, 27(11), Chiu, H. H., & Wagner, E. (2012). CEO compensation and credit default
2099–2120. https://doi.org/10.1016/S0378-4266(02)00319-9 swaps: Evidence from the US and Germany. In 25th Australasian
Bennett, R. L., Güntay, L., & Unal, H. (2015). Inside debt, bank default risk, Finance and Banking Conference.
and performance during the crisis. Journal of Financial Intermediation, Cho, C. H., Laine, M., Roberts, R. W., & Rodrigue, M. (2015). Organized
24(4), 487–513. https://doi.org/10.1016/j.jfi.2014.11.006 hypocrisy, organizational façades, and sustainability reporting.
Berle, A. A. Jr. (1930). Corporate powers as powers in trust. Harvard Law Accounting, Organizations and Society, 40, 78–94. https://doi.org/10.
Review, 44, 1049. 1016/j.aos.2014.12.003
Bharath, S. T., & Shumway, T. (2008). Forecasting default with the Merton Cho, S. Y., Lee, C., & Pfeiffer, R. J. Jr. (2013). Corporate social responsibility
distance to default model. The Review of Financial Studies, 21(3), performance and information asymmetry. Journal of Accounting and
1339–1369. https://doi.org/10.1093/rfs/hhn044 Public Policy, 32(1), 71–83. https://doi.org/10.1016/j.jaccpubpol.2012.
Bhojraj, S., & Sengupta, P. (2003). Effect of corporate governance on bond 10.005
ratings and yields: The role of institutional investors and outside direc- Cormier, D., Ledoux, M. J., & Magnan, M. (2011). The informational contri-
tors. The Journal of Business, 76(3), 455–475. https://doi.org/10.1086/ bution of social and environmental disclosures for investors. Manage-
344114 ment Decision, 49(8), 1276–1304. https://doi.org/10.1108/
Black, B. S., Kim, W., Jang, H., & Park, K.-S. (2015). How corporate gover- 00251741111163124
nance affect firm value? Evidence on a self-dealing channel from a nat- Cormier, D., Ledoux, M. J., Magnan, M., & Aerts, W. (2010). Corporate
ural experiment in Korea. Journal of Banking & Finance, 51, 131–150. governance and information asymmetry between managers and inves-
https://doi.org/10.1016/j.jbankfin.2014.08.020 tors. Corporate Governance: The International Journal of Business in Soci-
Bollen, N. P. (2007). Mutual fund attributes and investor behavior. Journal ety, 10(5), 574–589. https://doi.org/10.1108/14720701011085553
of Financial and Quantitative Analysis, 42(3), 683–708. https://doi.org/ Cornell, B., & Shapiro, A. C. (1987). Corporate stakeholders and corporate
10.1017/S0022109000004142 finance. Financial Management, 16, 5–14. https://doi.org/10.2307/
Boubaker, S., Cellier, A., Manita, R., & Saeede, A. (2020). Does corporate 3665543
social responsibility reduce financial distress risk? Economic Modelling, Correia, M., Kang, J., & Richardson, S. (2018). Asset volatility. Review of
91, 835–851. https://doi.org/10.1016/j.econmod.2020.05.012 Accounting Studies, 23(1), 37–94.
3956 ATIF AND ALI
Das, S. R., Hanouna, P., & Sarin, A. (2009). Accounting-based versus Godfrey, P. C. (2005). The relationship between corporate philanthropy
market-based cross-sectional models of CDS spreads. Journal of Bank- and shareholder wealth: A risk management perspective. Academy of
ing & Finance, 33(4), 719–730. Management Review, 30(4), 777–798. https://doi.org/10.5465/amr.
D'Aveni, R. A., & Ilinitch, A. Y. (1992). Complex patterns of vertical integra- 2005.18378878
tion in the forest products industry: Systematic and bankruptcy risks. Godfrey, P. C., Merrill, C. B., & Hansen, J. M. (2009). The relationship
Academy of Management Journal, 35(3), 596–625. between corporate social responsibility and shareholder value: An
Davis, K., & Blomstrom, R. (1975). Business and society: Environment and empirical test of the risk management hypothesis. Strategic Manage-
responsibility. McGrawHill. ment Journal, 30(4), 425–445. https://doi.org/10.1002/smj.750
DeAngelo, H., DeAngelo, L., & Stulz, R. M. (2006). Dividend policy and the Goyal, V. K., & Wang, W. (2013). Debt maturity and asymmetric informa-
earned/contributed capital mix: A test of the life-cycle theory. Journal tion: Evidence from default risk changes. Journal of Financial and Quan-
of Financial Economics, 81(2), 227–254. https://doi.org/10.1016/j. titative Analysis, 48(3), 789–817. https://doi.org/10.1017/
jfineco.2005.07.005 S0022109013000240
DeAngelo, H., DeAngelo, L., & Stulz, R. M. (2010). Seasoned equity offer- Grullon, G., Michaely, R., & Swaminathan, B. (2002). Are dividend changes
ings, market timing, and the corporate lifecycle. Journal of Financial a sign of firm maturity? The Journal of Business, 75(3), 387–424.
Economics, 95(3), 275–295. https://doi.org/10.1016/j.jfineco.2009. https://doi.org/10.1086/339889
11.002 Harjoto, M. A., & Jo, H. (2015). Legal vs. normative CSR: Differential
Deegan, C., & Rankin, M. (1996). Do Australian companies report environ- impact on analyst dispersion, stock return volatility, cost of capital,
mental news objectively? Accounting, Auditing & Accountability Journal, and firm value. Journal of Business Ethics, 128(1), 1–20. https://doi.
9(2), 50–67. https://doi.org/10.1108/09513579610116358 org/10.1007/s10551-014-2082-2
Dichev, I. D. (1998). Is the risk of bankruptcy a systematic risk? The Journal Hasan, M. M., & Habib, A. (2017). Corporate life cycle, organizational
of Finance, 53(3), 1131–1147. https://doi.org/10.1111/0022-1082. financial resources and corporate social responsibility. Journal of Con-
00046 temporary Accounting and Economics, 13(1), 20–36. https://doi.org/10.
Dickinson, V. (2011). Cash flow patterns as a proxy for firm life cycle. The 1016/j.jcae.2017.01.002
Accounting Review, 86(6), 1969–1994. https://doi.org/10.2308/accr- Hasan, M. M., Hossain, M., & Habib, A. (2015). Corporate life cycle and
10130 cost of equity capital. Journal of Contemporary Accounting and Econom-
Dodd, E. M. (1932). For whom are corporate managers trustees. Harvard ics, 11(1), 46–60. https://doi.org/10.1016/j.jcae.2014.12.002
Law Review, 45, 1145. https://doi.org/10.2307/1331697 Hermalin, B. E., & Weisbach, M. S. (2001). Boards of directors as an endog-
Duan, J. C., Kim, B., Kim, W., & Shin, D. (2018). Default probabilities of pri- enously determined institution: A survey of the economic literature
vately held firms. Journal of Banking & Finance, 94, 235–250. https:// (No. w8161). National Bureau of Economic Research.
doi.org/10.1016/j.jbankfin.2018.08.006 Hillegeist, S. A., Keating, E. K., Cram, D. P., & Lundstedt, K. G. (2004).
Duan, J. C., Sun, J., & Wang, T. (2012). Multiperiod corporate default Assessing the probability of bankruptcy. Review of Accounting Stud-
prediction—A forward intensity approach. Journal of Econometrics, ies, 9(1), 5–34. https://doi.org/10.1023/B:RAST.0000013627.
170(1), 191–209. https://doi.org/10.1016/j.jeconom.2012.05.002 90884.b7
Edmans, A.(2011). Does the stock market fully value intangibles? Horvathova, E. (2010). Does environmental performance affect financial
Employee satisfaction and equity prices. Journal of Financial Economics, performance? A meta-analysis. Ecological Economics, 70(1), 52–59.
101(3), 621–640. https://doi.org/10.1016/j.jfineco.2011.03.021 https://doi.org/10.1016/j.ecolecon.2010.04.004
El Ghoul, S., Guedhami, O., Kwok, C. C., & Mishra, D. R. (2011). Does cor- Hsu, P. H., Lee, H. H., Liu, A. Z., & Zhang, Z. (2015). Corporate innovation,
porate social responsibility affect the cost of capital? Journal of Bank- default risk, and bond pricing. Journal of Corporate Finance, 35,
ing & Finance, 35(9), 2388–2406. https://doi.org/10.1016/j.jbankfin. 329–344. https://doi.org/10.1016/j.jcorpfin.2015.09.005
2011.02.007 Jessen, C., & Lando, D. (2015). Robustness of distance-to-default. Journal
Eliwa, Y., Aboud, A., & Saleh, A. (2019). ESG practices and the cost of debt: of Banking & Finance, 50, 493–505. https://doi.org/10.1016/j.jbankfin.
Evidence from EU countries. Critical Perspectives on Accounting, 2014.05.016
102097. Jiraporn, P., Kim, J. C., & Kim, Y. S. (2011). Dividend payouts and cor-
Endrikat, J., Guenther, E., & Hoppe, H. (2014). Making sense of conflicting porate governance quality: An empirical investigation. Financial
empirical findings: A meta-analytic review of the relationship between Review, 46(2), 251–279. https://doi.org/10.1111/j.1540-6288.2011.
corporate environmental and financial performance. European Manage- 00299.x
ment Journal, 32(5), 735–751. https://doi.org/10.1016/j.emj.2013. Kabir, M. N., Dulal Miah, M., Ali, S., & Sharma, P. (2020). Institutional and
12.004 foreign ownership vis-à-vis default risk: Evidence from Japanese firms.
Faff, R., Kwok, W. C., Podolski, E. J., & Wong, G. (2016). Do corporate poli- International Review of Economics & Finance, 69, 469–493. https://doi.
cies follow a life cycle? Journal of Banking & Finance, 69, 95–107. org/10.1016/j.iref.2020.05.020
https://doi.org/10.1016/j.jbankfin.2016.04.009 Kennedy, P. (2003). A guide to econometrics (5th ed.). Blackwell Publishing.
Friedman, M. (1970). A theoretical framework for monetary analysis. Jour- Konstantinidi, T., & Pope, P. F. (2016). Forecasting risk in earnings. Con-
nal of Political Economy, 78(2), 193–238. https://doi.org/10.1086/ temporary Accounting Research, 33(2), 487–525. https://doi.org/10.
259623 1111/1911-3846.12158
Fukui, Y., & Ushijima, T. (2007). Corporate diversification, performance, Lev, B., Petrovits, C., & Radhakrishnan, S. (2010). Is doing good good for
and restructuring in the largest Japanese manufacturers. Journal of the you? how corporate charitable contributions enhance revenue growth.
Japanese and International Economies, 21(3), 303–323. https://doi.org/ Strategic Management Journal, 31, 182–200. https://doi.org/10.1002/
10.1016/j.jjie.2006.06.002 smj.810
Giesecke, K., Longstaff, F. A., Schaefer, S., & Strebulaev, I. (2011). Corpo- Limkriangkrai, M., Koh, S., & Durand, R. B. (2017). Environmental, social,
rate bond default risk: A 150-year perspective. Journal of Financial Eco- and governance (ESG) profiles, stock returns, and financial policy:
nomics, 102(2), 233–250. https://doi.org/10.1016/j.jfineco.2011. Australian evidence. International Review of Finance, 17(3), 461–471.
01.011 https://doi.org/10.1111/irfi.12101
Gippel, J., Smith, T., & Zhu, Y. (2015). Endogeneity in accounting and Liu, Y., Miletkov, M. K., Wei, Z., & Yang, T. (2015). Board independence
finance research: Natural experiments as a state-of-the-art solution. and firm performance in China. Journal of Corporate Finance, 30,
Abacus, 51(2), 143–168. https://doi.org/10.1111/abac.12048 223–244. https://doi.org/10.1016/j.jcorpfin.2014.12.004
ATIF AND ALI 3957
Liu, Y., Wei, Z., & Xie, F. (2014). Do women directors improve firm perfor- Sassen, R., Hinze, A. K., & Hardeck, I. (2016). Impact of ESG factors on firm
mance in China? Journal of Corporate Finance, 28, 169–184. https:// risk in Europe. Journal of Business Economics, 86(8), 867–904. https://
doi.org/10.1016/j.jcorpfin.2013.11.016 doi.org/10.1007/s11573-016-0819-3
Luo, X., & Bhattacharya, C. B. (2006). Corporate social responsibility, cus- Schultz, E. L., Tan, D. T., & Walsh, K. D. (2017). Corporate governance and
tomer satisfaction, and market value. Journal of Marketing, 70(4), 1–18. the probability of default. Accounting and Finance, 57, 235–253.
https://doi.org/10.1509/jmkg.70.4.001 https://doi.org/10.1111/acfi.12147
Luo, X., & Bhattacharya, C. B. (2009). The debate over doing good: Corpo- Stock, J., & Yogo, M. (2005). Testing for weak instruments in linear IV
rate social performance, strategic marketing levers, and firm- regression. In Identification and inference for econometric models: Essays
idiosyncratic risk. Journal of Marketing, 73(6), 198–213. https://doi. in honor of Thomas Rothenberg (pp. 80–108). Cambridge University
org/10.1509/jmkg.73.6.198 Press.
McGuire, J. B., Sundgren, A., & Schneeweis, T. (1988). Corporate social Switzer, L. N., & Wang, J. (2013). Default risk estimation, bank credit risk,
responsibility and firm financial performance. Academy of Management and corporate governance. Financial Markets, Institutions & Instruments,
Journal, 31(4), 854–872. 22(2), 91–112. https://doi.org/10.1111/fmii.12005
Tashman, P., Marano, V., & Kostova, T. (2019). Walking the walk or talking
McWilliams, A., Siegel, D. S., & Wright, P. M. (2006). Corporate social
the talk? Corporate social responsibility decoupling in emerging mar-
Rresponsibility: Strategic implications. Journal of Management Studies,
ket multinationals. Journal of International Business Studies, 50(2),
43(1), 1–18. https://doi.org/10.1111/j.1467-6486.2006.00580.x
153–171. https://doi.org/10.1057/s41267-018-0171-7
Merton, R. C. (1974). On the pricing of corporate debt: The risk structure
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and stra-
of interest rates. The Journal of Finance, 29(2), 449–470.
tegic management. Strategic Management Journal, 18(7), 509–533.
Miao, H., Ramchander, S., Ryan, P., & Wang, T. (2018). Default prediction
https://doi.org/10.1002/(SICI)1097-0266(199708)18:7<509::AID-
models: The role of forward-looking measures of returns and volatility.
SMJ882>3.0.CO;2-Z
Journal of Empirical Finance, 46, 146–162. https://doi.org/10.1016/j.
Vallascas, F., & Hagendorff, J. (2013). The risk sensitivity of capital require-
jempfin.2018.01.001
ments: Evidence from an international sample of large banks. Review of
Miglani, S., Ahmed, K., & Henry, D. (2015). Voluntary corporate gover-
Finance, 17(6), 1947–1988. https://doi.org/10.1093/rof/rfs042
nance structure and financial distress: Evidence from Australia. Journal
Valta, P. (2016). Strategic Default, Debt Structure, and Stock Returns. The
of Contemporary Accounting and Economics, 11(1), 18–30. https://doi.
Journal of Financial and Quantitative Analysis, 51(1), 197–229.
org/10.1016/j.jcae.2014.12.005
Waddock, S. (2008). The development of corporate respon-
Miller, D., & Friesen, P. H. (1984). A longitudinal study of the corporate life
sibility/corporate citizenship. Organization Management Journal, 5(1),
cycle. Management Science, 30(10), 1161–1183. https://doi.org/10.
29–39. https://doi.org/10.1057/omj.2008.5
1287/mnsc.30.10.1161
Waddock, S. A., & Graves, S. B. (1997). The corporate social performance–
Nadaraja, S., Huang, A., Liu, B., & Ali, S. (2020). Does Board Gender Diver-
financial performance link. Strategic Management Journal, 18(4),
sity Reduce Default Risk? A Global Analysis. Academy of Management
303–319. https://doi.org/10.1002/(SICI)1097-0266(199704)18:
Annual Meeting Proceedings. https://doi.org/10.5465/AMBPP.2020.
4<303::AID-SMJ869>3.0.CO;2-G
14718abstract
Yang, T., & Zhao, S. (2014). CEO duality and firm performance: Evidence
Nadarajah, S., Duong, H. N., Ali, S., Liu, B., & Huang, A. (2020). Stock liquid-
from an exogenous shock to the competitive environment. Journal of
ity and default risk around the world. Journal of Financial Markets.
Banking & Finance, 49, 534–552. https://doi.org/10.1016/j.jbankfin.
forthcoming, 100597. https://doi.org/10.1016/j.finmar.2020.100597
2014.04.008
Nakano, M., & Nguyen, P. (2013). Foreign ownership and firm perfor-
mance: Evidence from Japan's electronics industry. Applied Financial
Economics, 23(1), 41–50. https://doi.org/10.1080/09603107.2012.
705425 How to cite this article: Atif, M., & Ali, S. (2021).
Owen, S., & Yawson, A. (2010). Corporate life cycle and M&A activity.
Environmental, social and governance disclosure and default
Journal of Banking & Finance, 34(2), 427–440. https://doi.org/10.
1016/j.jbankfin.2009.08.003 risk. Business Strategy and the Environment, 30(8), 3937–3959.
Plumlee, M., Brown, D., Hayes, R. M., & Marshall, R. S. (2015). Voluntary https://doi.org/10.1002/bse.2850
environmental disclosure quality and firm value: Further evidence.
Journal of Accounting and Public Policy, 34(4), 336–361. https://doi.
org/10.1016/j.jaccpubpol.2015.04.004
Porter, M. E. (2008). Competitive strategy: Techniques for analyzing indus-
tries and competitors. Simon and Schuster.
AP PE NDIX A: SAMPLE SE LEC TI ON
Ramaswamy, V., Ueng, J. C., & Carl, L. (2007). Corporate governance char-
acteristics of growth companies: An empirical study. Academy of
Accounting and Financial Studies, 12(1), 71.
Rego, L. L., Billett, M. T., & Morgan, N. A. (2009). Consumer-based brand
equity and firm risk. Journal of Marketing, 73(6), 47–60. https://doi.
org/10.1509/jmkg.73.6.47 ESG
Rendleman, R. J. Jr. (1978). The effects of default risk on the firm's invest- All firms 9455
ment and financing decisions. Financial Management, 7, 45–53.
Less: Financial firms 2043
https://doi.org/10.2307/3664905
Renneboog, L., Ter Horst, J., & Zhang, C. (2011). Is ethical money finan- Nonfinancial firms 7412
cially smart? Nonfinancial attributes and money flows of socially Less: Missing values in key variables 2206
responsible investment funds. Journal of Financial Intermediation, 20(4),
Final observations 5206
562–588. https://doi.org/10.1016/j.jfi.2010.12.003
3958 ATIF AND ALI
APPENDIX B : VARIABLE DEF INITIONS
Notation Name Measure Source

Dependent variables
DD Distance to default Annual average of distance to default based on stock RMI-NUS
based on stock price variability
CDS Credit default swap spread Credit derivatives that allow the transfer of the firm's RMI-NUS
default risk between two agents for a predetermined
time period
Key independent variables
ESG Environmental, social and governance ESG is the aggregate environmental, social and Bloomberg
governance performance index.
E factor Environment The environmental factor refers to the firm's use of Bloomberg
energy, water, carbon emissions, pollution and
natural resources conservation. It includes the
appraisal of environmental risks that might affect the
firm's income and how the firm is managing those
potential risks.
S factor Social The social factor, which refers to the firm's relationship Bloomberg
with society, includes employee health, safety and
rights at the workplace, product safety and social
philanthropy.
G factor Governance The governance factor includes board independence, Bloomberg
diversity, shareholder rights and disclosure of
information. It evaluates the information required by
investors and stakeholders on accounting
transparency, conflict of interest and strength of
governance structure.
Control variables
SIZE Firm size Natural logarithm of total assets Bloomberg
ROA Firm profitability1 Net income divided by total assets Bloomberg
TLTA Firm leverage Total liabilities divided by total assets Bloomberg
LIQ Firm liquidity Current assets divided by current liabilities Bloomberg
TANG Asset tangibility Plant property and equipment divided by total assets Bloomberg
SALES Firm sales Total sales divided by total assets Bloomberg
TOBINQ Growth opportunities Fair market value and total liabilities divided by total Bloomberg
assets
Additional variables
ROE Firm profitability2 Net income divided by common shareholder equity Bloomberg
IQR Inter quantile range The 0.5 interquartile range calculated as Qi75 Qi25 Bloomberg
SKEW Skewness Skewness calculated as [(Qi75 Qi50) (Qi50 Qi25)]/IQRi Bloomberg
KURT Kurtosis Kurtosis calculated as [(Qi87.5 Qi62.5) Bloomberg
+(Qi37.5 Qi12.5)]/IQRi
UP Up The upside component of KURT, calculated as Bloomberg
[(Qi87.5 Qi62.5)]/IQRi
DOWN Down The downside component of KURT, calculated as Bloomberg
[(Qi37.5 Qi12.5)]/IQRi
COD Cost of debt Interest expense divided by average interest-bearing Bloomberg
debt
CLC Corporate life cycle A vector of dummy variables that capture firms' Bloomberg
different stages in the life cycle following the
Dickinson (2011) model
ATIF AND ALI 3959
Notation Name Measure Source

RETA Retained earnings Retained earnings divided by total assets. A high RETA Bloomberg
implies that the firm is more mature or old with
declining investment, while the firm with a low RETA
tends to be young and growing (DeAngelo
et al., 2006).
AGE Firm age Natural logarithm of the current year minus the year of Bloomberg
incorporation
ESG (IND) Industry-average ESG The average ESG of all the firms in firm i's industry
excluding firm i's ESG

Bus Strat Env - 2021 - Atif - Environmental Social and Governance Disclosure and Default Risk

Uploaded by

Copyright:

Available Formats

Bus Strat Env - 2021 - Atif - Environmental Social and Governance Disclosure and Default Risk

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bus Strat Env - 2021 - Atif - Environmental Social and Governance Disclosure and Default Risk

Uploaded by

Copyright:

Available Formats

Received: 5 May 2020 Revised: 19 May 2021 Accepted: 2 June 2021

Environmental, social and governance disclosure and

Muhammad Atif1 | Searat Ali2

1 | I N T RO DU CT I O N Corporate default, generally defined as a firm that is failing to make

TABLE 1 Descriptive statistics

Whole sample Low ESG High ESG

TABLE 2 Correlation analysis

TABLE 3 ESG and default risk—baseline results

DD DD DD CDS CDS CDS

SIZE 0.316*** (7.11) 0.409*** (9.22) 0.128** (1.98) 0.115* (1.79)

ROA 5.670*** (13.80) 5.262*** (12.10) 2.678*** (6.04) 2.733*** (5.88)

TLTA 4.363*** (13.54) 4.834*** (14.60) 1.825*** (5.01) 2.144*** (6.09)

TANG 0.159 (0.54) 0.082 (0.25) 0.700** (2.00) 0.959** (2.53)

SALES 0.039 (0.38) 0.049 (0.47) 0.256** (2.36) 0.333*** (2.61)

TOBINQ 0.796*** (13.32) 0.706*** (12.15) 0.272*** (5.01) 0.262*** (4.82)

Year effects No No Yes No No Yes

Industry No No Yes No No Yes

SE robust Yes Yes Yes Yes Yes Yes

N 5234 5206 5206 5234 5206 5206

Adj. R2 .065 .355 .513 .087 .188 .249

TABLE 4 ESG and default risk—endogeneity bias (lagged independent variables)

T A B L E 5 ESG and default risk—

TABLE 6 ESG and default risk—endogeneity bias (propensity score matching)

Panel A: Pre-match regression and post-match diagnostic test

Panel B: Mean differences

Treated Control Difference t test

Panel C: PSM estimator

Variables Treated Controls Difference t statistics

Panel D: Multivariate regression

TABLE 7 ESG and default risk—role of profitability

TABLE 8 ESG and default risk—role of performance variability

IQR SKEW KURT UP DOWN

T A B L E 9 ESG and default risk—role

COD 0.249*** (7.10) 0.170*** (4.79)

Table 11 reports the results between the components of ESG disclo-

TABLE 11 ESG components and default risk

DD DD DD CDS CDS CDS

G factor 0.042*** (3.82) 0.053*** (3.99)

E factor 0.016*** (2.79) 0.020*** (2.61)

S factor 0.015*** (2.62) 0.024*** (3.29)

Year Yes Yes Yes Yes Yes Yes

N 5206 2034 3429 5206 2036 3428

Adj. R2 .512 .608 .536 .240 .362 .250

TABLE 12 Alternative measure of ESG—Asset4

DD DD DD DD CDS CDS CDS CDS

ESG 0.007* (1.82) 0.013*** (3.77)

G factor 0.003* (1.85) 0.006** (1.97)

E factor 0.004*** (2.78) 0.007*** (2.63)

S factor 0.005*** (3.04) 0.011*** (3.26)

SE robust Yes Yes Yes Yes Yes Yes Yes Yes

N 4714 4714 4714 4712 4725 4725 4725 4723

Adj. R2 .553 .552 .552 .553 .338 .331 .333 .337

APPENDIX B : VARIABLE DEF INITIONS

Notation Name Measure Source

Notation Name Measure Source

You might also like

SIZE 0.316* (7.11) 0.409* (9.22) 0.128** (1.98) 0.115* (1.79)

ROA 5.670* (13.80) 5.262* (12.10) 2.678* (6.04) 2.733* (5.88)

TLTA 4.363* (13.54) 4.834* (14.60) 1.825* (5.01) 2.144* (6.09)

TANG 0.159 (0.54) 0.082 (0.25) 0.700 (2.00) 0.959 (2.53)

SALES 0.039 (0.38) 0.049 (0.47) 0.256 (2.36) 0.333* (2.61)

TOBINQ 0.796* (13.32) 0.706* (12.15) 0.272* (5.01) 0.262* (4.82)

COD 0.249* (7.10) 0.170* (4.79)

G factor 0.042* (3.82) 0.053* (3.99)

E factor 0.016* (2.79) 0.020* (2.61)

S factor 0.015* (2.62) 0.024* (3.29)

E factor 0.004* (2.78) 0.007* (2.63)

S factor 0.005* (3.04) 0.011* (3.26)