Revising the King Code: does it matter for board diversity and corporate performance?

Over the last three decades, corporate governance (hereafter, CG) codes have proliferated globally (Cuomo and Zattoni, 2023). The effectiveness of such codes, particularly relating to corporate boards, has been extensively researched in both established and emerging economies (Aguilera et al., 2019; Farah et al., 2021), with Fauver et al. (2017) finding that board reforms in 41 countries between 1992 and 2006 increased firm value. These codes are typically revised in response to economic crises, corporate/market failures, political and legal reforms, climate change, and public concerns on sustainability and social inequality, and the revised versions emphasise aspects such as board and corporate culture, ethical conduct, corporate purpose, decision-making processes, and social and environmental imperatives, alongside traditional rationales of market and shareholder protection, and resource allocation (e.g. OECD, 2023). The extent to which such revisions lead to substantive change continues to attract research and policymaking interest.

Specifically, we engage with the literature on code revisions (Chai-Aun et al., 2016, 2021; Wang et al., 2020) by focusing on the implications of recent board diversity or heterogeneity exhortations. Board diversity refers to the variety of backgrounds reflected in a board’s membership regarding generally observable characteristics such as age, gender, sexual orientation, ethnicity, race, religion, disability, nationality, education, tenure and professional experience (Khatib et al., 2023). From a functional perspective, boards ought to be heterogeneous to ensure a mix of competencies and skills for multidimensional decision-making, to embed viewpoints from a wide range of stakeholders and to foster constructive dialogue, underpinning extensive research, synthesised in recent literature reviews (Baker et al., 2020; Khatib et al., 2023). In parallel, concerns about sexism, racism and other forms of discrimination in accessing opportunities at the executive and non-executive director (NED) levels (Guest, 2019; Nguyen et al., 2020; Ghio et al., 2024) have highlighted the inequity of board representation. The evidence so far has mainly considered the outcomes of board gender diversity, with aspects such as race, age and expertise being largely ignored (Ntim, 2015; Guest, 2019; Khatib et al., 2023; Ghio et al., 2024).

Furthermore, while positive outcomes may be observed from greater board gender diversity, there remains significant variation in these findings owing to institutional, cultural and other structural factors (Carter et al., 2010; Nguyen et al., 2020; Martínez-García et al., 2022). In the case of board ethnic diversity, Ntim (2015) and Nguyen and Muniandy (2021) found a positive association with company market values in South Africa, while Guest (2019) found no evidence of such an association in a large sample of US companies. As more developed and emerging economies contemplate various board diversity initiatives and/or code reforms (Martínez-García et al., 2022; Mateos de Cabo et al., 2022; Nalukenge et al., 2024), and with a general lack of contemporary evidence from emerging economies (Adams et al., 2015), it seems timely to further examine the implications of revised board diversity pronouncements.

Empirically, we consider the case of the Report on Corporate Governance for South Africa, known as the King IV code, which underwent its third revision in 2016. Notable changes include integrated thinking and reporting, stakeholder inclusivity, switching to an “apply and explain” approach, and increasing emphasis on board diversity. King IV recommends that boards establish and monitor their progress towards racial ^[1] and gender diversity targets, distinguishing this version from its predecessor (IoDSA, 2016). This further revision provides a setting for considering the outcomes of code revisions concerning board diversity. Therefore, we raise the following questions in the South African context: (1) to what extent has board diversity, notably regarding gender, race, occupation and age, changed? (2) What was the association (if any) between board diversity and corporate performance (financial and sustainability dimensions)? (3) Did the CG code revision (King IV) moderate the relationship between board diversity and corporate performance?

Our study was motivated from two perspectives. Firstly, previous studies in the South African context (Ntim and Soobaroyen, 2013; Jonty and Mokoaleli-Mokoteli, 2015; Taljaard et al., 2015; Mans-Kemp and Viviers, 2015; Hassan and Miller, 2017; Scholtz and Kieviet, 2018; Bissett, 2020; Abdelkader et al., 2024) have explored separate aspects of diversity – notably, gender, race, age and occupation – and found mixed results when assessing market, financial, environmental or social performance outcomes of board diversity. Distinctively, we consider both specific dimensions and a composite measure of board diversity alongside the increasingly persuasive requirements of King IV (notably on race and gender), which, to the best of our knowledge, have not yet been examined.

Secondly, we consider a context that has undergone various iterations of a CG code, aiming to further delve into the consequences of subsequent CG code revisions, an issue that has not been extensively studied. A rare cross-country study by Chai-Aun et al. (2016) found that a higher cumulative number of code revisions significantly improves the positive relationship between CG score and company value. However, these results did not hold when separately examining the effect of the number of new practices in CG code revisions in developing countries (Chai-Aun et al., 2016). Given the scant empirical evidence on the national consequences of revising CG codes, notably about board diversity and even more so in developing country contexts (Areneke et al., 2022), the King IV changes offer an appropriate setting to deepen our understanding of this question.

This study, therefore, examines how the diversity requirements of the King IV code translate into change at the board level and whether such changes are associated with better financial and sustainability performance. In common with prior studies (e.g. Mahadeo et al., 2012), we conceptualise board diversity primarily as a combination of various characteristics (gender, race, occupation and age). Informed by the resource dependence perspective, we develop expectations of how combined and individual dimensions of board diversity are positively associated with financial performance and sustainability (environmental and social) performance. The study follows a longitudinal design to analyse board diversity data for companies listed on the Johannesburg Stock Exchange (JSE) that were included in the top 100 companies in any of the five years from 2015 to 2019, resulting in a sample of 111 unique companies. Board diversity was analysed for two time periods: the pre-King IV period (2015 and 2016) and the post-King IV period (2018 and 2019). Integrated (or annual) reports were obtained from company websites, and board diversity data were hand-collected from these reports. Financial information was extracted from the IRESS database, while sustainability performance data were collected from the LSEG Workspace (formerly Thomson Reuters).

It is found that board gender and racial diversity increased significantly after King IV, although the increase can be deemed to be incremental rather than revolutionary. Moderating for the effect of King IV, our analysis using the price-to-book ratio reveals that the market valued board diversity more before the effective date of King IV. Increased diversity was associated with increased social performance and lower environmental performance after the effective date of King IV. Like Guest (2019), our results suggest limited benefits and raise questions as to the effectiveness of code revisions. The mixed results may also indicate that South African companies have yet to reap the benefits, given the long-term nature of the process by which diverse boards are constituted and can affect outcomes. Contextual factors may influence the ability of board diversity reforms (e.g. targets) to generate material economic benefits from a resource dependence perspective.

We contribute to the literature on board diversity and its association with corporate performance by considering various dimensions of diversity that encompass the complex facets of social identity (Ghio et al., 2024) and use a diversity index, building on the work of Ararat et al. (2015), to encourage researchers to focus more on combined measures of board diversity than on individual dimensions. Our evidence of the moderating effect of CG code revisions contributes to the limited literature on the implications of subsequent code revisions (Wang et al., 2020; Chai-Aun et al., 2021) and brings into question whether the board structure aspects of governance codes may have become saturated (Chai-Aun et al., 2016).

The remainder of the paper is structured as follows: Section 2 explains the contextual setting of King IV in South Africa, and Section 3 provides a review of prior literature and development of the research hypotheses. The methodology is discussed in Section 4, while the findings and discussion are presented in Section 5. Section 6 concludes the article.

While the board of directors is usually portrayed as a singular and monolithic entity, each member brings unique experiences, views, biases, social constraints and power relations to corporate decision-making (Ferreira, 2010). Board heterogeneity or diversity favours problem-solving, enhances leadership effectiveness, increases external connections and improves global relationships (Frias-Aceituno et al., 2012). It brings different knowledge, perspectives and ideas (Harjoto et al., 2015; Hillman, 2015), which could lead to better decision-making and more effective monitoring (Barako and Brown, 2008; Adams and Ferreira, 2009; Bear et al., 2010).

In response to evolving business practices, codes of CG are updated and often increasingly address the issue of board diversity (Rossouw et al., 2002; MacNeil and Esser, 2022; Khatib et al., 2023). South Africa is widely regarded as a pioneer in the field of CG, specifically with the publication of the first King Report on CG in 1994 (Mans-Kemp et al., 2016). Since then, the King Report has been revised three times (IoDSA, 2002, 2009, 2016). While King III suggested that a board consider whether its size, diversity and demographics contribute to its effectiveness, no mention was made of diversity targets (IoDSA, 2009). In contrast, King IV recognises that a well-balanced board of directors, comprising members from diverse identity groups, brings varied perspectives and approaches. It introduces a broad definition of diversity that includes fields of knowledge, skills and experience as well as age, culture, race and gender. Specifically, King IV recommends that the board set diversity targets for race and gender representation in its membership and disclose progress against these targets (IoDSA, 2016). King III acknowledged the importance of diversity but focused on the independence of NEDs, while King IV places greater emphasis on achieving diversity of the entire board. Prior to the issuance of King IV, the Broad-Based Black Economic Empowerment Act 53 of 2003 (BBBEE Act) was one of the most significant drivers of board diversity in South Africa (Deloitte, 2017).

Although the King Report is a set of voluntary principles and leading practices about CG (IoDSA, 2016), companies listed on the JSE must comply with its requirements (JSE Limited, 2017). King IV was released for public comment in March 2016 and the final version was issued in November 2016 for application for reporting periods beginning on or after 1 April 2017 (IoDSA, 2016). However, the JSE required earlier application for documents, such as annual reports, submitted on or after 1 October 2017. Companies were granted additional time until 1 June 2018 to implement the disclosure requirement of a policy promoting racial diversity at the board level (JSE Limited, 2017). This informed our sample period, which included two years prior to and two years after the implementation of King IV in 2017, to capture the influence of this CG code revision on the association between board diversity and corporate performance.

Racial diversity, as a subcomponent of ethnic diversity, is pertinent in South Africa, given the country’s history of institutionalised racial segregation (Nguyen and Muniandy, 2021). Even after Apartheid and the instauration of majority rule, there remains a nationwide reliance on racial categorisations (Mans-Kemp and Viviers, 2015). Reflecting this, the JSE listing requirements reiterate the importance of board-level diversity. These requirements mandate that listed companies have a policy on promoting gender and racial diversity and report on how these policies have been considered and applied. Companies are further required to report on the progress towards achieving agreed voluntary targets for gender and racial diversity on their board (JSE Limited, 2017, 2023).

While slow progress is being made in South Africa, current board membership does not reflect the country’s demographics (Ntim, 2015; Deloitte, 2022). Mans-Kemp and Viviers (2015) found that the percentage of female directors of South African listed companies increased from 5% in 2002 to 16% in 2012, while black directors increased from 13% to 34% over the same period. Although their study period included the issuance of King III in 2009, Mans-Kemp and Viviers (2015) did not focus on the implications of the code revision as King III did not differ substantially from King II with regard to board diversity. Various global and South African studies exploring different kinds of board diversity have been conducted, and those considered relevant to the present study are discussed further in the next section.

There has generally been an increase in board gender diversity in countries where board diversity quotas or CG codes that emphasise board diversity were introduced, although the increase appears to have been faster in countries with mandatory quotas (Ferrary, 2024). Many countries (such as Norway, Germany, France, Italy, Netherlands, Austria and Belgium) have successfully increased gender representation on boards by introducing mandatory board diversity quotas subject to sanctions (Ahern and Dittmar, 2012; Pastore, 2018; Fedorets et al., 2019; Bennouri et al., 2020; Ferrary, 2024; Van Kints, 2024). Other countries have recommended voluntary diversity targets, often introduced by CG codes. While there is no legal obligation or CG code recommending board diversity in the US, there has still been a gradual but consistent increase in both gender and racial diversity on US boards over the past decade, resulting from pressure from shareholders and institutional investors (Rastad and Dobson, 2022; Gormley et al., 2023).

From 2013 to 2020, a voluntary gender diversity target was in force in the Netherlands. However, it did not yield the desired results (Van’t Foort-Diepeveen, 2021) and, therefore, stricter regulations were introduced in 2022, when a mandatory gender quota replaced the voluntary target. Spain has also not achieved its non-mandatory gender quota deadline of 2015, which was first introduced as a legal obligation in 2007, although with no sanctions for non-compliance (Safiullah et al., 2022). Several other CG reforms to emphasise board gender diversity in Spain over the years that followed proved to be unsuccessful (Kirsch, 2021; Safiullah et al., 2022), and in 2024 a mandatory board gender quota was introduced, this time with sanctions for non-compliance (ECIJA, 2024).

Several rounds of CG reforms have attempted to address the lack of board gender diversity in the UK In 2010, the UK’s Corporate Governance Code was amended to encourage board diversity (including gender diversity) (Doldor et al., 2012). The Lord Davies report on women on boards, published in 2011, recommended a new voluntary female board representation target in the UK (Davies, 2011). In 2017, the Hampton–Alexander report was published, which further increased board gender targets (Brahma et al., 2020). Despite the lack of legal obligation to increase the representation of women on UK boards, the proportion of women on UK boards increased significantly following the introduction of voluntary targets (Brahma et al., 2020). In Australia, there was a significant increase in female representation on boards following the introduction of the Australian CG code by the Australian Securities Exchange in 2010 (Klettner, 2016), and Sweden has similarly significantly increased the proportion of women on boards by using voluntary actions (Onwuka, 2022).

Many studies have consistently found a high degree of compliance with CG codes and their amendments (e.g. Dedman, 2002; Werder et al., 2005; Del Brio et al., 2006; Akkermans et al., 2007; Fauver et al., 2017, in Chai-Aun et al., 2016). Capital market dynamics positively influence the effects of code revisions, particularly in enforcing new practices regarding board structure (Chai-Aun et al., 2016). In line with this, the JSE listing requirements mandate compliance with King IV and its enhanced focus on board diversity. Notably, King IV’s emphasis on diversity targets and reporting is anticipated to improve board diversity and the disclosure thereof within JSE-listed companies. This led to the following hypothesis:

In this section, the potential positive and negative effects of board diversity on corporate performance are discussed, along with a discussion of theoretical perspectives (Section 3.2.1). Section 3.2.2 discusses previous studies on the impact of board diversity on financial performance and the potential moderating effect of CG code revisions. In contrast, Section 3.2.3 discusses earlier studies on the impact of board diversity on sustainability performance. Both Sections 3.2.2 and 3.2.3 lead to the development of further hypotheses (H2a, H2b, H3a and H3b).

The mainstream viewpoint is that CG mechanisms seek to address the agency problem, and having a sufficiently independent board is deemed beneficial for corporate performance (Muth and Donaldson, 1998). Yet, regarding board diversity, the outcome is less straightforward (Carter et al., 2010). At the same time, the value-in-diversity hypothesis (Cox and Blake, 1991) suggests that diversity in a workforce generally improves employee efficiency and, ultimately, corporate performance (Bader et al., 2019).

Informed by this perspective, this study relies on resource dependence theory (Pfeffer and Salancik, 1978), which posits that companies are dependent on resources provided by others to sustain growth and that inter-organisational relations are a means of securing resources and mitigating the uncertainties of accessing such resources. A board of directors provides the link between the company and its external environment to, for example, secure information and expertise, and hence generate positive outcomes (Hillman et al., 2000). A more diverse board can potentially access more varied resources for a company, particularly owing to the broader range of networks, perspectives and relationships that come with diversity, which might otherwise be inaccessible to more homogenous boards. Women and racial minorities might have connections in underrepresented business sectors, communities or regions that can be valuable for market expansion, recruitment or partnerships (Smith-Hunter, 2006; Korosteleva et al., 2024).

Diverse perspectives on the board may also lead to more creative solutions, better problem-solving and increased external connections (Frias-Aceituno et al., 2012). This can be valuable when accessing new markets or seeking new business opportunities. Board diversity can reduce the likelihood of groupthink (Kamalnath, 2017), encouraging critical thinking and increasing the possibility of a diverse board challenging consensus and exploring new avenues for resources and collaborations. In addition, companies with diverse boards may find it easier to attract funding, as a diverse board may signal strong governance and potential for innovation (An et al., 2021).

As such, achieving board diversity should go beyond mere compliance; it should be pursued to enhance board effectiveness and corporate performance (Brummer and Strine Jr, 2022). Failing to tap into a pool of diverse, well-educated and ambitious directorial talent may limit value creation, compromise sustainability and undermine long-term competitiveness. External pressures, like shareholder activism, amplify the need for sustainable value creation via board diversity (IoDSA, 2016, 2021). Additionally, companies may enhance their external perception by cultivating diverse boards (Bear et al., 2010), thereby projecting a strong commitment to corporate citizenship from the top down. This approach can create value by increasing investor confidence. Leveraging board diversity, not just for compliance but as a mechanism to enhance effectiveness, is likely to bolster corporate performance (Deloitte, 2017).

Yet prior literature indicates that diversity may be a double-edged sword, increasing opportunities for creativity and innovation, yet potentially leading to dissatisfaction and conflict and eventually increasing uncertainties in access to resources (Milliken and Martins, 1996; Bader et al., 2019; Brummer and Strine Jr, 2022). Diverse approaches and perspectives can introduce challenges and potentially hinder decision-making (Carter et al., 2010). Such reduced cohesion could fuel conflict and distrust if not appropriately managed (Deloitte, 2017; Água and Correia, 2024). In addition, identifying diverse board members, especially in niche industries, may be challenging, time-consuming and expensive (Ahern and Dittmar, 2012; Deloitte, 2017; Van Zyl and Mans-Kemp, 2020; Young et al., 2025).

Lastly, the type of institutional pressure – whether board diversity is mandated or merely recommended – may influence companies’ board appointment practices differently. For example, Mateos de Cabo et al. (2022), in a study focused on board gender diversity, found that voluntary approaches (through codes) to promote gender diversity led to higher visibility and the co-option of “connected” women among the prominent companies in the market. In contrast, where mandatory gender quotas are in place (such as those discussed in Section 3.1), all companies are expected to comply, leading to a search for female directors beyond the usual networks (Bertrand et al., 2019; Mateos de Cabo et al., 2022). Although the revision in the King IV code recommends that diversity targets be set (and compliance with King IV is mandated by the JSE listing requirements), the code does not quantify any thresholds for the targets and how to achieve them. Instead, it leaves diversity targets to the discretion of the board itself. Consequently, King IV can be characterised as a “middle of the road” approach in seeking to foster greater board diversity, but, equally, this raises the question of the influence of successive code changes on board composition and corporate performance.

Various studies in the US and emerging markets, such as Turkey, Mauritius and some Middle Eastern countries, suggest that a diverse board may enhance financial performance (Carter et al., 2003, 2010; Ararat et al., 2010, 2015; Sarhan et al., 2019). Diversity in these studies was defined as gender, race, multiple demographic diversity indicators or combined diversity indices. In contrast, age and educational diversity reduced corporate performance in Mauritius (Mahadeo et al., 2012), while Jhunjhunwala and Mishra (2012) found no association between board diversity and corporate performance in India. A meta-analysis of diversity studies found that gender, ethnicity and nationality dimensions of diversity positively contributed to financial performance, through diverse thinking styles and the availability of unique and private information (Torchia and Solarino, 2025).

South African studies examined board diversity before the implementation of King IV. Taljaard et al. (2015) found that increased gender diversity and younger boards had strong associations with improved share price performance in the largest 40 JSE-listed companies from 2000 to 2013, while no association was found between racial diversity and financial performance. Mans-Kemp and Viviers (2015) considered companies in the healthcare, consumer goods, consumer services, technology, telecommunications and industrials sectors between 2002 and 2012. They found a positive relationship between both female and black directors and earnings per share, yet a negative relationship with total shareholder return. Scholtz and Kieviet (2018) observed a positive relationship between both gender diversity and the number of directors with a business qualification, and corporate performance (measured using Tobin’s Q and return on assets (ROA)) of the largest 100 JSE-listed companies from 2013 to 2015, while racial diversity negatively affected performance. However, some studies for different periods or subsamples (Jonty and Mokoaleli-Mokoteli, 2015; Hassan and Miller, 2017; Bissett, 2020) reported non-significant relationships between board diversity and corporate performance. Given some support in prior literature for a resource dependence-related benefit (Brummer and Strine Jr, 2022), the following was hypothesised:

There is a scant understanding of the implications of code revisions (e.g. Chai-Aun et al., 2016; Wang et al., 2020; Chai-Aun et al., 2021; Gull et al., 2023). While it is legitimate to expect codes to be revised for various reasons, the incremental benefits of doing so could be limited. Fauver et al. (2017) found worldwide evidence in 1990–2012 that board reforms increased firm value. Chai-Aun et al. (2016) found, in 35 countries from 2007 to 2014, that more code revisions increased the benefit of having a high board structure score related to having well-balanced board membership. However, these results did not hold when the effect of the number of new practices in CG code revisions in developing countries was examined (Chai-Aun et al., 2016). In the case of Pakistan, Gull et al. (2023) found that CG code reforms in Pakistan, including a board gender diversity exhortation, positively moderated CG practices and risk disclosure relationships. However, Wang et al. (2020) failed to find evidence in the same country of board diversity improving financial performance, either before or after the code revision.

One of the implications of King IV’s diversity revisions may be that companies adopt a compliance approach to the JSE listing requirements, which mandate compliance with the King Code. This may be further strengthened by the shift in King IV to an “apply and explain” approach (Natesan, 2020). If such compliance-seeking leads to critical mass in terms of race or gender representation on boards (Joecks et al., 2013), diversity may reduce groupthink (Kamalnath, 2017), while increasing external connections and providing access to new markets and business opportunities (Frias-Aceituno et al., 2012; Korosteleva et al., 2024), thereby affecting financial performance. While increased diversity after King IV may be perceived as a compliance exercise rather than substantive engagement in enhancing board effectiveness and corporate performance, a higher level of board diversity before King IV may be positively perceived as a commitment to corporate citizenship risk/uncertainty mitigation and increased investor confidence (Adams and Ferreira, 2009; Brummer and Strine Jr, 2022). We therefore expected the implementation of King IV to moderate the relationship between diversity and financial performance, and formulated the following hypothesis:

In recent years, boards have faced immense pressure to prioritise stakeholder interest, create sustainable value and ensure long-term organisational resilience (Esser and Du Plessis, 2007; IoDSA, 2016, 2021). Boards must be knowledgeable about environmental and social matters and incorporate them into their decision-making process. While some studies have found a positive relationship between board gender diversity and sustainability performance in developed countries (Velte, 2016; Arayssi et al., 2019; Romano et al., 2020; Menicucci and Paolucci, 2022), others reported no such association for gender or age diversity (Cucari et al., 2017; Manita et al., 2018; Menicucci and Paolucci, 2022). In Italy, Cucari et al. (2017) found a negative correlation between board gender diversity and environmental, social and governance (ESG) disclosure, while a Chinese study found that, as board gender diversity increased, increased ESG activities led to poor financial performance (Dong et al., 2023).

A fundamental pillar of King IV is the creation of value in a sustainable manner, echoing the United Nations Sustainable Development Goals, the Africa 2063 Agenda and the South African National Development Plan 2030 (IoDSA, 2016). Ntim and Soobaroyen (2013) uncovered a positive relationship between racial and gender diversity and the extent of South African companies’ corporate social responsibility (CSR) practices from 2002 to 2009. Another study of all JSE-listed companies (excluding financial companies) in 2015–2020 focused only on gender diversity and found a negative relationship with sustainability performance (Abdelkader et al., 2024). Two recent South African studies considered the effect of board diversity on sustainability disclosures and transparency (Nel et al., 2022; Toerien et al., 2023). Toerien et al. (2023), in a study of South African companies during 2011–2021, found no statistically significant relationship between board gender diversity and ESG disclosures. Nel et al. (2022) found that board composition (including increased board gender and ethnic diversity or directors over the age of 50 years) improved South African companies’ voluntary disclosure of CG and transparency-related information on companies’ websites and the accessibility of information to all stakeholders. However, these prior South African studies neither examined sustainability performance (they considered disclosure practices) nor included all board diversity aspects emphasised by King IV (namely, gender, racial, age and occupation diversity), aspects that are considered in the present study.

As the literature provides some support for enhanced sustainability performance for companies with a more diverse board, we hypothesised that:

We expected King IV to improve board diversity and for board diversity to enhance sustainability performance. If compliance with King IV leads to a critical mass of, for example, women on the board, this may lead to real outcomes for sustainability performance. For example, it is argued that women have empathy and care for society, translating into a better management of stakeholder relationships and emphasis on environmental stewardship (Rao and Tilt, 2016). If women have sufficient representation on the board, they may be able to influence decision-making (Joecks et al., 2013) and affect sustainability performance (Jain and Zaman, 2020). If boards not only substantively engage with the diversity exhortations in King IV but also shift focus towards inclusive capitalism, an integrated approach to reporting and a capital market that is sustainable over the long-term (Esser and Delport, 2018), then the impact of board diversity on sustainability performance may be moderated by the implementation of King IV. Therefore, we hypothesised that:

To address the research questions and the role of King IV, the sample period had to include years before and after its implementation. For listed companies, the application of King IV was required for documents submitted to the JSE from 1 October 2017, except for racial diversity disclosure at the board level, effective from 1 June 2018. The sample period of 2015–2019, therefore, included pre-King IV (2015 and 2016), transitional (2017) and post-King IV (2018 and 2019) phases.

The top 100 JSE-listed companies were selected as the sample owing to their significant market capitalisation and systemic influence on the South African economy. Larger companies, especially those included in major indices, are more likely to be subject to public scrutiny and regulatory enforcement, leading to higher compliance with reporting and governance standards (Ntim et al., 2012). These companies are also more likely to disclose extensive and reliable information, making them suitable for empirical research (De Villiers and Marques, 2016). A total of 133 companies were selected but, to ensure a balanced panel dataset for the regressions, companies not listed for the entire five-year period were removed, resulting in a final sample of 111 companies from diverse industries. To compensate for possible differences between industries and companies with primary versus secondary listings on the JSE, industry and listing type dummies were included in the regressions executed (see Section 4.3). Moreover, robustness checks were executed for all regressions by excluding companies in the financial industry (the largest industry in the sample) and companies with secondary listings on the JSE. As these robustness checks mainly yielded similar results (although statistical significance sometimes decreased with smaller sample sizes), it was decided to include these companies to ensure comprehensive coverage of all major industries in South Africa.

Board diversity data, corporate performance measures, and control variables for the regression analyses were collected across the sample period. See Table 1 for their measurement and how (and from where) they were collected.

Corporate performance was viewed as both financial performance and sustainability performance. The price-to-book ratio was selected as a proxy for financial performance as it provides a market-based measure that captures investor sentiment and expectations, which are crucial when evaluating the impact of board diversity on corporate performance (Taljaard et al., 2015). For our study, two aspects of sustainability performance were measured: environmental and social performance. ESG performance is a multidimensional construct that reflects distinct organisational priorities and is influenced by different governance mechanisms (Endrikat et al., 2021). While environmental performance is often prioritised in developed economies, previous South African studies have emphasised the importance of social performance in South Africa, given the social problems existing here (Toerien et al., 2023; Mouton et al., 2024). Environmental and social performance were proxied by the environmental and social scores sourced from the LSEG Workspace database, a platform owned by the London Stock Exchange Group (LSEG, 2023). The ESG data within this platform were initially developed by Refinitiv, a financial data provider that became part of LSEG following its acquisition in 2021. The ESG scores are calculated based on company-reported information and cover over 450 metrics, standardised, weighted, and aggregated into consistent scores to facilitate cross-company and cross-sector comparisons (LSEG, 2023). Importantly, these scores are sector-relative, meaning that a company’s performance is evaluated against peers within the same industry sector, allowing for more meaningful benchmarking (Velte, 2016; LSEG, 2023). Categories such as resource use, emissions and environmental innovation are considered for environmental performance, while metrics are drawn from categories including workforce, human rights, community involvement, and product responsibility for social performance (LSEG, 2023).

Data on the four dimensions of board diversity emphasised in King IV – gender, race, age and occupation – were hand-collected from information disclosed in company integrated (or annual) reports. King IV required companies to disclose targets and progress towards meeting targets, after 1 October 2017 for gender and after 1 June 2018 for race representation. Where available, the explicitly disclosed data were collected. Similar to Dos Santos et al. (2024) and Mans-Kemp and Viviers (2015), where not specifically disclosed, a pragmatic approach was adopted of inferring race and gender from photographs, first names, surnames and the use of gender-specific pronouns, supplemented by other online sources such as business publications and LinkedIn. Where judgement was employed, data from later periods corroborated the data collected. Inferring demographic data from publicly available information is relevant when measuring market perceptions of board diversity, such as the price-to-book ratio (Lockhart et al., 2023). Although name-based demographic inferences may be challenged theoretically and ethically, the domain expertise of the researchers in the South African context may enhance the accuracy of such inferences (Lockhart et al., 2023). Like other studies, racial diversity is defined as the percentage of directors of non-white origin, such as Black African, Indian or mixed-race (often termed “Coloured” in a South African context) backgrounds (Ntim, 2015; Scholtz and Kieviet, 2018) and consistent with the use of the generic term “black people” in the BBBEE Act (RSA, 2003). The latter term is used in this paper to refer to the three aforementioned non-white groups.

We did not include critical mass variables as we did not focus on a single aspect of board diversity (such as gender or racial diversity, in relation to which the critical mass threshold would be relevant) but on all four dimensions of board diversity emphasised by King IV. In addition to the separate dimensions of board diversity, a board diversity index was also calculated, in line with Ararat et al. (2015). However, while Ararat et al. (2015) calculated their index by dividing by the maximum diversity (e.g. 100% black or 100% female directors), we divided by the average sample diversity. For example, an index of 120% signified that a company’s board diversity exceeded the sample’s board diversity by 20% in a specific year. The board diversity index was left blank for companies that had missing gender, racial, age or occupation diversity data, as a comparable diversity index could not be calculated in such cases, which resulted in the average board diversity index not equalling exactly 100%.

Control variables for all regression analyses performed (see Section 4.3) were identified based on prior studies (Carter et al., 2010; Ntim and Soobaroyen, 2013; Jonty and Mokoaleli-Mokoteli, 2015; Ntim, 2015; Scholtz and Kieviet, 2018; Abdelkader et al., 2024). Board size and board independence, similar to board diversity, can improve CG and lead to increased financial and sustainability performance (Carter et al., 2010; Jonty and Mokoaleli-Mokoteli, 2015; Scholtz and Kieviet, 2018; Abdelkader et al., 2024). Moderate leverage can boost financial and sustainability performance, but excessive debt can reduce performance because of higher interest costs and related risk (Jonty and Mokoaleli-Mokoteli, 2015; Ntim, 2015; Scholtz and Kieviet, 2018; Abdelkader et al., 2024). Larger companies are more likely to be subject to public scrutiny and provide more extensive ESG disclosures (Ntim et al., 2012). Such companies may also perform better financially because of their economies of scale and market power, while higher sales growth is anticipated to improve corporate performance by raising revenues and profitability (Carter et al., 2010; Jonty and Mokoaleli-Mokoteli, 2015; Ntim, 2015; Scholtz and Kieviet, 2018; Abdelkader et al., 2024). Higher institutional ownership was associated with greater public scrutiny and increased governance, which could impact financial performance and ESG disclosures (Ntim and Soobaroyen, 2013). ROA, an accounting-based performance measure, was also a control variable (Abdelkader et al., 2024). Furthermore, performance might differ depending on the industry and listing type (Jonty and Mokoaleli-Mokoteli, 2015). A binary variable was also used to indicate the pre-King IV (2015 and 2016) and post-King IV (2018 and 2019) periods. The year 2017 was deemed to be a transitional year.

After scrutinising the dataset, including the dependent, independent and control variables, we found that some variables (price-to-book ratio, ROA, leverage and sales growth) contained outliers (based on the range between minimum and maximum values). The outliers were winsorised at 1 and 99%. To test H1, descriptive statistics were used to describe board diversity data over the sample period. ANOVA F-tests were executed to test the significance of any observed improvement in board diversity in the post-King IV period compared to the pre-King IV period.

Panel regression analyses were executed to test H2a, H2b, H3a and H3b, with either financial performance (H2a and H2b) or sustainability performance (H3a and H3b) as dependent variables. The independent variables were either the four individual board diversity variables (the dimensions emphasised by King IV) (models 1 and 3 below) or the board diversity index (models 2 and 4 below). When the board diversity index was employed, the binary variable relating to before/after King IV was employed as a moderator (models 2 and 4) to address H2b and H3b specifically. The models employed, aligned with prior studies (Manita et al., 2018; Scholtz and Kieviet, 2018; Toerien et al., 2023), are shown below:

where Controls_it in all regressions included Board size + Board independence + Leverage + Company size + Sales growth + Institutional ownership + Accounting performance + Industry and Listing type dummies for company i at time t.

Hausman tests showed that a fixed effects model was most appropriate when the dependent variable was price-to-book ratio (financial performance). In contrast, based on Hausman tests, a random effects model was suitable when the dependent variable was the E or the S score (sustainability performance). Robust standard errors were employed in all regressions. Year fixed effects were excluded, as a binary variable pertaining to before/after King IV was already included. Industry dummies and controls for listing type were used when a random effects model was employed but were omitted when a fixed effects model was employed, owing to multicollinearity with the fixed effects per company.

In line with Manita et al. (2018) and Toerien et al. (2023), the researchers also considered the risk of endogeneity, arising from omitted variable bias or reverse causality, influencing the results. To reduce the risk of omitted variables, various control variables were included based on prior studies and a panel model was employed to control for unobserved company-specific heterogeneities (Manita et al., 2018; Ntim and Soobaroyen, 2013). Moreover, in model 2 we used a novel board diversity index, calculated as the composite of four board diversity metrics and scaled by the average board diversity of the sample in a specific year. The board diversity index is unlikely to be highly correlated to any one omitted variable contained in the error term owing to its calculation method, which combines various aspects of board diversity into a single measure and compares this measure to the average board diversity of the sample (i.e. it is presented as a relative performance index). To further address endogeneity concerns, especially relating to reverse causality, all regressions with significant results relating to the variables of interest (board diversity and the effective date of King IV) were re-executed using the lagged variable of interest, as was suggested by Manita et al. (2018) and Toerien et al. (2023) (i.e. using the lagged board diversity variables). The results (untabulated) did not change significantly when the lagged board diversity variables were employed, showing that reverse causality and endogeneity did not bias the results. In the main regressions reported, however, it was deemed essential to measure the variables in the same period as the theoretical hypothesis was that board diversity was associated with improved performance in the same period, in line with Toerien et al. (2023). The focus on the effect in a particular year was crucial given that the research focused on a specific period, namely just before and just after the effective date of King IV.

Section 5.1 presents results related to board diversity trends around the effective date of King IV (H1). Section 5.2 shows the results for the relationship between board diversity and financial (H2a and H2b) and sustainability (H3a and H3b) performance. Section 5.3 summarises the results and hypothesis testing.

Table 2 reveals mixed trends in board diversity for the sample period.

The average percentage of women and black board directors increased from 20% to 27 and 32%–38%, respectively. These increases follow the increase in gender and racial diversity from 2002 to 2012, as reported earlier by Mans-Kemp and Viviers (2015). Age diversity remained relatively constant, while occupation diversity declined slightly over this period. The number of observations indicates that more companies disclosed three diversity variables – gender, racial and age diversity – in later periods, while disclosures on occupation diversity decreased slightly. Table 3 presents the results of ANOVA F-tests to identify significant changes in diversity data after the effective date of King IV.

The F-statistics and accompanying p-values show a rise in gender and racial diversity after King IV’s implementation. These results support H1, suggesting that the King IV revision increased board diversity. Ntim (2015) reported lower gender (17–18%) and racial diversity (22–30%) levels a decade before this dataset, compared to 20–27% and 32–38% respectively in 2015–2019. This gradual inclusion of women and black people on boards aligns with the code’s adoption and dissemination over time in listed companies. Resistance to over-regulation or reluctance to bind themselves to targets may explain why, by 2019, only one-third of JSE-listed companies were reporting on their gender diversity targets (BusinessEngage, 2020). While efforts were being made to broaden the pool of potential directors, lack of diversity in executive directorships affected the pipeline of potential NEDs (only 14% of executive directors on the JSE in 2019 were female) (BusinessEngage, 2020). New appointments were limited by smaller boards (board size decreased from 12.34 to 11.91 over the sample period), and the regular reappointment of directors retiring on rotation (BusinessEngage, 2017) and the proportion of NEDs serving for nine years or longer (27% of NEDs on the JSE in 2019) (PWC, 2020).

After winsorisation, the descriptive statistics of the ratio-scale variables in the dataset are provided in Table 4. Table 4 shows that mean gender and racial diversity stand at 23 and 35%, although wide variations are reported in the sample. On average, almost four age bands and five occupational categories are represented on the boards. While mean board diversity index is close to 1 (as was expected), there are notable variations in the sample, with companies either underperforming relative to average performance or significantly exceeding it. A two-way correlation test to identify potential multicollinearity revealed no significant correlation among the independent variables, with the largest correlation of 0.43 between gender and racial diversity. Because Mans-Kemp and Viviers (2015) noted diversity variations across industries, industry classifications of the selected companies were collected. The two largest industries were financials (35% of companies) and basic materials (20%) (not tabulated). Out of 555 observations, 85 (15%) were classified as secondary listings on the JSE.

Regression 1 in Table 5 reports the regression analysis results using the individual dimensions of board diversity – gender, racial, age and occupation diversity – as separate independent variables. Due to missing data for certain variables, the number of companies included in the regression analyses was reduced to 94 from an initial sample of 111. Board diversity in all its guises did not show significant relationships with financial performance at a 5% significance level. The finding that board diversity (specifically gender diversity) was not associated with increased corporate performance contrasts with results reported by Scholtz and Kieviet (2018), who also considered the largest 100 JSE-listed companies, but before King IV. This may indicate that the relationship between board diversity and corporate performance changed after the effective date of King IV. This factor was considered by adding a moderator variable in the second regression executed (Regression 2 of Table 5) to test H2b. Leverage and ROA were positively related to price-to-book ratio, while company size and institutional ownership were negatively associated.

Regression 2 in Table 5 shows the effect of the diversity index on financial performance, moderating for the impact of the implementation date of King IV. The moderator variable was included to determine whether the implementation of King IV changed the relationship between board diversity and financial performance. A change in the relationship was hypothesised (see Section 3) and seems plausible when comparing the results of Regression 1 to those of prior studies before King IV. From Regression 2 in Table 5, it may be deduced that a company having a higher diversity index than its peers was associated with an increase in its price-to-book ratio before King IV, indicating a positive market response to companies with greater diversity (H2a). The coefficient also exhibited economic significance; an absolute increase of 10% in a company’s diversity index was associated with a 0.15 increase in its price-to-book ratio. However, after the effective date of King IV, when diversity expectations were codified, the benefits of surpassing peers in diversity seemed to have decreased (evidenced by the negative coefficient attached to the moderation effect in Regression 2 of Table 5) (H2b). Again, this coefficient was economically significant, with an absolute increase of 10% in a company’s diversity index associated with a 0.19 decrease in its price-to-book ratio. Figure 1 shows a graphical depiction of the moderation effect. Here, one can see that, before the implementation of King IV, high board diversity was associated with increased financial performance (a higher price-to-book ratio). However, after King IV, the previous positive relationship between board diversity and financial performance changed into a slightly negative one, where higher board diversity was not rewarded with increased market performance.

The results may indicate that the market perceived the increased board diversity efforts after King IV as less value-enhancing. Considering other theoretical perspectives (e.g. Guest, 2019; Nguyen et al., 2020), there is a possibility that the emphasis on gender and racial targets raised investor concerns about potential board conflict and pressures to select underqualified directors (Guest, 2019). Alternatively, if new appointments have the same characteristics as the in-group, they may not have added incremental value to board and corporate performance (Mendiratta and Tasheva, 2025). Considering the time it may take to develop “productive” diversity that balances cohesion and independence may also explain extant results (Chai-Aun et al., 2016). New board members need time to be inducted appropriately. At the same time, experienced directors appointed to too many boards may be unable to dedicate the requisite time to each board (Mans-Kemp et al., 2018). However, it is acknowledged that “over-boarded-ness” may be limited in the sample period, with 92–95% of directors holding only one or two directorships (PWC, 2020). In 2019, the market reacted positively to board changes at MTN, a South African telecommunications provider, attributed by market commentators to the local knowledge of two independent directors from African markets where the company operated and the orderly nature of the appointment process (BAO, 2019). This anecdotal evidence from the South African context supports the resource dependence perspective and underscores the importance of recruitment and onboarding processes (Hillman et al., 2000; Ferreira, 2010; Hillman, 2015).

Table 6 shows the effect of the individual dimensions of board diversity on sustainability performance (Regressions 1 and 2) and the impact of the diversity index (a composite measure) on sustainability while moderating for the period before or after the effective date of King IV (Regressions 3 and 4). Due to missing data on environmental and social performance, the number of companies included in the regressions was further reduced to 80 from the 94 companies used in Table 5.

No significant associations were noted between board diversity, whether measured by subcomponents or a composite index, and sustainability performance (H3a). This aligns with prior research from developed countries (Cucari et al., 2017; Manita et al., 2018; Menicucci and Paolucci, 2022). An earlier South African study found a positive relationship between board diversity and independence and company CSR practices (Ntim and Soobaroyen, 2013). This study covered 2002–2009, when King II was still effective, and the board diversity and corporate sustainability landscape has changed substantially since then. Moreover, Ntim and Soobaroyen (2013) focused on CSR, which is but one subcomponent of sustainability performance. Another study of all JSE-listed companies (excluding financial companies) in the period 2015–2020 focused only on gender diversity and found a negative association with sustainability performance (Abdelkader et al., 2024), attributing the negative association to attempts by female directors to disprove negative gender stereotypes by focusing on short-term rather than long-term sustainability performance.

These results, echoing Guest (2019) and Nguyen et al. (2020), suggest that either tokenism or selection bias (a situation where appointed female or black directors merely reflect the characteristics of the existing “in-group”) may be at play, thus having virtually no incremental influence on outcomes (Torchia et al., 2011; Yarram and Adapa, 2021; Nalukenge et al., 2024). No significant changes in sustainability performance were noted after King IV’s implementation. However, the implementation moderated the relationship between the diversity index and sustainability performance (H3b). After the effective date of King IV, a higher diversity index was associated with reduced environmental performance but with enhanced social performance. The coefficients showed economic significance; after King IV, a 10% increase in a company’s diversity index corresponded to a 1.5 decrease in environmental performance (measured out of 100) and a 0.9 increase in social performance. Figure 2 and 3 graphically depict the moderation effect. Figure 2 shows a positive relationship between board diversity and environmental performance before King IV but a slight negative relationship after it. Contrastingly, Figure 3 illustrates a negative relationship between board diversity and social performance before King IV but a positive relationship after it was implemented. These results suggest that, after King IV, more diverse boards might have increased their focus on social sustainability while decreasing their focus on environmental sustainability. Toerien et al. (2023) reported a specific focus on social disclosures in the South African context but failed to find a relationship between board gender diversity and ESG disclosures. The present study appears to have uncovered a significant association given its expanded definition of board diversity (by focusing on racial, age and occupation diversity in addition to gender diversity), considering environmental and social performance individually, and specifically investigating the impact of King IV.

The Global South tends to prioritise social over environmental issues, and South African policies such as BBBEE may also have influenced this shift in focus (Visser, 2008; Cheruiyot-Koech and Reddy, 2022). Ntim and Soobaroyen (2013) found a positive association between board diversity and elements of social performance. Drawing from the stakeholder salience perspective (Mitchell et al., 1997), increased board diversity in South Africa may increase attention towards social stakeholders (e.g. communities and workforce), whereas environmental concerns receive less attention in the South African context. Africa’s shortage of reliable energy and the transition to green energy pit environmental sustainability against social sustainability and economic performance, which boards may find challenging to balance.

Table 7 summarises the results of the hypotheses testing, King IV’s effect, and contextual factors’ influence.

Beyond the well-researched question of whether board diversity matters, the emphasis of our paper was on the extent to which (if at all) subsequent exhortations and revisions about board diversity in CG codes can lead to significant change and enhance performance. A concern among many commentators is that companies may tend to pay lip service to diversity expectations, in turn leading to calls for more specific and “unambiguous” requirements (e.g. racial and gender targets), whose performance could then be evaluated via disclosures based on an “apply and explain” basis. However, while codes of CG internationally appear to increasingly detail expectations about board diversity (MacNeil and Esser, 2022), recent research on African codes highlights disparities in their formulation and operationalisation, with very few requiring (or even defining) numerical targets (Nalukenge et al., 2024). The change to the 2016 version of the Report on Corporate Governance for South Africa (King IV) provided an opportunity to examine these implications.

Although our findings show a modest and significant increase in gender and racial representation in the sample of companies following King IV, their implications for financial and sustainability performance are mixed. Contrary to previous evidence of a positive association between dimensions of board diversity and performance (Taljaard et al., 2015; Mans-Kemp and Viviers, 2015; Scholtz and Kieviet, 2018), individual aspects (age, gender, race and occupation) do not seem to significantly influence financial, market and sustainability indicators – highlighting the limits of specific dimensions of board diversity. The observed relationship between the board diversity index was negative (for price-to-book ratio and environmental performance) after King IV and positive for social performance, noting that a holistic consideration of board diversity can be beneficial, albeit in a selective way (i.e. on social performance). Furthermore, one key takeaway relates to the consequences of multiple code revisions. While Chai-Aun et al. (2016) conjectured that such revisions (such as changes in board structure) may be welcomed by the market, our evidence about specific board diversity practices (i.e. gender and racial targets) suggests otherwise.

Without further in-depth and qualitative exploration of board structures after King IV, pinpointing the specific reason(s) for these results remains difficult. Given the context of a well-established CG setting experiencing substantial change over the last 20 years – while grappling with the lingering implications of addressing painful historical inequities – the results underscore the limits of code prescriptions or targets without a deeper appreciation of the structural barriers to achieving “authentic” engagements with board diversity (Guest, 2019). As noted by Nalukenge et al. (2024), making lofty commitments to greater board diversity without detailed strategies for implementation may result in “visible” responses (i.e. more women and black board members), without fostering genuine understanding of how positive outcomes could flow from the board. Without a holistic perspective towards board diversity, tokenistic and/or “in-group” selection strategies could take prominence to “game” the metrics, with little effect on, or even at the expense of, performance. The modest (yet statistically significant) change in board diversity for gender and race also chimes with Mateos de Cabo et al.’s (2022) insight that the use of codes to promote gender diversity can have limited impact among all companies, that is, beyond the prominent and more visible entities, with some companies still reporting no female or black representation on the board. Finally, it is worth noting that our analysis was constrained by the relatively short-term nature of performance measures. It is entirely plausible that more diverse boards may require time to generate material benefits (Chai-Aun et al., 2016). Nevertheless, our study seeks to contribute to CG research by revealing board diversity trends after a code revision and empirically examining its influence on corporate performance, contrasting with prior work (Ntim, 2015; Nguyen and Muniandy, 2021) and supporting the US-based insights (Guest, 2019). Theoretically, our results suggest that board diversity studies relying on the resource dependence perspective must acknowledge that efforts to increase board diversity can themselves be uncertainty-inducing, at least in the short term.

Our paper has key implications for boards, policymakers and researchers. Firstly, it is essential to couple board diversity targets with a rigorous, appropriately designed and well-resourced national programme of identifying, selecting, shadowing and mentoring potential directors. This programme should aim to substantially broaden the pool of diverse board members, rather than merely incentivising a “surface-level” attainment of targets, ensuring that all companies substantively engage with the expectations and that board members can contribute to organisational outcomes. Secondly, with the current revision of the King Code (King V), policymakers appear to be applying a principles-based approach to the code and asking companies to explain how recommended practices (including board diversity and racial/gender targets) are being implemented. Such explanations may help market players and stakeholders better appreciate the company’s value-in-diversity approach, provided companies provide meaningful explanations. Lastly, we suggest researchers focus more on combined measures of board diversity, given the often interrelated and dynamic nature of board composition decisions.

Our results should be interpreted with caution owing to the relatively short pre- and post-King IV sample periods. Other regulatory and societal developments may also have influenced this period, such as the 2015 revision to the BBBEE Act, which we addressed by including the moderation effect of King IV into our model. We acknowledge that our paper has not explicitly considered skills and experience. Admittedly, the study’s analysis was limited by the availability of publicly accessible board composition data, which affected sample size. Future research could analyse disclosure of targets and progress made in achieving targets, along with factors influencing progress, as well as incorporating, for example, experience or tenure, to further analyse the depth and breadth of the influence of board diversity on performance. Analysing, as espoused by the JSE Sustainability Disclosure Guidance (JSE, 2022), how board diversity influences materiality assessments in sustainability reporting may also be valuable, alongside a consideration of specific social and environmental performance measures. Lastly, a qualitative analysis of “apply and explain” narratives on board diversity may further illuminate the companies’ rhetorical commitment and motivations on this point versus the “reality” of their current board composition, aspects that could be easier to capture given King V’s proposal to require an “apply and explain” disclosure template. Finally, given the country’s political and social context and historical legacies, and our focus on listed companies, caution is advised when directly transposing our inferences to other emerging economies.

The authors wish to acknowledge the financial support from the Global Challenges Research Fund and the University of Essex. Any errors included in this paper are the sole responsibility of the authors.