The Impact of Corporate Social Responsibility Disclosure on Financial Performance: Evidence from the GCC Islamic Banking Sector

This paper examines the relationship between corporate social responsibility (CSR) and ﬁnancial performance for Islamic banks in the Gulf Cooperation Council (GCC) region over the period 2000–2014 by generating CSR-related data through disclosure analysis of the annual reports of the sampled banks. The ﬁndings of this study indicate that there is a signiﬁcant positive relationship between CSR disclosure and the ﬁnancial performance of Islamic banks in the GCC countries. The results also show a positive relationship between CSR disclosure and the future ﬁnancial performance of GCC Islamic banks, potentially indicating that current CSR activities carried out by Islamic banks in the GCC could have a long-term impact on their ﬁnancial performance. Furthermore, despite demonstrating a signiﬁcant positive relationship between the composite measure of the CSR disclosure index and ﬁnancial performance, the ﬁndings show no statistically signiﬁcant relationship between the individual dimensions of the CSR disclosure index and the current ﬁnancial performance measure except for ‘mission and vision’ and ‘products and services’. Similarly, the empirical results detect a positive signiﬁcant association only between ‘mission and vision’ dimension and future ﬁnan-cial performance of the examined banks.


Introduction
In recent years, the subject of corporate governance has begun to take an ever-increasing prominent space in the public sphere as a result of high profile corporate failures, such as Barings, Lehman Brothers and others, and the consequences of such failures. Taking into account the fact that poor corporate governance can negatively affect economies and the stability of financial systems and can also have tangible, serious social and environmental consequences, the focus has shifted from the conventional 'shareholders only' approach to corporate governance to a broader corporate governance model that identifies the issues and priorities of stakeholders (Dusuki 2011, p. 6). Such a shift, based on new model of good corporate governance, incorporates ethical considerations and values in the business strategy of corporations, including banks, making it necessary for corporations to be considerate of the wider environment within which the organisation operates and prioritise 'corporate social responsibility' (CSR). This means that corporations understand and address stakeholders' demands though their CSR practices as it is believed that ''CSR connects to governance at the values level, determining the boundaries and accountabilities of the company in relation to a broad universe of stakeholders and its social and environmental responsibilities'' (Strandberg 2005, p. 4).
In responding to the changing paradigm, there have been attempts to include different CSR practices on the agendas of an increasing number of corporations in various parts of the world and to adopt value-based governance in meeting the interests of primary and secondary stakeholders (Perrini et al. 2011, p. 59). This new trend for locating corporations within a broader stakeholder group through CSR activities is argued to yield positive results for firms and organisations through increased customer loyalty, willingness to pay premium prices and lower reputational risk during times of crisis. Consequently, it is argued that all these favourable results can have a positive effect on the financial performance of a corporation, such as improved profitability (Peloza and Shang 2011).
A number of banks and financial institutions have in recent years begun to incorporate CSR within their organisational and operational strategies. The banking sector is a unique industry in society and its role nowadays goes far beyond bringing financial stability to the economy; it now involves establishing new trends and strategies, providing necessary services for customers and reducing financial exclusion. The banking sector is at the heart of society and thus it is expected to be more socially responsible (Chambers and Day 2009, p. 4). Consequently, professionals and academics worldwide have acknowledged and researched the importance of CSR practices in the banking sector. In the emerging economic environment, it is vital for financial intermediaries to integrate moral, ethical and environmental concerns in their business operations (Evangelinos et al. 2009, p. 167). In other words, banks are driven by public demand to increase their transparency and accountability with respect to social responsibility as a result of changing norms and expectations in society. For the banking sector, it is accepted that being socially responsible is as a deep-rooted concept in the financial service industry (Scholtens 2009, p. 159).
In recent years, Islamic banking has emerged as a potentially alternative ethical method of banking and finance, shaped by the ontological and epistemological sources of Islam. Similar to other religions, Islam has at its heart 'social good', 'good governance', 'environmental concern' and 'ethical individual and organisational behaviour' (Asutay 2007(Asutay , 2012(Asutay , 2013. Therefore, the practice of CSR is, by definition an intended consequence of Islamic ethics because Islam as a religion suggests a proactive and expanded stakeholders' paradigm through enforcing a moral obligation of corporations towards society by substantive morality. Thus, CSR from an Islamic perspective, as part of the new paradigm, is an endogenised concept and practice, which by definition is expected to be an existential part of any Islamic corporation, including Islamic banks and financial institutions. In other words, the 'Islamicity' of 'Islamic banks' necessitates CSR as part of its moral substance beyond mechanical operations being Shari'ah or Islamic law compliant. While there is a plethora of research that has attempted to investigate the relationship between CSR and financial performance in conventional financial institutions, there is scant research examining this relationship in Islamic financial institutions (see Hassan et al. 2010;Arshad et al. 2012;Mallin et al. 2014). Thus, this research aims to explore the relationship between CSR disclosure and the financial performance of Islamic banks in the Gulf Cooperation Council (GCC) region over the period 2000-2014, assuming that there is a positive relationship between these two variables with causality running from CSR disclosure towards financial performance.
To assess the CSR disclosure of Islamic banks, the annual reports of the sampled GCC banks were scanned using content analysis. Annual reports have frequently been used in the CSR disclosure literature as they are the main documents aimed at communicating information with external stakeholders, they are prepared under the control of auditors and accountants, they are widely available and they offer a consistent measure (Tilt 1994, p. 57). In conducting the disclosure analysis, this research constructed an index reflecting CSR-related expectations as well as Islamic ethics. The construction of the dimensions and subdimensions of the CSR disclosure index is based on the CSR-related standards developed by the Accounting and Auditing Organisation for Islamic Financial Institutions (AAIOFI 2010). 1 Furthermore, the previous studies on CSR disclosure from an Islamic finance perspective studies are taken as a guide, which includes Haniffa and Hudaib (2007), Aribi and Gao (2012) and Aribi and Arun (2015). As a result, six major dimensions are selected: 'mission and vision statement', 'products and services', 'commitment towards employees', 'commitment towards debtors', 'commitment towards society'; 'zakah (compulsory alms giving by those beyond a threshold level of wealth in the sense of 'returning the right of society to society')'; charity and benevolent funding. It should also be noted that in measuring and calculating the CSR disclosure index, Haniffa and Hudaib's (2007) method was pursued.
The rest of the paper is organised as follows: it begins with a short review of the existing literature in ''Corporate Social and the Financial Performance Nexus: Theoretical Background'' section, followed by the theoretical background for the discussion of the empirical model and the development of hypotheses in ''Hypothesis Development'' section. ''Empirical Framework and Research Method'' section presents the research method and design used in this study, while in ''Empirical Model'' section, the proposed empirical model is described. Further, ''Results and Discussion'' section presents the results obtained based on the empirical model followed by discussion on the implications of the findings. ''Sensitivity Tests'' section provides new regression results in respect of robustness of emprical findings established in the previous section. Finally, the key policy implications and suggestions for the future conduct of Islamic banking industry are presented in ''Conclusion'' section of this paper.

Corporate Social and the Financial Performance Nexus: Theoretical Background
For over three decades, numerous theoretical and empirical studies have been conducted to understand the possible relationship between CSR and financial performance (see, for example, Marom 2006, p. 191;Makni et al. 2009, p. 409;Pava and Krausz 1996, p. 322). Among these, Ullman (1985), Preston and O'Bannon (1997), Roman et al. (1999), Margolis and Walsh (2003), Griffin and Mahon (1997) and Orlitzky et al. (2003) are considered seminal works in the field and have widely been cited as focusing on the relationship between CSR and financial performance in the corporate sector. These studies, which have tested the direction, strength and also causality of the relationship, have produced both confirmatory and contradictory results. A limited number of studies have assessed the relationship between CSR disclosure and financial performance in conventional banking sector (Simpson and Kohers 2002;Soana 2011;Ahmed et al. 2012), while only a limited number of empirical study on the subject matter is available for the Islamic banking sector. Among such studies, for instance, Arshad et al. (2012) analysed the impact of CSR disclosure on corporate reputation and performance of Islamic banks in Malaysia, the results of which indicated positive significant relationship between CSR activities disclosed in the annual reports of the sampled Islamic banks and their reputation and performance. In a similar study, Mallin et al. (2014) focused on examining the link between CSR disclosure and financial performance of Islamic banks across different countries covering the period between 2010 and 2011, whose study demonstrated a positive significant linkage between financial performance and CSR disclosure index of analysed Islamic banks.
Based on the empirical results, it is difficult to assign the direction of the aforementioned relationship in advance of testing any particular set of data, as the findings indicate that the correlation between CSR and corporate financial performance can be positive, negative, neutral, or non-significant; thus, no conclusive results have been attained.
As regards to negative association between CSR and financial performance, some of the empirical studies support the notion of a negative relationship between CSR and financial performance. According to the opponents of CSR, being socially active through engaging in charity projects, supporting and promoting staff welfare and minimising environmental damage can be expensive and give rise to an administrative burden Salomon 2006, p. 1103); therefore, it is argued that CSR activities create financial burdens for corporations.
The negative association between CSR and financial performance is supported by Preston and O'Bannon's (1997) 'trade-off hypothesis', which suggests that the better firms perform in terms of CSR practices, the lower their financial performance. Consequently, socially responsible firms will have less advantage compared to average firms. The proponents of this hypothesis argue that by taking socially responsible initiatives, firms undermine the main objective of the company: maximising profit (Friedman 1970). As argued by Waddock and Graves (1997), Preston and O'Bannon (1997) and Simpson and Kohers (2002), this hypothesis is based on the neoclassical argument that socially responsible strategies result in additional costs and therefore create a competitive disadvantage (Aupperle et al. 1985;Friedman 1970). Recent advocates suggest that 'resources dedicated to social programs or actions should be diverted-either spent on firms' efficiency or returned to shareholders' (Perrini et al. 2011, p. 69) as they argue that it is not the responsibility of firms to address social issues, these being matters which should be resolved by governments or the third sector.
Furthermore, those who consider that CSR should not be embedded in companies' agendas and operations highlight the potential adverse effect of excessive role diversification, as they believe that managers with multiple goals are managers with no objectives (Jensen 2001). Therefore, integrating CSR initiatives in business will create an obstacle to companies' competition for survival.
As for neutral association between CSR and financial performance, various studies that have tested the impact of CSR on financial performance have reported non-significant results (see, among others, Freedman and Jaggi 1986;Patten 1991;Ullman 1985;McWilliams and Siegel 2000). For example, Freedman and Jaggi (1982) detected no correlation between pollution measures and financial performance; and in their later study (Freedman and Jaggi 1986) also indicated a neutral association between the extent of pollution disclosure and financial performance.
It is argued that the neutrality occurs due to the existence of many variables that intervene in the relationship between social and financial performance. Therefore, it is posited that a direct linear relationship between CSR and financial performance is not possible (Waddock and Graves 1997). In line with this hypothesis, social and financial performances are not correlated and social responsibility does not affect financial outputs (Soana 2011, p. 135). Furthermore, based on a supply and demand theory of the firm, McWilliams and Siegel (2001, p. 125) argue for neutrality based on the fact that 'the firm chooses the level of the attribute that maximizes firms' performance, given the demand for the attribute and the cost of providing the attribute, subject to the caveat that this holds true to the extent that managers are attempting to maximise shareholder wealth'. Consequently, the supply of social performance of each firm depends on the demand for corporate social performance that each firm experiences. Therefore, at equilibrium, firms will be equally profitable, but the amount of socially responsible activities produced will be different. In support of this, recent empirical evidence of a neutral relationship between corporate social performance and financial performance in Italian banks was found by Soana (2011), who argues that investing in social projects does not result in financial advantage.
With regards to positive association between corporate social performance and financial performance, despite the lengthy discussions, analyses and ambiguous and mixed results regarding the relationship between corporate social performance and financial performance, a significant number of studies have detected a positive association between the two (Makni et al. 2009, p. 410). The positive relationship between corporate social performance and financial performance is theorised by the 'social impact hypothesis' (see Preston and O'Bannon 1997), which is derived from instrumental stakeholder theory (Freeman 1984;Cornell and Shapiro 1987;Donaldson and Preston 1995), the proponents of which argue that satisfying the needs of different groups of stakeholders will result in enhanced financial performance on the grounds of greater effectiveness and efficiency. In contrast, ignoring the interests of stakeholders might negatively affect corporate financial performance.
A stakeholder approach towards strategic management has an instrumental basis (Freeman 1984). According to this approach, 'if organizations want to be effective, they will pay attention to all and only those relationships that can affect or be affected by the achievement of the organization's purposes. That is, stakeholder management is fundamentally a pragmatic concept' (Freeman 1994, p. 234). However, this does not mean that instrumental stakeholder theory is value-free simply 'because it claims that consequences count' (Freeman 1994, p. 235).
Stakeholder management theory is used by instrumental stakeholder theory as a tool to attain predicted results, mainly profitability (Kakabadse et al. 2005, p. 292). According to Donaldson and Preston (1995, p. 74), the instrumental aspect of stakeholder theory provides an indication of the relationship between stakeholder management and corporate performance. The instrumental side is based on the suggestion that applying stakeholder management will positively affect corporate financial performance (Berman et al. 1999;Donaldson and Preston 1995). Jones (1995, p. 430) has also highlighted that instrumental stakeholder theory might clarify the association between corporate social performance and financial performance as: 'Certain types of corporate social performance are manifestations of attempts to establish trusting, cooperative firm/stakeholder relationships and should be positively linked to a company's financial performance'.
The positive association between corporate social performance and financial performance can also be explained by 'good management theory', which is in essence another articulation of stakeholder theory. 'Good management theory' implies better relationships with key stakeholders that in turn will result in improved performance (Waddock and Graves 1997, pp. 306-307). The simple idea advanced by this theory is that social responsibility can be an intangible asset that results in a more efficient use of resources, which in turn positively affects financial performance (Surroca et al. 2010, p. 465). Providing additional empirical evidence, Soana (2011, pp. 134-135) has argued in this vein that companies, by having a socially responsible agenda, mitigate the potential damage of their reputation from negative information they may face in the future and therefore protect their profits and financial results. Thus, a good CSR programme can help to generate valuable goodwill, which will protect companies from unexpected issues and open doors to new prospects that cannot be accessed by companies that are not as socially active. Consequently, good stakeholder management may lead to competitive advantage Salomon 2006, p. 1102). Empirical evidence for good management theory can be found in the studies by McGuire et al. (1988McGuire et al. ( , 1990, who used it as a dependent variable in their estimation with financial performance. According to the proponents of instrumental stakeholder theory and 'good management theory', corporate social performance results in better financial performance. Other studies contend that companies with better financial performance have more resources to invest in social projects. In addition, several supporters of 'slack theory' argue that enhanced social performance will arise as a result of allocating slack resources to CSR initiatives, which means that better financial performance predicts better social performance (Waddock and Graves 1997, p. 306;Preston and O'Bannon 1997, p. 423).
In addition to such theorisation, some other studies find a strong positive association between corporate social performance and financial performance, as in the case of Simpson and Kohers (2002) who undertook a study of the US banking sector. There are also several studies, such as Orlitzky et al. (2003) and Wu (2006) that have used metaanalysis and found support for a positive relationship.

Hypothesis Development
The previous section presented the theoretical arguments concerning the relationship between corporate social performance and financial performance. Based on the arguments of 'instrumental stakeholder theory' and 'good management theory', which suggest a positive relationship between corporate social performance and financial performance, it is expected that social performance will positively affect financial performance. Thus, the following hypothesis is tested in this study:

H1a
The higher the level of CSR disclosure, the better the financial performance of GCC Islamic banks.
In search of the nexus identified, empirical studies have examined social responsibility and concurrent financial performance. However, a number of studies have taken lead/lag issues into account and assessed the possible impact of corporate social performance on subsequent financial performance and vice versa (McGuire et al. 1988;Roberts 1992;Pava and Krausz 1996;Preston and O'Bannon 1997). Considering that the theoretical arguments proposed for the concurrent relationship between CSR and corporate financial performance can be valid for the relationship between CSR and subsequent financial performance (McGuire et al. 1998) and also that the potential advantages of employing CSR may appear later on and positively affect firms' financial performance, it is expected that CSR disclosure will have a positive impact on the future financial performance of GCC Islamic banks. Therefore, the following sub-hypothesis is developed: H1b The higher the level of CSR disclosure, the better the future financial performance of GCC Islamic banks.
In addition to this, a number of researchers (e.g. Mahoney and Roberts 2007;Makni et al. 2009;Mahoney and Thorne 2005;Hillman and Keim 2001;Backhaus et al. 2002;Waddock and Graves 1997;Griffin and Mahon 1997) have stressed the importance of individual components of the total social performance index when examining the impact of corporate social performance on financial performance, suggesting that 'interesting and explanatory information is lost' when an aggregate measure of corporate social performance is used (Johnson and Greening 1999, p. 574).
In explaining this, the findings of Makni et al. (2009), who tested the relationship between corporate social performance and financial performance of publicly held Canadian firms using measures of corporate social performance from the Canadian Social Investment Database, showed no statistically significant relationship between an aggregate measure of corporate social performance and financial performance. On the other hand, their empirical findings indicate the existence of a significant relationship between the individual dimensions of corporate social performance, such as the environment and employees, and financial performance, justifying the modelling of individual dimensions. It should be noted that similar results can be found in Mahoney and Roberts (2007), whose study tested the link between the corporate social performance and financial performance of Canadian firms over the period 1997-2000. Their findings also indicate a non-significant relationship between the combined measure of companies' corporate social performance and financial performance. Turning to the individual measures of CSR, Mahoney and Roberts (2007) reported that the measures for the environment and international activities in the CSR construct significantly affect financial performance.
The importance of taking into consideration the individual measures of corporate social performance is also stressed by Fisman et al. (2005), who found a positive significant correlation between the community dimension of corporate social performance and financial performance in advertising-intensive industries. Thus, as corporate social performance is multi-faceted, there is a need to look at the individual dimensions as they might affect financial performance differently (Buckingham et al. 2011, p. 13). This is also considered in this study.
Taking into account the outcomes of previous research, it is important to highlight the multidimensional nature of corporate social performance and the need to disaggregate it into sub-dimensions to gain an improved understanding of the relationship investigated. Therefore, it is expected that the individual dimension variables of CSR disclosure, namely 'mission and vision', 'products', 'zakah', 'employees', 'debtors' and 'community' will positively affect financial performance. As a result, this research develops another hypotheses as follows: H2a All the composite dimensions of CSR disclosure have an individual positive impact on the financial performance of GCC Islamic banks.
H2b All the composite dimensions of CSR disclosure have an individual positive impact on the future financial performance of GCC Islamic banks.

Empirical Framework and Research Method
This section presents the operationalisation of the research by identifying the specific aspects related to the research method.
The Impact of Corporate Social Responsibility Disclosure on Financial Performance: Evidence…

Research Sample
As this research assesses the CSR disclosure of Islamic banks in the GCC region and examines the impact of CSR disclosure on banks' financial performance, the sample is selected from GCC Islamic banks. The rationale for selecting the GCC region is that the GCC states are the global leaders in Islamic banking and finance (Wilson 2009, p. 2). Indeed, the GCC Islamic banking industry controls 31.29 % of the assets of the entire region's banking sector (The Banker 2013, p. 3). Furthermore, GCC Islamic banks operate under similar economic conditions, making the analysis more homogeneous.
The sample for this research consists of 24 fully fledged Islamic banks from five GCC countries, namely Bahrain, Saudi Arabia, Qatar, Kuwait and the United Arab Emirates (UAE). Despite being a GCC country, Oman was not included in the sample as, at the time of data collection, there were no Islamic banks in the country due to the negative approach of the Omani administration towards Islamic finance (Wilson 2009, p. 31).
The research sample covers 2000-2014, a period of 15 years, which as a period witnessed an increased awareness of CSR practices among policy makers and regulators worldwide, hence, it is important to examine the impact of such awareness of CSR on the level of CSR disclosure in Islamic banks and whether such disclosure has any impact on financial performance of the sampled banks. In addition, it can be argued that Islamic banks deemed to show more sophistication in their activities and enhance their databases starting from year 2000 and onward, as most of the annual reports of the GCC Islamic banks were not available prior to this year. Hence, the year 2000 was taken as a starting point of the analysis. Furthermore, it is true that Islamic banks have been around since 1975; however, the third wave of institutionalisation and internationalisation or globalisation 2 has taken place with the beginning of the new millennium. While policy circles and private financial sector in the world including the Muslim world were mostly hesitant to engage with Islamic finance, since the year 2000, the political and business will have overcome the initial hesitant attitude and has commenced to engage with Islamic finance all over the world. Considering that Malaysian Islamic finance industry is most developed and sophisticated market, the public policy sources of such expansion has been in the last 15 years, as before Islamic financial development was rather sluggish in Malaysia too. Thus, the new millennium constitutes an important turning point and corner stone in the history of relatively young Islamic banking industry.
The population of Islamic banks in the study for each country is as follows: nine banks in Bahrain, three banks in Kuwait, three banks in Qatar, four banks in Saudi Arabia and five banks in the UAE. The sample thus consists of 24 Islamic banks. The main motivation for choosing these Islamic banks was the availability of annual reports. The distribution of the banks in terms of countries is as follows: The annual reports of the banks in the sample were obtained from the websites of the banks in question to measure the CSR disclosure. The sample covers annual reports for the 24 banks over 15 years from the GCC region. However, some of the examined banks do not have all annual reports available on their databases for the period in question resulting into missing data, and as a result, the sample yielded 222 reports. As for the bank level financial data, Bankscope, which is the global database of banks, was used to collect the financial statements of banks, providing the financial data to measure the impact of CSR disclosure on the financial performance of the GCC Islamic banks.

Measuring Financial Performance
Empirical studies examining the link between corporate social performance and financial performance have used different measures of financial performance. Indeed, Griffin and Mahon (1997, p. 11) identified 80 financial measures of corporate financial performance adopted in 51 studies. Among the measures of financial performance widely used are accounting-based measures of profitability (see Aupperle et al. 1985;Freedman and Jaggi 1982;2 The first phase of institutionalisation from 1975 to mid 1980s was retail Islamic banking with one country one bank strategy mainly in the GCC, Malaysia and Pakistan; the second phase of institutionalisation, running from mid-1980s to the new millennium, was commercial Islamic banking with emerging competitive domestic environment; the third phase came with the development of Islamic financial and capital markets along with Islamic investment banks and international regulative and standard setting bodies since the new millennium. This last period has witnessed the increasing pace of Islamic banking and financial diffusion including its penetration into non-Muslim markets and secular Muslim societies. Waddock and Graves 1997;McWilliams and Siegel 2000;Simpson and Kohers 2002), such as return on assets (ROA), return on equity (ROE) and return on sales (ROS), or market-based measures, such as market return, price-toearnings ratio and market value to book value (see Vance 1975;Freedman and Jaggi 1986).
Accounting-based measures of corporate financial performance have generally been perceived in the literature as indicative of past or short-term financial performance. On the other hand, market-based measures capture future or long-term financial performance (Gentry and Shen 2010, p. 514). The proponents of accounting-based measures believe that market-based measures can be affected by several factors which are not related to the firm's activity. The proponents of market-based measures, however, dispute the objectivity of accounting figures and stress the relevance of value-based measures related to shareholders and investors (Brammer andMillington 2008, p. 1333). Griffin and Mahon (1997, p. 11), on the other hand, highlight the importance of employing traditional accounting performance measures as modern, value-based measures might reflect more than purely corporate financial performance. Despite the limitations of accounting-based measures of financial performance, they have been widely accepted in the banking sector as the most accurate in capturing the financial performance of banks (Simpson and Kohers 2002, p. 99). Furthermore, the results of previous studies advocate that corporate social performance is more likely to have strong correlations with accounting returns than with investor returns (Orlitzky et al. 2003;Peloza 2009).
Taking into consideration these arguments and the fact that not all the banks in the sample have common stock traded on stock exchanges, market returns have not been used here to assess financial performance. Therefore, in this study, accounting-based variables, in particular, return on average assets (ROAA), are adopted as a proxy of corporate financial performance. In addition, return on average equity (ROAE) is used for robustness checks throughout the analysis. In the estimation, the ROAA ratio is calculated by dividing net income by average total assets for the sampled Islamic banks, while ROAE is defined as the ratio of net income to average equity, both ratios being expressed as percentages.
It should be noted that ROAA is an indicator of banks' financial performance and managerial efficiency as it depicts how competent the management is in generating profits from assets and how efficient in managing assets to generate revenue. This study uses the value of average assets to make it possible to determine changes in assets during the fiscal year. This ratio has appeared in the financial literature as the fundamental and most frequently used ratio of bank performance (Dietrich and Wanzenried 2011, p. 311).
Another measure of profitability employed in this study is ROAE, which reflects the rate of return on base capital. Despite being widely employed in financial studies, this ratio is not been regarded as the best measure of financial profitability by some, who argue that banks with a lower leverage ratio or higher equity tend to have a higher ROAA but a lower ROAE (Dietrich and Wanzenried 2011, p. 311). The reason for this is that ROAE neglects the higher risk related to high leverage and the impact of regulations on leverage. Therefore, in this study ROAA is used as the key dependent variable.
It should also be noted here that ROAE is generally strongly associated with ROAA; however, it is useful to employ this additional measure of financial performance to gain an insight into the implicit opinion of investors regarding the common stock of companies (Pava and Krausz 1996). Thus, the results for ROAE are also reported.

Constructing the Measure for CSR Disclosure
Corporate social performance is a multidimensional concept (Waddock and Graves 1997, p. 304), 'with behaviors ranging across a wide variety of inputs (e.g. investments in pollution control equipment or other environmental strategies), internal behaviors or processes (e.g. treatment of women and minorities, nature of products produced, relationships with customers), and outputs (e.g. community relations and philanthropic programs)'. Therefore, by definition, measuring CSR disclosure has to take this multidimensionality into account. In the case of Islamic banking, further layers of dimensions have to be considered in the measurement of CSR performance due to peculiarities imposed by Islamic ethicality and the legal framework.
The first studies examining the relationship between corporate social performance and financial performance employed a dominant, single dimension of corporate social performance and thus suffered measurement problems (Griffin and Mahon 1997). Therefore, several researchers have stressed the importance of developing a multidimensional concept of corporate social performance measures (Griffin and Mahon 1997;Roman et al. 1999;Carroll 2000). However, as Simpson and Kohers (2002) highlighted in their study, the issues related to the measurement of corporate social performance have still not been resolved.
To bring clarity to corporate social performance measurement, Orlitzky et al. (2003, p. 408) classified their measurement strategy into four main groups for disclosures: 'reputation indicators', 'social audit', 'corporate social performance processes and observable outcomes' and 'managerial corporate social performance principles and values'. Exploring further, Peloza (2009) divided corporate social performance measures into three main groups: 'environmental', 'social' and 'broad', the latter consisting of both 'social' and 'environmental' matters. More recently, based on corporate social performance metrics, Perrini et al. (2011) categorised the existing studies on the link between corporate social performance and financial performance as follows: those using 'pollution indicators', which mainly employ a single dimension measure (see Bragdon and Marlin 1972;Bowman and Haire 1975;Fogler and Nutt 1975); those 'implementing environmental practices' (see Christmann 2000); those 'employing corporate reputation ' (Alexander and Buchholz 1978;Cochran and Wood 1984); those 'applying thirdparty social and environmental assessments', such as the KLD database developed by the firm Kinder, Lydenberg, Domini (KLD), which aims to evaluate corporate social performance throughout a variety of dimensions associated with the interests of stakeholders (see Waddock and Graves 1997;McWilliams and Siegel 2000).
For this study, the CSR disclosure index employed as the measure of corporate social performance for the GCC Islamic banking industry was generated through content analysis of the annual reports of the Islamic banks in the sample. As mentioned, this index represents a multidimensional construct, extracted into a single measure reflecting the CSR activities of the banks based on a benchmark derived from Islamic financial principles.
The use of the CSR disclosure index as a measure of corporate social performance is subject to several limitations. It should be taken into account that this measurement, similar to any other measurement of social reality, can be questioned on the grounds of its 'objectivity' (Font et al. 2012, p. 13); therefore, the individual measures developed and employed are a reflection of individual values through which the CSR is perceived and CSR practice is imposed on the corporations. Furthermore, an additional shortcoming is that financial institutions might potentially decide to misinform the users of annual reports to foster a better public image. Thus, the information revealed in annual reports can differ from real corporate activities (Turker 2009, p. 415). If this is the case, the published annual reports might not show a reliable and true picture of financial institutions; hence, any index developed based on such information will be open to debate.
However, the reputational indices and comprehensive databases used in the conventional banking sector as societal barometers and financial instruments are, unfortunately, not available for GCC Islamic banks. Therefore, it can be argued that the CSR disclosure index employed in this paper to measure corporate social performance is the most efficient method available.
Recalling that the aim of this research is to assess the impact of CSR disclosure practices on financial performance of Islamic banks, the dimensions and sub-dimensions of CSR were carefully drawn based on a social responsibility framework in line with the Islamic finance perspective, employing existing literature on Islamic finance and CSR. In providing a particular framework, certain industry standards were also employed, such as AAOIFI Standards, which developed the Governance Standard No. 7: Corporate Social Responsibility Conduct and Disclosure for Islamic Financial Institutions, addressing guidelines for CSR. Thus, in this study, in line with the guidance provided by AAOIFI's Standard No. 7 and the previous studies by Haniffa and Hudaib (2007), Aribi and Gao (2012) and Aribi and Arun (2015), who developed CSR categories and their items based on the Islamic financial perspective on CSR, the dimensions and sub-dimensions for the measurement of CSR disclosure for Islamic banks were identified and the CSR disclosure index was constructed.
The resulting CSR disclosure index for the Islamic banks, as mentioned before, comprises six dimensions: 'mission and vision statement'; 'products and services'; 'commitment towards employees'; 'commitment towards debtors'; 'commitment towards society'; 'zakah, charity and benevolent funding'. The last dimension is specific to Islamic banking; however, the other five dimensions of social responsibility are also applicable to the conventional banking industry. In this research, the environmental dimension, proposed and used by numerous studies that focus on CSR disclosure in conventional and Islamic financial institutions (Mallin et al. 2014;Aribi and Arun 2015), is not included due to the fact that information related to environmental issues was not disclosed in any of the annual reports of the GCC Islamic banks in the sample. The index used in this study can be found in the Appendix in Table 10.
The annual reports of Islamic banks in the GCC countries were carefully analysed by identifying statements describing each dimension and sub-dimension of CSR. In the analysis, this research adopted a dichotomous approach (see Cooke 1989;Hossain et al. 1995;Haniffa and Hudaib 2007) in developing a scoring scheme to determine the extent of CSR disclosure in the annual reports. If an Islamic bank in the sample disclosed an item included in the CSR disclosure index, it received a score of (1) and (0) otherwise. Added together, these scores are equal to the total amount of CSR disclosure per Islamic bank.
The CSR disclosure index constructed based on the dichotomous approach is unweighted and assumes that each item of disclosure is equally significant (Cooke 1989, p. 115). It is suggested that unweighted indices are suitable for studies (as in the case of this research), which do not focus on a particular group of users of annual reports, but target all users (Cooke 1989, p. 115). In other words, an unweighted index permits 'an analysis independent of the perception of a particular user group' (Hossain and Hammami 2009, p. 259).
For the estimation of the index, CSR disclosure is calculated as the ratio of points awarded over the total number of selected dimensions following Haniffa and Hudaib (2007, p. 103): where CSR disclosure index jt denotes the CSR disclosure index for dimension j and period t; X ijt is variable X (1, … n) for dimension j and time t; N is the number of variables/ statements.

Reliability of Content Analysis
Reliability and validity are two important issues that need to be addressed in content analysis. It must be demonstrated that the coded data are reliable; as Weber (1990, p. 12) states, 'to make valid inferences from the text, it is important that the classification procedure be reliable in the sense of being consistent: Different people should code the same text in the same way'. The reliability of the coding process is often demonstrated by the agreement of two or more coders concerning the coding of the content of interest, or by showing that the inconsistencies have been re-considered and the disparities resolved (Milne and Adler 1999, p. 238). Furthermore, (Krippendorff 1980) identified three forms of reliability for content analysis and disclosure measurement, namely stability, reproducibility and accuracy.
To evaluate the reliability of the content analysis conducted in this research, the following steps were undertaken. First, reliability was determined by conducting a pilot study using 13 randomly selected annual reports covering 10 % of the sample examined. Second, to check the consistency of the scoring, content analysis of the annual reports used in the pilot study was conducted again after a month. Third, the annual reports used in the pilot study were given to two independent coders. The dimensions and sub-dimensions of the CSR disclosure index were explained to the coders and they were asked to assess the content of CSR disclosure in the annual reports and assign related scores. Finally, decision rules were established and revised to facilitate the codification of data collected from the content analysis of the annual reports.
After completing the pilot study, repeating it over time and taking into consideration the evaluations of independent coders, minor changes in the disclosure dimensions and subdimensions and the decision rules defining them were accepted. Consequently, the accuracy of the coding process was assured and its objectivity and reliability enhanced.

Defining Control Variables
Based on previous theoretical and empirical studies, which have emphasised the importance of mediators in the link between social performance and financial performance, it has been suggested that the relationship between corporate social performance and financial performance should possibly be examined by controlling certain other variables, such as institutional size, industry specification, how closely they interact with each other and the socially responsible practices performed by institutions (Brammer andMillington 2008, p. 1326; see also Orlitzky 2001;Rowley and Berman 2000;Margolis and Walsh 2003).
Following the findings of other studies, this study includes bank size, risk, overhead expenses, capital ratio and loan ratio as control variables, as they are viewed as the most important factors in the linkage between corporate social performance and financial performance. For example, the log of total assets is used as a proxy of bank size (Claessens et al. 2002;Gorton and Schmid 2000). According to Demirguc-Kunt and Huizinga (2000), the bank profitability is affected by different factors including financial and legal factors that are related to bank size. It should be stated that bank size is strongly correlated with the capital adequacy as to some extend banks with large size appear to attract cheaper capital, which in return leads to higher profits (Short 1979). Therefore, it is expected that there should be a positive association between bank size and financial performance. The supporting evidence for the positive effect of size on a firm's profitability can be found in a number of empirical studies, which, among others, include Smirlock (1985), Bikker and Hu (2002), Goddard et al. (2004) and Athanasoglou et al. (2008).
In addition to size, capital ratio is another major internal determinant of bank profitability. Consistent with prior research, this ratio is defined as equity over average total assets (Simpson and Kohers 2002), which is an indicator that banks have sufficient equity relative to potential risk and shocks. Banks with a high capital ratio require less external funding and indicate higher profitability (Kosmidou 2008, p. 151).
Considering that well-capitalised banks are more efficient in following up on available business opportunities, are more flexible and are able to cover unforeseen losses generated during times of crisis, consequently attaining a higher level of profitability (Athanasoglou et al. 2008, p. 132), it is expected that a higher capital ratio for GCC Islamic banks will result in greater profitability. The positive correlation between the capital ratio and bank profitability has been shown empirically in several studies, such as Bourke (1989), (Demirguc-Kunt and Huizinga 2000), Athanasoglou et al. (2008) and Kosmidou (2008). Furthermore, the loan ratio is one of the bank-specific covariates that may have an effect on bank profitability (Chronopoulos et al. 2013). The loan to total assets ratio, which is one of the direct indicators of the earning power of the bank, is calculated by dividing average total loans by average total assets (Simpson and Kohers 2002). As GCC banks rely heavily on traditional financial methods, with deposits and loans being the main sources and uses of funds (Al Hassan et al. 2010), loans are the main sources of revenue (Demirguc-Kunt and Huizinga 2000) that positively affect bank profitability. Therefore, it is expected that the loan ratio will have a positive impact on profitability in the case of GCC Islamic banks and it is thus used in this study.
A number of empirical studies consider operating expenses to be one of the major determinants of bank profitability. Operating expenses can be defined as the expenses incurred in undertaking banks' everyday financial activities, calculated as total non-interest expenses divided by average total assets (Simpson and Kohers 2002, p. 103). The findings of numerous studies on the profitability of banks have suggested that variables related to expenses have to be included in the profit function (Bourke 1989;Athanasoglou et al. 2008). According to Athanasoglou et al. (2008, p. 128), overhead expenses have a negative effect on profitability and, therefore, an effective management of overhead expenses is required to boost the efficiency of banks and enhances their profitability. Hence, it is expected that overhead expenses will negatively affect profitability in the GCC Islamic banks.
Moreover, risk is used as a control variable in this study as it has an impact on the relationship between social and financial performance (Waddock and Graves 1997, p. 308). The debt ratio, which is calculated as long-term debt divided by total assets, is used as a proxy for the riskiness of a bank (Waddock and Graves 1997;Kapopoulos and Lazaretou 2007;Barnett and Salomon 2012). Taking into account that debt reduces managerial freedom and may restrict access to new business opportunities, it is expected that the debt ratio will have a negative impact on financial performance in the GCC banking sector.

Empirical Model
To test the hypotheses constructed above, the following models are constructed with the dependent and independent variables described above: Model (1): Model (2): Model (3): Model (4): where ROAA is the return on average assets ratio, calculated by dividing net income by average total assets; CSRD is the CSR disclosure score, measured as the ratio of disclosure content points over the maximum score a bank can achieve; individual dimension refers to six dimensions that composite the aggregate measure of CSR disclosure including 'mission and vision', 'products and services', 'zakat, charity and benevolent fund', 'commitment towards employees', 'commitment towards debtors' and 'commitment towards community'; Size is the log of total assets; Cap. ratio stands for the capital ratio, calculated by dividing equity capital by average total assets; Loan ratio denotes the loan ratio, which is calculated by dividing average total loans by average total assets; Ovhd exp. stands for overhead expenses measured by dividing total non-interest expenses by average total assets; Debt ratio is the debt ratio, calculated by dividing long-term debt by total assets; a is the intercept; b 1 …b n are the regression coefficients; e is the error term. Table 1 presents the definitions of the independent variables and their predicted signs.

Results and Discussion
This section presents the findings derived from the empirical process and modelling described above. Table 2 presents the descriptive statistics of the variables as part of the model examined, covering both the dependent and independent variables across the 222 observations collected.

Descriptive Statistics
As the results in Table 2 show, the highest ROAA value is 0.1643 and the lowest is -0.1429 with a standard deviation of 0.0295 and a mean of 0.0192. Thus, as the results show, the ROAA of Islamic banks in the GCC countries over the period 2000-2015 is quite stable. However, the mean of ROAA shows a sharp decline from 2008 to 2010 ranging between 0.028 and 0.0001, respectively. This indicates that during the recent financial crisis, Islamic banks earned lower returns on assets than before the crisis. It can also be argued that although Islamic banks were affected by a lesser degree from the global financial crisis of 2007-2009, their financial performance was negatively affected. From late 2010-2014, a modest positive trend can be observed in ROAA with a mean value ranging between 0.0093 and 0.012, respectively.
In terms of the CSR disclosure index, CSR was disclosed in the banks' annual reports at different concentrations and with varying levels of information. The maximum CSR disclosure score found was 0.8056 across the six dimensions of the CSR disclosure index, while the minimum was 0.0944. The mean of CSR disclosure 0.4956 for the entire sample of banks over the period covered indicates a relatively low level of CSR disclosure, despite Islamic ethics being at the heart of Islamic banking. The standard deviation is 0.1536, reflecting a small level of dispersion from the mean. The level of CSR disclosure for the sampled GCC Islamic banks is quite steady from 2000 to 2007 with a mean value ranging between 0.0112 and 0.05. Such an increase in the CSR disclosure level in 2007 could be an indicator that Islamic banks attempt to reduce the communication gap between them and stakeholders with an attempt to improve the image of the banks in the wake of the global financial crisis of 2007-2008. Bolstered with the confidence of customers, who remained confident in Islamic banks throughout the financial crisis, as they provided more stability, the level of CSR disclosure by the GCC Islamic banks slightly declined between the period of 2007-2009, but showed better results compared to the precrisis period. The positive improvements in disclosing socially responsible activities of the GCC Islamic banks over the period 2010-2014 can be observed. It can be argued that such improvements come as a result that Islamic banks recognise and accept an important role of CSR disclosure in shaping public opinion of the existence and performance of the corporation and assists companies to comply with the legal requirements of the countries where they operate (Deegan et al. 2000). Such a positive development can also be attributed to the impact of the global financial crisis, which called for a more ethical banking and finance including communication information and being transparent.
In terms of the control variables, the size of organisations ranges between 2.2011 and 6.2648 with a mean value of 3.7292. The loan ratio varies between 0.0181 and 0, having a mean of 0.0008, while the highest capital ratio is 0.9497 and the lowest is 0.0772, having a mean of 0.2155. In addition, as can be seen in Table 2, overhead expenses

Findings
To test the hypotheses, panel data regressions with a fixed effects model were used. As part of the empirical process, to examine the validity of the data used before conducting the multiple regression tests on the panel data, some essential statistical tests were performed. Skewness and kurtosis standards are explored to test the normality of the data. As some of the variables examined are not normally distributed as a result of outliers, this study normalised the data distribution through winsorising (Dhaliwal et al. 2012, p. 732;Artiach et al. 2010, p. 40). With winsorising, the data assessed are normally distributed, the skewness scores reaching ±1.96 and the coefficient of kurtosis in the range of ±3 (Haniffa and Hudaib 2006). The results are depicted in Table 3, demonstrating the normal distribution of data.
Based on the normally distributed data, the Pearson correlation matrix and VIF test were employed to check for the existence of multicollinearity between the independent variables examined. As shown in Table 4, the Pearson correlation matrix fails to detect a correlation value equivalent to or higher than 0.8 (Brooks 2008;Haniffa and Cooke 2005;Jing et al. 2008) and therefore the variables examined are not highly correlated. As can be seen in Table 4, the highest reported VIF value is 2.35 for the CSRD variable and the lowest is 1.02 for overhead expenses. Considering that the commonly accepted threshold for multicollinearity is a VIF value of 10 (Hair et al. 2010, pp. 204, 212), these results are considered statistically acceptable, showing that multicollinearity is not present.
To test for heteroscedasticity, the Breusch-Pagan/Cook-Weisberg test was used. The results are given in Table 5. As can be seen, the P value is equal to 1.20, enabling the acceptance of the null hypothesis implying no threat of heteroscedasticity.
In terms of the rationale for using fixed effects, based on the P value of the Hausman test (0.000; significant at 1 %) shown in Table 5, the null hypothesis is rejected, signifying that the difference in coefficients is systematic and hence suggesting that using fixed effects is most appropriate for the data examined.
The results of the regression estimates in Table 5 show the relationship between the CSR disclosure index and financial performance. The overall model is significant at P \ 0.01 and explains 31.07 % of the variation in profitability (ROAA) in the GCC Islamic banks. As can be seen in Table 5, the CSR disclosure variable is statistically significant at P \ 0.05. Based on these results, there is a positive relationship between the CSR disclosure index and ROAA. As the results show, four out of five control variables are found to be significant: size at 0.01, loan ration and capital ratio at 0.05 and debt ratio at 0.1. This indicates that the equation is reliable. The significant association between CSR disclosure and ROAA is consistent with the results evidenced in previous studies, such as those of Simpson and Kohers (2002) and Scholtens (2009).
The findings of the analysis support Hypothesis 1 and are in line with the theoretical frameworks discussed that predict a positive link between CSR disclosure and financial performance in the Islamic banking industry. Therefore, it can be inferred that the higher the level of CSR disclosure, the better an Islamic bank's profitability.
The results of this study are consistent with instrumental stakeholder theory, the 'social impact hypothesis' and 'good management theory'. The significance of CSR means that the results support instrumental theory, which suggests that applying stakeholder management will positively affect corporate financial performance (Marom 2006;Donaldson and Preston 1995), as is the case established by the findings in this study. Furthermore, the findings are in accordance with 'good management theory', meaning that maintaining a good relationship with stakeholders will lead to better financial performance (Waddock and Graves 1997). In other words, social responsibility results in the effective use of resources, in turn increasing financial performance. This study demonstrates such a relationship in the GCC countries during the period in question. The empirical results of this study agree with the results obtained by Preston and O'Bannon (1997) and McGuire et al. (1988McGuire et al. ( , 1990 but are contrary to those of Mahoney and Roberts (2007), who found no significant relationship between corporate social performance and financial performance. Table 6 displays the impact of corporate social performance on future financial performance. Using Model 2, future financial performance is measured as ROAA t?1 and all independent variables are lagged 1 year from Model 1. The empirical results shown in Table 6 indicate a positive and significant relationship between corporate social performance and future financial performance at 0.1 significance level. The results are consistent with the proposed hypothesis H1a. These results agree with the theoretical arguments proposed by McGuire et al. (1988), suggesting that current socially responsible activities carried out by firms might have a long-term impact on financial performance.
The positive effect of corporate social performance on future financial performance may be due to the positive impact of CSR disclosure on bank reputation. Consequently, banks that are more socially active may increase customer loyalty and receive the support of a wider range of stakeholders, which in turn may positively contribute to financial performance. The positive and significant empirical results may also indicate that investors take into consideration banks' CSR activities.
In the regression in Model 3, the composite measure of corporate social performance is substituted by the individual dimensions of the CSR disclosure index. The results in Table 7 show no statistically significant relationship between the individual dimensions of corporate social performance and the financial performance measure, except for 'mission and vision' and 'products and services'. Model 3 indicates a significant positive association at 0.05 between the individual dimension of 'mission and vision' and ROAA, and a significant positive association at 0.1 between the individual dimension of 'products and services' and ROAA. Accordingly, it can be stated that the result of this model is inconsistent with the proposed H2a, except for the dimensions 'mission and vision' and 'products and services'.
It should be noted that the significance of the relationship between ROAA and 'mission and vision' and the 'product and services' dimensions is consistent with instrumental stakeholder theory (Marom 2006;Donaldson and Preston 1995), the 'social impact hypothesis' and Groups 15 Observations 222 * P \ 0.1, ** P \ 0.05, *** P \ 0.01 Groups 15 Observations 200 * P \ 0.1, ** P \ 0.05, *** P \ 0.01 The Impact of Corporate Social Responsibility Disclosure on Financial Performance: Evidence… 'good management theory' (Waddock and Graves 1997), as explained earlier. The results suggest that Islamic banks show relatively stronger commitment to 'mission and vision' and the 'product and services' dimensions, as these two dimensions of CSR are purely related to the salient nature of Islamic banks and, in particular, Islamic financial principles, which essentially distinguish Islamic banks from their conventional counterparts, as also argued by Mallin et al. (2014) and Belal et al. (2014). In other words, Shari'ah compliancy and other Islamic ethical and Shari'ah commitments are expressed in 'mission and vision dimension' and articulated in the 'product and services' as a visible and everyday practice, their impact on the financial performance is found to be significant. Given that 'zakat, charity and benevolent funds' are considered as one of the key characteristics of Islamic banks, a possible justification of the lowest level of disclosure of this dimension and its insignificant impact on the financial performance could be explained by the fact that Islamic banks show relatively lesser inclination to disclose information related to this dimension that can be explained by Islamic ethics, which discourages individuals and institution in revealing their socially responsible activities and social giving, as such a revelation is considered against the dignity of those who receive and also is considered harming the initial intention of giving. In addition, it is well known that the distribution of such social giving mostly takes place in an unstructured manner implying that in distribution of such funds informal giving is more common.
With regards to the relationship between individual disclosure level of individual dimensions of corporate social disclosure and the future financial performance measure, the results are inconsistent with H2b, as the regression results detect no significant association, except for dimension of 'mission and vision' as shown in Table 8. This implies that the level of disclosure of these dimensions individually does not have a significant impact on future financial performance of the Islamic banks, except for the 'mission and vision' dimension. As suggested earlier, such a positive association could be result of the strong commitment of Islamic banks towards the 'mission and vision' dimension due to the fact that this dimension expresses the commitment to the essential principles of Islamic finance. Such a commitment can be evidenced from the disclosure score of 0.71 for 'mission and vision' dimension, which is the highest score compared to the disclosure scores of the other dimensions. Moreover, besides addressing the bank's philosophy and core principles, the mission and vision statement underlines the values and beliefs of an institution (Dermol 2012) that ''may describe desirable and possible future state or long term aims of a company and is also a backdrop of a purpose and company's strategy'' (Dermol 2012, p. 892). Therefore, by definition it defines the identity and the key priorities of a company by outlining, ''where a firm is headed; how it plans to get there; what its priorities, values, and beliefs are; and how it is distinctive'' (Williams 2008, p. 96).
Hence, communication of such information may in return affect a firm future financial performance, as the behavioural standards of that firm would be known by the investors. Accordingly, Islamic banks are obliged to operate, utilise and highlight instruments, roles and regulations of the Shari'ah in their 'mission and vision' statement that will positively result in a high level of disclosure, which leads to enhance its impact on future financial performance. In addition, 'mission and vision' dimension, containing the central principles of Islamic finance as a product of Islamic ethics such as the commitment towards Shari'ah compliancy, is more visible and tangible; and, therefore, the customers may move to such banks for their deposits and investments as the visibility provides them with trust. Since financial literacy and in particular emphasis on CSR is rather limited in the region, investors and depositors focus on the expression and articulation of Shari'ah compliancy and Islamic principles which are defined in 'mission and vision' statement.

Sensitivity Tests
To test the robustness of the empirical results of this study, two additional tests were conducted. First, ROAE was used as an alternative measure of profitability. Table 9 reports the regression results of the estimation of Model 1 using ROAE as a dependent variable and the same regression procedure adopted for Model 1 previously. The results presented in Table 9 for fixed effects regression in estimating the relationship between the CSR disclosure index and ROAE show general consistency with the results obtained when employing ROAA as the financial performance measure. The CSR disclosure index is significant at t = 2, P \ 0.1. These results are similar to those obtained from the main regression with ROAA used as the dependent variable. Second, to check for potential endogeneity between corporate social performance and financial performance, the Durbin-Wu test was applied. In a statistical model, endogeneity may occur as a result of joint determination between independent (corporate social performance) and dependent (financial performance) variables, or omitted variables, or if there is a correlation between explanatory variables and the error term (Greene 2003). The results in Table 9 indicate that the F-test is not significant and thus the null hypothesis of the Durbin-Wu test cannot be rejected, confirming that endogeneity does not represent a problem (Gujarati 2003) in this research.

Conclusion
This study examines the relationship between CSR disclosure performance and the financial performance of Islamic banks in the GCC region. The findings of this study indicate that there is a significant positive relationship between CSR disclosure and the financial performance of GCC Islamic banks. These results verify the hypotheses and are also in line with the theoretical framework that predicts a positive link between the corporate social performance and financial performance of the Islamic banking industry. Hence, it can be concluded that the higher the level of CSR disclosure, the better the bank's profitability in the case of GCC Islamic banking.
These results are supported by the findings of Model 2 in the empirical analysis, which tested the impact of social performance on future financial performance and indicated a positive relationship. Thus, the results align with the theoretical arguments advocating that current CSR activities carried out by financial institutions could have a longterm impact on financial performance (McGuire et al. 1988). The positive effect of corporate social performance on future financial performance may be explained by the positive impact of CSR disclosure on bank reputation. Islamic banks, which ensure that CSR is practised extensively beyond charity, may increase the loyalty of customers and receive the support of a wider range of stakeholders and in turn improve their financial performance. This is because CSR activities could provide legitimacy for banks; for Islamic banks, conducting extensive CSR activities could be perceived as fulfilling the requirements of Islamic ethics and norms as part of their existential reasoning. The positive significant relationship between CSR disclosure and future financial performance may also indicate that investors take into consideration banks' CSR activities. Hence, by being more socially responsible, banks can leverage new customers and more deposits, which will have a positive impact on their financial performance in the long term.
Furthermore, the empirical results indicate a significant positive relationship between the composite measure of the CSR disclosure index and financial performance. However, there is no evidence of a statistically significant relationship between the individual dimensions of the CSR disclosure index and the financial performance measure, except for 'mission and vision' and 'products and services'. With regards to the association between the individual dimensions of the CSR disclosure index and future financial performance, the empirical results show similar outcomes, where only one dimension, namely, 'mission and vision' has a positive and significant relation. The results of this study are consistent with instrumental stakeholder theory, the 'social impact hypothesis', and 'good management theory'.
The results also indicate that the majority of Islamic banks, despite the high expectations of full CSR disclosure, reveal significantly less than the level expected in terms of being shaped by 'Islamic ethics'. It seems that priority in the disclosure process is given to information related to their financial obligations towards their shareholders rather than to the interests of a wider range of stakeholders. This research thus confirms the results of previous studies in providing evidence that the CSR disclosure of Islamic banks lacks essential information.
The findings of this research suggest some of the key policy implications. To integrate the social dimension of the Islamic moral economy in Islamic banks' business strategy, it is important to make CSR objectives more explicit through reconsidering Islamic banks' approach Observations 222 * P \ 0.1, ** P \ 0.05, *** P \ 0.01 towards CSR activities, some dimensions of which are identified in this empirical research. As Islamic banks are obliged to operate according to the objectives of Islamic law and also Islamic morality, it is essential that they place CSR-related policies at the top of their business agenda to identify their 'Islamic moral economy' nature. Thus, without embedding social and ethical dimensions in their business strategy, these institutions could not be regarded as fulfilling the Islamic moral economy requirement of substance beyond form-oriented Shari'ah compliance, as substance distinguishes them from conventional banks in articulating and manifesting ethicality. It is the substance nature of Islamic banking which necessitates ethical business behaviour including disclosing information in relation to CSR and importantly conducting CSR activities, as Shari'ah compliancy relates only to rational-legal framework of negative screening relegated to interest or riba probation and limiting uncertainty or gharar. There is, hence, an urgent need to standardise and unify the financial and annual reports of Islamic banks. In this respect, AAOIFI released Governance Standard No. 7: Corporate Social Responsibility, Conduct and Disclosure for Financial Institutions, which provides guidelines for CSR. However, these standards are not obligatory for all jurisdictions and of the GCC countries. This voluntarily adherence to AAOIFI standards could be one of the main reasons for the low disclosure of the social performance of Islamic banks. Hence, it is vital to implement such standards as a mandatory policy for the Islamic banking industry, which would enhance Islamic banks' social performance and result in a higher level of CSR disclosure. Consequently, having CSR practices embedded into the working of Islamic banks is expected to generate positive contribution to social good or beneficence society and thus aid them in fulfilling their raison d'être as expressed by Islamic moral economy.
Furthermore, maintaining a good CSR policy would increase the capability of Islamic banks to cope with possible reputation-damaging events and the external negative news they may receive in the future and thus protect their profits and financial results. Therefore, having a comprehensive socially responsible agenda could assist Islamic banks in generating valuable goodwill that will safeguard them from unexpected challenges and give an access to new projects that are not available for companies with less CSR incentive. In essence, the 'Islamic prefix' in Islamic banking suggests an ethical identify and brand which should be upheld by Islamic finance industry rather than diverging from such ethical behaviour for the sake of more profit.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creative commons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Appendix See Table 10. Table 10 Dimensions and sub-dimensions of the CSR disclosure index (continued) Data Source: AAIOFI (2010), Haniffa and Hudaib (2007), Aribi and Gao (2012)