Disclosure of CSR Performance and Firm Value: New Evidence from South Africa on the Basis of the GRI Guidelines for Sustainability Disclosure

Abstract

Prior CSR and firm performance research has produced mixed results. Even so, numerous researches examining this relationship from the perspective of international standardisation have primarily concentrated on developed economics. This leaves an obvious gap within the extant literature with regards to evidence from sub-Saharan Africa. The aim of this study is to investigate the relationship between the extent of CSR disclosure performance and firm value, in an emerging institutional setting. Using hand collected data of South African listed companies, we apply the GRI G3.1 guidelines, as a measure of disclosure performance. Based on the panel data fixed effect model, we document a positive but insignificant relationship between CSR disclosure performance and firm value. Secondly, a negative and insignificant relationship was found between environmental disclosure performance and firm value. Lastly, we found a positive and statistically significant relationship between social disclosure performance and firm value. Overall, our findings suggest that CSR disclosure has a limited effect on firm value. Our findings hold for a set of robustness tests. Our findings suggest that the incorporation of sustainability disclosure, on the basis of GRI, is moderately high among the selected companies. Implications of our results suggest that CSR disclosure may not necessarily influence firm value, despite its numerous benefits. We contribute to this line of research from a multi-theoretical perspective.

1. Introduction

The concept of Corporate Social Responsibility (herein termed as CSR) and sustainable development is dynamic, multifaceted, and global, and has proven to be a contentious issue across the globe. For instance, it is assumed that CSR is largely a Western phenomenon, and that it is most likely to be found in countries with globally active organisations, democratic political structures, and active civil society organizations [1]. CSR as a concept has attracted worldwide attention and thus gained global recognition and prominence. The global corporate environment in recent times faces increasing pressure to take up initiatives and recognise its social and environment responsibilities. This has led to a growing interest in the question of whether the adoption of CSR disclosure and strategies by firms that actively support sustainable economic, social, and environmental development can play a key role in the generation of competitive advantage and firm value [2,3].

The notion that businesses have a responsibility of maintaining an impartial balanced working relationship among various stakeholders has been relevant to an understanding of a business strategy for the past half a century [4]. Performance of CSR and disclosure has therefore become an integral part of the corporate fabric [5]. Organisations disclose their social and environmental performance in many different ways, including annual reports, independent reports, websites and press releases, extra supporting documents for annual reports, advertisements, published articles and booklets regarding corporate social and environmental activities, and video tapes [6].

The rise of CSR and sustainability issues globally has led to a planetary rise in global standards for information disclosure over the past decades. Thus, there has been a worldwide diffusion of disclosure standards [7]. A study by Ligteringen and Zadek [8] identified over 300 global corporate standards that have evolved since the 1990s. These standards include the United Nations (UN) Development Goals and Global Compact, the directory of Organization for Economic Cooperation and Development (OECD) for multinational corporations, ISO 14001, the Global Reporting Initiative (GRI) guidelines, Sullivan Global Principles, SA 8000, Series AA1000, and ISO 26000 [7]. These standards encourage disclosures of social and environmental issues of corporate sustainability, among others. As a result of the global rise in standards, there have been calls for enhanced CSR and sustainability disclosures as well as any other information that can potentially impact on organisational performance. Consequently, there has been a rise in the number of organisations disclosing their social and environmental initiatives [9]. According to KPMG reports (KPMG, 2011; KPMG, 2013) 95% of the 250 largest companies in the world (based on the Fortune Global 500 ranking) produced a CSR report in 2011, a 14% increase from 2008. Also, a variety of companies appear to be increasingly engage in CSR disclosure related to environmental and social issues [10,11].

In this paper, we provide evidence in line with recent global upsurge of standards on debates within the broader framework of corporate sustainable development, and the social and environmental disclosure literature of CSR. We highlight the importance of expanding existing evidence on the relations among CSR disclosures and firm performance to different institutional settings that may allow for further findings in light of available inconclusive results within the CSR–FP nexus. The aim of this study is to investigate the relationship between the extent of CSR disclosure performance (and its components) on firm value in an emerging institutional setting where investor protection is relatively low, as a result of a weaker external corporate regulatory environment [12,13]. This study focus on South Africa, with the application of an internationally accepted standard, the Global Reporting Initiative (herein termed as GRI) 2011 (G3.1) guidelines related to sustainability disclosure, as a measure of CSR disclosure performance of sampled companies from the Johannesburg Stock Exchange (herein termed as JSE), an emerging market in sub-Saharan Africa.

Globally, recent trends indicate that many more organisations are paying increasing attention to CSR disclosure, its related organisational performances, and the issuance of CSR/sustainability reports based on international standards. This has become an important issue for both practitioners and academics, as a result of changing business dynamics and importance of organisations in today’s global environment, that is, the effects of globalisation and global capitalism [14,15]. As CSR and sustainability reporting are becoming effectively mandatory for global large multinational corporations (MNCs) it has also attracted a considerable amount of academic literature [16,17]. Various stakeholders continue to pressure organisations (i.e., both local and multinational) to improve upon their CSR performances and release the related reports. Such pressures include for instance, public concerns, media interest and advocacy, institutional and regulatory forces, consumer and industrial pressures, reputational concerns, and other perceived market advantages [18]. On the other hand, organisations may disclose their CSR as a way of sustainability development investment, and thus, an evaluation of their consequences should be relevant. Sustainability reporting guidelines developed by the Global Reporting Initiative (herein termed as GRI) provide a more organised approach for companies to disclose their performance on CSR dimensions of sustainability. The reason for the enormity of the subject is that reporting of a firm’s actions and inactions has become much more central because investors and other stakeholders are demanding greater transparency about business practices. To this end, it has been asserted that CSR disclosure is now a necessity for all kinds of organisations and is no longer a purview of only large corporations [19]. We argue that the extent of CSR disclosure and its associated transparency should have an impact on the firms’ wellbeing.

South Africa makes an interesting study case from sub-Saharan Africa, due to prior research and developments within its context. South Africa is noted for its long history of racial segregation termed as “apartheid”. Within apartheid, humanity was largely defined on the basis of race. It is well documented in literature that ‘no study of South Africa would be possible without an examination of the regime’s racial classifications’ [20]. Therefore, it has been argued that the practice of CSR in the context of South Africa was largely “apartheid-induced” [21]. The post-apartheid era therefore saw the formation of affirmative actions, policies, initiatives, and structures by the ANC government (i.e., ANC—the African National Congress is the Republic of South Africa’s governing political party and it has been the ruling party since the fall of apartheid in South Africa) after it took over power from the apartheid regime as a democratically elected government. One of the aims of the ANC government on assuming power was to rebuild the corporate image of South Africa, which occurred as a result of the racial imbalances that may have been created after the fall of apartheid. One of the most critical structures that have shaped the corporate and economic landscape of South Africa was the coming into force of the South African Corporate Governance Code of practice (popularly known as the King Reports). The first report was produced in 1994 (i.e., King I), the second in 2002 (i.e., King II), the third in 2009 (i.e., King III), and the most recent was produced in 2016 (i.e., King IV). It has been noted that these reports generally adopt the Anglo-Saxon style of corporate governance [12]; nevertheless, they also included some specific affirmative actions, rules, initiatives and stakeholder provisions.

Firstly, although empirical research on the disclosure of CSR and FP relations has attracted much interest among policy-makers and researchers in the emerging markets context, even so, this relationship seems to be erroneous and to some extent ambiguous. Hence, this relationship remains inconclusive [22]. As prior research in emerging markets on social and environmental disclosure grows in prominence relative to developed markets, a notable avenue by which emerging markets could demonstrate commitments to the global upsurge of standardisation in CSR disclosure trends and sustainable development is to promote a voluntary (and/or mandatory) use of such global standards among local and multinational corporations operating within their markets. It has been asserted that when organisations integrate, adopt, and disclosure on sustainability issues into their operations in a triple bottom context, it is termed as CSR [23]. However, social and environmental research has largely focus on the developed and industrialised economies (i.e., where better institutional settings, higher investor protection, relatively better levels of economic development, and sophisticated financial markets are in existence), with relatively limited prior research in emerging economies [24,25]. There is therefore a lingering need for improve research in social and environmental disclosure in emerging markets to commensurate the global call for corporate disclosures, in line with international standards.

Secondly, the call for CSR disclosures in line with international standards becomes even more desirable as issues of CSR and sustainability are well grounded in the Corporate Governance Codes of South Africa, making this research study unique from the context of South Africa, and for that matter, sub-Saharan Africa. The South Africa Corporate Governance Codes, since its inception, encourages corporations to integrate sustainability concerns into their operations and to disclosure based on the GRI. The King I and King II Corporate Governance Codes required listed companies on the JSE to report based on the GRI guidelines. Mitchell and Hill [26] assert that the GRI reporting in South Africa is endorsed by the King Report for Corporate Governance and the Johannesburg Stock Exchange for listed companies. However, the focal point of this study—the King III Corporate Governance Code—went further to encourage all entities in South Africa to integrate sustainability concerns into their operations, and to report based on the GRI guidelines in a triple bottom context of sustainable development. As such, all entities in South Africa are encouraged to integrate sustainability concerns into their operations, irrespective of the company size, nature, or type—see King III Corporate Governance of South Africa [27]. Therefore, it was timely to undertake research utilising internationally accepted standards (i.e., the GRI guidelines) to investigate the extent of CSR disclosure and its influence on firm value in sub-Saharan Africa, where research is relatively limited. Although literature notes that research on CSR (and disclosure) in emerging markets is relatively limited compared with advanced economies, it is worth recognising the rise in prominence [28] of CSR research in developing countries, and in particular South Africa [29,30,31,32]. However, and to the best of our knowledge, no research in today’s modern era in the South African context has utilised the GRI 2011 (G3.1) guidelines to investigate the relations among the extent of CSR disclosure performance (including components of CSR disclosure) of public-companies on firm value. Studies on CSR (and disclosure) in South Africa context have largely used affirmative action Acts [31] (for example, the Black Economic Empowerment Act—BEE Act and Broad-base Black Economic Empowerment Act—BBBEE Act), organisational survey data [29] (example, survey data by KPMG International) and other national level instruments [30] (example, JSE Socially Responsible Investment Index) to measure the extent of CSR disclosure. Also, we recognise that the past decade has witnessed research work exploring the usage of the GRI framework and GRI-guided reports.

Consequently, this study makes unique contribution to literature from South Africa context. To the best of our knowledge, this study is different from a study by Ntim [33], which examined the social responsibility role of public companies’ contribution towards reducing the negative effects of HIV/AIDS within the realms of social and environmental accounting (SEA), by adopting the GRI 2002 reporting guidelines specifically on HIV/AIDS and how the companies disclose their activities aimed at addressing the menace. Even so, this was a cross-country study within sub-Saharan Africa. Again, this present study is very different from a recent study by Horn et al. [29], which utilises survey data from KPMG International to investigate CSR disclosure in South Africa. Hence, this present study has a different scope which largely differs from the research works by Ntim [33] and Horn et al. [29]. This present study utilises a more recent data and the GRI 2011 (G3.1) of sustainability disclosure, as a measure to investigate the influence of the extent of CSR disclosure performance (and its components) on firm value. In particular, this present study manually collects CSR information disclosed from annual and CSR/sustainability reports of sampled companies to construct three separate disclosure indexes to examine their relations with firm value. In this case, this present study is very unique and different from other research works in the South African context, and as such, makes very substantial contribution to CSR literature in South Africa and in the context of emerging markets.

In this paper, we focus on listed companies on the JSE, South Africa for the following reasons. The rich corporate governance regime has, to a larger extent, shaped the corporate environment of South Africa towards sustainability reporting. In particular, the King III Code on Corporate Governance [27] explicitly encourages all entities in South Africa to integrate economic, social, and environmental issues of sustainability into their day to day operations, and to report based on the GRI. Lastly, South Africa is noted as one of the most ethnically diverse economies in the world [34]. For instance, in 2011, South Africa’s population stood at 79% Black Africans, 8.9% were coloured, 2.5% were of Indian/Asian origin, and 9.5% were White (Statistics South Africa, 2012) [35]; see also Gyapong et al. [36] p. 374 (to the best of our knowledge these are the most recent census statistics). This diverse nature makes it unique from its peers in the sub-Saharan region, accompanied with its continuous stability and economic landscape makes it a welcoming destination for a variety of multinational corporations (MNCs). The result of this is the significant rise in the number of company listings on the JSE, which has led to the attraction of more foreign direct investments into the country. Furthermore, South Africa has by far the largest financial system, relatively developed, and most sophisticated in Africa, and it compares well with the financial systems of the developed world, and has secured itself a spot among the top-20 stock markets in the world [37]. The JSE has remained a member of the Federation of International Stock Exchange since 1963, and is the only stock exchange in Southern Africa, but it has gained a mark as one of the largest in the world in terms of market capitalization [37]. With these dazzling features, South Africa compares well with some developed and middle-income emerging economies, hence exploring the extent of CSR disclosure performance of public companies on the JSE from an internationally accepted standard was not far from reach.

Using a hand collected dataset of 126 listed South African companies over the period from 2010 to 2015 we investigated the relationship between disclosure of CSR performance, based on G3.1 guidelines related to social and environmental performance indicators, and firm value. To do this, we employed a unique disclosure formula to construct three disclosure indexes. This study finds mixed and interesting results which are consistent with findings of prior studies. First, we examine the influence of the extent of CSR disclosure performance (as represented by social and environmental performance indicators of GRI—G3.1) on firm value. We find that CSR disclosure performance has a positive but insignificant influence on firm value. Even so, and with regards to the correlation matrix below, it also indicates a weak negative relation between CSR disclosure performance and firm value. Second, we examine the influence of the extent of environmental disclosure performance (as represented by environmental performance indicators of GRI—G3.1) on firm value. We find that environmental disclosure performance has a negative and an insignificant influence on firm value. Lastly, we investigated the influence of the extent of social disclosure performance (as represented by social performance indicators of GRI—G3.1) on firm value. We found a positive and statistically significant influence of social disclosure performance on firm value. Overall, our findings suggest that CSR disclosure has a limited effect on firm value. However, the influence of environmental disclosure on firm value may be weakening, if not becoming less important in driving firm performance. Our findings are robust after controlling for the potential effects of endogeneity. Our findings have significant implications in the context of an emerging market with weak investor protection. This study provides support for the theoretical reasoning in line with the stakeholder and legitimacy theories. For instance, the Stakeholder theory posits that a firm’s long-term existence depends upon addressing stakeholder concerns, and that CSR has thus expanded the firm’s responsibilities to the wider public. CSR disclosures are therefore ways of providing vital non-financial information of stakeholder needs about the firm, which could influence stakeholders’ perceptions about the firm and their investment pattern. Also, our findings support firms’ legitimacy, that is, where firms use CSR disclosure as a means of legitimising their actions and inactions. Legitimacy theory perceives CSR information as a strategy for managing perceptions of society regarding social and environmental impact on business operations [38]. This theory holds a widespread view that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, and beliefs [39]. Therefore the disclosure of CSR has become an integral part of corporate fabric [5].

This present study contributes to the CSR literature in developing countries, in several ways. Firstly, we consider that previous findings on the value relevance of CSR disclosure on firm performance in the developed/industrialised economies cannot be generalised to the emerging markets context, particularly sub-Saharan Africa. Even so, emerging economies, and for that matter sub-Saharan African countries, exhibit wider differences in development relative to other economies. Again, many local and multinational corporations can be perceived differently in the performance of CSR activities. Additionally, stakeholder demands for social and environmental disclosure may substantially differ within different institutional settings. Secondly, although CSR disclosure research has grown in prominence in South Africa, and that, corporate sustainability issues are well incorporated in Corporate Governance Codes of South Africa, CSR disclosure research utilising internationally accepted standards in a triple bottom context was still relatively limited. Therefore, this study shed light on a growing body of literature on international standardisation [40,41,42] of CSR disclosure towards sustainable development, by utilising the GRI guidelines. Third, and in support of the two points above, our study aims to broaden our understanding and fill an existing gap in literature, by providing new evidence, on the influence of CSR disclosure on firm performance, in line with global debates within the social and environmental disclosure literature, by providing new empirical evidence from a sub-Saharan African context, other than the developed and industrialised markets. Finally, this study makes theoretical contributions, in support of existing literature, that argue for a multi-theoretical perspective to explore the CSR disclosure practice of firms.

Our findings have important regulatory, policy, and research implications. This study suggests that CSR disclosure, on the basis of GRI guidelines, is moderately high for our sampled JSE companies over the period of study. This observation holds that CSR disclosure towards sustainable business operations was progressively being upheld in the South African context, and that, efforts by key stakeholders, like the JSE and the South African Corporate Governance Institute, to improve the adoption and incorporation of sustainable development have had positive impacts on companies’ operations. However, and like most administrative structures which are not without caveats; we suggest that attention be given to the credibility of some information that are disclosed by companies. From our findings we argue that the influence of CSR disclosure performance, on the basis of the GRI guidelines, serves to provide a common playing field to enable companies to be assessed from an internationally accepted standard, which is vital for performance evaluation.

The rest of this paper is organised as follows. Section 2 provides theoretical background, prior literature, and hypotheses development. Section 3 introduces the data, research design, and methodology. Section 4 presents the analyses and discusses the empirical results. Section 5 presents the robustness test. Finally, Section 6 presents a summary and conclusion.

2. Theoretical Background, Prior Research and Hypotheses Development

2.1. Multi-Theoretical View

Many different theories have been identified to explain CSR disclosure patterns, its importance, and necessity [43]. These theories and approaches are under the themes of economics, politics, social integration, and ethics [44,45]. Recognising that there may be limitations with respect to using a single theoretical perspective to explore the notion of CSR disclosure, as well as, given the multiplicity within corporate motivations for CSR disclosures [46], this present study adopts a multi-theoretical view to investigate the relationship between the extent of CSR disclosure performance and firm value of public companies on the JSE. Consistent with prior suggestions and evidence [31,47,48], we consider that these theories are complementary rather than opposing. Therefore, we propose that a combination of the theoretical perspectives is useful to explore CSR disclosure practices. Hence, this present study employs the Stakeholder theory and Legitimacy theory. Even so, it has been asserted that the Stakeholder theory is closely aligned with Legitimacy theory and that the two are often used to complement each other [49]. Again, recognising the existence of extensive literature on these theories, which is well documented in literature, we briefly describe these theories.

Stakeholder theory posits that a firm’s long-term existence depends upon addressing stakeholder concerns, and that CSR has expanded the firm’s responsibilities to stakeholders and to the wider public [50,51]. Porter & Kramer [52] document that CSR can be much more than a cost, constraint, or charitable deed and that it can, in fact, be a potent source of innovation and competitive advantage. Being socially responsible brings firms various potential benefits such as improved labour relations and employee productivity, less risk of litigation related to its products/services, fewer complaints from the community because of environmental issues, lower regulatory costs, and better brand recognition, among others. These potential benefits suggest that socially responsible firms will perform better. Stakeholder Theory [50], therefore, posits that corporate disclosure is an instrument for managing the informational needs of various powerful stakeholder groups. Theoretical literature has formulated competing hypotheses on the effect of CSR on firm performance. The first one is called “social impact hypothesis” and the second is the “shift of focus view”. While the former claims that CSR has a positive effect on firm performance, the latter suggests the opposite. According to the social impact hypothesis, three channels may explain the positive relationship between CSR and firm performance. These are improving employee productivity by providing a better working environment [53]; increasing managerial know-how and thus enhancing organizational efficiency [54]; and increasing social reputation, trust [55], brand image, and product competitiveness [56]. Lastly, modern Stakeholder theory contends that the value of the firm depends on the cost, not only of explicit claims, but also of implicit claims. From this viewpoint, the set of claimants on the firm’s resources goes beyond the stockholders and bondholders to include stakeholders who have explicit claims on the firm, like wage contracts and others with whom the firm has made implicit contracts, involving, for instance, quality service and social responsibility [57].

Legitimacy theory perceives CSR information as a strategy for managing perceptions of society regarding the social and environmental impact on business operations [38]. Legitimacy theory holds a widespread view that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values and beliefs [39]. Consequently, stakeholders can influence corporate actions by arguing that a given practice does not conform to societal expectations and/or lacks legitimacy [58,59]. Hence, when the legitimacy of an organisation is endangered, it could employ some legitimisation strategies to minimise those threats by seeking to influence stakeholder actions and perceptions [48,60]. For instance, within the context of South Africa, it has been argued that organisations could legitimise their actions by disclosing BEE information, notwithstanding its usefulness in explaining CSR disclosures motives [31]. De Villiers and Van Staden [38] asserted that in terms of legitimacy, when companies face changes in societal expectations they will change the extent and pattern of their disclosures in order to remain legitimate with their powerful stakeholders. These authors provide justification to support an understanding of legitimacy by adding a dimension of reduced disclosure levels by organisations as a legitimising strategy [38]. Therefore, organisational legitimacy theory predicts that corporations will do whatever they regard as needful in order to preserve their image as a legitimate organisation with legitimate aims and methods of achieving it [38]. Legitimacy has largely been used to support the idea that social disclosures will be maintained at present levels, or increased over time, to avert legitimacy crises. Proponents of legitimacy [38] have therefore predicted that maintenance and reductions, as much as increases, in social and environmental disclosure levels of firms are influenced by way of legitimacy. Woodward et al. [61] have shown that both stakeholder theory and legitimacy theory consider an organisation to be part of the wider social system. This notion is consistent with foundation pillars of the King III Corporate Governance Code. Whilst legitimacy theory looks at society as a whole, stakeholder theory recognises that some groups within the society are more powerful than others. Hence, disclosures are largely driven by stakeholder and/or societal pressure, and that such disclosures are likely to reduce information asymmetry.

2.2. GRI Guidelines and CSR Disclosure

The GRI guidelines are the first global standards for sustainability reporting. They feature a modular, interrelated structure and represent the global best practice for reporting on a range of economic, environmental, and social issues. The GRI reporting guidelines on CSR information has widely been acknowledged as a leader in the international standardisation of sustainability disclosure [62]. Further, it is regarded a primary tool for CSR disclosure, as it has worldwide applicability in multinational organisations that operate in a variety of industries [63]. It has been asserted that, firms that incorporate the GRI guidelines into their operations appear to have higher levels of commitment to CSR than those firms that do not incorporate these standards. Even so, opponents of the GRI framework have criticised it from a theoretical viewpoint that its sustainability principles seem to be widely dispersed [63]. Furthermore, the lack of specific, formal national and international regulations seems to allow organisations much flexibility on how they carry out their social and environmental disclosure activities, while allowing the biased use of these guidelines [64]. A research that used GRI guidelines as CSR reporting templates claimed that companies may increase their symbolic performance by being able to “tick more boxes”, thereby increasing the measured volume of their reporting framework. However, these authors note that such greater measured volume may not necessarily indicate managers’ intention with regards to social and environmental disclosure issues [65]. Nevertheless, supporters of social reporting and the GRI have long contended that disclosure offers reporting companies a wide spectrum of intangible benefits, such as employee loyalty and consumer reputation. Also, new research suggests that the value of CSR disclosure extends to the firm’s balance sheet [66,67]. An analysis of the results of more than 200 independent empirical studies examining the relations of corporate social and environmental performance to corporate financial performance suggested that companies might benefit from increased communication of their good deeds [68]. The studies in the sample specifically covering transparency and reporting indicated positive market reactions to sustainability reporting.

2.3. The Questionable CSR-Firm Performance Relationship

The debates on the relations between CSR and firm performance (herein termed as FP) have been part of the academic literature for decades, and therefore are not new [69]. Reviews of the some relevant studies on this questionable relationship have shown both positive, negative, and to some extent no relations at all [53,54,70,71,72,73]. Such inconclusive findings create a research gap for further examinations in an attempt at reaching conclusive results. The literature has attributed the differences in results to research design adopted which according to scholars may likely introduce error in specifying the models or issues of endogeneity. Usually, competition between the available and research adopted has also led to results that are incompatible and has helped perpetuate conceptual inconsistencies and contradictory arguments with the CSR and FP nexus [53,54]. The existing literature indicates that researches investigating the relationship among variables are associated with the endogeneity problem. This present study claims that causality and hence the problem of endogeneity exists. Therefore, we briefly explore the potential problem of endogeneity. Prior studies have examined the potential effects of endogeneity and their remedies [74]. Li [74] provides evidence by examining a variety of econometric methods that researchers commonly use to address endogeneity in empirical corporate finance. According to Li [74] the problem of endogeneity is one single biggest issue in the research line exploring the relations among variables of interest in the corporate finance literature, which is the main reason why results are largely mixed and inconclusive. The author asserts that the endogeneity problem has always been the main or possibly only obstacle to understanding the true relations between different aspects of constructs in empirical corporate finance. He observed further that variables are typically endogenous, instruments are scarce, and causality relations are complicated. Even so, researchers either use one or two simple methods to mitigate the issues of endogeneity or simply ignore it totally. Li [74] illustrates this by experimenting with a variety of econometric methods on how to mitigate the endogeneity problem. The author shows that all the prevailing econometric methods’ remedies are generally effective in mitigating the potential problem of endogeneity to some degree and to correct the sign from positive to negative. However, the effects vary considerably in quantitative terms. Among the remedies identified, Li [74] finds the generalised moments model (GMM) to having the greatest correction effect on the bias, followed by instrumental variables, fixed effect models, lagged dependent variables, and the addition of many more control variables. Even so, Li [74] argues that a combination of the methods, like firm fixed effects, year fixed effects, and the addition of more meaningful control variables appear to work well, even without a valid instrumental variable. As we will see in subsequent sections of this paper, we deal with this issue by employing several methods in an attempt to mitigate the potential effects of endogeneity in this present study.

Literature has further attributed the differences in results to the non-availability of a uniform definition on CSR. Thus, different scholars and researchers have used numerous different CSR definitions between different studies and settings to comprehend their research’s idea. Also, researchers contend that CSR disclosure and performance may vary across different institutional and industrial setting, because of the various costs and benefits associated with diverse institutional and industry characteristics [75,76]. For instance, it has been shown that industry sensitivity to the environment will affect disclosure responsibility of firms [77]. The use of different definitions for CSR by researchers, between different studies and settings, has been done to comprehend the researchers’ idea clearly, although there seems to be substantial common ground between them [78]. Therefore, this present study follows an acceptable concept of sustainable development as coined by the Brundtland Report, which claims for a development that “meets the needs of the present without compromising the ability of future generations to meet their own needs” [79] (p. 8). The report defines corporate sustainability under three umbrellas: economic attainment, social justice, and environmental stability [79]. As a result, CSR has now been viewed to achieve balance between the three dimensions; on an economic benefit level—businesses should be profitable and supply the society with products and services; on the ecological level—businesses must protect the environment and find new ways leading to eco-efficiency; and on the social level—businesses should safeguard human rights and promote social integration [80]. Werther and Chandler [81] state that the concept of CSR is about integrating the triple bottom line of social, economic, and environmental dimensions in a multi-stakeholder dialogue on a voluntary basis.

2.4. Prior Research on CSR Disclosure Performance and Firm Value

To date, the existing empirical studies on the CSR–FP relations in both advanced and emerging economies report mixed results. In contrast, it has also been found that CSR reputation index positively affects firms’ financial performance [82,83]. Furthermore, Orlitzky et al. [54] conducted a meta-analysis of 52 studies and showed that social responsibility disclosure and, to a lesser extent, environmental responsibility disclosure is likely to have a positive effect on firm performance. Theoretically, the formation and evolution of the Stakeholder Theory [84], and its instrumental aspects in particular [85], has provided the foundation for proponents of CSR to theorize that CSR can help in constructing and concretizing relations of trust with a variety of stakeholders (i.e., employees, consumers, local communities, environmental activists, and concerned citizens among many others) that are important to the long term success and value of a firm. Several researchers have for that matter documented a positive association between CSR–FP relationships [83,86]. On the other hand, there are opponents to this notion. Historically, Milton Friedman [87], an eminent economics professor, has been one of the most outspoken opponents of CSR. In his famous article entitled ‘The social responsibility of business is to increase its profits’, published in 1970, he argued that the application of CSR imposes unjustifiable and fundamentally undemocratic tax on shareholders, that its implementation cost outweighs any potential tangible benefits, and as such, it constitutes misappropriation and allocation of valuable resources of the firm [87]. These two opposing views (including those of other scholars that fall within the spectrum as defined), have led to the empirical investigations from different angles regarding the extent to which CSR can lead to demonstrably superior firm performance or not, by academic researchers. Therefore, with the recent increase in global attention on social and environmental issues toward sustainable development, this has led to an equal rise in global standardisation on social and environmental disclosures, towards the creation of a shared future. The rise in global standardisation has further provided academia with an even greater avenue to investigate deeper into the alleged relations between CSR disclosure performance and firm performance. Researchers have noted that the term “association” as used in the CSR and FP relations is a broad term, which is simultaneously used to represent the influence of independent variables on dependent variables, as well as, the relationships among them [69,88]. Therefore, the past several years witnessed scholars and researchers making use of several conceptual and theoretical foundations to develop and test hypotheses regarding the nature of the CSR–FP relations. Hence, results on the CSR–FP relations have been mixed; depending on theoretical orientations.

In support of a neutral CSR–FP relation, McWilliams and Siegel [89] argue that in equilibrium there should be no significant CSR–FP relations. Using the supply and demand theory of the firm, these authors [89] contend that managers’ main objective is to maximize shareholder wealth. Accordingly, managers will choose the level of the activities (including CSR activities) that maximize firm value, given the demand for various characteristics and the costs of supplying them. Thus, in equilibrium, there is a neutral CSR–FP relation. Linked to this neutral relation, is the argument that the environments in which societies and organisations exist are too complex that a simple and direct CSR–FP relationship seems questionable [82]. One of the bases for predicting a negative CSR–FP relationship derives from the neoclassical theory of the firm. This theoretical view posits that the opportunity cost of expenditures for corporate responsibility disclosure performance exceeds the profitability of such investment, so that a trade-off exists between CSR–FP relations. Therefore, where stakeholders exert effective pressure for CSR performance, the outcome is an observed diminished firm financial performance and firm value. Scholars [90] argue that, in spite of the trade-off between CSR and FP, managers may undertake CSR activities for their own benefit (e.g., public fame) at the expense of shareholders value maximisation. These scholars [90] argued that managers may seek to divert stakeholders’ attention of the poor firm financial performance by promoting the firms’ CSR activities. In contrast, managers may reduce expenditures on CSR programs in order to boost short-term profitability to enhance their personal compensation. Either way, CSR may signal to investors that managers are liable to acting in their own benefit. In this way, it is expected that stock prices will fall as investors come in the known that managers of socially responsible firms act in ways detrimental to shareholder interest [69].

Although researchers have observed mixed results in the literature regarding the CSR–FP relationship, prior evidence largely points towards a positive relationship in majority of the studies conducted in the developed countries context. However, and relating to developing economies, the evidence is still mixed and inconclusive. Some researchers found CSR to be positively and significantly associated with firm value/performance, but a large number of studies revealed an insignificant association. The difference in results, as noted earlier, may be due to differences in methodology adopted, study period examined and the wide variation in variables used in measuring CSR, the different institutional setting, and to some extent the operational constructs [91], and importantly the problem of endogeneity [74]. According to Carrots and Sticks [92], there exists a positive theoretical relationship between CSR disclosure and firm value/performance. Therefore, from the evidence above, we formulate the following first alternative hypothesis for empirical testing in the context of South Africa.

Hypothesis 1 (H1).

Ceteris paribus, there is a positive and significant relationship between the extent of CSR disclosure performance and firm value of selected JSE companies.

Regarding the relationship among the components of CSR disclosure performance (i.e., environmental disclosure performance and social disclosure performance) and firm value/performance, the results are also mixed and inconclusive. Although general and theoretical notions point towards a positive relationship, other researchers have found a negative relationship and, in some cases, no relationship at all. In all the different findings, researchers have provided theoretical basis to justify their results. In this present study therefore, we propose that the components of CSR disclosure performance (i.e., social and environmental disclosure performance) will have a positive relationship with firm value/performance, and hence, based on the above theoretical foundations and prior evidence, the following second and third alternative hypotheses are formulated for testing in the context of South Africa:

Hypothesis 2 (H2).

Ceteris paribus, there is a positive and significant relationship between the extent of environmental disclosure performance and firm value of selected JSE companies.

Hypothesis 3 (H3).

Ceteris paribus, there is a positive and significant relationship between the extent of social disclosure performance and firm value of selected JSE companies.

3. Data, Research Design and Methodology

3.1. Data: Sample Selection, Source and Description

The present study focus on companies listed on the JSE. Data were hand collected from annual reports, which were obtained from the African Financial Market database. In addition, we used CSR and sustainability reports which were collected from websites of sampled companies, for the period of 2010 to 2015. Consistent with prior disclosure studies, our final sample of 126 listed companies, is made up of only nonfinancial companies. In arriving at our final sample, we excluded companies with two or more missing years’ annual and CSR/sustainability reports. Our sample is made up of data over a six-year period for an unbalanced panel data analysis to explore the relationship between the extent of CSR disclosure performance and firm value. Therefore, our final sample of 126 listed companies, over a 6-year period which produced a total of 747 observations, within seven industry sectors. Researchers have noted that sample size is an essential piece for any empirical study. Thus, the sample size is supposed to represent a bigger population that would be too hard to measure. Where the sample is too small it may lose its significance, which means that the results may not be extendable to a larger population since the chosen sample was not largely representative [93]. Hence, and to the best of our knowledge, our final sample size of 126 listed companies over a six-year period, for this present study, provided a larger observation compared with prior CSR disclosure research work [29,31] in the context of South Africa.

3.2. Measurement of Variables and Model Specification

Our main independent variable is CSR disclosure performance. However, for this present study, we construct three disclosure indexes (i.e., CSR disclosure performance index, social disclosure performance index, and environmental disclosure performance index). In constructing our disclosure indexes, we manually extracted the CSR information disclosed (i.e., social and environmental disclosure information) from the annual reports and CSR/sustainability reports of sampled companies. We then analysed the CSR information collected, in line with G3.1 guidelines, to ascertain the extent of CSR disclosure performance. Thus, we used the G3.1 guidelines of 75 comprehensive provisions (i.e., measured on the basis of environmental and social performance indicators) as a measure of extent of disclosure performance. Binary coding using content analysis technique was employed for this case. The CSR disclosure performance index is developed based on 75 provisions of G3.1; the social disclosure performance index is developed based on 45 provisions of G3.1 and the environment disclosure performance index is developed based on 30 provisions of G3.1 guidelines. Thus, the CSR disclosure performance index is a combination of the social and environment performance indicators. Supplementary Material shows details of the G3.1 guidelines as provided by the GRI. Hence, consistent with prior disclosure research [36,94], we award a company 1 if a specific CSR item disclosed is in line with the G3.1 guidelines (i.e., either full or partially disclosed); otherwise 0. However, and consistent with prior research [95] where a performance indicator is found not to be applicable to a particular industry, we do not penalise nor award a score. Thus, using binary coding system of content analysis, we score the sampled companies based on their CSR information disclosed, measured against the G3.1 disclosure performance indicators, over the period of observation. Prior research has extensively used content analysis in disclosure research. Hence, this study used binary coding system of content analysis.

This present study employs the GRI (G3.1) guidelines because of its comprehensiveness in support of sustainability development and CSR disclosure from an international perspective. Also, it falls within our study period: 2010 to 2015. Our study is further guided by provisions of the South Africa Corporate Governance Codes, in particular, the King III Report, which encourages all entities in South Africa (including listed companies) to incorporate into their operations sustainability issues in a triple bottom context, and to report based on the GRI. The study period also falls within the period of the King III Report, which came into force from 2010 to 2016, following the inception of the King IV Report. We wish to state emphatically that in constructing the CSR disclosure index for this present study, we do not identify companies that issued GRI-guided reports, as have been used by some prior researchers [96,97,98]. Rather, and unique to this study, we manually review (measure) CSR disclosed information collected from the sampled companies’ annual reports and CSR/sustainability reports against the G3.1 guidelines, to ascertain the extent of disclosure with this guideline. For that matter, the authors thoroughly read all annual and CSR/sustainability reports twice to ensure consistency in the scoring.

Tobin’s Q is used as a measure of firm value and the dependent variable. This is because it has widely been used in prior research [99,100] as a measure of long term performance. Scholars [101] have noted that firm value incorporates both a firm’s future cash flow and the assessed inherent risk reflects in the cost of capital. Thus, firm value highlights the incremental portion of the market price that exceeds the book value. Therefore, Tobin’s Q is most suitable for the measure of firm value; even so, this study seeks to examine the influence of CSR disclosure performance on long-term firm value, which makes the Tobin’s Q a preferred indicator. Other financial performance variables like return on assets (ROA) and return on equity (ROE) are relatively unsuitable in this case because they are short-term measures of firm performance [102,103]. Also, unlike Tobin’s Q, other financial measures of firm value/performance (like ROA and ROE) could be easily subjected to several short-term earnings manipulation activities.

Researchers have noted that several variables could influence both independent and dependent variables. It is therefore not unusual to see important independent variables omitted in regressions for several reasons, such as when variables are immeasurable, the presence of an in-exhaustive list of potential independent variables, when the relationship between the omitted variables and variables that are included cannot be measured, and when the effect of the omitted variables are unknown [104]. Thus, omitted variable bias can cause regression estimates to be inconsistent [105]. Traditionally, omitted variables are dealt with by using either proxies [104]. Therefore, and consistent with prior studies, this present study employs several control variables to minimise potentially omitted variable bias. The following control variables are included in this study, board independence, presence of an external audit firm, firm size, return on asset (representing profitability), firm age, and leverage. For instance, Satish and Thompson [106] included in their study seven control variables. This is because previous studies have indicated that some variables may have the potential to influence both the dependent and independent variables. Therefore, and as argued, the control variables are adopted to account for factors other than the theoretical constructs of interest that could explain variations in the independent and dependent variables.

Empirical research in corporate finance considers firm size an important and fundamental firm characteristic, since they are usually the key variables that affect the independent and dependent variables simultaneously. Hence, the literature asserts that firm size is closely related to CSR and firm performance. Therefore, and importantly, the study by Dang et al. [107] provides empirical evidence for a “measurement effect” in the “size effect”. Specifically, these authors examine the influences of employing different proxies of firm size (i.e., total assets, total sales, and market capitalization) in areas of empirical corporate finance research. Dang et al. [107] highlights numerous empirical implications: first, the coefficients of firm size measures are robust in sign and statistical significance. Second, the coefficients on regressors other than firm size often change sign and significance when different size measures are employed. Third, the goodness-of-fit measured by R-squared also varies with different size measures, suggesting that some measures are more relevant than others in different situations. Last, different proxies capture different aspects of “firm size”, and thus have different implications. Therefore, the choice of size measures needs both theoretical and empirical justification [107]. Kumar et al. [108] provides a comprehensive review of literature and classified the theories into four categories, these are, technological theories, organizational theories, regulatory theories and financial theories (as cited in Dang et al. [107]). Existing literature therefore show that “firm size” is a most significant variable in most studies in this area. Therefore consistent with prior studies, we use total assets as our choice of firm size. We do this in line with the suggestion by Hart and Oulton [109] that the choice of a measure used depends on data availability (we do so not out of convenience), and also it is in line with the purpose of this study. However, to mitigate for the high skewness associated with the firm size data, we employ the natural logarithm of the total assets value. All variables are listed and defined in

Table 1. Definition of variables.

Dependent Variable
Q	Total market value of the firm to total asset value
Independent Variables
Env_DISC	Consisting of 30 disclosure performance provisions of GRI, G3.1
Soc_DISC	Consisting of 45 disclosure performance provisions of GRI, G3.1
CSR_DISC	Consisting of 75 disclosure performance provisions of GRI, G3.1
Control Variables
FSIZE	Natural logarithm of year-end of total assets
BIG4AUD	A dummy variable equal to ‘1’ if a firm is audited by a Big Four audit firm, otherwise ‘0’
FAGE	Number of years from a firm’s founding year to 2010, start year for the study period
BODIND	Number of independent executives and nonexecutive directors on corporate boards
ROA	Ratio of operating profit to total assets
LEV	Ratio between the volume of a firm’s short and long term debt to its total assets

Note: Q = Tobin’s Q; Env_DISC = environmental disclosure performance index; Soc_DISC = social disclosure performance index; CSR_DISC = CSR disclosure performance index; FSIZE = Firm size; BIG4AUD = an external big four audit and assurance company i.e., PricewaterhouseCoopers, Deloitte & Touché, Ernst & Young, and KPMG; FAGE = Firm age; BODIND = board independence; ROA = Return on Asset; LEV = leverage.

below. Consistent with prior disclosure research [91] that employed disclosure index, this study adopts the disclosure formula below to construct the three separate disclosure performance indexes.

$$DIS{C}_{ij}=\frac{{\displaystyle \sum _{i=0}^n{X}_{ij}}}{m_j}$$

(1)

where m_j = the maximum expected score for each category, j is the company, i = the items and X_ij assumes a value of 1 if a company disclosed an item, otherwise 0. However, we do not either penalize or award a company where an indicator is not applicable to a particular industry.

The issue of reliability and the approval of the quality of content analysis adopted for a particular study are very important. As such, prior studies [110] have observed different ways to improve the quality of content analysis. The process involves initial identification, selection, and definition of the disclosure items from related literature which is of utmost importance. Secondly, it is necessary to establish a reliable coding system with proper decision steps. Finally, the adoption of a uniform coding system is highly essential for a reliable content analysis [111]. Sustainability reporting on the basis of GRI guidelines for disclosure of CSR information has been noted as very reliable for analysing firms’ sustainable disclosure performances [7]. Therefore, in order to measure CSR disclosure performance, the G3.1 guidelines which provides a total of 75 well-defined specified performance indicators for all the components of CSR disclosure (i.e., environmental and social disclosure performance) was used, as a base to measure the CSR information disclosure collected from annual and CSR/sustainability reports. Moreover, it has been reported that, according to the GRI, only the validated contents are disclosed. Thus, the present study uses GRI-guided disclosure performance indicators as a unit of measure along with a uniform coding system for enhancing the reliability [91].

3.3. Model Specification

Consistent with prior research [112], and assuming that all relations are linear, the following regression model is employed to test the stated hypotheses and to establish the relationship among the dependent and independent variables.

$$Q_{it}=a_{it}+DIS{C}_{it}+{\displaystyle \sum _{i=1}^n{\beta }_{1}Control{s}_{it}+{\epsilon }_{it}}$$

(2)

where Q refers to Tobin’s Q; α refers to constant term; DISC refers to CSR_DISC, Soc_DISC, and Env_DISC indexes and; CONTROLS refers to the control variables of firm size (FSIZE), firm age (FAGE), presence of an external audit firm (BIG4AUD), board independence (BODIND), profitability (ROA), leverage (LEV), and Industry fixed effects.

4. Empirical Results and Discussion

4.1. Descriptive Statistics

Table 2. Summary of descriptive statistics.

Variable	Obs	Mean	Std. Dev.	Skewness	Kurtosis
Q	747	2.35	4.39	4.19	23.15
Env_DISC	747	0.61	0.16	−0.12	2.52
Soc_DISC	747	0.66	0.09	0.56	4.07
CSR_DISC	747	0.65	0.10	0.52	4.98
FSIZE	747	22.08	2.52	−0.23	3.49
BIG4AUD	747	0.76	0.43	−1.23	2.51
FAGE	747	36.76	29.64	0.91	2.94
BODIND	747	7.06	2.97	0.08	2.59
ROA	747	0.46	7.27	23.15	577.62
LEV	747	0.52	2.10	26.72	724.82

Authors construct from STATA. Note: All variables are as described and defined in Table 1.

, shows the descriptive statistics for all variables relating to mean, standard deviation, skewness, and kurtosis included in this present study. Here, our aim is to get an idea about how the variables are distributed, to examine the inherent interrelationships among them, and to know if there exists the presence of outliers in the data set.

Table 2, provides a summary of all study variables used. The table shows that Q has a mean of 2.35 and a standard deviation of 4.39. The mean score of Q suggests that on average the sampled companies have high market value/performance over the study period. However, the standard deviation for Q shows wider variations within the firm value of the sampled companies. On the other hand, the table also shows that the mean and standard deviations of our main independent variables respectively are: Env_DISC is 0.61 and 0.16; Soc_DISC is 0.66 and 0.09; and CSR_DISC is 0.65 and 0.10. These indicate that there exist a high degree of similarity among them, and that the differences among them are not very wide. The mean score for our main independent variables all indicate an average of 61% and above; this gives an indication that our sampled listed companies, for the period under observation, performed moderately high for CSR disclosure performance, as measured on the basis of the G3.1 guidelines for disclosure of CSR information. Therefore, this suggests that the extent of incorporation of sustainability among ompanies on the JSE is high. However, the presence of the wide variations within the scores suggests that there exist a high degree of heterogeneity among sampled companies. Additionally, the observed values of Skewness for some of the variables and the near equality of the scores of our main independent variables indicate that the distribution is not far away from symmetry.

Table 3. Correlation matrix.

	1	2	3	4	5	6	7	8	9	10
1.Q	1.000
2.Env_DISC	−0.116	1.000
3.Soc_DISC	−0.084	0.279	1.000
4.CSR_DISC	−0.126	0.837	0.758	1.000
5.FSIZE	−0.246	0.131	0.230	0.220	1.000
6.BIG4AUD	−0.101	0.090	0.114	0.126	0.278	1.000
7.FAGE	−0.116	0.040	0.188	0.134	0.228	0.200	1.000
8.BODIND	−0.127	0.041	0.137	0.106	0.542	0.336	0.257	1.000
9.ROA	−0.001	−0.014	0.017	−0.000	0.030	0.040	0.059	0.029	1.000
10.LEV	−0.018	−0.029	−0.039	−0.042	0.007	−0.063	−0.044	−0.056	−0.039	1.000

Authors construct from STATA. Note: All variables are as described and defined in Table 1.

below presents the correlation matrix among all the variables in this study. The correlation matrix provides understanding of the degree of relations among all the variables. First, we look at the relations amongst our variables of interests, the dependent and the independent variables. It shows weak negative relations between Q and Env_DISC of −0.12, Q and Soc_DISC of −0.08, and Q and CSR_DISC of −0.13, respectively. These results imply that CSR disclosures do not necessary lead to increased firm value. Next, we look at the relations between the dependent variable Q and the control variables. It further shows weak negative relations between Q and all the other variables as follows. Q and FSIZE at −0.25, Q and BIG4AUD at −0.10, Q and FAGE at −0.12, Q and BODIND at −0.13, Q and ROA at −0.001, and Q and LEV at −0.02. Relations among independent variables and control variables: FSIZE and Env_DISC is 0.13, FSIZE and Soc_DISC is 0.23, and FSIZE and Over_DISC is 0.22, all indicate a weak positive relation. FAGE and Env_DISC is 0.04, FAGE and Soc_DISC is 0.19, and FAGE and CSR_DISC is 0.13; all indicate a weak positive relation. From Table 3, we find some high correlation among the variables. These high figures may raise concerns of multi-collinearity [113] within the correlation matrix. Scholars [114] recommend ways of addressing multi-collinearity. First, regression techniques that use biased regressors coefficients can be used to achieve a substantial reduction in the stability and variance of these same coefficients. Second, the theoretical model can be altered either by combining the collinear independent variables or by including them in different regression models. Consistent with prior suggestion [114] that a safer approach to addressing multi-collinearity is to include the suspected variables in alternative regression models, hence, we adopt this approach. Therefore, we are of the view that the high correlation among variables in this study is not an issue. Hence, we run two separate regression equation models for this present study. Regression equation Model 1 is used to investigate the relationship between CSR disclosure performance and firm value. Also, regression equation Model 2 is used to investigate the relationship among social disclosure performance and environmental disclosure performance and firm value. In each of the regression models, we include all the control variables to minimise potentially omitted variable bias.

4.2. Panel Data Regression

In this study, both cross-sectional and time series components are involved. Hence, given the nature of data and to allow for time series and cross-sectional data observations, we employed the panel data technique. Prior studies in the context of South Africa have utilized both balanced panel [31] and unbalanced panel [36] data techniques. Therefore, consistent with these authors and based on data collected, this study employs an unbalanced panel data technique. Many different reasons have been associated for the use of unbalanced panels. Most importantly, and applicable to this study, in the real world most panels are unbalanced [115]. Therefore, to choose a balanced panel may introduce sample selection bias into the dataset, which may result in less representation of the population. Consistent with prior studies [116], this study employs two-panel data techniques, which are the fixed effects and the random effects models, to control for possible unobserved firm-level heterogeneities. However, before deciding on an appropriate panel estimation technique, we conduct the Breusch and Pagan test (test results not reported in this study) to determine the suitability of either a pooled OLS or random effects regression model. The test rejects the null hypothesis of zero variance across entities, which makes pooled OLS unsuitable. Afterwards, we conduct the Hausman test to identify if the individual effects are fixed or random, in cases where individual effects exist. Thus, we used the Hausman test to decide on the suitability of either the fixed effects or random-effects model. The result of the Hausman test rejects the null hypothesis that individual effects are uncorrelated with the regressors, for regression Models 1 and 2. Hence the fixed effects panel regression model is used for this present study to investigate the relationships among the variables.

To investigate the relationship among CSR disclosure performance (and its components of Social disclosure performance and environmental disclosure performance) and firm value, we ran the fixed effects model.

Table 4. Fixed Effects Model 1 (with CSR disclosure performance).

Q	Coef.	Std. Err.	t	p > |t|	[95% Conf. Interval]
CSR_DISC	0.313	1.130	0.28	0.782	−1.906	2.532
FSIZE	−1.232	0.227	−5.44	0.000	−1.677	−0.787
FAGE	0.234	0.044	5.34	0.000	0.148	0.320
BIG4AUD	0 (omitted)
BODIND	0.169	0.048	3.52	0.000	0.075	0.263
ROA	0.004	0.009	0.39	0.697	−0.015	0.022
LEV	0.012	0.035	0.35	0.729	−0.056	0.080
_cons	19.563	4.311	4.54	0.000	11.097	28.028
Industry-effects	YES
Sigma_u	8.165
Sigma_e	1.690
Rho	0.959 (fraction of variance due to u_i)
F test that all u_i=0:	F (125, 615) = 32.59			Prob > F = 0.0000

Note: sigma_u = SD of residuals within groups ui; sigma_e = SD of residuals (overall error term) ei; Rho = 95.9% of the variance is due to differences across panels; ‘rho’ is known as the intra-class correlation; t = t-values test the hypothesis that each coefficient is different from 0. To reject this, the t-value has to be higher than 1.96 (for a 95% confidence). If this is the case then you can say that the variable has a significant influence on your dependent variable (y). The higher the t-value the higher the relevance of the variable; p > |t|=two-tail p-values test the hypothesis that each coefficient is different from 0. To accept this, the p-value has to be ≤0.001, ≤0.05, and ≤0.10 (at 95% confidence level); if this is the case then you can say that the variable has a significant influence on your dependent variable (y). Source: Authors’ result of fixed effects model from STATA.

and

Table 5. Inputs for the Fixed Effects Model.

Fixed Effects (within) Regression	Number of Obs = 747
Group variable: Comid	Number of groups = 126
R2:	Obs per group:
Within = 0.0754	min = 5
Between = 0.0002	avg = 5.9
Overall = 0.0002	max = 6
	F (6, 615) = 8.35
Corr (u_i, Xb) = −0.8623	Prob > F = 0.000

below present the results of the panel regression using the fixed effects for Regression Model 1. The relevant statistics to take note of in the tables below are the coefficient (β) of the regressors and the p-values. The significance level is set to 95% significance, with p-values at 1, 5, and 10% interpreted to be statistically significant.

4.2.1. Fixed Effect Model: CSR Disclosure Performance and Firm Value

The results in Table 4 above shows that our first alternative hypothesis (H1) for this present study is not supported. This is represented by coefficient (β) of 0.313 and p-value of 0.782. It implies that CSR disclosure performance has a positive, but an insignificant relationship on firm value. Our results suggest that CSR disclosure has a limited effect on firm value. Even so, and with regards to the correlation matrix (Table 3), it shows a weak negative relationship between CSR disclosure performance and firm value of −0.13, which also suggests a weak negative relation between CSR disclosures and firm value. This indicates that CSR disclosure may not necessary lead to increasing firm value. Our findings go to support the line of argument by some researchers who assert that CSR disclosures are becoming largely regulated and routine and thus lead to providing narrow incremental information beyond other companies’ CSR disclosure [29]. Our findings is consistent with Horn et al. [29] research study which found a positive but no significant association between CSR disclosure (as measured by KPMG’s 2008, 2011, and 2013 survey data) and firm value (proxy by Tobin’s Q) in the context of South Africa. The result of this present study shows that there exist a direct relationship between the extent of CSR disclosure performance and firm value/performance of sampled companies. However, the existence of this direct relationship was insignificant; implying that this positive link may not necessarily lead to significant improvement in firm value/performance. This may be the outcome of the associated cost that may be related to the performance of CSR strategies by firms, which may eventually mitigate the overall returns that may be expected from CSR performance. Our result suggests that CSR disclosure performance has the potential to link firm’s future value, which is relevant for stakeholder decision making.

Regarding the influence of the control variables on the firm value in regression Model 1, some interesting results are found. The results show a negative, but significant relationship between firm size and firm value (i.e., β of −1.23; p-value of 0.000 at 1% significance). Prior researchers assert that firm size is a usual predictor of firm value/performance, which is associated with higher costs of monitoring as larger firms are more complex and have more arms’ length transactions [117]. Thus, our findings are consistent with prior studies that found a negative association between firm size and firm value/firm performance [118,119]. In the regression results for Model 1, the presence of an external audit firm (i.e., BIG4AUD) is omitted as a result of collinearity, because its value was the same over the period of our analysis. Thus, because there is no variability in its value, this variable was treated as a constant, and that fixed effect transformation eliminate such variables. That is why there were no coefficient values for this variable. The results further show a positive and significant relationship between firm age and firm value (i.e., β of 0.23; p-value of 0.000 at 1% significance). This is consistent with prior research [123] which found that firm age has a significant positive impact on the effects of CSR to firm performance. The results also show a positive and significant relationship between board independence and firm value, (i.e., β of 0.17; p-value of 0.000 at 1% significance). This result suggests that the more independent and nonexecutive directors are added to the board will lead to improve firm value. It also shows a positive, but insignificant relationship between return on asset (profit) and firm value (i.e., β of 0.003; p-value of 0.697). Similarly, our result shows a positive, but insignificant correlation between leverage and firm value (i.e., β of 0.012; p-value of 0.729). The result suggests that leverage although is positively related with firm value—it is not significant. This result is consistent with the mixed available evidence on leverage [112].

Table 6. Inputs for the Fixed Effects Model 2.

Fixed Effects (within) Regression	Number of Obs = 747
Group variable: Comid	Number of groups = 126
R2:	Obs per group:
Within = 0.0796	min = 5
Between = 0.0001	avg = 5.9
Overall = 0.0001	max = 6
	F (7, 614) = 7.59
Corr (u_i, Xb) = −0.8605	Prob > F = 0.0000

and

Table 7. Fixed Effects Model 2 (with Social and Environmental Disclosures).

Q	Coef.	Std. Err.	t	p > |t|	[95% Conf. Interval]
Env_DISC	−1.545	1.093	−1.41	0.158	−3.691	0.600
Soc_DISC	2.409	1.485	1.62	0.104	−0.506	5.325
FSIZE	−1.270	0.227	−5.58	0.000	−1.717	−0.823
BIG4AUD	0 (omitted)
FAGE	0.231	0.044	5.29	0.000	0.145	0.317
BODIND	0.165	0.048	3.44	0.001	0.071	0.259
ROA	0.004	0.009	0.39	0.697	−0.015	0.022
LEV	0.014	0.034	0.41	0.684	−0.054	0.082
_cons	20.066	4.315	4.65	0.000	11.593	28.539
Industry effects	YES
Sigma_u	8.121
Sigma_e	1.688
Rho	0.959 (fraction of variance due to u_i)
F test that all u_i=0:	F (125, 614) = 32.62			Prob > F = 0.0000

provide results for regression model 2 for this study. Model 2 was employed to investigate the relationship among social disclosure performance and environmental disclosure performance, and firm value. In Model 2, the two components along with the control variables are considered as explanatory variables. We checked for multicollinearity problem among these variables through correlation matrix (see Table 3). The result does not indicate any serious collinearity among these variables. Therefore, the two components of CSR disclosure performance are included in one regression model. Therefore, in Table 6, the relevant statistics to take note of are the coefficient (β) of the regressors and p-values. The significance levels are set to 95% with p-values ≤ 0.001, ≤0.05, and ≤0.10 interpreted to be significant.

4.2.2. Fixed Effects Model: Environmental Disclosure Performance and Firm Value

In Table 7 the result shows that our second alternative hypothesis (H2) for this present study is also not supported. This is represented by β of −1.545 and p-value of 0.158. It implies that environmental disclosure performance has a negative and insignificant relationship with firm value. Even so, and with regards to the correlation matrix (Table 3), it shows a weak negative relationship between environmental disclosure performance and firm value of −0.12, which also suggests a weak negative relation between environmental disclosures and firm value. Our result is consistent with Verbeeten F.H.M. et al.’s [121] findings of a negative association between environmental disclosure and firm value in a German context. Also, Cormier & Magnan [122] found that environmental reporting does not significantly influence the stock market valuation of Canadian and French firms’ earnings. Again, Li et al. [123] found an insignificant relationship between environmental performance and financial performance, and a negative relationship between environmental disclosure and financial performance in a Chinese context. Our results imply that environmental issues could be relatively less prioritised among stakeholder groups compared with social issues in South Africa. This insignificantly negative relationship may be of particular interest to players in the corporate sector of sub-Saharan Africa.

4.2.3. Fixed Effects Model: Social Disclosure Performance and Firm Value

Finally, in Table 6 the result shows that our third alternative hypothesis (H3) for this present study is supported. This is represented by β of 2.41 and p-value of 0.104 (which is significant at 10%). This implies that social disclosure performance has a positive and significant relationship with firm value. This result is consistent with Menassa [124] finding of a positive and significant relationship between social disclosure and firm financial performance (proxy by return on asset and net profit margin). Also, Richardson and Welker [125] found that social disclosures had a positive impact on return on equity. Our result may support the notion that social issues are more relevant to larger stakeholder groups in developing countries, particularly sub-Sahara African, compared with environmental issues, notwithstanding the fact that environmental issues could all connote into social issues from different perspectives. This supports the general assertion that greater attention has been paid towards social issues in the developing countries context by both governmental and corporate institutions; this is evident from various social interventions pursued by various corporate and governmental agencies in such countries. Visser’s [126] study of CSR in Africa noted that CSR is mostly associated with philanthropy or charity, i.e., through corporate social investment in education, health, sports, and other community services. It further includes tackling HIV/AIDS, improving working conditions, provision of basic services, supply chain integrity, and poverty alleviation [126].

5. Robustness Test

In this section, we attempt to deal with the problem of endogeneity and to examine the robustness of our findings. In doing this, and consistent with prior studies that used two-stage least squares (2SLS) [127] and one-year lag [112], [127] approach, we use one-year lagged of the three disclosure performance indexes in the robustness test. Thus, we introduced one-year lag between disclosure variables and firm value, such that a firm’s performance in any year (Q_t) depends on the previous year’s CSR disclosure performance (DISC_t−1). Hart and Ahuja [128] assert that the benefits from CSR might only become fully visible over a long term, rather than a short term. We therefore argue that disclosure of CSR performance by firms may not have an immediate effect on firms’ value/performance, and that it may take time to reflect on firm value/performance. Therefore, we see a one-year lag in the three disclosure indexes.

Table 8. Two-stage least-squares regression (Model 1).

TOBINSQ	Coef.	Std.Err.	t	p > |t|	[95% Conf. Interval]
lag_CSR_DISC	0.369	2.096	0.18	0.860	−3.747	4.485
FSIZE	−0.423	0.077	−5.47	0.000	−0.576	−0.271
BIG4AUD	−0.033	0.443	−0.07	0.941	−0.903	0.837
FAGE	−0.008	0.006	−1.27	0.204	−0.021	0.004
ROA	−0.004	0.022	−0.19	0.849	−0.048	0.039
LEV	−0.026	0.076	−0.34	0.735	−0.176	0.124
_cons	11.337	1.992	5.69	0.000	7.424	15.249
Industry-effects	YES
Endogenous variables:	TOBINSQ
Exogenous variables:	lag_CSR_DISC FSIZE 0.BIG4AUD 1.BIG4AUD FAGE ROA LEV 1.Sector 2.Sector 3.Sector 4.Sector 5.Sector 6.Sector 7.Sector
Equation	Obs	Parms	RMSE	“R-sq”	F-Stat	P
TOBINSQ	621	12	4.342	0.127	7.38	0.000

and

Table 9. Two-stage least-squares regression (Model 2).

TOBINSQ	Coef.	Std.Err.	t	p > |t|	[95% Conf. Interval]
lag_Env_DISC	−4.121	2.217	−1.86	0.163	−8.474	0.232
lag_Soc_DISC	3.545	1.653	2.14	0.062	0.299	6.791
FSIZE	−0.412	0.077	−5.33	0.000	−0.564	−0.260
BIG4AUD	−0.009	0.441	−0.02	0.984	−0.875	0.858
FAGE	−0.006	0.007	−0.95	0.344	−0.019	0.007
ROA	−0.004	0.022	−0.18	0.861	−0.047	0.039
LEV	−0.017	0.076	−0.22	0.823	−0.167	0.133
_cons	11.44583	1.985	5.77	0.000	7.547	15.34
Endogenous variables:	TOBINSQ
Exogenous variables:	lag_Env_DISC lag_Soc_DISC FSIZE 0.BIG4AUD 1.BIG4AUD FAGE ROA LEV 1.Sector 2.Sector 3.Sector 4.Sector 5.Sector 6.Sector 7.Sector
Equation	Obs	Parms	RMSE	“R-sq”	F-Stat	P
TOBINSQ	621	13	4.326	0.135	7.30	0.000

show the robustness test results. Table 8 shows the CSR disclosure performance index for Model 1; Table 9 shows the environmental and social disclosure performance indexes for Model 2. In all models, we included the control variables to account for factors other than the theoretical constructs of interest that could explain variations in both independent and dependent variables. In Table 8 and Table 9, the outcome still remains consistent with the main results for this study.

In the robustness test (Table 8 and Table 9) our particular concern is with the main variables of interest for this study. That is, the relationships among the dependent and independent variables. The results as indicated in Table 8 above remain positive and insignificant for Model 1; represented by a β of 0.401 and a p-value of 0.609. This implies that there is a positive, but an insignificant relationship between CSR disclosure performance and firm value. On the other hand, Table 9, Model 2 shows that environmental disclosure performance has a negative and insignificant relationship with firm value; represented by a β of −4.121 and a p-value of 0.163. For the social disclosure performance it shows a positive and a significant relationship with firm value; represented by β of 3.55 and a p-value of 0.062. These results are consistent with our earlier findings for this study. It is particularly worth noting that firm size is the most significant control variable in this study, as its significance (at 1%) is seen in all the regression models, including our robustness test. This is consistent with scholars that have argued that the firm size remains the most significant variable in corporate finance research [107].

6. Summary and Conclusions

Numerous research studies examining the CSR–FP relationship from the perspective of international standardisation, have mainly concentrated on the developed and emerging markets of America, Asia-Pacific, and Europe. This leaves a conspicuous gap within the extant literature with regard to evidence from the sub-Saharan Africa. Nevertheless, some sub-Saharan African countries, particularly South Africa, offer interesting context with regards to investigating CSR disclosure from an international viewpoint. Compared with its peers, South Africa has a matured corporate sector, is ethnically diverse, and has strong regulatory and governance frameworks comparable to those of some developed and other established emerging markets. Therefore this study has attempted to fill the existing gap in the literature from a sub-Saharan African context. This present study has investigated the relationship between CSR disclosure performance (and its components) and firm value. The authors construct three separate disclosure indexes, on the basis of an internationally accepted standard for disclosure of CSR information, to re-examine the questionable CSR–FP relations. Thus, we have investigated the influence of CSR disclosure performance, social disclosure performance and environmental disclosure performance on firm value, in a single study from an emerging economy. We have explored these relations from a multi-theoretical perspective, utilising the stakeholder and legitimacy theories. Hence, this study provides new evidence from an emerging market context with unique institutional setting.

This study supports existing literature which provides evidence that CSR continuous to gain much prominence in South Africa, and for that matter sub-Saharan Africa. Even so, from the descriptive statistics (Table 2) we find that the mean score of CSR disclosure (and the components of CSR) is approximately 60% and above, which gives an indication that on average, the extent of CSR disclosure performance (as measured by the social and environmental disclosure performance indicators on the basis of the G3.1 guidelines) towards sustainability development is moderately high in South Africa. Our findings of a positive but insignificant relationship between CSR disclosure performance and firm value and a negative and an insignificant relationship between environmental disclosure performance and firm value, may suggest a weakening of the influence of CSR disclosure on firm value. From the legitimacy theoretical perspective, we may conclude that CSR disclosure performances and environmental disclosure are avenues by which organisations turn to legitimise their actions and inactions, notwithstanding its usefulness in explaining the motives behind CSR disclosures. From the stakeholder theoretical perspective, we may conclude that disclosures are ways of providing vital nonfinancial information to stakeholder about the firm, which could influence stakeholders’ perceptions about the firm. Even so, we are of the view that CSR disclosures, on the basis of international standardisation in developing countries, particularly Sub-Sahara Africa, are largely ways by which firms legitimise their actions and inactions towards powerful stakeholders. Our finding of a positive and statistically significant influence of social disclosure on firm value is evident of many efforts by governments in sub-Saharan Africa and corporate bodies’ provision of many social interventions to their communities. For instance, the post-apartheid era saw the formation of many social interventions by the ANC government towards national and corporate rebuilding. Therefore, many different organisations of different sizes and from a variety of different sectors are actively utilising social responsibility policies, not only because of the growing global trends and other external pressures, but also because it could result in efficiency gains. Porter and Kramer [129] indicate that addressing societal concerns amongst others could increase the levels of company productivity, with the subsequent positive effects in profitability and share value.

Therefore, based on our findings the following insights can be drawn. We argue that disclosure of CSR performance, on the basis of the GRI guidelines, serves to provide a common playing field to enable companies to be assessed from an internationally accepted standard. However, this could merely be ceremonial, mainly done for legitimacy purpose. Prior studies [130,131] suggest that companies use several legitimisation strategies to legitimise their actions and inactions. The process of legitimisation can involve real, material change in an organisation’s output, methods, goals and disclosure pattern, which in this case, are needed to inform relevant stakeholders about these changes [38, 49]. On the other hand, disclosures can be symbolic, which means that they are part of a legitimisation process whereby companies try to influence the perception of the public without changing performance [132]. By fulfilling this however, it allows disclosing companies to benchmark their performance against others, from a global standard. Benchmarking the disclosure of performance is beneficial in allowing companies to communicate to stakeholders the wellbeing of their companies’ operations, as well as, enable their self-assessment [133].

In the wake of global upsurge in international standards and its adoption on debates within the broader framework of corporate sustainable, and the social and environmental disclosure literature, this study makes significant contributions to the existing literature by pushing forward our understanding of the importance of expanding empirical evidence on the CSR–FP relations to different institutional settings that may allow for further findings in light of available inconclusive results within the CSR–FP nexus. We do so following international standards from an emerging institutional setting, where investor protection is relatively low, as a result of a weaker external corporate regulatory environment [12,13].

Although we recognise that our study is not without limitations, nevertheless, our results have important policy implications. It is well recognised that the GRI guidelines is an international and a multi-stakeholder effort to create a common platform for CSR disclosure in line with the economic, environmental, and social impact of organisations. Thus, it seeks to improve the comparability and credibility of sustainability disclosure worldwide. However, GRI guidelines may be far from reach due to the differences in the effects associated with other individual national/institutional settings, culture and social value aspects of markets in developed and developing countries. For instance, research has particularly observed with concerns current CSR approaches and with their origins in developed countries, may not reflect and fully respond to the developing countries’ context and circumstances encountered [134]. Further, Jones [135] observed that the national sociocultural environment and level of national economic development are important factors influencing disclosure of CSR and its practice. Thus, international standards for sustainability disclosure, like the GRI, may not fully reflect these observations from different country perspective, and as such rendering such standards relatively less applicable in emerging markets, like sub-Saharan Africa. Thus, we argue that the notion of uniformity for GRI disclosure in line with sustainability, across the length and breadth of organisations globally, may casts serious doubts on its effectiveness to provide a common platform to assess organisations from different institutional contexts. This may lead to impaired outcomes to provide a common platform to assess organisations with similar sectors from different geographical locations. Thus, we find the disclosure parameters lacking specific national and sociocultural value dimensions from different national context, of developed and developing countries. Therefore, we suggest that current regulation and practice in line with international standards (in particular the GRI) should be improved by way of incorporation of other national/institutional outlook and sociocultural value dimensions from the perspective of developed and developing countries to fully reflect the big picture. Therefore, we argue that it is only when these several country level indicators are incorporated that an equitable performance assessment among countries (and within sectorial industries) was fairly possible. Therefore, we assert that the adoption of GRI guidelines for CSR disclosure is inevitably influenced by various institutional locations, regulatory controls, and national sociocultural value dimensions.

Our study has the following limitations, which provide possible future research avenues. First, the process of constructing the disclosure indexes may be bias and potentially subject to measurement error. As such, our results should be interpreted with caution. Even so, we have manually collected the CSR disclosure information from the annual and CSR/sustainability reports of sampled companies and, as such, may result in some omission on our part, although to the best of our knowledge, all relevant information has been captured. Secondly, as we align the CSR information collected with the GRI G3.1 guidelines, to ascertain the extent of disclosure, there could arise elements of subjectivity on our part which could prejudice our results. However, in an attempt to minimise these biases, we acknowledge that every effort were made by the authors to remain objective in the scoring and coding process. As such, all information collected were analysed twice by the authors to ensure that all relevant information are captured and interpreted correctly. Potential research avenues from this paper could be directed at exploring the relationship among the various subcomponents of the individual components of CSR disclosure performance (i.e., environmental and social) and firm value, in order to ascertain further how these other sub-components influence firm value/performance. Also, future research could investigate this relationship within industrial sectors. The future research could also consider other firm value indicators (i.e., both accounting-based and market-based performance, such as ROE and market value) to investigate the relationship between CSR disclosure performance and firm value. Even so, and consistent with prior literature, scholars [69], [82], [136] assert that the CSR–FP relationship depends on various interactions between many different factors, in particular the specific capabilities of the firm and other industrial context. To this end, we postulate that the dynamism that lies within the CSR–FP relations (and its disclosure) research remains inconclusive, owing to the various indirect relations of several indirect variables that have been perceived to influence such relations.