CI

Ca

b

a

ARRAA

JG

KCISS

1

scria

b

0

Journal of Economic Behavior & Organization 103 (2014) S21–S38

Contents lists available at ScienceDirect

Journal of Economic Behavior & Organization

j ourna l h om epa ge: w ww.elsev ier .com/ locate / jebo

orporate social responsibility and financial performance inslamic banks

hristine Mallinb,∗, Hisham Faraga, Kean Ow-Yonga

Birmingham Business School, University of Birmingham, Birmingham B15 2TT, UKNorwich Business School, University of East Anglia, Norwich NR4 7TJ, UK

r t i c l e i n f o

rticle history:eceived 17 November 2012eceived in revised form 28 February 2014ccepted 3 March 2014vailable online 6 April 2014

EL classification:21

eywords:orporate social responsibility (CSR)

slamic banksocial reportinghari’ah Supervisory Board

a b s t r a c t

This paper examines the relationship between corporate social responsibility (CSR) andfinancial performance in Islamic banks. Using a comprehensive CSR index covering tendimensions, we analyse the CSR disclosures in a sample of 90 Islamic banks across 13countries. The CSR disclosure index shows that Islamic banks engage across the range ofsocial activities, both as individual banks and as countries. However Islamic banks seem toshow more commitment to the vision and mission, the board and top management, andthe financial product/services dimensions, whilst least attention is paid to the environmentdimension. Islamic banks also show a considerable awareness of the mandatory disclosurerecommendations of the Accounting and Auditing Organisation for Islamic Financial Insti-tutions (AAOIFI) however, they pay less attention to the voluntary CSR disclosure. Moreover,we find a pronounced emphasis in Islamic banks strategy towards more universal disclo-sures, suggesting the legitimacy of these banks is reinforced through disclosure to the widerstakeholder community. The empirical analysis highlights a positive association betweenCSR disclosure and financial performance. We also find a positive and highly significant asso-ciation between the Shari’ah supervisory board (SSB) size and CSR disclosure index. Finally,the results of the three-stage least squares estimation show that the causality betweenthe two endogenous variables runs from financial performance to CSR disclosure. Thus CSRdisclosure is determined by financial performance.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Islamic banking has grown unabated since its inception in the mid-1970s. The industry has increasingly carved out aignificant slice of the global financial market. According to figures released by the Banker,1 global Islamic assets held byommercial banks exceeded US $1.8 trillion in 2013. All financial institutions, both conventional and Islamic, play a centralole in society. Hence they are expected to be responsive to the different needs of stakeholders. Due to their religious

dentity, Islamic banks are expected to be more socially responsible than their conventional counterparts whose operationsnd functions are primarily based on profit maximisation.

Islamic banking, from a theoretical perspective, is based on the principle of profit and loss sharing in place of the interestased deposit/lending found in conventional banks. Two conflicting juristic views have emerged in contemporary Islamic

∗ Corresponding author. Tel.: +44 01603597240.E-mail addresses: c.mallin@uea.ac.uk (C. Mallin), h.farag@bham.ac.uk (H. Farag), k.h.ow-yong@bham.ac.uk (K. Ow-Yong).

1 The Banker, March 2013.

http://dx.doi.org/10.1016/j.jebo.2014.03.001167-2681/© 2014 Elsevier B.V. All rights reserved.

S22 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

banking. Progressive Islamic scholars argue that there is no need to reinvent products offered by conventional banks in aglobally competitive banking industry. Instead, Islamic banks should adopt the minimal necessary modifications to theseconventional products to ensure Shari’ah compliance. This tendency to emphasise ‘form over substance’ (Warde, 2013 citedin Belal et al., 2014) is symptomatic of ‘big businesses’ driven by the profit making maxim. In contrast, those scholars whooppose conventional financial practice feel that the Islamic banking system needs to reconstruct pre-modern contracts bystrictly embedding Shari’ah and social responsibility into the banks’ business practices (El-Gamal, 2006). According to thisview, Islamic banks are expected to perform the role of redistributing wealth (through profit and loss sharing) to selectiveinvestments that contribute to the improvement and well-being of society (Farook, 2008). These Islamic banks practicethe ‘moral economy’ philosophy expounded for religious or secular ethical reasons and support the inclusion of social andenvironmental aims in their investment policies (Belal et al., 2014). Islamic banks, according to this view, should striveto achieve a balance between providing sufficient returns to their shareholders and depositors while at the same time notneglecting their social responsibilities and commitments to their various stakeholders (Ahmad, 2000). Recent studies suggestthat Islamic banks have failed to put the profit-loss principle into practice (Dar and Presley, 2000; Chong and Liu, 2009).Thus their findings point to the close similarity between Islamic and conventional banking practices and suggest that thealleged benefits of Islamic banking exist only in theory.

Concerning their social role, Islamic banks are expected to bring about economic and social benefits to their stakeholders;and to fulfil their Corporate Social Responsibility (CSR) including its disclosure. Farook (2008) argues that disclosure providesevidence of the Islamic banks involvement in social activities and hence earns legitimacy for their existence.

Islamic Financial Institutions (IFIs) may not be disclosing their social responsibility publicly, even though they are carryingout these activities. Therefore, to encourage disclosure, the international regulatory authorities such as the Accountingand Auditing Organisation for Islamic Financial Institutions (AAOIFI) developed reporting standards for Islamic banks. Inparticular, it issued Standard No.7 on Governance Standards for Islamic banks in relation to Corporate Social Responsibility(CSR) conduct and disclosure in 2010.2 In the Standard, CSR for (IFIs) is defined as ‘all activities carried out by an IFI to fulfilits religious, economic, legal, ethical and discretionary responsibilities as financial intermediaries for individuals and institutions.’Therefore, in complying with these standards, Islamic banks report aspects of their business activities and results differentlyfrom those of their conventional bank counterparts.3

The existing body of the literature on CSR in Islamic banks can be grouped into two broad strands. The first stranduses content analysis to explore the disclosure of CSR as described in the Islamic banks’ annual reports (Maali et al., 2006;Haniffa and Hudaib, 2007; Abdul Rahman et al., 2010; Hassan and Harahap, 2010; Aribi and Gao, 2012). The second strandinvestigates the determinants of CSR disclosure (e.g., Farook et al., 2011). This second strand is in the early research stageas – to the best of our knowledge – there is only one empirical study namely Farook et al. (2011). However, other possibledeterminants of CSR like financial performance (FP) have not yet been investigated empirically in Islamic banks.

Our paper is motivated to bridge a perceived gap between the two broad strands on CSR disclosure. We integrate Maali et al.(2006) and Haniffa and Hudaib (2007) benchmark CSR indices with AAOIFI Standard No.7 recommendations on mandatoryand voluntary CSR disclosure to produce a more comprehensive index covering ten dimensions to identify the type andthe extent of CSR disclosure for a sample of Islamic banks over the period 2010–2011. Furthermore, motivated by thestudy of Belal et al. (2014), we split the CSR index into two strands namely; ‘Particular’ reporting practices relating toShari’ah compliance issues and ‘Universal’ reporting practices which are more relevant to conventional banks and thewider stakeholder groups such as community, employees and customers. Secondly, we investigate the impact of financialperformance on CSR disclosure in Islamic banks which has not been empirically researched before; and finally, the paperinvestigates the direction of causality between financial performance and the CSR disclosure. Our data set covers a largesample of Islamic banks (90 Islamic banks4) and the CSR disclosure data is collected not only from the annual reports butalso from the information posted on the Islamic banks’ websites.5

This paper makes two incremental contributions to the literature on CSR and Islamic banks. First, although there havebeen a few empirical studies investigating the link between CSR and financial performance in the banking sector, as far as weare aware, this is the first study to empirically investigate this relationship in Islamic banks using a more comprehensive CSRdisclosure index which distinguishes between mandatory and voluntary disclosures as recommended in AAOIFI StandardNo.7. Secondly, we also classify the CSR index items into two main categories. The first includes items expected to be foundin Islamic banks and the second category includes those expected to be found in banks in general.

The paper is structured as follows. Section 2 discusses the literature review and hypotheses development. The sample

selection and research methodology are presented in Sections 3 and 4 respectively. Section 5 presents the results anddiscussion of the findings, whilst section 6 summarises and concludes.

2 This Standard provides both mandatory and recommended guidelines on CSR disclosure by Islamic banks. AAOIFI standards are compulsory for IBsoperating in Bahrain and Qatar but voluntarily applied to IBs in other countries.

3 For example, Islamic banks might be expected to include within their annual reports or within a separate report of the SSB, a section about the roleand function of their Shari’ah Supervisory Board (SSB), the sources and uses of zakah and charity funds and the unrestricted investment accounts held.Disclosing these activities will reveal the extent to which Islamic banks fulfil their socio-economic objectives for the benefit of the Islamic community.

4 Maali et al. (2006) and Farook et al. (2011) pointed out that further research with a larger sample of Islamic banks may be worthwhile.5 Previous studies covered a much smaller sample and collected data only from annual reports.

2

2

sbatsIaaUcTcc

batiurT‘w

aIItsisr

tii

sesneowoTsc

tp

r

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S23

. Literature review and hypotheses development

.1. Previous research on CSR disclosure

A number of studies have investigated the level of CSR disclosure in IFIs. Maali et al. (2006) investigated the extent ofocial activities disclosed by Islamic banks. They compared the disclosures of such activities with social practices that Islamicanks are expected to adopt that are particularly relevant to society. The expected social practices were then combined into

benchmark for social reporting by Islamic banks.6 Using content analysis, they found that the level of social disclosure byhe sample banks was well below their benchmark index. They concluded that banks complying with mandated mattersuch as paying Zakah and adopting AAOIFI standards tend to provide more disclosures than non-compliant banks. Also,slamic banks tend to accentuate disclosures that construct a positive Islamic image such as charitable activities. Haniffand Hudaib (2007) investigated the disclosure of information deemed crucial to Islamic ethics in business. They designedn ideal ethical disclosure benchmark based on the five features7 that distinguish Islamic banks from conventional banks.sing content analysis of annual reports to determine the extent of CSR disclosure, they found a significant gap between theommunicated and ideal ethical disclosure in the annual reports of a sample of seven Islamic banks over a 3-year period.hey surmised that this expectations gap could arise from an indifferent attitude by stakeholders or an underlying secretiveulture in the Arabian Gulf region where the sample banks were based. They concluded that for Islamic banks to remainompetitive, they need to communicate more effectively to enhance their image and reputation in society.

Using a longitudinal case study approach, Abdul Rahman et al. (2010) examined the CSR disclosure of a Malaysian Islamicank. They found that the volume and quality of CSR disclosure improved over the 14-year study period. Belal et al. (2014)lso adopted a longitudinal study but with two notably significant exceptions. First, using a longer time frame of 28 years,hey noted trends in CSR disclosure of a Bangladeshi Islamic bank over three distinguishing phases of the Islamic bankingndustry: pre 1990, 1990–2001 and post 2001. Secondly, they split ethical reporting requirements into two strands; ‘Partic-lar’ reporting practices relating to the nature of Islamic banks and especially to Shari’ah compliance issues and ‘Universal’eporting practices which are more relevant to the wider stakeholder groups such as community, employees and customers.hey found an overall increase in both ‘Particular’ and ‘Universal’ disclosures over the study period with a shift towards more

Universal’ disclosures after 2006. They interpret their findings as being the bank’s adherence to the minimalist approachithout explicitly contravening the wider perspective of Shari’ah.

Hassan and Harahap (2010) carried out a similar study to Haniffa and Hudaib (2007) focussing on the disclosure of socialctivities in the annual reports of seven Islamic banks. They found a significant expectations gap in all but one of the sevenslamic banks, and surmised that CSR issues are not the main concern for most Islamic banks. They concluded that someslamic banks pay scant attention to disclosing their social activities and thus argued for a standard on CSR disclosure relevanto IFIs. Farook et al. (2011) provided an a priori basis for CSR disclosure in Islamic banks. They empirically analysed the level ofocial disclosure in forty-seven Islamic banks’ annual reports based on a CSR benchmark derived from the Maali et al. (2006)ndex. They found substantial variation in CSR disclosures and this variation is best explained by the presence of Shari’ahupervisory board (SSB) governance and the preponderance of Muslims in their sample countries. They concluded that theegulatory bodies in Islamic banking should consider the SSB as being compulsory for all Islamic banks.

Aribi and Gao (2012) analysed the narrative disclosures of CSR in 21 IFIs operating in the Gulf countries. They found thathe main CSR disclosures were contained in the SSB reports with less disclosure in the annual reports on other Islamic basednformation such as Zakah, interest free loans and charitable donations. Table 1 presents the main studies on CSR disclosuren Islamic banks.

The CSR literature on Islamic banks is largely qualitative based studies that measure the volume of narrative CSR disclo-ures against an ideal benchmark drawn from Shari’ah based CSR objectives and AAOIFI standards. They generally find anxpectations gap between actual/communicated disclosures and ideal disclosures. However, the main limitations of thesetudies are their reliance on data collected from the annual reports to infer CSR disclosure. The annual report itself willot provide the true picture of CSR disclosure as Islamic banks may disclose some of their CSR separately in other reports,.g., SSB report, corporate governance report and on their websites. We collect data from the annual reports and all thether available reports as well as the Islamic banks websites. Secondly, the number of sample banks used in the literatureas limited as acknowledged by Maali et al. (2006) and Haniffa and Hudaib (2007). Our sample of Islamic banks is not

nly relatively large (90) compared to the largest study so far (47 in Farook et al., 2011) but also more recent data in 2011.he most recent year investigated was 2007 (Farook et al., 2011). Thirdly, although a number of studies referred to AAOIFI

tandards (Hassan and Harahap, 2010), none of them included AAOIFI Standard No.7 on CSR conduct in which CSR disclosurelassification distinguishes mandatory from voluntary CSR disclosure requirements.

6 Their social disclosure benchmark index, containing thirty items, covered the reporting of Shari’ah Supervisory Board opinion; unlawful transactions;he payment of Zakah; payment for Quad Hassan, charity and other social activities; policies on employees; policies on late repayment and insolvent clients;rotecting the environment; and other aspects of community involvement.7 These five features, discussed in detail in their paper, are the bank’s underlying philosophy and values; provision of interest-free products and services;

estriction to Islamic acceptable deals; focus on development and social goals; and reviews by the Shari’ah Supervisory Board.

S24 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Table 1Overview of corporate social responsibility disclosure studies in Islamic banks.

Authors Year Institution type and countries insample

Sample size and year Measure of CSR

Maali et al. (2006) Islamic banks around the world 29 (2000) Disclosure index containing 30 itemsHaniffa and Hudaib (2007) Islamic banks in Gulf countries 7 (2002–2004) Ethical Identity Index containing 78

itemsAbdul Rahman et al. (2010) Islamic banks in Malaysia 1 (1992–2005) Adapted Maali indexHassan and Harahap (2010) Islamic banks in Bahrain, Bangladesh,

Indonesia, Malaysia, Saudi Arabia,Kuwait, UAE

7 (2006) Adapted Haniffa and Hudaib index

Farook et al. (2011) Islamic banks around the world 47 (2007) Adapted Maali indexAribi and Gao (2012) Islamic banks in Gulf countries 21 (2004) Narrative disclosures

Belal et al. (2014) Islamic banks in Bangladesh 1 (1983–2010) Disclosure index containing 149 items

across 16 categories

2.2. Corporate social responsibility (CSR) and financial performance (FP)

Numerous theoretical and empirical studies have long investigated CSR impact measured by social performance and itsrelationship to FP. However there is a controversy in the results due to the discrepancies in the theoretical and methodologicalframeworks (see for instance, Preston and O’Bannon, 1997; Griffin and Mahon, 1997; and Waddock and Graves, 1997). Alimited number of studies focussed on the banking industry (Simpson and Kohers, 2002; Soana, 2011, Ahmed et al., 2012).However, there is no previous study that examines empirically the CSR-FP relationship specifically in the Islamic bankingindustry. Therefore a better understanding of this link would be invaluable, directly or indirectly, to all stakeholders includingmanagement, shareholders and the Islamic community. Various hypotheses sought to explain the link between CSR andFP. The main findings of these studies are controversial as they offer conceptual interpretations for a positive, neutral andnegative relationship between social and financial performance. See for example Preston and O’Bannon (1997) and Waddockand Graves (1997).

Neoclassical economists are the proponents of the negative association between social and financial performance (e.g.,Simpson and Kohers, 2002). They argue that firms which meet the social needs of their stakeholders will incur a competitivedisadvantage resulting in reduced profits because such social costs could be avoided or borne by others (e.g., the government).It could be argued from an Islamic bank perspective, that assisting to develop large scale environmental and communityprojects may have an adverse impact on its profitability.

Stakeholder theory assumes that there might be a positive relationship between social and financial performance. Waddockand Graves (1997) argue that the benefits from CSR are greater in comparison with its costs. Therefore, there should be apositive association between CSR and FP. Preston and O’Bannon (1997) argue that meeting the needs of various corporatestakeholders enhances a firm’s reputation in a way that will have a positive impact on its FP. The empirical findings ofSimpson and Kohers (2002), based on data from a sample of US commercial banks, support the notion of a positive social –financial performance link.

The central role played by Islamic banks within their communities suggest that being actively engaged in social and ethicalactivities enhances their reputation leading to an expected higher FP. Finally, the empirical finding of a simply neutral (non-existent) relationship between social and financial performance might be due to the complex relationship between societyand the firm which cannot be captured through a simple direct relationship. Islamic banks following Shari’ah principles areexpected to offer profit and loss sharing schemes for Investment Account Holders and depending on their policies may payZakah on behalf of their customers as well as dispensing discretionary benevolent loans (Qard Hassan) to the community.The effect of such business activities could be varied and complex resulting in a neutral relationship between CSR and FP.Based on the above discussion, we formulate our first hypothesis as follows:

H1. There is a relationship between an Islamic bank’s CSR activities disclosure and its financial performance.

2.3. The direction of causality between CSR and FP

Two perspectives can be contrasted and tested empirically. Waddock and Graves (1997) coined the first as the ‘slackresources theory’ and the second ‘good management theory’. Firms which have superior FP will then have slack resourcesavailable to spend on investing in socially responsible activities such as improving employee and community relations.Investing in these social domains would result in better social performance. It suggests that the surpluses generated byIslamic banks from strong FP will be invested in Shari’ah compliant socially responsible activities. As Shari’ah objectivesplace emphasis on promoting justice and welfare in society, Islamic banks which perform well financially are expected to do

good through undertaking socially responsible activities which benefit all stakeholders including the community. To sumup, under the “slack resources theory”, the CSR-FP relationship runs from FP to CSR. This suggests that CSR is determined byFP (Waddock and Graves, 1997).

sbaCdhms

H

3

mmbbbbKdstt

3

HatpCppfdie

wae

m

2

torrcp

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S25

Conversely, Waddock and Graves (1997) surmise that firms with good management who pay attention to socially respon-ible activities such as employee training and relations may expect to reap better FP later. So, publicity through disclosurey Islamic banks of their investment in activities that comply with Shari’ah such as sponsoring Islamic educational eventsnd supporting employees fulfilling their Haj may lead to more growth and profitability. The “social impact hypothesis” ofornell and Shapiro (1987) provides a consistent view with the good management theory as it suggests that satisfying theifferent needs of stakeholders will improve the reputation of the company and lead to better FP. However, the trade-offypothesis of Vance (1975) provides a consistent interpretation with the views of the neoclassical economists.8 Under goodanagement theory, social impact and the trade-off hypotheses, CSR would be a predictor of FP. Therefore we formulate our

econd hypothesis as follows:

2. The relationship between CSR and financial performance of Islamic banks is bi-directional.

. The sample and data

Our dataset is a cross-sectional analysis of the relationship between CSR disclosure and Islamic banks financial perfor-ance over the period 2010–2011.9 We use Bankscope and the Bankers databases for the sample selection. The Bankersagazine published a survey in November 2011 of the top Islamic financial institutions by country. There are 160 Islamic

anks with 100% Shari’ah compliant assets listed in this study. For the sake of consistency in our sample, we excluded Islamicanks in both Iran (18 Islamic banks) and Turkey (4 Islamic banks) as they do not have the SSB. We also excluded Islamicanks which provide only financial statements (37 banks). In addition, we excluded subsidiaries from our sample (11 Islamicanks). Therefore, we collect data for 90 Islamic banks from 13 countries namely Bahrain, Bangladesh, Indonesia, Jordan,uwait, Malaysia, Pakistan, Qatar, Saudi Arabia, Sudan, Syria, United Arab Emirates (UAE), and United Kingdom (UK). Theataset is hand collected from the annual reports and the websites of the respective banks. Data were collected from severalources including Bankscope, the Banker database, Perfect Information Navigator, and Companies House,10 in addition tohe annual reports and websites. Financial information was collected from Thomson One Banker and Bankscope in additiono the annual reports.

.1. Dependent variable

Our dependent variable is the CSR index. The CSR disclosure index incorporates items from the Maali et al. (2006) andaniffa and Hudaib (2007) studies. In addition, we reclassify the index items into two main categories namely mandatorynd voluntary aspects based on AAOIFI Standard No. 7.11 As a further study, and following Belal et al. (2014) we reclassifyhe index items into two categories namely particular and universal, the first category consisting of CSR items relatingurely to Shari’ah principles and the second category consists of CSR items generally expected in conventional banks. OurSR disclosure index covers a broad range of activities (e.g., compliance with Shari’ah principles, provision of interest freeroducts, treatment of Zakah, maintaining good relations with employees and involvement in charity, social activities androtecting the environment). The index consists of 10 dimensions. Our checklist consisting of 84 items12 also includes itemsrom AAOFI Standard No.7.13 We checked each item across the 90 annual reports and then we dealt with each item as aummy variable which takes the value of one if the item was found in the annual reports/websites and zero otherwise. Our

ndex is equally weighted to avoid any potential scoring and scaling biases following Haniffa and Hudaib (2007) and Maalit al. (2006) as shown in Eq. (1).

CSRIi =∑n

t=1Xi

n(1)

here CSRIi is corporate social disclosure index 0 ≤ CSRIi ≤ 1,ni is the number of items expected for bank i n ≤ 84 and Xi is dummy variable which takes the value of 1 if the item is disclosed and 0 otherwise. We took precautionary measures tonhance the validity and the reliability of our analysis. We checked that the index items generated from the classification

8 The trade-off hypothesis of Vance (1975) implies that companies engaging in CSR activities unnecessarily incur costs that may reduce their profitabilityeasured by earnings per share (EPS). Therefore, the trade-off hypothesis supports the negative association between CSR and FP.9 We use cross –section regression rather than panel data analysis as the vast majority of Islamic banks disclosed CSR in their annual reports in both

010 and 2011. We could not find enough data on IBs pre 2010.10 This source is used to collect data on Islamic banks in the UK.11 We thank our anonymous referee for the comments and suggestions regarding the classification of the index based on AAOIFI Standard No.7.12 For a full exposition of the 84 items please refer to Haniffa and Hudaib (2007) and Maali et al. (2006).13 The CSR responsibilities expected from IFIs fall into two categories, namely mandatory and voluntary. The mandatory category relates to IFI activitieshat are imperatives and bound by the command of God. These mandatory activities relate to screening and dealing responsibly with clients; actionn earnings and expenditure prohibited by Shar’iah; paying due attention to employees welfare and establishing terms and conditions for Zakah. Theecommended category relates to desirable activities that IFIs are encouraged to pursue and support from an Islamic perspective. These are policieselating to implementing Qard Hasan, reduction of adverse impact on the environment; investing in social, development and environment based projects;ommitment to par excellence customer service; assisting micro and small businesses; establishing and supporting charitable activities; and managingroperties dedicated to Muslim societies (Waqf).

S26 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

procedures represented what we intended to represent. The authors examined the items of the index and decided whatthat specific item was intended to measure (Beattie et al., 2004). In addition, we made sure that each item and the overallindex are closely related to CSR in Islamic banks as we carefully chose and developed the 10 dimensions of our CSR indexbased on the previous literature and the CSR Conduct and Disclosure for IFIs issued by AAOIFI. To enhance the reliability,the index items are coded and checked twice and we discussed any potential discrepancies. It is worth mentioning thateach bank is coded by two different authors to ensure consistency. We made sure that the same coder is consistent overtime when coding the same item of the index (stability), the coders produce the same results when coding the same item(reproducibility) and accuracy as well (Beattie et al., 2004). We then rank each bank based on its CSR index score. In additionwe rank each country based on the weighted average dimensions of the CSR index.

3.2. Independent variables

Return on assets and return on equity are widely used measures of bank performance and sustainable growth. However,during a period of financial uncertainty and low profitability, discretionary social responsibility expenditure has less prioritycompared to economic demands. Roberts (1992) argues that a satisfactory level of FP has a direct impact on the decisions bythe board of directors and their commitment to future CSR activities and expenditures. Based on the theoretical frameworkand the discussion presented in the hypotheses development section, we expect an association between CSR and FP. Aconsiderable strand of the CSR literature uses either the contemporaneous or 1-year lagged return on equity (ROE) and/orreturn on assets (ROA) as a measure of bank performance (for example, Mallin and Michelon, 2011; McWilliams and Siegel,2001; Balabanis et al., 1998; Waddock and Graves, 1997). Therefore, to avoid any endogeneity problems we use lagged inaddition to contemporaneous measures of performance. On the other hand, we argue that using a 1-year lagged ROE/ROAmight not be the optimal choice. This may lay the estimation open to performance bias in a particular year (2009 and 2010in our case). Following the seminal paper of Roberts (1992), we use the average annual change in ROE/ROA over the period2006–201014 as a measure of bank performance.15

The SSB has social influence and authority in monitoring the IB’s compliance with Shari’ah principles, and provides theconfidence to stakeholders about the legitimacy of the business transactions. Disclosure by the SSB may be seen as a keyaspect of accountability by the IB to its stakeholders. However, the degree to which social activities are disclosed dependson the underlying rationale of the SSB monitoring role on behalf of the Islamic bank’s investors. Moreover, the degree towhich the SSB influences the level of CSR disclosure may depend on characteristics such as its board size. Therefore, SSB sizeis expected to have a positive impact on CSR disclosure16 (Farook et al., 2011). The governance literature indicates the levelof disclosures could increase or decrease depending on whether mechanisms substitute or complement each other (Ho andWong, 2001). When governance mechanisms substitute each other, firms may not provide additional disclosures since themultiple governance mechanisms should have increased the level of monitoring. Therefore, the SSB may see no need to urgefor additional CSR disclosure if investors are already assured of the Islamic bank’s compliance in the Shari’ah compliancereport (Maali et al., 2006). Alternatively, if the governance mechanisms are complementary, agency theory suggests that ahigher level of disclosures is expected as more governance mechanisms will strengthen the monitoring aspect leading toa reduction in information asymmetry and opportunistic behaviour. The SSB in its role as an additional governance bodywould pressure Islamic bank to disclose more CSR activities in order to assure its investors that it is following Shari’ahlaws and principles. Notwithstanding the ambiguity raised, it is expected that the SSB acts as a complementary mechanismbecause compliance with Islamic laws and principles should not only be generally inferred within the Shari’ah report butalso reinforced with disclosures of specific CSR activities. Hence, it is generally expected that the SSB in an IB would lead tohigher disclosure levels of CSR activities.

We also control for board size and the proportion of non-executive directors (NEDs) to capture the effect of board structureand board independence on CSR disclosure.17 Pfeffer (1973) and Pfeffer and Salancik (1978) highlight the role of board sizeas an important indicator of a company’s need to link itself with the external environment. They argue that larger boardsmay form a greater number of committees, e.g., governance and CSR committees. In terms of resource dependence theory,outside shareholders may bring managerial know-how and hence help improve banks disclosure. Zahra et al. (1993) findthat the proportion of NEDs leads to racial, ethnic and gender diversity of the board and this in turn results in better impact

on CSR disclosure. Pfeffer and Salancik (1978) and Johnson and Greening (1999) argue that NEDs may enhance the reputationand credibility of the company and help to establish and maintain its legitimacy. Therefore, we may expect that there is apositive relationship between CSR and both board size and SSB size in addition to the proportion of NEDs.

14 We also used average annual change on ROE and ROA over the period 2006–2009 as a measure of lagged bank performance and obtained similar results.15 As not all the IBs in our sample are listed in the stock exchanges we could not use Tobin’s Q ratio as a measure of bank performance.16 The minimum number of any SSB is three according to AAOIFI standards (2008) and this is a common requirement in many Islamic banks. AAOIFI

recommends professionals other than religious scholars to sit on the SSB. These other professionals include bankers and economists. For this to happen,the size of the SSB would have to be large to accommodate the diversity of its members. The inclusion of business professionals with less religious standingmay increase the need for more CSR activities disclosure to convince the investors that the Islamic bank’s adherence to Shari’ah principles has not beencompromised.

17 Chair/CEO role duality is included in the CSR index.

aifiBaiI(tc

4

i

w

2

seliofttFf

a

wnm

FFvu2R

t

ah

s

t

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S27

Moreover, we use other control variables namely the natural logarithm of total assets as a proxy for bank size and bankge; it is argued that bank size has an impact on the level of corporate social activities. Big banks, we believe, being moren the public eye, are highly likely to monitor their activities towards wider society. A large strand of the literature on CSRnds an association between size and CSR activities (see for example, Mallin and Michelon, 2011; Al-Tuwaijri et al., 2004;rammer et al., 2006; McWilliams and Siegel, 2001; Roberts, 1992). Roberts (1992) finds a positive relationship between agend CSR; she argues that the older the company the more involvement there will be in CSR activities which have a positivempact on its reputation. We also control for the macro-economic factors by using the natural logarithm of country’s GDP.18

n addition we use the ratio of total finances/total assets to control for the differences in IBs operating business efficiency19

Andres and Vallelado, 2008). Different stock exchanges impose constraints on listed companies with regard to disclosure,herefore we control for whether the bank is private or listed on a stock exchange. Finally, country dummies are used toapture country heterogeneity and to control for country specific effects.

. Research methodology

To empirically investigate the relationship between CSR and FP in Islamic banks, we use the following OLS regression asn Eq. (2):

CSRIi = ˛i + ˇ1FPi + ˇ2SSB.Sizei + ˇ3B.Sizei + ˇ4NEDi + ˇ5lnAgei + ˇ6lnTAi + ˇ7Fin/TAi + ˇ8privatei

+ ˇ9LnGDPi+i

n∑C=1

ˇ10Di + εi (2)

here CSRIi: is corporate social responsibility disclosure index for bank i. FPi is the average annual change in ROE/ROA over

006–2010 determined as follows: FPi =(∑N

n=1(FPt − FPt−1)/N)

. . .N : 2006 − 2010, SSB.Size is the number of Shari’ah

upervisory board members for bank i, B.Size is the number of board members for bank i,NED: is the percentage of non-xecutive directors in the board for bank i, lnAge is the natural logarithm of bank age since foundation, lnTA is the naturalogarithm of bank’s total assets in US$ as a proxy for bank size, Fin/TA is the ratio of total bank finances/total assets20 for bank, Private is a dummy variable which takes the value of 1 if the bank is listed in a stock exchange of the respective country and 0therwise,21 lnGDP is the natural logarithm of the gross domestic product of country i as a proxy for country macroeconomicactors,

∑nC=1Di are country dummy variables take the value of 1 for the respective country and zero otherwise and εi is

he white noise error term. We use the first-order Taylor-series linearisation method to control for heteroskedasticity ando produce robust standard errors. In addition we use both lagged and contemporaneous independent variables in Eq. (2).inally we use the Ramsey RESET test for omitted variables and model mis-specification, we also use the variance inflationactors (VIF) to examine whether the independent variables are perfectly collinear.

To explore the interrelationship between CSR and FP, we formulate the following system of simultaneous equations thatddress the potential endogeneity issues in the estimation.

CSR = f1(FP, Z1, ε1) (3a)

FP = f1(CSR, Z2, ε2) (3b)

here Zi are the vector of control variables and instruments influencing the dependent variables; and εi are the whiteoise error terms associated with the unobservable effects resulting from firm heterogeneity i.e. unobservable features ofanagerial behaviour that explain heterogeneity in CSR and FP.Two concerns may arise regarding endogeneity when examining the relationship between CSR and financial performance.

irst, our findings may simply reflect some underlying omitted variable that influences both endogenous variable (CSR andP) (Nikolaev and Van Lent, 2005). The second concern is the potential reverse causality between the two endogenousariables. We estimate the instrumental variable two-stage least squares (2SLS) and the three-stage least squares (3SLS)22

sing a system of two simultaneous equations to investigate this bi-directional relationship using the pooled sample over010–2011 as in Eqs. (3a) and (3b). Eq. (3a) estimates the impact of bank financial performance measured by both ROA and

OE on the CSR disclosure index while Eq. (3b) estimates the influence of CSR on the bank financial performance.

The choice of instrumental variables is critical to a consistent estimation. The choice of our instruments is motivated byhe extant literature. A valid instrument should reasonably predict the endogenous variable and not be correlated with error

18 We also used the ratio of deposits/GDP following Andres and Vallelado (2008) and obtained similar results.19 We also used three other control variables namely the cross directorship in SSB, the proportion of Muslim population following Farook et al. (2011)nd Tier 1 capital expressed as a percentage of total risk weighted assets to control for IBs risk. However the results show that these three variables areighly insignificant and correlate with the other independent variables. Results are not presented but available from the authors upon request.20 Includes all IBs’ financing activities.21 We also use bank visibility defined as a dummy variable that takes the value of 1 if the bank is one of the main index constituents for the respectivetock exchange and 0 otherwise. We will discuss this variable in the next section as a valid instrument.22 The essential advantage of 3SLS estimation techniques is that it allows, not only for simultaneity among the CSR and FP, but also for correlations amonghe error components. Thus, it is believed that 3SLS estimators are asymptotically more efficient than two-stage least square (2SLS) estimators.

S28 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Table 2Descriptive statistics of the CSR Index by country.

Year Mean SD Min Max Obs

Aggregate CSR Index 2010 0.4356 0.1263 0.1309 0.7024 902011 0.4439 0.1248 0.1310 0.7024 90

Bahrain 2010 0.5054 0.0893 0.2857 0.6190 202011 0.5125 0.0860 0.3095 0.6429 20

Bangladesh 2010 0.4246 0.0975 0.2500 0.5238 62011 0.4345 0.1090 0.2500 0.5595 6

Indonesia 2010 0.5387 0.0393 0.4881 0.5833 42011 0.5387 0.0393 0.4881 0.5833 4

Jordan 2010 0.3929 0.0000 0.3929 0.3929 32011 0.4008 0.0137 0.3929 0.4167 3

Kuwait 2010 0.3929 0.0491 0.3333 0.4286 52011 0.3952 0.0515 0.3333 0.4405 5

Malaysia 2010 0.5082 0.0982 0.3571 0.7024 162011 0.5156 0.0934 0.4048 0.7024 16

Pakistan 2010 0.3000 0.1089 0.1667 0.4167 52011 0.3048 0.1125 0.1667 0.4167 5

Qatar 2010 0.3738 0.1367 0.2381 0.5714 52011 0.3738 0.1367 0.2381 0.5714 5

Saudi Arabia 2010 0.4762 0.1124 0.3571 0.6190 62011 0.4960 0.1129 0.3574 0.6190 6

Sudan 2010 0.3146 0.0356 0.2738 0.3690 72011 0.3146 0.0394 0.2738 0.3810 7

Syria 2010 0.3333 0.2862 0.1310 0.5357 22011 0.3571 0.2525 0.1786 0.5357 2

UAE 2010 0.3027 0.2166 0.1309 0.6190 7

2011 0.3486 0.1818 0.1310 0.6190 7

UK 2010 0.4256 0.0791 0.3571 0.5357 42011 0.4464 0.0731 0.3810 0.5357 4

terms. Previous studies have used stakeholders’ characteristics (power and legitimacy), corporate governance characteristics(board structure and independence) and company visibility as valid instruments for CSR; see for example Brammer andMillington (2006); Eesley and Lenox (2006); Mitchell et al. (1997); Rehbein et al. (2004) and Garcia-Castro et al. (2010).Garcia-Castro et al. (2010) define company visibility as whether the company is listed in the Standard & Poor’s 500 (S&P 500)or not. They argue that listed companies in the S&P 500 have more exposure to investors, media, and thus higher visibility.

Due to the data limitations, we are unable to use the stakeholders’ characteristics as an instrumental variable for the CSR.Moreover, although corporate governance characteristics may have an influence on CSR (Brammer and Millington, 2006),they are likely to be correlated with financial performance (Bhagat and Bolton, 2008). Therefore we use, following Garcia-Castro et al. (2010), bank visibility as our instrument for CSR. However, we redefine bank visibility as a dummy variable thattakes the value of 1 if the bank is listed in the main market index (constituent) of the respective country and 0 otherwise.We assume that index constituents are more visible to investors and media and are likely to adopt consistent policies withstakeholders such as engaging in CSR. Therefore, we expect that our instrumental variable is likely to be correlated with CSRand not with financial performance. We believe that this variable satisfies the necessary conditions for valid instrumentsassuming that the disturbance is not autocorrelated23 (Kennedy, 2003). Furthermore, we use the Breusch and Pagan (1980)test of independence to investigate whether cross-equation disturbances are indeed correlated and if the equations needto be estimated simultaneously. Beiner et al. (2006) argue that the advantages of 3SLS depend on instrument validity andthe correct specification of the system. To test the instrument validity, we use the Sargan (1964) mis-specification testwith the null hypothesis of “No mis-specification”. If the null hypothesis is rejected then the model is likely to be eitherincorrectly specified and/or some of the instruments are invalid. In addition, to test for the correct specification of thesystem of simultaneous equations, we apply the Hausman specification test (Hausman, 1978) to compare between 2SLSand 3SLS estimates. The null hypothesis of the Hausman test is “the 3SLS results are consistent and efficient while the 2SLSresults are also consistent but inefficient”.

5. Results and discussion

In this section, we analyse the results of the CSR disclosure index of the 90 Islamic banks over the period 2010–2011.

Table 2 presents the descriptive statistics of the CSR Index scores across counties. The results show that the average CSRindex increases from 43.5% in 2010 to 44.3% in 2011 with standard deviation 12% on average. Thus, there is no significantannual variation in CSR index scores between 2010 and 2011; however, the index scores show that the extent of disclosure

23 We test for the serial correlation in residuals using both the Breusch–Godfrey–Lagrange Multiplier and Durbin–Watson tests. The results of the twotests show that the residuals are not serially correlated.

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S29

Fig. 1. CSR Index across countries in 2011.

Table 3Weighted average CSR Index by dimension in 2011.

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 WeightedIndex (%)

Average Dimension 69.75 62.82 60.44 17.41 43.21 39.17 23.17 45.09 3.33 54.44 41.88Bahrain 83.33 76.15 75.00 18.00 37.78 31.25 33.57 53.33 0.00 68.33 56.74Bangladesh 70.37 64.10 58.33 11.11 68.52 50.00 11.90 33.33 0.00 44.44 48.11Indonesia 69.44 67.31 57.50 31.67 63.89 62.50 50.00 47.92 0.00 75.00 59.64Jordan 33.33 84.62 36.67 8.89 37.04 50.00 19.05 50.00 0.00 66.67 44.37Kuwait 55.56 55.38 50.00 16.00 48.89 50.00 5.71 48.33 0.00 33.33 43.76Malaysia 79.86 66.83 78.13 24.17 53.47 29.69 22.32 53.13 9.38 62.50 57.09Pakistan 33.33 56.92 44.00 5.33 22.22 40.00 11.43 35.00 0.00 46.67 33.74Qatar 80.00 43.08 52.00 1.33 31.11 40.00 34.29 35.00 20.00 60.00 41.39Saudi Arabia 83.33 66.67 63.33 22.22 57.41 45.83 40.48 31.94 8.33 66.67 54.92Sudan 47.62 46.15 40.00 17.14 22.22 50.00 4.08 41.67 0.00 9.52 34.83Syria 50.00 53.85 40.00 16.67 38.89 37.50 0.00 50.00 0.00 33.33 39.54UAE 66.67 39.56 42.86 15.24 38.10 42.86 16.33 34.52 0.00 38.10 38.60

Dc

aBF

de(aaatwIfs

Svtpt

e

f

UK 72.22 67.31 57.50 23.33 36.11 18.75 21.43 45.83 0.00 66.67 49.43

1, . . ., D10 are vision and mission statement; board of directors and top management; product/services; zakah, and benevolent loans; employees; debtors;ommunity; shari’ah supervisory board (SSB); environment and charitable and social activities dimensions respectively.

cross countries varies considerably. Indonesia has the highest CSR index score of 53.8%, followed by Malaysia at 51.5% andahrain at 51.2% in 2011. The lowest scores are achieved by Pakistan and the Sudan, 30.4% and 31.4% in 2011 respectively.ig. 1 shows the CSR Index scores by country in 2011.

Table 3 presents the weighted average24 CSR index scores in 2011. As reflected in the analysis of the country’s CSRisclosures, we find that the vision and mission statement dimension generally scores highly across all countries whilst thenvironment generally scores the lowest. The highest disclosure score relates to the vision and mission statement dimensionD1) which is 69.75% followed by the scores of board of directors (D2) and financial products/services (D3) dimensions (62.82%nd 60.44% respectively). On the other hand, the lowest disclosure score relates to environment (D9) which has a weightedverage score of 3.33%. This finding is consistent with the perception that Islamic banks pay relatively little attention to CSRctivities relating to the environment whereas the board and top management are areas that successful banks which wanto comply with best practice corporate governance would place significant emphasis on. Our findings are also consistentith those of Kamla (2007) who found very low levels of environmental related disclosures by Arab companies including

slamic banks. We notice that the weighted average index scores indicate that Indonesia has the highest score of 59.64%ollowed by Malaysia and Bahrain 57.09% and 56.74% in 2011 respectively. Fig. 2 presents the weighted average CSR indexcores by country and dimension in 2011.25

We also reclassify the CSR index into two main categories namely mandatory and voluntary disclosure based on thetandard No. 7 on CSR issued by AAOIFI in 2010. Table 4 presents the CSR index scores classified by mandatory versusoluntary disclosures in 2011. The results show that the average mandatory CSR index increased from 60.27% in 201026

o 61.2% in 2011, however the average voluntary disclosure index increased from 34.3% in 2010 to 35.4% in 2011. Fig. 3resents the weighted average CSR mandatory and voluntary disclosure index by country in 2011. It is clear from Fig. 3hat Malaysia has the highest mandatory CSR index scores in 2011 (76%) followed by Bahrain, UK and Indonesia (70%, 68%

24 We are grateful to our anonymous referees; we weight the CSR index by the number of items within each dimension and the number of banks withinach country. For more details on the number of items included in each dimension see Haniffa and Hudaib (2007) and Maali et al. (2006).25 There is no significant annual change in the weighted average index scores between 2010 and 2011. The results of 2010 are not presented but availablerom the authors upon request.26 2010 figures are not presented in Table 4.

S30 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Fig. 2. CSR Index by dimension.

Table 4Classifications of the CSR Index by country in 2011.

Aggregate Index Particular Universal Mandatory Voluntary

Aggregate CSR Index 0.4439 0.4062 0.4798 0.6127 0.3542Bahrain 0.5125 0.4805 0.5430 0.6982 0.4116Bangladesh 0.4345 0.3211 0.5426 0.5595 0.3690Indonesia 0.5387 0.4451 0.6279 0.6429 0.4821Jordan 0.4008 0.2764 0.5194 0.4762 0.3631Kuwait 0.3952 0.3659 0.4233 0.5786 0.3036Malaysia 0.5156 0.4959 0.5058 0.7589 0.3884Pakistan 0.3048 0.2293 0.3767 0.4214 0.2464Qatar 0.3738 0.3073 0.4372 0.4643 0.3214Saudi Arabia 0.496 0.4146 0.5736 0.5893 0.4315Sudan 0.3146 0.2962 0.3322 0.5051 0.2194Syria 0.3571 0.3902 0.3256 0.5714 0.2500UAE 0.3486 0.3275 0.3688 0.4898 0.2730UK 0.4464 0.4634 0.4902 0.6786 0.3259

The table presents the classifications of the CSR Index into two main classifications namely particular and universal, and mandatory and voluntary CSRdisclosures. The first classification consists of CSR items relating purely to Shari’ah principles (particular) and the items generally expected in conventionalbanks (universal). The second classification is based on the Standard No. 7 on CSR issued by AAOIFI.

Fig. 3. Weighted Average CSR Index by disclosure requirements of AAIOFI in 2011.

and 64% respectively). However, Pakistan has the lowest mandatory CSR index scores in 2011 (42%) followed by Qatar andJordan (46 and 48% respectively). On the other hand, Indonesia has the highest voluntary CSR disclosure index in 2011 (48%)followed by Saudi Arabia (43%); whereas the Sudan has the lowest voluntary CSR index score of 22% in 2011 followed byPakistan and Syria (24.6% and 25% respectively).

To summarise, the results of the CSR disclosure index show that Islamic banks engage across the range of CSR activities,both as individual banks and as countries. However Islamic banks seem to show more commitment to the vision and missionstatement, the board and top management, and the financial product/services dimensions, whilst least attention is paid tothe environment dimension. Islamic banks also show a considerable awareness of the mandatory disclosures stated bythe AAOIFI however, they pay less attention to the voluntary CSR disclosures. We interpret these results empirically in thefollowing section.

Following Belal et al., 2014 we reclassify our CSR index into two main categories namely particular and universal, the

first category consists of CSR items relating purely to the nature of Islamic banks and in particular Shari’ah principles andthe second category consists of CSR items generally expected in conventional banks. Fig. 4 presents the average CSR indexclassified into “Particular” and “Universal” disclosures by country in 2011. It is clear from Fig. 4 that the average “Particular”CSR index (40.62%) is lower than the average “Universal” CSR index (47.98%). This result is consistent with Belal et al. (2014).

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S31

TttbsbiTd

o7wtsip1mt

fi2tsPmea

TD

TrnfrP

Fig. 4. Average CSR index classified into “Particular” and “Universal” disclosures by country.

hey argue that there is a shift towards more universal disclosures over time especially after 2006. This shift is more relevanto the wider stakeholder groups such as community, employees and customers. They also argue that Islamic banks adhereo the minimalist approach without explicitly contravening Shari’ah principles. The CSR index scores are further analysedy country, the results show that the average “Universal” CSR index scores are higher compared to “Particular” CSR indexcores in all countries except for Syria. We agree with Belal et al. (2014) that Islamic banks firstly established their reputationased on Shari’ah compliance that distinguished them from the conventional banks; however, there is a remarkable shift

n Islamic banks strategy towards more universal disclosures which are more relevant to the wider stakeholder groups.able 4 presents the CSR index scores classified by Particular versus Universal disclosures and mandatory versus voluntaryisclosures in 2011.

In this section we present the empirical findings of the econometrics modelling. Table 5 presents the descriptive statisticsf the covariates in 2011. The sample size is 90 Islamic banks and the average CSR disclosure index ranges from 13.1% to0.2% with an average of 44.3%. The average return on equity (ROE) ranges from −52.7% to 39.3% with a mean of 5.71%,hereas, the average return on assets (ROA) ranges from −29.5% to 21.8% with an average of 0.47%. Table 5 also reports

hat SSB size ranges from 2 to 9 members with mean value of 4.17, whereas the average board size is 9.03 members withtandard deviation 3.31, in addition the proportion of NEDs is 65.2% on average. The average age of Islamic banks includedn our sample ranges from 6 to 36 years with an average age of 12.3 years. The average finance to total assets ratio and theercentage of private Islamic banks are 42.6% and 44.4% respectively. Finally, the average natural logarithm of bank size is5.32 (US$ 4.5 million) whereas the average natural logarithm of GDP is 25.67. Table 6 reports the outputs of the correlationatrix of the covariates used in the analysis. It is clear that there are no significant correlation coefficients greater than 50%,

herefore our estimation is not subject to multicollinearity problem.Table 7 presents the outputs of Eq. (2) in which we investigate the main determinants of CSR disclosure in particular

nancial performance using cross sectional analysis in both 2010 and 2011 in addition to the pooled sample over the period010–2011. Panels A and B report the estimation outputs using the average annual change in ROA and ROE respectively overhe previous 5 years as a proxy for Islamic banks financial performance. Results show that there is a positive and highlyignificant relationship between FP and CSR disclosure index in both 2010 and 2011 in addition to the pooled sample inanels A and B. This result supports the slack resources theory as better financial performance encourages banks to engage

ore in social activities. This result is consistent with Mallin and Michelon (2011), McWilliams and Siegel (2001), Balabanis

t al. (1998), Waddock and Graves (1997) and Roberts (1992). However, this result does not confirm the direction of causalitynd any potential endogeneity between CSR disclosure and FP. We investigate that using the 3SLS estimation.

able 5escriptive statistics.

Mean S.D Min Max Skewness kurtosis Obs.

CSR Index 0.4439 0.1248 0.1310 0.7024 −0.4025 2.7585 90ROE 0.0571 0.1369 −0.5271 0.3927 −1.3026 6.3265 90ROA 0.0047 0.0541 −0. 2955 0.2178 − 1.8238 9.2167 90SSB. size 4.1666 1.3759 2.0000 9.0000 1.1812 4.0995 90B. size 9.0333 3.3198 5.0000 23.000 1.9569 7.4666 90NED 0.6519 0.2800 0.0935 0.7125 −0.8411 2.1695 90ln TA 15.3196 1.6635 9.6881 17.9035 −0.2772 2.7799 90Age 12.3222 11.3848 6.0000 36.000 1.4990 4.9247 90Fin./TA 0.4265 0.2289 0.3358 0.9632 −0.2631 2.2159 90lnGDP 25.6743 1.2658 23.9813 28.5185 0.1667 2.3172 90Private 0.4444 0.4997 0.0000 1.0000 0.2236 2.0500 90

he table presents the descriptive statistics of the sample in 2011. CSR Index: corporate social responsibility index; ROE: return on equity before tax; ROA:eturn on assets before tax; SSB.size: the number of Shari’ah supervisory board members; B.Size the number of board members; NED: the percentage ofon-executive directors in the board of directors; lnTA: the natural logarithm of bank’s total assets in US$ as a proxy for bank size; Age: bank age since

oundation; Fin./TA: the ratio of total bank finances/total assets; private: dummy variable takes the value of 1 if the bank is listed in a stock exchange of theespective country and 0 otherwise; lnGDP: natural logarithm of the gross domestic product of country i as a proxy for country macroeconomic factors;rivate: dummy variable takes the value of 1 if the bank is listed in a stock exchange of the respective country and 0 otherwise.

S32 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Table 6Correlation matrix.

CSR Index �ROA �ROE lnTA lnAge SSB. size B. size NED Fin./TA lnGDP Private

CSR Index 1.0000�ROA 0.0481 1.0000�ROE 0.0391 0.9531 1.0000lnTA 0.2725*** −0.2285** −0.2285** 1.0000lnAge 0.2114** 0.0614 0.0614 0.1775* 1.0000SSB. size 0.2402** −0.0763 −0.0763 0.2986*** 0.1808* 1.0000B. size 0.0485 −0.0975 −0.0975 0.1530 0.2505** 0.4617*** 1.0000NED 0.2246** −0.0576 −0.0576 0.1524 −0.2854*** −0.1592 −0.4471*** 1.0000Fin./TA −0.1754* −0.0253 −0.0253 0.1468 −0.0401 0.3054*** 0.0693 −0.2534** 1.0000lnGDP −0.0079 −0.1430 −0.1430 0.3522*** −0.1559 0.0106 −0.2278** 0.0134 0.3742*** 1.0000Private −0.0612 −0.0950 −0.0950 0.1009 0.1105 0.2369** 0.3499*** −0.2319** 0.0806 −0.1233 1.0000

The table presents the correlation matrix for the covariates in 2011. CSR Index: corporate social responsibility index; AROA: average annual change in ROAover 2006–2010; AROE: average annual change in ROE over 2006–2010; lnTA: natural logarithm of bank’s total assets in US$ as a proxy for bank size;lnAge: the natural logarithm of bank age since foundation; SSB.size: number of Shari’ah supervisory board members; B.Size: number of board members;

NED: percentage of non-executive directors in the board of directors;, Fin./TA: ratio of total bank finances/total assets; Private: dummy variable takes thevalue of 1 if the bank is listed in a stock exchange of the respective country and 0 otherwise; lnGDP: natural logarithm of the gross domestic product ofcountry i as a proxy for country macroeconomic factors.

Interestingly, we find a positive and highly significant association between the SSB size and CSR disclosure index in PanelsA and B. However, board size is found to be insignificantly related to the CSR disclosure index. This implies the essential roleof the SSB in supporting Islamic banks social activities. This result is consistent with Farook et al. (2011) as they argue that

larger SSB size may lead to higher levels of CSR disclosure as the capacity of the monitoring role of the SSB increases. As aresult the SSB may efficiently allocate the designated tasks amongst larger group of members and hence achieve a greaterlevel of compliance. Moreover, the proportion of NEDs is positive in sign and marginally significant implying a positiveassociation with the CSR disclosure index. This suggests the positive impact of board independence on CSR disclosure and, in

Table 7Determinants of CSR disclosure.

Panel A Panel B

CSR Index 2010 2011 Pooled 2010 2011 Pooled�ROA 0.0003***

(0.0001)0.0004***(0.0001)

0.0002**(0.0001)

�ROE 0.0004***(0.0001)

0.0002**(0.0001)

0.0003***(0.0001)

SSB.size 0.0275***(0.0099)

0.0235**(0.0103)

0.0256***(0.0064)

0.0298***(0.0096)

0.0249**(0.0101)

0.0277***(0.0072)

B.size 0.0006(0.0047)

0.0008(0.0051)

0.0013(0.0031)

0.0003(0.0073)

0.0006(0.0051)

0.0011(0.0054)

NED 0.1085*(0.0615)

0.1057*(0.0588)

0.1249**(0.0637)

0.1184**(0.0542)

0.1066*(0.0588)

0.1141**(0.0589)

lnTA 0.0148*(0.0083)

0.0121*(0.0069)

0.0065*(0.0038)

0.0149*(0.0081)

0.0127*(0.0071)

0.0061*(0.0035)

lnAge 0.0274*(0.0161)

0.0272*(0.0161)

0.0241**(0.0121)

0.0269*(0.0159)

0.0275*(0.0157)

0.0239**(0.0118)

Fin./TA −0.0901(0.0736)

−0.1249(0.0923)

−0.1197(0.0820)

−0.0907(0.0731)

−0.1254(0.0924)

−0.1196(0.0967)

LnGDP −0.0048(0.0126)

0.0054(0.0113)

0.0089(0.0079)

−0.0034(0.0145)

0.0067(0.0122)

0.0081(0.0086)

Private −0.0390(0.0312)

−0.0205(0.0283)

−0.0236(0.0203)

−0.0399(0.0335)

−0.0207(0.0298)

−0.0241(0.0216)

Cons 0.2141(0.3283)

−0.0323(0.2985)

−0.0551(0.2051)

0.2136(0.3283)

−0.0328(0.2986)

−0.0556(0.2051)

Country Dummy No No No No No NoRamsey RESET test (p value) 0.157 0.257 0.324 0.168 0.213 0.297Mean VIF 1.48 1.61 1.59 1.53 1.52 1.36F stat. 6.85***

(0.0000)5.78***(0.0000)

11.21***(0.0000)

6.91***(0.0000)

5.83***(0.0000)

11.31***(0.0000)

R2 0.3601 0.4075 0.3934 0.3902 0.4276 0.3839Obs. 90 90 180 90 90 180

CSR Index: corporate social responsibility index; �ROA: average annual change in ROA over 2006–2010; �ROE: average annual change in ROE over2006–2010 SSB. size: number of Shari’ah supervisory board members; B.Size: number of board members; NED: percentage of non-executive directors inthe board of directors; lnAge: natural logarithm of bank age since foundation; lnTA natural logarithm of bank’s total assets in US$ as a proxy for bank size;Fin./TA: ratio of total bank finances/total assets; lnGDP: natural logarithm of the gross domestic product of country i as a proxy for country macroeconomicfactors; Private: dummy variable takes the value of 1 if the bank is listed in a stock exchange and 0 otherwise.***, **, * indicates significance at the 1%, 5%and 10% levels. Robust standard error between parentheses.

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S33

Table 8CSR disclosure dimensions and financial performance of Islamic banks.

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10

�ROA 0.0010***(0.0000)

0.0006***(0.0001)

0.0003*(0.0002)

0.0003***(0.0001)

0.0009***(0.0002)

0.0004***(0.0001)

0.0008***(0.0002)

0.0002*(0.0001)

−0.0002*(0.0001)

0.0012***(0.0003)

SSB.size 0.0126(0.0163)

0.0043(0.0170)

0.0283**(0.0118)

0.0184*(0.0101)

0.0217(0.0228)

0.0016(0.0135)

0.0483**(0.0205)

0.0030(0.0134)

0.0033(0.0101)

0.0675**(0.0336)

B.size 0.0558***(0.0112)

0.0247***(0.0085)

−0.0016(0.0112)

−0.0052(0.0055)

0.0293**(0.0114)

−0.0008(0.0060)

−0.0039(0.0123)

0.0016(0.0097)

0.0003(0.0069)

0.0035(0.0196)

NED 0.2746**(0.1355)

0.2901***(0.0923)

0.2877**(0.1261)

0.0061(0.0663)

0.5013***(0.1246)

−0.0559(0.0760)

0.1878(0.1262)

0.0369(0.0844)

−0.0520(0.0770)

0.2943(0.2339)

lnTA 0.0246*(0.0142)

0.0038(0.0144)

0.0366**(0.0172)

0.0073(0.0118)

0.0532**(0.0227)

0.0282**(0.0140)

0.0086(0.0244)

−0.0019(0.0098)

0.0147*(0.0085)

−0.0302(0.0330)

lnAge 0.0079(0.0325)

0.0440*(0.0258)

0.0588*(0.0358)

0.0060(0.0176)

0.0647*(0.0349)

−0.0174(0.0195)

−0.0328(0.0351)

0.0146(0.0242)

0.0376**(0.0172)

−0.0301(0.0550)

Fin./TA 0.1313(0.1109)

0.0466(0.1132)

0.2182*(0.1123)

0.0694(0.0796)

0.1973(0.1188)

0.0784(0.1017)

0.0058(0.1261)

0.0699(0.0884)

0.0229(0.0339)

−0.0523(0.2386)

lnGDP 0.0260(0.0249)

0.0161(0.0199)

0.0228(0.0242)

−0.0090(0.0164)

−0.0024(0.0347)

0.0554***(0.0204)

0.0156(0.0281)

0.0047(0.0130)

−0.0017(0.0143)

0.0753*(0.0421)

Private 0.0732(0.0505)

0.0430(0.0462)

0.0376(0.0529)

−0.0189(0.0302)

0.0612(0.0677)

0.0098(0.0349)

−0.0418(0.0613)

−0.0008(0.0378)

−0.0453**(0.0205)

−0.0235(0.0992)

Cons −0.2637(0.6816)

−0.0814(0.4813)

−0.1419(0.6526)

0.4058(0.4062)

0.4517(0.7919)

1.7492***(0.5116)

−0.2076(0.6941)

0.2466(0.3614)

0.0956(0.4004)

−1.2451(1.1143)

F stat.(p value) 23.44***(0.0000)

17.01***(0.0000)

3.51***(0.0011)

24.82***(0.0000)

36.65***(0.0000)

13.61***(0.0000)

27.65***(0.0000)

6.12***(0.0000)

1.7900*(0.098)

12.23***(0.0000)

R2 0.1670 0.1379 0.1419 0.1352 0.1881 0.1541 0.1279 0.0255 0.0670 0.1010Obs. 90 90 90 90 90 90 90 90 90 90

D1, . . ., D10 are vision and mission statement; board of directors and top management; product/services; zakah, and benevolent loans; employees; debtors;community; shari’ah supervisory board (SSB); environment and charitable and social activities dimensions respectively. �ROA: average annual changein ROA over 2006–2010; SSB.size: number of Shari’ah supervisory board members; B.Size: number of board members; NED: percentage of non-executivedirectors in the board of directors; LnTA natural logarithm of bank’s total assets in US$ as a proxy for bank size; lnAge: natural logarithm of bank age sincefm*

trt

ssa(twrConawcsr1avd

d

hh

oundation; Fin./TA: ratio of total bank finances/total assets; lnGDP: natural logarithm of the gross domestic product of country i as a proxy for countryacroeconomic factors; Private: dummy variable takes the value of 1 if the bank is listed in a stock exchange of the respective country and 0 otherwise.***,

*, * indicates significance at the 1%, 5% and 10% levels. Robust standard error between parentheses.

erms of resource dependence theory, NEDs may bring managerial know-how and hence help improve firm disclosure. Thisesult is consistent with Pfeffer and Salancik (1978) and Johnson and Greening (1999) as they argue that NEDs may enhancehe reputation and credibility of the company and help to establish and maintain its legitimacy.

Table 7 also reports – as expected- that lnTA (our proxy for bank size) and bank age are positive in sign and marginallyignificant; this suggests that there is a positive relationship between both bank size and bank age, and their CSR disclo-ure. This result is consistent with Roberts (1992) as she argues that the older the company the more involvement in CSRctivities as these relate to its reputation. This result is also consistent with Mallin and Michelon (2011), Al-Tuwaijri et al.2004), Brammer et al. (2006) and McWilliams and Siegel (2001) as they argue that big banks are highly likely to monitorheir activities towards wider society. In addition, a well-established bank may have a slightly different set of stakeholdersho may impose more constraints on financial institutions to disclose, communicate and to reflect their corporate social

esponsibility and ethical behaviour to their related parties. This result is consistent with the “social impact hypothesis” ofornell and Shapiro (1987) in which they suggest that satisfying different needs of stakeholders will improve the reputationf the company and lead to better FP. We notice that the “Private” variable is insignificant implying that the CSR disclosure isot determined by whether the bank is listed in the stock exchange or not. This may suggest that CSR disclosure regulationsre less mandatory for Islamic financial institutions in some of the countries included in our sample. However interestingly,hen we control for the Visibility dummy, it turns out to be positive and significant at the 5% level. This may imply that

ompanies included in index constituents may have a better CSR profile.27 The models presented in Panels A and B are wellpecified; the average R squared in Panels A and B are 38.7% and 40.1% respectively. The F statistics are highly significant andeject the null that the coefficients are insignificantly different from zero.28 The variance inflation factor (VIF) is well below0; this suggests that our models are not subject to multicollinearity. Finally, the Ramsey RESET test for omitted variablesnd model mis-specification fails to reject the null that the model is not mis-specified and concludes that there is no omittedariables bias and our models are well specified. However, the question on which dimension(s) of the CSR disclosure index

erive and impact the CSR-FP link remains unanswered; we try to answer this question in the following section.

Motivated by the study of Brown and Caylor (2006), we re-estimate Eq. (2) using the individual dimensions of the CSRisclosure index separately as a dependent variable (one at a time). Table 8 presents the results of the cross-sectional

27 Results are not presented in Table 4 as Visibility dummy is correlated with Private dummy variable.28 We also test the country effect (heterogeneity) and whether the country coefficients are jointly different from 0 with the null that country dummiesave no effect on the CSR disclosure index. We find no significant country effect and the results presented in Table 6 do not change if we control for countryeterogeneity.

S34 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Table 9Mandatory and voluntary CSR disclosure and financial performance.

CSR Index Panel A: mandatory Panel B: voluntary

�ROA 0.0004***(0.0001)

0.0002**(0.0001)

�ROE 0.0003***(0.0001)

0.0002**(0.0001)

SSB.size 0.0226*(0.0129)

0.0229*(0.0131)

0.0412***(0.0086)

0.0345***(0.0092)

B.size 0.0202**(0.0084)

0.0208**(0.0091)

0.0018(0.0053)

0.0022(0.0053)

NED 0.1730**(0.0838)

0.1749*(0.0922)

0.0978*(0.0601)

0.0921*(0.0561)

lnTA 0.0241**(0.0108)

0.0239**(0.0112)

0.0182**(0.0091)

0.0189*(0.0098)

lnAge 0.0004(0.0248)

0.0004(0.0248)

0.0239**(0.0141)

0.0225**(0.0117)

Fin./TA 0.1198(0.0896)

0.1200(0.0899)

0.0823(0.0657)

0.0824(0.0661)

lnGDP 0.0146(0.0158)

0.0141(0.0165)

0.0046(0.0131)

0.0055(0.0161)

Private 0.0262(0.0367)

0.0258(0.0371)

0.0126(0.0279)

0.0121(0.0284)

Cons 0.1095(0.4160)

0.1093(0.4161)

0.1302(0.3239)

0.1297(0.3240)

F stat. 8.57***(0.0000)

8.41***(0.0000)

27.85***(0.0000)

27.78***(0.0000)

R2 0.3231 0.2931 0.2720 0.2520Ramsey RESET test (p value) 0.149 0.374 0.125 0.268Mean VIF 1.99 1.45 1.89 1.24Obs. 90 90 90 90

CSR Index: corporate social responsibility index; �ROA: average annual change in ROA over 2006–2010; �ROE: average annual change in ROE over2006–2010 SSB.size: number of Shari’ah supervisory board members; B.Size: number of board members; NED: percentage of non-executive directors inthe board of directors; lnTA: natural logarithm of bank’s total assets in US$ as a proxy for bank size; lnAge: natural logarithm of bank age since foundation;

Fin./TA: ratio of total bank finances/total assets; lnGDP: natural logarithm of the gross domestic product of country i as a proxy for country macroeconomicfactors; Private: dummy variable takes the value of 1 if the bank is listed in a stock exchange of the respective country and 0 otherwise.***, **, * indicatessignificance at the 1%, 5% and 10% levels. Robust standard error between parentheses.

regression in 2011 using the annual average change in ROA as a measure of FP.29 Results show that there is a positive andsignificant association between FP and CSR dimensions except for dimension 9 (environment). This suggests that the betterthe FP the less disclosure on environmental activities. Therefore most of the Islamic banks included in our sample may notspend on environmental activities regardless of their FP. This result is consistent with Preston and O’Bannon (1997) andWaddock and Graves (1997) as they argue that these costs might be avoided or that they should be borne by governmentsor other stakeholders (i.e. investments in pollution control equipment). Therefore spending on environmental activitiesexceeds their potential benefits and this may explain the low scores of the environment dimension of the CSR disclosureindex. Interestingly, SSB size is found to be positive and significant with financial product/services, zakah, and benevolentloans, community, charitable and social activities dimensions. However, the main board and the proportion of NEDs arepositively and significantly associated with vision and mission, board of directors and top management and employeesdimensions. We also find that listed banks disclose more on the environmental activities compared with private banks. Themodels presented in Table 8 are well specified; the F statistics are highly significant and reject the null that the coefficientsare insignificantly different from zero.

Table 9 presents the estimation results of the relationship between Islamic banks FP and both mandatory and voluntarydisclosure requirements in the light of the Standard No.7 issued by AAOIFI in 2010. We re-estimate Eq. (2) by classifyingthe CSR index items into two main categories namely mandatory and voluntary CSR disclosure in 2011.30 Overall we findsimilar results to those presented in Table 7. The results presented in Table 9 confirm the positive and significant associationbetween both mandatory and voluntary CSR disclosure and Islamic banks FP. However, this relationship is less significantfor the voluntary disclosure. Interestingly, we notice that SSB size is marginally significant in Panel A (mandatory disclosure),

however it is highly significant in Panel B (voluntary disclosure). This suggests the fundamental role of the SSB in encouragingsocial and charitable activities. Moreover, board size is significant within the mandatory disclosure Panel and insignificantwithin the voluntary disclosure Panel. The models are well specified; the F statistics are highly significant and reject the null

29 We also estimate these models using the average annual change on ROE in 2010 and 2011 in addition to the pooled sample and obtained similar results.Results are available from the authors upon request.

30 We also estimate this link using 2010 and the pooled sample and obtained similar results to those presented in Table 7. Results are available from theauthors upon request.

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S35

Table 103SLS estimation results for CSR and FP equations.

Dependent variable Panel A: CSR equation Panel B: Performance equation

CSR index �ROA � ROE

System 1 System 2 System 1 System 2

�ROA 0.0011***(0.0002)

�ROE 0.0009***(0.0001)

CSR 2.3269(1.5282)

2.7665(1.8095)

SSB. size 0.0299***(0.0068)

0. 0284***(0. 0068)

1.9682**(1.0105)

1. 9572*(1. 0278)

B. size 0.0001(0.0032)

0.0002(0.0035)

−0.2868(0.6362)

−0. 3344(1.9374)

NED 0.0911**(0.0371)

0.0908**(0.0371)

0.4135**(0.2132)

0.4921**(0.1986)

lnTA 0.0059*(0.0032)

0.0048**(0.0021)

0.9751(2.1596)

1.1732(2.5572)

lnAge 0.0162*(0.0096)

0.0174*(0.0099)

1.2548(5.2699)

1.4711(1.2403)

Fin./TA −0.1060**(0.0409)

−0.1057**(0.0409)

0.2807(0.2705)

0.4567(0.3591)

lnGDP 0.0157**(0.0077)

0.0082(0.0071)

0.0084(0.6871)

0.0705(0.8121)

Cons 0.1991(0.1742)

0.1887(0.1689)

1.4557(2.9112)

1.6113(2.7498)

R.sq 0.272 0.187 0.074 0.141Chi. sq 41.35*** 42.11*** 6.06 9.015Sargan test (p.value) 0.348 0.301 0.348 0.301Hausman Specification Tests2SLS vs 3SLS (p.value) 0.246 0.272 0. 246 0. 272Breusch–Pagan test of independence(p.value)

0.02

CSR Index: corporate social responsibility index; �ROA: average annual change in ROA over 2006–2010; �ROE: average annual change in ROE over2006–2010; SSB.size: number of Shari’ah supervisory board members; B.Size: number of board members; NED: percentage of non-executive directors intFf

tieor

Tcjeoesct

h

O

o

f

he board of directors; lnTA natural logarithm of bank’s total assets in US$ as a proxy for bank size; lnAge: natural logarithm of bank age since foundation;in./TA: ratio of total bank finances/total assets; lnGDP: natural logarithm of the gross domestic product of country i as a proxy for country macroeconomicactors; ***, **, * indicates significance at the 1%, 5% and 10% levels. Standard error between parentheses.

hat the coefficients are insignificantly different from zero. To sum up, the results presented in Tables 7–9 suggest that theres a positive and significant relationship between Islamic banks FP and CSR disclosure.31 Moreover, our results highlight thessential role and the positive impact of the SSB size on the CSR disclosure. Based on the above discussion we cannot rejectur first hypothesis. However, the question on the direction of causality between CSR disclosure and the FP of Islamic banksemains unanswered; we try to answer this question in the following section.32

In this section we investigate the bi-directional relationship between CSR and Islamic bank FP.To account for a potential endogeneity between CSR and FP we use Durbin-Wu-Hausman test (e.g., Hausman, 1978).

he result the Durbin–Wu–Hausman test rejects the null hypothesis of no endogeneity at the 10% level. Therefore, weonclude that OLS may lead to biased and inconsistent estimates in our sample. To this end, we estimate Eqs. (3a) and (3b)ointly using three-stage least squares regression33 to deal with any potential endogeneity between CSR and FP. The 3SLSstimation results for the simultaneous system are summarised in Table 10. Panel A presents the results of the impact of FPn CSR as in Eq. (3a) whilst Panel B presents the impact of CSR on the FP as in Eq. (3b). The coefficient on �ROA in the CSRquation remains positive and significant at the 1% level. However, the coefficient on CSR in the FP equation turns out to betatistically insignificant. This suggests that the causality between the two endogenous variables runs from FP to CSR. It islear that replacing �ROE with �ROA as an FP measure does not alter our findings. The performance equation also shows

hat there is a positive and significant relationship between SSB size and the proportion of NEDs and FP.

The result of the Breusch–Pagan test shows that cross-equation residuals were not independent (p-value = 0.02) and,ence, the test rejects the null hypothesis of independence errors and indicates, therefore, that the equations need to be

31 As robustness check we re-estimate equation 2 excluding IBs from Malaysia and Bahrain as both countries represent 40% (36 banks) of our sample size.ur empirical results remain qualitatively unchanged.

32 We re-estimate equation 2 by classifying the CSR index items into other two categories namely Particular and Universal CSR disclosures in 2011. Webtained consistent results confirming the positive and significant association between both Universal and Particular CSR disclosure and Islamic banks FP.33 We use the 2SLS estimation as a robustness check following Al-Tuwaijri et al. (2004). The 2SLS estimation provides similar results to those obtainedrom 3SLS regarding the direction of causality between CSR and FP.

S36 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

estimated simultaneously. The system presented in Panel A is well specified as the Chi squared is highly significant. Onthe other hand, the system presented in Panel B is not well specified as the Chi squared is insignificant. The result of theSargan mis-specification test shows that we cannot reject the null hypothesis of “no mis-specification”, indicating that theinstruments of our system are orthogonal to the error terms of the respective equations. We also report the results of theHausman test; the results show that we cannot reject the null hypothesis of “the 3SLS results are consistent and efficientwhile the 2SLS results are also consistent but inefficient”. Hence, under the assumption that at least one of the equations ofour system is correctly specified, the specification of the entire system cannot be rejected and the most efficient estimatesare obtained by applying 3SLS. To sum up, the 3SLS results strongly suggest that the CSR disclosure index of the Islamicbanks is determined by their FP and the opposite is not true. Therefore we can reject our second hypothesis. We interpretthis result in the light of the slack resources theory which assumes that available slack resources might be allocated into thesocial domain. The higher profitability provides more opportunities for Islamic banks to subsequently promote and disclosetheir surplus resources invested into socially responsible activities in line with their religious values and beliefs. We arguethat the slack resources theory has been developed from the base of general and wider applicability and thus may also beapplicable to IFI. However, as the distinctive governance feature of IFI is the Shari’ah governance system and the presence ofthe SSB, we believe that the Shari’ah principles underlying IFI result in unique agency relationships.

6. Summary and conclusions

Islamic banking and CSR are both areas which have seen significant growth in recent years. In this paper we investigate thetype and the extent of CSR disclosure made by a sample of 90 Islamic across 13 countries over the period 2010–2011. Our paperrevises and constructs a comprehensive index created from CSR disclosure studies of Islamic banks and recommendationsfrom the AAOIFI Standard No.7. We investigate empirically the link between Islamic banks FP and CSR highlighting the impactof SSB size on the level of CSR disclosure and the direction of causality between FP and the CSR. The CSR index scores show thatthe extent of disclosure across countries varies considerably. Indonesia has the highest CSR index score of 53.8%, followed byMalaysia (51.5%) and Bahrain (51.2%) in 2011. The lowest scores are shown by Pakistan (30.4%) and the Sudan (31.4%). Wealso find that the vision and mission dimension generally scores highly across all countries whilst the environment generallyscores the lowest. Our findings are consistent with those of Kamla (2007) who found very low levels of environmental relateddisclosures by Arab companies including Islamic banks. We also find that Malaysia has the highest mandatory CSR indexscores (76%) whereas Pakistan has the lowest mandatory CSR index score (42%) in 2011. On the other hand, Indonesia has thehighest voluntary CSR disclosure index (48%) whereas the Sudan has the lowest voluntary CSR index score of 22% in 2011.Moreover, we find consistent results with Belal et al. (2014) as we find that Particular” CSR index (40.62%) is lower than theaverage “Universal” CSR index (47.98%). In addition, the CSR index scores are further analysed by country, the results showthat the average “Universal” CSR index scores are higher compared to “Particular” CSR index scores in all countries exceptfor Syria. Hence, there is a remarkable shift in Islamic banks strategy towards more universal disclosures which representthe wider perspective of Shari’ah and are more relevant to the wider stakeholder groups.

The empirical analysis highlighted a positive association between CSR disclosure index and banks’ performance. Thisresult supports the slack resources theory as better FP encourages banks to engage more in social activities. We also findthat there is a positive and significant association between FP and CSR dimensions except for the environment dimension.This result is consistent with Preston and O’Bannon (1997) and Waddock and Graves (1997) as they argue that these costsmight be avoided or that they should be borne by governments or other stakeholders (i.e. investments in pollution controlequipment). Therefore spending on environmental activities exceeds their potential benefits which contradicts the widerperspective of Shari’ah compliance. Interestingly, we find a positive and highly significant association between the SSB sizeand CSR disclosure index. This implies the essential role of the SSB in supporting Islamic banks social activities. This resultis consistent with Farook et al. (2011) as they argue that larger SSB size may lead to higher levels of CSR disclosure as thecapacity of the monitoring role of the SSB increases. As a result, the SSB may efficiently allocate the designated tasks amongstlarger group of members and hence achieve a greater level of compliance. The 3SLS results strongly suggest that the CSRdisclosure index of the Islamic banks is determined by their FP and the opposite is not true. This suggests that the causalitybetween the two endogenous variables runs from FP to CSR.

The findings have a number of policy implications. Firstly, the empirical evidence indicates that the level of CSR disclosurewas relatively low even though AAOIFI Standard No.7 provides a template for Islamic banks to adopt in terms of CSR conductand disclosure. Therefore policymakers might, in future, be more active in encouraging Islamic banks to adopt AAOIFI StandardNo.7 as a benchmark. Secondly, our empirical evidence suggests that an Islamic bank’s involvement in CSR tends to increaseas its SSB size grows larger. Therefore, policymakers should be encouraged to introduce policies to help increase the numberof eligible SSB members as a larger SSB may encourage disclosure of the Islamic Bank’s CSR and such disclosure providesevidence to the public that it is pursuing its social goals. Such policies might include more professional training programmeswhich help increase the supply of Shari’ah scholars. Additionally, policy makers should be encouraged to introduce specialisttraining for SSB members on the wider perspective of Shari’ah compliance that include universal CSR of environmental, social

and governance issues. Hopefully a larger pool of Shari’ah scholars may help Islamic banks to appoint more members ontotheir SSBs. Larger SSBs may approve more Islamic based products, thus contributing to higher profits which may thenencourage these banks to participate in more social activities in line with the Islamic Bank’s objectives in bringing social andeconomic benefits to all its stakeholders. Future research may consider other factors like an updated set of national culture

sai

A

F

R

AA

A

AA

A

AAB

B

B

B

BB

B

BBCCDEEFFG

GHH

HH

H

J

KKMM

MM

N

PPPRRS

C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38 S37

cores adapted from Hofstede’s (2001) model (Taras et al., 2012) to try and explain the effects of a society’s culture on CSRctivities. Furthermore, future research may study the determinants on social activities by collecting data through detailednterviews with management and other stakeholders. This might substantiate the findings from the CSR disclosure studies.

cknowledgements

The authors would like to express their appreciation to the paper discussant, participants and reviewers at the Islamicinance Conference, 29th September 2012, Aston University, Birmingham, UK.

eferences

AOIFI, 2008. Governance standards for Islamic financial institutions. Accounting and Auditing Organization for Islamic Financial Institutions, Bahrain.AOIFI, 2010. Corporate social responsibility conduct and disclosure for Islamic financial institutions. Accounting and Auditing Organization for Islamic

Financial Institutions, Bahrain.bdul Rahman, A., Md Hashim, M.F.A., Abu Bakar, F., 2010. Corporate social reporting: a preliminary study of Bank Islam Malaysia Berhad (BIMB). Iss. Soc.

Environ. Account. 4 (1), 18–39.hmad, K., 2000. Islamic finance and banking: challenges and prospects. Rev. Islam. Econom. 9, 57–82.hmed, S.U., Islam, M.Z., Hasan, I., 2012. Corporate social responsibility and financial performance linkage. Evidence from the banking sector of Bangladesh

1 (1), 14–21.l-Tuwaijri, S., Christensen, T.E., Hughes, K.E., 2004. The relations among environmental disclosure, environmental performance, and economic perfor-

mance: a simultaneous equations approach. Account. Organ. Soc. 29 (5–6), 447–471.ndres, D.P., Vallelado, E., 2008. Corporate governance in banking: the role of the board of directors. J. Bank. Finance 32, 2570–2580.ribi, Z.A., Gao, S.S., 2012. Narrative disclosure of corporate social responsibility in Islamic financial institutions. Manage. Audit. J. 27 (2), 199–222.alabanis, G., Phillips, H.C., Lyall, J., 1998. Corporate social responsibility and economic performance in the top British companies: are they linked? Eur. Bus.

Rev. 98 (1), 25–44.eattie, V., McInnes, W., Fearnley, S., 2004. A methodology for analysing and evaluating narratives in annual reports: a comprehensive descriptive profile

and metrics for disclosure quality attributes. Account. Forum 28 (3), 205–236.einer, S., Drobetz, W., Schmid, M.M., Zimmermann, H., 2006. An integrated framework of corporate governance and firm valuation. Eur. Financ. Manage.

12, 249–283.elal, A.R., Abdelsalam, O., Nizamee, S., 2014. Ethical reporting in Islami Bank Bangladesh Limited (1983–2010). J. Bus. Ethics,

http://dx.doi.org/10.1007/s10551-014-2133-8 (forthcoming).hagat, S., Bolton, B., 2008. Corporate governance and firm performance. J. Corp. Finance 14, 257–273.rammer, S., Brooks, C., Pavelin, S., 2006. Corporate social performance and stock returns: UK evidence from disaggregate measures. Financ. Manage.,

97–116.rammer, S., Millington:, A., 2006. Firm size, organizational visibility and corporate philanthropy: an empirical analysis’. Bus. Ethics: A Eur. Rev. 15 (1),

6–18.reusch, T.S., Pagan, A.R., 1980. The Lagrange multiplier test and its applications to model specification in econometrics. Rev. Econ. Stud. 47, 239–253.rown, L., Caylor, M., 2006. Corporate governance and firm valuation. J. Account. Public Policy 25, 409–434.hong, B.S., Liu, M.H., 2009. Islamic banking: Interest-free or interest based? Pac.-Basin Finance J. 17, 125–144.ornell, B., Shapiro, A.C., 1987. Corporate stakeholders and corporate finance. Financ. Manage. 16 (1), 5–14.ar, H.A., Presley, J.R., 2000. Lack of profit loss sharing in Islamic banking: management and control imbalances. Int. J. Islam. Financ. Serv. 2 (2), 3–18.esley, C., Lenox:, M.J., 2006. Firm responses to secondary stakeholder action. Strateg. Manage. J. 27, 765–781.l-Gamal, M.A., 2006. Islamic Finance: Law, Economics and Practice. Cambridge Publishers, Cambridge.arook, S., 2008. Social responsibility for Islamic financial institutions: laying down a framework. J. Islam. Econ. Bank. Finance 4 (1), 61–82.arook, S., Hassan, M.K., Lanis, R., 2011. Determinants of corporate social responsibility: the case of Islamic banks. J. Islam. Account. Bus. Res. 2 (2), 114–141.arcia-Castro, R., Arino, M.A., Canela, M.A., 2010. Does social performance really lead to financial performance? Account for endogeneity. J. Bus. Ethics 92

(1), 107–126.riffin, J.J., Mahon, J.F., 1997. The corporate social performance and corporate financial performance debate. Bus. Soc. 36 (1), 5–31.aniffa, R., Hudaib, M., 2007. Exploring the ethical identity of Islamic banks via communication in annual reports. J. Bus. Ethics 76, 97–116.assan, A., Harahap, S.S., 2010. Exploring corporate social responsibility disclosure: the case of Islamic banks. Int. J. Islam. Middle E Finance Manage. 3 (3),

203–227.ausman, J.A., 1978. Specification tests in econometrics. Econometrica 46, 1251–1271.o, S.S.M., Wong, K.S., 2001. A study of the relationship between corporate governance structures and the extent of voluntary disclosure. J. Int. Account.

Audit. Tax. 10 (2), 139–156.ofstede, G.H., 2001. Culture’s Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations, second ed. Thousand Oaks, CA,

Sage.ohnson, R.A., Greening, D.W., 1999. The effects of corporate governance and institutional ownership types on corporate social performance. Acad. Manage.

J. 5, 564–580.amla, R., 2007. Critically approaching social accounting and reporting in the Arab Middle East: a postcolonial perspective. Adv. Int. Account. 20 (8), 921–932.ennedy, P.A., 2003. Guide to Econometrics, fifth ed. MIT Press, Cambridge, MA.aali, B., Casson, P., Napier, C., 2006. Social reporting by Islamic banks. Abacus 42 (2), 266–289.allin, C.A., Michelon, G., 2011. Board reputation attributes and corporate social performance: an empirical investigation of the US best corporate citizens.

Account. Bus. Res. 41 (2), 119–144.cWilliams, A., Siegel, D., 2001. Corporate social responsibility: a theory of the firm perspective. Acad. Manage. Rev. 26 (1), 117–127.itchell, R.K., Agle, B.R., Wood:, D.J., 1997. Toward a theory of stakeholder identification and salience: defining the principle of who and what really counts.

Acad. Manage. Rev. 22 (4), 853–886.ikolaev, V., Van Lent, L., 2005. The endogeneity bias in the relation between cost-of-debt capital and corporate disclosure policy. Eur. Account. Rev. 14 (4),

677–724.feffer, J., 1973. Size, composition, and function of hospital boards of directors: the organization and its environment. Admin. Sci. Quart. 18, 349–364.feffer, J., Salancik, G.R., 1978. The External Control of Organizations: A Resource Dependence Perspective. Harper & Row, New York.

reston, L.E., O’Bannon, D.P., 1997. The corporate social–financial performance relationship. Bus. Soc. 36 (4), 419–429.ehbein, K., Waddock, S., Graves:, S.B., 2004. Understanding shareholder activism: which corporations are targeted? Bus. Soc. 43 (3), 239–267.oberts, R.W., 1992. Determinants of corporate social responsibility disclosure. Account. Organ. Soc. 17 (6), 595–612.argan, J.D., 1964. Wages and prices in the United Kingdom: a study in econometric methodology. In: Hart, P.E., Mills, G., Whitaker, J.K. (Eds.), Econometric

Analysis for National Economic Planning. Butterworth, London, pp. 25–63.

S38 C. Mallin et al. / Journal of Economic Behavior & Organization 103 (2014) S21–S38

Simpson, G.W., Kohers, T., 2002. The link between corporate social and financial performance: evidence from the banking industry. J. Bus. Ethics 35 (2),97–109.

Soana, M.-G., 2011. The relationship between corporate social performance and corporate financial performance in the banking sector. J. Bus. Ethics 104(1), 133–148.

Taras, V., Steel, P., Kirkman, B.L., 2012. Improving national cultural indices using a longitudinal meta-analysis of Hofstede’s dimensions. J. World Bus. 47,329–341.

Vance, S., 1975. Are socially responsible firms good investments risks? Manage. Rev. 64 (8), 18–24.Waddock, S.A., Graves, S.B., 1997. The corporate social performance–financial performance link. Strateg. Manage. J. 18 (4), 303–319.Warde, I., 2013. Islamic Finance in the Global Economy. CFA Institute Middle East Investment Conference, Dubai.Zahra, S.A., Oviatt, B.M., Minyard, K., 1993. Effects of corporate ownership and board structure on corporate social responsibility and financial performance.

Acad. Manage. Best Paper Proc., 336–340.