Received: 6 July 2018 Revised: 22 September 2018 Accepted: 9 October 2018

DOI: 10.1002/csr.1699

RESEARCH ARTICLE

Sustainable development and corporate governance in the

financial system: Are environmentally friendly banks less risky?
Francesco Gangi1 | Antonio Meles1 | Eugenio D'Angelo2 | Lucia Michela Daniele1

1
Department of Economics, Università degli
Studi della Campania Luigi Vanvitelli, Capua, Abstract
Italy This paper responds to the need for a deeper empirical investigation of the impact of
2
Università Telematica Pegaso, Naples, Italy
corporate social responsibility pillars on the financial performance of banks. To
Correspondence
Francesco Gangi, Associate Professor of
address this question, this study first analyzes the factors that encourage banks to
Management, Department of Economics, be more environmentally friendly and then investigates the relationship between a
Università degli Studi della Campania Luigi
Vanvitelli, Corso Gran Priorato di Malta, Capua
bank's environmental engagement and its risk. Using a sample of 142 banks from
81043, Italy. 35 countries covering the period from 2011 to 2015, we document the positive
Email: francesco.gangi@unicampania.it
impact of effective corporate governance mechanisms on banks' environmental
engagement. Moreover, by using the Heckman's two‐stage model for the treatment
of sample selection bias, we find that banks that are more sensitive to environmental
issues also exhibit less risk. Stakeholder theory and the conflict resolution hypothesis
are useful frameworks to overcome the trade‐off between economy and ecology in
the banking industry.

KEY W ORDS

banks, corporate governance, corporate social responsibility, environment, risk, sustainable

development

1 | I N T RO D U CT I O N In this paper, we argue that banks may assume environmental

responsibility not only to generate spillover benefits to the worldwide
In 2011, the report entitled “Bankrolling Climate Change” (Schücking, community but also to achieve their own strategic goals and address
Kroll, Louvel, & Richter, 2011) has pointed out that 20 of the world's the pressures of increasing regulatory requirements. As recently
largest commercial banks are among the actors that are the most reported by the Financial Times (Hook, 2018), European financial
responsible for the coal industry's emissions, because they provide supervisors, seeking to understand the exposure of banks to climate
74% of the industry's total funding. This evidence clearly controverts risks, have been increasingly scrutinizing the banking sector and are
the idea that banking activity is not environmentally relevant and pro- considering “increasing regulatory oversight to address climate‐related
vides new insights into the pivotal role that banks can play in mitigating risks to the financial system, including carbon stress tests for banks.”
climate change by breaking up with highly polluting industries. Stated In addition, analysts from some of the world's largest banks, such as
differently, although a bank's direct involvement in environmentally Citibank, Deutsche Bank, and Hongkong and Shanghai Banking
sustainable practices (such as adopting energy efficiency initiatives) Corporation, are questioning the financial justification for further
has been seen as a way to produce small benefits for the environment investments in coal (Schücking, 2013).
(e.g., Esteban‐Sanchez, de la Cuesta‐Gonzalez, & Paredes‐Gazquez, Although it is indisputable that bank supervisory authorities are
2017; Matute‐Vallejo, Bravo, & Pina, 2011), when banks stimulate an mainly concerned with the possibility that an abrupt climate change
environmentally friendly approach among their borrowers, they are may deteriorate banks' lending portfolios and create a systemic risk
able to achieve a type of “multiplier effect.” Obviously, although this to financial stability, the arguments discussed above contribute to
argument characterizes the banking system as a strategic element in shedding light on several additional questions that are of special inter-
achieving environmental sustainability, it raises a key question: Why est to both regulators and banks' stakeholders and that have been
do banks decide to act in an environmentally responsible manner? poorly investigated. What are the main characteristics of banks that

Corp Soc Resp Env Ma. 2018;1–19. wileyonlinelibrary.com/journal/csr © 2018 John Wiley & Sons, Ltd and ERP Environment 1
2 GANGI ET AL.

engage in environmentally friendly actions? More specifically, do economic system. Furthermore, we posit three main arguments that
banks with certain corporate governance (CG) characteristics exhibit corroborate the peculiarity of the relationship between a positive atti-
strong environmental responsibility? Finally, with regard to bank risk, tude towards the environment and bank risk.
what is the impact of engaging in environmentally friendly initiatives? First, if a bank incorporates environmental sustainability into its
We answer these questions by investigating a sample of 142 banks lending policy, then it will be less exposed to information risk and will
from 35 countries covering the period from 2011 to 2015. We report be able to overcome both adverse selection and moral hazard prob-
two main results. First, we find that banks with effective CG mecha- lems (e.g., Goss & Roberts, 2011), which are among the main causes
nisms are more likely to be engaged in environmentally friendly actions. of banks' nonperforming loans. To give a sense of risk perception
Second, while controlling for selection bias issues, we find that both about environmental items by banking industry, Hongkong and Shang-
corporate social responsibility (CSR) engagement and CG mechanisms hai Banking Corporation estimates that 40–60% of the market capital-
are negatively associated with banks' risk, as measured by the Z_score ization of major European oil firms could be at risk in a low emissions
(e.g., Fiordelisi & Mare, 2014; Goetz, Laeven, & Levine, 2016). scenario. Likewise, big players such as Citi Bank and Deutsche Bank
Our paper complements previous research in the field of sustain- are exploring the implications for the evaluation of companies that rely
able development by shedding new light on the relationship between on revenues from exploiting fossil fuel reserves (United Nations Envi-
CSR engagement and financial performance in the banking industry. ronment Programme Finance Initiative [UNEP FI], 2013).
There are at least two reasons that a new understanding in this Second, banks that are more environmentally friendly should also
research area is of special interest. First, previous studies have mainly reach a higher level of operational efficiency (e.g., Clarkson, Li, Richardson,
focused on nonfinancial companies (e.g., Barnett & Salomon, 2012; & Vasvari, 2011; Clarkson, Li, Pinnuck, & Richardson, 2015), which in
Margolis & Walsh, 2003; Santis, Albuquerque, & Lizarelli, 2016). How- turn may contribute to a reduction in bank risk, as documented by the
ever, as successful CSR strategies vary across industries (Dahlsrud, banking literature (e.g., Fiordelisi, Marques‐Ibanez, & Molyneux, 2011).
2008; Van Marrewijk, 2003) and each sector has “industry‐specific Third, by implementing environmentally friendly actions, banks
stakeholder pressures” (Melo & Garrido‐Morgado, 2012, p.18), the may improve both their reputation and the loyalty of their customers
results of these studies cannot be generalized and extended to all (Aramburu & Pescador, 2017; Gatzert, 2015; Ruiz, García, & Revilla,
industries. Second, although banks must manage sustainability risks 2016), which in turn should translate into lower bank funding costs
and opportunities similarly to firms operating in different sectors, their and higher funding stability (e.g., Bassen, Meyer, & Schlange, 2006;
involvement in CSR activities has a potential impact on the sustainabil- El Ghoul, Guedhami, Kwok, & Mishra, 2011). Furthermore, a greater
ity of other industries through the lending channel (e.g., Scholtens, sensitivity to environmental issues prevents the creation of a negative
2009; Weber, 2012). reputation of being connected with debtors that cause environmental
Therefore, it is not surprising that a growing number of studies damage (Weber, 2012) and can mitigate the reputational loss that the
have empirically addressed how a bank's CSR activism influences its recent financial crisis has caused the banking system as a whole
outcomes, primarily in terms of profitability, growth opportunity, (Venturelli, Cosma, & Leopizzi, 2018).
reporting quality, and customer satisfaction (e.g., Cornett, Erhemjamts, The remainder of this paper is organized as follows. Section 2 pre-
& Tehranian, 2016; Forcadell & Aracil, 2017; García‐Sánchez & García‐ sents the literature review and the hypotheses. Section 3 explains the
Meca, 2017; Shen, Wu, Chen, & Fang, 2016; Soana, 2011; Wu, Shen, design of the empirical study. Section 4 summarizes the results, and
& Chen, 2017). However, although these studies have widely docu- Section 5 presents the discussion. Section 6 concludes.
mented that CSR actions significantly influence various dimensions
of a bank's business, only a few papers (e.g., Aras, Tezcan, & Kutlu
Furtuna, 2018; Esteban‐Sanchez et al., 2017) have responded to the 2 | THEORETICAL BACKGROUND AND
call by Wu and Shen (2013) for deeper empirical investigations of HYPOTHESIS
the effects generated by CSR subcomponents, with very limited evi-
Given the aim of the current study, this section focuses on two rela-
dence of the relationship between the environmental engagement
tionships (Figure 1). First, we analyze the link between CG and envi-
and bank financial performance (viz., Finger, Gavious, & Manos,
ronmental engagement as pillars of CSR. Second, compared with
2018). We attempt to contribute to this debate motivated by the idea
prior studies, we deepen our analysis of the impact of environmental
that each individual CSR dimension has its own identity (Cai, Cui, & Jo,
engagement on bank risk. Both relationships are topics to which the
2016; Oikonomou, Brooks, & Pavelin, 2012). When CSR strategies are
literature has dedicated limited attention with reference to the bank-
broken down into qualitative areas (e.g., community relations,
ing industry.
employee relation, and environmental issues), each of the CSR dimen-
sions will have a different impact on performance (e.g., Melo &
Garrido‐Morgado, 2012). 2.1 | The relationship between CG and
More specifically, as far as can be ascertained, our paper provides environmental engagement
the first evidence on whether a bank's environmental activism may
influence its risk taking. Understanding this issue is of special interest Both CG and CSR have been subject to evolutionary trends, and aca-
because the existing literature (e.g., Bernanke, 1983; Calomiris & demic interest in their relationship has been increasing (Jamali,
Mason, 2003) and the 2007–2009 financial crises have highlighted Safieddine, & Rabbath, 2008). The classic approach depicts CG as
that bank risk taking may compromise the overall stability of the the set of mechanisms that align the interests of managers and owners
GANGI ET AL. 3

perspective (Lashgari, 2004), CG is a set of mechanisms to guarantee

responsible behavior towards a wider network of stakeholders. This
network includes but is not limited to shareholders and financial inves-
tors in general. CG is expected to maintain good relationships with
multiple interest groups that interact with the firm independent of
the existence of formal contracts (Kendall, 1999). Callan and Thomas
(2014) argue that in such a model, the CEO is compensated not only
FIGURE 1 Conceptual and empirical framework for improving the firm's financial position but also for meeting goals
such as lower levels of pollution releases, which help define CSR.
(Shleifer & Vishny, 1997). The premise is the separation between own- Thus, from the broader perspective of CG, the safeguarding of
ership and control (Berle & Means, 1932) and the consequent risk of stakeholders that are not equipped with formal power over managers
opportunistic behavior by insiders (managers) towards outsiders is fundamental. Businesses must be responsible and responsive
(owners). This conflict of interest is the focus of the agency model towards a wide spectrum of stakeholders even while they attend to
(Friedman, 1970; Jensen & Meckling, 1976). Building on the seminal the interests of shareholders. Trust relationships with internal and
position of Friedman (1970),1 studies identify an agency problem in external stakeholders and the community in general can be a
managers' decisions to involve the firm in CSR initiatives (e.g., Barnea strength. As Sacconi (2000) argued, effective CG solutions may help
& Rubin, 2010; Jensen, 2001). More specifically, the use of the a business deal with CSR by avoiding the risk that a manager will
company's resources for socially responsible actions is seen as detri- abuse his authority towards stakeholders that are not linked to the
mental to shareholders. Therefore, CG should either limit CSR efforts business through formal contracts. Caring about the broader require-
or prevent managers from overinvesting in CSR initiatives. ments of company's policy rather than just the rights of shareholders
The abovementioned perspective corresponds to a narrow vision is consistent with a vision of CG as a pillar of CSR (Elkington, 2006).
of CG. However, over the years, alternative positions have emerged. Renneboog, ter Horst, and Zhang (2008, p. 1723) argue that with the
In contrast to the controversial separation thesis of the Friedman concept of CSR, we can refer to “corporate decisions fostering social,
scholars (Harris & Freeman, 2008), CG cannot be limited to resolving CG, ethical and environmental issues.” In this sense, the effectiveness
the trade‐off between CSR and shareholders' interests. For a corpora- of CG mechanisms is a condition to guarantee responsible behavior
tion, being socially responsible might also mean fulfilling its economic towards all stakeholders (Huse, 2009). This perspective highlights
function (Montiel, 2008). Consistent with this perspective, Carvalho, the positive impact that CG may have on CSR. The aim of mitigating
Jonker, and Dentchev (2014) argue that the separation between the potential conflicts of interest goes beyond the relationship between
economic and the noneconomic dimensions of social responsibility is managers and ownership. Effective mechanisms of CG are useful in
artificial. Indeed, an organization can be defined as socially responsible implementing and sustaining effective CSR (Jamali et al., 2008).
to the extent that it exceeds the compulsory requirements dictated by The increasing diffusion and activism of investors who are sensi-
law provisions, thus integrating social, environmental, ethical, human tive to social and environmental issues further confirm and justify
rights, and consumer concerns into its strategies and operations (Crifo, the need for CG to support the company's involvement in socially
Escrig‐Olmedo, & Mottis, 2018). Sen, Bhattacharya, and Korschun responsible programs. Dimson, Karakaş, and Li (2015) argue that in
(2006) define CSR as the set of activities that firms implement to fulfill demonstrating their environmental, social, and governance (ESG) con-
their obligations to society. cerns, shareholders are increasingly exercising their right to influence
From a triple bottom line perspective, CSR includes the economic, the management of the business. Unlike traditional shareholder activ-
social, and environmental dimensions of corporate responsibility ism, effective CG must be responsive to ESG‐oriented activism that
(Bansal, 2005). A company that complies with the pillars of CSR has focuses on the interests of a broader range of stakeholders. Institu-
greater legitimacy than one that does not comply, and the former is tional investors,
therefore theoretically equipped with greater social capital (Adler &
Kwon, 2002). This latter consists of “connections among individuals Out of the public view on matters such as corporate
—social networks and the norms of reciprocity and trustworthiness strategy, board effectiveness, executive remuneration
that arise from them … that can improve the efficiency of society by and CEO succession …, act as an “early
facilitating co‐ordinated actions” (Putnam, 1993, p. 167). In socially warning system” on important strategic and governance
responsible banks, social capital sustains the construction of valuable issues, and, increasingly, are evaluating social and
goodwill (Fombrun, Gardberg, & Barnett, 2000; Stubbs & Cocklin, environmental risks facing the company. (Aguilera,
2008). Williams, Conley, & Rupp, 2006, p. 151)
If social responsibility cannot be separated from the economic
function of business, its relationship with CG assumes a new configu- Therefore, from a broader perspective, effective CG may be a
ration relative to agency conflicts. Consistent with a communitarian lever that influences a bank to act responsibly towards the multiple
stakeholders interacting with the bank's operations. Sensitivity to
1
environmental safeguards is common to all sectors independent of
Friedman (1970) argued that managers have to make as much money as possi-
ble while conforming to the basic rules of society embodied in the law and their pollution footprint. The achievement of legitimacy to survivor-
ethical customs. The unique social responsibility of a firm is its profit. ship, especially during the most turbulent phases of the financial
4 GANGI ET AL.

markets, is an incentive for banks' managers to protect the interests of themes into their strategies (Liu, Miletkov, Wei, & Yang, 2015; Porter
shareholders without limiting respect for social and environmental & van der Linde, 1995; Yang, Wang, Zhou, & Jiang, 2018). For a long
duties (Jo & Harjoto, 2012). From the conflict resolution approach time, the banking system has been considered marginal to this mana-
(Baron, 2009; Calton & Payne, 2003; Jo, Song, & Tsang, 2015), effec- gerial evolution. The frequent justification for this marginal status is
tive governance mechanisms enhance the status of environmental that the banking industry is not a polluting sector (Esteban‐Sanchez
engagement as a pillar of CSR. This creates a more systemic vision et al., 2017; Matute‐Vallejo et al., 2011; Thompson & Cowton,
of CSR components and their interactions. As argued by Walls, 2004). This narrow view helps to explain the scant analyses of banking
Berrone, and Phan (2012, p.885), “since the turn of the twenty‐first relative to other industries (Venturelli et al., 2018). However, aware-
century, CG discussions have shifted progressively toward contempo- ness of environmental issues is increasing, even among financial oper-
rary social issues (e.g., climate change, labour rights, and corruption) ators. In the field of socially responsible investment (SRI), for example,
that matter to lawmakers, consumers, shareholders, and corporate the manner in which firms engage with the protection of the natural
managers in the marketplace.”2 Thus, CG represents an additional tool environment is one of the most‐used screening criteria.4 In this sce-
to address the environmental and business pressures originating from nario, banks play an important role. On the one hand, they can be
diverse stakeholders. directly involved in projects that protect the environment or can pro-
Therefore, for the abovementioned reasons, we posit the mote socially responsible products. On the other hand, they can orient
following: funds according to the environmental risk of the target companies.
Consistent with this perspective, Scholtens (2009) asserted that
H1. CG mechanisms positively affect a bank's envi-
socially responsible banking has become a well‐established concept.
ronmental engagement.
Similarly, Soana (2011) argued that the banking sector was among
the first industries to implement CSR. Nevertheless, the adoption of
environmentally friendly approaches by banks raises similar questions
2.2 | The relationship between environmental
about the financial effects of SRI (e.g., Gangi & Varrone, 2018;
engagement and bank risk Renneboog et al., 2008). In the case of banks, the relationship
between environmental engagement and financial performance
Over the last few decades, protection of the natural environment has
deserves an analysis from at least three main perspectives: the finan-
become increasingly important to modern society. The 1995 Kyoto
cial benefits of financing environmentally friendly borrowers; the effi-
Protocol and, more recently, the Paris Conference (2015) confirm
cient use of resources within the bank as an organization; and the
the importance ascribed to this theme worldwide. Giannarakis,
lowering of reputational risk. As explained below, all these perspec-
Zafeiriou, Arabatzis, and Partalidou (2018) argue that climate change
tives may contribute to our understanding of why environmentally
disclosure is an effective tool to limit information asymmetry for
friendly approaches should lead to lower bank risk.
shareholders and stakeholders. Prior literature recognizes environ-
First, banks that incorporate environmental sustainability into
mental protection as a pillar of CSR (Renneboog et al., 2008). Eco-
their lending policy should be able to select less risky and more prof-
nomic growth can be considered sustainable if it preserves natural
itable borrower firms and to alleviate the lender–borrower information
resources for future generations (Halkos & Paizanos, 2016).3 Hence,
gap. Consistent with the natural‐resource‐based view (Hart & Dowell,
CSR and corporate sustainability have been considered converging
2011; Sharma, Aragon‐Correa, & Rueda‐Manzanares, 2007), the so‐
issues because “they share the same vision which intends to balance
called “green strategies” help firms to overcome the constraints
economic responsibilities with social and environmental ones”
imposed by the natural environment and generate a sustainable,
(Montiel, 2008). From this perspective, environmental engagement is
nonreplicable, or less‐imitable competitive advantage (Elijido‐Ten &
a subset of either CSR or corporate sustainability, although there is
Clarkson, 2017). Furthermore, several studies suggest that environ-
an increasing emphasis on corporate environmental responsibility
mentally friendly firms are less exposed to various risks that may cause
(Cai et al., 2016).
economic and financial difficulties. For instance, the introduction of
The progressive recognition of the economic and social relevance
stricter environmental regulation and/or the simple changes in con-
of environmental protection has exerted pressure on firms from both a
sumer preferences can weaken a firm's sales, causing, among other
regulatory and a managerial point of view. Companies have reacted by
things, the market withdrawal of one or more of its products (Thomp-
paying more attention to the environment (Albertini, 2018) to be com-
son & Cowton, 2004). At the same time, Mason (2012) argues that
pliant with higher societal expectations (Lee, Kim, & Kim, 2018). The
green customers are willing to pay a premium price for environmen-
managerial approach can be reactive (Delmas & Montes‐Sancho,
tally products. Hence, banks that finance the firms that make such
2010; Kassinis & Vafeas, 2006) if firms act to avoid sanctions, and it
products may indirectly benefit from this green premium. In addition,
can be proactive if companies consciously integrate environmental
a firm may be forced to stop its activity because of an inability to bear
2
“Many corporations on both sides of the Atlantic have also begun to include the costs of complying with increasingly onerous environmental regu-
social responsibility criteria in executive compensation” (Walls et al., 2012, p. lations. Finally, less environmentally friendly firms are more likely to be
886).
3 4
Halkos and Paizanos (2016) argue that ultimately, environmental macroeco- In the European SRI market, the first market for volumes of SRI (Global Sustain-
nomics contemplates the key role of physical constraints in economic growth able Investment Review, 2016), investment strategy based on environmental
and examines how decision makers can implement appropriate policies to pro- impact represents one of the most relevant approaches to investing, having
mote sustainable economic growth within these limits. grown 146% between 2013 and 2015 (Eurosif, 2016).
GANGI ET AL. 5

sanctioned by regulators or boycotted by other stakeholders (Trapp, industry (Dell'Atti, Trotta, Iannuzzi, & Demaria, 2017). According to
2014). Banks that include environmental considerations as part of this idea, Pérez and Rodríguez del Bosque (2015) find that only a rel-
the credit appraisal process should be better able to mitigate the infor- atively small portion of the banking industry's customers is
mation asymmetry that is among the main causes of banks' nonper- nonsupportive of banks' CSR engagement. Matute‐Vallejo et al.
forming loans. Indeed, as determined by various studies (Kim, Park, (2011) find that consumers in the banking industry do not perceive a
& Wier, 2012; Lopatta, Buchholz, & Kaspereit, 2016), a higher level bank's CSR engagement as an attempt to instrumentalize social issues
of firm environmental activism is associated with higher earnings qual- in a manner ultimately intended to increase prices.
ity, a greater degree of transparency, and higher moral standards. Therefore, in light of the above, we posit the following:
These factors help to mitigate the adverse selection caused by an
H2. Bank risk is inversely related to the bank's level of
ex‐ante information gap (i.e., hidden information) and the moral hazard
environmental engagement.
arising from an ex‐post information gap (i.e., hidden actions).
Second, banks are organizations that combine technical and human
resources. Therefore, similar to firms operating in other industries, they 3 | M E TH OD O LO GY
can invest in reducing energy consumption and lowering the heat dis-
persion of their buildings, and they can adopt procedures for recycling 3.1 | Data collection
their production factors. Today, banks are directly involved both in envi-
ronmental protection policies and in ecological initiatives at the institu- To test the prediction that environmental activism is valuable for
tional level.5 Moreover, most banks have adopted internal governance reducing a bank's risk, we start by collecting data on environmental
tools for climate issues and provided some disclosure on environmen- engagement from Thomson Reuters' ASSET4, which has been widely
tal themes such as low‐carbon products and services (Boston used in previous CSR studies (e.g., Cheng, Ioannou, & Serafeim,
Common Asset Management, 2018). A modern approach to bank risk 2014; El Ghoul, Guedhami, & Kim, 2017). Starting from the entire list
management reflects both the internal behavior of the bank and the of listed banks provided by Thomson Reuters' ASSET4, we exclude
behavior of borrower firms. A bank that cares only for the environ- banks for which we do not obtain information about environmental
mental risk of customers without addressing its own environmental engagement for one or more fiscal years between 2011 and 2016.
responsibility adopts a partial approach similar to window dressing This procedure yields a sample of 142 banks from 35 countries cover-
(Tang, Hull, & Rothenberg, 2012) that according to the prior literature ing the period from 2011 to 2015 (for a total of 710 firm‐year obser-
(Wu & Shen, 2013) has no effect on financial performance. Social vations). We then match the ASSET4 data with the Thomson Reuters
reports in the banking industry give space to direct and indirect “Datastream” to obtain financial and CG data and with World Bank
CO2 emissions (abstracts of qualitative environmental disclosures data to obtain data relative to concentration in banking systems and
are reported in Table A2). This is an important communication channel gross domestic product by country. However, because of missing data,
to strengthen trust relationships with stakeholders interested in the especially for several governance variables (such as board gender
environmental performance of organizations. Of course, banks invest diversity) and certain control variables (such as concentration of bank-
in reducing the pollution generated by their activity not only to pro- ing systems), the final sample used in our regressions changes across
tect the environment but also to achieve benefits for themselves. models, keeping more than 655 firm‐year observations.
Indeed, environmental protection capabilities can be valuable assets The remainder of this section first describes the main variables of
that contribute directly to the recovery of operational efficiency that interest in our analysis—environmental engagement (ENV) and bank
in turn may boost financial performance, as widely demonstrated in risk—and then describes the other variables we include in the esti-
the previous literature (e.g., Clarkson et al., 2011; Jung, Herbohn, & mates. A comprehensive set of variables is considered to control for
Clarkson, 2016). In the banking industry, then, a negative relation the effect of other factors on bank risk, such as bank‐specific charac-
between bank efficiency and bank risk has also been identified teristics, market concentration, and macroeconomic‐specific variation.
(e.g., Fiordelisi et al., 2011). Definitions of variables are also summarized in Table A1, whereas
Third, banks that are deeply involved in environmental actions can Table 1 provides the sample distribution by country, Table 2 reports
also receive benefits in terms of reputational capital (McWilliams & the Pearson's pair‐wise correlation matrix among independent vari-
Siegel, 2011). This argument is of special interest for our study ables, and Table 3 presents descriptive statistics.
because the ability of banks to fund their lending activity with suffi-
ciently stable sources, which in turn influences bank risk, highly
3.2 | Variables definition and operationalization
depends on their reputation. This statement is even more true
because the bursting of the subprime mortgage bubble has increased 3.2.1 | Bank risk
customer scepticism towards banks. Therefore, getting involved in
environmentally responsible practices has also been considered a Following the previous studies (e.g., Goetz et al., 2016; Laeven &
way to obtain social legitimacy and improve the image of the banking Levine, 2009), we proxy a bank risk using the Z‐score, which is calcu-
lated as follows:
5
More than 200 financial institutions are signatories to the UNEP FI statement,
a global partnership between UNEP and the financial sector to develop and pro- ðROA þ CARÞ
Z − score ¼
mote linkages between sustainability and financial performance. σðROAÞ
6 GANGI ET AL.

where ROA measures the return on total assets; CAR is the capital sources, such as annual reports and CSR reports6 that provide data
assets ratio (equity on total assets) and σ (ROA) is the standard devia- and information on both direct and indirect bank practices related to
tion of ROA. environmental protection. According to the Greenhouse Gas Protocol,
As highlighted by Roy (1952), the Z‐score can be interpreted as the most widely used international accounting tool to assess green-
the number of standard deviations that a bank's ROA has to fall below house gas emissions (UNEP FI Investor Briefing, 2013), the CSR
its expected value before equity is depleted. Thus, it is strongly related reports disclose information on emissions occurring from sources that
to the inverse of the probability of insolvency. Therefore, banks that are controlled by the bank and the emissions that are a consequence
show a higher Z‐score value are less risky and more stable relative of the activity of the bank but occur from sources not owned or con-
to banks that present a lower Z‐score value. trolled by the organization. The social reports of banks disclose infor-
mation on their environmental engagement and CO2 equivalent
3.2.2 | Environmental engagement indicator emissions related to in‐house actions (e.g., power purchases from
renewable sources, optimization and energy efficiency, and the
In assessing how a bank's Z‐score varies with its environmental responsible management of environmental and energy impacts) and
engagement, we use the score provided by Thomson Reuters' ASSET4 products aimed at environmental protection, in addition to the assess-
(ENV), which “measures a company's impact on living and non‐living ment of customers' environmental risks.
natural systems, including the air, land and water, as well as complete
ecosystems.” It reflects “how well a company uses best management
3.2.3 | CG variables
practices to avoid environmental risks and capitalize on environmental
opportunities in order to generate long term shareholder value”
Prior literature has shown that CG may influence both the attitude
(ASSET4 ESG Data Glossary, 2018). Hence, through ASSET4, we can
towards compliance with CSR and the items included under the
capture both the operational practices and the financing practices of
umbrella of CSR (e.g., Jamali et al., 2008; Jo & Harjoto, 2012). Consis-
banks. In more detail, the ENV score, which weights for the 33% of
tent with this perspective, we assume that CG positively impacts the
the overall ESG score, synthetizes and normalizes 70 variables related
bank's sensitivity to environmental topics. To account for this relation-
to three indicator scores of the environmental engagement, such as
ship, we adopt variables that are generally recognized as proxies of the
emissions reduction, product innovation, and resource reduction.
effectiveness of CG mechanisms. In particular, following Liu et al.
According to the ASSET4 ESG Data glossary, emissions reduction
(2015) and Reguera‐Alvarado, de Fuentes, and Laffarga (2017), we
Measures a company's management commitment and focus on the board of directors, its monitoring role, and various CEO
effectiveness towards reducing environmental emission characteristics.
in the production and operational processes. It reflects a First, we use the natural logarithm of the number of board direc-
company's capacity to reduce air emissions, waste, tors (B_size). Prior studies (e.g., Dalton, Daily, Johnson, & Ellstrand,
hazardous waste, water discharges, spills or its impacts 1999) argue that larger boards may be more efficient than smaller
on biodiversity and to partner with environmental ones, as the former assign more people to supervise managers'
organisations to reduce the environmental impact of the actions, thus reducing the risk of opportunistic behaviors. Moreover,
company in the local or broader community. a larger sized board means more debate and a greater variety of opin-
ions within the board. A larger board will have a greater mix of skills to
The product innovation indicator
use in managing complexity (Said, Hj Zainuddin, & Haron, 2009) and
Measures a company's management commitment and will have a broader vision, which may enhance the consideration of
effectiveness towards supporting the research and environmental concerns as bank objectives.
development of eco‐efficient products or services. It Second, we measure the incidence of independent directors to
reflects a company's capacity to reduce the environmental total board directors (B_ind), because it is generally accepted that a
costs and burdens for its customers, and thereby creating board with greater independence has a stronger monitoring function.
new market opportunities through new environmental Fama and Jensen (1983) argued that independent directors care more
technologies and processes or eco‐designed, about their reputations. Therefore, because CSR is a reputational tool,
dematerialized products with extended durability. nonexecutive directors may bring greater attention to environmental
issues. Moreover, studies show that board independence strengthens
Finally, the resource reduction indicator
attention to the interests of different stakeholders, improving financial
Measures a company's management commitment and performance (e.g., Black & Kim, 2012; Harjoto, Laksmana, & Lee,
effectiveness towards achieving an efficient use of 2015).
natural resources in the production process. It reflects a Third, we use the incidence of female directors to total board
company's capacity to reduce the use of materials, directors (B_GD), as this is a measure of attention to gender diversity.
energy or water, and to find more eco‐efficient solutions Prior literature highlights an association between the involvement of
by improving supply chain management. women on the board and its effectiveness. The gender diversity of
the board should strengthen the bank's capability to act in the
Therefore, as a whole, each bank's ENV score covers a wide spec-
6
trum of policies and strategies. ASSET4 adopts only publicly available (https://financial.thomsonreuters.com)
GANGI ET AL. 7

TABLE 1 Sample distribution by country compensation scheme can be used to discourage opportunism and

Country N Percentage
promote shareholder‐wealth‐maximizing behavior (Sanders, 2001;
Shleifer & Vishny, 1997). Equity‐based compensation is usually
Australia 4 2.82
assumed to be a long‐term incentive (Mahoney & Thorne, 2005). Con-
Austria 1 0.70
sistent with the vision of CSR efforts as drivers of operating perfor-
Belgium 1 0.70
mance in the long run, the variable we adopt in this study is a
Brazil 1 0.70
dummy (CEO_comp), which assumes the value of 1 if the CEO's remu-
Canada 7 4.93
neration is connected to total shareholder return and 0 otherwise.
Chile 3 2.11
Finally, our analysis considers whether the bank adopts a duality
China 5 3.52
solution by separating the role of CEO from the position of chairman
Colombia 1 0.70
of the board. Indeed, it is generally accepted that this mechanism is
Denmark 3 2.11
synonymous with more democratic and efficient board functioning
France 4 2.82
(Boyd, 1995). A director with the double role of CEO and chairman
Germany 2 1.41
means a higher concentration of power, to the detriment of opportu-
Greece 4 2.82
nities for wider discussion within the board (Cannella & Shen, 2001).
Hong Kong 2 1.41
Consistent with the agency theory, the presence of a dual chairman
Hungary 1 0.70
and CEO could impact the effectiveness of board monitoring. CEO
India 4 2.82
duality enhances the short‐term financial‐performance orientation of
Ireland 1 0.70
the board (Davidson, Jiraporn, Kim, & Nemec, 2004; Zhang, 2012).
Israel 3 2.11
Therefore, as socially responsible activities produce their results in
Italy 7 4.93
the long term (Godos‐Díez, Cabeza‐García, Alonso‐Martínez, &
Kuwait 3 2.11
Fernández‐Gago, 2018), concentration of power may discourage the
Malaysia 7 4.93
board from addressing environmental issues. To measure the effect
Norway 1 0.70
of CEO duality, we adopt a dummy variable (CEO_ power), which is
Oman 1 0.70
equal to 1 if a CEO simultaneously chairs the board or if the chairman
Philippines 4 2.82
of the board has been the CEO of the bank and 0 otherwise.
Puerto Rico 1 0.70
Russia 2 1.41
3.2.4 | Control variables
Singapore 3 2.11
South Africa 5 3.52
Both a bank's environmental engagement and its risk taking may be
Spain 5 3.52
explained by several bank‐specific characteristics. In particular, we
Sweden 4 2.82
control for bank size through the natural logarithm of total assets
Switzerland 1 0.70
(TA); level of coverage through the incidence of loan loss reserve to
Taiwan 7 4.93
gross loans (Coverage); economic efficiency through the incidence of
Thailand 4 2.82
overhead costs to income (CostIncome); bank specialization through
Turkey 3 2.11
the ratio of loans to total assets (LoansAssets); and market concentra-
United Kingdom 6 4.23
tion through the Herfindahl–Hirschman Index (HHI). Finally, we con-
United States 31 21.83
trol for macroeconomic variation by using the gross domestic
product per capita (GDPper) based on current price/population.
interests of different stakeholders (Harjoto et al., 2015). Studies sug-
gest that women adopt a more ethical and moral approach than men 3.3 | Descriptive statistics
do (Sikula & Costa, 1994). Sapienza, Zingales, and Maestripieri
(2009) found that due to their lower testosterone levels, women are Consistent with the aim of the current study, Table 3 shows the
more risk‐averse than men. This risk aversion can lead to a greater descriptive statistics and tests related to bank‐specific characteristics
awareness of operational risks, including environmental risks. More- and the CG variables for banks that show high environmental perfor-
over, as stated by Nielsen and Huse (2010), some typical characteris- mance (the ENV score provided by ASSET4 is higher than the median
tics ascribed to women are consistent with environmental value of the sample) and low environmental performance (the ENV
engagement, such as being concerned with the welfare of other peo- score provided by ASSET4 is lower than the median value of the
ple and being helpful and interpersonally sensitive. sample).
Fourth, we consider the link between the CEO's compensation Statistically significant differences exist between the two groups
and shareholder returns. This is an additional instrument used to bet- of banks as previously categorized. In particular, banks with high envi-
ter align the interests of top management and ownership. CEOs ronmental performance are larger and show a higher incidence of
whose compensation is linked to shareholder remuneration should overhead costs to income, a higher coverage ratio of loans, lower
have more incentives to pursue corporate goals rather than personal loan‐to‐assets ratios, and a higher market concentration index. Fur-
interests. The incentive alignment hypothesis suggests that this thermore, banks that are more environmentally sensitive have larger
 8

TABLE 2 Correlation matrix

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Variable TA CostIncome Coverage LoansAsset HHI GDPper B_ind B_GD B_size CEO_comp CEO_power ENV

TA 1.0000
CostIncome 0.1395*** 1.0000
Coverage 0.0115 0.0609 1.0000
LoansAsset −0.4575*** −0.2415*** 0.0406 1.0000
HHI 0.1783*** −0.0431 0.2231*** 0.0002 1.0000
GDPper 0.2571*** 0.3877*** −0.1788*** −0.1418*** −0.0998* 1.0000
B_ind 0.1785*** 0.1844*** −0.2158*** 0.0478 0.0542 0.4893*** 1.0000
B_GD 0.3962*** 0.1314*** −0.0798** −0.1393*** 0.1613*** 0.2428*** 0.2970*** 1.0000
B_size 0.2278*** 0.1613*** 0.1148*** −0.1683*** 0.1251*** 0.0149 0.0166 0.1543*** 1.0000
CEO_comp 0.3065*** 0.3363*** −0.2631*** −0.1951*** −0.0317 0.4602*** 0.4923*** 0.2953*** 0.0684* 1.0000
CEO_power −0.1341*** 0.1264*** −0.0140 0.0707* −0.1316*** 0.1924*** 0.2411*** −0.0894** 0.0088 0.1667*** 1.0000
ENV 0.6280*** 0.1530*** 0.1562*** −0.2971*** 0.2410*** 0.0876** 0.1239*** 0.3796*** 0.1841*** 0.2822*** −0.1893*** 1.0000

Note. This table shows the Pearson pair‐wise correlation matrix of independent variables.
*Statistically significant at 10% level.
**Statistically significant at 5% level.
***Statistically significant at 1% level.
GANGI
ET AL.
GANGI ET AL. 9

TABLE 3 Descriptive statistics and univariate test

Banks with low environmental performance Banks with high environmental performance Difference tests
Variable N Mean N Mean T stat
Bank characteristics
TA (mln $) 348 128,286.6 362 605,679.5 −11.675***
CostIncome (%) 348 55.92 362 61.61 −5.5558***
Coverage (%) 348 2.36 362 3.50 −4.6305***
LoansAssets (%) 348 62.25 362 52.95 5.7911***
HHI 336 0.52 344 0.63 −6.0527***
Governance characteristics
B_size 348 12.93 362 14.17 −4.4494***
B_ind (%) 348 50.16 362 55.26 −4.8842***
B_GD (%) 343 13.77 361 21.51 −9.4001***
CEO_comp (1,0) 348 21.55 362 51.66 −8.7346***
CEO_duality (1,0) 348 37.64 362 26.24 3.2801***

Note. This table reports descriptive statistics for the 348 firm‐year observations of banks with an ENV score below the median of the sample and the 362
firm‐year observations of banks with an ENV score above the median of the sample. The number of firm‐year observations (N) and the mean are reported
by types of banks. Difference in mean (t statistics) tests are reported. The definitions of variables are provided in Table A1.
*Statistically significant at 10% level.
**Statistically significant at 5% level.
***Statistically significant at 1% level.

boards, with a higher incidence of independent and female directors to (xi,t), including the bank's characteristics and CG practices; ε is the ran-
total board composition. These banks also have a higher propensity dom error term.
for duality in terms of the CEO's separation from the Chairman. More- Regarding Hypothesis H2, it may be that banks that are more
over, banks that are more environmentally friendly are also organiza- environmentally sensitive are less risky simply due to their better qual-
tions that are more inclined to link the CEO's compensation to total ity and more effective governance mechanisms than less environmen-
shareholder return. tally sensitive banks. Therefore, the estimates of probit Equation (1)
The significant differences above need to be considered in the are used to calculate the inverse Mills ratio, which is then included
context of selection bias when we estimate the impact of environmen- as an additional explanatory variable. The second stage of the Heck-
tal engagement on bank risk. man model consists of the following ordinary least squares estimation:

yi;t ¼ α þ βENV i;t þ γXi;t þ δIMRi;t þ εi;t ; (2)

3.4 | Empirical strategy
where y denotes the bank risk (the Z‐score); ENV is the environmental
We argue in Hypothesis H1 that effective CG mechanisms have a pos-
score provided by ASSET4; X is a vector of explanatory variables,
itive relationship with a bank's environmental engagement. Subse-
including firm characteristics and governance measures; IMR is the
quently, we test the hypothesis of a positive impact of
inverse Mills ratio; and ε is the random error term.
environmental engagement on bank risk (Hypothesis H2).
Prior literature has stated that the study of the relationship
between CSR and financial performance may be biased by methodo- 4 | E M P I R I C A L R ES U L T S
logical limits, among which the endogeneity from selection bias plays
a relevant role (Jo & Harjoto, 2011). Therefore, in line with earlier Table 4 reports the estimates of the impact of several CG variables on
analysis (Wu & Shen, 2013), the current study adopts the two‐stage the level of the bank's environmental engagement. In particular, the
Heckman model (1979) to treat this bias. dependent variable (env_high) has the value of 1 if the ENV score is
In particular, in the first stage, we apply a probit model to deter- higher than the median value of the sampled banks and 0 otherwise.
mine whether firms with better governance practices are more likely The probit models include several bank‐specific characteristics.
to engage in a high level of environmental activities (Hypothesis H1). The results indicate that the functioning and the composition of
The model is formalized as follows: the board have a statistically significant impact on environmental
engagement. First, Models 2 and 6 show that the sensitivity of a bank


PRðDENVi; t ¼ 1Þ ¼ Φ β0 þ βi;t xi;t þ εi;t ; (1) towards the environment is higher if the CEO's compensation
(CEO_comp) is linked to shareholder returns (at the 1% confidence
where D_ENVi,t is a dummy variable equal to 1 if bank i shows a high level). This result is interesting because it may mean that safeguarding
environmental engagement in year t and 0 if it shows low environ- ownership does not conflict with the managerial incentives for higher
mental engagement. βi,t represents the coefficients of the predictors environmental protection. Second, Models 3 and 6 show that the
10 GANGI ET AL.

TABLE 4 Propensity to engage in environmental activities

(1) (2) (3) (4) (5) (6)

Variable D_ENV
Coverage 0.081*** (4.74) 0.108*** (5.91) 0.115*** (5.85) 0.119*** (5.75) 0.120*** (5.71) 0.122*** (5.78)
CostIncome 0.019*** (3.96) 0.015*** (3.06) 0.016*** (3.26) 0.015*** (3.01) 0.014*** (2.96) 0.015*** (2.99)
TA 0.544*** (10.07) 0.541*** (9.73) 0.532*** (9.52) 0.493*** (8.66) 0.492*** (8.62) 0.495*** (8.68)
LoansAsset −0.016*** (−3.63) −0.015*** (−3.42) −0.014*** (−3.36) −0.015*** (−3.57) −0.016*** (−3.63) −0.016*** (−3.63)
HHI 1.000*** (3.97) 1.040*** (4.19) 0.976*** (4.03) 0.939*** (3.77) 0.921*** (3.59) 0.927*** (3.60)
GDPper 0.000* (1.83) 0.000 (0.22) 0.000 (0.47) 0.000 (0.15) 0.000 (0.03) −0.000 (−0.02)
CEO_comp 0.757*** (5.51) 0.869*** (6.17) 0.815*** (5.69) 0.794*** (5.32) 0.797*** (5.35)
CEO_power −0.386** (−2.96) −0.318** (−2.41) −0.328** (−2.41) −0.326** (−2.40)
B_GD 0.0145** (2.44) 0.0140** (2.33) 0.0143** (2.37)
B_ind 0.00119 (0.41) 0.00126 (0.43)
B_size −0.0907 (−0.41)
_cons −15.089*** (−10.33) −14.93*** (−9.96) −14.74*** (−9.84) −13.82*** (−9.20) −13.79*** (−9.16) −13.66*** (−8.78)
N 680 680 680 674 674 674
Pseudo R2 0.3375 0.3662 0.3742 0.3757 0.3759 0.3760

Note. This table shows the coefficient of estimates from the probit model explaining the determining factors of environmental engagement. The dependent
variable is D_ENV, which is a dummy variable that is set to 1 if a bank exhibits an ENV score above the overall sample median. T statistics are adjusted for
robust standard errors and reported in round brackets. Table A1 provides the definition of variables.
*Statistically significant at 10% level.
**Statistically significant at 5% level.
***Statistically significant at 1% level.

double role of CEO and board chair (CEO_power) is a negative predic- In the first stage, we run the probit model with the same specification
tor of higher environmental engagement (at a confidence level of 5%). as in Table 4. In the second stage of the Heckman model (1979), the
Hence, a more democratic and balanced board of directors is a driver dependent variable is the Z‐score, not winsorized (Z_score) and
of the bank's commitment to the environment. Third, Models 4 and 6 winsorized at 1% and 99% of its distribution (Z_score_w). This indicator
indicate that the incidence rate of female directors (B_GD) to the total exhibits high values when a bank is less risky.
number of board members is a positive predictor of higher environ- In Models 1 and 7, at the 10% confidence level or less, bank risk
mental engagement (at a confidence level of 5%). This result is consis- decreases with the increase in environmental engagement measured
tent with prior analysis of the positive effect of board gender diversity through the score provided by ASSET4 (ENV). For example, looking
on CSR engagement (e.g., Bear, Rahman, & Post, 2010). at the results reported in column (7), the variable ENV exhibits a pos-
Therefore, the overall results confirm Hypothesis H1 of a positive itive (0.679) and significant coefficient (t = 2.15), suggesting that an
impact of effective CG mechanisms on the environmental engagement increase in environmental activities is associated with an increase in
of banks. Our test supports the evidence of Jo and Harjoto (2011), the Z‐score. It is worth noting that we find similar results when we
who find that CSR engagement is positively associated with CG. Fur- introduce controls for governance in Models 2–6 and 8–12.
thermore, the findings of the current study are consistent with Jo Therefore, consistent with Hypothesis H2, the higher the bank's
and Harjoto (2012), who observe that monitoring by the board is pos- sensitivity to environmental protection, the lower the bank risk. Given
itively related to CSR activities; this finding contrasts with the overin- the positive relationship between CSR and reputation (De Castro,
vestment hypothesis (Barnea & Rubin, 2010). Jo, Song, and Tsang López, & Sáez, 2006; Kim, 2017; Melo & Garrido‐Morgado, 2012), the
(2015) arrive at similar conclusions about the significant and positive results of the current study seem consistent with prior investigations
association between board governance and the average environmen- that find both a positive impact of CSR on bank performance (Cornett
tal and social scores provided by ASSET4. et al., 2016; García‐Sánchez & García‐Meca, 2017; Shen et al., 2016;
Regarding the controls for bank characteristics, we can see that Wu & Shen, 2013) and a positive association between a bank's reputa-
the coverage rate of loans (Coverage), the incidence of overhead costs tion and its accounting‐based performance (Dell'Atti et al., 2017;
on total income (CostIncome), the bank's size (TA), and the market con- Forcadell & Aracil, 2017). Furthermore, our study is in line with Jo,
centration (HHI) are all positive predictors of higher environmental Kim, and Kwangwoo (2015), who find that reducing environmental
engagement (at a confidence level of 1%). Specialization (LoanAsset), costs has positive effects on the performance of financial services firms.
however, negatively predicts the bank's sensitivity to environmental At the same time, differences exist between these findings and prior
protection (at a confidence level of 1%). empirical results, which showed a negative association between the
Table 5 shows the results of the two‐step Heckman (1979) social and the operating performance of banks (Scholtens & Dam,
regression models we have run to estimate the impact of environmen- 2007). Moreover, our results are not in line with Finger et al. (2018),
tal activities on bank risk while controlling for selection bias concerns. who find that among banks from developed countries, the best
 GANGI
ET AL.

TABLE 5 Z_score regressions based on the Heckman two‐stage treatment effect model

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Variable Z_score Z_score_w
ENV 0.680* 0.678* 0.849** 0.816** 0.811** 0.879** 0.679** 0.659** 0.794** 0.723** 0.717** 0.774**
(1.87) (1.84) (2.30) (2.19) (2.17) (2.36) (2.15) (2.05) (2.44) (2.21) (2.19) (2.36)
CostIncome −0.306* −0.389** −0.388* −0.386* −0.386* −0.395** −0.213 −0.301** −0.303** −0.304** −0.300** −0.308**
(−1.72) (−2.31) (−2.30) (−2.30) (−2.30) (−2.41) (−1.36) (−2.05) (−2.06) (−2.10) (−2.06) (−2.16)
TA 7.131** 5.343** 5.329** 4.303** 4.212 4.535* 6.747*** 4.727** 4.605** 3.318 3.178 3.452
(2.31) (1.98) (2.02) (1.66) (1.62) (1.77) (2.48) (2.01) (2.00) (1.48) (1.41) (1.56)
LoansAsset −0.126 −0.084 −0.078 −0.095 −0.099 −0.077 −0.097 −0.043 −0.036 −0.054 −0.070 −0.052
(−0.81) (−0.55) (−0.51) (−0.62) (−0.64) (−0.51) (−0.71) (−0.32) (−0.27) (−0.41) (−0.52) (−0.39)
HHI 0.885 −0.521 3.686 2.094 2.035 6.020 −1.002 −2.946 0.083 −1.369 −1.223 2.086
(0.07) (−0.05) (0.32) (0.18) (0.18) (0.54) (−0.10) (−0.29) (0.01) (−0.14) (−0.12) (0.21)
GDPper 0.000** 0.000 0.000* 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 −0.000
(2.16) (1.51) (1.82) (1.44) (1.35) (0.97) (1.25) (0.46) (0.76) (0.28) (0.12) (−0.25)
CEO_comp 9.504** 9.388** 8.016* 7.756* 8.926* 10.016*** 10.058*** 8.572** 7.760* 8.731**
(2.17) (2.09) (1.80) (1.66) (1.92) (2.67) (2.59) (2.23) (1.89) (2.17)
CEO_power 4.805 5.892 5.897 5.299 3.694 4.818 4.768 4.274
(1.09) (1.34) (1.34) (1.23) (0.96) (1.27) (1.25) (1.14)
B_GD 0.441** 0.441** 0.443** 0.479*** 0.482*** 0.483***
(2.38) (2.38) (2.45) (2.99) (3.01) (3.09)
B_Ind 0.113 0.000 0.040 0.0303
(0.14) (0.00) (0.55) (0.43)
B_size −21.30*** −17.59***
(−3.02) (−2.87)
IMR 30.005*** 26.019*** 28.454*** 29.076*** 28.883*** 25.953*** 27.486*** 22.410*** 23.905*** 23.762*** 23.807*** 21.422***
(3.20) (3.19) (3.46) (3.65) (3.63) (3.28) (3.33) (3.14) (3.33) (3.44) (3.44) (3.11)
_cons −216.599** −164.226* −185.974** −161.575* −158.801* −116.9 −206.836*** −146.605*** −160.193** −125.650* −122.327* −87.94
(−2.22) (−1.93) (−2.20) (−1.94) (−1.90) (−1.41) (−2.41) (−1.98) (−2.18) (−1.74) (−1.69) (−1.22)
N 661 661 661 655 655 655 661 661 661 655 655 655
Wald 26.46 29.40 32.71 37.16 37.02 47.21 22.95 27.00 29.52 36.69 36.89 46.10
chi‐square

Note. This table presents two‐step Heckman

(1979) regression coefficients and
(in parentheses) associated t statistics. In the first step, we run the probit model with the same specification as in Table 4. The inverse Mills ratio
(IMR) estimated from the first‐step regression is used in the second stage with the ASSET4 environmental score
(ENV), governance variables, and control variables. The dependent variable in the second stage is the Z_score. Model 1 reports the results of the ENV indicator and control variables. Models 2–6 also include corporate
governance variables. In Models 7–12, the dependent variable is winsorized (Z_score_w).
*Coefficient estimates significantly different from 0 at 10% level.
**Coefficient estimates significantly different from 0 at 5% level.
***Coefficient estimates significantly different from 0 at 1% level.
11
12 GANGI ET AL.

practices of environmental risk management have little or no effect on through the creation of incentives, as well as checks and balances in
profitability. order to generate long term shareholder value” (ASSET4 ESG Data
We also find that several bank‐specific characteristics and gover- Glossary). Interestingly, the reported results (columns 1 and 2 of
nance variables are significant predictors of bank risk. In particular, the Table 6) of these regressions are similar to those of Tables 4 and 5
higher the incidence of overhead on total income (CostIncome), the and enable us to confirm Hypotheses H1and H2. Second, although
higher the bank risk seems to be (except for Model 7, at the 10% con- the Z‐score is the most widely accounting‐based measure used in
fidence level or less). Most of our regression models indicate that also the empirical banking literature to capture a bank's risk, there are sev-
the bank's size matters (TA). Indeed, larger banks seem to experience a eral caveats that may affect our empirical results. Indeed, as sug-
lower risk (e.g., in Model 6, TA exhibits a positive coefficient that is gested by previous studies (e.g., Chiaramonte, Liu, Poli, & Zhou,
significant at the 10% confidence level). Regarding governance mech- 2016) the Z‐score may suffer from measurement bias, as banks may
anisms, some CG variables also exhibit a significant influence on bank engage in earnings management practices. Furthermore, it ignores
risk. In particular, our results show that gender diversity (B_GD) the importance of regulatory capital constraints faced by banks
within the board (Models 4–6 and 10–12) and the link between the (Bouvatier, Lepetit, Rehault, & Strobel, 2018). To address these draw-
CEO's compensation and shareholder returns (CEO_comp; Models backs and further stress the resilience of our results, we also use as
2–6 and 8–12) are both negative predictors of bank risk (at the 10% dependent variable the Tier1 capital ratio that is a core measure in
confidence level or less). In contrast, Models 6 and 12 show that the the International Money Funds' Financial Soundness Indicators and
size of the board (B_size) is a positive predictor of bank risk (at the the main indicator used by Banking Supervisory Authorities to
1% confidence level). monitor the soundness of banks. Once again, the reported results
of these regressions are qualitatively similar to those discussed in
the previous section.
4.1 | Robustness tests

To further confirm the findings described in the previous section, we 5 | DISCUSSION

perform several robustness checks. First, because there is no consen-
sus on some variables that we have used as proxies of good gover- The relationship between CG and CSR has received relatively scant
nance practices (viz., the size of the board and its gender diversity), attention from the banking industry's perspective. At the same time,
we replace our CG variables with the CG score provided by ASSET4, the relationship between CSR and a bank's financial performance is
“which measures a company's capacity, through its use of best man- the subject of ongoing debate (Shen et al., 2016). To explore this topic,
agement practices, to direct and control its rights and responsibilities prior literature has suggested going beyond the application of an

TABLE 6 Robustness tests

(1) (2)
First step
Variable Pr(D_ENV = 1) Z‐score Tier1
ENV 0.673** (2.02) 0.044* (1.91)
Coverage 0.122*** (4.66)
CostIncome 0.009* (1.64) −0.308** (−2.15) −0.046*** (−4.44)
TA 0.446*** (7.48) 3.344 (1.62) −0.493*** (−3.32)
LoansAssets −0.209*** (−3.58) −0.038 (−0.30) −0.042*** (−4.54)
HHI 1.320*** (4.01) −3.977 (−0.40) −1.560** (−2.20)
GDPper −0.000 (−0.31) 0.000 (0.54) 0.000*** (3.29)
CG score 0.026*** (8.11) 0.254** (2.23) −0.017** (−2.05)
IMR 21.065*** (3.27) −1.635*** (−3.52)
_cons −13.248*** (−7.76) −121.310* (−1.80) 28.431*** (5.92)
No. of ob. 661 661 661
Pseudo R2 0.4385
Wald chi‐square 24.82 56.21

Note. This table presents two‐step Heckman (1979) regression coefficients and (in parentheses) associated t statistics (i.e., z‐statistics in the case of probit
regression) based on standard errors that are robust to heteroskedasticity. In the first step, a probit regression is estimated for the variable D_ENV, which is
a dummy that is set to 1 if a bank exhibits an ENV score above the overall sample median. The inverse Mills ratio (IMR) estimated from the first‐step regres-
sion is used in the second stage with the ENV score and CG score provided by ASSET4 and control variables. The dependent variables in the second stage
are Z_score (Model 1) and Tier1 (Model 2). CG: corporate governance.
*Coefficient estimates significantly different from 0 at 10% level.
**Coefficient estimates significantly different from 0 at 5% level.
***Coefficient estimates significantly different from 0 at 1% level.
GANGI ET AL. 13

aggregate index of social performance by using the subcomponents of gain the trust of their customers (Venturelli et al., 2018). People are
the CSR index (Wu & Shen, 2013). Moreover, other studies (Esteban‐ more inclined to deposit money in banks with a stronger reputation.
Sanchez et al., 2017) of banks' CSR have omitted to analyze, among Highly reputable banks also have the opportunity to pay lower interest
other subcomponents, the environmental subcomponent, arguing that on deposits, potentially increasing their intermediation margin. This
banks are not particularly polluting. The current study aimed to fill economic benefit mirrors the green premium for environmentally
these gaps. friendly companies (Mason, 2012). Moreover, banks with a friendly
First, we verified whether more effective CG mechanisms are pos- approach to the environment are more sensitive to the interests of
itively related to bank's environmental engagement. By confirming the community. This helps with the acquisition of a citizenship right
Hypothesis H1, the overall results support the conflict resolution (Matten & Crane, 2005). Banks' social legitimacy supports their eco-
model (Baron, 2009; Calton & Payne, 2003) and the conception of nomic legitimacy: They may inspire trust among stakeholders and, in
CG as a pillar of CSR (Elkington, 2006; Jamali et al., 2008). This is con- turn, may receive more attention from companies that want to be cer-
sistent with a broader view of CG (Lashgari, 2004), as opposed to the tified by the most highly reputable banks. This desirability offers rep-
narrow vision that limits CG to mitigating agency problems between utable banks an opportunity to increase their market share and to
ownership and management. Our estimates indicate that the best apply higher interest rates and commissions than less renowned
practices of CG are associated with the bank's higher level of commit- banks. Greater attention to the environment is a proxy of a bank's
ment to the environment. CG tools that are arranged to ensure open long‐term profitability vision and higher quality of services. All these
debate and democracy within the board lead to higher standards of factors may contribute to a reduction in bank risk, as this study has
environmental protection. The greater sensitivity of banks to the envi- documented.
ronment reflects the application of more effective CG tools. This find-
ing is consistent with a positive relationship between CG and CSR
(Jo, Song, & Tsang, 2015), whereas it contrasts with the agency per- 6 | CO NC LUSIO N
spective (Barnea & Rubin, 2010; Friedman, 1970). In our analysis,
the environmental engagement of banks is not a pathological form The current study analyzes the role of social responsibility in the bank-
of overinvesting that should be avoided through effective CG ing industry, with specific reference to the natural environment. In
mechanisms. particular, our research question is twofold. On the one hand, we
Second, consistent with Hypothesis H2, the overall results sug- deepen the understanding of the relationship between the best prac-
gest that banks that are more engaged in environmental protection tices of CG and the commitment of banks to environmental sustain-
are also less risky. There may be several explanations for this result. ability. On the other hand, we analyze the association between the
Banks that are more engaged in environmentally friendly actions are commitment to environmental responsibility and bank risk.
probably more efficient in their use of resources. This is a strictly oper- In the first case, the theoretical framework considers CG as a pillar
ational motive, as the prevention of environmental risk may generate of CSR. Therefore, more effective governance is a driver of greater
economic value. However, banks are financial intermediaries that not sensitivity to the environmental policies of banks. In the second case,
only provide money but also affect how money is invested by bor- the theoretical model considers the impact that greater efficiency in
rowers. In this sense, banks have a relevant social impact. Their financ- the use of resources and a strong reputation may have on the financial
ing modes may drive economic activities in a way that accounts for performance of banks. The results of our analysis are consistent with
environmental issues (Scholtens, 2009). Banks' level of care for the both frameworks.
environment may reflect additional requirements, such as lower envi- By focusing on the board's composition and its monitoring func-
ronmental risk, which relates to how an entrepreneur designs and tion, the present study shows that the governance solutions com-
manages a business (Weber, 2012). Banks' increased attention to the monly accepted as most effective at the board level are also
environment may be reflected in the selection of more sustainable positively associated with the bank's environmental commitment. This
and competitive firms that are equipped with a higher capacity for contradicts the narrow view of CG according to which effective gov-
stakeholder influence (Barnett, 2007) and high earnings quality, a high ernance would limit the commitment to CSR in the interest of share-
degree of transparency, and high moral standards (Lopatta et al., holders. In contrast, a board that is designed in a broad and more
2016). Of course, these factors help banks to mitigate the adverse democratic way, with a greater openness to gender diversity, contrib-
selection arising from an ex‐ante information gap and the moral haz- utes to increasing the bank's attention to the environment.
ard caused by an ex‐post information gap. Furthermore, banks' sensi- Concerning the relation between environmental commitment and
tivity to the environment helps to improve both their reputations and financial performance, this study presents results that are consistent
customer loyalty (Branco & Rodrigues, 2006; De Castro et al., 2006; with the reputation theory of financial intermediaries. That is, due to
Deng & Xu, 2017; Kim, 2017; Melo & Garrido‐Morgado, 2012). Repu- the positive relationship between CSR and reputation, considering
tation is a key asset for organizations such as banks, which base their the environment as a pillar of CSR, most environmentally committed
affairs on trust (Fiordelisi, Soana, & Schwizer, 2013; Pérez & Rodríguez banks are also less risky.
del Bosque, 2015). By improving their reputations, banks may enjoy Our study advances prior studies in many ways. First, as
several competitive advantages (Bushman & Wittenberg‐Moerman, requested by Esteban‐Sanchez et al. (2017), this study covers a more
2012; Dell'Atti et al., 2017). Consistent with the stakeholder view recent period than previous works. Second, unlike earlier investiga-
(Freeman, 1984), banks with the strongest reputation will more easily tions, our methodology was designed to establish the existence of a
14 GANGI ET AL.

relationship between CSR engagement and risk in the banking indus- Aguilera, R. V., Williams, C. A., Conley, J. M., & Rupp, D. E. (2006).
try. More specifically, as suggested by Wu and Shen (2013), the pres- Corporate governance and social responsibility: A comparative analysis
of the UK and the US. Corporate Governance: An International Review,
ent study analyzes this relationship by using the environmental 14(3), 147–158. https://doi.org/10.1111/j.1467‐8683.2006.00495.x
performance of banks as a subcomponent of the aggregated CSR
Albertini, E. (2018). The contribution of management control systems to
score to which the literature has frequently referred. In this sense, environmental capabilities. Journal of Business Ethics, 1–18. https://
the current study responds to the call to deepen our understanding doi.org/10.1007/s10551‐018‐3810‐9
of how banks' engagement in CSR is related to their performance Aramburu, I. A., & Pescador, I. G. (2017). The effects of corporate social
(Shen et al., 2016). Fourth, by examining a synthetic indicator of bank responsibility on customer loyalty: The mediating effect of reputation
in cooperative banks versus commercial banks in the Basque Country.
risk, our analysis adopts a measure of financial performance rarely Journal of Business Ethics, 1–19. https://doi.org/10.1007/s10551‐017‐
adopted by earlier studies on the relation between corporate financial 3438‐1
performance and corporate social performance. Aras, G., Tezcan, N., & Kutlu Furtuna, O. (2018). The value relevance of
The current study has both theoretical and practical implications. banking sector multidimensional corporate sustainability performance.
Corporate Social Responsibility and Environmental Management. https://
In particular, we confirm CSR engagement as a win‐win business
doi.org/10.1002/csr.1520
model that satisfies both the community's needs and the bank's com-
ASSET4 ESG Data Glossary (2018) http://lipas.uwasa.fi/~jaty/thomson/
petitiveness. Banks that are more committed to the environment ASSET4_ESG_Content_Datastream_Premium_Datatypes.xlsx
tend to outperform banks that pay less attention to environmental
Bansal, P. (2005). Evolving sustainably: A longitudinal study of corporate
responsibility. The reduction in environmental costs and risks may sustainable development. Strategic Management Journal, 26, 197–218.
support bank stability during turbulent scenarios such as the last https://doi.org/10.1002/smj.441
financial crisis (Ducassy, 2013). At the same time, banks dealing with Barnea, A., & Rubin, A. (2010). Corporate social responsibility as a conflict
reputational risk may rely on environmentally friendly policies to between shareholders. Journal of Business Ethics, 97(1), 71–86. https://
doi.org/10.1007/s10551‐010‐0496‐z
improve trust relationships with stakeholders. Thus, due to links
Barnett, M. L. (2007). Stakeholder influence capacity and the variability
among the individual pillars of CSR, the current study sheds light
of financial returns to corporate social responsibility. Academy of
on a systemic approach to CSR. Finally, consistent with the social Management Review, 32(3), 794–816. https://doi.org/10.5465/
impact view, our findings suggest that policy makers dealing with amr.2007.25275520
CSR should incentivize an environmentally centered approach in the Barnett, M. L., & Salomon, R. M. (2012). Does it pay to be really good?
banking industry. Addressing the shape of the relationship between social and financial
performance. Strategic Management Journal, 33(11), 1304–1320.
Given some of this work's limitations, further research could https://doi.org/10.1002/smj.1980
expand the current study in several ways. First, future studies could
Baron, D. (2009). A positive theory of moral management, social
disentangle the impact of internal environmental engagement and pressure and corporate social performance. Journal of Economics and
operational practices from the effect of lending money to more envi- Management Strategy, 18, 7–43. https://doi.org/10.1111/j.1530‐
9134.2009.00206.x
ronmentally friendly borrowers. Indeed, the measure of the bank's
environmental performance we adopt in this study does not allow us Bassen, A., Meyer, K., & Schlange, J. (2006). The influence of corporate
responsibility on the cost of capital (Vol. 984406). Available at SSRN:
to evaluate these effects separately. Second, the effectiveness of CG http://ssrn.com/abstract.
may depend on the combination of many solutions not limited to the
Bear, S., Rahman, N., & Post, C. (2010). The impact of board diversity and
board. Therefore, further research could deepen the relationship with gender composition on corporate social responsibility and firm reputa-
the bank's commitment to environmental issues by applying additional tion. Journal of Business Ethics, 97(2), 207–221. https://doi.org/
10.1007/s10551‐010‐0505‐2
proxies of effective CG mechanisms. Third, future studies could
deepen the issue of the reverse causality between an environmentally Berle, A., & Means, G. (1932). The modern corporation and private prop-
erty Macmillan. New York.
friendly attitude and bank risk. Finally, future surveys could focus on
Bernanke, B. S. (1983). Irreversibility, uncertainty, and cyclical investment.
customers' perception of the bank's environmental responsibility to
The Quarterly Journal of Economics, 98(1), 85–106.
obtain specific information for bank management. Indeed, all these
Black, B., & Kim, W. (2012). The effect of board structure on firm value: A
aspects may contribute to further clarifying the broader theme of multiple identification strategies approach using Korean data. Journal of
socially responsible banking. Financial Econonomics, 104, 203–226. https://doi.org/10.1016/j.
jfineco.2011.08.001
Boston Common Asset Management. Impact Report (2018). Banking on
ORCID
a low‐carbon future: Are the world's largest banks stepping up to
Francesco Gangi http://orcid.org/0000-0001-8598-3568 the risks & opportunities of climate change? http://news.
Antonio Meles http://orcid.org/0000-0002-9531-4517 bostoncommonasset.com/wp‐content/uploads/2018/02/Banking‐on‐
a‐Low‐Carbon‐Future‐2018‐02.pdf
Eugenio D'Angelo http://orcid.org/0000-0002-5532-6388
Bouvatier, V., Lepetit, L., Rehault, P. N., & Strobel, F. (2018). Bank insol-
Lucia Michela Daniele http://orcid.org/0000-0002-7394-8391
vency risk and Z‐score measures: Caveats and best practice.
Boyd, B. K. (1995). CEO duality and firm performance: A contingency
model. Strategic Management Journal, 16(4), 301–312. https://doi.org/
RE FE R ENC E S 10.1002/smj.4250160404
Adler, P. S., & Kwon, S. W. (2002). Social capital: Prospects for a new con- Branco, M. C., & Rodrigues, L. L. (2006). Corporate social responsibility and
cept. Academy of Management Review, 27(1), 17–40. https://doi.org/ resource‐based perspectives. Journal of Business Ethics, 69(2), 111–132.
10.5465/amr.2002.5922314 https://doi.org/10.1007/s10551‐006‐9071‐z
GANGI ET AL. 15

Bushman, R. M., & Wittenberg‐Moerman, R. (2012). The role of bank rep- An empirical analysis. Corporate Social Responsibility and Environmental
utation in “certifying” future performance implications of borrowers' Management, 24(6), 589–605. https://doi.org/10.1002/csr.1430
accounting numbers. Journal of Accounting Research, 50(4), 883–930.
Delmas, M. A., & Montes‐Sancho, M. J. (2010). Voluntary agreements to
https://doi.org/10.1111/j.1475‐679X.2012.00455.x
improve environmental quality: Symbolic and substantive cooperation.
Cai, L., Cui, J., & Jo, H. (2016). Corporate environmental responsibility and Strategic Management Journal, 31(6), 575–601.
firm risk. Journal of Business Ethics, 139, 563–594. https://doi.org/
Deng, X., & Xu, Y. (2017). Consumers' responses to corporate social
10.1007/s10551‐015‐2630‐4
responsibility initiatives: The mediating role of consumer–company
Callan, S. J., & Thomas, J. M. (2014). Relating CEO compensation to social identification. Journal of Business Ethics, 142(3), 515–526. https://doi.
performance and financial performance: Does the measure of compen- org/10.1007/s10551‐015‐2742‐x
sation matter? Corporate Social Responsibility and Environmental
Dimson, E., Karakaş, O., & Li, X. (2015). Active ownership. The Review of
Management, 21(4), 202–227. https://doi.org/10.1002/csr.1307
Financial Studies, 28(12), 3225–3268. https://doi.org/10.1093/rfs/
Calomiris, C. W., & Mason, J. R. (2003). Consequences of bank distress hhv044
during the Great Depression. American Economic Review, 93(3),
Ducassy, I. (2013). Does corporate social responsibility pay off in times of
937–947. https://doi.org/10.1257/000282803322157188
crisis? An alternate perspective on the relationship between financial
Calton, J. M., & Payne, S. L. (2003). Coping with paradox: Multistakeholder and corporate social performance. Corporate Social Responsibility and
learning dialogue as a pluralist sensemaking process for addressing Environmental Management, 20(3), 157–167. https://doi.org/10.1002/
messy problems. Business & Society, 42(1), 7–42. https://doi.org/ csr.1282
10.1177/0007650302250505
El Ghoul, S., Guedhami, O., & Kim, Y. (2017). Country‐level institutions,
Cannella, A. A., & Shen, W. (2001). So close and yet so far: Promotion ver- firm value, and the role of corporate social responsibility initiatives.
sus exit for CEO heirs apparent. Academy of Management Journal, 44(2), Journal of International Business Studies, 48(3), 360–385. https://doi.
252–270. org/10.1057/jibs.2016.4
Carvalho, J. M. S., Jonker, J., & Dentchev, N. (2014). What's in a word? An El Ghoul, S., Guedhami, O., Kwok, C., & Mishra, D. (2011). Does corporate
exploration of the changes in meaning of corporate social responsibility social responsibility affect the cost of capital? Journal of Banking &
over the last century with an emphasis on the last decades. In Finance, 35(9), 2388–2406. https://doi.org/10.1016/j.jbankfin.2011.
D. Turker, H. Toker, & C. Altuntas (Eds.), Contemporary issues in corpo- 02.007
rate social responsibility. Plymouth: Lexington Books.
Elijido‐Ten, E. O., & Clarkson, P. (2017). Going beyond climate change risk
Cheng, B., Ioannou, I., & Serafeim, G. (2014). Corporate social responsibility
management: Insights from the world's largest most sustainable corpo-
and access to finance. Strategic Management Journal, 35(1), 1–23.
rations. Journal of Business Ethics, 1–23. https://doi.org/10.1007/
https://doi.org/10.1002/smj.2131
s10551‐017‐3611‐6
Chiaramonte, L., Liu, H., Poli, F., & Zhou, M. (2016). How accurately can
Elkington, J. (2006). Governance for sustainability. Corporate Governance:
Z‐score predict bank failure? Financial Markets, Institutions & Instru-
An International Review, 14(6), 522–529. https://doi.org/10.1111/
ments, 25(5), 333–360. https://doi.org/10.1111/fmii.12077
j.1467‐8683.2006.00527.x
Clarkson, P., Li, Y., Pinnuck, M., & Richardson, G. (2015). The valuation rel-
Esteban‐Sanchez, P., de la Cuesta‐Gonzalez, M., & Paredes‐Gazquez, J. D.
evance of greenhouse gas emissions under the European Union carbon
(2017). Corporate social performance and its relation with corporate
emissions trading scheme. European Accounting Review, 24(3),
financial performance: International evidence in the banking industry.
551–580. https://doi.org/10.1080/09638180.2014.927782
Journal of Cleaner Production, 162, 1102–1110. https://doi.org/
Clarkson, P., Li, Y., Richardson, G., & Vasvari, F. (2011). Does it pay to be 10.1016/j.jclepro.2017.06.127
green? Determinants and consequences of proactive environmental
EUROSIF (2016). European SRI study. Available at: https://www.euros if.
strategies. Journal of Accounting and Public Policy, 30(2), 122–144.
org/sri‐study‐2016/. Accessed March 3, 2017.
https://doi.org/10.1016/j.jaccpubpol.2010.09.013
Cornett, M. M., Erhemjamts, O., & Tehranian, H. (2016). Greed or good Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control.
deeds: An examination of the relation between corporate social The Journal of Law and Economics, 26(2), 301–325. https://doi.org/
responsibility and the financial performance of US commercial banks 10.1086/467037
around the financial crisis. Journal of Banking & Finance, 70, 137–159. Finger, M., Gavious, I., & Manos, R. (2018). Environmental risk manage-
https://doi.org/10.1016/j.jbankfin.2016.04.024 ment and financial performance in the banking industry: A cross‐
Crifo, P., Escrig‐Olmedo, E., & Mottis, N. (2018). Corporate governance as a country comparison. Journal of International Financial Markets,
key driver of corporate sustainability in France: The role of board mem- Institutions and Money, 52, 240–261. https://doi.org/10.1016/j.
bers and investor relations. Journal of Business Ethics, 1–20. https://doi. intfin.2017.09.019
org/10.1007/s10551‐018‐3866‐6 Fiordelisi, F., & Mare, D. S. (2014). Competition and financial stability in
Dahlsrud, A. (2008). How corporate social responsibility is defined: An European cooperative banks. Journal of International Money and
analysis of 37 definitions. Corporate Social Responsibility and Environ- Finance, 45, 1–16. https://doi.org/10.1016/j.jimonfin.2014.02.008
mental Management, 15(1), 1–13. https://doi.org/10.1002/csr.132 Fiordelisi, F., Marques‐Ibanez, D., & Molyneux, P. (2011). Efficiency and
Dalton, D. R., Daily, C. M., Johnson, J. L., & Ellstrand, A. E. (1999). Number risk in European banking. Journal of Banking & Finance, 35(5),
of directors and financial performance: A meta‐analysis. Academy of 1315–1326. https://doi.org/10.1016/j.jbankfin.2010.10.005
Management Journal, 42(6), 674–686. Fiordelisi, F., Soana, M. G., & Schwizer, P. (2013). The determinants of rep-
Davidson, W. N., Jiraporn, P., Kim, Y. S., & Nemec, C. (2004). Earnings man- utational risk in the banking sector. Journal of Banking & Finance, 37(5),
agement following duality–creating successions: Ethnostatistics, 1359–1371. https://doi.org/10.1016/j.jbankfin.2012.04.021
impression management, and agency theory. Academy of Management Fombrun, C. J., Gardberg, N. A., & Barnett, M. L. (2000). Opportunity plat-
Journal, 47(2), 267–275. forms and safety nets: Corporate citizenship and reputational risk.
De Castro, G. M., López, J. E. N., & Sáez, P. L. (2006). Business and social Business and Society Review, 105(1), 85–106. https://doi.org/
reputation: Exploring the concept and main dimensions of corporate 10.1111/0045‐3609.00066
reputation. Journal of Business Ethics, 63(4), 361–370. https://doi.org/ Forcadell, F. J., & Aracil, E. (2017). European Banks' reputation for corpo-
10.1007/s10551‐005‐3244‐z rate social responsibility. Corporate Social Responsibility and
Dell'Atti, S., Trotta, A., Iannuzzi, A. P., & Demaria, F. (2017). Corporate Environmental Management, 24(1), 1–14. https://doi.org/10.1002/
social responsibility engagement as a determinant of bank reputation: csr.1402
16 GANGI ET AL.

Freeman, R. E. (1984). Strategic management: A stakeholder approach. Jo, H., & Harjoto, M. A. (2012). The causal effect of corporate governance
Boston: Pitman. on corporate social responsibility. Journal of Business Ethics, 106(1),
Friedman, M. (1970). The social responsibility of business is to increase its 53–72. https://doi.org/10.1007/s10551‐011‐1052‐1
profits. New York Times Magazine, September 13. Jo, H., Kim, H., & Kwangwoo, P. (2015). Corporate environmental respon-
Gangi, F., & Varrone, N. (2018). Screening activities by socially responsible sibility and firm performance in the financial services sector. Journal
funds: A matter of agency. Journal of Cleaner Production, 197, 842–855. of Business Ethics, 131, 257–284.
https://doi.org/10.1016/j.jclepro.2018.06.228 Jo, H., Song, H. M., & Tsang, A. (2015). Corporate social responsibility and
García‐Sánchez, I. M., & García‐Meca, E. (2017). CSR engagement and stakeholder governance around the world. Global Finance Journal,
earnings quality in banks. The moderating role of institutional factors. https://doi.org/10.1016/j.gfj.2015.04.003, 29, 1–28.
Corporate Social Responsibility and Environmental Management, 24(2), Jung, J., Herbohn, K., & Clarkson, P. (2016). Carbon risk, carbon risk aware-
145–158. https://doi.org/10.1002/csr.1405 ness and the cost of debt financing. Journal of Business Ethics, 150,
Gatzert, N. (2015). The impact of corporate reputation and reputation 1–21.https://doi.org/10.1007/s10551‐016‐3207‐6.
damaging events on financial performance: Empirical evidence from Kassinis, G., & Vafeas, N. (2006). Stakeholder pressures and environmental
the literature. European Management Journal, 33(6), 485–499. https:// performance. Academy of Management Journal, 49(1), 145–159.
doi.org/10.1016/j.emj.2015.10.001 https://doi.org/10.5465/amj.2006.20785799
Giannarakis, G., Zafeiriou, E., Arabatzis, G., & Partalidou, X. (2018). Deter- Kendall, E. A. (1999). Role models—Patterns of agent system analysis and
minants of corporate climate change disclosure for European firms. design. BT Technology Journal, 17(4), 46–57. https://doi.org/10.1023/
Corporate Social Responsibility and Environmental Management, 25(3), A:1009690824323
281–294. https://doi.org/10.1002/csr.1461
Kim, S. (2017). The process model of corporate social responsibility (CSR)
Global Sustainable Investment Review (2016). Retrived from: http://www. communication: CSR communication and its relationship with con-
gsi‐alliance.org/wp‐content/uploads/2017/03/GSIR_Review2016.F.pdf sumers' CSR knowledge, trust, and corporate reputation perception.
Godos‐Díez, J. L., Cabeza‐García, L., Alonso‐Martínez, D., & Fernández‐ Journal of Business Ethics, 1–17. https://doi.org/10.1007/s10551‐
Gago, R. (2018). Factors influencing board of directors' decision‐ 017‐3433‐6
making process as determinants of CSR engagement. Review of Mana- Kim, Y., Park, M. S., & Wier, B. (2012). Is earnings quality associated with
gerial Science, 12(1), 229–253. https://doi.org/10.1007/s11846‐016‐ corporate social responsibility? The Accounting Review, 87(3),
0220‐1 761–796. https://doi.org/10.2308/accr‐10209
Goetz, M. R., Laeven, L., & Levine, R. (2016). Does the geographic expan- Laeven, L., & Levine, R. (2009). Bank governance, regulation and risk tak-
sion of banks reduce risk? Journal of Financial Economics, 120(2), ing. Journal of Financial Economics, 93(2), 259–275.
346–362. https://doi.org/10.1016/j.jfineco.2016.01.020
Lashgari, M. (2004). Corporate governance: Theory and practice. The
Goss, A., & Roberts, G. S. (2011). The impact of corporate social responsi- Journal of American Academy of Business, 5(1/2), 46–51.
bility on the cost of bank loans. Journal of Banking & Finance, 35(7),
Lee, J. W., Kim, Y. M., & Kim, Y. E. (2018). Antecedents of adopting corpo-
1794–1810. https://doi.org/10.1016/j.jbankfin.2010.12.002
rate environmental responsibility and green practices. Journal of
Halkos, G. E., & Paizanos, E. A. (2016). Environmental macroeconomics: Business Ethics, 148(2), 397–409. https://doi.org/10.1007/s10551‐
Economic growth, fiscal spending and environmental quality. Interna- 016‐3024‐y
tional Review of Environmental and Resource Economics, 9(3–4),
Liu, Y., Miletkov, M. K., Wei, Z., & Yang, T. (2015). Board independence and
321–362. https://doi.org/10.1561/101.00000079
firm performance in China. Journal of Corporate Finance, 30, 223–244.
Harjoto, M., Laksmana, I., & Lee, R. (2015). Board diversity and corporate https://doi.org/10.1016/j.jcorpfin.2014.12.004
social responsibility. Journal of Business Ethics, 132(4), 641–660.
Lopatta, K., Buchholz, F., & Kaspereit, T. (2016). Asymmetric information
https://doi.org/10.1007/s10551‐014‐2343‐0
and corporate social responsibility. Business & Society, 55(3),
Harris, J. D., & Freeman, R. E. (2008). The impossibility of the separation 458–488. https://doi.org/10.1177/0007650315575488
thesis: A response to Joakim Sandberg. Business Ethics Quarterly,
Mahoney, L. S., & Thorne, L. (2005). Corporate social responsibility and
18(4), 541–548. https://doi.org/10.5840/beq200818437
long‐term compensation: Evidence from Canada. Journal of Business
Hart, S. L., & Dowell, G. (2011). Invited editorial: A natural‐resource‐based Ethics, 57, 241–253. https://doi.org/10.1007/s10551‐004‐5367‐z
view of the firm: Fifteen years after. Journal of Management, 37(5),
Margolis, J. D., & Walsh, J. P. (2003). Misery loves companies: Rethinking
1464–1479. https://doi.org/10.1177/0149206310390219
social initiatives by business. Administrative Science Quarterly, 48(2),
Heckman, J. J. (1979). Sample selection bias as a specification error. 268–305. https://doi.org/10.2307/3556659
Econometrica, 47, 153–161. https://doi.org/10.2307/1912352
Mason, C. F. (2012). The economics of eco‐labeling: Theory and empirical
Hook, L. (2018). Central bank chiefs sound warning on climate change. The implications. International Review of Environmental and Resource
Financial Times, April 2018. Economics, 6, 341–372.
Huse, M. (2009). Accountability and creating accountability: A framework Matten, D., & Crane, A. (2005). Corporate citizenship: Toward an extended
for exploring behavioural perspectives of corporate governance. In theoretical conceptualization. Academy of Management Review, 30(1),
The value creating board (pp. 51–72). New York: Routledge. 166–179. https://doi.org/10.5465/amr.2005.15281448
Jamali, D., Safieddine, A. M., & Rabbath, M. (2008). Corporate governance Matute‐Vallejo, J., Bravo, R., & Pina, J. M. (2011). The influence of corpo-
and corporate social responsibility synergies and interrelationships. rate social responsibility and price fairness on customer behaviour:
Corporate Governance: An International Review, 16(5), 443–459. Evidence from the financial sector. Corporate Social Responsibility and
https://doi.org/10.1111/j.1467‐8683.2008.00702.x Environmental Management, 18(6), 317–331. https://doi.org/10.1002/
Jensen, M. C. (2001). Value maximization, stakeholder theory, and the cor- csr.247
porate objective function. Journal of Applied Corporate Finance, 14(3), McWilliams, A., & Siegel, D. S. (2011). Creating and capturing value: Strate-
8–21. https://doi.org/10.1111/j.1745‐6622.2001.tb00434.x gic corporate social responsibility, resource‐based theory, and
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial sustainable competitive advantage. Journal of Management, 37(5),
behavior, agency costs and ownership structure. Journal of Financial 1480–1495.
Economics, 3(4), 305–360. Melo, T., & Garrido‐Morgado, A. (2012). Corporate reputation: A combina-
Jo, H., & Harjoto, M. A. (2011). CG and firm value: The impact of corporate tion of social responsibility and industry. Corporate Social Responsibility
social responsibility. Journal of Business Ethics, 103(3), 351–383. and Environmental Management, 19(1), 11–31. https://doi.org/
https://doi.org/10.1007/s10551‐011‐0869‐y 10.1002/csr.260
GANGI ET AL. 17

Montiel, I. (2008). Corporate social responsibility and corporate sustain- Schücking, H., Kroll, L., Louvel, Y., & Richter, R. “Bankrolling climate
ability. Organization and Environment, 21, 245–269. https://doi.org/ change”, (2011), Published by urgewald, groundWork, Earthlife Africa
10.1177/1086026608321329 Johannesburg and BankTrack.
Nielsen, S., & Huse, M. (2010). The contribution of women on boards of Sen, S., Bhattacharya, C. B., & Korschun, D. (2006). The role of corporate
directors: Going beyond the surface. Corporate Governance: An Interna- social responsibility in strengthening multiple stakeholder relationships:
tional Review, 18(2), 136–148. https://doi.org/10.1111/j.1467‐ A field experiment. Journal of the Academy of Marketing Science, 34(2),
8683.2010.00784.x 158–166. https://doi.org/10.1177/0092070305284978
Oikonomou, I., Brooks, C., & Pavelin, S. (2012). The impact of corporate Sharma, S., Aragon‐Correa, J.‐A., & Rueda‐Manzanares, A. (2007). The con-
social performance on financial risk and utility: A longitudinal analysis. tingent influence of organizational capabilities on proactive
Financial Management, 41(2), 483–515. https://doi.org/10.1111/ environmental strategy in the service sector: An analysis of North
j.1755‐053X.2012.01190.x American and European ski resorts. Canadian Journal of Administrative
Pérez, A., & Rodríguez del Bosque, I. R. (2015). How customer support for cor- Sciences, 24(4), 268–283. https://doi.org/10.1002/cjas.35
porate social responsibility influences the image of companies: Evidence Shen, C. H., Wu, M. W., Chen, T. H., & Fang, H. (2016). To engage or not to
from the banking industry. Corporate Social Responsibility and Environmen- engage in corporate social responsibility: Empirical evidence from
tal Management, 22(3), 155–168. https://doi.org/10.1002/csr.1331 global banking sector. Economic Modelling, 55, 207–225. https://doi.
Porter, M. E., & van der Linde, C. (1995). Toward a new conception of the org/10.1016/j.econmod.2016.02.007
environment‐competitiveness relationship. Journal of Economic Shleifer, A., & Vishny, R. W. (1997). A survey of corporate governance. The
Perspectives, 9(4), 97–118. https://doi.org/10.1257/jep.9.4.97 Journal of Finance, 52(2), 737–783. https://doi.org/10.1111/j.1540‐
Putnam, R. (1993). Making democracy work: Civic traditions in modern Italy. 6261.1997.tb04820.x
Princeton: Princeton University Press. Sikula, A., & Costa, A. D. (1994). Are women more ethical than men? Jour-
Reguera‐Alvarado, N., de Fuentes, P., & Laffarga, J. (2017). Does board nal of Business Ethics, 13(11), 859–871. https://doi.org/10.1007/
gender diversity influence financial performance? Evidence from Spain. BF00871700
Journal of Business Ethics, 141(2), 337–350. https://doi.org/10.1007/ Soana, M. G. (2011). The relationship between corporate social perfor-
s10551‐015‐2735‐9 mance and corporate financial performance in the banking sector.
Renneboog, L., ter Horst, J., & Zhang, C. (2008). Socially responsible invest- Journal of Business Ethics, 104(1), 133–148. https://doi.org/10.1007/
ments: Institutional aspects, performance, and investor behavior. s10551‐011‐0894‐x
Journal of Banking & Finance, 32(9), 1723–1742. https://doi.org/ Stubbs, W., & Cocklin, C. (2008). Conceptualizing a “sustainability business
10.1016/j.jbankfin.2007.12.039 model”. Organization & Environment, 21(2), 103–127. https://doi.org/
Roy, A. D. (1952). Safety first and the holding of assets. Econometrica: 10.1177/1086026608318042
Journal of the Econometric Society, 20(3), 431–449. https://doi.org/ Tang, Z., Hull, C. E., & Rothenberg, S. (2012). How corporate social respon-
10.2307/1907413 sibility engagement strategy moderates the CSR‐financial performance
Ruiz, B., García, J. A., & Revilla, A. J. (2016). Antecedents and consequences relationship. Journal of Management Studies, 49(7), 1274–1303. https://
of bank reputation: A comparison of the United Kingdom and Spain. doi.org/10.1111/j.1467‐6486.2012.01068.x
International Marketing Review, 33(6), 781–805. https://doi.org/ Thompson, P., & Cowton, C. J. (2004). Bringing the environment into bank
10.1108/IMR‐06‐2015‐0147 lending: Implications for environmental reporting. The British Account-
Sacconi, L. (2000). The social contract of the firm. Economics. In Ethics and ing Review, 36(2), 197–218. https://doi.org/10.1016/j.bar.2003.
organisation. Berlin: Springer Verlag. 11.005

Said, R., Hj Zainuddin, Y., & Haron, H. (2009). The relationship between Trapp, N. L. (2014). Stakeholder involvement in CSR strategy‐making?
corporate social responsibility disclosure and corporate governance Clues from sixteen Danish companies. Public Relations Review, 40(1),
characteristics in Malaysian public listed companies. Social Responsibil- 42–49. https://doi.org/10.1016/j.pubrev.2013.11.005
ity Journal, 5(2), 212–226. https://doi.org/10.1108/17471110 United Nations Environment Programme Finance Initiative (UNEP FI),
910964496 2013. http://www.unepfi.org/2013/10/
Sanders, W. G. (2001). Behavioral responses of CEOs to stock ownership Van Marrewijk, M. (2003). Concepts and definitions of CSR and corpo-
and stock option pay. Academy of 7Management Journal, 44(3), rate sustainability: Between agency and communion. Journal of
477–492. Business Ethics, 44(2–3), 95–105. https://doi.org/10.1023/A:102333
Santis, P., Albuquerque, A., & Lizarelli, F. (2016). Do sustainable companies 1212247
have a better financial performance? A study on Brazilian public com- Venturelli, A., Cosma, S., & Leopizzi, R. (2018). Stakeholder engagement:
panies. Journal of Cleaner Production, 133, 735–745. https://doi.org/ An evaluation of European banks. Corporate Social Responsibility and
10.1016/j.jclepro.2016.05.180 Environmental Management., 25, 690–703. https://doi.org/10.1002/
Sapienza, P., Zingales, L., & Maestripieri, D. (2009). Gender differences in csr.1486
financial risk aversion and career choices are affected by testosterone. Walls, J. L., Berrone, P., & Phan, P. H. (2012). Corporate governance and
Proceedings of the National Academy of Sciences, 106(36), environmental performance: Is there really a link? Strategic Manage-
15268–15273. ment Journal, 33(8), 885–913. https://doi.org/10.1002/smj.1952
Scholtens, B. (2009). Corporate social responsibility in the international Weber, O. (2012). Environmental credit risk management in banks and
banking industry. Journal of Business Ethics, 86(2), 159–175. https:// financial service institutions. Business Strategy and the Environment,
doi.org/10.1007/s10551‐008‐9841‐x 21(4), 248–263. https://doi.org/10.1002/bse.737
Scholtens, B., & Dam, L. (2007). Banking on the equator. Are banks that Wu, M. W., & Shen, C. H. (2013). Corporate social responsibility in the
adopted the equator principles different from non‐adopted? banking industry: Motives and financial performance. Journal of Banking
World Development, 35(8), 1307–1328. https://doi.org/10.1016/j. & Finance, 37(9), 3529–3547. https://doi.org/10.1016/j.jbankfin.2013.
worlddev.2006.10.013 04.023
Schücking, H. (2013). Banking on coal. Urgewald, BankTrack, CEE Wu, M. W., Shen, C. H., & Chen, T. H. (2017). Application of multi‐level
Bankwatch Network and Polska Zielona Sieć with Rainforest Action matching between financial performance and corporate social
Network, World Development Movement, and PowerShift. https:// responsibility in the banking industry. Review of Quantitative Finance
www.banktrack.org/download/banking_on_coal/banking_on_coal_ and Accounting, 49(1), 29–63. https://doi.org/10.1007/s11156‐016‐
updated.pdf 0582‐0
18 GANGI ET AL.

Yang, D., Wang, A. X., Zhou, K. Z., & Jiang, W. (2018). Environmental strat-
egy, institutional force, and innovation capability: A managerial How to cite this article: Gangi F, Meles A, D'Angelo E,
cognition perspective. Journal of Business Ethics, 1–15. https://doi.
org/10.1007/s10551‐018‐3830‐5 Daniele LM. Sustainable development and corporate gover-
nance in the financial system: Are environmentally friendly
Zhang, L. (2012). Board demographic diversity, independence, and corpo-
rate social performance. Corporate Governance: The International banks less risky?. Corp Soc Resp Env Ma. 2018;1–19. https://
Journal of Business in Society, 12(5), 686–700. doi.org/10.1002/csr.1699

APPENDIX A

TABLE A1 Description of variables

Variable Symbol Description
Bank risk Z‐score The ratio is a measure of overall banking risk. It is estimated as the sum of the current period t ROA and the
equity ratio divided by the standard deviation of ROA.
Environmental ENV Thomson Reuters ENV score. It is an overall company score based on self‐reported information in the
engagement indicator environmental governance pillar.
Board size B_size Natural logarithm of the number of board directors
Board independence B_ind The number of independent directors divided by the number of total board directors
Board diversity B_GD The number of female directors divided by the number of total board directors
CEO power CEO_power Dummy variable equal to 1 if a CEO simultaneously chairs the board or if the chairman of the board has been
the CEO of the company
CEO compensation CEO_comp Dummy variable equal to 1 if the CEO's compensation is linked to total shareholder returns
Size TA Natural logarithm of total assets
Cost efficiency CostIncome The ratio between overhead cost and total income
Coverage Coverage The ratio between loan loss reserve and gross loans
Specialization LoansAssets The ratio between loans and total assets
Market concentration HHI The Herfindahl–Hirschman Index
GDP per capita GDPper GDP based on current price/population

Note. GDP: gross domestic product. Source: Thomson Reuters, Thomson Datastream, and the World Bank.

TABLE A2 Abstracts of banks corporate social responsibility reports on environmental engagement (all the reported banks belong to the study
sample)
Bank Country Disclosure on environmental engagement (abstracts)

Intesa Sanpaolo Italy “… The Intesa Sanpaolo Group has been committed to reducing its emissions through: the purchase of
electricity from renewable sources; the reduction of consumption related to lighting and air conditioning, the
choice of low energy impact work tools, plant maintenance, waste recycling, use of recycled products, etc.;
responsible management of environmental and energy impacts; the promotion of products and services
aimed at environmental protection; the assessment of environmental risks related to loans to client
companies; the collaboration with international organizations for the development and application of
sustainability guidelines to the financial sector; the commitment and projects to reduce the emissions …” (for
details, see https://www.group.intesasanpaolo.com/scriptIsir0/si09/sostenibilita)
HSBC United “… HSBC plays a key part in supporting the transition to a low‐carbon economy, both via our own operations
Kingdom and by helping customers to reduce their own emissions. We believe it is important to lead by example. For
this reason, we have 10 goals to help us achieve a reduction in our environmental impact …” (for details, see
https://www.hsbc.com/our‐approach/sustainability/operations)
BNP Paribas France “… to combat climate change, in 2017 BNP Paribas became a carbon‐neutral bank for its operations through
three complementary actions: reduction of its CO2 emissions; use of low‐carbon electricity; counteracting
emissions that cannot be reduced. Financing for renewable energies: in 2017, the amount of financing in this
sector was €12.3 billion (compared with €9.3 billion in 2016). In total, in 2017 the Group participated in
financing or advising renewable‐energy projects totalling more than 6 GW of installed capacity. Carbon
content of kWh financed by the Group: BNP Paribas has committed to ensuring that the average carbon
content of each kWh of power it finances declines at the same rate as the global average according to the
International Energy Agency's (IEA) 450 scenario—a reduction of 85% between 2015 and 2040—which
means, for example, reaching the milestone of 350 g CO2/kWh in 2020. The kWh carbon content financed
by the Group is 342 g of CO2, compared with the world average of 544 in 2014 …” (for details, see https://
group.bnpparibas/en/group/corporate‐social‐responsibility)
Wells Fargo United “…. Our priority: Accelerate the transition to a lowcarbon economy and help reduce the impacts of climate
& Co. States change on our customers and communities … In 2017, we continued to strengthen our business processes to

(Continues)
GANGI ET AL. 19

TABLE A2 (Continued)

Bank Country Disclosure on environmental engagement (abstracts)

improve our environmental performance and build sustainability more deeply into our business, culture,
operations, and community engagement. To help guide responsible financing in our lines of business, we are
signatories of the Equator Principles, Green Bond Principles, and the UN Principles for Responsible Investing.
In 2017, we met 100% of our global electricity needs with renewable energy and we achieved our 2020
carbon‐reduction goal of 45 percent. As of 2017, 28 percent of our total square footage in leased and owned
buildings is LEED certified, keeping us on pace to achieve 35 percent by 2020 … Sustainable financing
includes: …Energy Efficiency … Sustainable Water Management … Pollution Prevention … Environmentally
Responsible Products … Other Green Business.” (for details, see https://www.wellsfargo.com/about/
corporate‐responsibility/environment)

Note. HSBC: Hongkong and Shanghai Banking Corporation.