Equity at the Top: Board Diversity and Executive Remuneration in South Africa

Abstract

For listed companies, board diversity is often associated with improved decision-making, sustainability and financial performance. However, prior studies have neglected the interplay between board diversity and executive remuneration, especially in developing countries, over extended time horizons, and at the level of individual executives. This study addressed this gap by examining the evolution of board diversity and executive remuneration in South African listed companies from 2002 to 2017. Specifically, it investigated trends in board diversity and the determinants of executive remuneration, with particular attention to gender and ethnic pay gaps. Descriptive and regression analyses were conducted on a dataset comprising 8835 executive-level observations. Findings reveal a steady increase in female and non-white executive representation, possibly to align with societal expectations and remain legitimate. However, persistent gender and ethnic pay gaps were also noted, which might indicate that white and/or male executives are more entrenched and able to extract additional remuneration in line with the managerial power theory. The study contributes to the literature by documenting long-term trends in diversity and remuneration, providing empirical evidence on the influence of demographic attributes on remuneration outcomes, and offering insights for regulators, investors and non-executive directors seeking to advance equity and effective governance.

1. Introduction

1.1. Research Background and Motivation

The board composition of listed companies has been scrutinised for many years, with board diversity increasingly recognised as a key driver of improved decision-making, sustainability and financial performance (Arfken et al., 2004; Stephenson, 2004; IoDSA, 2016). From a theoretical perspective, agency theory explains that more diverse boards are often better equipped to hold management accountable, as diverse perspectives reduce the risk of groupthink and make it harder for opportunistic behaviour to go unnoticed (Jensen & Meckling, 1976). However, legitimacy theory reminds us that companies may sometimes pursue diversity less for meaningful governance change and more to signal compliance with social expectations, thereby strengthening their public image and maintaining legitimacy (Dowling & Pfeffer, 1975). Diversity can take many forms, but demographic characteristics—such as gender and race (or ethnic background)—are widely acknowledged in the international literature as critical dimensions that influence group dynamics and outcomes (Milliken & Martins, 1996). Globally, board gender diversity is encouraged through quotas or voluntary targets (Ahern & Dittmar, 2012; Ferrary, 2024), and previous global studies have reported substantial increases in female representation on boards over the past two decades, although mostly at the non-executive rather than executive level (Fedorets et al., 2019; Pastore, 2019; Ferrary, 2024). However, existing research on board diversity has placed less emphasis on changes in ethnic diversity and evidence from developing countries—a gap which the present study seeks to remedy.

The present study focuses on the South African context, a developing country where ethnic board diversity is particularly poignant, given the country’s history of inequality (Seekings & Nattrass, 2005). Over the past two decades, corporate governance reforms such as the King Report and Black Economic Empowerment regulations have encouraged changes in board composition without introducing mandatory quotas (Terblanche et al., 2025). Prior studies have indicated that gender and ethnic diversity at the board level is increasing, but again, mostly at the non-executive level (Taljaard et al., 2015; Viviers et al., 2017).

From both global and local perspectives, it remains unclear whether these increases in board diversity (i.e., the appointment of female and non-white board members) have led to equitable remuneration outcomes for new appointees (Amore & Garofalo, 2021). Consistent with managerial power theory (Bebchuk & Fried, 2004), incumbent executives—who are predominantly male and white—may be more deeply embedded within the company’s governance structures and social networks (Cook & Glass, 2014b), affording them greater influence over remuneration-setting processes and potentially leading to higher compensation outcomes and executive pay gaps. While some evidence of gender and ethnic executive pay gaps has been reported in developed countries (Selody, 2010; Elkinawy & Stater, 2011; Gregory-Smith et al., 2014; Amore & Garofalo, 2021), only one academic study has considered the executive pay gaps of listed companies in the South African context (Bussin & Barrett, 2016). Bussin and Barrett (2016) considered only ethnic diversity, not gender, and included only chief executive officers (CEOs) in their sample.

1.2. Research Gap, Questions and Contribution

Although prior research has examined the gender and ethnic diversity of the boards of South African listed companies, less attention has been given to trends in executive director (hereafter, executive) diversity, particularly over extended timeframes. Additionally, the existence of gender and ethnic pay gaps has not been comprehensively examined in this context, as most prior work in this field is from the United States (US). This highlights a significant gap in understanding board diversity, executive remuneration, and pay gaps in developing countries.

This study addresses the identified gap by examining the evolution of board diversity and executive remuneration in South African listed companies from 2002 to 2017, with the following two research questions: (1) What were the trends in executive remuneration and diversity characteristics over the period from 2002 to 2017? (2) Does executive remuneration provide evidence of gender or ethnic pay gaps? The findings contribute to the existing literature by documenting trends in female and non-white executive representation and further evaluating how executive demographic characteristics affect remuneration outcomes in South Africa. The findings have practical significance for policymakers, regulators, investors, and non-executive directors (specifically remuneration committee members), as they provide evidence on the effectiveness of diversity initiatives, help identify persistent pay gaps, and inform strategies to promote equitable and effective governance in South Africa.

The remainder of this paper is structured as follows. Section 2 reviews the relevant literature on board diversity and gender or ethnic pay gaps. Section 3 outlines the methodology employed to answer the research questions, while Section 4 discusses the empirical findings. Finally, Section 5 concludes the paper by summarising the key insights and providing recommendations for policy and future research.

2. Literature Review and Hypothesis Development

2.1. Understanding Board Diversity

The board of directors is responsible for leading the company and setting its strategic direction (IoDSA, 2016). Diversity on boards can be conceptualised in several ways. Milliken and Martins (1996) suggest that demographic diversity refers to observable characteristics, such as ethnicity, gender, nationality, and age, while cognitive diversity refers to educational, functional and occupational backgrounds. This study focuses on demographic diversity, as it was more directly targeted by the governance codes and regulations in effect during the period under review (2002–2017). During this period, corporate governance for South African listed companies, i.e., those with primary listings on the Johannesburg Stock Exchange (JSE), was primarily guided by King II (effective from 2002) and King III (effective from 2010). King II recommended that boards be structured to include a balance of executive and non-executive members and encouraged companies to consider demographic representation, including gender and race (referred to as ethnicity in this study), to bring a variety of perspectives to decision-making (IoDSA, 2002). King III extended these principles by explicitly advocating for the inclusion of independent non-executive directors, the establishment of formal board committees, and greater transparency in reporting on board diversity and performance (IoDSA, 2009). Both reports underscore that diverse boards are better positioned to provide adequate oversight, improve strategic decision-making, and enhance accountability to shareholders and stakeholders. These principles were also emphasised by the Black Economic Empowerment (BEE) regulations, which promote racial representation on boards. King IV became effective for South African listed companies from 2018 onwards (JSE Limited, 2017) and further strengthened these recommendations (IoDSA, 2016). Given these regulations, it was expected that board diversity, specifically gender and ethnic diversity, would have increased gradually over the 2002–2017 period.

Board diversity introduces a variety of perspectives that can improve decision-making, risk management, audit control, and ethical oversight (Arfken et al., 2004; Stephenson, 2004). Companies with female board members are often associated with better financial performance and higher returns, reflecting improved governance practices (Stephenson, 2004). From a theoretical standpoint, agency theory suggests that diverse boards may enhance monitoring by bringing varied perspectives, reducing herd behaviour and limiting managerial opportunism (Jensen & Meckling, 1976). This suggests that boards with greater demographic diversity, including gender diversity, are better equipped to provide adequate oversight of management. In contrast to agency theory, legitimacy theory suggests that companies may adopt board diversity not solely to improve governance, but also to signal alignment with societal expectations, thereby maintaining legitimacy with stakeholders (Dowling & Pfeffer, 1975). In this context, diversity may sometimes serve a symbolic purpose, creating a more favourable external impression without necessarily reflecting substantive organisational change. Taken together, these perspectives highlight that while diversity has clear governance benefits, the motivations for its adoption can be both practical and symbolic. The next sections evaluate board diversity trends, focusing separately on gender and ethnic diversity trends.

2.1.1. Board Gender Diversity Trends

Gender diversity in companies has become a global priority, with many countries introducing gender quotas or voluntary targets to enhance diversity on corporate boards (Ahern & Dittmar, 2012; Ferrary, 2024). There has been a substantial increase in female representation on global boards over the past two decades, driven by gender quotas and targets (Fedorets et al., 2019; Pastore, 2019). Studies conducted in Norway, Italy, Germany, and France have documented significant and rapid increases in female representation on boards of directors between 2002 and 2018, owing to the implementation of mandatory affirmative action laws (Ahern & Dittmar, 2012; Fedorets et al., 2019). However, many of these studies found that increases in gender diversity occurred mainly at the non-executive level, which emphasises the slow growth in female representation at the executive level (Pastore, 2019; Ferrary, 2024). Moreover, countries (such as Malaysia) that implemented voluntary gender diversity targets have seen slower, more gradual increases in gender diversity (Zainal et al., 2013).

Global evidence indicated a steady rise in gender diversity on corporate boards between the early 2000s and 2017 (OECD, 2017). OECD (2017) showed that females held only around 15% of board seats in listed companies in 2013, increasing to roughly 20% by 2017 across member countries. In Europe, female board representation nearly doubled within a decade, from about 11% in 2008 to approximately 25% by 2016, primarily driven by regulatory reforms (CSRI, 2016). CSRI (2016) similarly documented a global doubling of female board participation over the 2000s and 2010s, with females occupying close to 20% of directorships in 2015–2016, compared with less than 10% in the early 2000s. Although regional variation persisted, with Nordic countries and parts of Western Europe leading and emerging markets lagging, the evidence consistently shows an upward trajectory in female board appointments across listed exchanges, attributed to governance reforms, investor activism, and broader societal pressure for inclusivity (CSRI, 2016).

In South Africa, Taljaard et al. (2015) found that female representation on boards, including both executives and non-executives, increased notably among the top 40 JSE-listed companies—from 10% in 2000 to 20% by 2012. Mans-Kemp and Viviers (2015) and Nguyen and Muniandy (2021) analysed larger samples of JSE-listed companies. They reported increases in the percentage of female directors from 5% in 2002 to 16% in 2012 (Mans-Kemp & Viviers, 2015) and from 11% in 2009 to 16% in 2013 (Nguyen & Muniandy, 2021). However, the increase in gender diversity was mainly attributable to non-executive rather than executive directors (Mans-Kemp & Viviers, 2015). Scholtz and Kieviet (2018) corroborated that female representation on boards increased from very low levels in 2002 to around 20% in 2015. Focusing on the percentage of female Chief Executive Officers (CEOs) across 50 JSE-listed companies, Bezuidenhout (2015) noted an increase from 0% in 2005 to 14% by 2012. However, a study by Viviers et al. (2017) emphasised that female representation on boards slowed from 2011 to 2016, despite increased emphasis on gender diversity through policies such as King III during this time. A more recent study of 92 JSE-listed companies between 2015 and 2021 (during the King III and King IV periods) showed that female representation on boards continued to grow from around 20% in 2015 to 32% by 2021 (Toerien et al., 2023). Based on this data, South Africa is considered one of the global leaders in gender diversity on company boards, despite the absence of mandatory gender quotas (Toerien et al., 2023). Multiple studies have therefore emphasised an increase in female representation on boards from 2002 to 2021, although these studies generally did not differentiate between executive and non-executive directors.

2.1.2. Board Ethnic Diversity Trends

Globally, there seems to be greater focus on gender inclusivity than on ethnic diversity in company policies (Viviers et al., 2017). A study by Brammer et al. (2007) found that corporate boards in the United Kingdom remained largely homogeneous in 2002, with limited ethnic diversity. The study found that, on average, boards had 0.2 non-white directors, with only 13% of companies having one. Ethnic diversity also showed minimal sectoral variation, which, compared to gender sectoral diversity, might indicate that companies respond more to external pressures for gender inclusion than to workforce demographics.

Ethnic diversity, however, is an important focus area in South Africa (often referred to as racial diversity here), with efforts to increase inclusivity in corporate governance so that board memberships reflect the country’s demographics (Viviers et al., 2017). The King Report and other policies, such as the BEE Act, aim to promote ethnic inclusivity in high-level corporate governance, encouraging the participation of non-white directors in corporate decision-making. In South Africa, there has been a significant increase in ethnic diversity on corporate boards (Scholtz & Kieviet, 2018; Nguyen & Muniandy, 2021). Taljaard et al. (2015) found a steady increase in non-white representation on the boards of the top 40 JSE-listed companies, from 13% in 2002 to 34% in 2012, while Muniandy (2022) reported an increase from 28% in 2003 to 35% in 2018 across all JSE-listed companies. Among CEOs in 2012, Bezuidenhout (2015) found that 25% of a sample of 50 JSE-listed companies had non-white CEOs, whereas no non-white CEOs had been reported in a previous study in 2004. However, Taljaard et al. (2015) attributed the increase in ethnic diversity mostly to non-executive directors. The following section will discuss executive remuneration and, specifically, the existence of gender and ethnic pay gaps that have been noted as board diversity increased.

2.2. Gender and Ethnic Pay Gaps Relating to Executive Remuneration

Executive remuneration varies greatly across companies, industries, sectors and countries; however, three basic components make up executive remuneration: guaranteed salary, annual bonuses and long-term share-based incentives (Murphy, 1999). Executive remuneration has traditionally been viewed through the lens of optimal contracting theory, which stems from agency theory and emphasises structuring executive remuneration to align managers’ personal financial outcomes with value creation for shareholders (Jensen & Meckling, 1976). From this perspective, an individual executive’s remuneration can be explained by the role they hold in the organisation (e.g., a CEO position) and their contribution to company performance, thereby creating efficient contracting (Fama, 1980; Murphy, 1999). However, previous research has shown that executive remuneration is not always related to performance, leading to the application of other theories in this field.

The primary opposing theory is the managerial power theory, which argues that executives with greater influence over boards and remuneration committees can shape their own compensation arrangements and extract ‘rents’ (excess remuneration) in this way (Bebchuk & Fried, 2004). Managerial power can arise from formal authority, tenure, entrenchment, or social networks (Finkelstein, 1992) and can lead to rent extraction in companies with weak governance where executives are monitored less closely (Bebchuk & Fried, 2004). Westphal and Zajac (1995) continued this work and highlighted the role of demographic attributes in shaping governance and remuneration outcomes. It could be argued that executives who resemble dominant board groups in terms of gender and race (i.e., white, male executives) are more embedded in their organisations and enjoy greater legitimacy accompanied by weaker scrutiny. These mechanisms align closely with managerial power theory and imply that demographic characteristics such as gender and race may constitute important yet often overlooked determinants of executive remuneration.

Legitimacy theory explains that companies may appoint female and non-white executives to align with societal expectations and thereby enhance organisational legitimacy with external stakeholders (Dowling & Pfeffer, 1975). However, drawing on decoupling theory (Meyer & Rowan, 1977), companies may symbolically adopt diversity practices when appointing executives without fundamentally changing their internal business practices or how such executives are remunerated. This could result in female and non-white executives receiving lower remuneration despite holding comparable positions. As such, legitimacy-seeking behaviour may produce a decoupling between board diversity and remuneration equality at the top of the company.

Research in the US shows that female executives often earn less than their male counterparts in the same role, with the same tenure, and in the same company (Selody, 2010; Elkinawy & Stater, 2011; Gregory-Smith et al., 2014). Similar gaps are evident along ethnic lines, with some minority directors facing systematically lower remuneration and fewer leadership opportunities (Cook & Glass, 2014a). These findings suggest that increased representation has not fully translated into equity in remuneration. Amore and Garofalo (2021) highlighted substantial gender pay gaps among executives in US-listed companies between 1992 and 2012. Their study found that, on average, female executive remuneration was 19% less than that of their male counterparts, with the average remuneration ratio between female and male executives being 0.81:1. Amore and Garofalo (2021) further observed that companies with greater disparities in executive remuneration were more likely to experience declines in female representation at senior levels, due to the lower remuneration. When female executives resigned, they were typically replaced by males (Amore & Garofalo, 2021). This emphasises that gender-based remuneration inequality remains a persistent feature in corporate leadership in the US and contributes to the continued underrepresentation of women in top executive positions (Amore & Garofalo, 2021).

In the South African context, Bussin and Barrett (2016) found no significant difference in overall remuneration (guaranteed salary, bonuses, and total remuneration) between white and non-white CEOs of JSE-listed companies. However, the same study found that non-white CEOs’ remuneration was significantly more sensitive to company performance than that of white CEOs, likely due to greater performance-based remuneration (Bussin & Barrett, 2016). However, two studies by PwC (2019, 2022) on practices and remuneration trends show that both gender and ethnic pay gaps remain prevalent among directors. Non-white executives of JSE-listed companies earned 78% to 93% of what their white counterparts earned (PwC, 2022). Similarly, female executives of JSE-listed companies are paid, on average, only 75% of what their male peers earn, indicating that there is still room for transformation in corporate remuneration structures (PwC, 2019). Therefore, despite the significant increase in board ethnic diversity (Scholtz & Kieviet, 2018), substantial disparities in remuneration outcomes persist, with persistent evidence of gender and ethnic pay gaps. This leads to the following hypotheses being developed relating to research question 2:

Hypothesis 1 (H1).

Executive remuneration is lower for female and non-white executives relative to male and white executives (i.e., gender and ethnic pay gaps are observed).

3. Methodology

This section addresses the research methodology, including the target period, sample, data collection and analysis. While the conceptual framing of board diversity (a company-level variable) was used, the level of analysis was at the individual director level (i.e., executive-level diversity demographics, such as gender and ethnicity, were employed) to evaluate gender and ethnic pay gaps. It was decided to focus on the period 2002 to 2017, as a previously collected database of per-executive remuneration data existed for this period (Steenkamp & Wesson, 2021). Except for the existing database employed by this study, per-executive remuneration data is not readily available for South African companies. Data for later periods (after 2017) has to be hand-collected, which would have limited the number of companies and years sampled. As the literature review identified a gap in studying a large number of executives over a lengthy period, it was decided to use the existing database covering 2002–2017 for most JSE-listed companies. Moreover, the period covered the entire time that King II and King III, which had relatively similar requirements regarding board diversity, were effective. Thus, the database created by Steenkamp and Wesson (2021) served as the starting point, and these 215 companies were included in the sample. The database comprised all companies listed on the JSE main board for three or more years during the target period, excluding those in the basic materials and financials industries. The existing database contained 8835 per-executive line items for 1472 executives over the 16 years. It detailed the company name, year, executive name, and comprehensive remuneration data for the executive (

Table 1. Measurement of study variables.

Variable Names and Abbreviations	Measurement	Reference to Literature
Remuneration variables (employed as separate dependent variables)
Guaranteed salary (SAL), Bonus (BON), Gain from share-based remuneration (SB) and Total remuneration (TOT_REM)	Thousands of rands	Steenkamp and Wesson (2021)
Demographic characteristics (employed as independent variables)
Male gender (GENDER)	Coded as a one for male, zero for female and left blank if gender was not disclosed	Bussin and Barrett (2016); Amore and Garofalo (2021)
White ethnicity (ETHNIC)	Coded as a one for white, zero for non-white and left blank if ethnicity was not disclosed
Age (AGE)	Age in years
Tenure (TENURE)	Number of years working for company
CEO (CEO)	Coded as a one for CEO or zero if not
Company characteristics (employed as control variables)
Size (SIZE)	Logarithm of market capitalisation	Faulkender et al. (2010); Conyon (2014); Blanes et al. (2020); Amore and Garofalo (2021); Ramgath and Bussin (2022)
Performance (PERF)	Return on assets ratio
Share return (SR)	Change in share price expressed as a percentage of the opening share price
Growth opportunities (GROW)	Price-to-book ratio
Leverage (LEV)	Debt-to-assets ratio

). Steenkamp and Wesson (2021) created the database by combining data from the IRESS financial database and hand-collecting data from published annual reports.

3.1. Data Collection

The literature review cited specific demographic characteristics that could affect executive remuneration, which were hand-collected from annual reports (see Table 1). Control variables were identified based on prior studies and collected from IRESS financial database, per company-year (Table 1).

3.2. Data Analysis

For the first research question, basic trend analyses (utilising descriptive statistics) were conducted relating to executive remuneration and board diversity data. For the second research question, a one-way fixed-effects panel regression was conducted after conducting a Hausman test to determine whether to use random or fixed effects (i.e., company fixed effects were included in all regressions). Separate regressions were executed for each of the following dependent variables: guaranteed salary, bonus, gain from share-based remuneration and total remuneration. Year dummies were added, and robust standard errors were employed to account for expected heteroscedasticity. The regression models are shown below (using the abbreviations from Table 1):

• SALi,t = β0 + β1GENDERi,t + β2ETHNICi,t + β3AGEi,t + β4TENUREi,t + β5CEOi,t + β6SIZEi,t + β7PERF,t + β8SRi,t + β9GROWi,t + β10LEVi,t
• BONi,t = β0 + β1GENDERi,t + β2ETHNICi,t + β3AGEi,t + β4TENUREi,t + β5CEOi,t + β6SIZEi,t + β7PERF,t + β8SRi,t + β9GROWi,t + β10LEVi,t
• SBi,t = β0 + β1GENDERi,t + β2ETHNICi,t + β3AGEi,t + β4TENUREi,t + β5CEOi,t + β6SIZEi,t + β7PERF,t + β8SRi,t + β9GROWi,t + β10LEVi,t
• TOT_REMi,t = β0 + β1GENDERi,t + β2ETHNICi,t + β3AGEi,t + β4TENUREi,t + β5CEOi,t + β6SIZEi,t + β7PERF,t + β8SRi,t + β9GROWi,t + β10LEVi,t

4. Results

4.1. Descriptive Statistics and Correlations

The descriptive statistics for the collected variables are shown in

Table 2. Descriptive statistics.

Variable	Number of Observations	Mean	Standard Deviation	Minimum	Maximum
Remuneration variables in R’000 (employed as separate dependent variables)
Guaranteed salary	8835	2212.358	2320.140	0	69,221
Bonus	8835	1247.687	2587.494	−5000	54,031
Gain from share-based remuneration	8835	923.391	7432.163	−260	594,500
Total remuneration	8835	4383.436	9134.912	0	627,526
Demographic characteristics (employed as independent variables)
Male gender	8642	0.930	0.255	0	1
White ethnicity	7903	0.886	0.318	0	1
Age	8142	50.201	8.782	22	84
Tenure	8007	8.366	7.534	0	55
CEO	8835	0.268	0.443	0	1
Company characteristics (employed as control variables)
Size	8835	21.124	2.179	12.221	28.046
Performance	8818	12.961	17.664	−198.082	205.559
Share return	8250	0.278	2.112	−0.996	119.200
Growth opportunities	8648	2.310	3.108	−5.986	120.175
Leverage	8818	0.537	0.236	0	2.325

. All remuneration data are shown in thousands of Rands (R’000). Minimal observations with negative remuneration amounts were noted but not excluded from the regression analyses. There are instances in which a bonus (or a share-based payment) from a prior year might be recouped in the next year, so it was not deemed appropriate to exclude negative remuneration amounts.

Table 3. Correlation matrix.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
(1) Male gender	1.000
(2) White ethnicity	0.136	1.000
(3) Age	0.193	0.114	1.000
(4) Tenure	0.104	0.129	0.499	1.000
(5) CEO	0.125	0.053	0.107	0.198	1.000
(6) Size	0.052	−0.046	0.224	0.035	−0.016	1.000
(7) Performance	−0.039	−0.023	−0.008	0.018	−0.002	0.217	1.000
(8) Share return	0.017	0.014	−0.041	−0.049	−0.004	0.019	0.085	1.000
(9) Growth opportunities	−0.023	−0.033	0.047	0.015	0.007	0.425	0.293	0.062	1.000
(10) Leverage	0.042	−0.083	−0.003	−0.079	−0.021	0.061	−0.204	0.012	0.145	1.000

shows the pairwise correlations among the independent and control variables employed in the regression analyses to address research question 2. None of the independent and control variables were highly correlated, and thus, no multicollinearity problems were noted.

4.2. Research Question 1: Trends in Executive Remuneration and Diversity Characteristics

Figure 1 shows the trends in the average values of remuneration components (guaranteed salary, bonus, gain on share-based remuneration) over the sample period.

The average guaranteed salary increased gradually over the period, quadrupling from R959,450 in 2002 to R3,861,900 in 2017 (a 303% increase). If the average value had increased only with inflation (at 5.5% per year), the 2017 figure would have been R2,140,082, indicating an above-inflation increase. Average bonuses showed a similar trend, increasing from R391,125 in 2002 to R2,208,410 (a 465% increase). The gain on share-based remuneration also showed an upward trend, but with a spikier pattern. During 2002–2006, a steep upward trend was observed, but decreases occurred in 2007–2009 during the global financial crisis. From 2010 to 2017, a more cyclical up-and-down pattern emerged. The differences above were accounted for by utilising year dummies in the regression employed to answer research question 2. Figure 2 shows the number of female executives employed by JSE-listed companies between 2002 and 2017 compared to the number of male executives.

Figure 2 shows a gradual decrease in the number of male executives from 2002 to 2017, from 624 to 365 per year, while the number of female executives remains essentially unchanged (ranging from 30 to 42 per year). Thus, the substantial decrease in the number of male executives is more due to an overall decline in the number of executives than to an increase in the number of female executives. There was, however, a gradual increase in the percentage of female executives employed by JSE-listed companies between 2002 and 2017, increasing from 4% to 9% (with a corresponding decrease in the rate of male executives). The increase in the percentage of female executives likely reflects the impact of diversity initiatives, employment equity legislation, and governance principles such as those recommended in the King Reports (Taljaard et al., 2015). It must be emphasised that these findings diverge from those reported in prior literature (for example, Taljaard et al. (2015) who reported 20% female directors by 2012 and Toerien et al. (2023) who reported 20% in 2015) as the present study exclusively focused on executive directors, whereas these previous studies included both executive and non-executive directors. This supports the proposition that board gender diversity in South Africa is driven mainly by the increasing number of female non-executives rather than by executive directors. The King Reports also emphasise board ethnic diversity (using the term ‘racial diversity’), which will be considered next. Figure 3 shows the number of executives from white and non-white ethnic backgrounds employed by JSE-listed companies between 2002 and 2017.

The number of white executives decreased from 516 in 2002 to 320 in 2017, reflecting the overall decline in the number of executives employed by JSE-listed companies over the period. Over the same period, the number of non-white executives increased from 38 in 2002 to 69 in 2010 and then decreased to 50 in 2017. However, overall, the percentage of non-white executives rose from 5% in 2002 to 12% in 2008, and remained approximately at that level (ranging between 11% and 13%) until 2017. This supports the idea of a slowing trend in board diversity in the early 2010s (Viviers et al., 2017). Again, the percentage of non-white directors in the present study, which focused solely on executive directors, is substantially smaller than that reported by previous studies that considered both executive and non-executive directors. For example, Taljaard et al. (2015) reported 13% non-white executives in 2002 and 34% in 2012, while Muniandy (2022) found an increase from 28% in 2003 to 35% in 2018. This indicates that non-executives mainly contribute to the increase in board ethnic diversity in South Africa (Taljaard et al., 2015).

4.3. Research Question 2: Determinants of Remuneration and Pay Gaps

Table 4. Average remuneration earned per gender and ethnic category.

	Number (Percentage) of Observations	Average Guaranteed Salary	Average Bonus	Average Gain on Share-Based Remuneration	Average Total Remuneration
Per gender category
Female	605 (7%)	1640.790	669.638	398.752	2709.180
Male	8037 (91%)	2282.043	1314.897	982.633	4579.573
Unallocated	193 (2%)	1102.212	260.912	101.010	1464.135
Per ethnic category
Non-white	900 (10%)	2373.700	1513.532	628.839	4516.071
White	7003 (79%)	2315.001	1332.327	1062.319	4709.646
Unallocated	932 (11%)	1285.307	354.989	163.933	1804.230

shows the average remuneration earned by gender and ethnic category over the entire 16-year period.

Table 4 shows that female executives were paid less than their male counterparts. Differences were observed in the remuneration of non-white and white executives, but no clear trend was noted. These matters were investigated further utilising regression analyses to answer research question 2, as other demographic factors (such as age, tenure, or position) also influence executive remuneration and need to be controlled for (

Table 5. Determinants of executive remuneration—regression results.

Variables	Dependent Variable Employed
	Guaranteed Salary	Bonus	Gain from Share-Based Remuneration	Total Remuneration
Demographic characteristics
Male gender	305.267 **	364.732 ***	−4.983	665.017
White ethnicity	364.521 ***	176.417	467.785 ***	1008.723 ***
Age	4.996	4.964	−7.489	2.47
Tenure	43.598 *	32.089 ***	71.91 *	147.596 **
CEO	1260.067 ***	1020.749 ***	990.138 ***	3270.954 ***
Company characteristics
Size	43.214	253.225 ***	451.266 ***	747.705 ***
Performance	1.368	2.12	14.855 *	18.343*
Share return	−24.76	61.528 **	19.475	56.243
Growth opportunities	1.103	55.368	214.617	271.088
Leverage	312.884	252.269	336.065	901.218
Constant	−1579.310	−6272.165 ***	−10,700.861 ***	−18,552.335 ***
Year dummies	Yes	Yes	Yes	Yes
Number of observations	6641	6641	6641	6641
R-squared	0.271	0.120	0.018	0.091
F-statistic	21.648	4.916	2.760	7.058
Prob > F	0.000	0.000	0.000	0.000

*** p < 0.01, ** p < 0.05, * p < 0.1.

). The dependent variables were the remuneration components (guaranteed salary, bonus, and gains from share-based remuneration) and total remuneration (the sum of these components). In the regression analyses, remuneration was analysed only in nominal terms, as year dummies were included. Where an executive’s gender or ethnicity could not be determined (termed ‘unallocated’ in Table 4), this resulted in missing data, and these observations were excluded from the regressions that follow in Table 5.

Table 5 shows that male executives received higher guaranteed salaries and bonuses than female executives. The reported gender pay gap showed economic significance: males received a salary of R305,267 more than females, compared to a mean salary of R2,212,358, while male bonuses were R364,732 higher, compared to a mean bonus of R1,247,687. The higher remuneration paid to male executives is consistent with Amore and Garofalo (2021) and Selody (2010), who documented persistent gender pay gaps in executive compensation. This also aligns with South African descriptive evidence from PwC (2019), which reported that female executives earned substantially less than male executives. Guaranteed salary, gain from share-based remuneration and total remuneration showed a positive correlation with white ethnicity. Again, the reported ethnic pay gap had economic significance: white executives earned R364,521 more in salary (against a mean salary of R2,212,358) and R467,785 more in share-based remuneration (against a mean gain of R923,391) than their non-white counterparts. The results are consistent with PwC (2022), which states that ethnic disparities may manifest differently across some remuneration components, but not others. The results, however, contrast with those of Bussin and Barrett (2016), who found no significant difference in mean remuneration between white and non-white CEOs, indicating that these differences might be more prevalent among non-CEOs. Note that the R-squared for the gain from share-based remuneration is very low (1.8%), signalling low explanatory power relating to this remuneration component. Share-based remuneration is inherently volatile and often driven by company- and contract-specific factors not captured in archival data.

The fact that a non-negligible portion of observations (11%) was unallocated in terms of ethnicity could have biased the results. Thus, as a robustness test, the regressions were re-executed when assuming that all the unallocated observations should have been allocated as white (the one extreme) or as non-white ethnicity (the other extreme). The results remained essentially unchanged in terms of the directions of the coefficients and which of the main variables of interest (gender and ethnicity) were significantly related to remuneration. Interestingly, under both assumptions (that all unallocated should have been coded white or all should have been coded non-white), the association between total remuneration and male gender became statistically significant (it currently has a p-value of 0.145). In some cases, the level of statistical significance increased or decreased slightly, but the same overall findings could be drawn from the results under these robustness tests. To mitigate potential endogeneity, additional robustness tests were conducted by lagging all independent variables (per executive). The results of these robustness tests were similar to those reported in Table 5, except that the association between guaranteed salary and male gender became statistically insignificant. Taken together, the robustness tests confirm the reliability of the reported results: gender and ethnic pay gaps exist for some remuneration components. Thus, hypothesis 1 is supported relating to certain remuneration components, but not for all components of executive remuneration.

Companies tended to pay higher remuneration to CEOs (compared to other executives) and to longer-tenured executives. Consistent with prior global studies by Conyon (2014) and Blanes et al. (2020), larger companies paid higher executive remuneration, particularly in bonuses and share-based remuneration. Company performance, measured as return on assets, showed a positive correlation with share-based and total remuneration, while share return was positively related to remuneration from bonuses, consistent with the idea of pay for company performance. This aligns with Faulkender et al. (2010), Conyon (2014) and Ramgath and Bussin (2022), who found positive relationships between company performance and variable or share-based remuneration. It is acknowledged that the analysis relied on accounting-based performance measures and share returns, but alternative indicators of company performance or risk could have provided a more nuanced understanding of pay-performance sensitivity.

To evaluate whether the pay gaps noted in Table 5 narrowed over time, the 16 years were divided into the King II period (2002–2009) and the King III period (2010–2017). The regressions were re-executed for these two sub-samples, and only the pertinent coefficients and p-values regarding the gender and ethnicity pay gaps are reported in

Table 6. Gender and ethnicity pay gaps for the King II and King III periods.

Variables	Dependent Variable Employed
	Guaranteed Salary	Bonus	Gain from Share-Based Remuneration	Total Remuneration
Male gender (2002–2009)	139.424 *	101.307	−93.75	146.98
Male gender (2010–2017)	393.661	513.594 ***	−191.551	715.704
White ethnicity (2002–2009)	308.97 ***	122.714	342.507 ***	774.192 ***
White ethnicity (2010–2017)	541.13 ***	373.886	422.591	1337.607 **

*** p < 0.01, ** p < 0.05, * p < 0.1.

.

For guaranteed salaries, a statistically significant gender pay gap was apparent only during the King II period, not during the King III period, indicating a less persistent pay gap. However, the gender pay gap for bonuses widened and became statistically significant during the King III period, providing evidence of a persistent pay gap in the latter years of the target period. Relating to the gain from share-based remuneration, the pay gap was not statistically significant in either the King II or King III period. Regarding the ethnic pay gap, the rand-value of the pay gap increased across all remuneration components, but the gain from share-based remuneration no longer showed statistical significance during the King III period. Seen together, the analysis in Table 6, though not causal, provides some evidence of a less persistent gender and ethnic pay gap in the South African context over the 2002–2017 period. However, it must be noted that this analysis did not explicitly test for a structural break in the data between the King II and King III periods.

5. Conclusions

This study set out to answer two key questions: (1) how executive remuneration and diversity characteristics changed in South African listed companies between 2002 and 2017, and (2) what the determinants of executive remuneration were, including whether gender and ethnic pay gaps existed (and changed over time) once differences in executive and company characteristics were taken into account. For research question 1, the results show a clear upward shift in total executive remuneration over the 16 years. Guaranteed salaries and bonuses grew well above inflation, while share-based remuneration rose more unevenly. Diversity trends showed slow, uneven progress: the number of male executives declined steadily, while the number of female executives remained almost unchanged, resulting in only a modest increase in the proportion of women in executive roles. A similar pattern emerged for ethnic diversity: the proportion of non-white executives rose during the early 2000s and then stabilised, reflecting some progress but also revealing that there has been limited transformation at the executive level. These trends align with the intentions of the King Reports, yet they also show that meaningful change in executive demographics happened gradually and modestly. For research question 2, the regression results revealed gender and ethnic pay gaps in South African executive remuneration. Male executives earned higher guaranteed salaries and bonuses than female executives, and white executives received higher remuneration than their non-white counterparts across several remuneration components. These differences persisted even after accounting for factors such as age, tenure, CEO status, company size and company performance. Theoretically, the results might indicate that white and/or male executives are more entrenched and able to extract additional remuneration in line with the managerial power theory. However, because the data are observational, the regression results identify only associations, not causality. Thus, one cannot assume that the pay gap arises from discrimination, as unobserved factors (such as qualifications and experience) might have affected the results. When the period was split into King II and King III, some evidence of a less persistent pay gap was noted.

This study contributes to the South African executive remuneration literature by providing a long-term view of 16 years, capturing changes in remuneration structures and shifts in executive demographics. It also aids the understanding of executive remuneration and pay gaps in developing countries, which is relatively under-researched. This study also highlights how South African patterns differ from global findings, offering insights into how policy frameworks, such as corporate governance codes, may have influenced executive-level outcomes.

The study is not without limitations. The results are contextually dependent and cannot be generalised beyond South Africa or the 2002–2017 period. Thus, different results might exist in other countries or time frames. A specific limitation is that only data up to 2017 are included. Moreover, only listed companies outside the financials and basic materials sectors were included in the sample, which may limit the external validity of the findings, particularly given the importance of these sectors in the South African economy. Gender and ethnicity were captured using binary indicators, which oversimplifies demographic diversity and may mask heterogeneity within non-white groups or across different career trajectories of executives. Additionally, the data captured only broad demographic categories such as gender and ethnicity, while qualifications and experience were not measured. Although a dummy variable for CEO (versus non-CEO) was included, the study does not control for differences arising from other roles, such as chief financial officer or chief operating officer. The discussion of legitimacy versus substantive change remains largely interpretative, as the empirical design does not directly distinguish symbolic compliance from genuine governance improvements.

The limitations open several possibilities for future research. The study could be expanded to other (developing) countries; cross-country studies are specifically encouraged. Moreover, studies could examine more detailed demographic characteristics, executives’ career paths, or examine how remuneration committees make decisions. Finally, extending the South African dataset beyond 2017 would enable evaluation of the impact of King IV (and King V) and whether more recent governance reforms have supported greater equity in executive remuneration.