Artikel 6

J. Int. Financ. Markets Inst.
Money 52 (2018) 240–261
Contents lists available at ScienceDirect
Journal of International Financial

Markets, Institutions & Money
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / i n t fi n
Environmental risk management and financial performance in

the banking industry: A cross-country comparison q
Maya Finger a,b,1, Ilanit Gavious a,1,⇑, Ronny Manos b,1
a
Guilford Glazer Faculty of Business and Management, Department of Business Administration, Ben-Gurion University of the Negev, PO Box 653, Beer-Sheva
84105, Israel
b
The School of Business Administration, The College of Management Academic Studies, 7 Yitzhak Rabin Blvd., Rishon Lezion 75190, Israel
a r t i c l e i n f o a b s t r a c t
Article history: The Equator Principles (EP) provide banks with environmental guidelines for project
Received 2 July 2017 finance. Distinguishing between banks from developed and developing countries, we anal-
Accepted 14 September 2017 yse the effect of EP adoption on performance. Two sets of hypotheses for each sub-sample
Available online 18 September 2017
of banks are developed and tested using comparison analyses, event-study methodology
and two-stage selection modelling. We find that in developed (developing) countries, EP
JEL classification: adoption is associated with an increase (decrease) in funding activity and in the share of
G21
income from interest. These results indicate that EP adoption is a strategic decision for
M14
banks in developing countries, and a form of greenwashing in developed countries.
Keywords: Ó 2017 Elsevier B.V. All rights reserved.
Corporate social responsibility
Cross-country
Environmental risk management
Equator principles
Financial institutions
Financial performance
1. Introduction
Banks and other financial institutions are exposed to various financial risks related to their business of lending (e.g.,
credit, counterparty and interest rate risks). These risks are well-known and thus financial models have been developed
to evaluate, manage and hedge their impact on business (see, e.g., Freixas and Rochet, 1998; Shen, 2002). More recently,
there has been a growing awareness of social, environmental and sustainability hazards as playing a role in the overall risks
to which financial institutions are exposed (e.g., Bing et al., 2011; Nandy and Lodh, 2012). From this perspective, banks are no
different from other firms. Indeed, the social and environmental responsibilities of firms – generally referred to as Corporate
q
This paper was presented at the 2016 Cross Country Perspectives of Finance conferences held in Taiyuan and Pu’er, China. The authors thank the Editors
Gady Jacoby and Zhenyu Wu as well as two anonymous referees. We have also benefited from the comments of Diane Romm, Israel Drori, workshop
participants at Guilford Glazer Faculty of Business and Management, Ben-Gurion University, Israel as well as the Journal of International Financial Markets,
Institutions and Money Special Issue Conferences and Forum – Financial Innovation in Yunnan Province conference on Cross Country Issues on Credit,
Banking, Asset Pricing, and Market Liquidity, China. We gratefully acknowledge the financial support of the Guilford Glazer School of Business and
Management at Ben-Gurion University and of The School of Business Administration, The College of Management Academic Studies. All errors remain our
responsibility.
⇑ Corresponding author.
E-mail addresses: mayafinger1@gmail.com (M. Finger), madaril@bgu.ac.il (I. Gavious), rmanos@colman.ac.il (R. Manos).
1
Equal authorship.
http://dx.doi.org/10.1016/j.intfin.2017.09.019
1042-4431/Ó 2017 Elsevier B.V. All rights reserved.
M. Finger et al. / J. Int. Financ. Markets Inst. Money 52 (2018) 240–261 241
Social Responsibility (CSR) – is the topic of a growing literature. (See for example, Margolis and Walsh, 2003; or the meta-
analysis studies by Margolis et al., 2007; Orlitzky et al., 2003; and Wang et al., 2015). Some studies focus on CSR within the
banking industry. For example, Chih et al. (2010) study the determinants of banks adopting CSR. Scholtens (2009) provides a
framework for assessing CSR by international banks, and Cuesta-González et al. (2006) examine the social performance of
Spanish banks.
This study investigates the relationship between corporate environmental responsibility and financial performance
within the financial sector, comparing between banks from developed and developing countries. Comparison literature on
the effect of implementing CSR in general – and environmental responsibility in particular – on firm performance in devel-
oped versus developing countries, is limited. This is particularly true in the banking industry. Indeed, recent literature on CSR
note that little is known about social responsibility practice in developing countries and highlight the need for more research
in this area, given the different socio-economic context in these countries (e.g. Jamali and Mirshak, 2007; Jones, 1999; Belal,
2001). We aim to fill this gap in the literature.
Our comparison study of the relationship between CSR and bank performance focuses on the Equator Principles (EP), a set
of best practice principles to guide banks that are involved in project finance, on social and environmental risk management
(see Macve and Chen, 2010; Scholtens and Dam, 2007). Since their inception in 2003, banks from developed and developing
countries have been gradually adopting the EP, and in this study, we analyse the effect of EP adoption on their performance.
Using data over the period 2003–2015 on EP-adopters worldwide, we start by showing that the characteristics of the coun-
tries and of the banks which adopt the EP, differ significantly in developed and developing countries. This reinforces the call
by Wu and Shen (2013) to control for regional differences. Thus, our theory development, hypotheses and empirical inves-
tigation control for the development stage of the country of origin of our sample banks.
Specifically, we ask three questions relating to the effect of EP adoption on bank performance in developed and develop-
ing countries: what type of banks adopt the EP; how does the market react to the adoption; and what are the short-term and
long-term performance effects of adopting the EP. These three questions are formulated into two sets of hypotheses, one for
developed and one for developing countries. We argue that in most developed countries, awareness of environmental issues
is well established and the legislation systems, institutional capacity, and country and corporate governance are designed to
protect society and the natural environment.2 Adopting the EP is a form of window-dressing or greenwashing in the sense that
on the whole banks already adequately address environmental concerns. In contrast, in developing countries the awareness to –
and regulation of environmental issues is not as well established as in developed markets (Jamali and Mirshak, 2007; Chapple
and Moon, 2005; Dögl and Behnam, 2015). This means that adopting the EP is a strategic choice which is expected to require
costly investments and to change the adopters’ operations and performance. Based on this rationale, we develop hypotheses to
answer the three questions set above, to which we empirically seek answers.
The empirical procedure includes comparison analyses, event study and multivariate analyses using the two-stage Heck-
man selection model procedure (Heckman, 1979). Our measures of bank performance include five alternative measures that
are particularly relevant to our setting: funding activities; net interest income; return on equity; return on assets; and non-
performing loans. We also estimate the cumulative abnormal returns around EP adoption, to assess the market reaction to
this event. The results of our empirical procedures, indicate that EP-banks from developed countries are different from those
from developing countries. First, while banks from developed countries that opt for EP adoption, tend to be of poor quality
relative to non-adopters, in developing countries it is high quality banks that opt for EP adoption. Second, while there is no
significant reaction to EP-adoption in developed countries, in developing countries the reaction is positive and significant.
Third, while in developed countries the change in performance following EP adoption is an increase in the bank’s funding
activity and fraction of interest income, in developing countries the change is a decrease in the bank’s funding activity
and fraction of interest income. We explain these results in terms of our hypotheses and the underlying motivation to adopt
CSR.
The contribution of this paper is fourfold. First, our study focus on performance in the banking sector. Understanding
this sector is important given that often banks are excluded from samples in empirical work due to their special charac-
teristics. For example, banks have different reporting incentives, accounting requirements, and risk exposures compared to
other industries. They also operate under unique regulatory codes. As such, studies in finance and accounting generally
investigate financial institutions separately from other industries (e.g. Abreu and Gulamhussen, 2013; Riedl and
Serafeim, 2011). Second, we investigate the CSR/financial performance relationship in the banking sector. As noted by
Wu and Shen (2013), banks use public resources paid for by society and are often criticized for exploiting naïve or trusting
customers to generate short-term profits and inflated bonuses. Thus, studying CSR in the banking industry can help inform
this debate.
Third, our study contributes to previous studies of CSR in the banking sector, in that we focus on a contemporary proxy for
CSR behaviour – environmental protection – rather than relying on aggregate CSR indices. Previous studies, such as Simpson
and Kohers (2002) and Wu and Shen (2013), investigate general CSR performance in the banking industry using published
ratings of social performance. Wu and Shen (2013) refer to the weakness of this approach of measuring CSR, citing
Chatterji and Levine (2008) and Chatterji et al. (2010) who ‘‘. . .argue that the organizations rating the social performance
of enterprises cannot truly discern which firms are socially responsible, resulting in metrics that are often invalid and can
be misleading to stakeholders” (Wu and Shen, 2013, p. 3531). Moreover, Wu and Shen (2013) call for studies that focus
2
Equator Principles III, June 2013 (www.equator-principles.com).
242 M. Finger et al. / J. Int. Financ. Markets Inst. Money 52 (2018) 240–261
on specific aspects of CSR – such as governance, social or environmental – to deepen our understanding of the CSR/financial
performance relationship. The current study does not depend on an aggregate CSR index or the accuracy of published CSR
ratings. Instead, we focus on environmental performance and use a direct measure – adopting the EP, which involves com-
mitment to forgo the financing of large projects that do not meet predetermined standards. Indeed, while the EP were crit-
icized for not going far enough in the direction of achieving sustainable development (Watchman, 2005; Watchman et al.,
2006), there is more recent evidence to suggest that banks which sign the EP are rated significantly higher than those that
do not (Scholtens and Dam, 2007).
Fourth, we conduct a cross-country comparison, differentiating between developed and developing countries. This is in
line with McWilliams et al. (2006a,b) who point to the lack of research on CSR that has a global perspective, as much of the
extant literature is US or Europe based. Moreover, we show that the relationship between CSR and financial performance
differ between firms in developed and developing countries, and explain these differences in terms of the motivation to
engage in CSR (Baron, 2001, 2009; Wu and Shen, 2013; McWilliams et al., 2006a,b). For example, the adoption of CSR by
companies in developing countries, particularly banks in these countries, has been criticized as ‘‘greenwashing” (see, e.g.,
BankTrack, 2005). We show this not be the case, at least with relation to EP-adoption.
The remainder of the paper is organized as follows. Section 2 provides background on the EP while Section 3 summarizes
the relevant literature and develops the hypotheses. Section 4 describes the sample and methods and provides descriptive
statistics. Section 5 presents the empirical tests and discusses the results. Finally, Section 6 summarizes the key findings and
concludes.
2. The Equator Principles (EP)3
The EP are a set of principles that provide a benchmark and a risk management framework for financial institutions that
finance large scale infrastructure and industrial projects4 (see Macve and Chen, 2010; Scholtens and Dam, 2007). Adopting the
EP is voluntary, and the adopting financial institutions are referred to as Equator Principles Financial Institutions (EPFIs).5 EPFIs
use the EP to guide them in determining, assessing and managing environmental and social risks in projects that can have a
significant impact on people and the environment. The aim of the EP is to ensure that projects financed or advised upon by
EPFIs, follow good environmental management practices and are executed in a socially responsible manner.
The EP cover four financial products: project financing, project financing advisory services, project-related corporate loans
and bridge loans. They were initially launched on June 4, 2003, when eight banks from developed countries signed up and
became members of the EP. The Principles have been updated since, most recently on June 4, 2013 (EP III). The ten principles
that make up the EP are based on the International Finance Corporation Performance Standards (IFCPS) and World Bank
Group Environmental, Health, and Safety Guidelines (EHSG). As EPFIs operate in diverse locations, the EP classifies countries
on the basis of their environmental and social governance. Specifically, countries with robust environmental and social gov-
ernance are classified as designated countries while other countries are classified as non-designated. Based on this classifi-
cation, only projects located in non-designated countries need to refer to the applicable IFCPS and EHSG. For projects located
in designated countries (mainly developed countries such as the US, Canada, Western Europe, Japan, etc.), it is assumed that
the legislation and institutional frameworks relating to protection of people and the natural environment, exceed the
requirements of the IFCPS and EHSG. Thus, compliance with local or national law in designated countries is considered to
be an acceptable substitute for the IFCPS and EHSG.
The body through which EPFIs are organized is the EP Association, an unincorporated organization that was formed on
July 1, 2010. The Association is responsible for the management, administration and development of the EP, which it coor-
dinates through a 12-member steering committee and working groups. While initially only eight financial institutions from
developed countries adopted the EP, by the end of its first year (2003) the EP framework included 14 EPFIs, all from devel-
oped countries. However, since 2004 more and more financial institutions from developing counties have joined the EP, and
today, there are 89 EPFIs in 36 countries. These institutions are committed to the EP and will not finance or otherwise sup-
port projects that do not comply with these principles. Moreover, according to the UN System Task Team, the EP cover over
70% of international project financing debt in emerging markets (UNTT, 2013).
3. Literature review and hypotheses development
Corporate social responsibility (CSR) has recently been gaining the attention of the business community. Companies’
reporting on ethical, social and environmental practices is regularly published in detail and is closely monitored by both
firms’ immediate stakeholders and the public at large. Research on CSR has also been growing (e.g. Jo and Kim, 2008;
Renneboog et al., 2008; Kempf and Osthoff, 2008; Schroder, 2007). In particular, scholarly interest in the relationship
between firm performance and its social impact and costs, has led to a view of CSR as a tool for aligning corporate profits
3
This section is based on the information published by the EP Association at: http://www.equator-principles.com/.
4
These are infrastructure or industrial projects that exceed US$10 million. The financing is often via corporate loans, although the EP also guide financial
institutions that provide advice on large scale projects that meet certain criteria.
5
From here on, we use ‘‘EPFIs” interchangeably with ‘‘EP-banks” or ‘‘EP-adopters”, to refer to Equator Principles Financial Institutions.
and social goals (Geva, 2008; Heal, 2005). For example, Heal (2005) argues that CSR helps businesses to manage conflicts
with society, regardless of whether these conflicts are due to differences between the private and social costs of corporate
activity (e.g. pollution) or to discrepancies over distributional issues (e.g. exploitation of employees).
Numerous studies have investigated the impact of implementing CSR, on financial performance (e.g. Allouche and
Laroche, 2006; Cochran and Wood, 1984; Lockett et al., 2006; Scholtens, 2008). One important insight to emerge from this
literature is the lack of consensus regarding the definition of CSR (Kakabadse et al., 2005; Frankental, 2001). Also lacking con-
sensus, is an answer to the question of what are the implications – to firms and to society – of adopting CSR (see for example
Silberhorn and Warren, 2007; McWilliams and Siegel, 2001; McWilliams and Siegel, 1997; de-los Ángeles et al., 2007). One
view is that implementing CSR is a complex, bureaucratic, expensive and time-consuming process such that the economic
performance of firms that decide to act responsibly should deteriorate (Friedman, 1970). Consequently, CSR initiatives wor-
sen the firm’s ability to effectively compete in the market with other firms which are less CSR inclined (Aupperle et al., 1985).
Empirical evidence to support this view is provided for example by Lopez et al. (2007) who investigate the short term finan-
cial performance of European firms engaged in CSR. For the period 2002–2004 a negative and significant relation is recorded
between CSR and performance, which the authors explain in terms of CSR-related costs such as on training or maintaining
product quality and safety. Scholtens and Dam (2007) investigate the adoption by banks of the Equator Principles (EP), a vol-
untary set of guidelines for promoting social and environmental responsibility in project finance. The authors find that there
are real costs involved in adopting the EP. For example, EP adoption requires investment in screening, monitoring and pro-
cedures for periodic reporting as well as having to decline profitable projects. Furthermore, the diversion of resources
towards CSR projects may clash with the firm’s core activity and expertise. Similarly, lack of knowledge in implementing
CSR projects could hurt the firm’s financial performance (Preston and O’Bannon, 1997).
However, CSR and social responsible investing are becoming increasingly more popular in recent years (e.g., Jo and Kim,
2008; Renneboog et al., 2008). The question thus arises as to why firms adopt CSR policies and programmes if such behaviour
destroys shareholder value. The literature discusses three motives that drive firms to adopt CSR (see Baron, 2001, 2009; Wu
and Shen, 2013; McWilliams et al., 2006a,b): strategic, altruistic and greenwashing. First, strategic CSR is based on the idea
that CSR creates a competitive advantage which will improve the firm’s long-term performance and its sustainability (e.g.
Orlitzky et al., 2011). Thus, McWilliams and Siegel (2001) contend that managers should consider CSR decisions precisely
as they treat all other investment decisions. For example, Heal (2005) argues that using CSR to manage potential conflicts
with society, firms can improve their profits and guard against reputational risks. McWilliams et al. (2006a,b) discuss other
ways by which strategic CSR create competitive advantage, including attracting green investors and employees who are will-
ing to accept lower salaries from socially responsible firms. Scholtens and Dam (2007) investigate the adoption by banks of
the EP, and find that banks which adopted the principles were significantly larger than non-adopters. They conclude that
adoption of the EP is an act of CSR which signals the responsible conduct of adopters, and that for large banks, the reputa-
tional benefits to be gained from adopting the EP, outweigh the costs of adoption.6 This ties in with the general literature
about the importance of reputation to firm value and performance. For example, Brønn and Vrioni (2001) explain that the asso-
ciation between CSR and firm performance is by and large through reputation. Sanchez and Sotorrio (2007) argue that firms
create value from their reputation of being socially responsible. Using Spanish companies, they find positive and significant rela-
tionship (albeit non-linear) between financial and social performance. This is in line with a view of CSR as being strategically
motivated. Indeed, based on an overview of recent empirical evidence, Orlitzky et al. (2011) conclude that economic theories of
strategic CSR have the greatest potential for explaining the relationship between CSR and firm performance. More specifically to
the banking industry, Wu and Shen (2013) find that strategic choices provide the primary motivation for banks to engage in CSR.
The second motive that may drive firms to adopt CSR is altruistic or morally-motivated CSR (Baron, 2001). It asserts that
firms conduct CSR to create social value – that is value for all stakeholders in the firm rather than focusing on shareholders
alone (Renneboog et al., 2008). Given that altruistic CSR is not specifically concerned with financial performance, Wu and
Shen (2013) argue that when CSR is morally-driven rather than being a strategic decision, implementation costs negatively
affect financial performance, and there are no financial related benefits that can offset this effect. However, to the extent that
altruistic CSR aims to meet the needs of key stakeholders, it should also contribute to the survival and long-term perfor-
mance of the firm. For example, engaging in altruistic CSR may contribute to the firm’s legitimacy and leads to its endorse-
ment by stakeholders with direct bearing on its economic position (Orlitzky et al., 2003). Sethi (1979) suggests that some
firms may adopt CSR norms in their search for legitimacy by various stakeholders, even when these are detrimental to short
term profits. Thus, Baron (2001) notes that altruistic CSR may be indistinguishable from strategic CSR.
The third motive that drive firms to adopt CSR is greenwashing (Wu and Shen, 2013). Firms that engage in greenwashing
will present themselves, their products and policies as environmentally-concerned while not actually committing to the
cause. Dam et al. (2009) note that committing to CSR should result in substantial costs. Thus, if no cost differences are
observed between CSR adopters and other firms, then it must be that those declaring to be socially responsible, are in effect
engaged in greenwashing. Frankental (2001) observes that CSR amounts to a mere PR exercise if it does not result in a real
change in the way the firm is operating or in the way it is governed. Wu and Shen (2013) describe greenwashing as aiming to
enhance corporate image without significantly changing the firm’s operations.
6
The importance of bank reputation in mitigating information asymmetries and increasing bank profitability has also been investigated by Bushman and
Wittenberg-Moerman (2012). These authors study the role of bank reputation in the context of syndicated loans, and find that higher reputation is associated
with higher profitability and credit quality.
The different motives that drive firms to engage in CSR may be explained in terms of differences in the socio-economic
and institutional environment. For example, Dögl and Behnam (2015) find that the link between corporate environmental
responsibility and positive business outcomes is stronger in developing countries (India and China) than in developed ones
(Germany and the US). Jamali and Mirshak (2007) examine eight Lebanese firms that are actively engaged in CSR, and find
their CSR approach to be unsystematic and unfocused. The authors conclude that in Lebanon CSR is still associated mainly
with philanthropy, and highlight the value to be added from further exploring CSR conceptions and perceptions in develop-
ing countries. Belal (2001) conducts a survey of corporate social reporting practices in Bangladesh. The study explores the
socio-political and economic context in which CSR disclosures take place, and stresses the need for more research on CSR
in developing countries, given the socio-economic context. Jones (1999) discusses the importance of the socio-cultural envi-
ronment and level of national economic development in explaining CSR practice, while Chapple and Moon (2005) show that
CSR in Asian countries is different to that in Western countries.
Turning to the relationship between CSR and financial performance in the banking sector, Wu and Shen (2013) focus on
environmental performance of mainly European banks, noting that regional differences may explain inconsistencies
between their findings and previous work. For example, while Soana (2011) finds insignificant relationship between the
social and financial performance of international and Italian banks, Simpson and Kohers (2002) find this relationship to
be significantly positive in the case of US banks. These contradicting findings for banks in different regions, is likely related
to restrictions on banking activities, control over corruption, sector development and other macroeconomic factors which
vary across countries at differing levels of economic development. Indeed, these factors have been shown to influence bank
performance (e.g., Beck et al., 2010; Shen and Lee, 2005; Shen and Lin, 2012; Wu and Shen, 2011, 2013). Given the above
discussion, we consider the socio-economic and institutional differences of the countries in which EP adopters operate.
Our hypotheses are thus developed separately for banks in developed and developing countries, starting with developed
countries.
In developed countries awareness of environmental issues is well established and has been institutionalized, so that it is
taken for granted as an integral part of the firm activity (Scott, 2001). Consequently, most firms have in place mechanisms
and procedures to ensure environmental risks are managed. Indeed, the EP requirements generally recognize EPFIs from
developed countries as designated countries, i.e. countries with robust environmental and social governance, legislation sys-
tems and institutional capacity. Accordingly, Principle 3 of the EP (Equators Principles III, 2013) states that projects located in
these countries need to comply with host country laws rather than with the IFC Performance Standards on Environmental
and Social Sustainability (Performance Standards) and the World Bank Group Environmental, Health and Safety Guidelines
(EHS Guidelines). This means that for most banks from developed countries, adopting the EP is not expected to: (a) involve
costly investments; (b) dramatically change their operations; or (c) impact performance. Thus, with respect to banks from
developed countries, we hypothesize that adopting the EP is a form of window dressing in the sense that on the whole banks
already adequately address environmental concerns. Adopting the EP is the expected step to be taken by banks involved in
project finance. It is used to boost or maintain reputation, and thus is taken up mainly by risky or poor quality banks or by
large banks with high exposure. The latter is in line with Scholtens and Dam (2007) who argue that it is more likely that large
banks will adopt the EP, because for larger firms with more exposure, maintaining reputation is particularly important.7 Con-
sequently, we formulate Hypothesis 1, relating to EPFIs from developed countries:
Hypothesis 1.
H1(a). Prior to EP-adoption: Banks that opt for EP adoption, are large, risky, and poorer performers, compared with non-
adopters.
H1(b). Share price reaction: EP adoption is ‘business as usual’, not requiring extra commitment or investment. Therefore,
the share price reaction is insignificant.
H1(c). Performance effects: EP adoption is ‘greenwashing’ because the regulatory, institutional, and governance systems
are robust in protecting people and the natural environment. Therefore, there are no short-term or long-term perfor-
mance effects to EP adoption.
Developing countries are substantially different from developed countries. For example, although financial market regula-
tion is often stricter, awareness and regulation of environmental issues is not as well established as in developed markets
(Jamali and Mirshak, 2007; Chapple and Moon, 2005). Dögl and Behnam (2015) note that upholding high standards of social
responsibility by firms in developing countries, is typically challenging because the regulatory infrastructure and general
practices are not always sufficiently effective in controlling environmental and social risks. Thus, many firms in developing
countries do not have in place mechanisms and procedures to deal with environmental and social risks. This can explain why
the EP require the financing of big projects in non-designated countries to comply with the IFC Performance Standards and
the EHS Guidelines, rather than with local law (Equators Principles III, 2013). Thus, adopting the EP is expected to: (a) require
costly investments; (b) dramatically change the adopters’ operations; and (c) impact performance. If the drive to adopt the
7
See also Ioannou and Serafeim (2010) who find that firms with higher visibility receive more favorable security analysts’ recommendations for their CSR
strategies.
EP is strategic and efficiently executed, then it should lead to improved performance.8 For example, as EP require screening of
borrowers, adoption may lead to a reduction in the growth rate of EPFIs’ loan portfolios and hence reduce income from interest,
at least in the short to medium term. However, the impact on profitability may be positive, negative or none. This depends on
the balance between the additional costs and revenues relating to the adoption of the EP. Likewise, if adoption is driven by altru-
istic motives, which will eventually create reputation, it may also lead to improved long-term performance. Moreover, in devel-
oping countries, banks that choose to adopt the EP, are likely to be large and of high quality. This is due to the high
implementation costs which large, fast-growing, better performers will find easier to finance. Thus, in developing countries,
EP adoption is driven by strategic (or altruistic) motives, and involves high implementation costs. Consequently, we formulate
Hypothesis 2, relating to EPFIs from developing countries:
Hypothesis 2.
H2(a). Prior to EP-adoption: Banks that opt for EP adoption are large, fast-growing, and better performers, compared with
non-adopters.
H2(b). Share price reaction: EP adoption requires commitment and investment to create value. Therefore, the share price
reaction is positive and significant.
H2(c). Performance effects: EP adoption is a strategic decision with short-term and long-term effects on performance, the
direction of which depends on the balance between the costs and benefits to accrue from EP adoption. But EP-adopters
should experience a drop in their funding activity.
4. Sample and methods
Our sample selection procedure begins with all financial institutions from 34 countries in which the EP were adopted by
at least one EPFI during the period from 2003, when the EP were initiated, and until the end of 2015. 80 banks have signed
the EP during the sample period, of which we exclude those which signed the EP after 2013, due to insufficient information
about their long-term post-EP adoption performance. Our sample of EPFIs thus includes 78 banks which are listed in Appen-
dix A by name, country and year of adoption. Table 1 displays the distribution of the EPFIs by country. Of the 78 EPFIs, 52
(67%) are from developed countries and 26 (33%) are from developing countries. Of the banks from developed countries, 27
(35%) are from Northern and Western Europe, 12 (15%) from North America, 6 (8%) from the UK, 4 (5%) from Australia and 3
(4%) from Japan. Banks from developing countries include 14 (18%) from South America, 6 (8%) from countries in Asia exclud-
ing Japan, and 6 (8%) from Africa. Comparing the various regions, most banks signing the EP are from Northern and Western
Europe (notably, none of the adopters is from Eastern Europe). This high prevalence of European banks, compared to other
continents, may be related to the fact that the initiation of the EP was in Europe, specifically by the Dutch bank ABN AMRO.
This conjecture is supported by a statement made in 2014 by the CEO of a bank in the Netherlands, Eksportkreditt Norge:
‘‘Many of the international banks we work closely with have adopted the Equator Principles, and the initiative has contributed
in raising awareness on social and environmental issues. We therefore see it as a natural step for us to join the initiative.”9
While the highest prevalence of EPFIs is from Europe, the largest EPFIs come from Asia (Japan), North America (the US and
Canada) and Oceania (Australia) as displayed in Table 2. Notwithstanding these observations, the EP have reached beyond
these developed regions, as reflected in the fact that about a third of EPFIs come from developing countries. Moreover, EPFIs
from developing countries come from different continents (Asia, Africa, and Latin America) and include major countries (e.g.,
Argentina, Brazil, China, India, Nigeria and South Africa) as well as small/medium countries (e.g. Morocco, Togo, Mauritius,
Chile, Uruguay, and Panama). If we assess the distribution of EPFIs from developing countries in terms of the ease of doing
business in their country of origin, we find that most of them fall into the top half (e.g., Mauritius, Mexico, Peru, Colombia,
Costa Rica, and Uruguay).10 Amongst the countries ranked lower on the 2015 Index for Ease of Doing Business are India, Egypt
and Nigeria.
For each EPFI, information was gathered on a yearly basis for the period beginning in the three years preceding the year of
EP adoption (t 3) and up until 2015, the most recent year for which annual financial statements were available at the time
of the study. This allowed us to analyse the relationship between EP adoption and performance in the short-term as well as
in the long-term (although for recent adopters of the EP, the post EP adoption period is shorter). Specifically, for 73 EPFIs in
our sample, at least five post-adoption years are available, and for 5 EPFIs only four post-adoption years are available. Over-
all, we have 929 bank-year observations of EPFIs with the required data, of which 625 are from developed countries and 304
are from developing countries. We supplement this sample with 6335 bank-year observations of non-EFFIs with required
data.
8
In line with Baron (2001), we do not distinguish between altruistic and strategic CSR. This is because altruistic CSR may create value and improve (long-
term) performance (for example through reputation). We thus use the term strategic CSR to cover both strategic and altruistic CSR, which require investment in
CSR. This is distinguishable from greenwashing motivated CSR which does not involve real commitment or investment.
9
Source: http://www.equator-principles.com/index.php/all-news-media.
10
The Ease of Doing Business Index ranks economies from 1 to 189 based on their ease of doing business. A high score means the regulatory environment is
more conducive to the starting and operating of a local firm. The ranking here relates to the June 2015 index of Ease of Doing Business. (http://
www.doingbusiness.org/rankings).
Table 1
The distribution of 78 EPFIs by country.
Developed countries 52 (67%) Developing countries 26 (33%)

# (%) Banks # (%) Banks
Northern & Western Europe 27 (35%) South America 14 (18%)
Belgium 1 Argentina 1
Denmark 1 Brazil 5
France 4 Chile 1
Germany 4 Colombia 1
Italy 1 Costa Rica 1
Norway 1 Mauritius 1
Portugal 1 Mexico 2
Spain 5 Peru 1
Sweden 2 Uruguay 1
Switzerland 1 Asia 6 (8%)
The Netherlands 6 China 1
North America 12 (15%) Egypt 1
Canada 7 Kingdom of Bahrain 1
US 5 India 1
UK 6 (8%) Morocco 1
Australia 4 (5%) Sultanate of Oman 1
Asia 3 (4%) Africa 6 (8%)
Japan 3 Nigeria 2
South Africa 3
Togo 1
Table 2
The four largest EPFIs measured by real assets as of the signing date. Source: http://www.equator-principles.com/
Bank Country Region Incorporation year Assets in US$ (Billions)

Bank of Tokyo-Mitsubishi UFJ Japan Asia 2005 164,180,681
Bank of America Corporation US North America 2006 1,463,685
Bank of Nova Scotia Canada North America 2006 291,992
ANZ Australia Oceania 2006 246,629
Our dataset includes bank-level and country-level data. To obtain bank-level data, we use the Bloomberg Professional
Database and Bankscope.11 To obtain country-level data, we use the World Bank database and The Worldwide Governance
Indicators (WGI).12
4.1. Performance measures
We have five measures for financial performance: (1) change in funding activity as measured by the annual growth
rate in the value of total loans (LG); (2) net interest income scaled by the sum of net interest and non-interest
income13 (NII); (3) net income scaled by total equity (ROE); (4) net income scaled by total assets (ROA); and (5) non-
performing loans scaled by total loans (NPL). These five measures are particularly relevant to our setting, because the
adoption of the EP implies that the bank will not provide loans to projects that do not comply with the EP. Thus, on
the one hand, the restrictions set by the EP, as well as the costs associated with their implementation, may shrink
the bank’s loan activities and consequently reduce its interest income and profitability. On the other hand, EP adoption
may increase income and profitability through improving market access, avoiding costs associated with negative external-
ities, and/or creating reputational effects (Scholtens and Dam, 2007). Indeed, the positive reputational impact of adopting
corporate social and environmental responsibility has been highlighted in empirical studies of the relationship between
CSR and performance (e.g., Orlitzky et al., 2003).
11
https://www.bloomberg.com/professional/ and http://www.bvdinfo.com/en-gb/home. The Bloomberg database provided all the information required for
our analyses except for overhead cost – a control variable in the performance models – which is missing from Bloomberg. We obtained this item from
Bankscope. Importantly, we compared the data available from these two sources, to make sure that they match. With very few exceptions, data in Bloomberg
matched with that in Bankscope.
12
http://data.worldbank.org/ and http://info.worldbank.org/governance/wgi.
13
Non-interest income includes fee earnings from fiduciary activities, service charges on deposits accounts, trading account gains and fees, as well as
commissions.
4.2. Control variables
We include control variables at the bank level and at the country level based on the banking literature (e.g., Athanasoglou
et al., 2008; DePrince et al., 2011; Ellul and Yerramilli, 2013; Wu and Shen, 2013). Specifically, we include the following
seven bank level control: size (Total assets); Capitalization (equity/total assets); Tier1 ratio (core equity capital/total risk-
weighted assets); loan-to-deposit ratio (Loan/Deposit); Coverage (loan loss reserve/non-performing loans); and overhead
cost-to-income ratio (Cost/Income).
Controls at the country level are included to capture cross country differences in our sample. We include five institutional
factors and two macroeconomic determinants, consistent with Wu and Shen (2013). To capture institutional differences
between countries, we control for the following factors. First, we use the World Bank database to measure the degree of
restrictions on banks, related to activities in securities (Rest_S), insurance (Rest_I) and real estate (Rest_E). These restrictions
measure the degree of regulatory restrictions on banks activities (see Barth et al., 2006, 2008). The level of restrictions on
each activity is measured on a scale from 1 to 4, with 4 representing the greatest level of restrictions. Second, we control
for the degree of corruption (Corrupt), using the WGI as in Kaufman et al. (2009). Third, to capture financial sector develop-
ment (see Brewer et al., 2008; Levine, 1999) we use the World Bank database to obtain the domestic credit to the private
sector as a fraction of GDP (Credit/GDP). Finally, we also control for two macroeconomic determinants, GDP growth and
GDP growth per capita (GDPgr and GDPgr p/c), which are also obtained from the World Bank database.
Table 3 presents descriptive statistics. Panel A of Table 3 splits the sample into developed and developing countries, and
presents statistics for EP banks and for the countries in which they reside. Indeed, as shown in Panel A of Table 3, the two
groups differ significantly (at the 1% level) across all characteristics, at both the bank and country levels (except for Cost/
Income). Specifically, EP banks from developing countries enjoy higher growth in their loan portfolios/funding activity
(LG). This is consistent with higher economic growth as reflected by the macroeconomic variables (GDPgr and GDPgr p/c)
and with a higher potential for growth in the credit market as reflected by a significantly lower Credit/GDP ratio. However,
this higher growth in loan portfolio (or funding activity, LG), comes with higher fraction of non-preforming loans (NPL) for
EPFIs from developing compared to developed countries. The higher NPL for EPFIs from developing countries may be
explained by lower degree of control over debt collection, differences in payment ethics or more corruption, as reflected
by the higher degree of control over corruption (Corrupt) in developed countries. The higher level of NPL could have been
alarming because a greater fraction of the income of EPFIs from developing countries comes from interest-bearing activities
(NII). However, it does not seem to impact their profitability, which is higher compared with EPFIs from developed countries
(ROE and ROA). Moreover, while (as may be expected), bank size, (total assets), is significantly smaller in developing coun-
tries, it is the EPFIs from developed countries that appear to be riskier as indicated by lower Capitalization, Tier1 and Coverage
ratios and a higher Loan/Deposit ratio. It thus appears that EPFIs from developing countries are more profitable while main-
taining lower risk, compared with EPFIs from developed countries. This is consistent with the observation that more of the
EPFIs come from developed than from developing countries, possibly as EP adoption has become an industry norm in devel-
oped countries.14 In contrast, adopting the EP in developing countries is a more recent phenomenon, and banks that do so tend
to be of higher quality.
The institutional and macroeconomic variables also differ significantly between developed and developing countries. As
noted above, developing countries have lower degree of control over corruption (Corrupt), enjoy higher economic growth
(GDPgr and GDPgr p/c) and tend to have less developed credit markets (Credit/GDP). In addition, regulatory restrictions on
the banking industry is higher in developing countries (Rest_S, Rest_I and Rest_E). This is consistent with the findings in
Hafeez (2003) who explains it in terms of the move towards free market reforms which developing countries have been
more slow to adopt.
The results of Table 3, Panel A, indicate that the institutional and macroeconomic environment is significantly different in
developed versus developing countries, as well as the EPFIs characteristics in these countries.15 This reinforces our argument
that empirical investigations of EPFIs must control for whether they come from a developed or developing country. It is also in
line with previous studies which document the significant effects on bank performance, of restrictions on banking activities,
control over corruption, sector development and macroeconomic factors (e.g., Beck et al., 2010; Shen and Lee, 2005; Shen
and Lin, 2012; Wu and Shen, 2011, 2013). We repeat the descriptive analysis of Panel A, Table 3, excluding the years preceding
the EP adoption and obtain similar results. Additionally, we repeat the descriptive analysis excluding the sub-prime crisis years
(2007–2009) and obtain the same qualitative results.16 Importantly, the differences between developed and developing coun-
tries that we document, are robust to economic cycles.
Panel B of Table 3 provides the first examination of the type of banks which opt for EP adoption. Presented are the results
of comparing EP-adopters and non-adopters in the period before the establishment of the EP, i.e. in the period prior to 2003.
Scholtens and Dam (2007) argue that if EP adoption is related to reputation, then it is more likely that large banks will sign
up to the EP, because for larger firms with more exposure, reputation is particularly important. Indeed, we see that in devel-
oped countries, during the period before the establishment of the EP, banks that later adopted the EP, tended to be signif-
icantly larger (total assets) than non-adopters. This is also the case for banks from developing countries although the
14
Scholtens and Dam (2007) argue that if consumers are willing to pay higher prices for products that are produced by socially responsible methods, then CSR
becomes profitable. As this becomes apparent, more and more firms in the industry adopt the practice which eventually becomes ‘business as usual’.
15
Excluding the ratio of cost to income and according to both the means and the medians of these measures.
16
These additional statistics are available from the authors upon request.
Table 3
Descriptive statistics.
Developed countries Developing countries

(# bank-years = 625) (# bank-years = 304)
Mean Median St. Dev Mean Median St. Dev
Panel A: Developed versus developing countries
Characteristics of banks that signed the EP (EPFIs)
LG (%) 9.513 4.473 21.063 16.928*** 14.802*** 24.350
NII (%) 68.098 68.775 14.449 71.336*** 72.881*** 13.533
ROE (%) 10.182 10.983 9.873 16.555*** 16.421*** 10.249
ROA (%) 0.652 0.522 1.040 1.600*** 1.492*** 1.091
NPL (%) 2.851 1.864 3.113 5.156*** 3.773*** 7.931
Total assets ($ billions) 800.036 572.544 725.192 146.765*** 23.574*** 432.257
Capitalization (%) 5.980 4.958 6.895 10.337** 8.638*** 7.794
Tier1 ratio (%) 10.244 10.100 3.925 13.250*** 12.210*** 7.610
Loan/Deposit (%) 119.358 114.073 49.400 101.936*** 92.355*** 34.386
Coverage (%) 102.119 68.540 94.534 201.348*** 87.717*** 663.904
Cost/Income (%) 262.484 222.230 815.917 232.881 218.603 268.612
Institutional factors
Rest_S 1.500 1.000 0.607 1.900*** 2.000*** 0.710
Rest_I 2.140 2.000 0.441 2.610*** 3.000*** 0.729
Rest_E 2.310 2.000 1.192 3.060*** 3.000*** 0.976
Corrupt 1.590 1.616 0.467 0.103*** 0.113*** 0.617
Credit/GDP (%) 134.370 129.469 38.580 52.199*** 52.765*** 27.765
Macroeconomic variables
GDPgr (%) 1.645 2.079 1.962 4.502*** 4.048*** 3.826
GDPgr p/c (%) 0.824 1.332 1.844 2.666*** 2.481*** 3.972

EP-adopters Non-adopters EP-adopters Non-adopters
(# bank-years = 38) (# bank-years = 584) (# bank-years = 18) (# bank-years = 193)
Panel B: EP adopters vs. non-adopters in the period prior to 2003 (means)
Bank characteristics
LG (%) 5.954 32.762* 21.674 14.875
NII (%) 68.999 80.802*** 69.595 73.547
ROE (%) 10.480 7.655 7.686 4.350
ROA (%) 0.624 0.770 5.729 0.081
NPL (%) 2.736 0.764*** 10.759 10.342
Total assets ($ billions) 356.382 15.375*** 54.719 13.995
Capitalization (%) 4.894 8.799*** 8.748 11.422
Tier1 ratio (%) 7.766 11.908*** 9.662 8.370
Loan/Deposit (%) 141.071 101.144 93.877 116.181
Coverage (%) 161.726 145.280 98.360 92.099
Cost/Income (%) 420.661 700.441 299.431 292.881
Pre-adoption (# bank- Post adoption (# bank- Pre-adoption (# bank- Post adoption (# bank-
years = 182) years = 443) years = 142) years = 162)
Panel C: The performance of EPFIs: Pre-adoption versus post-adoption (means)
LG (%) 9.911 8.293 25.173 16.471***
NII (%) 69.591 67.487* 72.231 70.577
ROE (%) 13.086 8.986*** 16.102 16.944
ROA (%) 0.841 0.574*** 1.741 1.476**
NPL (%) 2.238 3.051*** 6.759 4.355
Notes: This table provides descriptive statistics for 78 banks that adopted the EPs during the period starting in 2003, the year the EPs were initiated, and
until the end of 2015. Of the total of 78 banks, 52 come from developed countries and 26 from developing countries. The descriptive analysis presented in
the table is based on data collected for each bank starting from three years preceding EP adoption and until 2015, resulting in 625 bank-year observations
for developed countries and 304 for developing countries. Panel A compares between developed and developing countries, and asterisks indicate that the
developing countries’ value is significantly different from the corresponding developed countries’ value. Panel B compares the mean characteristics of EP-
adopters with that of non-adopters in the pre-EP period (i.e. prior to 2003), in developed and developing countries. Asterisks indicate that the mean value
for EP-adopters is significantly different than the corresponding mean for non-adopters. Panel C compares the mean performance measures in the pre- and
post-EP periods, for developed and for developing countries separately. Asterisks indicate that the post-adoption value is significantly different from the
corresponding pre-adoption value.
Variable definitions: LG is loan growth rate or funding activity, measured as the annual percentage change in the value of loans. NII is net interest income
divided by net interest income plus non-interest income. Non-interest income includes fee earnings from fiduciary activities, service charges on deposits
accounts, trading account gains and fees, as well as commissions. ROE is net income divided by total equity. ROA is net income divided by total assets. NPL is
non-performing loans divided by total loans. Total assets is the bank’s total assets (measured in $ billions). Capitalization is equity divided by total assets.
Tier1 ratio is the bank’s core equity capital divided by total risk-weighted assets. Loan/Deposit is loans divided by deposits. Coverage is loan loss reserve
divided by non-performing loans. Cost/Income is overhead costs divided by total income. Rest_S (I, E) is the degree of restrictions on banking activities in
securities (insurance, real estate), ranging from 1 (low restrictions) to 4 (high restrictions) as per Barth et al. (2006, 2008). Corrupt is the degree of control
over corruption, ranging from 3 (high level of corruption) to +3 (low level of corruption), consistent with Kaufman et al. (2009). Credit/GDP is domestic
credit to the private sector, divided by GDP, measuring the dominance of the financial sector (Brewer et al., 2008; Levine, 1999). GDPgr is GDP growth rate.
GDPgr p/c is the growth rate of GDP growth per capita based on current price/population.
*
Significance levels (two-tailed) at the 10% level.
**
***
results for them are insignificant.17 It also appears that in the pre-adoption period, EP-adopters from developed countries,
tended to be more risky than non-adopters (significantly lower Capitalization and Tier 1 ratio). EP-adopters from developed
countries also exhibited poorer performance in the pre-adoption period, compared with non-adopters (significantly lower LG
and NII, and significantly higher NPL). Thus, in line with Hypothesis 1(a), it appears that in developed countries, it is large but
lower quality banks that tend to adopt the EP, perhaps in an attempt to improve their reputation.
In contrast, and with the caveat that the results are mainly insignificant, banks from developing countries that opt for EP
adoption, appear to be of higher quality (in the period prior to adoption) compared with non-adopters. Specifically, we see in
Panel B of Table 3, that EP-adopters from developing countries enjoyed a higher rate of growth in loan portfolio/funding
activity (LG) although with higher fraction of non-performing loans (NPL) but with higher profitability (ROE, ROA) compared
with non-adopters. These results suggest that in developing countries, it is large and higher quality banks that tend to adopt
the EP, in line with Hypothesis 2(a). Reasons may include an attempt by large and high quality banks to maintain their rep-
utation or because it is these banks that can afford the additional costs involved in implementing the EP requirements.
Panel C of Table 3 presents comparison of the means of the five performance measures (LG, NII, ROE, ROA, NPL) of EPFIs,
before and after adoption of the EP, and separately for developed and developing countries. For EPFIs from developed coun-
tries, there is strong evidence to suggest that performance has decreased in the post-adoption period, relative to the pre-
adoption period. Specifically, for these banks, interest income (NII) and profitability (ROE, ROA) are significantly lower, non-
performing loans (NPL) are significantly higher, and the growth rate in loan portfolio/funding activity (LG) is insignificantly
lower in the post-adoption period compared with the pre-adoption period. Moving to EPFIs from developing countries, we
observe a significant reduction in LG and some evidence of a reduction in profitability (ROA) between the pre- and post-
adoption periods. Consistent with the reduction in their loan portfolio/funding activity (LG), EPFIs from developing countries
experienced an insignificant decrease in the percentage of net interest income (NII) and of non-performing loans (NPL),
between the pre- and post-adoption periods. Thus, while there is strong evidence of reduction in performance by EPFIs from
developed countries in the post-relative to the pre-adoption period, for EPFIs from developing countries, it appears that the
impact on performance was limited to a reduction in the growth of their loan portfolio/funding activity, possibly as these
banks tightened their lending policies in line with EP requirements. This is consistent with Hypothesis 2(c) which asserts
that in developing countries EP adoption not a ‘greenwashing’ exercise but a strategic choice that involves additional costs
(Wu and Shen, 2013).
The comparison in Panel C of Table 3, of the performance of EPFIs in the pre- and post-adoption periods is a crude way of
assessing the impact of EP-adoption. Specifically, that comparison does not distinguish between short-term and long-term
effects. A more sophisticated approach to study the direct effect of EP adoption on performance, must control for the time
aspect and for the effects of other confounding factors. A common approach, which accounts for both, is to study the share
price and accounting performance around the short-term period surrounding the event – EP adoption, in this case (Brown
and Warner, 1985; Mackinlay, 1997). Another way, is by conducting a multivariate analysis that controls for the effect of
various long-term determinants of performance, including bank-specific, institutional, and macroeconomic factors
(Athanasoglou et al., 2008; DePrince et al., 2011; Wu and Shen, 2013). A multivariate analysis, however, must control for
possible endogeneity resulting, for example, from a tendency by high performing banks to adopt the EP. The results of these
analyses are discussed in the empirical tests of the next section.
5. Empirical tests and results
5.1. Short-term performance analysis
In this section, we test the short-term performance effects of signing up to the EP. First, to test hypotheses 1(b) and 2(b),
we use an event study methodology (Brown and Warner, 1985; Mackinlay, 1997), examining the share price reaction to the
EP-adoption. Specifically, we use the market model to obtain the cumulative abnormal returns (CARs) around the EP-
adoption date, and examine the average CAR for EP-banks from developed and developing countries. Second, we examine
the short-term changes in the five performance measures (LG, NII, ROE, ROA, NPL) from the year preceding the adoption
(t 1) to the year following the adoption (t + 1). We compare the averages of the performance measures before and after
17
The insignificant results in panel B of Table 3 in the case of developing countries may be due to the small number of observations.
EP adoption, for the sub-samples of EP-banks from developed and developing countries. This goes some way towards testing
hypotheses 1(c) and 2(c) for developed and developing countries, respectively.
5.1.1. Event study analysis

Table 4, Panel A presents the mean and median CARs around EP-adoption date for the sub-samples of EPFIs from devel-
oped and developing countries. The abnormal returns are computed from the market model, by summing the size-adjusted
abnormal returns in the two (0, +1) and four (0, +3) days surrounding the adoption date. We examine whether the reaction to
the event begins prior to the adoption day, and find no significant abnormal returns prior to day 0. This is true for banks from
developed countries as well as for banks from developing countries. As Panel A of Table 4 shows, the average and median
CARs are positive in both developed and developing countries. However, while in developed countries, the positive market
reaction is statistically insignificant, which is consistent with Hypothesis 1(b), in developing countries it is significant. Indeed,
in line with Hypothesis 2(b), the median developing countries’ CAR in the (0, +1) window, is positive and significant. This
effect persists in the two subsequent days, resulting in larger and more significant mean and median CARs in the (0, +3) win-
dow (p-value <5% and <1%, respectively).
Thus, despite the costs and restrictions associated with implementing the EP, investors in EPFIs operating in developing
countries, appear to expect that signing up to the EP will have a positive impact on these banks’ future cashflows. In devel-
oped countries, the reaction to EP adoption is positive, but insignificant. A possible explanation for the difference in reaction
in developed and developing countries to EP adoption, is that in developing countries the prevalence of dubious environmen-
tal management practices of borrowers is more common than in the former. This may be due to a weak regulatory infras-
tructure and ineffective enforcement related to the control over environmental and social risks (Dögl and Behnam, 2015).
Such explanation is consistent with the assumption underlying the EP, that the regulatory environment in designated coun-
tries (mainly developed, high-income countries) generally exceeds the minimum requirements for environmental and social
responsibility, and consequently, projects in these countries are not subject to the same guidelines as projects located in
other countries (see Section 2). Thus, in developing countries, the stock market views EP-adoption as a strategic decision
by the bank to make a real commitment to CSR, which should lead to the creation of value (e.g., through enhanced reputation
and consequently stronger profitability and better credit quality of borrowers). This is in line with Ioannou and Serafeim
(2010) who find that CSR strengths are perceived by the market as value creating. In contrast, in developed countries, adopt-
ing the EP is the norm, (that is business as usual) for any bank involved in project finance (Scholtens and Dam, 2007). It is not
expected to lead to changes in behaviour or performance, and hence the stock market does not react to EP adoption. These
explanations are in line with our hypotheses 1(b) and 2(b) in Section 3.
Alternatively, it may be the case that financing environmentally harmful projects is only a small fraction of the overall
activity of developed countries’ banks, meaning that the net economic impact of EP adoption is limited (see also,
Scholtens and Dam, 2007). In contrast, the significant reaction to EP adoption by banks in developing countries, indicates
that accepting to adhere to the EP requirements, is viewed as a commitment to high moral standards, which investors asso-
ciate with good business and higher future cashflows, despite the burden of EP implementation.
5.1.2. Short-term changes in performance following EP adoption

Panel B of Table 4 reports the means of the five performance measures (LG, NII, ROA, ROE, NPL) for the sub-samples of EPFIs
from developed and developing countries, in the pre-adoption year, year of adoption, and year post-adoption. The perfor-
mance of EPFIs from both developed and developing countries does not change significantly in the adoption year (t) com-
pared to the pre-adoption year (t 1). This is not surprising because the adoption date within a calendar year varies with
each bank, hence the remaining period from the adoption date until the end of the year may be relatively short for substan-
tial changes to transpire. We thus focus on comparing the pre-adoption year (t 1) with the post-adoption year (t + 1).
As presented in Panel B of Table 4, EPFIs from developed countries enjoyed a significant improvement in the rate of
growth of their loan portfolio/funding activity from the pre-adoption year to the post-adoption year (a significant change
of 7.382% in LG). There is also some evidence of improvement in their short-term profitability (ROE has changed significantly
by 4.135%). In contrast, EPFIs from developing countries, appear to have experienced a significant reduction in the rate of
growth of their loan portfolio/funding activity in the period from one year pre-adoption to one year post-adoption (LG
has dropped by a significant 9.598%). The ROE of EPFIs from developing countries also seems to have decreased in the
short-term (1.632%), however this reduction is statistically insignificant. For both groups of EPFIs, the short-term perfor-
mance analysis does not detect significant changes in ROA, NII or NPL.
Thus, the results of Panel B of Table 4 indicate that the short-term performance of EPFIs relating to interest income, prof-
itability, and non-performance loans (NII, ROA, ROE, NPL) have not changed much in both developed and developing coun-
tries. However, there is a striking difference in the means of the short-term growth rate of loan portfolio/funding activity (LG)
between EPFIs from developed and developing countries. Specifically, EPFIs from developed countries appear to have expe-
rienced a significant positive increase in their loan portfolios/funding activity, while EPFIs from developing countries have
experienced a decrease.
The short-term increase in the growth of the loan portfolio/funding activity (LG) of EPFIs in developed countries contra-
dicts the expectations of investors as captured by the CARs (Panel A of Table 4) and is also inconsistent with Hypothesis 1(c),
which predicts no change in performance. It appears that, at least in the short-run, EPFIs from developed countries managed
Table 4
Short-term performance analysis.
Developed countries Developing countries Difference between developed and developing countries
Panel A: Mean (median) Cumulative Abnormal Returns [CAR] for EPFIs in developed versus developing countries surrounding the adoption of the EP
CAR (0, +1) 0.420% 0.559% 0.139%
(0.095%) (1.019%**) (0.924%)
CAR (0, +3) 0.385% 1.856%** 1.471%*
***
(0.355%) (1.603% ) (1.248%*)
Performance variable Pre-adoption Year of adoption Post adoption Pre-adoption Year of adoption Post adoption
(t 1) (t) (t + 1) (t 1) (t) (t + 1)
Panel B: Performance of EPFIs across developed and developing countries: Comparison of the means before, after and around the time of the adoption of the EP
LG 8.096 10.366 15.478 24.658 18.627 15.060
NII 0.578 0.563 0.579 0.619 0.645 0.624
ROE 10.619 9.822 14.754 17.019 16.991 15.387
ROA 0.753 0.590 0.601 1.511 1.576 1.440
NPL 2.719 2.490 2.610 4.127 4.234 4.197
Change from t 1
LG 2.270 7.382** 6.031 9.598**
NII 0.015 0.001 0.026 0.005
ROE 0.797 4.135* 0.028 1.632
ROA 0.163 0.152 0.065 0.071
NPL 0.229 0.109 0.107 0.070
Notes: This table reports the results of our short-term performance analysis, by developed versus developing countries. Panel A shows the mean (median in
parentheses) cumulative abnormal returns (CARs) around the date of the bank’s adoption of the EP. The CARs are accumulated over two (0, +1) and four (0,
+3) days surrounding EP-adoption date. The daily abnormal returns are computed based on the market model and are size adjusted. Panel B shows the
means of five performance measures, (LG, NII, ROE, ROA, and NPL) in the three years surrounding the adoption of the EP (t 1, t, t + 1, where t is the adoption
year). In addition to the annual levels of LG, NII, ROE, ROA, and NPL, the change between each year (t, t + 1) and the pre-adoption year (t 1) are displayed.
LG, NII, ROE, ROA, and NPL are as defined in Table 3.
*
Significance (two-tailed) at the 10% level.
**
***
to boost their funding activities following adoption of the EP. This may be a short-term reputational effect, although there is
only weak evidence that this improvement leads to higher short-term profitability (ROE).
In contrast, while investors in EPFIs from developing countries expect EP adoption to have a substantial economic impact
on the adopting bank (Panel A of Table 4), the evidence in Panel B of Table 4 indicates that at least one aspect of their per-
formance (namely, LG) has deteriorated. The deterioration in LG of EPFIs from developing countries is consistent with
Hypothesis 2(c) and with the notion that for these banks, EP adoption requires changes in operations, such as implementa-
tion of harsher screening of borrowers leading to a reduction in the growth rate of loan portfolio/funding activity. Moreover,
the inconsistency between the share price reaction (Panel A of Table 4) and the short-term shifts in performance (Panel B of
Table 4), may be explained in that changes in share prices reflect investors’ expectations for the long-run. In what follows, we
expand the performance analysis to the long-run (five years following EP adoption).
5.2. Long-term performance analysis
In this section, we estimate multivariate performance regressions which control for various determinants of performance
as well as for possible endogeneity effects. In our setting, endogeneity results if well-performing banks choose to adopt the
EP. Hence, to estimate the direct effect of EP adoption on a bank’s long-term performance, we estimate a Heckman-type, two-
stage treatment effect (Heckman, 1979) as follows: In the first stage, we estimate a Probit-type, EP adoption model in which
the likelihood of a bank adopting the EP, denoted by EP_adoption, is regressed on a set of variables that are expected to affect
this decision:
EP Adoptionikt ¼ b0 þ b1 Ln TAikt þ b2 Capitalizationikt þ b3 Loan Depositikt þ b4 Cov erageikt þ b5 Cost Incomeikt

þ b6 Rest Skt þ b7 Rest Ikt þ b8 Rest Ekt þ b9 Corruptkt þ b10 Credit GDP kt þ b11 GDPgrkt
þ b12 GDPgr p=ckt þ v ikt ð1Þ
EP_adoption is a binary variable that equals 1 if the bank adopted the EP, and 0 otherwise. The subscripts i, k, and t, denote
the ith bank in the kth country in year t, where t begins in 2003, the year the EP were initially launched. In line with previous
research, we include a set of controls at the bank and country levels (see, e.g., Wu and Shen, 2013). Controls at the bank level
include total assets after the logarithmic transformation (Ln_TA), equity divided by total assets (Capitalization),18 loan-to-
deposit (Loan_Deposit), loan loss reserve to non-performing loans (Coverage), and overhead costs to income (Cost_Income). Insti-
18
We examine and find that adding Tier1 ratio does not make an incremental contribution to explaining performance over Capitalization.
tutional factors at the country level include the three restriction factors (Rest_S, Rest_I, Rest_E), the degree of control over cor-
ruption (Corrupt), and credit-to-GDP (Credit_GDP). Also at the country level are two macroeconomic factors including GDP
growth rate (GDPgr) and GDP growth per capita (GDPgr p/c). To capture the differences between developed and developing
countries, each variable, including the intercept, are also interacted with a dummy variable that takes the value of 1 if the coun-
try is developing and zero otherwise. Our model also includes time and country fixed effects.
The Inverse Mills ratio (Heckman, 1979), calculated from the Probit model which predicts the banks’ EP-adoption choice,
is included in the performance regressions to control for possible endogeneity effects. The inverse Mills ratio (Inv. Mills)
included in the performance regressions captures unobservable characteristics that affect both selection and performance
equations.
Performikt ¼ b0 þ b1 EP Adopterikt þ b2 Ln TAikt þ b3 Capitalizationikt þ b4 Loan Depositikt þ b5 Cov erageikt

þ b6 Cost Incomeikt þ b7 Rest Skt þ b8 Rest Ikt þ b9 Rest Ekt þ b10 Corruptkt þ b11 Credit GDPkt
þ b12 GDPgr kt þ b13 GDPgr p=ckt þ b14 Inv :Millsikt þ v ikt ð2Þ
Perform is the bank’s performance, measured alternatively by LG, NII, ROE, ROA and NPL. Thus, we run five regressions for the
five performance proxies. EP_Adopter is an indicator for a bank that has adopted the EP.19 The set of control variables is in line
with prior research (see variable definitions above). The Inverse Mills ratio (Inv. Mills) is computed from our first-stage Probit
regression (Eq. (1)). Finally, v is a white noise error term. All the regressions include time and country fixed effects.
Table 5 presents the correlation coefficients among the independent variables of Eqs. (1) and (2), for developed coun-
tries (Panel A) and for developing countries (Panel B). Spearman’s rho and Pearson’s rho are displayed above and below
the diagonal respectively. Looking at Table 5, we see that the correlation coefficients are generally less than 0.5 in absolute
value, removing worries over potential multicollinearity problems in our multivariate model.20 To further assess whether
potential multicollinearity exists, we calculated the variance inflation factors (VIFs) for each of the independent variables in
our model. The results indicate that most of the VIFs are below 2, and all are below 3, providing confidence that our sample is
generally free from the problem of multicollinearity.21 Moreover, we apply the White’s (1980) heteroscedasticity test to the
residuals of the regressions and find that none of the chi-square statistics are significant, suggesting no issues of
heteroscedasticity.
The estimates of the Probit model (Eq. (1)), displayed in Table 6, indicate that the probability of a bank adopting the EP is
influenced by factors at the country level and at the bank level. At the country level, the probability of a bank adopting the EP
is decreasing with the level of restrictions on banking activities involving securities (Rest_S) and with growth in GDP (GDPgr).
It is increasing with GDP growth per capita. Moreover, although the probability of a bank adopting the EP is increasing when
it resides in a developing country (Developing), the effect of the other country-level controls appears to be the same in devel-
oped and developing countries. At the bank level, the probability of a bank adopting the EP is increasing with its size (Ln_TA),
and equity to assets (Capitalization) and decreasing with the ratio of loans to deposits (Loan_Deposit). The direction of these
effects on the probability of EP adoption is similar regardless of whether they reside in a developed or a developing country
(though the magnitude of the effect may differ between regions). The only exception is the effect of the ratio of equity to
assets (Capitalization) which appears to have a positive effect on the probability of EP adoption in developed countries,
and a negative effect in developing countries. The regression’s pseudo R-square, at nearly 80%, is high, thus the Inverse Mills
ratio, obtained from this regression, is saved to be included in the performance regressions of Eq. (2), as presented in Table 7.
Table 7 presents the results for the second of the two-stage Heckman selection model, which investigates the determi-
nants of the long-term performance of EPFIs. Like the results of the Heckman’s first-stage Probit regression – which models
the probability of a bank adopting the EP (Table 6) – also the results of Table 7 indicate that factors at both the bank and
country levels are important determinants of the performance of EPFIs.
Starting with EPFIs from developed countries, the coefficient on EP_Adopter, our variable of interest, is significantly pos-
itive (at the 1% level) in the NII performance regression. This suggests that, after controlling for various bank-specific, insti-
tutional and macroeconomic effects on long-term performance, EP adopters have a higher interest income (as a percentage
of total income) compared to non-adopters. This may be the result of the short-term increase in loan portfolio/funding activ-
ity (LG, as shown in Panel B of Table 4). Consistent with Hypothesis 1(c), and with the non-significant share price reaction
(the CARs of Panel A, Table 4) the other performance measures (LG, ROE, ROA, NPL) do not differ significantly for adopters and
non-adopters in developed countries.
Moving on to EPFIs from developing countries, the results of Table 7 indicate that in the long-term, EP adopters tend to
experience lower growth in their loan portfolios/funding activity (LG) relative to non-adopters. This result is similar to the
results for the short-term performance analysis, which showed that EPFIs from developing countries tend to experience a
19
We define a bank as an EP-adopter only from the year of adopting the EP. However, the results are robust to an alternative definition of a bank being an EP-
adopter throughout the sample period (i.e. in all years), if at some time it adopted the EP.
20
In the sub-sample of EPFIs from developing countries, there are a number of correlation coefficients that exceed 0.5 (in absolute values). These include
0.596 Spearman’s rho between Ln(TA) and Capitalization, and 0.616 (0.518) Spearman’s (Pearson) rho between the degree of control of corruption and GDP
growth per capita.
21
The VIFs are not reported for the sake of parsimony, but are available from the authors upon request.
Table 5
Correlations matrix.
Ln(TA) Capitalization Loan/Deposit Coverage Cost/Income Rest_S Rest_I Rest_E Corrupt Credit GDP GDPgr GDPgr p/c
Panel A: Developed countries
Ln(TA) 1 0.108** 0.132** 0.013 0.240*** 0.157*** 0.056 0.250*** 0.314*** 0.217*** 0.145*** 0.049
Capitalization 0.390*** 1 0.023 0.020 0.053 0.070 0.032 0.118** 0.159*** 0.065 0.054 0.023
M. Finger et al. / J. Int. Financ. Markets Inst. Money 52 (2018) 240–261

Loan/Deposit 0.294*** 0.293*** 1 0.016 0.037 0.209*** 0.373*** 0.200*** 0.108* 0.142** 0.085 0.183***
Coverage 0.157*** 0.091 0.082 1 0.047 0.352*** 0.052 0.365*** 0.193*** 0.343*** 0.280*** 0.127**
Cost/Income 0.006 0.024 0.039 0.004 1 0.175*** 0.065 0.000 0.101* 0.005 0.147*** 0.114**
Rest_S 0.195*** 0.050 0.184*** 0.394*** 0.000 1 0.343*** 0.392*** 0.156*** 0.194*** 0.084 0.382***
Rest_I 0.096* 0.016 0.270*** 0.071 0.094* 0.292*** 1 0.139** 0.121** 0.380*** 0.039 0.100*
Rest_E 0.262*** 0.041 0.170*** 0.303*** 0.022 0.385*** 0.165*** 1 0.330*** 0.121** 0.012 0.126**
Corrupt 0.276*** 0.033 0.148*** 0.068 0.011 0.145*** 0.160*** 0.369*** 1 0.310*** 0.191*** 0.495***
Credit GDP 0.186*** 0.029 0.039 0.190*** 0.017 0.072 0.241*** 0.119** 0.275*** 1 0.005 0.108**
GDPgr 0.071 0.050 0.103* 0.209*** 0.030 0.074 0.052 0.004 0.201*** 0.005 1 0.074
GDPgr p/c 0.012 0.037 0.046 0.068 0.039 0.246*** 0.015 0.062 0.425*** 0.301*** 0.056 1
Panel B: Developing countries
Ln(TA) 1 0.596*** 0.186** 0.223*** 0.089 0.306*** 0.242*** 0.218** 0.107 0.316*** 0.076 0.079
Capitalization 0.119 1 0.077 0.153* 0.259*** 0.348*** 0.062 0.276*** 0.019 0.301*** 0.044 0.084
Loan/Deposit 0.154* 0.077 1 0.135 0.099 0.191** 0.030 0.001 0.369*** 0.287*** 0.194** 0.370***
Coverage 0.226*** 0.019 0.369*** 1 0.306*** 0.046 0.339*** 0.296*** 0.020 0.267*** 0.264*** 0.269***
Cost/Income 0.222** 0.051 0.154* 0.021 1 0.004 0.321*** 0.118 0.025 0.070 0.217** 0.212**
Rest_S 0.467*** 0.015 0.248*** 0.094 0.114 1 0.070 0.231*** 0.402*** 0.372*** 0.204** 0.458***
Rest_I 0.195** 0.038 0.034 0.263*** 0.213** 0.041 1 0.409*** 0.282*** 0.054 0.005 0.328***
Rest_E 0.130 0.017 0.174** 0.237*** 0.021 0.013 0.321*** 1 0.143 0.176** 0.213** 0.029
Corrupt 0.229*** 0.084 0.222** 0.330*** 0.047 0.349*** 0.411*** 0.179** 1 0.385*** 0.160* 0.616***
Credit GDP 0.305*** 0.071 0.200** 0.201** 0.081 0.389*** 0.064 0.186** 0.273*** 1 0.190** 0.051
GDPgr 0.192** 0.088 0.217** 0.159* 0.166* 0.298*** 0.020 0.250*** 0.122 0.129 1 0.116
GDPgr p/c 0.176** 0.081 0.273*** 0.172** 0.136 0.486*** 0.385*** 0.146* 0.518*** 0.051 0.033 1
Notes: Spearman’s (Pearson) rhos are displayed above (below) the diagonal. Ln(TA) is the natural logarithm of the bank’s total assets (measured in $billions). All the other variables are as defined in Table 3.
*
**
***
253
Table 6
First-stage of the Heckman selection model: Probit model for EP adoption.
Intercept 11.284***
Developing 7.999**
Ln_TA 2.234***
Ln_TA * Developing 0.764***
Capitalization 13.863**
Capitalization * Developing 18.077**
Loan_Deposit 1.425***
Loan_Deposit * Developing 1.556***
Coverage 0.138
Coverage * Developing 0.248
Cost_Income 0.003
Cost_Income * Developing 0.007
Rest_S 1.371***
Rest_S * Developing 1.072
Rest_I 0.393
Rest_I * Developing 0.105
Rest_E 0.482
Rest_E * Developing 0.485
Corrupt 0.513
Corrupt * Developing 1.383
Credit_GDP 0.395
Credit_GDP * Developing 0.972
GDPgr 84.584**
GDPgr * Developing 59.660
GDPgr p/c 78.107*
GDPgr p/c * Developing 56.167
Time fixed effects Included
Country fixed effects Included
Pseudo R-sq 0.796
No. of Obs. 7264
Notes: The dependent variable, EP_adoption, is a binary variable that equals 1 if the bank adopted the EP, and
0 if it did not. Controls at the bank level include total assets after the logarithmic transformation (Ln_TA),
equity divided by total assets (Capitalization), loan-to-deposit ratio (Loan_Deposit), loan loss reserve to non-
performing loans (Coverage), and overhead cost to income (Cost_Income). Institutional factors at the country
level include the three restriction factors (Rest_S, Rest_I, Rest_E), the degree of control over corruption
(Corrupt), and the credit-to-GDP ratio (Credit_GDP). The macroeconomic factors are GDP growth rate (GDPgr)
and GDP growth per capita growth (GDPgr p/c). The variables are as defined in Table 3. To capture the
differences between developed and developing countries, each variable, including the intercept, are inter-
acted with a dummy variable, Developing, that takes the value of 1 if the country is developing and zero
otherwise. Time and country fixed effects are included.
*
Significance at the 10% level.
**
***
decrease in their loan portfolios/funding activity (LG) following the year of adoption (Panel B of Table 4). The continuing
reduction in their loan portfolios/funding activity may be the reason that net interest income (as a percentage of total
income, NII) tends to be lower for EPFIs in developing countries, compared to non-adopters. Furthermore, the significant neg-
ative effect of EP adoption on the growth in loan portfolio (i.e. on funding activity, LG) is in line with Hypothesis 2(c) for EPFIs
in developing countries. In contrast, the effect of EP adoption on profitability (ROE, ROA), and on the percentage of non-
performing loans (NPL) is insignificant. Although adoption of the EP by banks in developing countries involves costs of imple-
mentation, including a decline in funding activity (LG) and a reduction in income from interest (NII), the benefits that arise
from this investment seem to offset these costs. It is possible that over a longer period, (i.e. if performance was measured
over longer than five years), the benefits arising from increased reputation or from better quality of borrowers, would more
than offset implementation costs, leading to an increased in profitability (ROE, ROA) and a reduction in the fraction of non-
performing loans (NPL). Alternatively, it may be the case that like in developed countries, CSR will become ‘business as usual’.
Accordingly, firms will be pushed to implement CSR, simply to maintain their reputation (i.e. not for the purpose of creating
value, but merely to avoid destroying it).
Table 7
Second-stage of the Heckman selection model: Performance regressions.
LG NII ROE ROA NPL

Intercept 29.913* 0.728*** 0.135*** 0.017*** 0.257***
Developing 27.587 0.021 0.244*** 0.015*** 0.340***
EP_Adopter 4.418 0.076*** 0.012 0.001 0.001
EP_Adopter * Developing 11.481** 0.095** 0.014 0.002 0.046
Ln_TA 0.112 0.021*** 0.009*** 0.001*** 0.002
Ln_TA * Developing 4.299** 0.004 0.009** 0.000 0.033
Capitalization 8.801 0.057 0.581*** 0.058*** 0.151*
Capitalization * Developing 3.562*** 1.244*** 0.210** 0.090*** 0.094
Loan_Deposit 0.080 0.000 0.016*** 0.000 0.004
Loan_Deposit * Developing 1.060*** 0.052** 0.001 0.001* 0.024
Coverage 0.000 0.000 0.000 0.000 0.000
Coverage * Developing 0.241 0.001 0.002* 0.000 0.008**
Cost_Income 0.000 0.000 0.000 0.000 0.000
Cost_Income * Developing 0.027 0.002** 0.000 0.0001*** 0.000
Rest_S 0.228 0.027* 0.043*** 0.003*** 0.022*
Rest_S * Developing 3.142 0.079*** 0.038*** 0.002** 0.057**
Rest_I 2.207 0.109*** 0.023*** 0.002*** 0.005
Rest_I * Developing 7.201* 0.082*** 0.021** 0.001 0.033*
Rest_E 4.507** 0.095*** 0.012*** 0.000 0.006
Rest_E * Developing 0.922 0.100*** 0.017** 0.001* 0.001
Corrupt 5.365* 0.085*** 0.037*** 0.002*** 0.053***
Corrupt * Developing 10.001* 0.056* 0.007 0.000 0.035
Credit_GDP 5.511 0.167*** 0.031*** 0.000 0.003
Credit_GDP * Developing 13.871 0.308*** 0.049** 0.013*** 0.363***
GDPgr 0.667 15.030*** 1.464* 0.126** 1.773
GDPgr * Developing 2.374 13.291*** 2.654*** 0.236*** 2.029
GDPgr p/c 5.210 17.677 0.661 0.073 3.609*
GDPgr p/c * Developing 7.188* 16.051*** 1.907** 0.188*** 3.332*
Inv. Mills 2.120 0.064 0.005 0.002 0.002
Inv. Mills * Developing 3.486 0.025 0.012 0.002 0.052
Time fixed effects Included Included Included Included Included
Country fixed effects Included Included Included Included Included
R_squared 0.178*** 0.145*** 0.172*** 0.301*** 0.123***
No. Obs. 7022 7260 7029 7069 7263
Notes: This table presents the estimation results for the long-term performance regressions. Five regressions are run, one for each of the five performance
measures (LG, NII, ROE, ROA, and NPL), using the pooled sample of EP and non-EP banks from developed and developing countries, and excluding the year of
adoption. EP_Adopter is a dummy set to 1 for banks that adopted the EP, and zero otherwise. The control variables include bank and country characteristics.
Bank characteristics include Ln_TA (total assets after the logarithmic transformation), Capitalization, Loan_Deposit, Coverage and Cost_Income. Country
characteristics include institutional and macroeconomic factors. The institutional factors include Rest_S, Rest_I, Rest_E, Corrupt, and Credit_GDP. The
macroeconomic factors include GDPgr and GDPgr p/c. All these control variables are defined in Table 3. Inv. Mills is the inverse Mills ratio generated from the
first-stage of the Heckman selection model. To capture the differences between developed and developing countries, we include a dummy variable,
Developing, which takes the value 1 if the bank is operating in a developing country, and zero otherwise. Each variable, including the intercept, is interacted
with the Developing dummy. Time and country fixed effects are included.
*
**
***
The estimated coefficients on the control variables are generally consistent with the literature. Differences between our
results for the control variables, and those of previous studies – to the extent that they exist – may be attributed to differ-
ences in the choice of countries and regions. As discussed above, previous studies on the relationship between CSR and bank
performance, use either US banks (Simpson and Kohers, 2002) or banks mostly located in Europe (Wu and Shen, 2013). In
contrast, our sample is spread across different continents as well as across developed and developing regions. The interaction
terms between the explanatory variables and the developing country dummy, exemplify these differences. Finally, the coef-
ficients on the inverse Mills ratio are consistently insignificant for developed and for developing countries, suggesting that
endogeneity does not affect our findings. Thus, as suggested in previous studies (see Wang et al., 2015), socially responsible
performance affects subsequent financial performance rather than vice versa.
5.3. Robustness tests
The sample period in this study includes the sub-prime crisis (2007–2009), which had negative effects on the perfor-
mance of banks worldwide. In addition to a potential impact on performance, the crisis could have influenced social perfor-
mance. Karaibrahimoglu (2010) shows that CSR projects declined due to the financial downturn of the sub-prime crisis (in
the US more than in Europe and other countries). We thus seek to explore whether our results are robust to crisis periods.
We repeat the multivariate regressions analysis, adding an indicator for the sub-prime crisis years, 2007–2009 (Crisis), as
well as an interaction variable between EP_Adopter and Crisis.
Performikt ¼ b0 þ b1 Crisis þ b2 EP Adopter ikt þ b3 ðEP Adopter CrisisÞikt þ b4 Ln TAikt þ b5 Capitalizationikt

þ b6 Loan Depositikt þ b7 Cov erageikt þ b8 Cost Incomeikt þ b9 Rest Skt þ b10 Rest Ikt þ b11 Rest Ekt
þ b12 Corrupt kt þ b13 Credit GDPkt þ b14 GDPgrkt þ b15 GDPgr p=ckt þ b16 Inv : Mills þ v ikt ð3Þ
Each variable in Eq. (3), including the intercept, are further interacted with a developing country dummy (Developing). The
coefficient on Crisis is expected to capture the effect of the 2007–9 crisis on performance. The coefficient on
EP_Adopter ⁄ Crisis is included to measure the particular effect of the crisis on the performance of EPFIs. The results of
Eq. (3) are displayed in Table 8. We find that the observed higher (lower) NII for EP adopters in developed (developing)
countries compared to non-adopters is robust to controlling for the 2007–9 crisis. Likewise, the statistically negative effect
of EP adoption on funding activity (LG) in developing countries is robust to controlling for the 2007–9 crisis. Notably, the
reduction in lending activities for EP adopters in developing countries seems to worsen in crisis years. During these years,
in concomitance with the reduction in borrowing, the proportion of nonperforming loans has increased, providing an addi-
tional evidence that borrowers in these regions are generally of low credit quality. Moreover, the results reported in
Table 8 indicate that once we control for the 2007–9 crisis, the difference in the effect of EP adoption on bank performance
between developed and developing countries becomes more apparent. Indeed, the developing country dummy (Develop-
ing) is now statistically significant in all but the NII regression. Overall, the results suggest that our inferences are by and
large robust to market cycles.
6. Discussion
We study the effects of assuming environmental responsibility, on bank performance in developed and developing coun-
tries. A summary of the research questions, hypotheses, tests and results are given in Table 9.
Central to this study are the Equator Principles (EP), a set of best practice principles first drawn in 2003 to provide guid-
ance on environmental risk management to banks involved in project finance. Since 2003, banks from developed and devel-
oping countries have been gradually adopting the EP, and the aim of this study was to analyse the effect of EP adoption on
bank performance. Thus, our sample includes 78 EP-member banks from 34 countries over the period 2003–2015. For some
of our tests, the sample is supplemented by a sample of banks from the same countries that did not adopt the EP. Moreover,
our sample of EP-member banks is split into a sub-sample of 52 EP-banks from 16 developed countries and a sub-sample of
26 EP-banks from 18 developing countries. We draw this distinction between developed and developing countries because
we show that country and bank characteristics differ significantly between the two sub-samples. A summary of the results of
these tests is presented in Panel A of Table 9.
To study the effect of EP-adoption on bank performance, we seek to answer three questions, namely, what type of banks
adopt the EP; how does the market react to EP adoption; and what are the short-term and long-term performance effects of
EP adoption. To answer these questions, we formulate two sets of hypotheses, one for EP-banks from developed countries
and one for EP-banks from developing economies. A summary is provided in Panel B of Table 9.
Our first set of hypotheses concerns EP-banks from developed countries. We argue that in developed countries, environ-
mental risk management is well established. Accordingly, most firms (and banks) have in place mechanisms and procedures
that ensure social and environmental concerns are adequately addressed, regardless of the EP. Consequently, we hypothesize
that EP-adoption is a form of greenwashing which is used by poor quality banks with high public exposure, in an attempt to
influence public and investors’ perceptions. We also hypothesize the EP adoption does not involve high costs, nor is it
expected to change operations or cashflows and hence neither the market reaction nor the short term or long term perfor-
mance effects of EP adoption should be significant.
The empirical results for the sub-sample of EP-banks from developed countries are largely in line with the first set of
hypotheses. First, we find that in the pre-EP era (i.e. before 2003), EPFIs from developed countries were large, risky, and
poor performers compared with banks in the same countries that did not subsequently adopt the EP. Second, in line
with the idea that in developed countries EP adoption is ‘business as usual’, not involving much investment or costs,
we find – by way of an event study methodology – that the share price reaction to EP adoption is insignificant. Third,
in line with the greenwashing hypothesis, we find – by way of performance comparison and a two-stage Heckman selec-
tion model – minor performance effects to EP adoption. Specifically, we record a short-term increase in funding activity
and weaker evidence of improved profitability, both of which can be attributed to short-term reputational effects of EP
adoption. However, the positive performance effects of EP adoption appear to disappear in the longer term. Indeed, the
results of our two-stage Heckman selection procedure show that in the 5 years following EP-adoption, none of the prof-
itability or other performance indicators are significantly influenced by EP-adoption. The only exception is the positive
effect of EP-adoption on the fraction of income from interest, which may be the result of the short-term increase in
funding activity.
Our second set of hypotheses concerns EP-banks from developing countries. We assume that in developing countries the
institutional and regulatory settings are less supportive of social and environmental risk management. Thus, we argue that
Table 8
Robustness tests: Controlling for the sub-prime crisis years (2007–9).
LG NII ROE ROA NPL

Intercept 8.187 0.750*** 0.101*** 0.014*** 0.167**
Crisis 0.491 0.021 0.020*** 0.000 0.015
Developing 28.012* 0.081 0.240*** 0.015*** 0.312***
Developing * Crisis 32.700*** 0.026 0.014 0.002** 0.158***
EP_Adopter 3.210 0.058** 0.006 0.002 0.006
EP_Adopter * Crisis 0.861 0.056 0.010 0.001 0.002
EP_Adopter * Developing 8.602* 0.094* 0.008 0.003 0.087
EP_Adopter * Developing * Crisis 23.326* 0.051 0.028 0.002 0.189***
Ln_TA 0.040 0.020** 0.010*** 0.001*** 0.003
Ln_TA * Developing 4.172** 0.004 0.008** 0.000 0.034***
Capitalization 5.560 0.067 0.582*** 0.058*** 0.144*
Capitalization * Developing 357.823*** 1.296 0.225** 0.088*** 0.045
Loan_Deposit 0.567 0.008 0.013*** 0.000 0.007
Loan_Deposit * Developing 106.729*** 0.048** 0.004 0.001* 0.018
Coverage 0.000 0.000 0.000001* 0.0000001* 0.000
Coverage * Developing 0.275 0.001 0.001 0.000 0.008***
Cost_Income 0.000 0.000 0.0001*** 0.000 0.000
Cost_Income * Developing 0.003 0.002** 0.000* 0.0001*** 0.000
Rest_S 0.512 0.015* 0.001 0.000 0.012
Rest_S * Developing 3.644 0.057** 0.005 0.001 0.048**
Rest_I 0.001 0.082*** 0.009 0.001* 0.002
Rest_I * Developing 5.735 0.051** 0.011 0.000 0.035*
Rest_E 2.122 0.070*** 0.003 0.001*** 0.002
Rest_E * Developing 4.031 0.081*** 0.004 0.000 0.015
Corrupt 0.156 0.028** 0.039*** 0.003*** 0.018
Corrupt * Developing 14.151** 0.006 0.006 0.000 0.065**
Credit_GDP 1.326 0.092*** 0.022** 0.000 0.028
Credit_GDP * Developing 14.344 0.227*** 0.072*** 0.015*** 0.370***
GDPgr 56.738 14.159*** 4.576*** 0.324*** 1.943
GDPgr * Developing 230.728 12.250*** 5.709*** 0.428*** 2.107
GDPgr p/c 51.305 15.563*** 2.350*** 0.176*** 1.561
GDPgr p/c * Developing 277.330 13.940*** 3.809*** 0.305*** 1.401
Inv. Mills 1.221 0.062 0.002 0.002 0.001
Inv. Mills * Developing 4.630 0.040 0.001 0.003 0.063
Time fixed effects Included Included Included Included Included
Country fixed effects Included Included Included Included Included
R_squared 0.178*** 0.139*** 0.138*** 0.274*** 0.123***
No. Obs. 7022 7260 7029 7069 7263
Notes: This is the Second-stage of the Heckman selection model for the performance regressions with controls for the sub-prime crisis years (2007–2009).
Crisis is a dummy variable equal to 1 in the crisis years and zero otherwise. All other definitions are as provided in the notes to Table 7.
EP adoption is a strategic choice that involves costly investments, dramatically changes adopters’ operations; and impacts
performance. Consequently, we hypothesize that the decision to adopt the EP, will be taken by large banks of high quality.
This is due to the high implementation costs which large, fast-growing, better performers will find easier to finance. We fur-
ther hypothesize that the market reaction should be strong and positive due to the signal in the adoption (Spence, 1973).
Lastly, we hypothesize that because EP require screening of borrowers, adoption should lead to a reduction in funding activ-
ity. The effect on profitability may be positive, negative or none. This depends on the balance between the additional costs
and revenues relating to the adoption of the EP.
The empirical results for the sub-sample of EP-banks from developing countries are largely in line with the second set
of hypotheses. First, we find that in the pre-EP era (i.e. before 2003), EPFIs from developing countries were large, and of
high quality compared with banks in the same countries that did not subsequently adopt the EP. Second, in line with the
idea that in developing countries EP adoption is a strategic choice that involves substantial investment, the event study
methodology reveals a significantly positive share price reaction to EP adoption. Third, the short-term performance com-
parison and the two-stage Heckman selection model reveal that the EP adoption is followed by a reduction in lending
activity, in line with the idea that adoption leads banks to better screening of potential borrowers. Consequently, we also
find that compared with other banks, in the long-term, EP-adopters from developing countries generate a lower fraction of
their net income from interest. The adoption of CSR by companies, particularly banks, in developing countries, has been
Table 9
Summary of hypotheses, tests and findings.
Panel A: Developed versus developing countries

Table 3A shows the significant differences between developing and developed countries in terms of both the characteristics of the markets and the
characteristics of EP-banks:
Compared with EP-banks from developed countries, EP-banks from developing Compared with developed countries, developing countries are
countries are characterised by: characterised by:
higher growth in their loan portfolios/funding activity (LG). higher economic growth (GDP growth and GDP growth per
higher fraction of non-preforming loans (NPL). capita)
higher fraction of income from interest-bearing activities (NII). higher potential for growth in the credit market (lower
higher profitability (ROE and ROA). Credit/GDP ratio)
smaller bank size, (total assets) lower degree of control corruption (Corrupt)
lower degree of risk (higher Capitalization, Tier1 and Coverage ratios and a more regulatory restrictions on the banking industry
lower Loan/Deposit ratio). (Rest_S, Rest_I and Rest_E).
Panel B: EP-adopters in developed and developing countries
Research questions Test Developed countries Developing countries
Hypotheses Findings Hypotheses Findings
(a) What type of Table 3B: H1(a). Prior to EP- Adopting banks are H2(a). Prior to EP- Adopting banks are
banks adopt Comparing EP- adoption: bigger but of lower adoption: bigger and of higher
the EP? adopters with non- Banks that opt for EP quality: Banks that opt for EP quality:
adopters in the adoption, are large, Bigger (signifi- adoption are large, Bigger (total assets)
pre-2003 period risky, and poorer cantly larger value fast-growing, and higher rate of
performers, compared for total assets) better performers, growth in loan
with non-adopters. Riskier (signifi- compared with non- portfolio/funding
cantly lower Capi- adopters. activity (LG) with
talization & Tier 1) higher fraction of
Poorer performers non-performing
(significantly lower loans (NPL)
LG and NII, and sig- higher profitability
nificantly higher (ROE, ROA)Results
NPL) are insignificant,
possibly due to
small #obs.
(b) How does Table 4A: H1(b). Share price Insignificantly positive H2(b). Share price Significantly positive
the market Event study reaction: reaction: reaction: reaction:
react to bank methodology to EP adoption is ‘business Both mean and EP adoption requires Both mean and
adoption of analyse the share as usual’, not requiring median CAR(0, +1) commitment and median CAR(0, +1)
the EP? price reaction of extra commitment or are positive but investment to create are positive but
EP-banks around investment. Therefore, insignificant. value. Therefore, the insignificant.
EP-adoption the share price reaction Both mean and share price reaction is Both mean and
is insignificant. median CAR(0, +3) positive and median CAR(0, +3)
are positive but significant. are positive and
insignificant. significant.
(c) What are the Table 4B: H1(c). Performance Short-term change H2(c). Performance Short-term change
short term Comparing the effects: in performance fol- effects: in performance fol-
and long short-term EP adoption is lowing EP adoption EP adoption is a lowing EP adoption
term perfor- performance of ‘greenwashing’ because is an increase in strategic decision with is a reduction in
mance EPFIs before and the regulatory, funding activity short-term and long- the growth of loan
effects of after EP-adoption. institutional, and (LG) and weaker term effects on portfolio/funding
adopting the Tables 7 and 8: governance systems evidence of performance, the activity (LG)
EP? Second-stage of are robust in protecting improvement in direction of which Long-term effect of
the Heckman people and the natural profitability (ROE) depends on the balance EP adoption is to
selection model, environment. Long-term effect of between the costs and lower growth in
estimating the Therefore, there are no EP adoption is to benefits to accrue from loan portfolio/fund-
effect of EP- short-term or long- increase the frac- EP adoption. But EP- ing activity (LG),
adoption on the term performance tion of income that adopters should and the fraction of
long-term effects to EP adoption. comes from inter- experience a drop in net income that
performance of est (NII) their funding activity. comes from inter-
banks est (NII)
criticized as ‘‘greenwashing” (see, e.g., BankTrack, 2005). Our study provides contradicting evidence. We show that EP
adoption by banks from developing countries, involves real commitment and leads to a reduction in funding activities.
This commitment is appreciated by investors, even though the performance effects must take longer than five years to
materialize.
Appendix A
The list of banks that adopted the EP by country and year of adoption.
Name Year of adoption Country

Banco de Galicia y Buenos Aires 2007 Argentina
ANZ 2006 Australia
EFIC 2009 Australia
National Australia Bank 2007 Australia
Westpac Banking Corporation 2003 Australia
KBC Group 2004 Belgium
Banco Bradesco 2004 Brazil
Banco do Brasil 2006 Brazil
Banco PINE 2012 Brazil
CAIXA Economica Federal 2009 Brazil
Itaú Unibanco 2004 Brazil
Bank of Montreal 2005 Canada
Bank of Nova Scotia 2006 Canada
Canadian Imperial Bank of Commerce 2003 Canada
Export Development Canada 2007 Canada
Manulife Financial 2005 Canada
Royal Bank of Canada 2003 Canada
TD Bank Financial Group 2007 Canada
CORPBANCA 2007 Chile
Industrial Bank 2008 China
Bancolombia 2008 Colombia
CIFI 2007 Costa Rica
Eksport Kredit Fonden 2004 Denmark
Arab African International Bank 2009 Egypt
Crédit Agricole Corporate and Investment Bank 2003 France
Société Générale 2007 France
BNP Paribas 2008 France
Natixis 2010 France
UniCredit Bank 2003 Germany
KfW IPEX-Bank 2008 Germany
DekaBank 2011 Germany
DZ Bank 2013 Germany
IDFC 2013 India
Intesa Sanpaolo 2006 Italy
Bank of Tokyo-Mitsubishi UFJ 2005 Japan
Mizuho Bank 2003 Japan
Sumitomo Mitsui Banking Corporation 2007 Japan
Ahli United Bank 2011 Kingdom of Bahrain
Mauritius Commercial Bank 2012 Mauritius
Banco Mercantil del Norte 2012 Mexico
CIBanco 2012 Mexico
BMCE Bank 2010 Morocco
Access Bank Plc 2009 Nigeria
Fidelity Bank 2012 Nigeria
DNB 2008 Norway
Banco de Crédito 2013 Peru
Banco Espírito Santo 2005 Portugal
FirstRand 2009 South Africa
Nedbank 2005 South Africa
Standard Bank 2009 South Africa
Banco Bilbao Vizcaya Argentaria 2004 Spain
CaixaBank 2007 Spain
Banco Santander 2009 Spain
Banco Sabadell 2011 Spain
(continued on next page)

Appendix A (continued)
Name Year of adoption Country

Banco Popular Español 2013 Spain
Bank Muscat 2007 Sultanate of Oman
Nordea Bank 2007 Sweden
Skandinaviska Enskilda Banken 2007 Sweden
Credit Suisse Group 2003 Switzerland
ING Bank 2003 The Netherlands
Rabobank Group 2003 The Netherlands
FMO 2005 The Netherlands
ABN Amro 2009 The Netherlands
ASN Bank NV 2009 The Netherlands
NIBC Bank 2010 The Netherlands
Ecobank Transnational Incorporated 2012 Togo
Barclays 2003 UK
HSBC 2003 UK
Standard Chartered 2003 UK
Lloyds Banking Group 2008 UK
UK Green Investment Bank 2013 UK
Royal Bank of Scotland 2003 UK
Banco de la República Oriental del Uruguay 2008 Uruguay
Bank of America Corporation 2006 US
Citigroup 2003 US
Ex-Im Bank 2011 US
JPMorgan 2006 US
Wells Fargo Bank 2005 US
References
Abreu, J.F., Gulamhussen, M.A., 2013. Dividend payouts: evidence from U.S. bank holding companies in the context of the financial crisis. J. Corp. Financ. 22,
54–65.
Allouche, J., Laroche, P., 2006. The relationship between corporate social responsibility and corporate financial performance: a survey. In: Allouche, J. (Ed.),
Corporate Social Responsibility: Performance and Stakeholders. Palgrave Macmillan, Basingstoke, pp. 3–40.
Athanasoglou, P.P., Brissimis, S.N., Delis, M.D., 2008. Bank-specific, industry-specific and macroeconomic determinants of bank profitability. J. Int. Financ.
Mark. Inst. Money 18 (2), 121–136.
Aupperle, K.E., Carroll, A.B., Hatfield, J.D., 1985. An empirical examination of the relationship between corporate social responsibility and profitability. Acad.
Manage. J. 28 (2), 446–463.
BankTrack, 2005. Unproven Principles. The Equator Principles at Year Two. BankTrack, Amsterdam.
Baron, D.P., 2001. Private politics, corporate social responsibility, and integrated strategy. J. Econ. Manage. Strategy 10 (1), 7–45.
Baron, D.P., 2009. A positive theory of moral management, social pressure and corporate social performance. J. Econ. Manage. Strategy 18, 7–43.
Barth, J.R., Caprio, G., Levine, R., 2006. Rethinking Bank Regulation: Till Angels Govern. Cambridge University Press, New York.
Barth, J.R., Caprio, G., Levine, R., 2008. Bank regulations are changing – but for better or worse. Comp. Econ. Stud. 50, 537–563.
Beck, T., Demirguc-Kunt, A., Levine, R., 2010. Financial institutions and markets across countries and over time: the updated financial development and
structure database. World Bank Econ. Rev. 24 (1), 77–92.
Belal, A.R., 2001. A study of corporate social disclosures in Bangladesh. Manage. Audit. J. 16 (5), 274–289.
Bing, Z., Yan, Y., Jun, B., 2011. Tracking the implementation of green credit policy in China: top-down perspective and bottom-up reform. J. Environ. Manage.
92, 1321–1327.
Brewer, E., Kaufman, G.G., Wall, L.D., 2008. Bank capital ratios across countries: why do they vary? J. Financ. Serv. Res. 34 (2 & 3), 177–201.
Brønn, P.S., Vrioni, A.B., 2001. Corporate social responsibility and cause-related marketing: an overview. Int. J. Advert. 20, 207–222.
Brown, S.J., Warner, J.B., 1985. Using daily stock returns: the case of event studies. J. Financ. Econ. 14 (1), 3–31.
Bushman, R.M., Wittenberg-Moerman, R., 2012. The role of bank reputation in ‘‘certifying” future performance implications of borrowers’ accounting
numbers. J. Acc. Res. 50 (4), 883–930.
Chapple, W., Moon, J., 2005. Corporate social responsibility (CSR) in Asia: a seven-country study of CSR web site reporting. Bus. Soc. 44 (4), 415–441.
Chatterji, A.K., Levine, D.I., 2008. Imitate or Differentiate? Evaluating the Validity of Corporate Social Responsibility Ratings. Working Paper Series, pp. 1–55.
Chatterji, A.K., Levine, D.I., Toffel, M.W., 2010. How well do social ratings actually measure corporate social responsibility? J. Econ. Manage. Strategy 18 (1),
125–169.
Chih, H.L., Chih, H.H., Chen, T.Y., 2010. On the determinants of corporate social responsibility: international evidence on the financial industry. J. Bus. Ethics
93 (1), 115–135.
Cochran, P.L., Wood, R.A., 1984. Corporate social responsibility and financial performance. Acad. Manage. J. 27 (1), 42–56.
Cuesta-González, M., Muñoz-Torres, M., Fernández-Izquierdo, M.Á., 2006. Analysis of social performance in the Spanish financial industry through public
data: a proposal. J. Bus. Ethics 69 (3), 289–304.
de-los Ángeles, M., Estallo, G., de-la Fuente, F.G., Gríful-Miquela, C., 2007. The importance of corporate social responsibility and its limits. Int. Adv. Econ. Res.
13 (3), 379–388.
DePrince, A., Ford, W., Morris, P., 2011. Some causes of interstate differences in community bank performance. J. Econ. Financ. 35 (1), 22–40.
Dögl, C., Behnam, M., 2015. Environmentally sustainable development through stakeholder engagement in developed and emerging countries. Bus. Strategy
Environ. 24, 583–600.
Ellul, A., Yerramilli, V., 2013. Stronger risk controls, lower risk: evidence from U.S. bank holding companies. J. Financ. 68 (5), 1757–1803.
Frankental, P., 2001. Corporate social responsibility – A PR invention? Corp. Commun.: Int. J. 6 (1), 18–23.
Freixas, X., Rochet, J.C., 1998. Microeconomics of Banking. MIT Press.
Friedman, M., 1970. A Friedman Doctrine: Social Responsibility of Business is to Increase its Profits. The New York Times Magazine, Sep. 13, 32–33, and 123–
125.
Geva, A., 2008. Three models of corporate social responsibility: interrelationships between theory, research, and practice. Bus. Soc. Rev. 113 (1), 1–41.
Hafeez, S., 2003. The Efficiency of Regulation in Developing Countries. UN DESA Discussion Paper 31.
Heal, G., 2005. Corporate social responsibility: an economic and financial framework. Geneva Pap. Risk Insurance: Iss. Pract. 30 (3), 387–409.
Heckman, J.J., 1979. Sample selection bias as a specification error. Econometrica 47 (1), 153–162.
Ioannou, I., Serafeim, G., 2010. The impact of corporate social responsibility on investment recommendations. In: Best Paper Proceedings, Academy of
Management 2010 Annual Meeting (Montreal Canada). Social Issues in Management (SIM) Division.
Jamali, D., Mirshak, R., 2007. Corporate social responsibility (CSR): theory and practice in a developing country context. J. Bus. Ethics 72 (3), 243–262.
Jo, H., Kim, Y., 2008. Ethics and disclosure: a study of the financial performance of firms in the seasoned equity offerings market. J. Bus. Ethics 80 (4), 855–
878.
Jones, M.T., 1999. The institutional determinants of social responsibility. J. Bus. Ethics 20 (2), 163–179.
Kakabadse, N.K., Rozuel, C., Lee-Davies, L., 2005. Corporate social responsibility and stakeholder approach: a conceptual review. Int. J. Govern. Ethics 1 (4),
227–302.
Karaibrahimoglu, Y.Z., 2010. Corporate social responsibility in times of financial crisis. Afr. J. Bus. Manage. 4, 382–389.
Kaufman, D., Kraay, A., Mastruzzi, M., 2009. Governance Matters VIII: Aggregate and Individual Governance Indicators for 1996–2008. World Bank Policy
Research Working Paper, No. 4978.
Kempf, A., Osthoff, P.C., 2008. SRI funds: Nomen est Omen. J. Bus. Financ. Acc. 35 (9–10), 1276–1294.
Levine, R., 1999. Law, finance, and economic growth. J. Financ. Intermed. 8, 8–35.
Lockett, A., Moon, J., Visser, W., 2006. Corporate social responsibility in management research: focus, nature, salience and sources of influence. J. Manage.
Stud. 43 (1), 115–136.
Lopez, M.V., Garcia, A., Rodriguez, L., 2007. Sustainable development and corporate performance: a study based on the Dow Jones sustainability index. J. Bus.
Ethics 75 (3), 285–300.
MacKinlay, A.C., 1997. Event studies in economics and finance. J. Econ. Liter. 35 (1), 13–39.
Macve, R., Chen, X., 2010. The ‘‘Equator Principles”: a success for voluntary codes? Acc. Audit. Acc. J. 23 (7), 890–919.
Margolis, J., Elfenbein, H., Walsh, J., 2007. Does It Pay to be Good? A Meta-analysis and Redirection of Research on the Relationship between Corporate Social
and Financial Performance. Harvard Business School Working Paper.
Margolis, J.D., Walsh, J.P., 2003. Misery loves companies: rethinking social initiatives by business. Adm. Sci. Q. 48 (2), 263–305.
McWilliams, A., Siegel, D.S., Wright, P.M., 2006a. Corporate social responsibility: strategic implications. J. Manage. Stud. 43 (1), 1–18.
McWilliams, A., Siegel, D., 1997. Event studies in management research: theoretical and empirical issues. Acad. Manage. J. 40, 626–657.
McWilliams, A., Siegel, D.S., 2001. Corporate social responsibility: a theory of the firm perspective. Acad. Manage. Rev. 26 (1), 117–127.
McWilliams, A., Siegel, D.S., Wright, P.M., 2006b. Introduction by guest editors corporate social responsibility: international perspectives. J. Bus. Strat. 23 (1),
1–7.
Nandy, M., Lodh, S., 2012. Do banks value the eco-friendliness of firms in their corporate lending decision? Some empirical evidence. Int. Rev. Financ. Anal.
25, 83–93.
Orlitzky, M., Schmidt, F.L., Rynes, S.L., 2003. Corporate social and financial performance: a meta-analysis. Org. Stud. 24 (3), 403–441.
Orlitzky, M., Siegel, D.S., Waldman, D.A., 2011. Strategic corporate social responsibility and environmental sustainability. Bus. Soc. 50 (1), 6–27.
Preston, L.E., O’Bannon, D.P., 1997. The corporate social-financial performance relationship: a typology and analysis. Bus. Soc. 36 (4), 419–429.
Riedl, E.J., Serafeim, G., 2011. Information risk and fair values: an examination of equity betas. J. Acc. Res. 49, 1083–1122.
Renneboog, L., Horst, J.T., Zhang, C., 2008. Socially responsible investments: institutional aspects, performance, and investor behavior. J. Bank. Financ. 32 (9),
1723–1742.
Sanchez, J.L.F., Sotorrio, L.L., 2007. The creation of value through corporate reputation. J. Bus. Ethics 76 (3), 335–346.
Scholtens, B., 2008. A note on the interaction between corporate social responsibility and financial performance. Ecol. Econ. 68 (1–2), 46–55.
Scholtens, B., 2009. Corporate social responsibility in the international banking industry. J. Bus. Ethics 86 (2), 159–175.
Scholtens, B., Dam, L., 2007. Banking on the equator: are banks that adopted the equator principles different from non-adopters? World Dev. 35 (8), 1307–
1328.
Schroder, M., 2007. Is there a difference? The performance characteristics of SRI equity indices. J. Bus. Financ. Acc. 34 (1&2), 331–348.
Scott, W.R., 2001. Institutions and Organizations. Sage, Thousand Oaks, CA.
Sethi, S.P., 1979. A conceptual framework for environmental analysis of social issues and evaluation of business response patterns. Acad. Manage. Rev. 4 (1),
63–74.
Shen, C.H., 2002. Credit rationing for bad companies in bad years: evidence from bank loan transaction data. Int. J. Financ. Econ. 7, 261–278.
Shen, C.H., Lee, C.C., 2005. Same financial development yet different economic growth—and why? J. Money Credit Bank. 38, 1907–1944.
Shen, C.H., Lin, C.Y., 2012. Why government banks underperform: a political interference view. J. Financ. Intermed. 21, 181–202.
Silberhorn, D., Warren, R.C., 2007. Defining corporate social responsibility; a view from big companies in Germany and the UK. Eur. Bus. Rev. 19 (5), 352–
372.
Simpson, W., Kohers, T., 2002. The link between social and financial performance: evidence from the banking industry. J. Bus. Ethics 35, 97–109.
Soana, M., 2011. The relationship between corporate social performance and corporate financial performance in the banking sector. J. Bus. Ethics
(November), 104–133.
Spence, M., 1973. Job market signaling. Quart. J. Econ. 87 (3), 355–374.
UNTT (UN System Task Team), 2013. Financing for Sustainable Development: Challenges in Raising Private Sector Resources for Financing Sustainable
Development (Chapter 3).
Wang, Q., Dou, J., Jia, S., 2015. A meta-analytic review of corporate social responsibility and corporate financial performance: the moderating effect of
contextual factors. Bus. Soc., 1–39 http://dx.doi.org/10.1177/0007650315584317.
Watchman, P., 2005. Beyond the equator. Environ. Financ. 6 (June), 16–17.
Watchman, P., Freshfields, P., Deringer, B., 2006. Banks, business & human rights. Butterworths J. Int. Bank. Financ. Law, 46–50.
White, H., 1980. A heteroskedastic-consistent covariance matrix estimator and a direct test of heteroskedasticity. Econometrica 48, 817–838.
Wu, M.W., Shen, C.H., 2011. The elusive effect of bank size on profits. Serv. Ind. J. 31 (11), 1703–1724.
Wu, M.W., Shen, C.H., 2013. Corporate social responsibility in the banking industry: motives and financial performance. J. Bank. Financ. 37 (9), 3529–3547.

Artikel 6

Uploaded by

Copyright:

Available Formats

Artikel 6

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Artikel 6

Uploaded by

Copyright:

Available Formats

J. Int. Financ. Markets Inst.

Money 52 (2018) 240–261

Contents lists available at ScienceDirect

Journal of International Financial

Environmental risk management and financial performance in

2. The Equator Principles (EP)3

3. Literature review and hypotheses development

4. Sample and methods

Developed countries 52 (67%) Developing countries 26 (33%)

Bank Country Region Incorporation year Assets in US$ (Billions)

4.1. Performance measures

4.2. Control variables

Developed countries Developing countries

Developed countries Developing countries

5. Empirical tests and results

5.1. Short-term performance analysis

5.1.1. Event study analysis

5.1.2. Short-term changes in performance following EP adoption

5.2. Long-term performance analysis

EP Adoptionikt ¼ b0 þ b1 Ln TAikt þ b2 Capitalizationikt þ b3 Loan Depositikt þ b4 Cov erageikt þ b5 Cost Incomeikt

Performikt ¼ b0 þ b1 EP Adopterikt þ b2 Ln TAikt þ b3 Capitalizationikt þ b4 Loan Depositikt þ b5 Cov erageikt

M. Finger et al. / J. Int. Financ. Markets Inst. Money 52 (2018) 240–261

LG NII ROE ROA NPL

5.3. Robustness tests

Performikt ¼ b0 þ b1 Crisis þ b2 EP Adopter ikt þ b3 ðEP Adopter CrisisÞikt þ b4 Ln TAikt þ b5 Capitalizationikt

LG NII ROE ROA NPL

Panel A: Developed versus developing countries

Name Year of adoption Country

(continued on next page)

Name Year of adoption Country

You might also like