Executive performance evaluation and remuneration: Disclosure and practices of selected listed South African companies (2002−2015)

Abstract

Given growing inequality in South Africa, shareholders are increasingly questioning the size and composition of executive remuneration packages. They are also demanding greater transparency on the criteria and processes used to award performance incentives. The researchers hence investigated the extent and depth to which a sample of companies listed on the Johannesburg Stock Exchange (JSE) disclosed details on their executive performance evaluations. Attention was furthermore given to whether these companies reported a link between their executives’ pay and performance. The criteria and time frames used to evaluate and reward executive performance were also explored. Content analysis was performed on 2 136 annual/integrated reports over the period 2002 to 2015. Semi-structured personal interviews were also conducted with six directors serving on local remuneration committees. Controlling for company size, a significant increase was noted in the number of JSE-listed companies that disclosed information on their executive performance evaluations over the research period. The depth of these disclosures, however, remains superficial. As such, shareholders are constrained in their ability to hold remuneration committees accountable. These committees are encouraged to adopt a wider range of performance criteria and re-assess their focus on short-term performance. It is also suggested that commerce educators cultivate an appreciation for long-term, sustainable value creation among graduates.

Keywords:

• Executive remuneration
• Performance-based incentives
• Performance evaluations
• King II
• King III
• Remuneration committees
• Shareholder activism

JEL Classification:

• M12 Personnel management
• Executives
• Executive compensation

Introduction

Reports of executives receiving large bonuses often elicit strong reactions from the public. This was the case in October 2016 when Shoprite Holdings Limited announced that their chief executive officer (CEO) would receive a performance bonus of R50 million which was almost equal to his base salary (R49.7 million) (Kew, 2016). The use of cash bonuses and share options as performance incentives was first recommended by Jensen and Meckling (1976) in their seminal paper on agency theory. They argued that these incentives could be used to align managers’ interests with those of shareholders. Cash bonuses and share options subsequently became standard components of executives’ pay packages (Brookfield & Ormrod, 2000).

As far back as the 1980s, executives in the United States (US) came under fire for receiving what the public considered to be ‘excessive’ remuneration. Jensen and Murphy (1990) investigated this claim and discovered that dollar values were not excessive per se. The problem rather lay with executives being rewarded amidst deteriorating company performance. Lee (2002) reported similar findings more than a decade later.

The dangers of using share options as performance incentives were also highlighted by other scholars. Studies show that executives who receive share options tend to focus on short-term results, are more inclined to take excessive risks and engage in speculative behaviour (Bolton, Scheinkman, & Xiong, 2006; Kirkpatrick, 2009). This behaviour contributed, in part, to the dot.com bubble in 2002 and the 2007−2009 global financial crisis (Bhagat & Bolton, 2014; Fahlenbrach & Stulz, 2011; Faulkender, Kadyrzhanova, Prabhala, & Senbet, 2010).

While some shareholder activists engaged with companies on the topic of executive remuneration prior to the global financial crisis, their endeavours paled in comparison with those which took place during the subsequent ‘shareholder spring’. Despite being rescued from bankruptcy in 2008, a number of banks in the US and the United Kingdom (UK) awarded their top executives bonuses at the end of that year. These bonuses, which former US President Barack Obama called ‘shameful’ (Brady & Randall, 2009), angered many investors and taxpayers (Corkery & Medarevic, 2013).

Since the global financial crisis, shareholder engagements on executive remuneration have grown in frequency and intensity (Faulkender et al., 2010). Not only are shareholders questioning the size and composition of executive pay packages, but they are also demanding greater transparency on the criteria and processes that are used to award performance-based incentives. Whilst some shareholders are voting against remuneration-related resolutions at annual general meetings, others are using traditional and social media to convey their discontent. A growing number of shareholders are also raising their concerns directly with remuneration committees (Del Guercio, Seery, & Woidtke, 2008). Research shows that some of these engagements have been effective in changing executive remuneration policies and practices (Ferri & Maber, 2013; Hooghiemstra, Kuang, & Qin, 2017). Whereas most shareholder engagements are motivated by poor financial performance, others focus on fairness and social justice (Wade, O’Reilly, & Pollock, 2006).

Executives in South Africa have not escaped investor scrutiny. Local shareholders and policymakers are beseeching remuneration committees to evaluate executives’ performance and to link it to sustainable, long-term value creation (Hartzell & Starks, 2003; Institute of Directors in Southern Africa (IoDSA), 2016). The well-known shareholder activist Theo Botha has long been critical of the executive remuneration policies and practices of companies listed on the Johannesburg Stock Exchange (JSE). Botha frequently lambasts companies for failing to disclose sufficient details on executive emolument and for rewarding executives generously despite poor financial performance, retrenchments and dividend cuts (Viviers, 2015). Questions have also been raised in the media about the unfair distribution of wealth between executives and lower-level employees in a country that has one of the highest levels of inequality in the world (International Labour Organization, 2017).

Most of the studies that have been conducted on the topic of executive remuneration in South Africa have centred on the association between executive pay and a company’s financial performance (Bradley, 2013; Bussin & Blair, 2015; Bussin & Modau, 2015; Bussin & Nel, 2015; Crafford, 2015; De Wet, 2012; Scholtz & Smit, 2012; Shaw, 2011; Van Blerck, 2012). Limited academic research has, however, been undertaken on the nature and depth of information that is publicly disclosed on executives’ performance evaluations and the link between their pay and performance (be it their own or that of the company).

In the light of the research gap and the relevance of the topic from a social justice point of view, the purpose of the study was three-fold. Firstly, to investigate the extent and depth to which selected JSE-listed companies disclosed details on their executive performance evaluations. Secondly, to determine whether they reported a link between their executives’ pay and performance. Thirdly, to explore the criteria and time frames that selected JSE-listed companies use to evaluate and reward executive performance.

A mixed-methods approach was adopted to collect and analyse quantitative and qualitative data. Content analysis was conducted on the annual/integrated reports of 260 JSE-listed companies over the period 2002 to 2015. The analysis was informed by specific guidelines set out in the King II and III reports. Semi-structured personal interviews were furthermore conducted with six directors serving on local remuneration committees.

Shareholders have rights and responsibilities (Bower & Paine, 2017). To properly exercise their rights, they require detailed information. Given South Africa’s well-developed corporate governance framework, it could be expected that listed companies would report details on executive performance evaluations and remuneration to enable shareholders to make informed decisions. This is, however, not necessarily the case. Unless JSE-listed companies become more responsive to calls for improved disclosure on executive remuneration policies and practices, it is likely that they will encounter increasing resistance from shareholders and other key stakeholders.

Pertinent theories relating to executive remuneration are explored next, followed by an overview of executive performance evaluation and the link between executives’ pay and performance. Details are then provided on the methods used to collect and analyse quantitative and qualitative data. Next, key findings are presented followed by pertinent conclusions and recommendations for local remuneration committees and academicians.

Prominent theories related to executive compensation

Executive compensation is underpinned by several theories of which the most prominent are discussed in this section. According to the agency theory, executives are likely to concentrate on their own wealth creation, rather than maximise shareholders’ wealth (Jensen & Meckling, 1976). Various corporate governance codes and guidelines have been introduced over time to better align the interests of principals and agents.

Executive compensation could be an instrument to address the agency problem, but it could also be regarded as part of the problem. According to the optimal contracting theory, incentives could be used to encourage executives to maximise shareholders’ wealth. The managerial power approach postulates that executives are in a position to partly influence their pay. As such, they could benefit from exorbitant packages which are not necessarily linked to their performance (Bebchuk & Fried, 2003; Edmans & Gabaix, 2009).

In contrast to the agency theory, the stakeholder theory suggests that the competing interests of stakeholders such as employees, customers and local communities should get attention (Freeman, 1994). If executives receive excessive compensation, the interests of a wide range of stakeholders could be adversely affected.

While the agency theory is based on the premise that executives are primarily driven by self-interest, the stewardship theory suggests that they are intrinsically motivated to accomplish the tasks and responsibilities with which they have been entrusted. Trustworthy managers, who are devoted to shareholders and prominent stakeholders, would thus be more motivated by being a good steward than by personal monetary gain. In practice, only some managers are intrinsically motivated (Donaldson & Davis, 1991; Kiel & Nicholson, 2003).

Another often-cited theory in the executive compensation literature is the resource dependency theory. Pfeffer and Salancik (1978) emphasised that all companies rely on limited resources to ensure their survival. Executives’ ability to acquire, maintain and optimise these resources is thus central to success. Focus is therefore placed on the composition, performance and ability of a board to optimally allocate resources (Klettner, 2017). According to this theory, board performance evaluations are essential and could lead to the removal of poorly performing executives. Attractive compensation packages are often used to attract and retain highly skilled directors (Klettner, 2017). Above-average emolument might inspire high performance given the threat of dismissal (Lee, 2016). The social comparison theory also suggests that pay-benchmarking practices could result in above-average compensation (O’Reilly, Main, & Crystal, 1988).

Based on the tournament theory, executive pay could be regarded as a prize for reaching the highest corporate echelons (Lazear & Rosen, 1981). This so-called ‘prize’ does not incentivise current executives, but rather lower level managers who are still climbing the corporate ladder. This school of thought suggests that limited attention is given to the performance of executives, post-appointment, unless they compete to become the CEO. Executives might therefore strive to seek tenure to maximise their emolument, rather than their performance (Lee, 2016).

Evaluating executives’ performance

Performance evaluations can take different forms, including formal, self- and peer-assessments (Stybel & Peabody, 2005). Regular evaluations of the board and individual directors (especially CEOs and chief financial officers) are important, as most investors review the quality of a directorate before committing funds (Conger, Finegold, & Lawler, 1998). Research also reveals that effective performance evaluations can enhance a board’s overall functioning and the company’s subsequent financial performance (Minichilli, Gabrielsson, & Huse, 2007). Performance evaluations can furthermore highlight areas of concern and critical training and development needs at board level (Deloitte, 2014; IoDSA, 2009).

Companies can report on the following aspects of executives’ performance evaluations: criteria and time frames used, the frequency of evaluations, measurement tools used and the persons/committees conducting these reviews. In most instances, executive performance is evaluated against financial metrics such as return on assets (ROA), return on equity (ROE), earnings per share (EPS), total share return (TSR), economic value added (EVA) and market value added (MVA) (Bradley, 2013; Kuboya, 2014; Scholtz & Smit, 2012; Zakaria, 2012). Despite being widely used, all these criteria have shortcomings. For example, a company’s EPS should be viewed with caution as executives can manipulate reported profits (Bussin & Modau, 2015). Not only does TSR fail to adjust for risk, but it could also promote an excessive focus on short-term results (Li & Young, 2016). While EVA is a measure of historic performance, MVA reflects market expectations about a company’s prospects. The latter, however, ignores the opportunity cost of capital (Pietge, 2007).

In addition to these financial benchmarks, individual directors can be evaluated on their knowledge of the company, their effective fulfilment of board tasks and their preparation, attendance and participation during meetings (Van den Berghe & Levrau, 2004). When evaluating the board in its entirety, attention can be given to meeting frequency (Ntim & Oseit, 2011), collective knowledge and experience, board composition and succession planning (Minichilli et al., 2007; Preston & Brown, 2004).

A new category of performance benchmarks has emerged in recent years centring on environmental, social and governance (ESG) considerations. ESG performance can be assessed by, inter alia, examining risk audits and the number of fines and complaints received from employees, customers and/or suppliers in a particular year (Epstein & Roy, 2004). This category of performance criteria is expected to become more important in future as investors are increasingly incorporating ESG considerations into their investment analyses and ownership practices (Van Duuren, Plantinga, & Scholtens, 2016).

Performance criteria should be regularly assessed to assure that they remain appropriate (IoDSA, 2009). The same applies to the time frames over which executives’ performance is assessed and rewarded. Whereas ‘short-term’ is generally understood to be a period of one year, ‘long-term’ can range from three to five years or from five to ten years (Chiu, 2015).

Most reporting guidelines urge companies to publicly disclose whether or not they have conducted performance evaluations in the year under review. Evaluations can be conducted by the remuneration and/or nomination committee, the CEO, chairperson, the directors themselves, industry peers and/or independent experts (Conger et al., 1998; Deloitte, 2014).

Linking executives’ pay to performance

Numerous scholars have investigated the relationship between executive remuneration and company performance over several decades. International researchers primarily measured performance in terms of financial metrics and reported a positive relationship between these variables (e.g. Abdullah, 2006; Conyon, Peck, & Sadler, 2000; Duffhues & Kabir, 2008; Fatemi, Desai, & Katz, 2003; Kato, Kim, & Lee, 2007; Merhebi, Pattenden, Swan, & Zhou, 2006).

As indicated earlier, very few studies have focused on the reporting of executive remuneration. In the UK, Dulewicz and Herbert (2008) noted some ambivalence in disclosures on board evaluations among FTSE 350 companies. Hooghiemstra et al. (2017) investigated the impact of this ambivalence and discovered that less readable reports resulted in fewer ‘say-on-pay’ votes. The influence of obscure reporting, however, diminished as institutional ownership increased and even resulted in increased voting dissent in cases where the majority of the company’s shares were held by institutional investors. Concerns about the quality of reporting on board performance evaluations were also noted in the US. Roy (2008) found that the majority of the Fortune 500 companies that she analysed disclosed generic information. She argued that such reporting gave shareholders little reassurance that proper processes were in place to promote their long-term interests.

For the purpose of this research, focus was placed on studies conducted locally of which the most prominent are presented in Table 1.

Table 1. Selected South African studies on the relationship between executive remuneration and performance.

Author(s) and year of publication	Research period	Measure(s) of executive pay	Measure(s) of performance	Empirical findings
Bradley (2013)	2006−2010	CEO salaries, bonuses and other compensation	ROA, ROE and EPS	No correlation between CEO compensation and any of the performance measures.
Shaw (2011)	2005−2010	Total remuneration (fixed pay and short-term incentives)	Profit after tax, earnings before interest, tax, depreciation and amortisation (EBITDA), ROE, EPS and total assets	Positive correlation between fixed pay and short-term incentives and total assets, EBITDA and profit after tax for financial services companies.
De Wet (2012)	2006−2010	Total remuneration	ROA, ROE, EVA and MVA	Positive associations with all the performance measures and total remuneration.
Scholtz & Smit (2012)	2003−2010	Short-term cash compensation	Total assets, turnover, EBITDA and volume weighted average share price	Positive link between cash compensation and total assets, turnover and volume weighted average share price for companies listed on the JSE’s Alternative Exchange.
Van Blerck (2012)	2002−2011	Total remuneration (excluding long-term incentives)	ROE, EVA and share price at year-end	Positive correlation between EVA and remuneration of selected executives of South African banks.
Bussin & Blair (2015)	2008−2012	Long-term CEO incentives	Capital employed, change in fixed assets, change in turnover, dividend per share, EBITDA, EPS, profits after interest and tax, change in share price and TSR	Positive associations noted at industry level between long-term CEO incentives and EBITDA, fixed assets, profit after tax and interest, capital employed, changes in share price and TSR.
Bussin & Nel (2015)	2006−2011	Guaranteed CEO cost-to-company (excluding long-term incentives)	Du Pont ROE	Negative relationship with ROE and the guaranteed CEO cost-to-company.
Bussin & Modau (2015)	2006−2012	Total CEO remuneration (excluding long-term incentives)	EPS, ROE, EVA and MVA	Inverse correlation between fixed pay and short-term incentives and both ROE and MVA. Direct and indirect correlations noted between fixed pay and EVA in several years. Direct correlation between EPS and short-term incentives.
Deysel & Kruger (2015)	2006−2012	Total CEO compensation	Share price, ROE, EBITDA, headline EPS	Positive correlation between CEO remuneration and the considered accounting and market-based performance measures.
Crafford (2015)	2002−2010	Total remuneration	Turnover, EPS, MVA and TSR	Positive links between total remuneration and turnover and MVA. Negative correlation between total remuneration and TSR.

Source: Authors’ own compilation

Although limited studies have been conducted in South Africa on the relationship between selected measures of executive pay and performance (Lee, 2016), the majority of authors reported a positive correlation. This observation is in line with international findings. In addition to the pay-performance studies mentioned in Table 1, Ulrich (2010) examined compliance with executive emolument reporting (as set out in King II, the JSE’s listing requirements and legislation); the section in the annual report where this information was presented; disclosure of the company’s remuneration philosophy; and criteria used to evaluate the performance of the board and/or individual directors. Although most of the considered companies provided details on the absolute level of executive emolument in 2007 (the year under analysis in the Ulrich 2010 study), very few disclosed particulars on the targets that were used to allocate performance incentives. Similar findings were also reported in a more recent PwC (2015) study of the same nature.

By analysing the annual/integrated reports of JSE-listed companies, other researchers also noted a growing misalignment between executives’ pay and performance (Bussin & Modau, 2015; Bussin & Nel, 2015). Bussin (2015) observed that local remuneration committees were placing more emphasis on fixed pay rather than performance-based incentives, thereby widening the disconnect between executive emolument and performance.

Although the reporting requirements of King II and King III have resulted in greater transparency in pay reporting, they might have contributed to the wider adoption of pay benchmarking. Research shows that remuneration committees of local companies are increasingly aligning their executives’ emolument with the median or upper quartile of their peer group (Bussin, 2017; PwC, 2011). Benchmarking has inflated executive pay beyond what could be regarded as ‘reasonable’ relative to company performance (Armstrong, Ittner, & Larcker, 2012; Cadman & Carter, 2014). The upward spiral in executive remuneration is clearly illustrated in the following quote by the chairperson of a local remuneration committee that approved a substantial increase in their CEO’s emolument (Barron, 2014, p. 1):

No matter how immoral executive remuneration may seem, it has to be competitive. We simply can’t afford to be bold in terms of morals, because we would then run a real risk of losing some of our talented executives who have specialised knowledge and skills and who are also very mobile.

In the light of shareholders’ calls for improved reporting to make more informed investment decisions, the following alternative hypotheses were postulated:

Table

H1:	There was a change in disclosure on performance evaluations conducted by selected JSE-listed companies over the research period.
H2:	There was a change in the acceptability of disclosures on performance evaluations conducted by selected JSE-listed companies over the research period.
H3:	There was a change in the depth of reporting on performance evaluations conducted by selected JSE-listed companies over the research period.
H4:	There was a change in disclosure on the executive pay-performance link by selected JSE-listed companies over the research period.

Methodology

The following sections present details on the mixed-methods approach adopted in this study. The quantitative dataset consisted of disclosures, acceptability and depth scores related to executive performance evaluations and the pay-performance link of selected JSE-listed companies over a fourteen-year period. The qualitative data were sourced by conducting interviews with six remuneration committee members serving JSE-listed companies.

Quantitative data collection and analysis

The population consisted of all companies that were listed in the considered industries (industrials, consumer services, consumer goods, technology, healthcare and telecommunications) from 2002 to 2015 (see Table 2).

Table 2. Population and sample.

Year	Population^(a)	Sample
		Listed	Delisted^(b)	Total
2002	220	163	22	185
2003	200	175	11	186
2004	177	149	14	163
2005	169	150	8	158
2006	179	139	6	145
2007	212	133	6	139
2008	209	141	8	149
2009	204	144	6	150
2010	212	137	3	140
2011	208	137	7	144
2012	196	130	12	142
2013	188	143	9	152
2014	178	140	4	144
2015	170	139	0	139

(a) Data were sourced from the Bloomberg database as at 31 December of each year. The population included all companies that were listed on the JSE in the industrials, consumer services, consumer goods, technology, healthcare and telecommunications industries.

(b) A company was classified as delisted if it delisted from the JSE at any point during a particular year or in the last year that an annual/integrated report was available. These companies were included in the analysis to reduce survivorship bias. Companies that were suspended from the JSE at any point during a particular year were also classified as delisted.

Companies listed in the basic materials, financial services and oil and gas industries were excluded from the sample as data were initially collected as part of a broader study focusing on corporate governance and financial performance. As financial reporting in these industries differs from the considered industries, they were omitted. The exclusion of these three industries represents a limitation of the study, given their size and impact on the South African economy.

A total of 254 unique companies and 2 136 company-year observations were analysed. Attention was given to name changes as a result of merger and acquisition activities (e.g. Caxton Ltd. becoming Caxton & CTP Publishers & Printers Ltd.) and to companies that delisted and relisted at a later stage (e.g. Adcock Ingram Holdings Ltd.). In many of these cases, ticker codes changed although the company continued with the same core operations. Changes to financial year ends were also noted. As indicated in Table 3, most of the considered companies operated in the industrials and consumer services industries.

Table 3. Considered industries.

Industry^(a)	% of sample
Industrials	39.65
Consumer services	25.98
Consumer goods	17.37
Technology	10.96
Healthcare	3.79
Telecommunications	2.25
Total	100.00

(a) During the research period, the JSE used two different systems to classify companies in terms of their business activities. From 1 June 2002 to 31 December 2005 the Exchange used the FTSE Global Classification. On 1 January 2006 they switched to the Industry Classification Benchmark. Given that there are considerable differences between the two classification systems, a number of the considered companies that were classified in one industry before 1 January 2006 had to be reclassified thereafter. The sample consisted of 260 unique companies.

Market capitalisation (in millions of rand) was used as a measure of company size. The data were downloaded from the IRESS database and reflect a company’s market capitalisation (number of shares × closing share price) on 31 December of a particular year. In instances where companies delisted or were suspended, the market capitalisation on the final day of being listed on the JSE was used. The log of market capitalisation was computed as this variable differed significantly across the sample.

It could be argued that companies with a primary listing on a stock exchange other than the JSE, such as London or New York, might pay their executives more than companies with a primary listing on the JSE. These companies tend to be larger and thus more complex to manage. On average, about seven per cent of the sample had primary listings in other countries. Due to the limited number of observations, further statistical tests could not be conducted for this variable.

The companies’ annual/integrated reports were downloaded from the IRESS database. Content analysis was used to assign disclosure, acceptability and depth scores relating to executive performance evaluations. The following keywords were used to identify relevant sections in the companies’ annual/integrated reports: board(s), director(s), directorate, committee(s), committee member(s), perform, performance, review(s), assessment(s), effectiveness, evaluate(s), evaluation(s), appraise and appraisal.

If a company disclosed that a performance evaluation took place during a particular year – for the directorate as a whole or for individual directors – a disclosure score of one (1) was allocated. The same applied to cases where companies acknowledged the importance of these assessments. A disclosure score of zero (0) was assigned if no mention was made of such evaluations. Where a company had a disclosure score of 1, an acceptability score of 1 or 0 was allocated, reflecting whether or not the reported performance evaluation information was in line with the King guidelines in a particular year (see Table 4). An acceptability score of 0 was assigned where companies reported that they were considering conducting performance evaluations in future or where it was unclear if the reported performance review(s) took place during the year under consideration. If the disclosure score was 0, the acceptability score was, by implication, also 0.

Table 4. King II and King III recommendations on performance evaluations.

King II (applicable 2002−2009)

King III (applicable 2010−2015)

• The board, directors and board committees should conduct regular (at least annual) performance evaluations. • Self-evaluations should be reviewed by the nomination (or similar) committee. • The board should appraise the performance of the chairperson. • The chairperson (or independent deputy chairperson or appropriate committee) should assess the performance of the CEO.

Annual performance evaluations of the board, directors and board committees should be performed by the chairpersons (with the assistance of the company secretary) and/or independent service provider. The nomination committee should review the evaluation process and outcomes. The board should appraise the performance of the CEO against criteria developed based on his/her role and responsibilities. An overview of the appraisal process, results and action plans to address weaknesses should be disclosed in the company’s integrated report. A director should only be nominated for re-appointment after his/her performance and meeting attendance has been evaluated.

Source: Authors’ own compilation based on IoDSA (2002) and IoDSA (2009)

For those companies that had acceptability scores of 1, a subjective depth score of 1, 3 or 5 was assigned to indicate the level of detail divulged. Where companies only mentioned performance evaluation(s) in passing, a depth score of 1 was allocated. Where certain particulars were provided (for example on the performance benchmarks used or the process followed), a depth score of 3 was assigned. Where an extensive explanation was provided, a depth score of 5 was allocated.

To investigate reporting on the executive pay-performance link, an additional set of keywords were used, namely: pay, remuneration, compensation, emolument, incentive(s), incentivised, reward(s), rewarded, share option(s), bonus(es), link(ed), acknowledge(d), encourage(d), recognise(d) and recognition. A score of 1 was allocated if there was evidence in a company’s annual/integrated report that executives’ pay was linked to performance. Where no indication of this practice could be found, a score of 0 was assigned.

A large number of cross-checks was conducted, given that a certain degree of subjectivity was involved in the coding process and that the process occurred over an 18-month period.

Descriptive statistics were used to analyse trends in the disclosure, acceptability and depth scores pertaining to performance evaluations and disclosure on the link between executive pay and performance. Generalised estimating equations (GEEs) were computed to test H₁, H₂ and H₄, given the binary nature of the panel data. A mixed-model analysis of variance (ANOVA) was used to test H₃. The considered fixed effects factors were ‘Year’, ‘Size’, ‘Listing status’, and ‘Year-listing status interaction’. The random effects factor was ‘Company’. The GEE and mixed-model ANOVA methods take the dependence of responses over time into consideration. Significant findings over the entire research period prompted the researchers to employ the Fisher’s least significant difference test to determine the significance of annual differences.

Qualitative data collection and analysis

Six semi-structured personal interviews were conducted with non-executive directors who were serving on or had served on remuneration committees of companies operating in the local financial services industry. This industry was selected as it did not form part of the quantitative analysis. Furthermore, executives in this industry are often criticised for accepting ‘considerable bonuses whilst profits are dwindling’ (PwC, 2015).

Convenience and snowball sampling techniques were used to identify eligible participants. The least experienced director served on a remuneration committee for one year, whereas the most experienced director held this position for seven years. All the interviewees were white males who had at least one tertiary qualification and served on multiple boards and board committees. A considerable gender bias was not foreseen as the vast majority of directors in South Africa are white males (Mans-Kemp & Viviers, 2015).

Biographic information was requested in Section A of the interview guide, followed by questions related to executive performance evaluations and remuneration practices (Sections B and C, respectively). Data saturation occurred after six interviews, i.e. no new information or insights were forthcoming in the last interview. All the interviews were conducted in July 2016. Thematic analysis was employed to analyse the qualitative data (Clarke & Braun, 2013). Credibility was assured by selecting experienced non-executive directors.

Findings and discussion

The extent of corporate reporting on director performance evaluations

After controlling for size, performance evaluation disclosure and acceptability, scores changed significantly over the research period (Table 5).

Table 5. GEE results: Performance evaluation disclosure and acceptability scores over the research period.

	Degrees of freedom^(a)	Disclosure		Acceptability
		Wald statistic	p-value	Wald statistic	p-value
Year	12	109.800**	0.000	124.114**	0.000
Size	1	78.107**	0.000	77.446**	0.000
Listing status	1	0.402	0.526	1.399	0.237
Year-listing status interaction	12	8.719	0.727	8.766	0.723

* Significant at the 5% level; ** Significant at the 1% level

(a) The limited number of delisted companies during the last part of the study period reduced the degrees of freedom.

Based on the results reported in Table 5, H₁ and H₂ were supported. Not only did more of the sampled companies report on performance evaluations towards the end of the research period, but the acceptability of these disclosures in terms of the King recommendations also increased notably (see Table 6). Several significant annual differences were noted for the disclosure and acceptability scores as highlighted by the letters in Table 6 (columns 5 and 8).

Table 6. Details on performance evaluation disclosure and acceptability scores.

Year	Sample	Disclosure			Acceptability
		N	%	LS mean^(a)(b)	N	%	LS mean^(a)(b)
2002	185	41	22.16	0.340^f	33	17.84	0.240^f
2003	186	82	44.09	0.617^d,e	68	36.56	0.486^e
2004	163	87	53.37	0.665^d,e	80	49.08	0.581^c,d
2005	158	86	54.43	0.625^e	81	51.27	0.550^d,e
2006	145	85	58.62	0.618^e	80	55.17	0.545^d,e
2007	139	89	64.03	0.665^e	82	58.99	0.580^d,e
2008	149	103	69.13	0.709^d	97	65.10	0.638^c
2009	150	117	78.00	0.820^c	110	73.33	0.749^b
2010	140	114	81.43	0.845^b,c	111	79.29	0.803^a
2011	144	126	87.50	0.880^a,b,c	121	84.03	0.825^a
2012	142	129	90.85	0.904^a,b	123	86.62	0.841^a
2013	152	138	90.79	0.918^a	130	85.53	0.846^a
2014	144	133	92.36	0.924^a	128	88.89	0.871^a
2015	139	128	92.09	0.919^a	124	89.21	0.874^a

(a) Least square means ranged between 0 and 1. The LS mean score of each year was compared to the LS mean score of every other year over the research period for disclosure and acceptability respectively. Letters were assigned to show significant differences between years. If the same letter is present for two or more years, it implies that there was no significant difference between the year in question and other years with the same letter. For example: The letter ‘a’ in column 5 shows that the LS mean disclosure scores did not differ significantly in any years between 2011 and 2015. The same interpretation is applicable to the LS mean acceptability scores column. The letter ‘f’ in column 8 shows that the LS mean scores differed significantly between 2002 and all other years.

(b) Size was used as a control variable.

Perusal of the disclosure scores in Table 6 (column 4) reveals that less than half of the considered companies disclosed whether or not the performance of their boards and/or individual directors was appraised in 2002 and 2003. The initial improvements might be attributed to the introduction of King II in 2002. Significantly more companies also reported on their performance evaluations in anticipation of and in response to the release of the King III Report in 2009. The disclosure and acceptability scores started stagnating after 2011. As indicated in Table 7, depth scores changed significantly over the research period (taking size into account). As such, H₃ was supported.

Table 7. Mixed model ANOVA results: Depth scores over the research period.

	Numerator degrees of freedom^(a)	Denominator degrees of freedom	F	p-value
Year	12	1014	5.150**	0.000
Size	1	1014	23.841**	0.000
Listing status	1	188	0.153	0.696
Year-listing status interaction	12	1014	0.709	0.743

* Significant at the 5% level; ** Significant at the 1% level

(a) The limited number of delisted companies during the last part of the study period reduced the degrees of freedom.

Significant differences in the depth scores were noted for several years (Table 8).

Table 8. Details on depth scores.

Year	Sample^(a)	1 (mentioned briefly)		3 (provided some detail)		5 (provided extensive explanation)
		N	%	N	%	N	%	LS mean^(b)(c)
2002	33	23	69.70	10	30.30	0	0.00	1.316^g
2003	68	45	66.18	22	32.35	1	1.47	1.661^f,g
2004	80	54	67.50	24	30.00	2	2.50	1.606^f,g
2005	81	49	60.49	29	35.80	3	3.70	1.735^e,f
2006	80	43	53.75	26	32.50	11	13.75	1.998^d,e
2007	82	40	48.78	29	35.37	13	15.85	2.117^c,d
2008	97	48	49.48	31	31.96	18	18.56	2.236^b,c,d
2009	110	54	49.09	30	27.27	26	23.64	2.395^a,b,c
2010	111	51	45.95	32	28.83	28	25.23	2.466^a,b
2011	121	53	43.80	49	40.50	19	15.70	2.346^b,c
2012	123	46	37.40	60	48.78	17	13.82	2.437^a,b
2013	130	58	44.62	52	40.00	20	15.38	2.302^b,c
2014	128	58	45.31	50	39.06	20	15.63	2.323^b,c
2015	124	42	33.87	58	46.77	24	19.35	2.644^a

(a) Depth scores were only assigned to companies that had an acceptability score of 1.

(b) Least square means ranged between 1 and 5. Letters were assigned to show significant differences between years. The same interpretation as explained previously applies.

(c) Size was used as a control variable.

Over the period 2002 to 2004, several of the considered companies acknowledged the importance of performance evaluations and showed an interest in conducting such reviews in future. However, many of these companies simply repeated this intention year after year. The remainder of the companies merely mentioned that performance appraisals took place without providing details. Some companies only evaluated their CEO’s performance, whereas others evaluated the board in its entirety, and then only on a bi-annual basis. An example of a basic disclosure (depth score of 1) was observed in the 2004 annual report of Allied Electronics Corporation. They stated that ‘the board has adopted a system whereby it evaluates its own effectiveness’ and ‘the review takes place on an annual basis’.

As pointed out by Stybel and Peabody (2005), companies can use various types of performance assessments. However, only after 2010 did more of the considered companies begin to publish particulars of their performance evaluations, including the types of assessments used (earning them depth scores of 3). During the 2000s, some companies mentioned whether the reviews were conducted in-house or by external service providers. For example, in their 2006 annual report, AVI stated that a combined nomination and remuneration committee appraised the chairperson and the CEO’s performance. Their board assessed its performance against objectives and used self-assessments to identify individual director’s development needs. The importance of recognising areas for improvement was also highlighted by Deloitte (2014).

Over time, more companies published details on the measurement tools they used and the challenges they identified in the appraisal process (resulting in depth scores of 5). They also explained how these challenges would be addressed in future. Woolworths Holdings was an example of a company that scored a 5 for depth in various years. For example, they reported in 2015 that informal and formal evaluations were conducted during the year. Informal appraisals were applicable to all directors. These evaluations included, amongst others, progress on key strategic objectives and critical capabilities required at board level. In addition to these informal assessments, the chairperson also conducted formal performance appraisals of all directors who made themselves available for re-election. The chairperson’s views were shared with the board and the nominations committee. Similarly, the lead independent director, in consultation with the nomination committee, evaluated the chairperson’s performance.

As illustrated in Table 8, less than one-fifth of the sampled companies obtained depth scores of 5 during the latter part of the research period (2011 to 2015). This finding seems to be in line with claims by shareholder activists that JSE-listed companies are ignoring their requests for greater transparency (Viviers, 2015).

Frequency, criteria and time frames used to evaluate executives’ performance

In line with performance-related King III recommendations (IoDSA, 2009), all interviewees agreed on the importance of conducting regular performance evaluations and mentioned that more of these reviews had taken place in their industry in recent years. Remuneration committees in the financial services industry mostly used financial targets, notably earnings before interest, tax, depreciation and amortisation (EBITDA), headline EPS, ROA, ROE and TSR. These performance measures were used by previous South African authors (refer to Table 1).

A few interviewees questioned the use of TSR. One director argued that ‘total share return does not always reflect the performance of the [entire] executive team’. Another suggested that ‘market-based performance is only an indication of what the market perceives as performance’. They argued that political events, such as ‘Nene-gate’ had a significant adverse effect on share prices in 2015. One participant remarked that ‘we [directors] should not have to be evaluated on factors beyond our control, political events being a case in point’.

This claim also featured in a study conducted by PwC (2015). The company highlighted the negative impact that economic conditions had on executives whose compensation was mostly performance-based. Another interviewee said: ‘We do not use predetermined benchmarks to determine short-term incentives, as these benchmarks can change due to shifting market conditions’.

Some participants pointed out that different sectors in the financial services industry (such as banking, insurance and asset management) had different value drivers and therefore used different financial metrics to evaluate executive performance. A number of director and board-specific performance criteria were also used. One participant remarked that ‘individual performance targets play an important role’ and argued that these performance targets are ‘often evaluated on a qualitative, subjective basis’ at the companies where he served as a director. Another director reiterated the importance of individual performance measures as ‘viable alternatives to company performance when evaluating individual directors’. He continued by saying ‘measuring executive director performance on an individual basis can be useful in determining what skills and competencies the said director brings to the board and board committees. It can also serve as a measure of effort exerted’.

Despite the growing use of ESG considerations by investors (Van Duuren et al., 2016), none of the remuneration committees on which interviewees served employed ESG benchmarks. Similar observations were made while conducting the content analysis.

One of the participants was of the opinion that ‘long-term incentives did not require performance criteria as employment was the only prerequisite’. This misalignment between executive performance and remuneration was also noted by Bussin and Modau (2015) and Bussin and Nel (2015). This view is representative of the argument set out in the tournament theory (Lazear & Rosen, 1981). By definition, long-term incentives aim to motivate employees to excel beyond the minimum expectations. If they receive incentives irrespective of their performance, these incentives do not serve their purpose. In line with the agency theory, all the interviewees supported the notion that executive remuneration was the best way to align managers’ interests with those of shareholders.

Reporting on the executive pay-performance link

No significant change was observed in the number of considered companies that reported on the link between executive pay and performance over the research period after controlling for size (Table 9). As a result, H₄ could not be supported.

Table 9. GEE results: Reporting on the executive pay-performance link over the research period.

	Degrees of freedom^(a)	Wald statistic	p-value
Year	11	13.691	0.251
Size	1	50.979**	0.000
Listing status	1	1.456	0.228
Year-listing status interaction	11	13.138	0.284

* Significant at the 5% level; ** Significant at the 1% level

(a) The limited number of delisted companies during the last part of the study period reduced the degrees of freedom.

While Table 10 (column 4) shows that 63.78 per cent of the sampled companies linked their executives’ pay to performance in 2002, only 22.16 per cent of those companies evaluated their directors’ performance in the same year (refer to Table 6 column 4). This discrepancy raises questions regarding the basis on which the remaining 42 per cent of companies rewarded their executives in 2002. This discrepancy was no longer detected by 2015.

Table 10. Disclosure on the link between executives’ pay and performance.

Year	Sample	N	%	LS mean^(a)(b)
2002	185	118	63.78	0.792
2003	186	128	68.82	0.821
2004	163	119	73.01	0.828
2005	158	122	77.22	0.837
2006	145	115	79.31	0.831
2007	139	107	76.98	0.808
2008	149	120	80.54	0.834
2009	150	126	84.00	0.857
2010	140	121	86.43	0.877
2011	144	129	89.58	0.905
2012	142	127	89.44	0.887
2013	152	132	86.84	0.876
2014	144	129	89.58	0.888
2015	139	126	90.65	0.889

(a) Least square means ranged between 0 and 1. No significant differences were noted between years, given the insignificant ‘Year’ Wald statistic for the overall period (refer to Table 9).

(b) Size was used as a control variable.

A notable increase in the number of companies linking their executives’ pay to some measure of performance is encouraging as most prior studies in South Africa (see Table 1) established a positive association between this practice and financial performance. Whereas most companies mentioned the types of performance incentives offered (such as cash bonuses and share options), few clarified the performance benchmarks they used, specifically pertaining to individual directors. Where details were given, it was almost exclusively financial in nature. Several of the sampled companies allocated bonuses without providing any justification.

Many of the interviewed directors stated that the base salaries of executives in the financial services industry were set according to the ‘market mean’. As indicated earlier, benchmarking against peers is common practice in South Africa (Bussin, 2017; PwC, 2011). All interviewees viewed performance-based incentives as critical to attract and retain high-calibre executives in the financial services industry. They mentioned that the remuneration committees on which they served were increasingly tying executives’ pay to performance. One director remarked ‘the use of equity is a particularly effective way of ensuring that executives in our industry [banking sector to be more specific] do not take unnecessary risks when it comes to managing the company’. This finding seems to contradict Bussin (2015) who claimed that local companies are moving away from performance-based incentives.

The participants remarked that short-term cash bonuses were extensively used in their industry to align executives’ interests with those of shareholders. Some remuneration committees based these bonuses on financial and director-specific targets, whilst others did so irrespective of performance. One interviewee noted, ‘Although it is key to understand what drives a company’s bottom line, the absence of pre-determined benchmarks [for short-term incentives] is not uncommon’. Directors indicated that share options are often used as long-term incentives in this industry. Although ‘long-term’ could refer to a period of three to five years or five to ten years (Chiu, 2015), most of these share options only had two-year vesting periods. Several directors justified this practice by claiming that executives ‘cannot see the carrot if the evaluation period is too long’. Another participant argued that shorter rather than longer assessment periods were preferred as executives’ tenure in their industry was relatively short. Lee (2016) warned against this practice.

Conclusions and recommendations

As South Africa has one of the most unequal societies in the world, it comes as no surprise that executives are often criticised for receiving what could be considered ‘exorbitant’ pay packages. As elsewhere in the world, local stakeholders are requesting greater transparency on how executives are rewarded. The majority of previous studies conducted in the local context reported on the link between JSE-listed companies’ financial performance and their executive directors’ remuneration. Limited knowledge, however, exists on the nature and depth of executives’ performance evaluations and the link between their performance and pay.

Although the evidence in this study shows that the vast majority of the sampled JSE-listed companies evaluated the performance of their boards and/or individual directors in 2015, the depth of these disclosures was superficial. As a result, shareholders and other key stakeholders are constrained in their ability to hold remuneration committees accountable. The narrow focus on short-term performance is also disconcerting as previous researchers warned that it could fuel short-termism and manipulative behaviour among executives (Bolton et al., 2006; Kirkpatrick, 2009). The same argument applies to the over-reliance on financial performance metrics such as EPS. The managerial power approach postulates that executives could partly influence their personal pay packages. Executive directors often serve on remuneration committees where emolument packages are discussed, and decisions are made regarding the pay-performance link. Due to possible earnings management, they might benefit financially if their pay is linked to reported EPS. Such a gain might not be linked to individual directors’ performance.

Based on the resource dependency theory, directors who can optimise their companies’ limited resources are central to their success. Local remuneration committees are accordingly encouraged to evaluate and reward the performance of individual directors, as opposed to the board in its entirety. Individual evaluations will furthermore enable them to identify and remove poorly performing individuals and highlight specific development needs.

Remuneration committees are also urged to critically assess whether the performance criteria and time frames which they are using have the desired effect. In line with Lee’s (2016) warning against the consequences of the tournament theory, remuneration committees are encouraged to avoid formulating ‘long-term’ criteria which are based on an executive’s expected tenure rather than true long-term measures. Executive pay might otherwise not serve its intended purpose to reward directors for their contribution to their companies’ success. Performance-based incentives should reward only above-average performance and, where justified, punish poor performance. Pay benchmarking should not be blindly embraced, since it could result in above-average remuneration that is not performance-linked. Remuneration committees could consider smoothing the progression of pay through the corporate ranks, rather than paying top-level executives exorbitant amounts (Lee, 2016).

By focusing exclusively on financial performance, remuneration committees ignore the impact that ESG risks could have on a company’s survival. A wider adoption of ESG metrics, along with financial targets, is hence suggested. A more balanced approach will bode well as shareholders are increasingly incorporating non-financial considerations into their investment analyses and ownership practices. Using a broader spectrum of criteria will most likely require additional director development (Deloitte, 2014).

Commerce educators should cultivate an appreciation for long-term, sustainable value creation among future generations of shareholders and executives (Giacalone & Thompson, 2006; Leveson & Joiner, 2014; Rezaee & Homayoun, 2014). They need to sensitise students on the importance of social justice, transparency and accountability. Future researchers could include the three industries that were omitted from the quantitative analysis in a follow-up study. Research could also be undertaken on the impact that shareholder activists could have on the size and composition of executive remuneration packages. The use and effectiveness of long-term incentives and value-based performance metrics could furthermore be explored.

This study shows that many of the considered JSE-listed companies are not reporting sufficient information on how their executives’ performance is evaluated and rewarded. By disclosing more particulars on performance benchmarks and time frames, future shareholder engagements are likely to be more constructive.

Acknowledgements

The authors would like to thank Mr Bryce Harding, Mr Niel de Kock and Ms Blanche-Mari Staal for their assistance in collecting some of the data. We are also grateful for the valuable inputs of Prof. Martin Kidd, Ms Ruth Albertyn, Prof. Pierre Erasmus and Mr Dewald Smit. A word of thanks is furthermore extended to Ms Michele Boshoff for her assistance with the language editing.